Animals with Novel Genes

This book is about transgenic animals — animals into which new genes have been artificially introduced. It contains chapters by leading authorities on the present state of the art regarding the application of transgenic technology to the great variety of animal types - ranging from protozoan cells, through nematode worms and fruit flies, to many higher vertebrates - that have been used in this experimental way. The impact of transgenic animals on the development of agriculture and medicine is likely to be very great; at the same time they provide an unrivalled experimental system for the study of gene regulation, genetic aspects of disease and gene therapy. One of the objectives of the book, therefore, is to place transgenic animals in the context of their present and future contributions to science, medicine and agriculture. There is also some discussion of the ethical implications of this work. This up-to-date, comprehensive and authoritative book is ideal for advanced undergraduates and graduate students wishing to familiarize themselves with the field of transgenic animals.

ANIMALS WITH NOVEL GENES

ANIMALS WITH NOVEL GENES Edited by

NORMAN MACLEAN University of Southampton

CAMBRIDGE UNIVERSITY PRESS

CAMBRIDGE UNIVERSITY PRESS Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, Sao Paulo Cambridge University Press The Edinburgh Building, Cambridge CB2 2RU, UK Published in the United States of America by Cambridge University Press, New York www.cambridge.org Information on this title: www.cambridge.org/9780521432566 © Cambridge University Press 1994 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 1994 This digitally printed first paperback version 2006 A catalogue recordfor this publication is available from the British Library Library of Congress Cataloguing in Publication data Animals with novel genes / edited by Norman Maclean, p. cm. Includes index. ISBN 0-521-43256-1 1. Transgenic animals. I. Maclean, Norman, 1932-. QH442.6.A54 1994 591.1'5'0724-dc20 94-11728 CIP ISBN-13 978-0-521-43256-6 hardback ISBN-10 0-521-43256-1 hardback ISBN-13 978-0-521-02472-3 paperback ISBN-10 0-521-02472-2 paperback

Contents

List of contributors Preface

page vii ix

1

Transgenic animals in perspective

Norman Maclean

2

Transgenic insects Julian M. Crampton and Paul Eggleston

21

3

Transgenic fish Norman Maclean and Azizur Rahman

63

4

Transgenic birds K. Simkiss

5

Transgenic rodents

106

Martin J. Evans, Darren T. Gilmour

and William H. Colledge 6

Large transgenic mammals

7

Minor transgenic systems Index

1

138 G. Brem and M. Mtiller Norman Maclean

179 245 255

Contributors

G. Brem Department of Molecular Animal Breeding, LMU Munich, Veterinarstr. 13, 80539 Munich, Germany William H. Colledge Wellcome I CRC Institute, Tennis Court Road, Cambridge CB2 1QR, England Julian M. Crampton Wolf son Unit of Molecular Genetics, Liverpool School of Tropical Medicine, Pembroke Place, Liverpool L3 5QA, England Paul Eggleston Wolf son Unit of Molecular Genetics, Liverpool School of Tropical Medicine, Pembroke Place, Liverpool L3 5QA, England Martin J. Evans Wellcome I CRC Institute, Tennis Court Road, Cambridge CB2 1QR, England Darren T. Gilmour Wellcome I CRC Institute, Tennis Court Road, Cambridge CB2 1QR, England Norman Maclean Department of Biology, University of Southampton, Hampshire SO5 3TU, England M. Muller Department of Molecular Animal Breeding, LMU Munich, Veterinarstr. 13, 80539 Munich, Germany Azizur Rahman Department of Biology, University of Southampton, Hampshire, SO9 3TU, England K. Simkiss Department of Pure and Applied Zoology, University of Reading, Reading RG6 2AJ, England vu

Preface

The explosion of knowledge in the biological sciences is now widely recognized. This intellectual revolution has involved some remarkable examples of lateral thinking, and the production of transgenic animals is one such application. The development of these new animal types has depended on the exploitation of cloned genes and techniques capable of achieving the integration of transgenes into the chromosal DNA of the relevant animals. Although the dream of directly modifying an animal's genetic endowment was long cherished, no one, at least until ten years ago, could have guessed how rich the benefits of the transgenic technology would be. We now stand at a scientific watershed, and interesting and productive applications have already appeared in genetics, medicine and agriculture. Obviously they are only the beginning: The technology is now proven and the future looks promising in innumerable ways. This book attempts to provide the reader with a picture of what has been achieved and where the methodology is most likely to go from here. We have chosen to divide the topic by animal groups, partly because the technology varies from group to group, but even more because present and future applications differ strikingly among organisms. Thus, whereas the greatest appeal of transgenic insects is their power to throw light on problems like mutagenesis and gene regulation, the greatest appeal of transgenic mammals is their contribution to agricultural productivity and pharmaceutical research. I am most grateful to the contributors to this volume for their expenditure of time and energy. All are active molecular biologists who have a worldwide reputation in their area of expertise and could be diverted from their experiments only with difficulty. Some of the

IX

x

Preface

chapters have been carefully checked by other knowledgeable colleagues in the field - Chapter 2 by Dr. Michael Ashburner, Cambridge, England; Chapter 3 by Professor Lazlo Orban, Godollo, Hungary, Mr. Ferenc Mtiller, Godollo, Hungary, and Dr. Arati Iyengar, Southampton, England; and Chapter 4 by Dr. K. Powlett, Oxford, England. To these people we are particularly grateful.

ANIMALS WITH NOVEL GENES

1 Transgenic animals in perspective NORMAN MACLEAN

1.1 Introduction 1.1.1 What are transgenic animals?

Since the earliest days of animal domestication, people have sought to improve their flocks and herds and companion animals by selective breeding. Although the results of such breeding are impressive, the procedure is by nature slow and to some extent imprecise. So the dream existed that perhaps one could take a more direct hand in stock improvement by recovering the genetic factors involved and adding these selectively to individual animals. The first attempts to do this took place some thirty years ago, but since the methods of DNA purification were suboptimal and methods had not been developed for gene isolation and cloning, complete genomic DNA preparations were used. These were simply mixed with animal eggs or embryos in the hope that transfer might occur, and then the resulting adults were screened for phenotypic features previously present only in the donors. Many of us can remember such experiments being undertaken with DNA from pigmented Ambystoma (the Mexican axolotl), such DNA being added to or injected into albino embryos of this species in the hope that the pigment gene would be transferred. Since that gene was likely to be swamped by the DNA for at least 1 million alternative sequences, one can see with the benefit of hindsight that the odds against success were enormous. The developments which have now tipped the balance in our favour are chiefly threefold. The first, along with improvements in methods of DNA extraction and purification, was the discovery of restriction enzymes, which made it possible to dissect DNA into manageable fragments of precise length and sequence. Such fragments are easily separated and recovered from agarose gels. 1

2

Norman Maclean

The second development was the discovery of bacterial plasmids, which, together with bacterial viruses, provided easy vector systems into which genes could be inserted for bulk replication and subsequent recovery. These two discoveries together made gene cloning feasible and even straightforward. However, they did not offer a means of recognizing and identifying gene sequences. Hence, the third development was nucleic acid hybridization, with the possibility of using a tagged probe sequence to search out and hybridize with any other molecule having partial sequence homology. With these three developments, the old dreams and experimental visions suddenly became a reality. By purifying and cloning individual gene sequences, it became possible to add many copies of a single gene sequence to eggs or embryos, and then to use DNA hybridization with a radio-tagged probe DNA to search out any persisting or progeny molecules from the initial introduction now present in the adult animals. So it is that the science of transgenic induction in animals has arisen within the past decade, and in the pages that follow, we endeavour to provide a clear account of what has been achieved in this period of intense experimental effort. We also attempt some predictions of where the future lies. What began as a pipe dream or an attempt to improve stock characteristics by selective breeding has now generated other and sometimes unexpected rewards, such as gene therapy for the treatment of inherited disease, new insights into gene regulatory mechanisms and the use of farm animals as production units for the synthesis of desirable pharmaceutical products. Transgenic animals can be defined as individuals to which copies of a gene sequence have been artificially added. The novel DNA is referred to as the transgeney since it is delivered by artificial means to the individual animal, usually but not invariably by insertion into the fertilized egg. If the sequence is demonstrated to have been incorporated into the chromosomal DNA of the animal, the individual can be regarded as being stably transgenic with respect to the added sequence. More frequently, the added DNA sequences are not integrated into the chromosomal DNA but are gradually lost from the embryonic cells over the first few days of development. This is not to say that the genes are not expressed, and indeed such transiently transgenic animals often display transient expression of these temporarily resident

Transgenic animals in perspective

3

gene copies and thus display novel proteins in their temporarily modified phenotype. When a novel gene has been integrated into the chromosomal DNA of an animal, the gene may be active in certain cells and tissues, specific proteins coded by the gene may be produced and consequent changes in the phenotype of the animal may result. Such expression of the transgene is frequent but not invariant. Also, the transgene is likely to be transmitted in the sexual gametes to progeny. Since the initial transgenic individual is almost invariably heterozygous for the novel sequence, no more than 50% of the initial Fl generation progeny are likely to inherit the transgene. The percentage can be lower if the initial transgenic animal is mosaic, in that only some of its cells have arisen from a transgene-inheriting cell (i.e., if incorporation of the transgene copies has occurred after a few rounds of cell replication in the early embryo). Alternatively, the percentage can be higher if more than one chromosome has inherited transgene copies. The three crucial aspects of transgene biology are therefore integration, expression and transmission. Although we are here using the term 'transgenic animal', much of the legislation which has arisen to regulate the experimentation and release of such organisms refers to them as 'genetically modified' or 'genetically engineered'. However, since this terminology is ambiguous, inviting confusion with animals whose genomes have been modified by mutation, polyploid or aneuploid induction or whatever, the term 'transgenic animal' is clearer and more straightforward. 1.1.2 Why produce transgenic animals? There are at least two quite separate reasons for the vigorous efforts in many laboratories to produce transgenic animals. The first is simply that such animals can be used as 'living test tubes' into which a novel sequence is introduced in order to learn about the mechanisms governing its regulation. In this way problems and questions relating to gene promoter characteristics, enhancer sequences, transcription factor site interactions, mutations, position effects, DNA methylation, gender imprinting, tissue specific expression and so on are readily addressed. The second reason is the long-cherished desire to develop strains of animals with new and desirable genetic traits which will prove useful in agriculture, aquaculture or the domestic environment. Essentially

4

Norman Maclean

the logic here is that breeding by genetic selection is usually slow and imprecise, whereas the development of an animal strain carrying and expressing a specific novel gene might be more rapid and more specifically goal-oriented via the transgenic route. There can be no question that transgenics have richly fulfilled their promise in the first regard. Whether they will do so in the second remains for the most part to be demonstrated, although there have already been promising successes. Since this is a book about transgenic animals, we will not dwell on the extensive experimental work on the production of transgenic bacteria, fungi and plants. But it is worth noting that all of these systems have natural plasmids which can if necessary act as vector systems to convey and express new genes in the cells of these organisms. Such natural plasmids are almost unknown in animal cells, even in those of protozoans, and therefore transgenic induction in animals is more difficult than it is in other living systems. The ability to grow entire plants from single somatic cells has also made transgenic induction in the plant kingdom somewhat more straightforward than in the animal kingdom. Some interesting books and reviews covering transgenic animal biology are the following: Church, R. B., ed. (1990). Transgenic Models in Medicine and Agriculture. New York: Wiley-Liss. First, N. L. & Haseltine, F. P., eds. (1991). Transgenic Animals. London: Butterworth-Heinemann. Grosveld, F. & Kollias, G., eds. (1992). Transgenic Animals. London: Academic Press. Jaenisch, R. (1988). Transgenic animals. Science 240, 1468-74. Kingsman, S. M. & Kingsman, A. J. (1988). Genetic Engineering. Oxford: Blackwells. Multiauthor review (1991). Transgenic vertebrates. Experientia 47, 865-935. Murray, J. A. H., ed. (1992). Transgenesis: Applications of Gene Transfer. New York: Wiley. Peters, P. (1993). Biotechnology: A Guide to Genetic Engineering. Dubuque, Iowa: Brown. Watson, J. D., Witkowski, J., Gilman, M. & Zoller, M. (1992). Recombinant DNA, 2d ed. San Francisco: Freeman.

1.2 A brief look at the methodology

In general, the easiest way to achieve transgene induction in an animal is to introduce the DNA as multiple copies of the chosen sequence into the fertilized egg. For example, in the mouse system, eggs are

Transgenic animals in perspective

5

recovered from a superovulated mouse, and the transgene copies are injected directly into one of the pronuclei (usually the male pronucleus). The early embryos are then reintroduced into the uteri of pseudopregnant surrogate females. Variants of this approach are used with insects, fish and amphibians where the nuclei of the egg are too small to be readily visualized and the transgene copies are injected into the perinuclear cytoplasm. In certain fish species, oocytes rather than eggs are injected and in vitro fertilization follows injection. However, injection is slow and often requires considerable skill; a variety of alternative methods exist. These include electroporation of sperm or fertilized eggs (in some cases after egg dechorionation); lipofection, in which the transgene copies are prepackaged in lipid micelles and fusion of these packages with the eggs follows; and bombardment of the eggs or embryos with gold or tungsten particles previously coated with the DNA. If gametes and zygotes are inaccessible, as they are in birds, infective retro viruses may be used as vectors and the embryos exposed to the infective virus. Integration will, of course, be mosaic and germ-cell transmission rather infrequent. The exploitation of embryonic stem cells, especially in mammals, has opened up altogether new and beneficial ways of inducing transgenesis. Here cells of embryonic origin (in mammals these are cells originating from the inner cell mass of the early blastula) are cultured in vitro in conditions that allow proliferation without differentiation. While in culture, they may be transfected with the transgenes by electroporation or facilitated uptake. What is so good about this system is that the cell population can then be exposed to rigorous selection in culture, but not just in terms of transgenic versus nontransgenic cells; cells can also be selected which have fortuitously incorporated the transgene into a particular locus. This also opens the possibility of gene targeting either by homologous recombination or by selective ablation. Once selected, the desired cells can then be replaced in a suitable embryo and the embryo grown to term. Resulting transgenic individuals will be mosaic, but a percentage will be partially germline-transformed . It will be obvious to many readers that advances in gene cloning permit the test-tube construction of new combinations of DNA sequences, in which the promoter of one sequence is spliced to the coding sequence of another and so on. Such constructs also make it possible to use transgenic animals as test systems in order to investigate the precise roles of regulatory sequences such as TATA boxes

6

Norman Maclean

(consensus sequences found in most promoter regions and recognized by RNA polymerase enzymes and transcription factors) and enhancers, especially in terms of tissue-specific expression. An important part of this methodology is the use of reporter genes. These are protein coding genes which encode enzymes not present in the transgenic animal and for which a sensitive assay procedure exists. Popular reporter genes include lac Z, CAT and LUC, which are, respectively, bacterial /3-galactosidase, bacterial chloramphenicol acetyltransferase and luciferase (commonly of firefly origin). Transgene expression can be profoundly affected by the characteristics of the construct, the site of integration (so-called position effects), the integration of single copy or concatemer tandem copies and endogenous responses such as transgene methylation. The presence of the transgene is commonly checked by such techniques as the polymerase chain reaction and Southern blotting; it is of great importance to demonstrate that chromosomal integration of transgene copies has actually occurred. Transient expression of nonintegrated transgene copies is routinely observed, so expression is no indication of integration. Integration into mitochondrial DNA is theoretically possible with such techniques as cytoplasmic injection but has been little investigated. The method for investigating transmission is to look for integration and/or expression in progeny following interbreeding of transgenics with nontransgenics. Transgenes may be lost as well as gained, and this presumably becomes more likely if incorporation has been facilitated by terminal repeat sequences of a transposable element. Although the phenomenon has been reported, most transgenes have proved to be stably inherited once integrated. 1.3 Some important advances

Since the field of transgenic animals is now so large, it is not possible to discuss all the important advances in one short section. Instead, a few key developments will be highlighted and briefly discussed. 1.3.1 Introducing the transgene to the animal

In this fundamental methodology, injection into eggs or very early embryos by microneedle remains the most reliable choice in all but a very few systems. Lipofection and electroporation are probably the

Transgenic animals in perspective

7

closest rivals in terms of promise, but fusion of the liposome, containing the packaged DNA, with the fertilized egg cell is often difficult and sometimes impossible. Electroporation requires finding a balance between survival of the eggs after exposure to the electrical field conditions and electrical pulse strength which will transfer the DNA into the eggs. If very few copies of the transgene are transferred, the chances of integration and subsequent transmission remain poor. The characteristics of different electroporation kits are highly variable, and it may yet become clear that the technique has much to offer. (The great benefits of exploiting embryonic stem cells for gene transfer are discussed separately in Section 1.3.5.) Another interesting possibility is the introduction of transgenes directly into gonads by localized microinjection or ballistic delivery via DNA-coated particles propelled by an explosive charge. If the DNA were incorporated into progenitor cells which would later produce gametes, then both integration and germ-line transmission might be achieved in one step by the use of the gametes after such targeted introduction. 1.3.2 Transgene integration Although many studies using transgenic animal technology require only transient expression of nonintegrated copies of the transgenes to provide useful data, most experimental programmes attempt the chromosomal integration of one or more transgene copies. Some interesting methods have been invoked in order to facilitate integration. These chiefly involve the attempted use of terminal repeat sequences, as are found in retroviruses and other transposable elements. Such terminal repeats, in the form of a series of short repeats, can be added to the termini of gene constructs in the hope of facilitating integration after transgene delivery. However, the efficiency of such repeats in achieving integration is probably partly dependent on the existence of similar repeats within the genome being targeted and on specific transferase enzymes which facilitate the actual integration. For example, not only do the P elements used so efficiently to achieve transgene integration in Drosophila melanogaster carry their own specific transposase gene, but the splicing of the exon is appropriately achieved only in germ-line tissue, thus preventing normal transposition in somatic cells (see review by Rio 1990). Harnessing the machinery of transposition to help achieve

8

Norman Maclean

integration in other systems is clearly an attractive proposition, and already the co-injection of transposase enzyme has been shown to be just as effective as co-injecting the P element in Drosophila (Kaufman & Rio 1991). Retroviruses and some bacteriophages effect their integration into host genomes by using recombinase enzymes. One family of sitespecific recombinases are termed integrases, requiring as they do short regions of genetic homology between the partner molecules (Argos et al. 1986). The yeast FL recombinase has been used in mammalian cells: Cells were transfected with an expression construct together with a reporter plasmid that was a substrate for the recombinase (O'Gorman et al. 1991). The FL recombinase system has also been shown to catalyze site-specific recombination in Drosophila (Golic & Lindquist 1989). More recently, the phage PI Cre recombinase has been used in tandem microinjection along with a /oxP-flanked j8galactosidase reporter gene, and the resulting transgenic mice shown to have integrated the reporter transgene at specific sites (Orban et al. 1992). Of particular interest is the work of Dale and Ow (1991), who achieved the integration of a luciferase gene in combination with a hygromycin phosphotransferase gene (to act as a selectable marker gene) in tobacco plants. The phage PI Crellox recombination system was used to provide transgene flanking sequences, and the Cre recombinase was subsequently activated to excise the selectable marker but to leave the reporter gene. Such a system might be desirable for application in transgenic animal systems. 1.3.3 Transgene expression As mentioned earlier, when transgenic animals are used as 'living test tubes' in which to study gene expression by novel constructs, integration may not be required. When transgene copies are introduced into egg cytoplasm rather than into the nucleus, as they are, for example, in insects, fish and amphibians, a pseudonucleus frequently forms around the exogenous DNA. Replication of the transgene copies may then occur, with or without attendant gene expression. Eventually, most or all of this foreign DNA is degraded or lost, presumably because, since it is not associated with centromeres, it cannot be readily segregated at mitosis. (See discussion of this topic in Chapter 3, Section 3.4.1.) This specific temporary amplification of cytoplasmically located transgene copies is therefore a bonus to transient expres-

Transgenic animals in perspective

9

sion studies. Temporary replication and expression of nuclearly located transgenes may occur but are apparently less dramatic. Transient-expression studies often involve reporter genes, and the widespread use of neo, CAT, ft-gal and LUC amply demonstrates their effectiveness. Such genes are particularly useful for studies on regulatory elements. The reporter gene of choice varies with the transgenic system, neo is often used as a selection system, since it provides resistance to the toxic drug neomycin or its analogue G418. Unfortunately, this approach will not work if the transgenic animals are mosaic with respect to the transgene, since nontransgenic cells will not be resistant. The sensitivity of the assay for particular reporter genes is often highly variable. In my own laboratory we find the procedure for detecting fi-gal particularly sensitive and error free. The use of some of these reporter genes is bedevilled by confusion with endogenous expression, either of galactosidase in the foetal gut, or even of contaminating bacteria in the sample. The use of luciferase enzyme in living organisms has stirred the imagination, so that transgenic plants or fish may glow in the dark when provided with the substrate luciferin. Living zebra fish transgenic for this reporter gene may be placed in vials of water containing a trace of luciferin and be scored as positives by scintillation counting! Maximal transient expression is probably best afforded by the use of viral promoters such as the Rous sarcoma virus long terminal repeat, since viral promoters have presumably evolved to function under these precise conditions. Turning to integrated expression, the expression of genes is profoundly affected by a host of regulatory factors, one of which is the specific attachment of the transcribing DNA to the chromosomal scaffold so as to best effect its looping out to facilitate transcription. Such matrix-attachment regions (MARs) have been isolated, and there is some evidence that their inclusion in a novel construct will indeed facilitate transcriptional expression. A useful discussion of the role of such MAR sequences, especially in connection with the chicken lysozyme gene, is to be found in Sippel et al. (1992) and McKnight et al. (1992). Endogenous activity of MAR elements may also explain the finding that co-integration of a poorly expressed transgene in the vicinity of an actively transcribed transgene can improve the expression efficiency of the former (Clark et al. 1992b, 1993). The inclusion of tissue-specific locus control regions is also important (see discussion by Dillon 1993), and there is a good deal of evidence that genomic

10

Norman Maclean

genes with introns are often better expressed than their intronless counterparts (Clark et al. 1993). The expression levels of integrated transgenes are most profoundly affected by the associated 5' regulatory region. Until recently it was commonly assumed that if sufficient 5' promoter sequence was provided, maximal expression would be obtained. Sadly, things are clearly not so simple. As is clear from the work of Wada-Kiyama et al. (1992) on human e-genes, silencer elements may also be found in upstream sequences, and rates of expression can be dramatically increased by actually shortening the 5' regulatory region in some cases. It now looks as though each promoter in turn will have to be dissected and manicured by a substantial series of experiments involving a good reporter gene before maximal expression in any one tissue is obtained. Let us hope that not too much tissue specificity and species specificity will turn out to involve silencer as well as enhancer elements. 1.3.4 Optimizing transmission A common complication in ensuring a high rate of transgene transmission, provided that a reasonable incidence of integration is being achieved, is mosaicism, and anything that can be done to reduce a mosaic status in the original transgenic animals, or at least to ensure that the germ-cells are themselves at least partially transgenic, will ensure a reasonable incidence of transmission. If the germ-line is set aside very early in development, as it is in determinative embryos, any integration events occurring after the first few rounds of division are bound to result not only in mosaicism but in poor or zero rates of transmission to progeny. Regulative embryos like those of fish are more amenable to transmission but still often lead to mosaic transgenics (perhaps because of the cytoplasmic injection procedure). Transgene mosaicism of germ-cells is commonly found in the gonads of transgenic fish. In mammals, where gender imprinting occurs, this may pose problems for transmission if transgenes have been integrated into the sequences of other crucial genes. Gender imprinting also precludes another useful genetic trick that may prove beneficial in fish and amphibians, namely the induction of gynogenetic or androgenetic development in the eggs of transgenic individuals. This should ensure

Transgenic animals in perspective

11

that the progeny are homozygous for the transgene in the Fl generation, without waiting for the F2. 1.3.5 The use of embryonic stem cells Various alternatives to egg or oocyte injection have been considered in the history of animal transgenics; most of these have been mentioned or are discussed in later chapters. One alternative is to exploit cell culture by transfecting the cells in culture with the transgene copies and then replacing a fertilized egg cell nucleus with the nucleus of the transfected cell. Clearly, such donor cells must have totipotent nuclei, but the experiments of John Gurdon and his collaborators have demonstrated that, at least in Xenopus, the experiment is feasible (see review by Gurdon & Melton 1981). In mammalian systems it seemed that one might be able to achieve similar success by the use of, say, leukaemic cells, but the real breakthrough has come with the development of a culture system which will allow the growth and proliferation of embryonic stem cells (ES cells) without differentiation and their subsequent reintroduction to the inner cell mass of the mammalian embryo (see review by Evans 1989). Given that the transfection of ES cells in culture is straightforward, that gene targeting and homologous recombination are proving more efficient each year and that cell selection by drugs is feasible in culture, it becomes clear that this system offers great potential. For the moment the success seems limited to some mammalian species, but there is promise of a parallel system being developed in zebra fish (see Chapter 3, Section 3.2.8). 1.3.6 Gene targeting and homologous recombination A useful account of this topic can be found in the review by Doi et al. (1992). It is now clear that, although homologous recombination is rare in tissue cells as compared with random integration, the use of long, absolutely identical isogenic DNA sequences can provide targeting frequencies as high as 78% of all the integration events (Riele et al. 1992). This opens up two interesting possibilities. The first is gene ablation by targeting modified transgenes to the homologous chromosomal sequence. The second is accomplishing high rates of incorporation by a double induction process. In such a process, lines of animals

12

Norman Maclean

would first be produced that were transgenic for a novel target sequence, such as a sequence of repetitious DNA or a phage genome, and then the transgenes of choice would be introduced into the fertilized eggs of these founder transgenics. The second-round transgenes would include a long sequence precisely homologous to the first incorporated transgene. 1.3.7 Gene therapy in humans One of the most interesting by-products of the development of transgenic technology is the possible treatment or cure of some human diseases by gene therapy (see the review by Mulligan 1993 and the special volume of Trends in Biotechnology 1993). A number of possibilities exist for effective gene therapy (Tables 1.1 and 1.2). A comprehensive list has been compiled by Van Beusechem and Valerio (1992). One is the recovery of a stem cell population from the patient - say, lymphoid cells from marrow or spleen - transfection of these cells in culture with the chosen transgene and selection of transgenic cells, followed by reintroduction of the new transgenic cells into the patient. A second possible strategy is the use of retroviral vectors to accomplish gene integration, either with tissue cells recovered as described earlier or with some sort of in vivo gene transfer (see Valerio 1992; Van Beusechem & Valerio 1992). A third, overlapping possibility is to use this in vivo gene transfer to achieve short- or long-term transgene expression within a target tissue, the gene product being effective in relieving the diseased state. Surprisingly, it is now clear that genes can be injected directly into such tissues as mammalian heart muscle or mouse brain, and transient or even long-term expression of the transgene obtained (Kitsin et al. 1991; Ono et al. 1991). The administration of cloned genes coding for cystic fibrosis transmembrane conductance regulator to the lung by nebulizer is envisaged as treatment for cystic fibrosis (Mulligan 1993), and protection against influenza virus has been demonstrated in mice following muscle injection of a relevant viral protein-coding gene (Ulmer et al. 1993). Gene therapy has already been authorized in the United Kingdom (see Brown 1992 and Clothier 1992) and the United States. Early attempts at gene therapy in humans have included the introduction of a gene coding for tumour necrosis factor (TNF) into tumour-infiltrating lymphocytes and administration of the lymphocytes, now transgenic for the TNF gene,

Transgenie animals in perspective

13

Table 1.1. Diseases in which gene therapy may be useful and for which cloned genes are available Adenosine deaminase deficiency Argininosuccinic aciduria (argininosuccinase) Citrullenemia (arginosuccinate synthetase) Coagulation factors VIII, IX, X, XHIa Complement factors C2, C4, C9 Elliptocytosis 1 (protein 4.1) Elliptocytosis 2 (spectrin) Gaucher disease type I (glucocerebrosidase) Granulocyte actin deficiency Haemoglobinopathies Hereditary angioneurotic oedema (Cl inhibitor) Phenylketonuria (phenylalanine hydroxylase) Purine nucleoside phosphorylase deficiency Thalassaemia

Table 1.2. Diseases potentially treatable by gene therapy but in which serious targeting problems remain arAntitrypsin deficiency Carbamylphosphate synthetase deficiency Cystic fibrosis Fabry disease (a-galactosidase) Fucosidosis (a-fucosidase) Gaucher disease types II and III Hypophosphatasia (alkaline phosphatase) Metachromatic leukodystrophy, variant (SAP 1) Lesch-Nyhan disease (HPRT) Ornithine transcarbamylase deficiency Propionyl CoA-carboxylase deficiency Sandhoff disease (hexosaminidase A and B) Tay-Sachs disease (hexosaminidase)

back to the patient whence they came (see Gershon 1991) and administration of the adenosine deaminase (ADA) gene copies to patients with immune deficiency, by transfecting some of the patients' cells in vitro with a retroviral vector containing a human ADA gene (which also expresses neomycin resistance for use as a selection system). An

14

Norman Maclean

outline of this and other proposed work on gene therapy can be found in the reviews of Miller (1992), Mulligan (1993) and Trends in Biotechnology (1993). There is a medium- to long-term prospect of treating insulindependent diabetes mellitus by gene therapy (see Newgard 1992). There is also the possibility that animals such as pigs could be transgenically modified so that their tissues, if used for heterologous grafting into humans, would be less immunogenic. Gene therapy is, however, a two-edged sword in many ways. It is yet to be determined whether proto-oncogenes in the genome can be activated by the random integration of therapeutically administered cloned genes. The possibility is obviously there, but it has yet to be quantified. A second and more worrisome aspect is the future misuse of gene therapy, especially with genes whose products would modify behaviour. It is not difficult to foresee the benefits of administering therapeutic genes to overaggressive companion animals or troublesome strains of farm stock. But would it then be advocated for some human behavioural disorders, or recommended for certain prisoners? In the long view, gene therapy is probably the area of gene cloning and manipulation which should give human society the greatest cause for concern, and its uses must be monitored with the utmost caution. 1.3.8 Pharmaceutical products from transgenic animals

As well as a human therapy spinoff, transgenic animal science has provided a possible technique for producing pure pharmaceutical products economically. The possibility follows from the easy combination of promoter sequence with structural gene, so giving a construct which, when used as a transgene, will ensure the production of a genecoded protein in an animal and/or an organ in which that product would not normally be expressed. A search is now under way for appropriate producer systems to rival bacteria and yeast. A few have already been exploited. The most celebrated is that of the sheep, in which human factor IX (important in blood clotting in certain types of haemophilia) and human arantitrypsin (a protease inhibitor important in lung function and deficient in emphysemic patients) have been hooked up to the sheep /3-lactoglobulin promoter. Sheep transgenic for these constructs express the products in their milk, from which they may be readily purified (see Clark et al. 1989, 1992a). Clearly, these

Transgenic animals in perspective

15

successful experiments in sheep are only the beginning. There have been parallel attempts to produce human proteins in transgenic silkworms (Bombyx mori larvae). 1.4 Legislation, exploitation and release

Various ethical and legal problems arise in the context of the development, exploitation and release of transgenic animals. The problems are chiefly associated with four aspects of the subject, namely the morality of making transgenic animals in the first place, possible dangers to human health posed by some of these organisms, the acceptability of transgenic animal products as food and dangers to the environment posed by the accidental or intentional release of transgenic animals. I will discuss each of these problem areas briefly. 1.4.1 Is transgenic animal production immoral?

Evolution works through the natural selection of novel animal forms which arise as a result of gene mutation and the continuing variation generated by sexual reproduction and genetic recombination. Evolution thus not only fixes species as interbreeding populations of organisms, but helps to generate new species and thus increase the diversity of life. In addition to this is the great variation which has been produced by hundreds of years of 'artificial' human selection and contrived crossing of domesticated animals in the context of agriculture, aquaculture and companion animals. Against this background, transgenic induction can be seen simply as a more directed and focused artificial breeding programme, where a desirable phenotypic feature can be introduced or emphasized at a single gene level. But a somewhat novel dimension is introduced by transgenics because unique genetic combinations of promoter sequence, coding sequence and animal genome can be chosen; as a result, the target species may come to possess and express genetic traits which could probably never have resulted from natural selection or artificial breeding owing to the barriers of frequent interspecific sterility. The chances of producing a sheep expressing human a r antitrypsin gene in its milk by means other than transgenic induction are certainly very remote. Thus, we can see that transgenic animal production is both essentially similar to more conventional methods of generating genetic variation and yet singularly different.

16

Norman Maclean

Public response to the science has been mixed, which is not surprising. Members of the Green Party have consistently voiced opposition, except perhaps in the medical context of transgenic mice, and so German science includes little such work. One result has been the movement of committed transgenic researchers in Germany to laboratories in other countries. Most countries have adopted more liberal views than Germany, varying from permission with strict control in the United Kingdom and United States to apparent lack of concern in China and Russia. In the United Kingdom each experiment requires the express permission of the Health and Safety Executive of the Department of the Environment. Organizations like Greenpeace keep a distrustful and generally disapproving eye on the work, but in the main the system works well. Permission is clearly withheld if the resulting transgenic animal would be genetically traumatized or if its existence or accidental escape could be seen as a threat to public health or the environment. Laboratories carrying out the work are regularly inspected and have to adopt stringent safety standards and working conditions. The comparative difficulty of doing the work is itself a control on the unregulated production of transgenic animals in many laboratories outside the developed world, but this will no doubt change. It is not easy to imagine how individuals with ill intent could exploit transgenic animals as weapons of war, but the history of human aggression and malice certainly gives strong grounds for extreme caution and control, more indeed than is currently being exercised in such countries as China and Russia. 1.4.2 Are transgenic animals a threat to human health or welfare? Some animals pose a threat to human health by their very existence. Examples frequently cited include foxes carrying rabies or badgers carrying bovine tuberculosis (although the latter threat to human health is indirect). There is no doubt that animals could be engineered by the transgenic pathway to secrete toxins or to act as effective carriers of pathogenic bacteria. (Here the carrier status, not the bacterium itself, would result from transgenic induction.) But transgenic animals do not pose any greater or different threat than nontransgenics in this connection. Provided that the authorization control is effective, no potentially dangerous transgenic animals should be produced. Thus, the simple answer to our question is no, transgenic animals do

Transgenie animals in perspective

17

not as such pose a threat to human health, and are unlikely ever to do so, even in the wrong hands. 1.4.3 Is it safe to eat transgenic animals? That it is safe to eat some transgenic organisms is evidenced by the fact that transgenic tomatoes have been authorized for sale in the United States. These tomatoes are transgenic for a sequence of DNA which modifies the speed of ripening. On the other hand, since domestic tomatoes could readily be made transgenic for a toxin gene, it is clearly not safe to eat all transgenic forms of normally edible organisms. What I want to discuss here is whether there are any special dangers in eating, say, a cow or a chicken which is transgenic for, say, a growth hormone sequence derived from the DNA of, say, a rabbit. Two separate questions are involved. First, is it safe to eat the animal, considering that it is carrying a novel rabbit gene in all of its cells? Second, is it safe to eat a cow or a sheep which has an excess of growth hormone of rabbit origin in its blood and tissues? The answer to the first question is that surely there is no possible hazard. There is no genetic transfer of cow genes to a human who eats beef and therefore no additional threat to eating beef containing a mix of cow and rabbit DNA. DNA is effectively degraded in the human gut, whether it is of bovine or of rabbit origin. The second question is more complex, and is relevant since growth hormone transgenics may be considered in the future for farming. Since growth hormone itself is a proteinaceous hormone, it is also digested and does not threaten human health. In some situations, such as those of fish or mammalian new-born, polypeptide hormones such as growth hormone could cross the gut intact and modify the physiology of the eater. But all present evidence indicates that the growth hormone content of our meat is not a threat to our well-being and that the consumption of meat from growth hormone transgenics is entirely safe. Food from transgenic animals therefore carries no conceivable new threat to health, unless, of course, the novel gene generates a product which is itself toxic. That scarcely poses a new problem, however, since it is simply the counterpart of the threat to our health from the consumption of parasitized meat or meat products contaminated with botulinus toxin.

18

Norman Maclean

1.4.4 What problems are posed by release? Many transgenic animals pose few if any additional problems to society, since they would not survive in the wild, or if they accidentally escaped, interbreeding with wild species would be unlikely or impossible. However, quite important problems are associated with the release of such species as insects or fish, and these have been considered by, among others, Bruggemann (1993). There is already in place a procedure for monitoring and assessing risk to the environment before authorization for release in the United Kingdom. This is presented in the January 1993 publication of the Department of the Environment entitled 'The Genetically Modified Organisms (Deliberate Release) Fees and Charges Scheme 1993 and Guidance Notes'. Industrialists have voiced objections to the high cost and restrictive nature of the scheme (see Coghlan 1992). Monitoring of the release of transgenic plants is at an advanced stage, and a recent analysis of a U.K.-authorized release of oilseed rape is, in the main, reassuring (Crawley et al. 1993). Many of the most serious problems are associated with the accidental or intentional release of transgenic insects, molluscs or fish. Elaborate and expensive containment facilities have been constructed in the Unites States to house transgenic carp undergoing growth trials. Hallerman and Kapuscinski (1992) have discussed some of the uncertainties associated with transgenic fish release, and list relevant publications on the topic. One probably cannot significantly improve on the conclusions of the U.K. Government Royal Commission on Environmental Pollution in its report entitled The Release of Genetically Engineered Organisms to the Environment' (July 1989). This resulted in the publication by the U.K. Advisory Committee on Genetic Manipulation of new guidelines in January 1990 entitled The Intentional Introduction of Genetically Manipulated Organisms to the Environment' . One of the chief recommendations of the Royal Commission was to take each transgenic animal bearing a novel gene construct as a separate case, rather than try to generalize for a class or species of animal. Any threat to the environment is as likely to arise from the gene construct, its probable expression levels and stability as from the habits of the animal species concerned. Another recommendation was to regard each case as roughly the equivalent of a new species. Most countries now have a great deal of experience with escape or inten-

Transgenie animals in perspective

19

tional release of nontransgenic species. The commission concluded that, of all such releases in the United Kingdom, both plant and animal, approximately 1 in 10 had become established in the environment on a more or less permanent basis, and of those so established, approximately 1 in 10 had become a pest. Transgenic animals will probably pose an equivalent threat. Whether we can spot the pests before they are released is surely the crucial question. References Argos, P. et al. (1986). The integrase family of site-specific recombinases. EMBO Jr. 5, 433-40. Brown, P. (1992). Gene therapy wins official blessing: Report on the Committee on the Ethics of Gene Therapy. New Scientist, January, 18. Bruggemann, E. P. (1993). Environmental safety issues for genetically modified animals. / . Anim. Sci. 71 (Suppl. 3), 47-50. Clark, A. J., Bessos, H., Bishop, J. O., Brown, P., Harris, S., Lathe, R., McClenaghan, M., Prowse, C , Simons, J. P., Whitelaw, C. B. A. & Wilmut, I. (1989). Expression of human anti-hemophilic factor IX in the milk of transgenic sheep. Biotechology 7, 487-92. Clark, A. J., Simons, J. P. & Wilmut, I. (1992a). Germ line manipulation: Applications in agriculture and biotechnology. In Transgenic Animals, ed. F. Grosveld & G. Kollias, pp. 247-71. London: Academic Press. Clark, A. J., Cowper, A., Wallace, R., Wright, G. & Simmons, J. P. (1992b). Rescuing transgene expression by co-integration. Biotechnology 10, 1450-5. Clark, A. J., Archibald, A. L., McClenaghan, M., Simons, J. P., Wallace, R. & Whitelaw, C. B. A. (1993). Enhancing the efficiency of transgene expression. Phil. Trans. R. Soc. Lond. B 339, 225-32. Clothier, C. (1992). Report of the Committee on the Ethics of Gene Therapy. London: HMSO. Coghlan, A. (1992). Industry slams draft law on novel organisms. New Scientist, January, 20. Crawley, M. J., Halles, R. S., Rees, M., Kohn, D. & Buxton, J. (1993). Ecology of transgenic oilseed rape in natural habitats. Nature 363, 620-3. Dale, E. C. & Ow, D. W. (1991). Gene transfer with subsequent removal of the selection gene from the host genome. Proc. Natl. Acad. Sci. 88, 10558-62. Dillon, N. (1993). Regulating gene expression in gene therapy. Trends Biotech. 3, 167-73. Doi, S., Campbell, C. & Kucherlapati, R. (1992). Directed modification of genes by homologous recombination in mammalian cells. In Transgenic Animals, ed. F. Grosveld & G. Kollias, pp. 27-46. London: Academic Press. Evans, M. J. (1989). Potential for the genetic manipulation of mammals. Mol. Biol. Med. 6, 557-65. Gershon, D. (1991). Cancer trial starts. Nature 349, 445. Golic, K. G. & Linquist, S. (1989). The FLP recombinase of yeast catalyzes site-specific recombination in the Drosophila genome. Cell 59, 499-509.

20

Norman Maclean

Gurdon, J. B. & Melton, D. A. (1981). Gene transfer in amphibian eggs and oocytes. Annu. Rev. Genet. 15, 189-218. Hallerman, E. M. & Kapuscinski, A. R. (1992). Ecological and regulatory uncertainties associated with transgenic fish. In Transgenic Fish, ed. C. L. Hew & G. Fletcher, pp. 209-28. World Scientific. Kaufman, P. D. & Rio, D. C. (1991). Germline transformation of Drosophila melanogaster by purified P element transposase. Nucl. Acids Res. 19, 6336. Kitsin, R. N., Butlrick, R. N., McNally, E. M., Kaplan, M. L. & Leinwand, L. A. (1991). Hormonal modulation of a gene injected into rat heart in vivo. Proc. Natl. Acad. Sci. 88, 4138-42. McKnight, R. A., Shamay, A., Sankaran, L., Wall, R. J. & Hennighausen, L. (1992). Matrix-attachment regions can impart position-independent regulation of a tissue-specific gene in transgenic mice. Proc. Natl. Acad. Sci. 89, 6943-7. Miller, A. D. (1992). Human gene therapy comes of age. Nature 357, 455-60. Mulligan, R. C. (1993). The basic science of gene therapy. Science 260, 92632. Newgard, C. B. (1992). Cellular engineering for the treatment of metabolic disorders: Prospects for therapy in diabetes. Biotechnology 10, 1112-20. O'Gorman, S., Fox, D. T. & Wall, G. M. (1991). Recombinase mediated gene activation and site specific integration in mammalian cells. Science 251, 1351-5. Ono, T., Fujino Y., Tsuchiya, T. & Tsuda, M. (1991). Plasmid DNA directly injected into mouse brain with lipofectin can be incorporated and expressed by brain cells. Neurosci. Lett. 117, 259-63. Orban, P. C , Chui, D. & Marth, J. D. (1992). Tissue and site-specific DNA recombination in transgenic mice. Proc. Natl. Acad. Sci. 89, 6861-5. Riele, H., Maandag, E. R. & Berns, A. (1992). Highly efficient gene targeting in embryonic stem cells through homologous recombination with isogenic DNA constructs. Proc. Natl. Acad. Sci. 89, 5128-32. Rio, D. C. (1990). Molecular mechanisms regulating Drosophila P element transposition. Annu. Rev. Genet. 24, 543-78. Sippel, A. E., Sanevessig, H., Winter, D., Grewal, T., Faust, N., Hecht, A. & Bonifer, C. (1992). The regulatory domain organization of eukanjotic genomes: Implications for stable gene transfer, in Transgenic Animals, ed. F. Grosveld & G. Kollias, pp. 1-26. London: Academic Press. Trends in Biotechnology. (1993). 11 (5), 156-215. Special volume devoted to gene therapy. Ulmer, J. B. et al. (1993). Heterologous protection against influenza by injection of DNA encoding a viral protein. Science 259, 1745-9. Valerio, D. (1992). Retrovirus vectors for gene therapy procedures. In Transgenic Animals, ed. F. Grosveld & G. Kollias, pp. 211-46. London: Academic Press. Van Beusechem, V. W. & Valerio, D. (1992). Prospects for human gene therapy. In Transgenesis: Applications of Gene Transfer, ed. J. A. H. Murray, pp. 283-322. New York: Wiley. Wada-Kiyama, Y., Peters, B. & Constance, T. (1992). The epsilon globin gene silencer. J. Biol. Chem. 267, 111532-8.

2 Transgenic insects JULIAN M. CRAMPTON AND PAUL EGGLESTON

2.1 Introduction

For many years, transgenic research in eukaryotic organisms has been dominated by only three species, namely Drosophila, the mouse and yeast. Perhaps at the forefront, research with the fruit fly Drosophila melanogaster has, to an unprecedented degree, shown how transgenic technology can be used not only to manipulate genes, but also to extend our understanding of molecular genetics. We therefore feel it appropriate to devote the first half of this chapter to a description of the development of transgenic technology in Drosophila melanogaster and an overview of the ways in which this technology has been applied. The second section concentrates on attempts to apply the same technology to non-drosophilid insects. The future of transgenic insects is then briefly discussed in order to explore the potential that transgenic technology may have in insects of economic, agricultural and medical significance.

2.2 Drosophila melanogaster 2.2.1 Historical introduction

One chance event has, more than any other in recent times, generated an explosive increase in what is known of the molecular biology of Drosophila melanogaster. That chance event was the finding of male recombination among the progeny of crosses between wild caught males and laboratory stock females, reported by Hiraizumi in 1971. Male recombination is not normally seen in this species, and this aberration, once associated with work elsewhere, ultimately led to the unearthing of the P transposable element family. We now know that 21

22

Julian M. Crampton and Paul Eggleston

male recombination is one of a syndrome of germ-line-specific genetic aberrations, collectively referred to as 'hybrid dysgenesis' (Kidwell et al. 1977), which result from transposition of the P element. Other such traits include reduced fertility and sterility, high mutation frequencies, chromosome rearrangements and meiotic drive (Engels 1983, 1989). The P element now forms a routine part of the armoury in molecular genetic research with Drosophila. It has greatly facilitated the cloning of genes through insertional mutagenesis (transposon tagging) and has been used to generate novel alleles following remobilization of an integrated element. There is no question, however, that the ability to use P elements to transform the germ-line of intact insects has had the most far-reaching consequences. Stably transformed 'transgenic strains' have now been used to study a number of genetic problems which would have been much more difficult to tackle without this facility. In his excellent review, Engels (1989) cites a list which includes gene regulation, amplification, chromosome puffing, population variability and evolutionary divergence. This is by no means exhaustive, and the field is expanding rapidly, for example, into the analysis of gene expression and developmental regulation. The advances made in Drosophila genetics through exploitation of the P element represent a watershed, but drosophilists, resourceful as they are, had been trying for some time to develop transformation techniques in an attempt to revert specific mutations. These first attempts involved soaking permeable mutant embryos in solutions of wild-type DNA (Fox & Yoon 1966, 1970). Occasionally, these experiments yielded somatic mosaics in which certain cells expressed the transforming wild-type sequence rather than the mutant allele. The most likely explanation for this is that the transforming DNA was subject to episomal transmission to a subset of the available cells rather than to stable chromosomal integration. However, a small proportion of these transformation events did involve the germ-line tissue and were inherited in a Mendelian fashion. Subsequently, attempts were made to improve upon the crude delivery of transforming DNA. Rather than the embryos being soaked in wild-type DNA, this solution was microinjected into preblastoderm embryos (Germeraad 1976; Liu et al. 1979). This, of course, was the technique which was later to become routine in P-element-mediated transformation. At this stage, the Drosophila embryo has a syncitial blastoderm with many cleavage nuclei around the periphery and no individual cellular membranes. There are, therefore, many nuclei available to interact with the trans-

Transgenic insects

23

forming DNA. Germeraad (1976) used homozygous vermilion - brown mutants which are unable to synthesize either the ommochrome or drosopterin eye pigments and which, therefore, have white eyes. These were injected with wild-type DNA, and a proportion emerged as adults with pigmented eyes. We would now interpret these results as another example of unstable extrachromosomal transmission, but a small proportion of Germeraad's (1976) transformed adults reportedly transmitted the phenotypic alteration to their progeny. These may, in fact, have represented stable transgenic lines, but the molecular basis of these events was not determined. Similar experiments had also been carried out with other insects and, in fact, the earliest report of a successful somatic gene transformation was not in Drosophila, but in the meal moth Ephestia kuhniella, again through a microinjection technique (Caspari & Nawa 1965). Wild-type DNA was introduced into embryos carrying a wing scale mutation, and a proportion of these embryos gave rise to adults with wild-type wing scales. The experiments were repeated with red-eye Ephestia mutants and also with white-eye Bombyx mod mutants (Nawa & Yamada 1968). Once again the majority of these events involved extrachromosomal transmission to somatic tissue, and phenotypic alterations were not subject to Mendelian inheritance. Those rare occasions in which transformed progeny were produced probably represented chance germ-line transformation events through illegitimate recombination. This procedure is the simplest method by which transforming DNA can become incorporated into a recipient genome, although it is technically demanding and in some cases rather inefficient. In essence, it can occur whenever DNA is introduced (usually as a plasmid) into any plant or animal cell nucleus. Such techniques have been used to transfect cultured cells and to generate transgenic mice, plants, nematodes, slime moulds, frogs and sea urchins (Spradling 1986). Often, these techniques yield reasonably high transformation frequencies, and the poor return of stable transformants in early Drosophila microinjection experiments possibly reflects the fact that Drosophila embryos do not routinely incorporate microinjected DNA through such unspecific mechanisms. Another disadvantage of illegitimate, random recombination mechanisms is that the transforming DNA is often integrated into the recipient chromosomes as multiple, tandemly repeated copies (Constantini & Lacy 1981). This can complicate subsequent molecular analyses. Real advances in Drosophila transformation were achieved only with the linking of the transforming DNA to an appropriate vector

24

Julian M. Crampton and Paul Eggleston

which could, through properties of inherent mobility, facilitate high frequencies of integration into the recipient chromosomes. In other organisms, this approach has involved the use, for example, of retroviruses, retrotransposons and Ti plasmid infection. In Drosophila, however, although early consideration was given to the use of retrotransposons, such as copia (whose low mobility makes it a poor candidate for a vector), the primary route was through the exploitation of P elements. Drosophila melanogaster has, in fact, yielded several transposable elements which may be used in ways similar to those just described. At least three families of elements are known to generate hybrid dysgenic phenomena (P, I and hobo elements), and the hobo element itself has been used as a transformation vector (Blackman et al. 1989). Other Drosophila transposable elements which are not associated with dysgenic phenomena have also been used as transformation vectors. These include the mariner element from Drosophila mauritiana (Garza et al. 1991) and the minos element from Drosophila hydei (Franz & Savakis 1991), both of which have been used to transform Drosophila melanogaster. The mariner element is known now to be widespread (Robertson 1993), but it is not clear whether its transposition is phylogenetically restricted. Clearly, however, the study of transposable genetic elements, and in particular the P element, has done much to advance the cause of insect molecular genetics. 2.2.2 Germ-line transformation In essence, the procedure now routinely employed in Drosophila is a simple one. The cloned gene of interest is incorporated into a transformation vector, which carries a mobilization P element sequence. This resulting plasmid is then introduced into a preblastoderm-stage embryo of an appropriate host strain. DNA is introduced by microinjection into the posterior of the embryo prior to, and at the site of, pole cell formation. These germ-line precursor cells can then take up the transforming DNA, which, if stably integrated into the germ-line chromosomes by means of a transposition event, will be subject to normal hereditary transmission through subsequent generations. The identification of such transformed progeny is simpler if the transformation vector also carries an appropriate selectable marker. This is important because, even with P-mediated transformation in Drosophila, stable transformants are relatively rare among the progeny of adults arising

Transgenic insects

25

from microinjected embryos. An appropriate selectable marker therefore allows the rare transgenic offspring to be distinguished from their 'normal' or untransformed siblings. In order to maximize the efficiency of the procedure, however, appropriate consideration must be given to the characterizations of the host strain, the collection and preparation of embryos, the microinjection procedure itself and the characteristics of the transformation vector. What follows is an overview of these procedures, leading to a consideration of the applications of germ-line transformation techniques in Drosophila genetics. 2.2.2.1 Host strain characteristics The primary consideration is that the host recipient genome be free of P elements and therefore classified as an M strain within the P-M hybrid dysgenic nomenclature. This will facilitate the high-frequency transposition of P elements introduced into the germ-line through microinjection. M strains have never encountered P elements and have not involved mechanisms to regulate P element transposition. This is not true of P strains, which carry functional (intact) P elements and which have efficient regulation mechanisms that, under normal circumstances, reduce transposition effectively to zero. The presence of even a single functional P element in the host strain can seriously limit the efficiency of P-mediated transformation. There are also strains, designated M', which carry multiple internally deleted copies of the P element but no intact transposons. Host strains of this kind have been transformed using P element vectors but, as is always the case, efficiency must be determined empirically (Engels 1989). In general, it is recommended that the genomic DNA of potential host strains should be analysed by Southern blotting to verify the absence of P homologous sequences. This is especially true of laboratory strains which have been maintained as large mass-mated populations. Beyond this consideration, little research has been carried out on the effect of different genetic backgrounds or chromosomal rearrangements on transformation efficiency. It is wise to check the overall vigour of the potential host strain with simple measures of fecundity and fertility before embarking on transformation studies. 2.2.2.2 Collection of embryos The principal aim is to collect a sufficient quantity of synchronous embryos so that they can be prepared for microinjection before pole

26

Julian M. Crampton and Paul Eggleston

cell development, which, at 25°C, occurs 80-90 min after oviposition. A variety of techniques exist for the collection of embryos, and most Drosophila laboratories favour some over others. Most involve a laying chamber of some kind with replaceable collection surfaces. These might be small petri dishes containing agar or Drosophila growth medium or plastic partitions (which can be placed into standard glass vials) covered with thickened starch paste. The laying surface should be firm enough for easy embryo retrieval but soft and moist enough to encourage embryo deposition. It can be helpful to include a dark colourant (charcoal powder or food dye is ideal) to make embryos more visible. The number of flies transferred to the laying chamber will depend on the number of embryos required but is usually several hundred. These should be kept at low density to avoid retention of embryos in utero before transferring to the laying chamber. Freely laying females around 4-8 days old should be used. In all cases an oviposition stimulus, such as live yeast paste, is required. This can be a small central patch or a petri dish or a supply at the base of a vial into which plastic partitions are inserted. Where possible, the yeast paste should be kept away from the immediate oviposition surface, as it can make the embryos hard to see. Laying periods of around 30 min generally provide sufficient embryos of the appropriate age for microinjection. 2.2.2.3 Preparation of embryos Drosophila embryos are covered by two protective layers, the outer chorion (which is opaque) and the inner, transparent, vitelline membrane. Before microinjection, the chorion must be removed either by rolling on double-sided sticky tape or chemically in a sodium hypochlorite (bleach) solution. The time for chemical treatment depends on bleach concentration and must be determined empirically. Once the chorion is removed, the shiny vitelline membrane will be visible and the embryos become hydrophobic. Dechorionated embryos can then be arranged, with their posterior tips aligned, on 1-mm-wide strips of double-sided sticky tape held on a coverslip. With practice and experience, the researcher can align embryos of the correct age (under 90 min) in groups of 20-30 for injection before pole cell formation. Each group of embryos must be desiccated for the appropriate period of time in a chamber containing silica crystals. Desiccation times may vary from 5 to 20 min depending on

Transgenic insects

27

environmental conditions (longer times following chemical dechorionation). This stage is important for the efficiency of the overall procedure and can be determined only by practice. Overdesiccation results in visibly wrinkled embryos which die, even though they might take up large amounts of transforming DNA. Underdesiccated embryos have high turgidity and leak excessive quantities of cytoplasm during and after injection, again resulting in embryo death. Once appropriately desiccated, the embryos are covered with water-saturated halocarbon oil, which prevents further drying but which facilitates normal gaseous exchange and development. 2.2.2.4 Microinjection DNA for microinjection should comprise supercoiled plasmids, purified on caesium chloride gradients. The DNA concentration is not critical and generally falls within the range of 200-500 /x,g/ml for the vector and 50-100 jug/ml for the helper (Spradling & Rubin 1982). Injection needles are pulled from glass capillary tubing of around 1 mm outside diameter, and these may be siliconized to increase their usable life. Ideal needle-pulling conditions will vary with the equipment used, but the aim is a needle which tapers rapidly to increase rigidity and with a tip of around 2-5 jiim. The tip can be further prepared by either controlled breakage or grinding to produce a sharp bevel. Needles can be filled either by sucking up the DNA solution from a drop in the halocarbon oil or by backfilling with a drawn-out capillary. It can then be attached to the tool holder of a micromanipulator (which need not be excessively sophisticated) to control the injection process itself. Various mechanisms may be used for microinjection. The simplest is to connect the needle to a 10-ml syringe with air-filled tubing and use the syringe plunger to expel the DNA solution into the embryo. However, either water- or oil-filled delivery systems may be used, in some cases in conjunction with a gas, pressure-regulated, delivery mechanism. If the embryos are appropriately desiccated, about 500 pi (corresponding to 1-2% of the embryo volume) can be injected posteriorly, through the vitelline membrane, and seen directly to enter the embryo under the compound microscope. After the needle is introduced, it can be withdrawn slightly to allow the DNA to be delivered into the polar cytoplasm at the very posterior tip of the embryo, where the pole cells which yield the germ-line tissue will bud.

28

Julian M. Crampton and Paul Eggleston

As the procedure has become more routine, the necessity of such precise delivery has been questioned, but it may improve overall efficiency. The microinjection itself is more easily accomplished by using the micromanipulator to bring the needle into alignment with the embryos, but the microscope stage to impale the embryo onto the needle tip. After microinjection, the embryos (still covered with halocarbon oil) are maintained in a moist environment until they hatch. Larvae can then be transferred to normal growth medium. Once the adults emerge, they can be crossed to males or females of the host strain (as appropriate) and the progeny screened for transformants. 2.2.2.5 Transformation vector constructs Wild-type P strains of Drosophila melanogaster generally contain around 50 copies of the P element per haploid genome. Only a proportion of these, however, are intact, functional, 2.9-kb P elements. The remainder are internally deleted and unable to express their own transposase enzyme. They can be mobilized, however, if they are supplied with an active transposase in trans, either by a resident intact P element or by a functional P element which is transformed into that genome. Transposition involves an interaction between the transposase enzyme and the 31-bp inverted terminal repeats (ITRs) of the P element. The transposase itself appears to bind to sequences internal to the ITRs, and there is evidence that a second protein not encoded by the P element, and known as the ITR binding protein, is responsible for binding to the ITRs (Rio & Rubin 1988). The ITR binding protein has now been purified and partially sequenced, and the cDNA which encodes it isolated (A. Handler, pers. commun.). It remains to be seen whether the incorporation of this accessory gene into P element vectors will facilitate P element transposition outside of the Drosophilidae. Whatever the precise nature of the transposition mechanism, the ITRs are an absolute requirement for mobility. Their interaction with tranposase and ITR binding protein mediates excision of the P sequence from one genomic location and its integration at another. Similarly, it can mediate the excision of a P sequence from a plasmid construct, such as that introduced during a microinjection experiment, and then subsequent integration of this P sequence into a recipient or host genome. This was the basis of the first P-mediated transformation experiments reported by Rubin and Spradling (1982). They showed

Transgenic insects

29

that a suitably engineered plasmid carrying a P element could be used as a vector to carry any cloned sequence into a recipient genome, following microinjection into developing preblastoderm embryos. In this way, the basic strategy for all subsequent Drosophila transformation studies was established. The requirements for a P transformation vector are threefold: 1. The necessary P element machinery to mediate excision and reintegration, that is, intact P-element-inverted terminal repeats and a functional P element transposase coding sequence. The transposase can, of course, be supplied in trans by existing genomic P elements, rather than forming part of the exogenous injected DNA. 2. A suitable polylinker with a variety of restriction sites, for the introduction of cloned DNA sequences into the vector. 3. A suitable dominant selectable marker for the identification of transformed individuals. In practice, the transposase provision function is invariably separated from the machinery for transposition proper. Thus, two engineered plasmids are generally co-injected into the embryo. The first is the transformation vector itself, which carries the inverted terminal repeats of the P element but not the transposase coding sequence. The polylinker is placed between the inverted terminal repeats, since this is where a cloned DNA sequence might be inserted, and all sequences carried within the ITRs will be integrated into the recipient genome. The dominant selectable marker, along with any necessary controlling sequences, is also carried within the ITRs. Functional transposase is supplied in trans to this transformation vector by a 'helper element' which carries all of the transposase coding sequence but is rendered incapable of transposition itself by deletions to one ITR or both. This separation of 'vector' and 'helper' functions means that vector sequences can be integrated, generally as single copies, into the recipient genome, where they will remain stable, unless remobilized either by a further microinjection of 'helper' elements or purified P transposase (Kaufman & Rio 1991) or by the introduction of active P elements into the genome by way of a conventional cross. In principle, vector and 'helper' functions can be combined, but any introduced sequence would be capable of autonomous transposition and therefore unstable. The first such vectors to be prepared were the Carnegie vectors (Rubin & Spradling 1983). The vectors Carnegie 1, 2, 3 and 4 were

30

Julian M. Crampton and Paul Eggleston

identical except for the provision of a different set of restriction sites in each polylinker. Each was based on the plasmid pUC8 into which a nonautonomous P element, isolated from the white gene, was inserted. After further modifications to the poly linkers, the resulting vectors were 3.6 kb in length. None of these four vectors, however, carries a selectable marker for the identification of transformed individuals. This was remedied by the construction of Carnegie 20 (Rubin & Spradling 1983), which is based on Carnegie 2 but which carries the wild-type coding sequence of the rosy eye colour gene. The advantage of this is that when rosy mutants are microinjected with Carnegie 20, transformed individuals are identified by the restoration of wild-type eye colour. The appearance of individuals with wild-type eye colouration among the progeny of adults arising from injected rosy mutant embryos therefore indicates normal Mendelian transmission of the vector sequence - that is, a germ-line transformation event. In fact, because only a fraction of the wild-type level of rosy+ activity is required to complement the rosy~ mutation (Spradling 1986), a proportion of the adults arising directly from injected embryos may exhibit Go expression, which restores the wild-type phenotype. All evidence suggests that this is due to transient expression from the introduced plasmid rather than from integration into somatic chromosomes. A variety of other P-element-based transformation vectors are now available, all of which operate along lines similar to those already described. They include vectors with selectable markers based on white (Klemenz et al. 1987; Pirrotta 1988), Adh (Goldberg et al. 1983), yellow (Smith & Corces 1991), rough (Lockett et al. 1992) and vermilion genes (Fridell & Searles 1991) and cuticle pigmentation (Patton et al. 1992). An additional advantage of the white selectable marker is the intermediate level of expression generated at most integration sites. This makes the establishment of lines homozygous for the vector sequence easier, since they usually have eyes that are significantly more pigmented than those of heterozygotes. If the introduction of very large cloned sequences is required, there are cosmid-based P transformation vectors available (Haenlin et al. 1985). However, it should be realized that there is generally an inverse relationship between transformation vector size and efficiency. All of the vectors described here require the availability of defined mutant host strains (e.g., rosy~, white~, Adh~ or yellow') with transformants identified on the basis of reversion to a wild-type phenotype.

Transgenie insects

31

In 1985, Steller and Pirrotta introduced the pUChsneo vector, which has a dominant selectable marker gene (the bacterial neomycin resistance gene from Tn5) that circumvents this requirement. This gene codes for a monophosphotransferase which inactivates G418, neomycin, kanamycin and gentamycin. Most protocols involving this vector employ G418 as the selective agent, since it appears to be universally toxic among eukaryotic cells. The pUChsneo vector is smaller than the other vectors, and its construction offers certain advantages during subsequent analysis of transformants. However, care has to be exercised in determining the appropriate concentration of G418 during selection; this toxin can often slow development. There may also be problems with repeated heat treatment of the embryos and larvae to induce the heat shock promoter (hsp70) which controls expression of the neo gene. However, the basal expression from hsp70, without heat shocking, may be sufficient in many cases to render selection effective, and alternative promoters are, of course, available for driving neo expression. More recently, transformation vector constructs based on the hobo transposable element have been established (Blackman & Gelbart 1989; Blackman et al. 1989). These constructs also employ the rosy+ selectable marker and appear to mediate efficient germ-line integration at frequencies similar to those observed with the P element. Some consideration should be given to the choice of 'helper' element which supplies the active transposase. Almost any intact P element would suffice, but a helper which is disabled through deletions of the 31-bp ITRs is a better choice. Elements such as PTT25.7WC, where the we (or 'wings clipped') symbol designates a deletion of the 3' ITR (Karess & Rubin 1984), have been described in detail. Similar helpers include phsTr (Steller & Pirrotta 1986), in which the transposase gene is expressed under the control of the hsp70 heat shock promoter. The use of a stable (immobile) genomic transposase source, rather than a co-injected helper element, has also been reported (Robertson et al. 1988). One particular element, P [ry+A2-3](99B), has proved to be a particularly efficient source of transposase. In summary, virtually all of the parameters described here, as well as the skill of the individual researcher, can affect the efficiency of transformation. Not all embryos survive desiccation and injection to hatch as larvae. Of those that do, some will have suffered injection damage and may not yield fertile adults. Finally, not all fertile survivors will yield transformed progeny. In addition, the proportion of

32

Julian M. Crampton and Paul Eggleston

transformed progeny will vary depending on when, during the proliferation of germ-cells, integration of the vector takes place. As a guide to the number of embryos which might have to be injected, Pirrotta (1988) suggests 100-200 for smaller (plasmid-based) vectors, but perhaps 500-1,000 for cosmid vectors. However, this procedure is now routine in all major Drosophila laboratories and generally presents few problems in experienced hands. 2.2.2.6 Experimental design and analysis Fertile adults arising from microinjected embryos (Go adults) are generally mated to two or three host strain partners, and the progeny (Gj generation) are screened for transformants as appropriate to the selectable marker involved. Transmission of an altered phenotype through the Go gametes to the Gx progeny is indicative of a germ-line transformation event. It should be noted, however, that, with certain selectable markers, transient expression from the plasmid construct may cause an altered phenotype directly in the surviving Go adults. With the pUChsneo vector, it would be inappropriate to select with G418 before the Gx generation, as Go somatic mosaics would not be expected to survive. Putative transformed Gj adults are mated individually to host strain partners. DNA from these G2 lines can be analysed to determine the molecular basis of the transformation, and the chromosomal location of the insertion can be determined by in situ hybridization. In subsequent generations, particular integration events can be maintained against appropriate balancer chromosomes using standard techniques. With an efficient vector-helper system, those sequences maintained between the ITRs of the vector are usually integrated as a single copy. This can be virtually anywhere in the euchromatic part of the genome, and there is little evidence for 'hot spots' or preferential sites. Heterochromatic integration sites are also found, though perhaps more rarely (Karpen & Spradling 1992), and the presence of homologous selectable marker genes does not appear to precipitate integration at homologous sites. In general, transgenic lines maintain the normal qualitative expression of the introduced gene, including appropriate temporal and spatial activity. This is particularly true if reasonable 5' and 3' flanking sequences are included as a buffer in the design of the transformation vector. The experimenter should be wary, however, of 'position ef-

Transgenic insects

33

fects', which can alter levels of expression of the introduced sequence depending on the site of integration (Engels 1989). However, at least two recently described methods specifically buffer transgenes from position effects. These include the use of specialized chromatin structures (Kellum & Schedl 1991) and the su(Hw) suppressor of Hairywing protein (Roseman et al. 1993).

2.2.3 Applications of transgenic technology to Drosophila genetics 2.2.3.1 Insertional mutagenesis Mutations are a fundamental requirement for most research in genetics, and P elements have played a substantial role in the generation of new variability through novel alleles at many loci. In a strict sense, germ-line transformation is not a prerequisite for using P elements in the generation of new mutations, but there are a number of ways in which it can be used to optimize specific procedures. In order to consider the generation ofP element insertion mutations, it can be helpful to think in terms of primary and secondary mutagenesis. Primary mutagenesis involves the integration of a P element into a novel site which was previously devoid of P sequences. If this occurs at, or near, the coding or controlling regions of a gene, it may result in an altered phenotype. Secondary mutagenesis involves the remobilization of an existing P element at a given site. In this case, the excision of the P sequence is not always precise, or there may be chromosomal rearrangements, inversions or deletions, which again result in an altered or mutant phenotype. In both situations, hybrid dysgenesis can be expected to generate high mutation rates, and there is the considerable advantage that the extent of dysgenesis is under genetical control (by controlling mating and/or developmental temperature). There may also, however, be disadvantages with such systems. The nature of P-mediated mutations precludes base-pair substitutions, and there is some evidence for sitespecificity such that certain loci may be intractable. Both primary and secondary P element mutagenesis can be induced by a normal dysgenic cross in which a P strain male is mated to an M strain female. However, the major disadvantage with the use of natural P strains is the large number of independent P elements generally involved. This can make it difficult to identify the P element associated with the new mutation, and it can be difficult to stabilize the new

34

Julian M. Crampton and Paul Eggleston

mutation when there is potential for further transposition. One way of avoiding the presence of unwanted 'natural' P elements is to use transgenic strains which have been transformed with a single P sequence of known potential. The most useful of these are transgenic strains carrying strong, but nonmobile P transposase sources which are able to generate high levels of transposition but which are themselves stable. The most fully characterized of these transposase sources is P [ry+A2-3](99B), which generates high transposase levels in both germ-line and somatic tissue (Robertson et al. 1988), and the Jumpstarter element (Cooley et al. 1988), which makes transposase only in the germ-line but which carries no selectable marker. These transposase source elements can be combined with nonautonomous (deficient) P elements, either through appropriate genetic crosses or by germ-line transformation, to give exquisite control over the insertional mutagenesis procedure. By appropriate crosses, the insertion mutation can be isolated in a stable, transposase-free background and confirmed as a P element mutation either by in situ hybridization to the polytene chromosomes or by reversion rates in a dysgenic background. 2.2.3.2 Transposon tagging The biology of P elements renders them particularly useful for gene cloning experiments (Bingham et al. 1981; Searles et al. 1982). The initial step is the generation of a P element insertion mutation in the gene of interest, either through 'natural' dysgenic crosses or through germ-line transformation, as already described. In such a situation DNA sequences of the gene being sought will flank the P element, and a probe consisting of P element DNA can therefore be used as a 'handle' or tag with which to retrieve the desired gene. The procedure is relatively simple. The target is first confirmed as a P element insertion mutation (see Section 2.2.3.1); then a gene library of the mutant strain is constructed and screened with a P element probe. Positively hybridizing clones should carry flanking DNA of the gene of interest, although this will not always be the case if extraneous P elements are involved in the experiment. This highlights the importance of minimizing the number of such P elements involved and perhaps argues against using 'natural' P strains in dysgenic crosses. In any case, it is useful to screen positively hybridizing clones isolated from a transposon tagging experiment by in situ hybridization to M strain polytene

Transgenic insects

35

chromosomes. Only those which carry flanking DNA from the target locus are, of course, of interest (Engels 1989). Recent variations of this technique are described by Kaiser and Goodwin (1990) and Ballinger and Benzer (1989), in which insertions of P sequences next to any previously cloned DNA can be identified. DNA from flies subjected to P element insertional mutagenesis is subjected to a polymerase chain reaction (PCR) in which one primer is located in the target sequence and one in the P element. Such a PCR strategy does not strictly require an altered phenotype - that is, a P element insertion within the target gene or its controlling sequences. It is sufficient that the insertion lies close enough to the target gene for the intervening sequence to be amplified. Should mutations in the target gene be required, they may arise from secondary P element mutagenesis if the insert is remobilized. 2.2.3.3 Enhancer trapping The techniques described so far have also been used to identify genes with particular temporal or spatial expression patterns. Such specificity is often determined by cis-acting regulatory elements, or enhancers, which may operate over considerable distances (Finnegan 1992). P element transformation vectors have now been developed which carry the lac Z gene from Escherichia coli under the control of a weak promoter. When such vectors integrate into a recipient chromosome at a site which is under the sphere of influence of an enhancer, the lac Z gene will be expressed in the tissue or at the developmental stage for which the enhancer is active (O'Kane & Gehring 1987; Bellen et al. 1989; Wilson et al. 1989). In such transgenic strains, the enhancer-trap vector is often adjacent to sequences which are also regulated by the enhancer, and these can therefore be retrieved by the tagging protocols described earlier. Enhancer trapping is therefore a powerful technique for isolating genes with tissue-specific expression patterns. 2.2.3.4 Cell ablation studies Cell ablation is a powerful technique for studying the origin and fate of specific cell types. Although physical ablation techniques have previously been described, transgenic technology now provides alternative genetic ablation techniques. One such system described for Dro-

36

Julian M. Crampton and Paul Eggleston

sophila involves a temperature-sensitive toxin (ricin A chain), which produced ablation of eye cells when expressed under the control of the eye-specific sevenless enhancer (Moffat et al. 1992). At 29°C cell death occurred rapidly in the developing eye, whereas at 18°C very little toxicity was observed. An alternative toxin-based cell ablation system which employed the diphtheria toxin A-chain with expression controlled by the photorecepter cell-specific chaoptic promoter has been described by Kunes and Steller (1991). 2.3 Non-drosophilid insects 2.3.1

Introduction

As indicated in the first part of this chapter, dramatic advances have been made in the genetic manipulation of the genome of the fruit fly, Drosophila melanogaster. It seems an appropriate time to consider how this technology may be applied to economically important insects, including those of direct commercial value, such as the honeybee and silk moth, and insects that have a profound impact on agriculture and human health. Perhaps surprisingly, this area of research has, until recently, attracted scant attention. The remainder of this chapter will consider what has been done to develop methods for manipulating insect populations of economic significance and will highlight the potential that this research may have for the future. The majority of work to develop transgenic techniques in non-drosophilids has so far been focused on mosquito vectors of disease (Fig. 2.1). For this reason this section will concentrate on the developments in the field of transgenic mosquitoes, although results obtained with other insects will be mentioned in the relevant sections. 2.3.2 The organization and complexity of non-drosophilid genomes If the genomes of insects are to be manipulated in a controlled and directed fashion, it is important to determine the size of the genomes involved. Genome organization - that is, the nature and dispersion pattern of repetitive sequences and how they are organized in relation to the coding sequences - is also important. This is because it will have a profound influence on the types of manipulation which can be envisaged and the approaches to be adopted in order to identify and to clone sequences of interest. Until recently, very little was known about the size and organization of many insect genomes, including

Transgenic insects

37

Fig. 2.1. An Aedes aegypti mosquito shortly after a blood meal. those of mosquitoes. The DNA of all higher eukaryotes is conveniently subdivided into three components, namely highly repetitive, moderately repetitive and unique or single-copy sequences, and the organization of these components can be determined by experiments which involve denaturing the DNA and measuring the reassociation of complementary strands over time. Generally, highly repetitive sequences reassociate most quickly since there are so many copies, whereas unique sequences reassociate more slowly. The complexities of these components and of the total genome are determined by comparing the reassociation values with that of a single-copy sequence of known complexity, namely E. coli DNA. Of equal importance, however, is the way in which the different components are distributed throughout the genome. Black and Rai (1988) suggest that the existence of two basic organization patterns throughout all higher eukaryotes is indicative of a set of rules governing the establishment and spread of repetitive elements. The first pattern is known as shortperiod interspersion (SPI), in which 1- to 2-kb segments of single-copy sequence alternate regularly with short (0.2- to 0.6-kb) or moderately long (1- to 4-kb) repetitive sequences. This pattern is characteristic of the majority of animal species. The second pattern, long-period interspersion (LPI), is characterized by long (5- to 6-kb) repetitive

38

Julian M. Crampton and Paul Eggleston

sequences alternating with very long (>13-kb) uninterrupted stretches of unique-sequence DNA. Clearly, the evolution of such organization is an interesting phenomenon in itself, and the family Culicidae (to which the mosquitoes belong) may be of particular interest. According to Black and Rai (1988), this is the only family so far shown to contain species exhibiting both patterns of organization, and it may therefore explain how the transition from one to the other occurs. Black and Rai (1988) have analysed the genomic DNA from four species of mosquito, namely Anopheles quadrimaculatus, Culex pipiens, Aedes albopictus and Aedes triseriatus. More recently, Cockburn and Mitchell (1989) have shown that the genomes of anopheline mosquitoes are generally relatively small and exhibit the LPI pattern of repetitive sequences. For example, the genome of Anopheles gambiae, the major mosquito vector of malaria in Africa, has a genome size of 2.7 x 108 bp (Besansky & Powell 1992). This is in marked contrast to Aedes aegypti, which has the more complex SPI pattern (Warren & Crampton 1991). The Aedes aegypti genome is also relatively large (8.3 x 108 bp), being five times the size of the Drosophila genome, or, to place this in context, one-third the size of the human genome. Thus, almost every clone isolated from Aedes aegypti genomic libraries will contain repetitive DNA, and this may well mask the hybridization characteristics of sequences of interest (particularly single-copy sequences) when complex genomic probes are used. This is in direct contrast to organisms like Drosophila and Anopheles, which display the LPI pattern, where cloned sequences are more likely to consist entirely of either repetitive or unique DNA. In insect studies to date, Anopheles quadrimaculatus, Drosophila melanogaster, Apis mellifera, Sarcophaga bullata and Chironomous tentans have been shown to exhibit the LPI pattern. Conversely, Aedes aegypti, Culex pipiens, Aedes triseriatus, Aedes albopictus, Lucilia cuprina, Musca domestica, Bombyx mod and Antherea pernyi exhibit the SPI pattern (Black & Rai 1988; Crampton et al. 1990b; T. Howells, pers. commun.). Given that these species represent a wide range of genome sizes, there appears to be no simple relationship between genome size and organization. Finally, the molecular mapping of a number of insect genomes other than Drosophila has now been initiated. Foremost among these is the programme which aims to map the genome of the mosquito Anopheles gambiae. A preliminary, low-resolution map for part of this genome has been generated using PCR-amplified, microdissected segments

Transgenic insects

39

from polytene chromosomes as segment-specific probes and markers (Zheng et al. 1991). It is hoped that a complete molecular map will be generated which can then be correlated with the genetic map and the very detailed cytological map which is available for this insect. Such maps will be of considerable benefit for testing the mobility of potential DNA vector systems and for exploring the nature of transgene integration events and their long-term stability in insect genomes. 2.3.3 The early development of non-drosophilids in relation to embryo microinjection In relation to embryo microinjection, the early development of all insects is relatively similar in general terms to that of Drosophila. There may, however, be significant differences in developmental rate and embryo physiology. For example, mosquito embryos differ from those of Drosophila in having opaque, rigid chorions which cannot be removed without loss of viability. In addition, the mosquito embryo is extremely susceptible to dessication during the period at which injection must take place. 2.3.4 Methods for introducing DNA in the embryos and adults of non-drosophilids 2.3.4.1 Microinjection of embryos Here we describe the system developed in our own laboratory for the transformation of the mosquito Aedes aegypti (Morris et al. 1989). Similar techniques have been used elsewhere both for Anopheles and for other Aedes species, and all are based in general on the methods developed for Drosophila melanogaster. As already indicated, the rigid, opaque endochorion of the mosquito embryo cannot be removed, and the embryos are extremely sensitive to desiccation. However, glass capillaries with tips of 100-300 /mi x 4-10 ixm can be used to puncture the rigid endochorion without tearing it and deliver the DNA solution without damage to the embryo (Fig. 2.2). The slightly viscous DNA solution cannot be expelled manually from such a fine needle and is therefore injected by means of a two-phase nitrogen supply. The lower pressure prevents backflow and the higher pressure delivers 160-180 pi of DNA solution (corresponding to 1-5% of the embryo volume) into the posterior pole of the embryo at the syncitial blastoderm stage, before cell partitioning occurs. This is where the

40

Julian M. Crampton and Paul Eggleston

Fig. 2.2. Microinjection of Aedes aegypti embryos. pole cells, which are the germ-line primordia, develop. As indicated in Section 2.2.2.4, injection of DNA close to the site of pole cell formation is not critical to germ-line incorporation, but the timing is clearly important if the DNA is to be taken up by the developing germ-line cells. All of our injections are normally completed within 2 hr of oviposition. Both during and after injection, the embryos are covered with a water-saturated halocarbon oil, which permits the normal uptake of water until they are returned to standard insectary conditions. In this way DNA has been introduced into mosquito embryos (Miller et al. 1987; McGrane et al. 1988; Morris et al. 1989), with survival rates comparable to those obtained with Drosophila (Spradling & Rubin 1982). Similar methods have also been employed to introduce DNA into the embryos of the silk moth (Mahalingam et al. 1992), the med-fly (Malacrida et al. 1992), the sheep blow-fly (Atkinson et al. 1992) and the house-fly (Atkinson et al. 1992). 2.3.4.2 The biolistics technique The biolistic technique, which uses high-velocity, DNA-coated, tungsten microprojectiles to deliver DNA directly into the nuclei (Sandford et al. in press), has been successfully applied to introduce DNA into large numbers of dechorionated Drosophila embryos (Baldarelli &

Transgenie insects

41

Lengyel 1990). Transient expression from the P160 construct, a plasmid containing the actin 5C promoter fused to the E. coli /3-galactosidase gene, was detected. In addition, one germ-line transformation event was obtained when embryos were bombarded with the P7r25.7we helper and the P element vector 2pCasPeR in a ratio of 1:1.7. Once this method has been optimized, the advantage will be that a very large number of embryos (between 3,000 and 5,000) may be bombarded in one experiment, thus eliminating the need for the timeconsuming microinjection process. This will be particularly advantageous in a situation where microinjection survival rates are low and/or the microinjection process is difficult. These considerations led Mialhe et al. (1992) to test the biolistic method for gene transfer in Anopheles gambiae mosquitoes. In these experiments a construct made up of the luciferase gene under the control of the Drosophila heat shock promoter hsp70 was used as the reporter gene system. Transient expression of the reporter gene was detected, and experiments are currently under way to select for stable transformants using the biolistics method and a construct which incorporates a selectable marker. 2.3.4.3 Sperm-mediated transformation Considering reports of sperm-mediated gene transfer in rabbits, sea urchins, mice and chickens, some efforts have been directed towards determining whether insect sperm can transfer genes. The honeybee, Apis mellifera, has been used to examine this transformation technique. Honey-bee queens were inseminated with sperm that had been incubated with a 1-kb linear DNA fragment, and the DNA of Go progeny was assayed by the PCR (Milne 1992). Preliminary evidence indicates that DNA incubated with bee sperm was present in about 30% of the progeny, appeared to be in the genome, occurred in only a fraction of the cells and was probably not transferred by incorporation into the sperm genome. These preliminary results indicate that honey-bee sperm can transfer genes into the genome. This method may represent an attractive alternative to microinjection in gene transfer technology in terms of its ease, success rate and broad applicability to other insects, but to date it has not been attempted with other insects. This technique may work in insects for which instrumental insemination with sperm has not been perfected if a concentrated DNA-containing solution is placed in the reproductive tracts of females before natural mating to treat the

42

Julian M. Crampton and Paul Eggleston

sperm in vivo. However, it is clear that there is still a considerable amount of work to be done before sperm-mediated transformation can be considered an established technique for the creation of transgenic insects. 2.3.4.4 Transformation of symbionts as a means of introducing genes into insects Microinjection of embryos is the system for producing transgenic insects which has received the most attention to date. Another very interesting possibility is the idea of transforming symbionts isolated from the insect gut or ovary with gene constructs and reintroducing them into the insect vector. This approach has received particular attention in tsetse flies, where the gut symbionts have been isolated, cultured and transformed and shown to express the introduced DNA (Beard et al. 1993). The introduction of gene constructs into the insect via symbionts may be a very attractive alternative to direct genome manipulation, particularly in the case of gut symbionts, where they would be ideally placed for the expression of antiparasitic agents. Clearly, however, this form of manipulation would not be stable unless the symbionts were inherited transovarially. While this may be true for the rickettsia-like microorganisms present in the ovaries of tsetse flies and Culex mosquitoes, symbionts such as these would be difficult to reintroduce into the adult fly. Gut symbionts, however, while easy to isolate, transform and reintroduce into the insect midgut, nevertheless would prove unsatisfactory for introducing specific genes into large populations of insects, as they are unlikely to be passed on to subsequent generations. 2.3.5 DNA vector systems used in non-drosophilid transformation

The experimental design for the creation of transgenic insects, including mosquitoes, is based on that developed for Drosophila melanogaster (see Section 2.2.2.5). Go individuals which survive microinjection with the P element vector/helper DNA, pUChsneo/pUCs7r(A2-3), are mated inter se and allowed to produce progeny. These G j individuals are the first which might be expected to express antibiotic resistance throughout all tissues, and larvae are therefore subjected to selection with the neomycin derivative G418. The molecular nature of any transformation event is determined by DNA analysis using radioactively

Transgenic insects

43

labelled transformation vector DNA to probe Southern blots of genomic DNA extracted from the putative transformants and their progeny (Miller et al 1987; Morris et al 1989). Intact vector P elements have been detected in 5-10% of adults that have developed from injected embryos (Go), confirming that the introduced DNA is not immediately broken down by the mosquito. Furthermore, we have detected the chromosomal integration of vector DNA in several Go individuals. This probably reflects direct incorporation into a proportion of the somatic cell nuclei, since it is only in the following generation (Gj) that we might expect a germ-line integration event to have been transmitted to every nucleus. More promisingly, vector DNA has been identified in the chromosomes of the Gx and G2 progeny of injected embryos, suggesting that integration has occurred in the germ-line of the mosquito and that this DNA shows normal Mendelian inheritance. Some of these events, however, appear to be unstable from one generation to the next, and this phenomenon, together with the molecular basis of the transformation events, awaits further investigation. As indicated earlier, chromosomal integration of the introduced P element DNA has been observed in both Anopheles and Aedes mosquitoes. The integration events appear, in some cases, to be heritable and clearly to involve the germ-line of the transgenic mosquitoes (Miller et al. 1987; McGrane et al 1988; Morris et al 1989). Although these events did not result from normal P element transposition, some functional role of the P sequences cannot be ruled out. This is particularly true since similar experiments in Lucilia cuprina (Atkinson et al. 1992; T. Howells, pers. commun.) and Ceratitis capitata (Malacrida et al. 1992) have failed to produce any integration of vector sequences. Research in other laboratories is now being directed towards the identification of these accessory Drosophila proteins, and the cloning of the genes involved may facilitate high-efficiency P transposition in non-drosophilids. It is clear, however, from this work and from other experiments involving the transfection of the same DNA into cultured mosquito cells (Lycett et al. 1989) that the P element system in its present form is not suitable for routine use in the mosquito. Thus, while the means are currently available for introducing DNA into both mosquito embryonic germ-lines and cultured cells, a major stumbling block is the lack of an appropriate, high-efficiency, DNA transformation vector system for manipulating the mosquito genome. In addition to the use of the P-element-derived DNA vector systems, experiments to assess the mobility of a number of other trans-

44

Julian M. Crampton and Paul Eggleston

posable elements in non-drosophilid systems have been attempted. These experiments have included the use of fully processed cDNA copies of both the Ac and Spm elements from Zea mays (Comley et al. 1992), which transpose actively in a number of evolutionarily disparate organisms and may prove to act autonomously in mosquitoes (Kunze & Starlinger 1989). In addition, experiments involving the introduction of hobo, gypsy and mariner into Anopheles gambiae embryos are currently under way (K. Vernick, pers. commun.). 2.3.6 The search for alternative DNA vector systems for the transformation of non-drosophilids 2.3.6.1 The search for mobile genetic elements in non-drosophilid genomes It is clear that the germ-line integration events so far observed in mosquitoes do not involve normal P element transposition. The absence of this controlled mobility poses certain limitations - for example, with respect to transposon tagging for functional cloning (see Sections 2.2.3.2 and 2.3.7.1). Research elsewhere is concentrating on the precise mechanism of P transposition, and attempts are being made to modify the P element system for more general use (O'Brochta 1990). Such research may yet lead to the 'universal vectors' originally envisaged. At the same time, there remains the possibility that P elements may never function as efficient transposition-mediated transformation vectors in non-drosophilids. We and others are therefore actively searching for endogenous transposable elements which may yet prove to be the most suitable transformation vectors. The isolation of endogenous transposable genetic elements may ultimately be central to the development of efficient transformation and transposon tagging systems in mosquitoes. A number of approaches have been taken to identify such mobile elements. One of these was to analyse specific gene systems, such as the ribosomal DNA of mosquitoes, in an attempt to isolate variants of these genes which may have arisen from the insertion of transposons. No such insertions have as yet been identified in Aedes aegypti DNA (Gale & Crampton 1989), but insertion events have been detected in the rDNA of Anopheles gambiae and these elements are being fully defined (Paskewitz & Collins 1989). The elements appear to resemble a particular class of mobile element known as nonviral retroposons. It is unlikely, how-

Transgenic insects

45

ever, that these elements will prove useful as transformation vectors because of the ill-defined nature of their mode of transposition. We have adopted an alternative strategy to identify directly a specific class of mobile elements, known as retrotransposons, in the mosquito DNA. The approach relies on utilizing the characteristic biochemical and structural properties of these elements. This work has led to the successful isolation of several retrotransposon-like elements from the Aedes aegypti genome (Crampton et al. 1900a,b). More recently, we have used the PCR to develop a particularly rapid method for identifying endogenous retrotransposon-like elements in mosquito DNA (Warren & Crampton 1992). Others (Booth 1993) have used a similar approach to analyse the conserved reserve transcriptase motif from retrotransposon-like sequences in sand-fly vectors of leishmaniasis. A similar approach has also been employed to identify sequences related to the manner transposable element from Drosophila mauritiana in the genomes of a wide range of insect vector species (Robertson 1993). Sequences related to P and hobo have also been identified in Lucilia cuprina (Perkins & Howells 1992). Once such elements have been isolated and fully characterized, and their ability to transpose autonomously established, they may be engineered to form the core of a transformation vector system.

2.3.6.2 The FLPIFRT recombinase system in non-drosophilids Low-frequency illegitimate recombination events such as those isolated from previous attempts to transform mosquitoes could be utilized effectively if the integrated sequence served as a target for a heterologous high-frequency recombination system. Morris et al. (1991) have reported the activity in mosquito embryos of a yeast recombinase, FLP, acting on a specific target DNA sequence, FRT, isolated from the yeast 2-)Ltm plasmid. In a series of experiments, plasmids containing the FLP recombinase under the control of the Drosophila melanogaster hsp70 heat shock promoter were co-injected with target plasmids containing FRT sites. FLP-mediated recombination was detected between (a) target sites located on separate plasmids, resulting in the formation of dimers or higher-order multimers, and (b) target sites located on the dimers reformed in (a), leading to resolution of the dimers to their original monomeric forms. Synthetic FRT sites were also used and gave rise to results similar to those obtained using

46

Julian M. Crampton and Paul Eggleston

the FRT sites originally isolated from the yeast 2-/u,m plasmid. This successful demonstration of yeast FLP recombinase activity within the mosquito embryo suggests a possible future application of this system in establishing transformed lines of mosquitoes. 2.3.6.3 The search for gene systems with potential as phenotypic markers for incorporation into non-drosophilid transformation vectors A very important feature to be incorporated into a DNA transformation vector for use in mosquitoes is an improved selectable marker system for the identification of transformed individuals. A number of research laboratories have now reported problems with the existing neomycin resistance technique. In our experience, the main problem is the low activity of the neomycin phosphotransferase produced by the Tn5 neo gene in the transformation vector. The system is also less than satisfactory in that mosquitoes exhibit a spectrum of sensitivities to G418, and the survival of transformed mosquitoes relies on high levels of expression of the resistance gene. In addition, Aedes aegypti cells appear to have a tendency to retain intact vector plasmids which do not integrate into the recipient chromosomes but which transiently express antibiotic resistance. Alternative selectable markers, including the Tn903 neomycin resistance gene, which has been reported to show much higher phosphotransferase levels in yeast cells (Langhinrichs et al. 1989) and hygromycin resistance, together with alternative promoters to drive their expression, are now being investigated. A considerable body of work now suggests that the Drosophila heat shock promoter currently incorporated in the P element transformation vector constructs is not entirely satisfactory for driving the expression of selectable marker gene systems in mosquitoes (Lycett et al. 1992). Ultimately, therefore, cloned phenotypic markers available for use in transgenic mosquitoes, such as the eye colour mutations used routinely in Drosophila, would appear to offer the most efficient selection system for incorporation into a mosquito DNA transformation vector. Eye colour mutations exist in both Aedes and Anopheles, and recently the gene coding for white has been isolated and characterized from Anopheles gambiae (F. Collins, pers. commun.). In addition, a number of genes controlling eye colour have been cloned from the sheep blow-fly, Lucilia cuprina (Atkinson et al. 1992; Patterson et al. 1992).

Transgenic insects

47

2.3.6.4. The search for stage- and tissue-specific promoters for use in non-drosophilid transformation vectors At some point it will be desirable to express defined genes in mosquitoes and other insects in a tissue- or stage-specific fashion. For this to be envisaged, stage- and tissue-specific promoters have to be defined. None are as yet available, but attempts are being made to characterize the DNA sequences responsible for expressing certain genes, particularly in mosquito systems, by identifying genes which are expressed in a tissue-specific fashion and then defining the upstream, putative tissue-specific promoter sequence. One example of this approach has been the identification of an Aedes aegypti sequence which is expressed only in the female salivary gland (James et al. 1989). The expectation is, therefore, that this will make it possible to define a salivary gland-specific promoter sequence which may eventually allow the controlled expression of an introduced gene sequence in this tissue. More recently, trypsin genes have been cloned and characterized from Aedes aegypti (Barillas-Mury et al. 1991), the black-fly, Simulium vittatum (Ramos et al. 1993), and Anopheles gambiae (Muller et al. 1993). In each case, the expression of one or more of these trypsin genes is induced in the insect midgut by a blood meal. It is therefore likely that gut-specific, blood-meal-inducible promoters will shortly be available for each of these insects. Such promoters are clearly of interest, as they will allow the expression of antiparasitic agents in the insect gut when it takes a blood meal - that is, when the insect first comes in contact with the organisms which it can transmit to the human population. However, it will be necessary to develop methods for establishing the functionality of putative promoter sequences. To this end, we and others have begun to develop the methodologies for transfecting mosquito cells in culture. The simple, controlled environment of cultured cells enables one to follow the expression of cloned genes, and so delineate promoter and enhancer sequences. Genetically manipulated cell cultures are also able to overproduce specific proteins, which facilitates their isolation and purification. Cultured mosquito cells have been used to examine different transfection techniques and vectors and to help establish a suitable system for germ-line transformation of Aedes aegypti (Lycett 1990). Initially, experiments involved introducing the P element vector and helper constructs into several cell lines by a variety of techniques

48

Julian M. Crampton and Paul Eggleston

devised to generate transient cell membrane pores, including calcium phosphate precipitation (Wigler et al. 1977), dextran sulphate (Lopata et al. 1984), polybrene (Durbin & Fallon 1985), electroporation (Chu et al. 1987) and lipofection (Feigner et al. 1987). Much of this work has concentrated on the immortal Mos20 fibroblast cell line, which was derived from minced, trypsinized, neonate larvae of the Aedes aegypti London strain in 1969. Polybrene- and electroporationmediated transfection have proved to be most successful for these cells, producing approximately 30 and 4,000 transformants per 106 cells, respectively (Fig. 2.3). Subsequently, constructs incorporating the chloramphenicol acetyltransferase (CAT) reporter gene system has been utilized to optimize expression of the CAT gene under the control of the Drosophila heat shock promoter, hsp70, in the Mos20 mosquito cultured cells (Lycett et al. 1992). This type of approach, in addition to experiments utilizing a range of reporter genes and microinjection of embryos, will eventually make it possible to define functional constitutive, stage- or tissue-specific mosquito promoters. Table 2.1 provides a summary of the methods and constructs used in attempts to transform non-drosophilid embryos, organs and cells in culture. 2.3.7 The potential application of transgenic techniques in nondrosophilid insects

Once the systems necessary to create transgenic insects have been developed, how can this technology be applied? Two aspects will be discussed to illustrate the potential of the technology. The first deals with the use of the technique for analytical purposes, and the second with applying transgenic technology to economically significant insect populations. 2.3.7.1 Transgenic technology as an analytical tool The introduction and insertion of a mobile genetic element at or near a particular locus can cause that allele to mutate, producing a structural or developmental effect. As indicated in Sections 2.2.3.1 and 2.2.3.2, transposable genetic elements have been used as mutagens in the Drosophila system in order to clone genes or gene clusters of interest via transposon tagging (Bingham et al. 1981). This approach is an extremely powerful and immediate application of the technology, as it allows the cloning of genes purely on the basis of their mutant phe-

Transgenic insects

49

Fig. 2.3. Fluorescent in situ hybridization of a hspneo construct to metaphase chromosome preparations from an Aedes aegypti Mos20 cell line following stable transformation with the construct. The hspneo DNA can be seen integrated towards the telomeric region of one of the chromosomes. notype. It will therefore greatly extend our understanding of insect biology and, in the case of insect vectors of disease, considerably enhance our knowledge of the interactions between the insect and the disease-causing organisms they transmit to human and animal populations.

2.3.7.2 Transgenic technology in insects of economic significance In the long term, perhaps the most exciting applications of transgenic technology will be with insects of economic significance. Clearly, the type of manipulation to be envisaged will depend on the target insect and the scale of its economic impact. For example, insect pest populations have perhaps the most significance in terms of their commercial importance, and here the aim may be either to suppress or to control specific pest populations by transgenic means. Also in the agricultural sphere, there are an increasing number of beneficial insect populations. Such insects may be predators of pest insects and so form an important means of biological control. In this instance, the beneficial insects may be manipulated in such a way as to confer resistance to

Table 2.1. Summary of the methods and DNA constructs used to transform non-drosophilid embryos, organs and cells in culture

TARGET

DNA CONSTRUCT USED

METHOD OF INTRODUCTION

P element Microinjection

EMBRYO

Electroporation Biolistics ORGAN

Lipofection Virus

CELLS EN CULTURE

Cell translection

INSECT Mosquito Aedes

/

+/-

Ac

• ?

FLP recomb

• +

hsp/cat

• +

actin/cat

• +

hsp/neo

S +

hsp/luc

• +

Mosquito Anopheles

Medfly Ceratitis capitata

Sheep blowfly Lucitia cuprina

Silkworm moth Bombyx mori

Locust Locusta migratoria

Housefly Musca domestica

/ +/-

• -

/ -

• +

• -

• -

/ +

• -

• -

hobo

• -

/-

mariner

/ -

/-

/ +

Sindbis/cat

• +

hsp/neo

• +

hsp/cat

• +

hsp/luc

/ +

actin/cat

/ +

met/fi-gal

• +

actin/luc

• +

ElAlluc

• +

K74/cat

/ +

hsp/Ityg

• +

hsp/p-gal

• +

• +

• + • +

hsp/cat

hsp/luc

Cotton boll weevil Anthonomus grandts

/ +

Note: A J indicates that the method or construct has been tried in the relevant insect species; + indicates that the outcome was at least somewhat successful; - , unsuccessful; + / - , successful introduction of the element but not as the result of.

52

Julian M. Crampton and Paul Eggleston

insecticidal compounds which may be applied as part of an integrated pest management programme. Insects are also important vectors of disease to both humans and agricultural animals. In this way, vector populations may have a profound impact on the economy of regions which in many cases, particularly in the tropics, are the most fertile and potentially productive areas. Transgenic technology may eventually play a role in controlling vector-borne disease by providing the means to suppress vector populations by rendering them vulnerable to subsequent control measures, such as insecticide susceptibility, temperature sensitivity or ability to survive diapause. A second possibility, and perhaps a more exciting approach, would be to alter the ability of the insect to transmit the disease. Clearly, such possibilities are for the future, but it is quite feasible to consider genetic manipulation of insect populations of direct commercial value, such as the honey-bee or silk moth. Here transgenic technology may be employed to confer a number of beneficial characteristics on these insects to create novel and highly productive strains. For example, one might manipulate the insect to increase the yield of the product by increasing the growth rate or by enhancing the resistance of the insect to infection, temperature shock or other detrimental factors.

2.3.7.3 Potential target genes for manipulation Having discussed the types of manipulation of insect genomes which may be beneficial in economic or health terms, it is worth considering the types of gene systems which may be potential targets for manipulation to achieve these aims. In this respect, there are a number of obvious targets for manipulation, including the genes involved in the insect immune system, developmental control genes and insecticide resistance genes. Genes influencing all of these factors have been characterized at the molecular level for a number of insects, and it is now feasible to consider manipulating them in the germ-line of these insects. In addition, a number of genes are of particular interest because they are directly implicated in the ability of insects to transmit disease-causing organisms. Examples include the filarial susceptibility (f™) and Plasmodium susceptibility (pis) loci of the mosquito Aedes aegypti. The /m locus is genetically well defined and there are good data on its linkage relationships. Refractoriness to infection is due to a partially sex-linked, dominant gene (Macdonald & Ramachandran

Transgenic insects

53

1965). There is marked variation in the susceptibility of this mosquito to different filarial worms, although all of the alleles concerned map at about the same place on the sex chromosome. Also of particular interest is a strain of Anopheles gambiae which has been selected for refractoriness to the malaria parasite and characterized genetically (Collins et al. 1986). Attempts are currently under way to clone these genes, but it is difficult to undertake such a cloning exercise in the absence of any knowledge of the gene product. Recently, however, considerable progress has been made towards the molecular characterization of this gene, with the development of a linkage map for the X chromosome of Anopheles gambiae (Zheng et al. 1993). Once fully refined, this map should allow the genes influencing refractoriness to be cloned and fully characterized. Clearly, the use of transgenic technology through transposon tagging will assist in the characterization of refractory genes and their products. An important genotypic characteristic not met by the majority of genes encoding refractoriness is that any such gene introduced into the insect would have to be capable of altering the phenotype through the expression of a single gene copy. Unfortunately, at present, there is no gene or gene product defined at the molecular level which is known to affect phenotype directly in relation to pathogen development in, or transmission by, any insect. However, in the mosquito system a number of molecules affect the transmission of malaria by anophelines. Foremost among these are the so-called transmissionblocking vaccines, which can achieve a total transmission blockade (Winger et al. 1987). These vaccines attack antigens present on the gametes and ookinetes of the malaria parasite, and antibodies which recognize these antigens are able to block the development of the parasite in the mosquito midgut. A very exciting possibility, therefore, is to introduce the genes coding for such antibodies into the mosquito genome, thus directly conferring the transmission-blocking phenotype to the insect. In this case, a transgenic mosquito would be created incorporating an antibody gene which would be expressed in the insect midgut in response to a blood meal and which would therefore block the transmission of malaria. This type of approach is attractive for a number of reasons. It eliminates the need for the detailed molecular analysis of refractory mechanisms in mosquitoes and it would be a 'dominant' gene system (i.e., one gene copy only would be needed in each cell of the mosquito). The antigen target on the stage of the malaria parasite present

54

Julian M. Crampton and Paul Eggleston

in the mosquito is highly conserved, suggesting that the parasite may be less able to avoid this type of transmission control mechanism. Finally, the use of transgenic insects incorporating an antibody gene could be applied to any vector-transmitted pathogen (parasite or viral) where a target antigen can be identified as being inhibited by the expressed molecule. To date, mouse antibody genes have been cloned and introduced into mosquito cells in culture, and mouse Fab molecules have been expressed and detected using immunohistochemical staining techniques. Efforts are currently under way to clone the transmissionblocking antibody genes themselves and to assess their functionality in vivo. The trypsin genes from the mosquito Anopheles gambiae have recently been cloned (Muller et al. 1993). The expression of two of the genes has been found to be induced in the female mosquito midgut when it takes a blood meal. It may therefore be possible in the near future to create a mosquito expressing the transmission-blocking antibody genes in such a way as to block or disrupt the transmission of malaria. If researchers succeed in doing this, transgenic mosquitoes expressing antimalarial antibodies may represent a potential strategy for controlling malaria and may establish a precedent for a wide range of new antidisease strategies. 2.3.7.4 Transgenic insects in natural populations Once transgenic insects with the necessary characteristics have been created, there remains the question, what next? Clearly, if the manipulated insects are themselves to be cultivated for production, it may be possible to directly apply novel strains created by transgenic means. However, where this is not the case, it will be necessary to consider the problems likely to be faced in applying the technology to experimental and natural populations. It may well be that such a situation would disrupt the normal adaptive process and therefore be opposed by natural selection. If this is so, then some form of drive mechanism may be needed to force the desired gene through the population. This is not a new concept to those who have worked on the genetic control of insect populations. However, the testing of such mechanisms has been limited since, in reality, they have awaited the advent of recombinant DNA technology to provide the necessary raw material. Two types of drive mechanism have been suggested. One is meiotic drive, whereby a given chromosome is transmitted to more than the

Transgenic insects

55

expected 50% of offspring. Any desirable genes linked to the driven chromosome would eventually approach fixation, even with the release of relatively few individuals. There is experimental evidence to support the use of meiotic drive in Aedes aegypti. This mechanism, driven by the MD locus, has been used to force the marker gene re (red eye) into a laboratory cage population (Wood et al. 1977). Interestingly, meiotic drive may form part of the spectrum of hybrid dysgenic traits, and it might therefore be possible to exploit this phenomenon by using either the P element itself, or a mobile element with properties similar to those of P, as an efficient mechanism to drive a specific gene construct through an insect population. The second type of drive mechanism is the exploitation of genetic traits that reduce heterozygote fitness (Curtis & Graves 1988). For example, the gene to be driven could be introduced into a translocation chromosome such that viable and fertile homozygotes were formed, whereas heterozygotes would display reduced fertility or viability. In this way, translocation, pericentric inversions, interracial hybrid sterility, cytoplasmic incompatibility and compound chromosomes all have potential since, in each case, hybrids have reduced fitness. Such mechanisms require larger release numbers, since there is no exponential increase in the frequency of the driven chromosome as with meiotic drive. However, fixation of desirable genes would occur more quickly than with meiotic drive because of the reduced fitness of heterozygous combinations. Efficiency could be improved by providing the released individuals with some form of temporary advantage. For example, insecticide resistance could be incorporated into the genome and then insecticide applied (Whitten 1970). Ideally, the insecticide resistance gene would be fused to the desirable gene and introduced as a unit to prevent disruption of useful combinations by meiotic recombination. Certainly in the case of vector populations, the most useful end result of such programmes would be the progressive replacement rather than the eradication of disease-transmitting populations, since an emptied ecological niche might be colonized rapidly by migration of wild types. 2.4 Transgenic insects: the future

Eventually, embryo transformation will provide the raw material to test the proposed drive mechanisms in laboratory and natural populations. The questions posed by considering the release of transgenic insects emphasize the need to assess the biological consequences of

56

Julian M. Crampton and Paul Eggleston

such a release. It is difficult, however, to gauge the possible hazards of such a release in the absence of experimental evidence, and these ethical and safety considerations must be faced at an early stage. In order to undertake an informed appraisal where the possible net benefits may be balanced against the potential hazards, considerable effort will have to be devoted to utilizing caged populations and the controlled release of molecularly tagged individuals together with mathematical modelling of these populations. There is clearly some way to go before any release of transgenic insects can be considered. The power of the technology is, however, so enormous that it must be explored, and there is every indication that over the next few years the potential of transgenic technology in insects will be fully exploited.

Acknowledgements

The authors acknowledge the financial support provided by the Wolfson Foundation, Wellcome Trust, Lister Institute of Preventive Medicine, Medical Research Council, Liverpool University and the UNDP / World Bank / WHO Special Programme for Research and Training in Tropical Diseases. J.M.C. is a Wellcome Trust Senior Research Fellow in Basic Biomedical Sciences and P.E. is a Lister Institute Research Fellow. We are grateful to Michael Ashburner for his helpful comments on this manuscript.

References Atkinson, P. W., Coates, C. J., Howells, A. J. & O'Brochta, D. A. (1992). Prospects for nondrosophilid germ-line transformation. In Proceedings of the International Symposium on Management of Insect Pests: Nuclear and Related Molecular and Genetic Techniques, p. 11. Vienna: IAEA/ FAO Publication. Baldarelli, R. M. & Lengyel, J. A. (1990). Transient expression of DNA after ballistic introduction into Drosophila embryos. Nucl. Acids Res. 18, 5903-4. Ballinger, D. G. & Benzer, S. (1989). Targeted gene mutations in Drosophila. Proc. Natl. Acad. Sci. USA 86, 9402-6. Barillas-Mury, C , Graf, R., Hagedorn, H. H. & Wells, M. A. (1991). cDNA and deduced amino acid sequence of a blood meal-induced trypsin from the mosquito, Aedes aegypti. Insect Biochem. 21, 825-31. Beard, C. B., O'Neill, S. L., Mason, P., Mandelco, L., Woeses, C. R., Tesh, R. B., Richards, F. F. & Aksoy, S. (1993). Genetic transformation and phylogeny of bacterial symbionts from tsetse. Insect Mol. Biol. 1, 123— 32.

Transgenic insects

57

Bellen, H. J., O'Kane, C. J., Wilson, C , Grossniklaus, U., Pearson, R. K. & Gehring, W. J. (1989). P-Element-mediated enhancer detection: A versatile method to study development in Drosophila. Genes Dev. 3, 1288— 1300. Besansky, J. N. & Powell, J. R. (1992). Reassociation kinetics of Anopheles gambiae (Diptera: Culicidae) DNA. J. Med. Entomol. 29, 125-8. Bingham, P. M., Levis, R. & Rubin, G. M. (1981). The cloning of DNA sequences from the white locus of Drosophila melanogaster using a novel and general method. Cell 25, 693-704. Black, W. C. & Rai, K. S. (1988). Genome evolution in mosquitoes: Intraspecific and interspecific variation in repetitive DNA amounts and organisation. Genet. Res. (Camb.) 51, 185-95. Blackman, R. K. & Gelbart, W. M. (1989). The transposable element hobo of Drosophila melanogaster, In Mobile DNA, ed. D. Berg & M. Howe, pp. 523-9. Washington, DC: American Society for Microbiology. Blackman, R. K., Macy, M., Koehler, D., Grimaila, R. & Gelbart, W. M. (1989). Identification of a fully functional hobo transposable element and its use for germ line transformation of Drosophila. EMBO J. 8, 211-17. Booth, D. R. (1993). Retrotransposons in phlebotomine sandflies. Ph.D. thesis, University of London. Caspari, E. & Nawa, S. (1965). A method to demonstrate transformation in Ephestia. Z. Naturforsch. 206, 281-4. Chu G., Hayakawa, H. & Berg, P. (1987). Electroporation for the efficient transfection of mammalian cells with DNA. Nucl. Acids Res. 15, 1311— 26. Cockburn, A. F. & Mitchell, S. F. (1989). Repetitive DNA interspersion patterns in Diptera. Arch. Insect. Biochem. Physiol. 10, 105-13. Collins, A. F., Sakai, R. K., Vernick, K. D., Paskewitz, S., Seeley, D. C , Miller, L. H., Collins, W. E., Campbell, C. C. & Gwadz, R. W. (1986). Genetic selection of a Plasmodium refractory strain of the malaria vector Anopheles gambiae, Science 236, 607-10. Comley, I., Crampton, J. M. & Eggleston, P. (1992). In vivo heterologous gene expression in the mosquito Aedes aegypti. In Insect Molecular Science, ed. J. M. Crampton & P. Eggleston, pp. 252-3. London: Academic Press. Constantini, F. & Lacy, E. (1981). Introduction of a rabbit /3-globin gene into the mouse germ line. Nature 294, 92-4. Cooley, L., Kelley, R. & Spradling, A. C. (1988). Insertional mutagenesis of the Drosophila genome with single P elements. Science 239, 1121-8. Crampton, J. M., Morris, A. C , Lycett, G. J., Warren, A. M. & Eggleston, P. (1990a). Transgenic mosquitoes: A future vector control strategy. Parasitology Today 6: 31-6. Crampton, J. M., Morris, A. C , Lycett, G. J., Warren, A. M. & Eggleston, P. (1990b). Molecular characterisation and genome manipulation of the mosquito, Aedes aegypti. In Molecular Insect Science, ed. H. H. Hagedorn, J. G. Hildebrand, M. G. Kidwell & J. H. Law, pp. 1-11. Oxford: Plenum Press. Curtis, C. F. & Graves, P. M. (1988). Methods for replacement of malaria vector populations. J. Trop. Med. Hyg. 91, 43-8. Durbin, J. E. & Fallon, A. M. (1985). Transient expression of the chloramphenicol acetyl transferase gene in cultured mosquito cells. Gene 36, 173-8.

58

Julian M. Crampton and Paul Eggleston

Engels, W. R. (1983). The P family of transposable elements in Drosophila. Annu. Rev. Genet. 17, 315-44. Engels, W. R. (1989). P elements in Drosophila. In Mobile DNA, ed. D. Berg & M. Howe, pp. 437-84. Washington, DC: American Society for Microbiology. Feigner, P. L., Gadek, T. R., Holm, M., Roman, R., Chan, H. W., Wenz, M., Northrop, J. P., Ringhold, G. M. & Danielson, M. (1987). Lipofection: A highly efficient, lipid mediated DNA transfection procedure. Proc. Natl. Acad. Sci. USA 84, 7413-17. Finnegan. D. J. (1992). Transposable elements and their biological consequences in Drosophila and other insects. In Insect Molecular Science, ed. J. M. Crampton & P. Eggleston, pp. 35-48. London: Academic Press. Fox, A. S. & Yoon, S. B. (1966). Specific genetic effects of DNA in Drosophila melanogaster. Genetics. 53, 897-911. Fox, A. S. & Yoon, S. B. (1970). DNA induced transformation in Drosophila: Locus specificity and the establishment of transformed stocks. Proc. Natl. Acad. Sci. USA 67, 1608-15. Franz, G. & Savakis, C. (1991). Minos, a new transposable element from Drosophila hydei, is a member of the Tc-1-like family of transposons. Nucl. Acids Res. 19, 6646. Fridell, Y. W.-C. & Searles, L. L. (1991) Vermilion as a small selectable marker gene for Drosophila transformation. Nucl. Acids Res. 19, 5082. Gale, K. R. & Crampton, J. M. (1989). The ribosomal genes of the mosquito. Aedes aegypti. Eur. J. Biochem. 185, 311-17. Garza, D., Medhora, M., Koga, A. & Hartl, D. L. (1991). Introduction of the transposable element mariner into the germ line of Drosophila melanogaster. Genetics 128, 303-10. Germeraad, S. (1976). Genetic transformation in Drosophila by microinjection of DNA. Nature 262, 229-31. Goldberg, D., Posakony, J. & Maniatis, T. (1983). Correct developmental expression of a cloned alcohol dehydrogenase gene transduced into the Drosophila germ line. Cell 34, 59-73. Haenlin, M., Steller, H., Pirrotta, V. & Mohier, E. (1985). A 43 kilobase cosmid P transposon rescues the fs(l)K]° morphogenetic locus and three adjacent Drosophila developmental mutants. Cell 40, 827-37. Hiraizumi, Y. (1971). Spontaneous recombination in Drosophila melanogaster males. Proc. Natl. Acad. Sci. USA 68, 268-70. James, A. A., Blackmer, K. & Racioppi, J. V. (1989). A salivary glandspecific, maltase-like gene of the vector mosquito, Aedes aegypti. Gene 75, 73-83. Kaiser, K. & Goodwin, S. F. (1990). 'Site selected' transposon mutagenesis of Drosophila. Proc. Natl. Acad. Sci. USA 87, 1686-90. Karess, R. E. & Rubin, G. M. (1984). Analysis of P transposable element functions in Drosophila. Cell 38, 135-46. Karpen, G. H. & Spradling, A. C. (1992). Analysis of subtelomeric heterochromatin in the Drosophila minichromosome Dpi 187 by single P element insertional mutagenesis. Genetics 132, 737-53. Kaufman, P. D. & Rio, D. C. (1991). Germline transformation of Drosophila melanogaster by purified P element transposase. Nucl. Acids Res. 19, 6336.

Transgenic insects

59

Kellum, R. & Schedl, P. (1991). A position-effect assay for boundaries of higher order chromosomal domains. Cell 64, 941-50. Kidwell, M. G., Kidwell, J. F., & Sved, J. A. (1977). Hybrid dysgenesis in Drosophila melanogaster: A syndrome of aberrant traits including mutation, sterility and male recombination. Genetics 86, 813-33. Klemenz, R., Weber, U. & Gehring, W. J. (1987). The white gene as a marker in a new P element vector for gene transfer in Drosophila. Nucl. Acids Res. 15, 3947-59. Kunes, S. & Steller, H. (1991). Ablation of Drosophila photoreceptor cells by conditional expression of a toxin gene. Genes Dev. 5, 970-83. Kunze, R. & Starlinger, P. (1989). The putative transposase of transposable element Ac from Zea mays L interacts with sub-terminal sequences of Ac. EMBOJ. 8, 3177-85. Langhinrichs, C , Berndorff, D., Seefeldt, C. & Stahl, U. (1989). G418 resistance in the yeast Saccharomyces cerevisiae: Comparison of the neomycin resistance genes from Tn5 and Tn903. Appl. Microbiol. Biotechnol. 30, 388-94. Liu, C. P., Kreber, R. A. & Valencia, J. I. (1979). Effects on eye colour mediated by DNA injection into Drosophila embryos. Mol. Gen. Genet. 172, 203-10. Lockett, T. J., Lewy, D., Holmes, P., Medveczky, K. & Saint, R. (1992). The rough (ro + ) gene as a dominant P element marker in germ line transformation of Drosophila melanogaster. Gene 114, 187-93. Lopata, M. A., Cleveland, D. W. & Sollner-Webb, B. (1984). High level transient expression of a chloramphenicol acetyl transferase gene by DEAEdextran mediated DNA transaction coupled with a dimethyl sulfoxide or glycerol shock treatment. Nucl. Acids Res. 12, 5707-17. Lycett, G. J. (1990). DNA transfection of mosquito cells in culture. Insect. Mole. Genet. Newsletter 4, 1-3 (Royal Entomological Society of London Publication). Lycett, G. J., Eggleston, P. & Crampton, J. M. (1989). DNA transfection of an Aedes aegypti mosquito cell line. Heredity 63, 277. Lycett, G., Crampton, J. M. & Eggleston, P. (1992). The expression of reporter genes in cultured cells of Aedes aegypti. In Insect Molecular Science, ed. J. M. Crampton & P. Eggleston, pp. 247-8. London: Academic Press. Macdonald, W. W. & Ramachandran, C. P. (1965). The influence of the gene f" (filarial susceptibility, Brugia malayi) on the susceptibility of Aedes aegypti to seven strains of Brugia, Wuchereria and Dirofilaria. Ann. Trop. Med. Parasit. 59, 64-73. Mahalingam, S., Banno, Y. & Iatrou, K. (1992). Stable silkmoth embryo transformation by DNA injection. In Insect Molecular Science, ed. J. M. Crampton & P. Eggleston, p. 250. London: Academic Press. Malacrida, A., Gasperi, G., Baruffi, L., Torti, C. & Guglielmino, C. (1992). Basic studies for genetic manipulation in Ceratitis capitata Wied (Diptera, Tephritidae). In Insect Molecular Science, ed. J. M. Crampton & P. Eggleston, p. 256. London: Academic Press. McGrane, V., Carlson, J. O., Miller, B. R. & Beatty, B. J. (1988). Microinjection of DNA into Aedes triseriatus ova and detection of integration. Am. J. Trop. Med. Hyg. 39, 502-10. Mialhe, E., Laughinghouse, A. & Miller, L. (1992). Biolistic for gene transfer

60

Julian M. Crampton and Paul Eggleston

to mosquito embryos. In Proceedings of the International Symposium on Management of Insect Pests: Nuclear and Related Molecular and Genetic Techniques, pp. 18-19. Vienna: IAEA / FAO Publication. Miller, L. H., Sakai, R. K., Romans, P., Gwadz, R. W., Kantoff, P. & Caon, H. G. (1987). Stable integration and expression of a bacterial gene in the mosquito Anopheles gambiae. Science 237, 779-81. Milne, C. (1992). Sperm-mediated honey bee transformation. In Insect Molecular Science, ed. J. M. Crampton & P. Eggleston, p. 251. London: Academic Press. Moffat K. G., Gould, J. H., Smith, H. K. & O'Kane, C. J. (1992). Inducible cell ablation in Drosophila by cold sensitive ricin A chain. Development 114, 681-7. Morris, A. C , Eggleston, P. & Crampton, J. M. (1989). Genetic transformation of the mosquito Aedes aegypti by micro-injection of DNA. Med. Vet. EntomoL 3, 1-7. Morris, A. C , Schaub, T. L. & James, A. A. (1991). FLP-mediated recombination in the vector mosquito, Aedes aegypti. Nucl. Acids Res. 19, 5895900. Muller, H.-M., Crampton, J. M., della Torre, A., Sinden, R. & Crisanti, A. (1993). Members of a trypsin gene family in Anopheles gambiae are induced in the gut by blood meal. EMBO J. 12, 2891-2900. Nawa, S. & Yamada, S. (1968). Hereditary change in Ephestia after treatment with DNA. Genetics 58, 573-84. O'Brochta, D. A. (1990). Genetic transformation and its potential in insect pest control. Bull. Ent. Res. 80, 241-4. O'Kane, C. J. & Gehring, W. J. (1987). Detection in-situ of genomic regulatory elements in Drosophila. Proc. Natl. Acad. Sci. USA 84, 9123-7. Paskewitz, S. M. & Collins, F. H. (1989). Site-specific ribosomal DNA insertion elements in Anopheles gambiae and A. arabiensis: Nucleotide sequence of gene-element boundaries. Nucl. Acids Res. 17, 8125-33. Patterson, C , Elizur, A., Garcia, R., Perkins, R. & Howells, A. (1992). Cloning eye colour genes from Lucilia cuprina. In Insect Molecular Science, ed. J. M. Crampton & P. Eggleston, pp. 246-7. London: Academic Press. Patton, J. S., Gomes, X. V. & Geyer, P. K. (1992). Position independent germ line transformation in Drosophila using a cuticle pigmentation gene as a selectable marker. Nucl. Acids Res. 20, 5859-60. Perkins, H. & Howells, A. (1992). Genomic sequences with homology in the P element of Drosophila melanogaster occur in the blowfly Lucilia cuprina. Proc. Natl. Acad. Sci. USA 89, 10753-57. Pirrotta, V. (1988). Vectors for P-mediated transformation in Drosophila. In Vectors: A Survey of Molecular Cloning Vectors and Their Uses, ed. R. L. Rodriguez & D. T. Denhardt, pp. 437-56. Stoneham, MA: Butterworths. Ramos, A., Mahowald, A. & Jacobs-Lorena, M. (1993). Gut-specific genes from the black fly Simulium vittatum encoding trypsin-like and carboxypeptidase-like proteins. Insect Mol. Biol. 1, 149-64. Rio, D. C. & Rubin, G. M. (1988). Identification and purification of a. Drosophila protein that binds to the terminal 31-base-pair inverted repeats of the P transposable element. Proc. Natl. Acad. Sci. USA 85, 8929-33. Robertson, H. M. (1993). The mariner transposable element is widespread in nature. Nature 362, 241-5.

Trans genie insects

61

Robertson, H. M., Preston, C. R., Phillis, R. W., Johnson-Schlitz, D., Benz, W. K. & Engels, W. R. (1988). A stable genomic source of P element transposase in Drosophila melanogaster. Genetics 118, 461-70. Roseman, R. R., Pirrotta, V. & Geyer, P. K. (1993). The su(Hw) protein insulates expression of the Drosophila melanogaster white gene from chromosomal position effects. EMBO J. 12, 435-42. Rubin, G. M. & Spradling, A. C. (1982). Genetic transformation of Drosophila with transposable element vectors. Science 218, 348-53. Rubin, G. M. & Spradling, A. C. (1983). Vectors of P element mediated gene transfer in Drosophila. Nucl. Acids Res. 11, 6341-51. Sandford, J. C , Smith, F. D. & Russell, J. A. (in press). Optimizing the biolistic process for different biological applications. Meth. Enzymol. Searles, L. L., Jokerst, R. S., Bingham, P. M., Voelker, R. A. & Greenleaf, A. L. (1982). Molecular cloning of sequences from a Drosophila RNA polymerase II locus by P element transposon tagging. Cell 31, 585-92. Smith, P. A. & Corces, V. G. (1991). Drosophila transposable elements. Adv. Genet. 21, 229-300. Spradling, A. C. (1986). P element-mediated transformation. In Drosophila: A Practical Approach, ed. D. B. Roberts, pp. 175-97. Oxford: IRL Press. Spradling, A. C. & Rubin, G. M. (1982). Transposition of cloned P elements into Drosophila germ line chromosomes. Science 218, 341-7. Steller, H. & Pirrotta, V. (1985). Transposable P vector that confers selectable G418 resistance to Drosophila larvae. EMBO J 4, 167-71. Steller, H. & Pirrotta, V. (1986). P transposons controlled by the heat shock promoter. Mol. Cell. Biol. 6, 1640-9. Warren, A. M. & Crampton, J. M. (1991). The Aedes aegypti genome: Complexity and organisation. Genet. Res. (Camb.) 58, 225-32. Warren, A. M. & Crampton, J. M. (1992). Retrotransposon reverse transcriptase-like sequences in the genome of mosquitoes. In Insect Molecular Science, ed. J. M. Crampton & P. Eggleston, pp. 244-5. London: Academic Press. Whitten, M. J. (1970). Use of chromosome rearrangements for mosquito control. In The Sterility Principle for Insect Control or Eradication, International Atomic Energy Agency Symposium, pp. 399-410. Athens. Wigler, M., Siverstein, S., Lee, L. S., Pellicer, A., Cheng, Y. & Axel, R. (1977). Transfer of purified herpes virus thymidine kinase gene to cultured mouse cells. Cell 11, 223-32. Wilson, C , Pearson, R. K., Bellen, H. J., O'Kane, C. J., Grossniklaus, U. & Gehring, W. J. (1989). P-element-mediated enhancer detection: An efficient method for isolating and characterising developmentally regulated genes in Drosophila. Genes Dev. 3, 1301-13. Winger, L., Smith, J. E., Nicholas, J., Carter, E. H., Tirawanchai, N. & Sinden, R. E. (1987). Ookinete antigens of Plasmodium berhei: The appearance of a 21kd transmission blocking determinant on the developing ookinete. Parasite Immunol. 10, 193-207. Wood, R. J., Cook, L. M., Hamilton, A. & Whitelaw, A. (1977). Transporting the marker gene re (red eye) into a laboratory cage population of Aedes aegypti (Diptera: Culicidae), using meiotic drive at the MD locus, /. Med. Entomol. 14, 461-4. Zheng, L., Saunders, R. D. C , Fortini, D., della Torre, A., Coluzzi, M., Glover, D. M. & Kafatos, F. C. (1991). Low resolution map of the ma-

62

Julian M. Crampton and Paul Eggleston

laria mosquito, Anopheles gambiae, Proc. Nad. Acad. Sci. USA 88, 11187-91. Zheng, L., Collins, F. H., Kumar, V. & Kafatos, F. C. (1993). A detailed genetic map for the X chromosome of the malaria vector, Anopheles gambiae. Science 261, 605-8.

3 Transgenic fish NORMAN MACLEAN AND AZIZUR RAHMAN

The introduction of novel genes into fish of many species has become a common procedure. It has been reviewed by Ozato et al (1989), Chourrout et al (1990), Cloud (1990), Maclean and Penman (1990), Fletcher and Davies (1991), Guise et al. (1991), Houdebine and Chourrout (1991), Powers et al (1992a) and Hew and Fletcher (1992). A comprehensive review would now be a major task. Instead we attempt a shorter appraisal of certain critical aspects of this research area. 3.1 Fish species selection

The species of fish which have been subjected to transgenic induction include Atlantic salmon (Salmo Salar; see, e.g., Shears et al 1991), rainbow trout (Oncorhynchus mykiss; e.g., Guyomard et al 1989), tilapia (Oreochromis niloticus; Brem et al 1988; Rahman & Maclean 1992a), carp (Cyprinus carpio; Zhang et al 1990), channel cat-fish (Ictalurus punctatus; Dunham et al 1987), African cat-fish (Clarias gariepinus; Muller et al 1992) and northern pike (Esox lucius; e.g., Guise et al 1992), and attempts have been made with sea bass {Spams auratus; Cavari et al 1993). In addition to these commercially important species, certain "model" fish have been used, notably goldfish (Carassius auratus; see, e.g., Yoon et al 1990), loach (Misgurnus fossilis; e.g., Kozlov et al 1988), medaka (Oryzias latipes; e.g., Ozato et al 1986) and zebra fish {Brachydanio rerio; e.g., Stuart et al 1990). Unless there are compelling local or strategic reasons for choosing salmonids or other slow-maturing species, it seems that small, fastmaturing species make better initial choices in the present stage of development of this field. Zebra fish and medaka make excellent models, especially the former species, for which a substantial genetic 63

64

Norman Maclean and Azizur Rahman

background exists. The shortcomings of these species lie chiefly in their very rapid embryonic development, so that transgene mosaicism in the parental generation results in poor rates of transmission to the Fl generation. Tilapia species offer some compromise advantages, since they mature sexually within 6 months, but more aquarium space is required to hold large numbers than is the case with either of the former species, and the genetic background information on tilapia is poor. We would argue that the use of both medaka and zebra fish is to be encouraged, since they have some complementary advantages and disadvantages and allow useful comparison between species when used for transgenic work. 3.2 Methods of introducing the transgene 3.2.1 Oocyte injection

Almost all of the microinjection procedures used to accomplish transgenism in fish involve fertilized eggs. However, a notable exception is the work of Ozato et al. (1986), Inoue et al. (1989) and Matsumoto et al. (1992) in which medaka oocytes were injected. Since the oocyte nucleus is very large, these workers have developed a procedure for direct nuclear injection, in contrast to the cytoplasmic injection invariably used with fertilized eggs. 3.2.2 Fertilized egg injection This is the procedure most commonly employed to establish a transgenic state in fish, it being assumed that injection of unfertilized eggs will lead to premature activation and/or damage to the micropyle and jeopardize later fertilization. Eggs are normally injected as soon as possible after fertilization; usually eggs and milt are stripped from ripe fish and fertilization effected by mixing in vitro. Fish egg chorions are often very tough and may be removed before injection, as is done with goldfish and loach eggs, or softened in some cases by agents such as glutathione. Injection may be accomplished in intact eggs via the micropyle, as with eggs of tilapia (Fig. 3.1) and Atlantic salmon, or by direct injection through the chorion. Invariably the perinuclear cytoplasm is the injection target, and often large numbers (e.g., 106 copies) of the transgene may be delivered to each egg. In fast-developing eggs such as those of zebra fish or medaka, it may be difficult to inject one-cell eggs routinely, and procedures often entail injection

Transgenie

fish

65

Fig. 3.1. Needle microinjection of a two-cell-stage egg of the tilapia fish Oreochromis niloticus. The flask-shaped egg has a thick and dense chorion and a wide micropyle. As can be seen, a flexible Teflon tube is used to hold the egg by suction, and the microneedle is passed down the micropyle by mechanical micromanipulation. The multiple gene copies can be precisely injected into either cell. Microinjection is cytoplasmic, since the egg nuclei cannot be visualized. into the early multicelled embryos. This, not surprisingly, results in major problems with transgene mosaicism and germ-line transmission rates with the parental generation that are lower than those obtained in work with species such as salmonids where development is comparatively slow (see, e.g., Stuart et al. 1988 on zebra fish and Penman et al. 1991 on rainbow trout). 3.2.3 Fertilized egg electroporation

This procedure, which involves subjecting fertilized eggs to electric pulses in the presence of the transgene copies, has recently become very popular, and very high transgenic induction efficiencies have been claimed in some cases (Powers et al. 1992). There remains some uncertainty about the true efficiency of integration with this method, however. With the firefly luciferase reporter gene, driven by a viral promoter, electroporation of dechorionated eggs certainly results in a significant percentage of the embryos expressing the transgene (Muller et al. 1993). This has been demonstrated in a number of laboratories, including our own. What is still somewhat in doubt is whether eggs

66

Norman Maclean and Azizur Rahman

with intact chorions can be efficiently transfected by this method, and whether, with or without chorions, sufficient transgene copies are introduced to permit subsequent integration and germ-line transmission (for references see Buono & Linser 1991, 1992). It may be that very particular electroporation conditions are necessary to achieve success with eggs having intact chorions (Powers et al. 1992b). Although electroporation is an attractive mass transfer procedure, it requires careful optimization to achieve the right balance between embryo viability and delivery of transgene copies. A supply of fairly large volumes of transgene suspension is also required, which can be expensive in terms of restriction enzymes and work input. In our view, the worth of electroporation as a means of inducing transgene integration remains to be proved for most fish species and is still complicated by the frequent need for prior egg dechorionation. Although this section focuses on electroporation of fertilized eggs, the technique has been successfully applied to oocytes (Inoue et al. 1990). 3.2.4 Sperm electroporation A Hungarian group including Muller, Orban and Horvath have had some success with the electroporation of fish sperm in the presence of transgenes before its use for fertilization. Initial rates of DNA transfer are low, but any attempts to evaluate the method would be premature (see Muller et al 1992; Sin et al. 1993). 3.2.5 Gene guns Various attempts are being made to use gold or tungsten particles propelled explosively to carry DNA into cells, and fish eggs are one of the chosen targets (Zelenin et al. 1991). Again the procedure is too novel for a precise evaluation to be made. 3.2.6 Chromosome-mediated gene transfer This interesting approach, used by Disney et al. (1988), involves fertilization of trout eggs with gamma-irradiated sperm (to induce fragmentation of the chromosomes) followed by heat-shock-induced gynogenetic development. A proportion of these eggs develop and retain minichromosomes or integrated chromosome fragments from the male. Although the identity of the genes transferred is unknown ex-

Transgenie

fish

67

cept by deduction following extensive experimentation, large-scale screening of the resulting mosaic fish could identify the acquisition of desirable genetic traits. 3.2.7 Lipofection This procedure involves the inclusion of the transgene DNA in lipid or lipoprotein micelles and subsequent exposure of dechorionated eggs to the liposomes in a medium favouring fusion between them. The technique has been used successfully in nonfish systems (SchafferRidder et al. 1982), and its successful exploitation with fish eggs has been described (Szelei et al. in press). 3.2.8 Embryonic stem cells This procedure is proving extremely valuable in the work with transgenic mammals, since it allows transfection of the stem cells (derived from the embryo), in culture (Evans & Kaufman 1981), followed by selective screening for cells with the appropriate copy number and expression characteristics. These cells are then reintroduced into the inner cell mass of a developing mammalian embryo (Evans 1989). Although the resulting animal will be mosaic, germ-line transmission is frequent and the Fl animals are, of course, nonmosaic. Efforts are being made to use the procedure with fish; Collodi et al. (1992) are currently experimenting with zebra fish embryos in this way. The procedure may be more difficult with fish than with mammals because of the larger cell numbers in the embryos and early loss of totipotency in the embryonic cells. However, Lin et al. (1992) have already successfully introduced midblastula cells from genetically pigmented zebra fish into midblastula embryos of albino zebra fish. Twenty-three out of 70 fish were chimeric, and some were germ-line-transformed. There is therefore hope that, in at least some fish species, the huge advantages of an embryo stem cell (ES cell) system will become available. 3.3 Design and expression of transgene constructs 3.3.1 The choice of promoter

It is clearly necessary to couple structural genes with suitable and efficient promoters, unless their own promoters are deemed adequate

68

Norman Maclean and Azizur Rahman

or desirable. Viral promoters such as the enhancer/promoter from the long terminal repeat section of Rous sarcoma virus (RSV) have been used successfully in fish (Chong & Vielkind 1989; Powers et al. 1991), although Bearzotti et al. (1992) obtained only aberrant expression of a bovine growth hormone cDNA driven by an RSV promoter in trout. This finding may reflect a problem with the gene rather than the promoter. While viral promoters provide constitutive or nearly constitutive expression, they afford no information about normal gene regulation in fish and are undesirable for use in aquaculture. Indeed, in some countries there are already laws forbidding the release of any transgenic organism containing a DNA sequence of viral origin. Heterologous promoters from mammalian genes may or may not sustain expression in transgenic fish. Thus, Chourrout et al. (1990) and Penman et al. (1991) found no detectable expression of certain all-mammalian constructs in transgenic fish. However, the situation is complex, since success is recorded using the same promoter (mouse metallothionien 1) when spliced to a human growth hormone gene (Chen et al. 1990). One of the clearest demonstrations of the activity of mammalian promoters in transgenic fish is provided by the work of Chourrout (pers. commun.), in which a mouse immunoglobulin enhancer/promoter effectively directed chloramphenicol acetyltransferase (CAT) expression exclusively in white blood cells of transgenic trout. In our laboratory we have investigated transient expression levels in tilapia fry with two comparable reporter gene constructs, one driven by a fish /3-actin promoter and the other by an equivalent rat promoter (Fig. 3.2); the fish promoter is convincingly more efficient (unpublished data). Matsumoto et al. (1992) have dramatically demonstrated the expression of a mouse tyrosinase gene, driven by its own promoter, in cells of transgenic albino medaka. Inoue et al. (1989) have studied the expression of a chicken crystallin gene driven by its own promoter in medaka embryos. Expression was Fig. 3.2. Restriction maps of two gene constructs used in our laboratory to make comparative studies of equivalent promoters from fish and mammals. Each has a bacterial lac Z reporter gene driven by a large upstream promoter sequence from a j8-actin gene, one from the common carp and the other from a rat. The first exon and intron of the /3-actin gene are also included in each case, and each sequence also contains a poly A signal sequence from SV40. Constructs made by Dr. Andrew Popplewell at Edinburgh Centre for Genome Research, using a carp sequence originally isolated and kindly provided by Professor Perry Hacket, University of Minnesota, and a carp sequence originally isolated and kindly provided by Professor Yaffe, Rehovot, Israel.

Trans genie fish

69 Spel

Sful

Kspl

Muni

Xhoi/Sall Fusion

ACTAGTTTCGAACCGCGGCAATTGCTCGAC

pSP72/poly-3 Carp 8-actin promoter

T7 Promoter

SV40 polyadenylation Signal

pP3PAIacZ-CarpBA 10.306kb

Hpal BamHI/Bglli Stop Codons EcoRI

Exon 1 untranslated

Sail

LacZ

Inton A Exon 2 Nco|P artial

Hpal

Bglll/BamHI INSERT: A 4.6kb Sall/Ncot carp 8-actin promoter fragment.

Spel

Sful

Kspl

Munl/EcoRI fusion

iCTAGTTTCGAACCGCGGCAAT

SP6 Promoter

\

^

—

pSP72/poly-3

Rat B-actin promoter

P

" To'»>\ A Spel\\\ M ~ Xbal\^\\ m p

SV40 polA P signal sjmm Hpai/W m BamHI/Bglll// Stop Codons / EeoRl

\\

PP3PAIacZ-RatBA 10.100kb

^k ^ ^

\ \ _Ncol

II //

1 /

Hpal

Exon 1

Intron A Exon 2 ^.(partial) Bglll/BamHI ™ c c l INSERT: A 4.6kb EcoRI/Ncol rat B-actin promoter fragment.

70

Norman Maclean and Azizur Rahman

measured by colorimetric immunoassay. Although the gene was expressed in many embryos, the tissue specificity was anomalous, with expressing tissues including muscle, bone and gill. It may be that this pattern of expression is a result of aberrant recognition of the chicken promoter by transcription factors. Alternatively, some or all of the detected crystallin may result from the activity of long-lived message derived from the oocyte injection procedure employed. However, now that a range of fish-derived promoter sequences are available, there seems little point in persisting with sequences of mammalian origin. Promoter sequences now in clone from fish include a range of growth hormone promoters, others from H4 histone, several antifreezes, metallothionien, j3-actin, myosin and several more. These are listed in the Appendix (Section 3.9.1). It is clearly desirable to test for promoter and enhancer activity in fish tissue cell lines before introducing them into fish, and this has been done in many cases (Friedenreich & Schartl 1990; Huang et al. 1990). The cell line must be permissive for the particular promoter, and this approach is therefore limited by the present scarcity of established fish cell lines. As emphasized in the excellent review by Sippel et al. (1992), the extent of the promoter sequence upstream is just as important as its identity, and no doubt many problems of aberrant expression can be attributed to inadequate 5' sequence in the chosen promoter. The optimization of promoter sequences is discussed in Chapter 1 (Section 1.3.3) of this book and is reviewed by Moav et al. (1992). 3.3.2 The mechanics of integration The mechanisms involved in stable transgene integration are poorly understood. Initially it was often assumed that the use of retroviral terminal repeat sequences to flank transgenes might be necessary to accomplish satisfactory rates of integration, much as transposable elements have proved necessary in insect transgenic induction (Bello & Couble 1990), but this has not proved essential in fish. In any event, no clear examples of retroviruses have as yet been identified in piscine animals, although a possible endogenous retrovirus has been identified in salmonids (Stuart et al. 1992). Enhanced incorporation of transgene DNA into zebra fish has been reported by Ivies et al. (1993) who used a retroviral integrase protein. Bishop and Smith (1989) have considered the mechanisms likely to

Trans genie

fish

71

be involved in both cellular transformation following transfection with foreign DNA and transgene integration after cytoplasmic or nuclear injection of transgenes. They concluded that these processes have a common mechanism, namely opportunistic repair ligation, probably occurring during DNA replication at the cell cycle S phase. This would sometimes, but not invariably, involve deletion and loss of some chromosomal DNA sequence. All of the evidence considered in Bishop and Smith's article is drawn from work on mammals and mammalian cells, but similar mechanisms may be obtained in fish. 3.3.3 Use of reporter genes When particular promoter sequences are being assessed either in vivo or in vitro, it is often effective and desirable to use a reporter gene which provides an easy assay. The choice is quite wide, although various reporter gene sequences do have inherent weaknesses or attendant problems. Possible choices include /3-galactosidase (lac Z) (Fig. 3.3), chloramphenicol acetyltransferase (CAT), neomycin phosphotransferase (neo), all of bacterial origin, and firefly luciferase. Some of the shortcomings of these sequences include possible bacterial contamination of the samples, with some or all of the bacterial reporter genes, the occurrence of lac Z-like activity in some fish cells, especially in the developing gut of embryos or fry (K. Inoue, pers. commun.), the frequency of false positives with CAT even with thinlayer chromatography assay and the toxicity of the antibiotic G418, a neomycin analogue used in the assay for neo. The advantages offered are the sensitivity of many of the assays, especially with luciferase and fluorometric detection of lac Z, the use of a colorimetric reaction to indicate tissue-specific expression in the case of lac Z, the possible use of neo as a viability screen if a large number of potential expressing cells or animals are available and the possible use of luciferase in vivo, since living fry take up luciferin (the substrate) from the water and display light emission from tissues containing the gene (unpublished work in our own and other laboratories). A general drawback of the work with reporter genes is that commonly only transient expression is monitored. It is obviously desirable to follow this up with longer-term expression assays involving the activity of incorporated sequences. Another drawback is that, if plasmid sequences are injected as part of a construct, reporter gene ex-

72

Norman Maclean and Azizur Rahman

Fig. 3.3. Three 48-hr yolk sac fry of the zebra fish (Brachydanio rerio) which demonstrate variable expression of injected reporter transgenes. The reporter gene is bacterial lac Z; the product of this gene is the enzyme /3-galactosidase, which here has reacted with a substrate to provide a blue colour. Note the intense expression in the yolk sac. This is not in the yolk itself but in the yolk sac endoderm surrounding the yolk. Some expression is also visible in strips of muscle fibre cells in the caudal area. Expression is mosaic and probably transient, revealing a partial distribution of the transgene copies to some cells and tissues and not others. From the top down, the three fry are (1) a control which has developed from an uninjected egg, (2) a fish which has developed from an egg injected with 106 copies of a construct in which the lac Z gene is driven by a rat /3-actin promoter and (3) a fish which has developed from an egg injected with 106 copies of a construct in which the lac Z gene is driven by a carp /3-actin promoter (see Fig. 3.2 for restriction maps of these constructs). The second construct is expressed more strongly than the first. The experiment was carried out by Darren Williams in our laboratory (unpublished).

pression may be driven by plasmid promoters and not the promoter being tested. It is desirable to remove all plasmid sequences from constructs before using them for introduction into fish tissues. It is even rumoured that some very active reporter genes such as CAT are expressed weakly in the absence of any promoter sequence (Shiokawa et al. 1990). A comparatively novel way of determining promoter activity with a reporter gene involves direct injection of gene constructs into skeletal muscle of live fish, followed by assays of biopsy samples recovered from the same site (Hansen et al. 1991; Rahman & Maclean 1992b).

Transgenie

fish

73

This follows from similar work in small mammals (Thomason & Booth 1990; Wolffs al. 1990). Some fish structural genes may display obvious phenotypic effects which render them natural reporter genes, and the frequent use of fish growth hormone sequences is probably partly explained by this observation. However, we should emphasize that the difficulty of carrying out effective growth trials in fish because of their considerable growth plasticity implies that claims for increased growth of fish transgenic for growth hormone genes should always be supported by strong molecular evidence for the presence of the novel growth hormone protein in the fish serum. 3.3.4 Copy number injected We have already suggested that plasmid DNA sequences should not be included with the injected sequence. This also implies that the sequence be linearized. It seems likely that circularized sequences will survive longer after injection than linear ones because of their resistance to exonuclease attack. However, there is evidence that many linear sequences are rapidly circularized after injection (Bishop & Smith 1989) with or without attendant concatamer formation. Since fish eggs and embryos vary greatly in size, there is some logic in varying the number of copies of transgenes introduced per egg at microinjection. Such numbers as 105 and 106 per egg are commonplace. There is some evidence that larger numbers may lead to more transgenics but lower survival ratios (Penman et al. 1990). The number of transgenes injected per egg is very high when compared to the number introduced into mammalian eggs (often 102-104), but it must be remembered that in fish eggs the nucleus cannot be targeted and so many copies are presumably lost or degraded and are never included within the nucleus. However, in transgenic work with Xenopus DNA injection is often lethal due to the toxic effects of the DNA on the embryo (R. Patient, pers. commun.), and some batches of rainbow trout eggs suffer high mortality at gastrulation following injection with more than 106 copies of a 7-kb transgene (Penman et al. 1990). A compromise must therefore be reached between injecting sufficient DNA to ensure a reasonable percentage of transgenic fish and injecting so much that embryo death becomes a problem. It is widely supposed, but without good evidence, that transgene copies are taken up into the nucleoplasm before re-formation of the nuclear envelope after each mitosis.

74

Norman Maclean and Azizur Rahman

3.3.5 The complication ofDNA methylation There is strong evidence that DNA methylation, especially within promoter regions, is correlated with gene inactivity. Thus, transgene DNA methylation may account for many of the examples of transcriptional inactivity and lack of transgene expression. A preliminary study of this aspect of transgene expression in fish is that of Maclean et al. (1992) (Fig. 3.4). A correlation between incorporation of transgene concatamers and transgene hypermethylation has been established in mice by Mehtali et al. (1990). This may provide an explanation for the poor or nil expression often observed with transgene concatamers.

3.3.6 Transgene expression It is essential to distinguish between the expression of integrated and that of nonintegrated DNA sequences. Although the latter is frequently referred to as transient expression, the activity of nonintegrated sequences may persist for many days in fish embryos. Expression is normally presumed to be transient until good evidence for integration is available. If transgene sequences become associated with centromeric DNA, they might become minichromosomes, especially if existing as concatamers, and this situation would then be a third possible outcome. We know of no reported examples in work with transgenic fish of such a situation. As discussed in Section 3.3.1 the promoter sequence is likely to be the chief factor determining both whether expression occurs and when and where it occurs. The complexity of the transcriptional event, as indicated by the ever-increasing knowledge of transcription factors, enhancer sequences and interaction with other cytoplasmic molecules such as hormones (see review by Latchman 1991) is obviously considerable. There are many aspects to the monitoring of transgene expression in the transgenic animal. There is a temporal one, namely the developmental stages, or even cell cycle stages, at which expression occurs. There is a spatial aspect, namely the cell types and tissues in which the transgene is expressed, given that most animal tissues comprise more than one type of cell. There is a quantitative aspect, involving determination of the variation in the rate at which the transgene may be expressed at different times. Finally, there is the crucial question of how expression is being monitored. This may involve detection of

Transgenie fish

A R C D E F G H

75

I

J K L M N O P Q R S T U

Fig. 3.4. DNA from a rainbow trout transgenic for the sequence mMTrGH (mouse metallothionein promoter spliced to a rat growth hormone gene) subjected to Southern blotting following digestion with enzymes which are or are not methylation sensitive. Thus, Msp I will cut the sequence CCGG even if the internal C is methylated, whereas Hpa II will cut it only if this C remains unmethylated. The gel was probed with a 32P-labelled probe of the entire sequence. Lane A: control DNA + mMTrGH (equivalent to x2 copies per genome) and Pst I; B: as in A but digested with Msp I; C: muscle DNA from transgenic fish, undigested; D: muscle DNA from transgenic fish, digested with Msp I; E: muscle DNA from transgenic fish digested with Hpa II; F: heart muscle DNA from transgenic fish digested with Msp I; G: heart muscle DNA digested with Hpa II; H: kidney DNA from transgenic fish (undigested); I: kidney DNA from transgenic fish digested with Msp I; J: kidney DNA from transgenic fish digested with Hpa II; K: liver DNA from transgenic fish, undigested; L: liver DNA from transgenic fish digested with Msp I; M: liver DNA from transgenic fish digested with Hpa II; N: spleen DNA from transgenic fish, undigested; O: spleen DNA from transgenic fish, digested with Msp I; P: spleen DNA from transgenic fish, digested with Hpa II; Q: brain DNA from transgenic fish, undigested; R: brain DNA from transgenic fish, digested with EcoRI + Bam HI; S: brain DNA from transgenic fish, digested with Pst I; T: brain DNA from transgenic fish, digested with Msp I; U: brain DNA from transgenic fish, digested with Hpa II. Note that all lanes were loaded with 5 mg of DNA except F and G, which were loaded with only 3 mg. Reproduced with permission from Maclean et al. (1992).

76

Norman Maclean and Azizur Rahman

transgene message by probe hybridization on Northerns, detection of protein as a cellular or secreted product by histochemical or immunofluorescent recognition or assessment of some physiological parameter such as pigment formation or growth rate if the gene product has a satisfactory observable phenotypic effect. Scoring a phenotypic effect such as growth should always be accompanied by parallel assays for specific mRNA or protein, while the detection of specific message cannot be taken to mean that the transgene product is being correctly processed and expressed as a functional protein. One of the better examples of transgene expression studies in fish is provided by Shears et al. (1991) in which Atlantic salmon carrying a winter flounder antifreeze gene, driven by its own promoter, were shown by immunoblotting to have the antifreeze precursor protein in sera. Unfortunately, the production of the antifreeze protein in these fish seems insufficient to make them resistant to freezing-cold Arctic waters in winter. The same group (Du et al. 1992) has also demonstrated growth enhancement in Atlantic salmon using an entirely fish gene construct (chinook salmon growth hormone cDNA driven by the ocean pout antifreeze promoter). It is often desirable to monitor the expression of a construct in tissue culture before using it for transgenic induction, although unless a ubiquitous promoter is used a cell line must be chosen which will demonstrate the tissue-specific expression required by the construct. Unfortunately, rather few established (immortalized?) fish cell lines exist, and this has undoubtedly impeded such experiments. However, some laboratories have used available lines; a good example is the work of Daniel Chourrout's group in collaboration with others (Bearzotti et al. 1992) which demonstrates the ability of a human heat shock 70 promoter to drive CAT reporter gene activity in both carp and trout cells. 3.3.7 Transgene markers A desirable goal of work with transgenic fish would be to develop a marker gene such that fish transgenic for that gene would have an altered body pattern or distinctive behavioural trait. The luciferase reporter gene when driven by a ubiquitous promoter such as that of SV40 virus or carp /3-actin may come close to this ideal situation, since in at least some species the embryos or fry can be assayed alive in the presence of luciferin (see Section 3.3.3). Unfortunately, most

Trans genie fish

11

interesting transgene constructs are tissue specific in expression and not expressed in skin, thus preventing the use of such an easy selection assay. The use of a mouse tyrosinase gene for introduction into albino medaka constitutes an excellent use of a valuable marker gene (Matsumoto et al. 1992), and no doubt the system will be taken up elsewhere. The neomycin resistance gene neo has proved very useful in mice, especially in the positive-negative selection procedure used by Thomas and Capecchi (1990) in which G 418 and gancyclovir are used for the dual selection system. In the absence of an ES cell system, the question is frequently posed whether selection on entire fish fry could be used, especially if electroporation allowed the production of a large number of potential transgenics. The difficulty here is that almost all such fish are transgene mosaics, and the cells and tissues not carrying or expressing the transgene are not drug resistant. The drugs kill almost all the fish, whether or not they are transgenic. 3.4 The fate of injected DNA Because, apart from oocyte injections in which the nucleus can be targeted, all DNA injections into fish eggs are into the perinuclear cytoplasm, some interesting questions arise about what happens to these novel sequences. 3.4.1 Cellular containment of the DNA There is good evidence that, in Xenopus, heterologous DNA injected into the cytoplasm is rapidly (within 90 min) surrounded by membrane and packaged as numerous cytoplasmic nucleus-like structures (Forbes et al. 1983; Shiokawa et al. 1986, 1987). Following gene injection into Xenopus fertilized eggs, these pseudonuclei become distributed among the blastomeres until the blastula stage. It seems a fair assumption that a similar process occurs in the fish egg. Just how these structures form is unknown, though Shiokawa et al. (1986, 1987) report that they have apparently normal nuclear envelopes with pores, but rather homogeneous contents, as visualized in the electron microscope. The capacity of eggs to form these pseudonuclei and to facilitate so many cellular and enzymic activities such as transcription, DNA replication, chromatin formation and finally DNA degradation is presumably at least partly a result of the fact that these eggs function

78

Norman Maclean and Azizur Rahman

as natural storehouses of molecules to be used in the early stages of development (Etkin & Pearman 1987). 3.4.2 Concatamerization Injected DNA is often rapidly concatamerized, either as head to head, head to tail or tail to tail, or as random arrangements, and irrespective of whether it is originally injected in a linear or supercoiled form (see discussion in Vielkind 1992). This indicates that DNA ligase enzyme is abundantly available in the proximity of the DNA, presumably within the pseudonuclei. Like all enzymes, DNA ligase is synthesized in the cytoplasm but probably is normally present only in the nucleus. The precise form of the concatamers seems to vary somewhat depending on whether the original DNA was supercoiled or linear. If supercoiled, most concatamers are head-to-tail multimers, perhaps as a result of homologous recombination between linearized molecules (Vielkind 1992; A. Iyengar in our own laboratory, unpublished data), while the linear DNA concatamerizes into head to tail chiefly, but head-to-head and tail-to-tail arrangements also occur. 3.4.3 Replication Many authors have reported that DNA injected into fish eggs undergoes rapid replication until the blastula or even the gastrula stage, followed by slow degradation (Stuart et al. 1988; Winkler et al. 1991). It seems highly likely that this replication occurs within the pseudonuclei, and as with the ligase referred to in the preceding section, the necessary DNA polymerase enzymes are present within these structures. It is not known whether it is in synchrony with the normal cellular S phase, but there is evidence to suggest that transgene replication may certainly be dependent on pseudonucleus formation (Etkin & Pearman 1987). The fact that transgene replication is not a continuing process is presumably due to the fact that the pseudonuclei are eventually diluted out by cell division. 3.4.4 Integration In some transgene-injected fish eggs, integration into chromosomal DNA occurs. The percentage is often low, less than 10% of viable embryos and fish, and is commonly a mosaic situation, indicating that

Trans genie

fish

79

integration occurred after one or more rounds of cell division. It is frequently assumed that the DNA copies become internalized into the nuclei following breakdown and re-formation of the nuclear envelopes at mitosis. But it may not be as simple as this if the DNA copies are already packaged up in pseudonuclei. Rather, internalization into the nucleus may follow from fusion between nucleus and one or more pseudonuclei. The situation deserves careful analysis. Integration may not always follow from nuclear internalization, and long-term persistence of nonintegrated DNA may occur within the nucleus and within cytoplasmic pseudonuclei (Etkin & Pearman 1987). 3.4.5 Degradation The fate of most of the injected transgene DNA in most eggs is degradation, presumably by cellular nucleases. The efficiency of the process is impressive when one considers that often up to 106 copies are injected per egg. We assume that this digestion affects transgenic DNA whether packaged into pseudonuclei or not. 3.4.6 Transient expression Transgenes introduced into fish eggs are very frequently transiently expressed. It is widely supposed that this follows from nuclear inclusion. We suggest that it follows rather from inclusion within pseudonuclei and that these structures are equipped with all of the necessary transcriptional machinery. The localization of transient expression in the yolk (F. Muller, pers. commun.) is perhaps also sometimes a result of migration of pseudonuclei into the yolk compartment of the egg. 3.4.7 Linear or supercoiled conformation There is often disagreement about whether transgenes should be injected in linear form (with or without attached plasmid sequences) or as supercoiled circles. We favour using linear copies without plasmids, as discussed earlier. But the fact that the supercoiled molecules, like linear ones, are rapidly concatamerized surely indicates that supercoils are converted to linear molecules even if they are subsequently recircularized (Bishop & Smith 1989). A useful discussion of this topic can be found in the review of Vielkind (1992), which concludes that both physical forms of DNA yield equivalent percentages

80

Norman Maclean and Azizur Rahman

offish carrying integrated copies. In our own laboratory, we find that better integration rates follow from the injection of linear as compared with supercoiled DNA, and similar observations have been made with transgenic mice (Brinster et al. 1985).

3.4.8 The form of the injected DNA To our knowledge there is no information about whether the injected DNA is packaged as chromatin with nucleosomes. However, it is suggestive that exogenous DNA injected into Xenopus oocyte nuclei is rapidly converted to chromatin (Trendelenburg & Gurdon 1978).

3.5 Verifying integration 3.5.7 Methods used to verify integration Dot or slot blotting followed by hybridization with a suitably labelled probe is the customary way of monitoring the presence of transgene copies, and provided that appropriate controls verify that there is no cross-hybridization to endogenous sequences, the method works well. It can also be made semiquantitative by the inclusion of blots of standard numbers of transgene copies, estimated to coincide with the number of genome equivalents in the loaded DNA, or the DNA concentration used together with the known genome size to provide standards of, say, 1 copy or 10 copies over diploid genome. An increasingly popular alternative to slot/dot blotting is the use of the polymerase chain reaction (PCR). This technique is exquisitely sensitive but suffers from the fact that it is nonquantitative, since the concentrations of the amplified sequences are not expressive of the original copy number of the transgene. Cross-contamination must also be studiously avoided. Dot/slot blotting and PCR do not in themselves indicate integration of transgene copies, only their presence. This must be assumed to be transient in many cases. Careful Southern blotting after appropriate restriction enzyme digestion is therefore necessary to demonstrate true chromosomal integration. Even the existence of transgenes in high molecular weight DNA without restriction digestion is complicated by the frequent presence of transgene concatamers. It may therefore be essential to demonstrate carefully that putative junction fragments include fish sequences, either by their greatly increased

Transgenie

fish

81

molecular weight in the Southern blot or, even better, by cloning and partially sequencing the junction fragments. Although conventional PCR using template primers from the transgene cannot be used to prove integration, the use of an alternative PCR strategy, inverted PCR (Ninomiya et al. 1990), permits the amplification of junction fragments and, subsequently, their identification and verification. Scrutiny of Southern blotted restriction fragments of DNA recovered from putative transgenics sometimes yields evidence suggestive of the long-term persistence of nonintegrated transgene concatamers (Penman et al. 1990). Recently, however, Tewari et al. (1992) have elegantly demonstrated that, at least in some cases, this finding may result not from lack of integration, but from head-to-tail concatamers forming DNA palindromes which, after restriction, have molecular weights identical to nonintegrated flanking sequences. These authors also used in situ hybridization in chromosomal spreads to verify integration into a single chromosomal locus of the large concatamer followed in these studies (Fig. 3.5). Although concatamers readily form from cloned linear DNA sequences, with or without attendant modification of the terminal sequences, the recovery of transgenic fish which have large integrated concatamers may not be ideal for expression studies. There is mounting evidence that transgene expression is often best from single-copy transgenics and minimal or absent from concatamer transgenics. This may be partly or entirely a result of widespread DNA methylation of integrated concatamers, as discussed earlier. As a generalization, transgene integration is readily obtained in fish following oocyte or egg microinjection or egg electroporation, but the resulting fish often contain more than one transgene copy per genome. 3.5.2 Transgene mosaicism The most common result of transgenic induction in fish is an individual which is mosaic with respect to the transgene. This may imply the presence of a single copy in some tissues and an absence of copies in others, or a variable copy number in different tissues as a result of multiple integration events, or some combination of the two. In assessing mosaicism it must be borne in mind, first, that most tissues, such as liver, spleen and brain, are composed of many different cell types, some of which are and some of which are not transgenic and,

Norman Maclean and Azizur Rahman

82

, i

(a)

Genome

•

'

Genome

(b)

Fig. 3.5. (a) Microscopic demonstration of chromosomal integration of a concatamerized transgene in rainbow trout. The chromosome spread is from blood leucocytes, and the points marked by arrows were positive for a digoxigenin-labelled probe of the injected sequence. The positive chromatin is visible in both interphase nuclei and a mitotic spread, and is localized in a single metacentric chromosome. This argues strongly that the concatamer is truly integrated, although the Southern blot results for this fish suggested persistence of the transgene as an extrachromosomal concatamer. Reproduced with permission from Tewari et al. (1992). (b) Diagrammatic representation of how concatamers, including inversely oriented copies, could form hairpin loops by intrastrand pairing and thus yield a Southern blot without apparent evidence of chromosomal integration of the concatamerized DNA. The probability of chromosomal integration in this case is indicated by the localization demonstrated in panel (a). Reproduced with permission from Tewari et al. (1993).

Trans genie

fish

83

second, that tissues are rarely clonal even with respect to one cell type, and so may be mosaic even with respect to a single cell type. It is likely that fish with rapid early development such as medaka and zebra fish display significantly more transgene mosaicism than species such as salmonids with slow embryonic development (compare Stuart et al. 1988, 1990 with zebra fish and Chourrout et al. 1986 with rainbow trout). 3.6 Transmission of transgenes to Fl and subsequent generations Assuming transgene integration, mosaicism is presumably the chief obstacle to transmission. Despite widespread problems with transgene mosaicism, transmission has now been obtained with many fish species: Atlantic salmon (Fletcher et al. 1992), rainbow trout (Penman et al. 1991; Yoshizaki et al 1991b), common carp (Zhang et al. 1990), zebra fish (Stuart et al. 1988, 1990; Khoo et al. 1992), medaka (Inoue et al. 1989; Matsumoto et al. 1992) and others. Since the gonads and germ-line cells are themselves often mosaic for the transgene, transmission rates of less than 50% are frequently obtained in the PI to Fl transmission. However, percentages well above 50% may also be obtained as a result of using fish with transgenes incorporated into more than one chromosome (Penman et al. 1991; Tewari et al. 1992). As expected, crossing of Fl transgenics with nontransgenics results in Mendelian segregation of transgenes to approximately 50% of the F2 progeny (Maclean et al. 1992). Breeding between sibling transgenic Fl with similar Southern patterns should result in F2 progeny homozygous for transgenes. This may also reveal some degree of transgenic lethality due to the absence of expression of DNA sequences lost in a recombinational event when transgene integration occurred. Such studies have not, to our knowledge, been undertaken. 3.7 Gene ablation and function reduction Whether to achieve sterility or increased growth, or purely as a laboratory procedure, there is currently much interest among fish-oriented molecular biologists in possible methods for knocking out individual cells or silencing specific genes. However, until an effective fish ES cell system is available (see Section 3.2.8), gene targeting and homologous recombination seem a distant prospect. It is possible to antagonize gene expression using techniques involving antimessage or cus-

84

Norman Maclean and Azizur Rahman

tom-made ribozymes, but in most fish laboratories these techniques are still either hypothetical or at the developmental stage. Specific cell ablation with toxins such as ricin or diphtheria toxins has proved effective in mice (Breitman & Bernstein 1992) but is unlikely to commend itself to laboratories allied to commercial aquaculture. The alternative procedure of using the herpes symplex thymidine kinase gene can be made selective; moreover, the gene product is not itself toxic. Some manipulation is required to combine the tK gene with a fish promoter which is specific to the cells to be ablated. Following selective administration of an inducing drug such as Gancyclovir, the viral tK gene is expressed and knocks out the cells in which it is so expressed (reviewed in Breitman & Bernstein 1992). Here, then, is a potentially acceptable way of ablating specific cells following drug administration to the targeted fish. This procedure may prove useful in the application of transgenic technology to practical aquaculture, but requires treatment of the fish with acyclovir.

3.8 Prospects for transgenic fish

Probably no one writes a review only to conclude that the topic has no future. However, what is interesting about this topic, we believe, is not simply that it has a bright future, but that many scientists remain unaware of the undoubted advantages of the system.

3.8.1 Studies of gene regulation and fish physiology

As animals for the study of gene regulation, transgenic fish have very special attractions. Many species lay copious eggs at regular intervals, fertilization is easily effected in vitro and development is external to the female's body. Even in mouth-brooding species such as tilapia, the eggs and embryos are easily maintained in glass or plastic vessels with a throughput of water. Especially with such species as zebra fish, medaka and tilapia, eggs can be obtained throughout the year. Initially, the difficulties of nuclear injection posed problems, but it now seems that both electroporation and microneedle injection into the perinuclear cytoplasm yield good percentages of transgenics, and even the frequency of transgene mosaicism does not prevent good proportions of germ-line transmission from PI transgenics to Fl prog-

Transgenie

fish

85

eny (Maclean et al. 1992). Coupled with this progress is the proven effectiveness of reporter genes such as CAT, lac Z and LUC in fish, and the promise of marker transgenes such as tyrosinase in albino lines. But if fish are ever to rival Drosophila as a developmental/ genetic model organism, transgenic induction in zebra fish or medaka will have to be done on a huge scale. Fish transgenic technology still suffers because of the absence of an effective ES cell system. The attendant problems and possible solutions are discussed in an excellent review by Rossant and Hopkins (1992). As these authors emphasize, transgenic technology could theoretically be used to conduct a saturation mutagenesis programme, with mutations being effectively tagged by the transgene insertion, but at most only 1 in 19 transgenics could be expected to display insertional mutagenesis (extrapolating from the mouse model). However, one of the great advantages of the zebra fish system is the survival of haploid embryos, at least for a few days of development. Assuming that microinjected DNA will prove as mutagenic in zebra fish as in the mouse, Rossant and Hopkins (1992) calculate that the production and identification of the 30,000 mutations needed for genome saturation would take some 300 scientist-years of effort. As these authors also remark, the development of a fish retroviral transposable element might greatly reduce this effort. Although there is no known fish retrovirus, there are fish retroposons (Kido et al 1991). Two other interesting questions related to the future use of transgenic zebra fish are, first, the development of gene and enhancertrap systems using a gene like lac Z as the expressing gene and, second, the possible future of targeted mutagenesis involving homologous recombination. Both systems, to become at all comprehensive, require an efficient ES cell selection system. 3.8.2 The use of transgenic fish in aquaculture Another future application of transgenic fish lies in aquaculture. Given the world shortage of animal protein, the healthy dietary properties of fatty fish and the gross overfishing of the world's oceans, it seems inevitable that aquaculture will be of ever-increasing significance worldwide. It is possible, however, that some government bodies will fail to authorize the use of transgenic fish in aquaculture unless the fish are sterile or they are of a species which is incapable of breeding

86

Norman Maclean and Azizur Rahman

in the local aqueous environment (see Devlin & Donaldsen 1992; Hallerman & Kapuscinki 1992; Hindar 1993). This implies the use of a non-native species or the use of fish which are sterile. The latter requirement makes the development of methods of hormonal sterilization of fish an important goal. It is also clear that the sexual maturation offish in aquaculture often takes place at the expense of useful growth, so providing an additional incentive for transgenic interference with gonadal development. Although transgenic fish, because of their capacity for mobility, will have to be used with greater caution, it seems to us that there should be controlled use of nonsterile transgenic fish, following an appropriate analysis of the likely environmental impact of escapees. Other biological parameters that may be usefully addressed in the future include disease resistance, especially in view of the havoc that diseases such as salmonid furunculosis continue to wreak among farmed and wild fish, and the undesirability of adding antibiotics to open water systems. With many fish diseases there is no clear or wellunderstood genetic basis for resistance, and so development in this area will be slow. The cloning of fish growth hormone genes immediately raises the prospect of bigger or faster-growing transgenic strains. Caution must be exercised, however, since undesirable side effects of overproduction of growth hormone are possible. The interesting experiments of Hew's group (Fletcher et al. 1992) in the development of Atlantic salmon which express antifreeze genes point to another promising future development. The probable future use of transgenic fish in aquaculture provides an interesting second string to the plausible use of transgenic fish in basic molecular research. 3.8.3 Ethical considerations Given the broad agreement that there is no ethical objection to creating a transgenic fish per se (assuming that the novel gene does not confer some disability on the recipient fish), the question remains as to whether the consumption of transgenic fish as food will be authorized or accepted. We would suggest that each new transgenic strain will have to be considered on its own merits, but as a general rule and especially when the transgene constructs used are themselves of 'allfish' origin, it should be accepted that the transgenic technology does not necessarily produce an end-product which differs from that obtained by evolutionary or agricultural selection. Cattle selected over time for high milk yield differ from the original wild type in genetic

Transgenic fish

87

terms just as profoundly as transgenics with extra copies of casein genes, for example. Only the time scale and the technology differ. Since transgenic yeast has already been authorized for bread making, and the production and sale of transgenic tomatoes are already under way in the United States there seems a strong possibility that transgenic fish will be accepted as food in the near future. (For further discussion of this topic see Chapter 1.)

3.9 Appendix: availability of fish genomic and cDNA clones 3.9.1 Fish genes isolated Vitellogenin gene (cDNA) Vitellogenin gene (genomic) Histone gene (genomic) Histone gene (genomic) Myosin gene (genomic) Metallothionein gene (cDNA) Metallothionein gene (cDNA) Metallothionein B gene (genomic) Metallothionein-like promoter Metallothionein A gene 5' end Antifreeze gene (genomic) Antifreeze protein precursor gene (genomic) Growth hormone gene Growth hormone gene (cDNA) Pregrowth hormone gene (cDNA) Growth hormone gene (cDNA) Premature growth hormone (cDNA) Growth hormone gene (cDNA) Growth hormone gene (genomic) Growth hormone gene (cDNA) Growth hormone gene (cDNA) Growth hormone gene Growth hormone gene Growth hormone gene Growth hormone gene Growth hormone gene

(cDNA) (cDNA) (cDNA) (genomic)

Rainbow trout: Le Guellec et al. (1988) Tilapia: Ding et al. (1990) Rainbow trout: Winkfein et al. (1985) Tilapia: Englander and Moav (1989) Carp: G. Goldspink (pers. commun.) Rainbow trout: Bonham et al. (1987) Plaice: S. George (unpublished) Trout: Zafarullah et al. (1988) Xiphophorus: Friedenreich and Schartl (1990) Trout: Murphy et al. (1990) Winter flounder: Scott et al. (1985) Winter flounder: Davies et al. (1982) Atlantic salmon: Johansen et al. (1989) Chum salmon: Sekine et al. (1989b) Red sea-bream: Momota et al. (1988a) Flounder: Momota et al. (1988b) Flounder: Mori et al. (1989) Tilapia: Rentier-Delrue et al. (1989a) Tilapia: Raphael and Violet (1992) Yellow fin sea-bream: Tsai et al. (1991) Rainbow trout: Agellon and Chen (1986) Common carp: Chiou et al. (1990) Rainbow trout: Lorens et al. (1989) Atlantic salmon: Lorens et al. (1989) Grass carp: Ho-Walter et al. (1989) Chum salmon: Shen et al. (1991)

88

Norman Maclean and Azizur Rahman

Growth hormone gene (cDNA) Growth hormone gene (cDNA) Growth hormone gene (cDNA) Pregrowth hormone gene (cDNA) Protamine gene Protamine gene (genomic) Prolactin gene Prolactin gene (genomic) Prolactin gene (cDNA) Prolactin gene Prolactin cDNA Prolactin gene Isotocin (cDNA) Isotocin precursor (cDNA) Vasotocin (cDNA) Vasotocin precursors (cDNA) Vasotocin precursors (cDNA) j8-Crystallin gene (cDNA) y-Crystallin gene (cDNA) Major repetitive DNA family Repetitive satellite DNA Repetitive satellite DNA Repetitive satellite DNA Repetitive DNA Repetitive retroposon-like gene Corticotropin releasing factor precursor (cDNA) High-mobility group proteins (HMGD) (cDNA) Homeobox-containing gene Homeobox-containing gene Homeobox-containing gene Dynein heavy chain (cDNA) Immunoglobin heavy chain (cDNA) MHC antigen gene MHC class I + II and T-cell receptor b chain (genomic)

Coho salmon: Gonzalez-Villasenor et al. (1988) Northern pike: Schneider et al. (1992) Striped bass: Cheng et al. (1991) Gilthead sea-bream: Funkenstein et al. (1992) Rainbow trout: G. H. Dixon (unpublished) Chum salmon: Moir and Dixon (1988a) Rainbow trout: Dixon Chinook salmon: Xiong et al. (1992) Chum salmon: Song et al. (1988) Chinese carp: Chang et al (1992b) Common carp: Chao et al. (1988) Tilapia: Poncelet et al. (1993) Catostomus commersoni: Heierhorst et al. (1989) Masu salmon: Suzuki et al. (1992) Catostomus commersoni: Heierhorst et al (1989) Chum salmon: Heierhorst et al (1990) Masu salmon: Suzuki et al (1992) Common carp: Chang and Chang (1987) Common carp: Chang et al (1991) Tilapia: Wright (1989) Common carp: Datta (1988) Netropis lutrensis: Moyer et al. (1988) Salmonids: Moir and Dixon (1988b) Zebra fish: He et al (1992) Trout: Winkfein et al. (1988) Teleosts: Okawara et al (1988) Trout: P. Pentecost (unpublished) Atlantic salmon: Fjose et al. (1988) Zebra fish: Eiken et al (1987) Zebra fish: Njolstad et al (1990) Trout (testes): Garber et al (1989) Channel cat-fish: Ghaffari and Lobb (1989) Common carp: Keiichiro et al (1990) Carp: S. Kohosawa (unpublished)

Transgenie fish Apopolysiologlycoprotein (cDNA) Apopolysiologlycoprotein genes Insulin-like growth factor (cDNA) Insulin-like cDNA Insulin cDNA Preproinsulin gene /3-Actin (genomic) j8-Actin (genomic) Gonadotropin gene (cDNA) Preprogonadotropin releasing hormone gene Gonadotropin II j8-gene (genomic) Gonadotropin /3-gene (genomic) Gonadotropin a-gene (genomic) Melanin-concentrating hormone genes a-Globin gene (genomic) Serum albumin (cDNA) a-Globin cDNA Opsin gene (genomic) Hsp (heat shock promoter), 90 genes (cDNA)

89 Cherry salmon: Sorimachi et al. (1990a) Trout: Sorimachi et al. (1990b) Coho salmon: Cao et al. (1989) Rainbow trout: Shamblott and Chen (1991) Chum salmon: Sorokin et al. (1983) Chum salmon: Koval et al. (1989) Common carp: Liu et al. (1990a) Grass carp: Liu et al. (1989) Chum salmon: Sekine et al. (1989a) Salmon: Klungland et al. (1992) Salmon: Xiong and Hew (1991) Common carp: Chang et al. (1992a) Common carp: J. H. Chang et al. (1992) Chum salmon: Takayama et al. (1989) Common carp: Miyata et al. (1991) Atlantic salmon: Byrnes and Gannon (1990) Common carp: Takeshita et al. (1984) Goldfish: Johnson et al. (1993) Zebra fish Krone (1993)

3.9.2 Gene constructs used in producing transgenic fish CA T gene + RSV LTR + SV40 promoter Zebra fish

Medaka CAT gene + RSV LTR Goldfish

Linear and circular, germ-line transmission, expression (Stuart et al. 1990) Linear, circular, sperm vector, integration, germ-line transmission (Khoo et al. 1992) Possible integration; expression observed up to 4 weeks after injection (Chong & Veilkind 1989) Persistence and expression in different tissues; level of expression is high in muscle and low in brain and gonad (Hallerman et al. 1990)

90

Norman Maclean and Azizur Rahman

Zebra fish Trout CAT gene + carp/3-actin promoter Zebra fish Tilapia CAT gene + CMV promoter Medaka CAT gene + mMT-1 promoter Medaka CAJ and /3-gal + rainbow trout MTA promoter Medaka CAT + rainbow trout MT-B promoter Medaka CAT gene + Xiphophorus metallothionein-like promoter Medaka

CAT gene + trout histone promoter Tilapia CAT gene + antifreeze promoter Medaka j8-Galactosidase + chicken /3-actin promoter + RSV-LTR Trout /3-Galactosidase gene + mouse metallothionein promoter Salmon /3-Galactosidase gene + RSV promoter Common carp, African cat-fish

Linear, persistence, expression (Buono & Linser 1991) Linear, 2.0 kb, integration, expression (Yoshizaki et al. 1992) Circular, 4.7 kb, expression (Liu et al. 1990b) Linear, 4.7 kb, expression (Rahman & Maclean 1992a) Integration, expression (Winkler et al. 1991) Expression (induced) (Inoue et al. 1992) CAT expression (induced) (Inoue et al. 1992) Transient expression in embryos (Hong et al. 1993) Temporal expression in embryos, noninducible expression by heavy metals (Winkler et al. 1991) Linear, persistence, expression (Muller et al 1992) Expression in embryo (Gong et al. 1991) Persistence, tissue-specific transient expression (Inoue et al. 1991) Linear; no integration but expression in embryos (McEvoy et al. 1988) Linear, supercoiled, persistence, expression (Muller et al. 1992)

Transgenie fish Zebra fish

Loach Sea-bream /3-Galactosidase gene + CMV promoter Zebra fish and cat-fish j8-Galactosidase gene + mouse heat shock promoter Zebra fish Luciferase gene + salmon GnRH promoter Zebra fish and medaka Luciferase gene + CMV/RSV/SV promoter Zebra fish and trout Luciferase gene + RSV promoter Medaka neo gene + RSV promoter Goldfish Hygromycin gene + SV40 promoter Zebra fish Chicken 5-crystallin gene + own promoter Medaka

91 Supercoiled, integration, expression (Fl), germ-line transmission (Bayer & Campos-Ortega 1992) Integration, germ-line transmission (Culp et al. 1991) High-velocity microprojectiles technique, expression (Zelenin et al. 1991) Transient expression in embryos (Cavari et al. 1993) Linear, expression (Volckaert et al. 1991) Linear, 4.8 kb, expression (Fl), germ-line transmission (Bayer & Campos-Ortega 1992) Expression (Alestrom et al. 1991) Expression in embryos (Gibbs et al. 1991) Expression observed during embryogenesis (Tamiya et al. 1990) Linear, 5.3 kb, integration, expression by detecting mRNA (Yoon et al. 1990) Linear, integration, germ-line transmission (Stuart et al. 1988) Circular, expression at different embryonic developmental stages (Inoue et al. 1989) Integration, expression in different tissues in 7-day-old embryo (Ozato et al. 1986) Expression in lens (Ozato 1991)

92

Norman Maclean and Azizur Rahman

Mouse tyrosinase + own promoter Medaka Flounder antifreeze protein gene + own promoter Atlantic salmon

Carp a-globin gene + own promoter Trout Bovine growth hormone cDNA + RSV promoter Northern pike Bovine growth hormone gene + avian retroviral LTR Tilapia Human growth hormone + mouse metallothionien promoter Goldfish Loach

Channel cat-fish Tilapia Atlantic salmon

Trout Crucian carp Red crucian carp

Linear, 4.5 kb, integration, expression, germ-line transmission (Matsumoto et al 1992) Linear, 7.8 kb, integration (Fletcher et al 1988) Integration, germ-line transmission (Hew et al 1991) Linear, 2.2 kb, integration, germ-line transmission (Yoshizaki et al 1991a, b) Circular, integration, expression in serum and growth enhancement (Guise et al 1992) Linear, 8.5 and 3.5 kb, integration (Phillips et al 1992) Linear, 9.4 kb, integration (Zhu et al 1985) Linear, 3.49 kb, expression assumed (Zhu et al 1986) Integration (Chen et al 1990) Electroporation, integration (Xie et al 1993) Linear, 2.9 kb, integration (Dunham et al 1987; Hayat et al 1991) Linear, 4.0 kb, integration (Brem et al 1988) Linear and circular, 2.5 and 6.3 kb; integration, expression in embryos (Rokkones et al 1989) Linear and circular, 6.65 kb, integration, germ-line transmission (Guyomard et al 1989) Integration (Chen et al 1990) Integration (Chen et al 1990) Electroporation, integration (Xie et al 1993)

Trans genie fish Silver crucian carp Mirror carp Red carp Common carp Common carp Medaka Indian major carp Human growth hormone gene + SV40 promoter Trout Human growth hormone gene + chicken /3-actin promoter Medaka Rat growth hormone + mouse metallothionein promoter Trout

Tilapia Trout growth hormone cDNA + RSV promoter Common carp

Channel cat-fish Trout growth hormone cDNA + mouse metallothionein promoter Medaka

93 Integration, expression (Chen et al. 1990) Integration (Chen et al. 1990) Integration (Chen et al. 1990) Linear, integration (Powers et al., 1991) 4.1 kb, linear, integration (Hernandez et al. 1991) Persistence, growth increase (Lu et al. 1991) 16% survival in injected embryos (Alok & Khillan 1989) Linear, circular, integration (Chourrout et al 1986) Integrated, growth enhancement, germ-line transmission (Lu et al. 1991) Linear and circular, 6.6 and 8.9 kb, integration, germ-line transmission (Maclean et al. 1987; Penman et al. 1990) Linear, integration (Guyomard et al. 1989) Linear, 6.6 kb, integration (Rahman & Maclean, 1992a) Linear, 5.2 kb, integration, expression and germ-line transmission in fish developed from eggs injected at 2- and 4-cell stage; expression observed in fish developed from eggs injected at 1-cell stage (Zhang et al. 1990) Linear, 5.2 kb, integration (Powers et al. 1991) Linear, 5.2 kb, electroporation technique, integration, germ-line transmission (Inoue et al. 1990)

94

Norman Maclean and Azizur Rahman

Coho salmon growth hormone cDNA + RSV-LTR Channel cat-fish Gold fish Chinook salmon growth hormone gene + antifreeze promoter Atlantic salmon

Linear, 5.5 kb, integration (Powers et al. 1991) Integration (Zhang et al. 1993) Linear, integration, expression in blood serum and growth enhancement (Du et al. 1992)

References Agellon, L. B. & Chen, T. T. (1986). Rainbow trout growth hormone: Molecular cloning of cDNA and expression in E. coli. DNA 5(6), 463-71. Alestrom, P., Klungland, H., Kisen, G. & Anderson, O. (1991). Fish GnRH gene and molecular approaches for control of sexual maturation. In Second International Marine Biotechnology Conference, p. 65. Baltimore, MD (abstract). Alok, D., & Khillan, J. S. (1989). Gene transfer in fish embryos for increased productivity. Theoriogenology 31(1), 167. Bayer, T. A. & Campos-Ortega, J. A. (1992). A transgene containing lacZ is expressed in primary sensory neurons in zebrafish. Development 115, 421-46. Bearzotti, M., Perrot, E., Michard-Vanhee, C , Jolivet, G., Attal, J., Theron, M.-C, Puissant, C , Dreano, M., Kopchick, J. J., Powell, R., Gannon, F., Houdebine, L. M. & Chourrout, D. (1992). Gene expression following transfection offish cells. J. Biotechnol. 26, 315-25. Bello, B. & Couble, P. (1990). Specific expression of a silk encoding gene Bombyx in the anterior salivary gland of Drosophila. Nature 346, 480-2. Bishop, J. O. & Smith, P. (1989). Mechanism of chromosomal integration of microinjected DNA. Mol. Biol. Med. 6, 283-98. Bonham, K., Zafarullah, M. & Gedamu, L. (1987). The rainbow trout metallothionein: Molecular cloning and characterization of two distinct cDNA sequences. DNA 6(4), 519-28. Breitman, M. L. & Bernstein, A. (1992). Engineering cellular deficits in transgenic mice by genetic ablation. In Transgenic Animals, ed. F. Grosveld & G. Kollias, pp. 127-46. London: Academic Press. Brem, G., Brenig, B., Horstgen-Schwark, G. & Winnacker, E. L. (1988). Gene transfer in tilapia (Oreochromis niloticus). Aquaculture 68, 209-19. Brinster, R. L., Chen, H. Y., Trumbauer, M. E., Yagle, M. K. & Palmiter, R. D. (1985). Factors affecting the introduction of foreign DNA into transgenic mice by microinjecting eggs. Proc. Natl. Acad. Sci. USA 82, 4438-42. Buono, R. J. & Linser, P. J. (1991). Transgenic animals by electroporation. Biorad Bulletin: Bio-Radiations 79, 1-2. Buono, R. J. & Linser, P. J. (1992). Transient expression of RSV/CAJin transgenic zebrafish made by electroporation. Invest. Ophthalmol. Vis. Sci. 33, 1006. Byrnes, L. & Ganon, F. (1990). Atlantic salmon (Salmo salar) serum albu-

Trans genie

fish

95

min: cDNA sequence, evolution and tissue expression. DNA Cell. Biol 9(9), 647-55. Cao, Q. P., Duguay, S. J., Plisetskaya, E., Steiner, D. F. & Chan, S. J. (1989). Nucleotide sequence and growth hormone regulated expression of salmon insulin-like growth factor I mRNA. Mol. Endocrinol. 3(12); 2005-10. Cavari, B., Funkenstein, B., Chen, T. T., Gonzalez-Villasenor, L. I. & Schartl, M. (1993). Effect of growth hormone on the growth rate of the gilthead seabream (Sparus aurata), and use of different constructs for the production of transgenic fish. Aquaculture 111, 189—97. Chang, J. H., Huang, F. L., Wang, Y. C , Chang, Y. S. & Lo, T. B. (1992). Organization and nucleotide sequence of carp gonadotropin a sub unit gene. Biochim. Biophys. Ada 1192(2), 239-42. Chang, T. N. & Chang, W. C. (1987). Cloning and sequencing of carp /3-crystallin cDNA. Biochim. Biophys. Ada 910(1), 89-92. Chang, T., Lin, C. L., Chen, P. H. & Chang, W. C. (1991). Gamma crystallin genes in carp: Cloning and characterization. Biochim. Biophys. Ada 1090(2); 261-4. Chang, Y. S., Huang, F. L. & Lo, T. B. (1992a). Isolation and sequence analysis of carp gonadotropin B subunit gene. Mol. Mar. Biol. Biotechnol. 1(2), 97-105. Chang, Y. S., Huang, F. L. & Lo, T. B. (1992b). Molecular cloning of silver carp and bighead carp prolactin. Gen. Comp. Endocrinol. 87(2), 260-5. Chao, S. C , Pan, F. M. & Chang, W. C. (1988). Nucleotide sequence of carp prolactin cDNA. Nucl. Acids Res. 16 (19), 9350. Chen, T. T., Lin, C. M., Zhu, Z., Gonzalez-Villasenor, L. I., Dunham, R. A. & Powers, D. A. (1990). Gene transfer, expression and inheritance of rainbow trout and human growth hormone genes in carp and loach. In Transgenic Models in Medicine and Agriculture, ed. R. B. Church, pp. 127-39. New York: Wiley-Liss. Cheng, C. M., Lin, C. M. & Chen, T. T. (1991). Isolation and characterization of striped bass GHcDNA sequence. In Second International Marine Biotechnology Conference, p. 80. Baltimore, MD (abstract). Chiou, C. S., Chen, H. T. & Chang, W. C. (1990). The complete nucleotide sequence of the growth hormone gene from common carp (C. carpio). Biochim. Biophys. Ada 1087(1), 91-4. Chong, S. S. C. & Veilkind, J. R. (1989). Expression of fate of CAT reporter gene microinjected into fertilized medaka (Oryzias latipes) eggs in the form of plasmid DNA, recombinant phage particles and its DNA. Theor. Appl. Genet. 78, 369-80. Chourrout, D. R., Guyomard, R. & Houdebine, L. M. (1986). High efficiency gene transfer in rainbow trout (Salmo gairdneri Rich.) by microinjection into egg cytoplasm. Aquaculture 51, 143-50. Chourrout, D. R., Guyomard, R. & Houdebine, L. M. (1990). Techniques for the development of transgenic fish: A review. In Transgenic Models in Medicine and Agriculture, ed. R. B. Church, pp. 89-99. New York: Wiley-Liss. Cloud, J. G. (1990). Strategies for introducing foreign DNA into the germ line offish. /. Reprod. Fert. Suppl. 41, 107-16. Collodi, P., Kamei, Y., Ernst, T., Miranda, C , Buhler, D. R. & Barnes, D. W. (1992). Culture of cells from zebrafish (Brachydanio redo) embryo and adult tissues. Cell Biol. Toxicol. 8, 43-61.

96

Norman Maclean and Azizur Rahman

Culp, P., Niisslein-Volhard, C. & Hopkins, N. (1991). High frequency germline transmission of plasmid DNA sequences injected into fertilized zebrafish eggs. Proc. Natl. Acad. Sci. USA 88, 7953-7. Datta, U. (1988). Cloning and characterization of a highly repetitive fish nucleotide sequence. Gene 62(2), 331-6. Davies, P. L., Roach, A. H. & Hew, C. L. (1982). DNA sequence coding for an antifreeze protein precursor from winter flounder. Proc. Natl. Acad. Sci. USA 79, 335-9. Devlin, R. H. & Donaldson, E. M. (1992). Containment of genetically altered fish with emphasis on salmonids. In Transgenie Fish, ed. C. L. Hew & G. L. Fletcher, pp. 229-65. Singapore: World Scientific. Ding, J. L., Ho, B., Valotaire, Y., LeGuellec, K., Lim, E. H., Tay, S. P. & Lam, T. J. (1990). Cloning characterization and expression of vitellogenin gene of Oreochromis aureus. Biochem. Int. 20(4), 843-52. Disney, J. E., Johnson, K. R., Banks, D. K. & Thorgaard, G. H. (1988). Maintenance of foreign gene expression and independent chromosome fragments in adult transgenic rainbow trout and their offspring. J. Exp. Zool. 248, 335-44. Du, S. J., Gong, Z., Fletcher, G. L., Shears, M. A., King, M. J., Idler, D. R. & Hew, C. L. (1992). Growth enhancement in transgenic Atlantic salmon by the use of an 'all fish' chimeric growth hormone gene construct. Bio/ Technology, 10(10), 176-81. Dunham, R. A., Eash, J., Askins, J. & Townes, T. M. (1987). Transfer of the metallothionein-human growth hormone fusion gene into channel catfish. Trans Am. Fish. Soc. 116, 87-91. Eiken, H. G., Njolstrad, P. R., Molven, A. & Fjose, A. (1987). A zebrafish homeobox containing gene with embryonic transcription. Biochem. Biophys. Res. Commun. 149, 1165-71. Englander, E. & Moav, B. (1989). Cloning and characterization of a histone gene family in tilapia fish. Biochem. Biophys. Acta 1007(3), 277-82. Etkin, L. D. & Pearman, B. (1987). Distribution, expression, and germ line transmission of exogenous DNA sequences following microinjection into Xenopus laevis eggs. Development 99, 15-23. Evans, M. J. (1989). Potential for the genetic manipulation of mammals. Mol. Biol. Med. 6, 557-65. Evans, M. J. & Kaufman, M. H. (1981). Establishment in culture of pluripotential cells from mouse embryos. Nature 292, 154-6. Fjose, A., Molven, A. & Eiken, H. G. (1988). Molecular cloning and characterization of homeobox containing genes from Atlantic salmon. Gene 62(1), 141-52. Fletcher, G. L. & Davies, P. L. (1991). Transgenic fish for aquaculture. In Genetic Engineering, ed. J. K. Setlow, Vol. 13, pp. 331-70. Singapore: World Scientific. Fletcher, G. L., Shears, M. A., King, M. J., Davies, P. L. & Hew, C. L. (1988). Evidence for antifreeze protein gene transfer in Atlantic salmon (Salmo salar). Can. J. Fish Aquat. Sci. 45, 352-7. Fletcher, G. L., Davies, P. L. & Hew, C. L. (1992). Genetic engineering of freeze-resistant Atlantic salmon. In Transgenic Fish, ed. C. L. Hew & G. L. Fletcher, pp. 190-208. Singapore: World Scientific. Forbes, D. J., Kirschner, M. W. & Newport, J. W. (1983). Spontaneous formation of nucleus-like structures around bacteriophage DNA microinjected into Xenopus eggs. Cell 34, 13-23.

Trans genie

fish

97

Friedenreich, H. & Schartl, M. (1990). Transient expression directed by homologous and heterologous promoter and enhancer sequences in fish cells. Nucl. Acids Res. 18, 3299-305. Funkenstein, B., Chen, T. T., Powers, D. A. & Cavari, B. (1992). Cloning and sequencing of the gilthead sea bream (Spams aurata) growth hormone-encoding cDNA. Gene 103(2), 243-7. Garber, A. T., Relief, J. D. & Dixon, G. H. (1989). Isolation of dynein heavy chain cDNA's from trout testis which predict an extensive carboxyl-terminal helical coiled domain. EMBO J. 8(6), 1727-34. Ghaffari, S. H. & Lobb, C. J. (1989). Nucleotide sequence of channel catfish heavy chain cDNA and genomic blot analysis: Implications for the phylogeny of Ig heavy chains. /. Immunol. 143(8), 2730-9. Gibbs, P. D. L., Peek, A. & Thorgaard, G. H. (1991). An in vitro screen for transgenic fish utilizing the luciferase reporter gene. In Second International Marine Biotechnology Conference, p. 79. Baltimore, MD (abstract). Gong, Z., Hew, C. L. & Veilkind, J. R. (1991). Functional analysis and temporal expression of promoter regions from fish antifreeze protein genes in transgenic Japanese medaka embryos. Mol. Mar. Biol. Biotechnol. 1(1), 64-72. Gonzalez-Villasenor, L. I., Zhan, P., Chen, T. T. & Powers, D. A. (1988). Molecular cloning and sequencing of coho salmon growth hormone cDNA. Gene 65, 239-46. Guise, K. S., Kapuscinski, A. A. R., Hackett, P. B. & Faras A. J. (1991). Gene transfer in fish. In Transgenic Animals: Proceedings of a Symposium on Transgenic Technology in Medicine and Agriculture, ed. N. L. First & F. P. Haseltine, pp. 295-306. London: Butterworth-Heinemann. Guise, K. S., Hackett, P. B. & Faras, A. J. (1992). Transfer of genes encoding neomycin resistance, chloramphenicol acetyl transferase and growth hormone into goldfish and northern pike. In Transgenic Fish, ed. C. L. Hew & G. L. Fletcher, pp. 142-64. Singapore: World Scientific. Guyomard, R., Chourrout, D., Lerox, C , Houdebine, L. & Pourrain F. (1989). Integration and germline transmission of foreign genes microinjected into fertilized trout eggs. Biochimie 71, 857-63. Hallerman, E. M., Schneider, J. F., Gross, M., Liu, Z., Yoon, S. J., He, L., Hackett, P. B., Faras, A. J., Kapuscinski, A. R. & Guise, K. S. (1990). Gene expression promoted by the RSV long terminal repeat element in transgenic goldfish. Anim. Biotechnol. 1(1), 79-93. Hallerman, E. M. & Kapuscinski, A. R. (1992). Ecological and regulatory uncertainties associated with transgenic fish. In Transgenic Fish, ed. C. L. Hew & G. L. Fletcher, pp. 190-208. Singapore: World Scientific. Hansen, E., Fernandes, K., Goldspink, G., Butterworth, P., Umeda, P. K. & Chang, K. (1991). Strong expression of foreign genes following direct injection into fish muscle. FEBS Lett. 290, 73-6. Hayat, M., Joyce, C. P., Townes, C. M., Chen, T. T., Powers, D. A. & Dunham, R. A. (1991). Survival and integration rate of channel catfish and common carp embryos microinjected with DNA at various developmental stages. Aquaculture 99, 249-55. He, L., Zhu, Z., Faras, A. J., Guise, K. S., Hackett, P. B. & Kapuscinski, A. R. (1992). Characterization of Alu repeats in zebrafish genome. Mol. Mar. Biol. Biotechnol. 1(2), 125-35. Heierhorst, J., Morley, S. D., Figueroa, J., Krentler, C , Lederis, K. &

98

Norman Maclean and Azizur Rahman

Richler, D. (1989). Vasotocin and isotocin precursors from white sucker Catostomus commersoni: Cloning and sequence analysis of the cDNAs. Proc. Natl. Acad. Sci. USA 86(14), 5242-6. Heierhorst, J., Mahlmann, S., Morley, S. D., Coe, I. R., Sherword, N. M. & Richter, D. (1990). Molecular cloning of two distinct vasotocin precursor cDNAs from chum salmon (Onchorynchus keta) suggests an ancient gene duplication. FEBS Lett. 260(2), 310-14. Hernandez, O., Castro, F. O., Aquilar, A., Uliver, C , Perez, A., Herrera, L. & de la Fuente, J. (1991). Gene transfer in common carp {Cyprinus carpio L.) by microinjection into the germinal disc. Theriogenology 35(3), 625-32. Hew, C. L., Fletcher, G. L. & Davies, P. L. (1991). Antifreeze protein gene transfer in Atlantic salmon. In Second International Marine Biotechnology Conference, p. 72. Baltimore, MD (abstract). Hew, C. L. & Fletcher G. L., eds. (1992). Transgenic Fish. Singapore: World Scientific. Hindar K. (1993). Genetically engineered fish and their possible environmental impact. NINA Oppdragsmelding 215, 1-48. Hong, Y., Winkler, C , Brem, G. & Schartl, M. (1993). Development of a heavy metal-inducible fish specific expression vector for gene transfer in vitro and in vivo. Aquaculture 3, 215-16. Houdebine, L. M. & Chourrout, D. (1991). Transgenesis in fish. Experientia 47, 891-7. Ho-Walter, K. K., Tsang, W. H. & Dias, N. P. (1989). Cloning of the grass carp growth hormone cDNA. Biochem. Biophys. Res. Commun. 161(3), 1239-43. Huang, R. C. C , Price, J. L. & Gourlie, B. (1990). Regulation of antifreeze gene expression in winter flounder and in transgenic fish cells. In Transgenic Models in Medicine and Agriculture, ed. R. B. Church, pp. 109-26. New York: Wiley-Liss. Inoue, K., Ozato, K., Kondoh, H., Iwamatsu, T., Wakamatsu, Y., Fujita, T. & Okada, T. S. (1989). Stage-dependent expression of the chicken delta crystallin gene in transgenic fish embryos. Cell. Diff. Dev. 27, 57-68. Inoue, K., Yamashita, S., Hata, J., Kabeno, S., Asada, S., Nagahisa, E. & Fujita, T. (1990). Electroporation as a new technique for producing transgenic fish. Cell. Diff. Dev. 29, 123-8. Inoue, K., Yamashita, S., Akita, N., Mitsuboshi, T., Nagahisa, E., Shiba, T. & Fujita T. (1991). Histochemical detection of foreign gene expression in rainbow trout. Nippon Suisan Gakkaishi 57(8), 1511-17. Inoue, K., Akita, N., Shiba, T., Satake, M. & Yamashita, S. (1992). Metal-inducible activities of metallothionein promoters in fish cells and fry. Biochem. Biophys. Res. Commun. 185(3), 1108-14. Ivies, Z., Izsvak, Z. & Hackett, P. B. (1993). Enhanced incorporation of transgenic DNA into zebrafish chromosomes by a retroviral integration protein. Mol. Mar. Biol. Biotechnol. 2(3), 162-73. Johansen, B., Johansen, O. C. & Valla, S. (1989). The complete nucleotide sequence of the growth hormone gene from Atlantic salmon. Gene 77(2), 317-24. Johnson, R. L., Kathryn, B. G., Todd, C. Z., Marcus, F. B., Shannath, L. M., Jeremy, N. & Koji, N. (1993). Cloning and expression of goldfish opsin sequences. Biochemistry 32(1), 208-14. Keiichiro, H., Teruyuki, N. & Yoshikazu, K. (1990). Isolation of carp genes

Trans genie

fish

99

encoding major histocompatibility complex antigen. Proc. Natl. Acad. Sci. USA 87(17), 6863-7. Khoo, H. W., Ang, L. H., Lim, H. B. & Wang, K. Y. (1992). Sperm cells as vectors for introducing foreign DNA into zebrafish. Aquaculture 107, 119. Kido, Y., Aono, M., Yamaki, T., Matsumoto, K. I., Murata, S., Saneyoshi, M. & Okada, N. (1991). Shaping and re-shaping of salmonid genomes by amplification of tRNA-derived retroposons during evolution. Proc. Natl. Acad. Sci. USA 88, 2326-30. Klungland, H., Lorens, J. B., Anderson, O., Kisen, G. O. & Alestrom P. (1992). The Atlantic salmon prepro-gonadotropin releasing hormone gene and MRNA. Mol. Cell Endocrinol. 84(3), 167-74. Koval, A. P., Petrenko, A. I., Dmirenko, V. V. & Kavasan, V. M. (1989). Nucleotide sequence of chum salmon preproinsulin gene. Mol. Biol. (Moscow) 23(2), 473-80. Kozlov, A. P., Reshetnikov, V. L., Korzh, V. P. & Neifakh, A. A. (1988). Foreign DNA in the developing loach embryo (Misgurnus fossilis). Mol. Biol. 22, 1286-93. Krone, P. H. (1993). Several unique Hsp 90 genes are expressed during embryonic development of zebrafish. In Symposium on Advances in Molecular Endocrinology of Fish, p. 28. Toronto, 23-5 May (abstract). Latchman, D. S. (1991). Eukaryotic Transcription Factors. London: Academic Press. Le Guellec, K., Lawless, K., Valotaire, Y., Kress, M. & Tenniswood, M. (1988). Vitellogenin gene expression in male rainbow trout (Salmo gairdneri). Gen. Comp. Endocrinol. 71, 359-71. Lin, S., Long, W., Chen, J. & Hopkins, N. (1992). Production of germ line chimeras in zebrafish by cell transplants from genetically pigmented to albino embryos. Proc. Natl. Acad. Sci. USA 89, 4519-23. Liu, Z., Zhu, Z., Roberg, K., Faras, A., Guise, K., Kapusciniski, A. & Hackett, P. B. (1989). The j8-actin gene of carp (Ctenopharyngodon idella). Nucl. Acids Res. 14, 5850. Liu, Z., Zhu, Z., Roberg, K., Faras, A., Guise, K., Kapuscinski, A. R. & Hackett, P. B. (1990a). Isolation and characterization of j8-actin gene of carp (Cyprinus carpio). DNA sequence. Journal of DNA Sequencing and Mapping 1, 125—36. Liu, Z., Moav, B., Faras, A. J., Guise, K. S., Kapuscinski, A. R., Hackett, A. R. & Hackett, P. B. (1990b). Development of expression vectors for transgenic fish. Bio/Technology 8, 1268-72. Lorens, J., Nerland, A. H., Male, R., Lossius, I., Telle, W. & Todland, G. (1989). The nucleotide sequence of Atlantic salmon growth hormone cDNA. Nucl. Acids Res. 17(6), 2352. Lu, J. K., Chrisman, C. L. & Chen, T. T. (1991). Integration and germline transmission of human growth hormone gene in medaka (Oryzias latipes). In Second International Marine Biotechnology Conference, p. 77. Baltimore, MD (abstract). Maclean, N. & Penman, D. J. (1990). The application of gene manipulation to aquaculture. Aquaculture 85, 1-20. Maclean, N., Penman, D. & Talwar, S. (1987). Introduction of novel genes into rainbow trout. Proceedings of the World Symposium on Selection, Hybridization and Genetic Engineering in Aquaculture, Vol. 2, pp. 32533. Berlin: Heineman.

100

Norman Maclean and Azizur Rahman

Maclean, N., Iyengar, A., Rahman, A., Sulaiman, Z. & Penman, D. (1992). Transgene transmission and expression in rainbow trout and tilapia. Mol. Mar. Biol. Biotechnol. 1(4-5), 355-65. Matsumoto, J., Akiyama, T., Hirose, E., Nakamura, M., Yamamoto, H. & Takeuchi, T. (1992). Expression and transmission of wild type pigmentation in the skin of transgenic orange-colored variants of medaka (Oryzias latipes) bearing the gene for mouse tyrosinase. Pigment Cell Res. 5, 3227. McEvoy, T., Stack, M., Kaene, B., Barry, T., Sreenan, J. & Gannon, F. (1988). The expression of a foreign gene in salmon embryos. Aquaculture 68, 27-37. Mehtali, M., Le Meur, M. & Lathe, R. (1990). The methylation-free status of a house keeping transgene is lost at high copy number. Gene 91, 179-84. Miyata, M., Hirong, I. & Akoi, T. (1991). Nucleotide sequences of the carp (Cyprinus carpio) a-globin gene family. In Second International Marine Biotechnology Conference, p. 92. Baltimore, MD (abstract). Moav, B., Liu, Z., Groll, Y. & Hackett, P. B. (1992). Selection of promoters for gene transfer into fish. Mol. Mar. Biol. Biotechnol. 1(4-5), 338-45. Moir, R. D. & Dixon, G. H. (1988a). Characterization of protamin gene from the chum salmon. J. Mol. Evol. 27(1), 8-16. Moir, R. D. & Dixon, G. H. (1988b). A repetitive DNA sequence in the salmonid fishes similar to a retroviral long terminal repeat (rainbow trout and chum salmon). /. Mol. Evol. 27(1), 8-16. Momota, H., Kosugi, R., Hiramatsu, H., Ohgai, H., Hara, A. & Ishioka, H. (1988a). Nucleotide sequence of cDNA encoding the pregrowth hormone of red sea bream. Nucl. Acids Res. 16(7), 3107. Momota, H., Kosugi, R., Ohgai, H., Hara, A. & Ishioka, H. (1988b). Amino acid sequence of flounder growth hormone deduced from a cDNA sequence. Nucl. Acids Res. 16(2), 10362. Mori, H., Kimura, T., Tsunenari, T., Ogura, T., Niki, H., Ezaki, B., Shigesada, K. and Hiraga, S. (1989). The complete cDNA sequence for the premature form of growth hormone of flounder. Nucl. Acids Res. 17(10), 3977. Moyer, S. P., Ma, D. P., Thomas, T. L. & Gold, J. R. (1988). Characterization of a highly repetitive satellite DNA from the cyprinid fish Notropis lutrensis. Comp. Biochem. Physiol. B 91b(4), 639-46. Muller, F., Ivies, Z., Erdelyi, F., Papp, T., Varadi, L., Horvath, L., Maclean, N. & Orban, L. (1992). Introducing foreign genes into fish eggs with electroporated sperm as a carrier. Mol. Mar. Biol. Biotechnol. 1, 276-81. Muller, F., Lele, Z., Varadi, L., Menczel, L. & Orban, L. (1993). Efficient transient expression system based on square pulse electroporation and in vivo luciferase assay of fertilized fish eggs. Fed. Eur. Biochem. Soc. 324(1), 27-32. Murphy, M. F., Collier, J., Koutz, P. & Howard, B. (1990). Nucleotide sequence of the trout metallothionein A gene 5' regulatory region. Nucl. Acids Res. 18(15), 4622. Ninomiya, T., Iwabuchi, T., Soga, Y. & Yuki, A. (1990). Direct sequencing of flanking regions of a transgene amplified by inverted PCR. Agric. Biol. Chem. 54(7), 1869-72. Njolstad, P. R., Molven, A., Apold, J. & Fjose, A. (1990). The zebrafish ho-

Trans genie

fish

101

meobox gene hox-2.2: Transcription unit, potential regulatory regions and in situ localization of transcripts. EMBO J. 9, 515-24. Okawara, Y., Morley, S. D., Burzio, L. O., Zwiers, H., Lederis, K. & Richter, D. (1988). Cloning and sequence analysis of cDNA for corticotropinreleasing factor precursor from the teleost (Catostomus commersoni). Proc Natl. Acad. Sci. USA 85(22), 8439-43. Ozato, K. (1991). Exogenous gene transfer and expression in medaka embryos. In Second International Marine Biotechnology Conference, p. 64. Baltimore, MD (abstract). Ozato, K., Kondoh, H., Inohara, H., Iwamatsu, T., Wakamatsu, Y. & Okada T. S. (1986). Production of transgenic fish: Introduction and expression of chicken delta crystallin gene in medaka embryos. Cell Diff. 19, 23744. Ozato, K., Inoue, K. & Wakamatsu, Y. (1989). Transgenic fish: Biological and technical problems. Zool. Sci. 6, 445-57. Penman, D. J., Beeching, A. J., Penn, S. & Maclean, N. (1990). Factors affecting survival and integration following microinjection of novel DNA in rainbow trout eggs. Aquaculture 85, 35-50. Penman, D. J., Iyengar, A., Beeching, A. J., Rahman, A., Sulaiman, Z. & Maclean, N. (1991). Patterns of transgene inheritance in rainbow trout (Oncorhynchus mykiss). Mol. Reprod. Dev. 30, 201-6. Phillips, C. P., Cohler, C. C. & Muhlach, W. (1992). Procedural protocol, survival to hatching and plasmid DNA fate after microinjection into tilapia zygotes. J. World Aqua. Soc. 23(2), 98-113. Poncelet, A. C , Sekkali, B., Swennen, D., Brim, H., Martial, J. A. & Belayew, A. (1993). Transcriptional regulation of growth hormone and prolactin gene expression in tilapia. In Symposium on Advances in Molecular Endocrinology in Fish, p. 14. Toronto, 23-5 May (abstract). Powers, D. A., Gonzales-Villasenor, L. L, Zhang, P., Chen, T. T. & Dunham, R. A. (1991). Studies on transgenic fish: Gene transfer, expression and inheritance. In Transgenic Animals: Proceedings of a Symposium on Transgenic Technology in Medicine and Agriculture, ed. N. L. First & F. P. Haseltine, pp. 98-113. London: Butterworth-Heinemann. Powers, D. A., Chen, T. T. & Dunham, R. A. (1992a). Transgenic fish. In Transgenesis: Applications of Gene Transfer, ed. J. A. H. Murray, pp. 233-50. New York: Wiley. Powers, D. A., Hereford, L., Cole, T., Creech, K., Chen, T. T., Lin, C. M., Kight, K. & Dunham, R. (1992b). Electroporation: A method for transferring genes into the gametes of zebrafish (Ictalurus punctatus), and common carp (Cyprinus carpio). Mol. Mar. Biol. Biotechnol. 1(4-5), 301-8. Rahman, M. A. & Maclean, N. (1992a). Production of transgenic tilapia (Oreochromis niloticus) by one-cell-stage microinjection. Aquaculture 105, 219-32. Rahman, A. & Maclean, N. (1992b). Fish transgene expression by direct injection into fish muscle. Mol. Mar. Biol. Biotechnol. 1(4-5), 286-9. Raphael, B. & Violet, D. (1992). Structure and sequence of the growth hormone - encoding gene from Tilapia nilotica. Gene 113(2), 245-50. Rentier-Delrue, F., Swennen, D., Philippart, J. C , Hoir, C. L., Lion, M., Benrubi, O. & Martial, J. A. (1989a). Tilapia (Oreochromis niloticus) growth hormone: Molecular cloning of cDNA and expression in E. coli. DNA 8(4), 271-8.

102

Norman Maclean and Azizur Rahman

Rentier-Delrue, F., Swennen, D., Prunet, P., Lion, M. & Martial, J. A. (1989b). Tilapia prolactin: Molecular cloning of two cDNA's and expression of E. coli. DNA 8, 261-70. Rokkones, E., Alestrom, P., Skjervold, H. & Gautvik, K. M. (1989). Microinjection and expression of a mouse metallothionein fusion gene in fertilized salmonid eggs. J. Comp. Physiol. B 158, 751-8. Rossant, J. & Hopkins, N. (1992). Of fin and fur: Mutational analysis of vertebrate embryonic development. Genes Dev. 6, 1-13. Schaffer-Ridder, M., Wang, Y. & Hofschneider, P. H. (1982). Liposomes as gene carriers: Efficient transformation of mouse L cells by thymidine kinase gene. Science 215, 166-8. Schneider, J. F., Myster, S. H., Hackett, P. B., Guise, K. S. & Faras, A. J. (1992). Molecular cloning and sequence analysis of the cDNA for northern pike (Esox lucius) growth hormone. Mol. Mar. Biol. Biotechnol. 1(2), 106-12. Scott, G. K., Hew, C. L. & Davies, P. L. (1985). Antifreeze protein genes are tandemly linked and clustered in the genome of the winter flounder. Proc. Natl. Acad. Sci. USA 82, 2613-17. Sekine, S., Akiko, S., Hiromichi, I., Hiroshi, K. & Seiga, I. (1989a). Molecular cloning and sequence analysis of chum salmon gonadotropin cDNAs. Proc. Natl. Acad. Sci. USA 86(22), 8645-9. Sekine, S., Mizukami, T., Saito, A., Kawauchi, H. & Iton, S. (1989b). Isolation and characterization of a novel growth hormone cDNA from chum salmon (O. keta). Biochim. Biophys Acta 1009(2), 117-20. Shamblott, M. J. & Chen, T. T. (1991). Characterization of three cDNA of insulin-like genes from the liver of rainbow trout. In Second International Marine Biotechnology Conference, p. 92. Baltimore, MD (abstract). Shears, M. A., Fletcher, G. L., Hew, C. L., Ganthier, S. & Davies, P. L. (1991). Transfer, expression and stable inheritance of antifreeze protein genes in Atlantic salmon (Salmo salar). Mol. Mar. Biol. Biotechnol. 1(1), 58-63. Shen, X. Z., Wang, Y., Welt, M., Liu, D. M. & Leung, F. C. (1991). Molecular cloning and sequence analysis of the chum salmon growth hormone genomic gene. In Second International Marine Biotechnology Conference, p. 93. Baltimore, MD (abstract). Shiokawa, K., Sameshima, M., Tashiro, K., Muira, T., Nakakura, N. & Yamana, K. (1986). Formation of nucleus-like structure in the cytoplasm of lambda-DNA injected fertilised eggs and its partition into blastomeres during early embryogenesis in Xenopus laevis. Dev. Biol. 116, 539-42. Shiokawa, K., Tashiro, K., Yamana, K. & Sameshima, M. (1987). Electron microscopic studies of giant nucleus-like structure formed by lambda DNA introduced into the cytoplasm of Xenopus fertilized eggs and embryos. CellDiff 20, 253-61. Shiokawa K., Fu, Y., Hosokawa, K. & Yamana, K. (1990). Temporally uncontrolled expression of linearized plasmid DNA which carries bacterial chloramphenicol acetyltransferase gene with Xenopus xenopus fertilized eggs. Roux's Arch. Dev. Biol. 199, 174-80. Sin, F. Y. T., Barkley, A. C , Walker, S. P., Sin, P. L., Symonds, J. E., Hawke, L. & Hopkins, C. C. (1993). Gene transfer in chinook salmon by electroporating sperm in the presence of PRSV-/ac Z DNA. Aquaculture 111, 57-69.

Trans genie

fish

103

Sippel, A. E., Saueressig, H., Winter, D., Grewal, T., Faust, N., Hecht, A. & Bonifer, C. (1992). The regulatory domain organization of eukaryotic genomes: Implications for stable gene transfer. In Transgenie Animals, ed. F. Grosveld & G. Kollias, pp. 1-25. London: Academic Press. Song, S., Trinh, K., Hew, C. L., Hwang, S. J., Belkhode, S. & Idler, D. R. (1988). Molecular cloning and expression of chum salmon prolactin cDNA. Eur. J. Biochem. 172(2), 279-85. Sorimachi, H., Emori, Y., Kawasaki, H., Suzuki, K. & Inoue, Y. (1990a). Molecular cloning and characterization of cDNAs coding for apopolysialoglycoproteins in cherry salmon (Onchorynchus mason) eggs. J. Biochem. Tokyo 107(1), 61-7. Sorimachi, H., Emori, Y., Kawasaki, H., Suzuki, K. & Inoue Y. (1990b). Organization and primary sequence of multiple genes coding for the apopolysialoglycoproteins of rainbow trout. J. Mol. Biol. 211(1), 35-48. Sorokin, A. V., Mironov, A. A., Zlocheviskii, M. L., Deniyanova, N. G., Rebentish, B. A., Kozlov, Yu. I., Kavasan, V. M., Sova, V. V. et al. (1983). Nucleotide sequence of cloned insulin cDNA of O. keta. Dokl. AN. SSSR 269(3), 758-62. Stuart, G. R., Dixon, B. & Pohajdak B. (1992). Isolation of a putative retrovirus POL gene fragment from trout. Comp. Biochem. Physiol. 102B, 137— 42. Stuart, G. W., Mcmurray, J. V. & Westerfield M. (1988). Replication, integration and stable germline transmission of foreign sequences injected into early zebrafish embryos. Development 103, 403-12. Stuart, G. W., Veilkind, J. R., Mcmurray, J. V. & Westerfield M. (1990). Stable lines of transgenic zebrafish exhibit reproducible patterns of transgene expression. Development 109, 577-84. Suzuki, M., Hyodo, S. & Urano A. (1992). Cloning and sequence analysis of vasotocin and isotocin precursor's cDNA's in the masu salmon O. masou: Evolution of neurohypophysial hormone precursors. Zool. Sci. (Tokoyo) 9(1), 157-67. Szelei, J., Varadi, L., Muller, F., Erdelyi, F., Horvath, L., & Duda, E. (in press). Liposome mediated gene transfer in fish embryos. Takayama, Y., Wada, C , Kawauchi, H. & Ono, M. (1989). Structures of two genes coding for melanin-concentrating hormone of chum salmon. Gene 80(1), 65-73. Takeshita, S., Aoki, T., Fukumaki, Y. & Takagi, Y. (1984). Cloning and sequence analysis of a cDNA for the a-globin mRNA of carp Cyprinus carpio. Biochim. Biophys. Acta 783(3), 265-71. Tamiya, E., Sugiyama, T., Masaki, K., Hirose, A., Okoshi, T. & Karube, I. (1990). Spatial imaging of luciferase gene expression in transgenic fish. Nucl. Acids Res. 18(4), 1072. Tewari, R., Michard-Varhee, C , Perrot, E. & Chourrout, D. (1992). Mendelian transmission, structure and expression of transgenes following their injection into the cytoplasm of trout eggs. Transgen. Res. 1, 250-60. Thomas, K. R. & Capecchi, M. R. (1990). Targeted disruption of the mouse Int-1 proto-oncogene results in severe abnormalities in mid-brain and cerebellar development. Nature 346, 847-50. Thomason, D. B. & Booth, F. W. (1990). Stable incorporation of a bacterial gene into adult rat skeletal muscle in vivo. Am. J. Physiol. 258(3/1); 57882. Trendelenburg, M. F. & Gurdon, J. (1978). Transcription of cloned Xenopus

104

Norman Maclean and Azizur Rahman

ribosomal genes visualised after injection into oocyte nuclei. Nature 276, 292-4. Tsai, H. J., Lin, K. L., Lee, L. H. & Liu, J. J. (1991). Molecular cloning, sequencing and expression of yellowfin seabream growth hormone cDNA. In Second International Marine Biotechnology Conference, p. 94. Baltimore, MD (abstract). Vielkind, J. R. (1992). Medaka and zebrafish: Ideal as transient and stable transgenic systems. In Transgenic Fish, ed. C. Hew & G. L. Fletcher, pp. 72-91. Singapore: World Scientific. Volckaert, F., Hellemans, B., Daemen, D. & Ollevier, F. (1991). Gene transfer and expression of the CMV-/3 gal fusion gene in embryos of the African catfish and the zebrafish. In Second International Marine Biotechnology Conference, p. 91. Baltimore, MD (abstract). Winkfein, R. J., Connor, W., Mezquita, J. & Dixon, G. H. (1985). Histone H4 and H2B genes in rainbow trout. J. Mol. Evol. 22(1), 1-19. Winkfein, R. J., Moir, R. D., Krawetz, S. A., Blanco, J., States, J. C. & Dixon, G. H. (1988). A new family of repetitive transposon-like sequences in the genome of rainbow trout. Eur. J. Biochem. 176(20), 25564. Winkler, C , Veilkind, J. R. & Schartl, M. (1991). Transient expression of foreign DNA during embryonic and larval development of the medaka fish (Oryzias latipes). Mol. Gen. Genet. 226, 129-40. Wolff, J. A., Malone, R. W., Willams, P., Chong, W., Acsadi, G., Jani, A. & Feigner, P. L. (1990). Direct gene transfer into mouse muscle in vivo. Science 247, 1465-8. Wright, J. M. (1989). Nucleotide sequence, genomic organization and evolution of a major repetitive DNA family in tilapia (Oreochromis mossambicuslhornorum). Nucl. Acids Res. 17(13), 5071-80. Xie, Y., Liu, D., Zou, J., Li, G. & Zhu, Z. (1993). Gene transfer via electroporation in fish. Aquaculture 3, 207-13. Xiong, F. & Hew, C. L. (1991). Chinook salmon (Oncorhynchus tshawytcha) gonadotropin II p subunit gene encodes multiple mRNA. Can. J. Zool. 69(10), 2572-8. Xiong, F., Chin, R. A. & Hew, C. L. (1992). A gene encoding chinook salmon (Oncorhyncus tshawytcha) prolactin: Gene structure and potential cis-acting regulatory elements. Mol. Mar. Biol. Biotechnol. 1(2), 155— 64. Yoon, S. J., Hallerman, E. M., Gross, M. L., Liu, Z., Schneider, J. F., Faras, A. J., Hackett, P. B., Kapuscinski, A. R. & Guise, K. S. (1990). Transfer of the gene for neomycin resistance into goldfish (Carassius auratus). Aquaculture 85, 21-33. Yoshizaki, G., Oshiro, T. & Takashima, F. (1991a). Introduction of carp aglobin gene into rainbow trout. Nippon Suisan Gakkaishi 57(5), 819-24. Yoshizaki, G., Oshiro, T., Takashima, F., Hirono, I. & Aoki, T. (1991b). Germline transmission of carp a-globin gene introduced in rainbow trout. Nippon Suisan Gakkaishi 57(12), 2203-9. Yoshizaki, G., Kobayashi, S., Oshiro, T. & Takashima, F. (1992). Introduction and expression of CAT gene in rainbow trout. Nippon Suisan Gakkaishi 58(9), 1659-65. Zafarullah, M., Bonham, K. & Gedamu, L. (1988). Structure of the rainbow trout metallothionein B gene and characterization of its metal responsive region. Mol. Cell. Biol. 8(10), 4469-76.

Trans genie

fish

105

Zelenin, A. V., Alimov, A. A., Barmintezev, V. A., Beniumov, A. O., Zelenina, I. A., Krasnov, A. M. & Kolesnikov, V. A. (1991). The delivery of foreign genes into fertilized fish eggs using high velocity microprojectiles. FEESLett. 287(1-2), 118-20. Zhang, P., Hayat, M., Joyce, C , Gonzalez-Villasenor, L. I., Lin, C M., Dunham, R. A., Chen, T. T. & Powers, D. A. (1990). Gene transfer, expression and inheritance of pRSV-rainbowtrout-GH cDNA in the common carp, Cyprinus carpio (Linnaeus). Mol Reprod. Dev. 25, 3-13. Zhang, P., Zhou, J. & Wang, R. (1993). Gene transfer in goldfish (Carassius auratus), by oocyte microinjection. Aquaculture 111, 311. Zhu, Z., Li, G., He, L. & Chen, S. (1985). Novel gene transfer into fertilized eggs of goldfish {Carassius auratus, L.1758). Z. angew. Ichthyol. 1, 314. Zhu, Z., Xu, K., Li, G., Xie, Y. & He, L. (1986). Biological effects of human growth hormone gene microinjected into the fertilized eggs of loach Misgurnus anguillicaudatus (Cantor). Xexue Tongbao, Academia Sinica, (Wuban, P. R. China) 31, 988-90.

4 Transgenic birds K. SIMKISS

4.1 Introduction

Birds are virtually unique among domesticated species in being bred for the nutrient value of their ova, and few organisms have been subjected to as much genetic selection as poultry. These two features encapsulate both the attractions and the frustrations of producing transgenic birds. Current breeding practices in the poultry industry depend on natural genetic variation that is exploited by selective reproduction. New variants are introduced from mutations or translocations, but there is a progressive loss of 'wild-type' characteristics. By inserting foreign DNA into organisms it is possible to bypass reproductive barriers and induce gene flow between vastly different organisms from quite distinct populations. Thus, the potential benefits of genetic engineering are extremely attractive, for they could dramatically increase both the genetic repertoire and the rate of change of a particular breed. Unfortunately, one of the characteristics of birds that makes them so commercially attractive - their large eggs - is also one of the features that makes it necessary to devise new approaches to their genetic engineering. The most common way of introducing foreign DNA into mammals is microinjection into the male pronucleus (Palmiter & Brinster 1986). The large size of birds' eggs makes it virtually impossible to see these structures and, to make matters worse, polyspermy is common. There may be up to 20 male pronuclei in a fertilized fowl egg with no indication as to which will eventually fuse with the female pronucleus. For these reasons microinjection is difficult unless special methods are used, and as a result most approaches to avian transgenesis have used other techniques, based mainly upon gaining access to 106

Transgenie birds

107

the avian genome via a variety of embryonic cells. Fortunately, the bird's embryo is probably the best-studied example of vertebrate development, and this has been of great benefit to these approaches. Unfortunately, other aspects of avian biology are not as well documented. Avian karyotypes are characterized by a few megachromosomes and a large number of microchromosomes. The first attempts have been made at producing a genetic map of the fowl (Bumstead & Palyga 1992), and it is estimated that there are between 50,000 and 100,000 genes present (Shuman 1990). Only about one-thousandth of these have been cloned, however, so that a great deal of the potential value of genetic engineering remains unidentified (Bulfield 1990). It will be apparent, therefore, that it has not been possible simply to adapt the technology of mammalian transgenesis to birds. Instead, new methods have had to be devised, often with considerable ingenuity. The subject is now at the stage where progress is being made and where many of the results from previous studies of avian embryology and biology can be capitalized upon with all the potential benefits of disease resistance, metabolic manipulation and enhanced understanding.

4.2 Avian reproduction

4.2.1 Gametes 4.2.1.1 Spermatozoa Birds do not possess accessory reproductive organs that produce seminal fluid. There is, however, a lymphlike fluid generated during the erection of the phallic structures in the cloaca that is referred to as 'transparent fluid'. The role of this fluid during ejaculation is not clear, but fertilization can occur in its absence. Such semen contains about 6xlO 8 spermatozoa per cubic millimeter (Lake 1971). A number of attempts have been made to use the sperm as a vehicle for transferring adventitious genetic material into the germ-line. The simplest approach to this technique was that described by Pandey and Patchell (1982). Sperm were disabled by radiation and then used to inseminate hens of a different strain. After 24 hr the hens were inseminated with fresh sperm from cocks of their own strain. The theory was that the irradiated sperm would be carried into the ovum by the active sperm, and it was claimed on the basis of feather and egg colour markers that this occurred in between 3 and 5% of cases. These results have been

108

K. Simkiss

confirmed by Tomita et al. (1988) and by Bumstead et al. (1987), who used this method to introduce the major histocompatibility haplotype marker into progeny. Work by Shoffner et al. (1987), however, failed to reveal any evidence for the transfer of marker genes from irradiated sperm, and it is not entirely clear what gene insertion mechanism is actually envisaged as occurring with this approach. Even more controversial results followed a report by Lavitrano et al. (1989) that the sperm of mice could be coated with strands of foreign DNA, which were then carried into the ovum on fertilization. A number of attempts were made to repeat this experiment, but Brinster et al. (1989) were unable to produce any transgenic mice by such sperm-mediated DNA transfer despite analyses of 1,300 mice produced in this way. Similar detailed attempts to transfer linearized plasmid (5.5-kb pCK17 or 2.7kb pUC19) using fowl sperm showed that from 350 to 4,000 plasmid modules attached to each sperm but that none of this was incorporated into the progeny (Gavora et al. 1991). This result confirmed the observation that foreign DNA is rapidly attached to the head of sperm but that there is no evidence for the production of transgenic birds from such gametes. Abstracts are still being published claiming that between 30 and 60% of offspring from DNA-coated rooster sperm contain foreign DNA that is transmitted into the Fl generation (Gruenbaum et al. 1991), but doubt must be expressed about these results until the full experimental details are published. The idea of using sperm as a vector to transfer exogenous DNA into the ovum is so attractive that liposomes have also been used to introduce genes into sperm. This technique was used successfully to incorporate foreign DNA into mouse sperm, but no transgenic offspring were produced (Bachiller et al. 1991). This approach has not been tried in birds. 4.2.1.2 Ova Most birds have only one functional ovary, usually the left one, which lies deep in the body cavity. Relatively few attempts have been made to insert genes directly into the developing follicle, although Shuman (1986) used replication-competent reticuloendotheliosis virus (REV) for that purpose. Fertilized eggs from such birds were incubated and virus was found in 26% of cases. Encouraged by these results, Shuman repeated the experiments using a replication-deficient retrovirus, and 8% of the offspring showed signs of this DNA.

Trans genic birds

109

Similar experiments were undertaken by Kopchick et al. (1991), who used surgery to puncture the air sacs and expose the avian follicles. These were injected with 1-25 /xg of a plasmid (pBGH). Of the 400 offspring that were eventually produced, none was found to possess the foreign DNA, and it was suggested that the inserted DNA was rapidly degraded in the oocyte. Because of the surgical procedures that are necessary and the relatively poor results that were obtained, few attempts have been made to use the ovary as a site for DNA insertion. It is possible to fertilize avian oocytes in vitro, and avian sperm does not require capacitation in the female tract, so that this process is relatively easy to perform (Howarth 1971). Unfortunately, however, the avian oocyte rapidly degenerates during the 24 hr it takes from the time it is ovulated to the time it is laid. It is therefore not possible to attempt either fertilization or genetic engineering on an infertile newly laid egg, and consequently the prospects for using the female gamete in such studies are virtually nonexistent. 4.2.2 The embryo The embryology of the domestic fowl is very well described for both normal and pathological states (Lillie 1919; Romanoff 1972). The requirements of genetic manipulation, however, have directed attention to two relatively neglected aspects. These are, first, the earliest stages of cell proliferation and differentiation that occur before egg laying and, second, the development of the germ-cell lineage. The interest in these stages relates directly to the possibility of using either embryonic stem (ES) cells or primordial germ-cells (PGCs) as vehicles for producing transgenic birds by producing germ-line chimeras. 4.2.2.1 Development before oviposition About 2 hr before the oocyte is released from the ovary of the fowl, it undergoes its first reduction division to form a secondary oocyte and the first polar body. Sperm, introduced into the oviduct at copulation, are partly stored in special glands at the uterovaginal junction. Some, however, rapidly pass up the reproductive tract so that fertilization occurs in the infundibulum region almost immediately after ovulation. Polyspermy is common, with an average of about 10 sperm entering the small region of yolk-free cytoplasm that forms the germinal disc.

110

K.Simkiss

Here they lose their cytoplasm and swell to about 50 times their volume to form the male pronuclei (Perry 1987). Shortly after this, the second maturation division of the oocyte occurs to form another polar body and the female pronucleus. Fertilization, in the sense of the fusion of the male and female pronuclei, occurs shortly afterwards, and the supernumerary male pronuclei slowly degenerate. The fertilized egg passes down the oviduct and is covered in a number of tertiary 'membranes' secreted by the surrounding epithelium and its glands. The magnum region secretes the albumen, and the egg passes through this region in about 2.5 hr to reach the isthmus, or shellmembrane-forming region. At this stage it has probably completed its first cell division, and it then enters the pouchlike shell gland at the distal end of the oviduct, where it is 'plumped' by watery secretions until it is roughly twice its original size. During this process, which takes about 18 hr, the egg is rotated and the eggshell is secreted. Cell division continues throughout this period, so that when the egg is laid it consists of about 60,000 cells, of which about half are destined to be extraembryonic (Spratt & Haas 1960). Four aspects of this phenomenon are important. The first is that the albumen is converted from a high Na/K secretion of about 10:1 to a product with an Na/K ratio of about 2:1 (Mongin & Sauveur 1970). Second, the rotation of the egg at a rate of about 1 revolution per 4 min twists the albumen to form the chalazae. Third, the effect of gravity during this process determines the anterioposterior axis of the future embryo. Finally, the subembryonic fluid develops, separating an overlying cellular layer, or area pellucida, from the underlying yolk (Eyal-Giladi 1984). Understanding these events and their relationship to both embryo culture and the establishment of cell commitments is crucial to many of the approaches involved in producing transgenic birds (Simkiss 1993). This has been greatly facilitated by the description of 14 stages in the initial cleavage and development of the blastoderm. These are identified and described as I to XIV by Eyal-Giladi and Kochav (1976). The main components of this analysis are the recognition of three developmental periods starting with cleavage (stages I-VI), which takes about 10 hr and results in the entire germinal disc being converted to a uniform epithelium, the blastoderm. The second period involves the formation of the area pellucida (stages VII-X), in which an upper layer of smaller cells separates from a lower layer of larger cells. This area pellucida appears as a transparent zone that becomes

Transgenic birds

111 area pellucida-

Fig. 4.1. Diagrammatic vertical sections of the blastoderm at stage X (top), stage XII (middle) and stage 6 (bottom). Anterior to left. The initial uniform blastoderm consists of a transparent area pellucida overlying a subembryonic fluid and an opaque area opaca (top). By stage XII this layer has separated into an upper epiblast (ep) and a forming hypoblast (hp, middle). Primordial germ-cells (PGCs), originating from the epiblast, are carried anteriorly on the hypoblast and accumulate at the germinal crescent (gc, bottom).

progressively more clearly demarcated from the surrounding area opaca due to the formation of the subembryonic fluid between the blastoderm and the underlying yolk. This stage of development has normally been reached by the time the egg is laid, which in descriptive embryology is considered day 0 (i.e., no incubation). It is important to realize, however, that cell division has been occurring and the blastoderm has been developing for between 20 and 24 hr by this stage. The period of hypoblast formation (stages XI-XIV) involves the separation of the blastoderm into two layers, an upper epiblast and a lower hypoblast. The process of gastrulation involves the subsequent inward migration of cells from the epiblast to form the notochord and mesoderm of the embryo. A simplified composite drawing of this system and the way the primordial germ-cells are involved over the period of the first 2 days of incubation is shown in Fig. 4.1. It should be reallized, however, that there is still some controversy over the details of the involvement of the posterior section of the area pellucida in hypoblast formation (Stern 1990; Eyal-Giladi et al. 1992).

112

K.Simkiss

4.2.2.2 Development during incubation The initial events in embryo development have already been outlined, since they extend over stages XI-XIV of the Eyal-Giladi and Kochav scheme. This period, however, is also covered by the classification of Hamburger and Hamilton (1951), which starts with the initiation of incubation and progresses through primitive streak formation (stage 2, 6-7 hr) and the formation of the first somite (stage 7, 23-6 hr incubation) to the formation of the lung primordia (stage 15, 50-5 hr), appearance of the allantois (stage 18, 65-9 hr) and eventual hatching (stage 46, 480-504 hr). Stages XI-XIV of Eyal-Giladi and Kochav overlap stages 1-2 of the Hamburger and Hamilton classification. Much of the interest in the Hamburger and Hamilton scheme is associated with organogenesis. For the purpose of this review, this involves the formation of the gonads and, by implication, the separation of the germ-line. In all vertebrates, the cells that will form the gametes arise at sites some distance away from the developing gonads. The explanation for this is not clear, but it may be a strategy to protect the germ-line from the consequences of differentiation in the somatic cells. One of the results of this extragonadal origin of these PGCs, however, is that they must migrate, during embryogenesis, to find and populate the gonadal stroma (Nieuwkoop & Satasurya 1979). This implies some complex cellular signalling between the migrating cell line (PGCs) and the target organ (gonadal stroma) with all its implications for gene activation and expression. In the birds the PGCs have been identified as arising from the blastoderm of freshly laid eggs at stage X (Muniesa & Dominguez 1990) and leaving the epiblast at stages XII-XIII (Ginsburg & EyalGiladi, 1987) to be carried anteriorly on the hypoblast and mesodermal edges to a position to the anterior of the embryo, the germinal crescent (see Fig. 4.1). Thus, by the first day of incubation (stages 4-8) they occupy a position that is not only extragonadal but actually extraembryonic. In the duck, Fargeix (1969) counted between 50 and 100 PGCs in the germinal crescent of stage 2-5 embryos, and by stages 10-15 this had increased to between 200 and 400. In the quail, it is possible to use monoclonal antibodies to detect a few isolated cells in the unincubated blastoderm (stage X), with about 90 PGCs at stages 2-3 and several hundred by stages 8-9 (Pardanaud et al. 1987a). Similar numbers were found in the embryo of the fowl by Al-Thani and Simkiss (1991) until around stage 12 of incubation, when, as in other birds, they begin to migrate from this site into the vascular

Transgenic birds

113

j

Fig. 4.2. Primordial germ-cells (tPGC) from the germinal crescent region of a stage 12 embryo (left), from the vascular system (bPGC) of a stage 16 embryo (centre) and from the germinal ridge (gPGC) of a stage 18 embryo (right). L, lipid droplet; Y, yolk granule; G, glycogen deposit; N, nucleus; gc, germinal crescent; gr, germinal ridge, nu, nuage. (After Simkiss, 1991.)

system. The PGCs appear to be capable of some active movement at this time, but their incorporation into the circulatory system is largely dependent on the formation of a vascular network and the onset of cardiac activity. A population of migrating PGCs in the blood was first identified by Meyer (1964) at about stage 16 of development; shortly after this stage, the germ cells can be found in increasing numbers in the gonadal anlagen (Fig. 4.2). The PGCs are attracted to the gonad by chemical signals (Kuwana et al. 1986), but their exit from the circulation may be facilitated by the constriction of the blood vessels in this region. Once the PGCs have entered the gonadal anlagen, they divide rapidly to produce more than half a million oocytes in the fowl. This number then crashes, so that only a few thousand persist in the 4-dayold hatchling (Hughes 1963). It will be apparent from this brief description that it is possible to isolate the cells that will form the gametes from (a) presumptive blastoderm (pPGC), (b) germinal crescent tissue (tPGC), (c) circulating blood (bPGC) or (d) gonadal sources (gPGC). In the case of (a) it is possible to culture pieces of blastoderm to produce PGCs (Ginsburg & Eyal-Giladi 1989), while in the cases of (b), (c) and (d) it is possible to

114

K. Simkiss

harvest a considerable number of these cells from embryos. Such cells provide an opportunity for the insertion of foreign DNA into the germline if such genetically manipulated cells are subsequently injected into the recipient embryo to form a germ-line chimera. This possibility was explored by Simkiss et al. (1989), who isolated PGCs from the blood of a stage 15 embryo with a DNA marker and injected them into a recipient embryo without this DNA. The resulting chimera had a gonad containing germ-cells from both embryo sources. In pursuing this approach it should be recognized that the different developmental stages of PGCs (pPGC, tPGC, bPGC and gPGC) not only are morphologically distinct (see Fig. 4.2) but also have a variety of membrane receptors and cellular properties that may be very important for producing transgenics.

4.3 Approaches to avian transgenesis 4.3.1 Microinjection and embryo culture

As already mentioned, the injection of foreign DNA into the pronucleus of the avian egg suffers from two difficulties. The first is the problem of identifying this structure in the cytoplasm of an extremely large and opaque cell. The second is determining which of the supernumerary pronuclei in the germinal disc will be involved in the syngamic fusion. Neither of these problems has been resolved, but a number of interesting experiments have been performed. Cloned DNA was injected into the germinal disc, about 25 /mm below the plasma membrane in the region of the female pronucleus of a freshly fertilized egg. In Xenopus such experiments have resulted in a low frequency of incorporation of the foreign DNA into the host genome (Etkin & Pearman 1987). In similar experiments with fowl eggs using a bacterial gene encoding chloramphenicol acetyltransferase under the control of the Rous sarcoma virus promoter (pRSV cat), the DNA appeared to remain episomal. The linearized plasmid replicated up to 20-fold in the first 24 hr, but the rate then declined and by day 7 was restricted to the extraembryonic membranes (Sang & Perry 1989). In similar experiments using lac Z as a reporter gene, histochemical methods could be used to demonstrate /3-galactosidase expression (Naito et al. 1991a). In this case a chicken fi-actin-lac Z hybrid gene (MiwZ) was used and tissue expression remained in 40% of the embryo at day 4 of

Transgenic birds

115

incubation. It was suggested that some exogenous DNA may have integrated into the genome of these cells in a mosaic fashion. In order to pursue this possibility, Perry et al. (1991) examined the fate of injected DNA in greater detail by studying the temporal and spatial expression of the lac Z gene under the control of the cytomegalovirus promoter (pHFB GCM). Analysis of the DNA in the 26 hr blastoderm showed that in 80% of the embryos only 10% or less of the injected DNA remained, while in the other 20% the whole injected dose was recovered. Maximum expression of the reporter gene occurred around stage X, and normal cell death and dilution by cell division were probably the major influences in its subsequent decline. By day 7 only 8% of the surviving embryos contained any cells expressing the foreign gene. Experiments which involve the microinjection of the freshly fertilized ovum require eggs that have not passed down the oviduct. This usually means killing the bird to obtain this material and culturing the ovum in a surrogate egg (Rowlett & Simkiss 1985; Rowlett 1991). Devising such a culture system has been difficult, involving several changes of fluid (Perry 1988), the removal of the albumen capsule (Naito et al. 1990), the provision of an eggshell as a source of embryonic calcium (Ono & Wakasugi 1984; Rowlett & Simkiss 1987) and rocking the cultures (Deeming et al. 1987). One reason for these difficulties is the fact that under normal circumstances the albumen that is secreted by the bird has an Na/K ratio that is extensively modified during its passage down the oviduct (Section 4.2.2.1) and it is difficult to simulate this in a static system (Simkiss et al. 1993; Simkiss 1993). These culture problems have been largely overcome by removing much of the original albumen, and it is now possible to obtain hatchability rates of almost 50% (Naito et al. 1990). The procedures, however, are difficult and time-consuming and add to the problems of using microinjection of the pronucleus to produce transgenic birds. For these reasons most experimenters have subsequently attempted to introduce foreign DNA into partially incubated eggs.

4.3.2 The blastoderm By far the most successful way of producing transgenic birds to date has been to infect the blastoderm with either competent or replicationdefective retro viruses.

116

K. Simkiss

Fig. 4.3. A virion (v) attaches by its envelope proteins (ep) to receptor (r) molecules on the cell surface and inserts its double-stranded RNA into the cytoplasm, where it is transcribed into a double-stranded DNA. This enters the nucleus (n) as a circular structure that inserts as a provirus into the host genome. The retrovirus terminal sequences (RV5, U3R) are replaced by the long terminal repeats that are essential for the transcription of the gag (antigen), pol (polymerase) and env (envelope) genes.

4.3.2.1 Retroviruses Retroviruses are RNA viruses that replicate through a region of chromosomally integrated proviral DNA (Fig. 4.3). The viral sequences include the trans-acting regions that encode viral proteins and the cisacting components involved in replication. The trans-acting region of the provirus contains three protein coding regions, gag, pol and env. The gag gene is involved with viral encapsidation, pol encodes for reverse transcriptase, protease and integrase enzymes, while env produces the viral envelope. The cis-acting regions include the promoter sequences, the termination signal, the primer binding site and the packaging signal (Shuman 1991). The retrovirus enters specific cells by binding to a membrane receptor protein that is recognized by the viral envelope. Upon entering the cell, the RNA genome is released and transcribed into a linear DNA form. This is transported to the nucleus, where it is ligated into a circular form that integrates into the host chromosome as a provirus. During subsequent transcription and translational activity, new viral particles are produced through the

Trans genic birds

117

host cell pathways. These viral components form new virions that bud from the cell membrane. After entering the cell the retrovirus usually blocks the surface receptor so that only one copy of the viral genes is present in any one cell (Varmus 1988). Pro virus integration can occur at multiple sites, but these are probably the more transcriptionally active regions and highly preferred sites have been discovered in the genome (Shih et al. 1988). 4.3.2.2 Retroviral transgenics The first transgenic birds were produced by injecting avian leukosis virus (ALV) into the blastoderm region of newly laid eggs from a retrovirus-free strain of White Leghorns (so-called line zero birds). The eggs were sealed after the injection and incubated normally. Twenty-one males were raised to sexual maturity and shown to contain the retrovirally derived DNA in their sperm. There were 23 different proviral inserts in these birds, which were stably inherited for at least two generations (Salter et al. 1986, 1987). Subsequent studies revealed some interesting aspects of this approach. Of the 23 proviral inserts produced in such experiments, 21 coded for the complete ALV but 2 were incomplete. One coded for only the envelope protein while the other also included a group-specific antigen that induced resistance to these subgroup A viruses (Crittenden & Salter 1990). Since these 2 retrovirus inserts (a/v6 and alv\3) were genetically incomplete, they were not infective and only parts of the viral genome were transmitted with that of the bird. They were, however, able to express the genes for the viral coat protein and thus were resistant to subsequent infection of the bird by the same virus type. Processes similar to the type detected in these experiments are presumably responsible for the occurrence of endogenous viral (ev) regions in the normal avian genome. By these processes replicationcompetent retro viruses generate, by chance deletions, transgenic birds that are capable of expressing some remnants of the original viral genome but without the ability to infect other birds. Male birds do not transmit ALV congenitally to their progeny (Spencer et al. 1980), so that any evidence for the transfer of such genetic material is interpreted as germ-line transmission. A replication-competent strain of RSV was constructed by Chen et al. (1990) that contained the bovine growth hormone gene (bGH).

118

K.Simkiss 5

Approximately 10 virus particles were injected into the blastoderm to produce viraemic chicks that, on maturation, produced hens that laid infected eggs in roughly 50% of cases (Kopchick et al. 1991). Male offspring from this generation were tested for their ability to transmit pro viral sequences. Among those cocks that were able to do this, the transmission rate was between 0 and 4%, indicating that not all of the germ-cells had been infected and that integration had occurred at different sites. Subsequent breeding experiments showed that stable germ-line integration of the exogenous retro viral DNA had occurred, but Southern blots for bGH demonstrated that this sequence had been lost. With a second construct the reverse transcriptase gene was replaced and bGH was linked to murine metallothionein promoter. Infected blastoderms produced hatchlings containing the construct but with poor correlations in serum bGH levels. It appeared that many of the cells secreting the growth hormone were in a variety of somatic tissues. 4.3.2.3 Replication-defective vectors Replication-defective retroviruses are so called because they are missing either trans-acting or cis- and trans-acting regions (Shuman 1991). As a result they are unable to complete more than one round of the life cycle and cannot produce infectious particles. They are therefore produced in specially constructed helper cells that provide the deleted genes necessary for virus production (Watanabe & Temin 1983). By inserting foreign genes that are driven from either long terminal repeats (LTRs) or internal promoters, it is possible to use replicationdefective viruses to introduce and express a variety of foreign genes in the host cell. A replication-defective spleen necrosis virus (SNV) was used by Bosselman et al. (1989a,b) to introduce two marker genes into the blastoderm of fowl eggs. The resulting 760 offspring showed evidence of vector-derived DNA in 23% of cases, and four cocks were obtained that contained these sequences in their sperm. All four birds produced transgenic offspring representing 34 different sites of integration. In a subsequent experiment the chicken growth hormone gene (cGH) was introduced into the vector. Embryos injected with this vector did not hatch well, but thirty 15-day-old embryos were analysed for serum growth hormone levels, and in 16 cases these were at least 10 times the level of normal controls (Bosselman et al. 1990). Germ-line transmission of a replication-defective SNV vector has

Transgenie birds

119

also been obtained in quail. Of 16 male birds produced by blastoderm injection, only one showed germ-line transmission, although this was stable over five generations and expressed the genes in a wide variety of tissues (Lee & Shuman 1990). 4.3.3 Embryonic stem cells The inner cell mass of the preimplantation mouse embryo contains pluripotent cells which, if injected into other embryos, produce viable chimeras (Gardner & Papaioannou 1975). Culturing such cells was extremely difficult, however, until Evans and Kaufman (1981) induced delayed implantation of these embryos, which then showed a retarded development. Cells obtained from the inner cell mass of these isolated blastocysts could be maintained in culture by using feeder cells. Permanent cell lines were subsequently established from these ES cells, and have been used as cellular vectors for experimental manipulation of the mouse germ-line (Thomas & Capecchi 1987). A number of techniques have been developed using differentiation inhabitant factor to retain these ES cells in a pluripotent state suitable for use in germline chimeras (Nichols et al. 1990). The normal preimplantation inner cell mass of the mouse blastocyst contains relatively few cells, and it is not clear what the equivalent stage is in an avian embryo. In reviewing the literature on this subject, Eyal-Giladi (1984) concluded that the stage X blastoderm of the newly laid egg had a concealed bilateral symmetry but that individual cells were pluripotent and epiblast cells may even be totipotent. Evidence to support this derives from the attempts of Marzullo (1970) to produce chick chimeras by mixing blastoderm cells from White Leghorn and Barred Plymouth Rock to Rhode Island Red embryos. It was demonstrated that the transferred cells contributed to feather pigmentation, although no viable chicks were produced. Petitte et al. (1990) repeated and extended these observations. Of 53 dwarf White Leghorns injected with Barred Plymouth Rock blastoderm cells, 6 were chimeric for feather colour and 1 male survived to hatching. Mating this bird produced 719 chicks, of which 2 indicated that some germ-line incorporation of cells had occurred; this was confirmed by DNA fingerprinting. This degree of chimerism is low compared with that obtained in mice (Robertson 1986), which raises concern about how pluripotent such stage X blastoderm cells actually are. This question was pursued by Naito et al. (1991b), who transferred quail blastoderm cells from a newly laid egg into the stage X blastoderm of the fowl. With an

120

K. Simkiss

injection volume of 3-5 /xl only 38 out of 441 embryos survived, but when the volume was reduced to 1 /xl, 48 out of 199 chicks hatched. A total of 7 chimeric chicks were produced, warranting further analysis of such birds. The experiments were repeated by injecting between 700 and 2,000 quail cells into the centre of state XI-XIII blastoderms or into the centre or posterior of stage XIV blastoderms (primitive streak stage) of the fowl. The embryos were killed at 6 days incubation. Histological analysis showed that ectodermal chimerism was limited to the head region when injections were made into the centre of stage XI-XIII blastoderms. Ectodermal and mesodermal chimerism occurred in both types of stage XIV injection. In all three experimental conditions, quail PGCs were found in the embryos and 16 out of 26 individuals (i.e., 62%) were germ-line chimeras. It is not clear from this experiment whether this was due to the random incorporation and differentiation of pluripotent cells or to the selective incorporation of cells that were already developmentally committed. The use of ES cells for producing transgenic birds has two attractions. The first is that by isolating such cells it will be possible to use nonretroviral vectors to introduce foreign DNA into them. The second is that the ES cells will be capable of being cultured and selected in vitro before being used to produce chimeras. Chicken blastoderm cells have been transfected with a variety of plasmids, including heat shock or metallothionein promoters and lac Z genes (Brazolot et al. 1991). Transfection rates of up to 1 in 25 cells were obtained depending on the quantity of Lipofectin used and the subsequent identification by X gal staining of the enzyme /3-galactosidase. Injecting such blastoderm cells into 58 stage X embryos produced 36 specimens at 65 hr in which gene activity was detected, although it was mainly extraembryonic. In one case, however, /3-galactosidase-expressing cells were found in the brain, head and ventricle. There is good reason to believe that blastoderm cells can be transfected by a variety of standard techniques to produce chimeras, but the second requirement, of producing a stem cell that is pluripotent for the germ-line and capable of being maintained in cell culture, remains the main obstacle to this approach.

4.3.4 Primordial germ-cells

There is an extensive literature on the transfer of a variety of cells between embryos to form chimeras (Le Douarin & McLaren 1984). In

Transgenie birds

121

Fig. 4.4. The chimeric Rhode Island Red cockerel (centre) contains genetically manipulated White Leghorn PGCs. Most of its offspring are brown (left), but some have the phenotype of the White Leghorn and are transgenic (right).

some elegant experiments using embryos in culture, Simon (1960) devised a variety of parabiosis experiments to introduce PGCs into individual chick embryos and also to exchange them between the circulations of duck and fowl embryos. In an extension of this work, Reynaud (1969) used intravascular injections of PBCs to create both fowl and turkey-fowl chimeras (Reynaud 1976). The difficulty in all these experiments, however, was to establish the unequivocal survival of the transferred PGCs, and for this reason Simkiss et al. (1989) used specific DNA markers to demonstrate the formation of germ-line chimeras in the fowl. Subsequently, a replication-defective retrovirus, based on SNV, was prepared (Meyers et al. 1991) and shown to infect fowl PGCs. These manipulated bPGCs were injected into the vasculature of 3-day-old embryonic fowl and specific SNV DNA sequences were recovered from the gonads. Using a defective retrovirus based on a defective avian leukosis virus, Vick et al. (1993) transfected PGCs of the White Leghorn and transferred them to a Rhode Island Red embryo. Cocks produced in this way produced some normal brown offspring and some white transgenics, clearly derived from the genetically manipulated PGCs (Fig. 4.4). It is clear, therefore, that PGCs can be used as vehicles for introducing foreign DNA into the

122

K. Simkiss

genome of the bird, and subsequent experiments have been directed at characterizing this phenomenon and increasing its efficiency. The protocol for producing transgenics by such a procedure would, in outline, consist of (a) isolating PGCs, (b) integrating foreign DNA into their genome, (c) screening the cells for suitable gene expression, (d) introducing these genetically manipulated cells into a recipient embryo to form a chimera and (e) breeding from this chimera and selecting those offspring that contained the introduced genes. 4.3.4.1 Isolating primordial germ-cells There are three well-characterized sources of PGCs in the embryo. These are the germinal crescent tissue (tPGC), blood-borne PGCs (bPGCs) and gonad-derived cells (gPGCs). These correspond to the three progressive stages in the origin, migration and settlement of these cells, and as such they might be expected to show different physiological properties. Morphologically, the germinal crescent PGCs are roundish in shape, with large numbers of yolk granules, ribosomes and mitochondria (see Fig. 4.2). Once they enter the vascular system they develop a few microvilli, and bPGCs typically have a large eccentric nucleus with increasing numbers of glycogen granules in the cytoplasm. On entering the germinal ridge and becoming gPGCs, they often become elongated (up to 20 jum) with numerous cytoplasmic projections and microvilli (Ukeshima & Fujimoto 1984). These features can be associated with increased motility, a conversion of cytoplasmic reserves to glycogen and an increase in protein metabolism. Corresponding changes occur in the membrane proteins. The lectin concanavalin A binds to glucose- and mannose-containing receptors, and it shows a strong affinity for the PGC membrane (Lee et al. 1978). In its presence the normal migration of PGCs becomes inhibited (Al-Thani & Simkiss 1991). A number of other lectins (Yoshinaga et al. 1992), monoclonal antibodies, such as EMA-1 (Urven et al. 1988), and carbohydrate antigens (Loveless et al. 1990) are associated with PGC membranes at various stages of this maturation process. This suggests that the membrane receptors on PGCs undergo considerable changes during their migration, and confirmation of this has come from studies on mammals. In the mouse, the dominant-white spotting (W) and steel (SI) loci have been known for many years to be involved in the migration

Trans genie birds

123

of PGCs. It is now apparent that these loci are responsible for a transmembrane tyrosine kinase receptor (stem cell receptor) on the surface of the PGC and a secretion (stem cell factor) from cells associated with the migration path (Witte 1990; Motro et al. 1991). The stimulation of such membrane receptors is probably also responsible for inducing the proliferation and differentiation of PGCs and accounts for the fact that migratory and postmigratory PGCs behave quite differently (Donovan et al. 1986). It appears that a wide range of these growth factors, including basic fibroblast growth factor (bFGF) and soluble leukaemia inhibitory factor (sLIF), influence the PGC during this period (Resnick et al. 1992). It might be expected, therefore, that attempts to produce chimeras by transferring tPGCs, bPGCs and gPGCs into recipient embryos would have different effects. There is evidence that both tPGCs and bPGCs can be transferred from donor to recipient embryos and settle in the germinal ridge. These results come from histological examination of both intra- (Simon 1960) and interspecific (Reynaud 1969) transfers and from DNA analyses of recipient embryos (Simkiss et al. 1989; Savva et al. 1991). Gonocytes released from avian gonads after the PGCs have settled are of three types, which can be grown in large numbers in culture (Wentworth & Wentworth 1991). These cells appear, therefore, to be released from their mitotic block, but the question as to whether they can be reintroduced into embryos to form chimeric gonads is, as yet, unresolved. Long-term culture of PGCs, to form what Resnick et al. (1992) call embryonic germ-cells, produced the speculation that these would rarely, if ever, colonize the germ-line (McLaren 1992), but this does in fact appear to be possible in mice (Stewart et al. 1994). 4.3.4.2 Introducing foreign DNA To date the main method of introducing foreign DNA directly into PGCs has involved the use of normal or replication-defective retro viruses (Simkiss et al. 1990). There is no inherent reason why electroporation, calcium phosphate precipitation, DEAE-dextran or Lipofectin liposomes should not be used (Wentworth et al. 1992), except that during the development of procedures using PGCs it is most convenient to use vectors that are likely to give the highest rates of integration and at the present time these are retro viruses.

124

K. Simkiss

43.43 Screening PGCs One of the central attractions of using ES cells in producing transgenics is that they can be genetically manipulated, maintained in long-term culture and then screened for gene expression before being used to produce germ-line chimeras. In this they are superior to PGCs, which at present can be maintained in culture for only short periods and proliferate very poorly (Donovan et al. 1986). A number of extracellular matrix components have been tested to determine their influence on avian PGC growth, but they appear to have little effect until 7 days after application (Bellin et al. 1985; Wentworth 1989). Attention has therefore shifted to the effects of feeder cells and in particular a number of growth factors (SCF, FGF, LIF; see Section 4.3.4.1). These may represent the normal factors that stimulate the rapid proliferation that occurs with PGCs once they have settled in the germinal ridge. The results of such experiments are very encouraging (Resnick et al. 1992) and indicate that it may soon be possible to maintain and screen PGCs in culture.

43.4.4 PGC transfer and proliferation The expression and stimulation of a variety of membrane receptors on PGCs appear to be an essential part of their development, migration to the germinal ridge and subsequent proliferation in the gonad. It is not clear what the temporal and spatial relationships are in these interactions. The system appears, however, to be fairly robust in that a large number of workers have shown, histologically, that both tPGCs and circulating bPGCs can be transferred to recipient embryos and shown to develop in the host germinal ridge (Section 4.3.4). Indeed, several groups have shown that PGCs from mice will populate the germinal ridge of chick embryos (Rogulska et al. 1971; FerhaneTachinante 1975). It is therefore rather surprising that Petitte et al. (1991) were unable to detect any offspring among 3,177 chicks derived from dwarf White Leghorn PGCs that had been injected into 59 Barred Plymouth Rock embryos. These results suggest either that the PGC transfers were not effective for technical reasons (age of embryos, too few PGCs or poor injections) or that there is considerable sorting of PGC cells within the stroma of the developing gonad. In a study of the ovary of the fowl, Hughes (1963, 1964) found that the number of germcells rose to 27,000 on day 9 of incubation, reached 680,000 on day 17

Trans genie birds

125

and then crashed to only a few thousand in the 4-day-old chick. The fate of individual cells and their relationship to the original 400 gPGC clones that form in the fowl gonad is not known. Clearly, any discrimination in cell sorting in the gonad would have major implications for the production of transgenics by any procedure that involves these chimeras. 4.4 Perspectives

The location of the gonads deep in the body cavity and the size and constitution of the oocytes, with a long delay in oviposition, mean that both the anatomy and physiology of the bird present serious obstacles to the microinjection of DNA into the egg. Thus, unless there is some major technological advance involving, for example, sperm coating with DNA or a pronucleus-seeking vector, it is unlikely that avian transgenics will be produced from the zygote. There are therefore only two viable targets for avian genetic engineering: the blastoderm in situ and the embryo chimera. In both cases this means that transgenics will not be produced until the Fl generation, and then they will occur in only a proportion of the offspring. Clearly, if this approach is to be used in the future, technological efficiency will be a major concern. 4.4.1 Manipulating the blastoderm in situ

Manipulating the blastoderm in situ has been both the most popular and most successful way of introducing foreign DNA into the avian genome. Most of the transgenics that have been produced have been obtained by direct infection of blastoderm cells in ovo using either competent (Section 4.3.2.2) or replication-defective retroviruses (Section 4.3.2.3). The advantages of this approach are (a) technical simplicity of the injection and (b) good success rate. The disadvantages are (a) the use of retroviruses that are difficult to construct and that have a number of inherent problems (size of DNA insert, stability, resistance to use in food products, etc.) (Temin 1989) and (b) somatic as well as germ-line insertion. Of the pros and cons of this approach the use of retroviruses dominates the discussion. An attempt has been made to use cationic liposomes to introduce the RSV LTR and the firefly luciferase gene (pRSVL) into the blastoderm (Rosenblum & Chen 1991). High luciferase activity was found in 3-day-old embryos with detectable activity in 8-day-old embryos, but the results again

126

K. Simkiss

suggested an episomal rather than an integrated fate for the foreign DNA. 4.4.2 Embryo chimeras There are two approaches to producing chimeras by isolating, manipulating and then reintroducing cells into embryos in the hope that they might be involved in forming the germ-line. In their monograph on germ-cells in chordates, Nieuwkoop and Satasurya (1979) wrote, "Although we have a reasonable understanding of the early development of the avian embryo we know nothing about the actual mode of origin of the PGCs in this group." Two types of experiment have shed a light on this problem in the past 24 years. The first have indicated that a specific gene product (oskar) may determine the number of PGCs in Drosophila. The suggestion is that this gene product may be identified as granules in the cytoplasm of germ-cells (i.e., nuage) and that this material may protect the germ-line from somatic differentiation signals (Ephrussi & Lehman 1992). If the same system operates in chordates, we clearly need to know which cells contain this cytoplasmic factor (or nuage or germ-plasm) and what effect it has on foreign DNA insertion and expression. In the second set of experiments, electron microscopists and immunohistochemists have claimed to have identified PGCs in stage X blastoderms - that is, before incubation (Pardanaud et al. 1987b) and before they descend onto the hypoblast (see Fig. 4.1). To make the matter even more significant, at least two groups of workers have claimed that stage 3-4 avian PGCs contain nuage (Climent et al. 1979; Muniesa & Dominguez 1990). Those workers who are looking for ES cells have taken the stage X blastoderm as their starting material. It appears at least possible that in doing this they have actually selected a group of cells which include PGCs or precursors that are already committed to forming them in the next few hours. By breaking up the blastoderm, they have partially delayed this process. According to Ginsburg and Eyal-Giladi (1989), "In cultures of stage X blastoderms the morphologic expression of PGCs is related to the level of differentiation and organization of the somatic cells in the culture which, in turn, is dependent on the initial concentration of cells in the culture". It is therefore worth considering that experiments described as using 'stage X embryonic stem' and 'stage 10 primordial germ cell' may be dealing with almost the same stages of the same cell line.

Transgenie birds

Ml

4.4.2.1 Plasmid insertion and cell screening The attraction of using isolated cell lines is that it is possible to use reporter genes to identify and select those cells that have successfully integrated and expressed the foreign DNA. Thus, the relative inefficiency of inserting plasmid DNA into embryonic cells can be overcome by forming chimeras only from those cells that survive the screening process. Blastoderm cells have been transfected with a number of plasmids using the liposome Lipofectin (Brazolot et al. 1990). Transfection rates of up to 5% were obtained, and these cells were then injected into recipient embryos. The reporter gene was, however, lac Z so that, although it was possible to trace the transfected cells in the embryo, no attempt was made to select these cells before forming a chimera. By contrast, Page et al. (1991) used a plasmid vector containing the neo gene to transfect embryo fibroblasts again using Lipofectin. The neo gene confers resistance to the amino glycoside antibiotic G418, which normally kills these fibroblasts within 2 weeks. Transformed cells were maintained in culture for 2 months in the presence of this antibiotic but not reintroduced into the embryo. Thus, the complete experiment, on which the stem cell concept is based, has not yet been undertaken in birds. Embryonic cells have not, to date, been transfected, screened and then used to form even somatic chimeras, let alone germ-line chimeras. An additional reason for introducing plasmid vectors into cell lines is that it opens up the possibility of targeted integration of DNA into the avian genome. Thus, by including a 1.8-kb fragment of the ovalbumin gene in their plasmid, Page et al. (1991) raised the possibility of using homologous recombination to get site-specific insertion. This is normally considered to be a relatively rare event in mammalian cells (1 in 104), but Buerstedde and Takeda (1991) obtained targeted integration in about 80% of transformed chick B cell lines. In this work constructs containing either chicken /3-actin or ovalbumin genes were used. These results are extremely surprising, especially as they were not repeated in other cell lines such as T-cells, myoblasts or erythroblasts.

4.4.2.2 Cell insertion and chimera screening The crucial factor determining the efficiency by which genetically manipulated cells are incorporated into a chimera is clearly the ratio of donor cells (Cd) to endogenous cells (Ce) for any specific tissue. If

128

K. Simkiss

totipotent blastoderm cells are used with no spatial preference for particular sites of embryo incorporation, the ratio obviously becomes simply one of relative cell numbers. In this situation Carscience et al. (1992) used a clever procedure of exposing recipient blastoderms to 540-660 rads of 60Co radiation before adding donor cells. The effect of the radiation was to disrupt normal cell biology, so that the donor cells were at an advantage in populating the embryo. As a result the degree of chimerism increased by three- to four-fold. A general cellular disruption of this type is clearly the only approach to increasing efficiency when tissues have not differentiated, but once differentiation has progressed far enough to establish specific properties other possibilities exist. A number of attempts have been made to remove PGCs in 2-day-old fowl embryos by surgical methods (McCarrey & Abbott 1982) or by destroying them with ultraviolet (Reynaud 1976), X-ray (Fargeix 1976) or laser (Mimms & McKinnell 1971) radiation. In all cases, however, it is difficult to avoid damage to surrounding tissues. An alternative approach is to use a chemosterilant such as the drug Busulphan (1,4butanediol dimethane sulphonate), which destroys migrating PGCs in both mammals (Hemsworth & Jackson 1962) and birds (Reynaud 1981; Hallett & Wentworth 1991). This drug is a potential teratogen (Bishop & Wassom 1986), but by manipulating the dose and the route of uptake Aige-Gil and Simkiss (1991) were able to destroy 95% of the endogenous PGCs with minimal side effects. These approaches make it possible, therefore, to manipulate specifically the CJCe ratio for germ-line cells. In the formation of most chimeras, relatively little attention is paid to the sex of the donor and recipient cell lines. When chimeras are being used to form germ-line transgenics, however, there is clearly a potential problem. Thus, male PGCs (ZZ) inserted into a female embryo (WZ) or female PGCs (WZ) introduced into a male embryo (ZZ) may undergo abnormal processing in the developing gonad. All the evidence indicates that the type of gonad (i.e., ovary or testis) is determined by the genotype of the germinal ridge, but in quail-chick chimeras the stroma of the gonad was also able to harbour germ-cells of the opposite sex, at least for up to 12 days of incubation (Hajji et al. 1988). A somewhat different result was obtained by Haffen (1975), who made chimeras between anterior and posterior halves of chick embryos in culture before grafting them into coelomic sites. Under these conditions, male PGCs degenerated in ovaries, apparently be-

Trans genie birds

129

cause they could not cross the premeiotic prophase, while PGCs of either sex survived in the testis. There is, of course, a considerable difference in the onset of meiosis during the development of male and female gametes, and this is initiated during embryonic development (Dieterlen-Lievre et al. 1985). In the quail, it was possible to feminize male embryos so that the PGCs gave rise to oocytes. In the chick embryo, injections of an aromatase inhibitor (which blocks the synthesis of oestrogen from testosterone) caused females to develop testes and undergo full spermatogenesis (Elbrecht & Smith 1992). At present, no successful breeding experiments have been undertaken on such birds, so that it is not known if these gametes are functional. It is therefore not at all clear what will happen if mixed-sex chimeras are produced in attempts to produce germ-line chimeras. 4.4.3 Conclusions The current indications are that the size of the avian oocyte, the deep abdominal location of the gonads and the extended time between fertilization and oviposition all militate against the efficient microinjection of the pronucleus as a way of producing transgenics. The alternative route towards this end is to produce some kind of chimera by either an in ovo or an in vitro manipulation of embryonic cells. This suggests the following considerations and problems: 1. Manipulation of blastoderm cells: To date this has been the most successful approach but with two constraints. These are (a) a dependence on retroviral vectors and (b) an inability to select for inserted DNA, resulting in great variability in offspring. 2. Manipulation of ES cells: This is by far the most intellectually exciting possibility, raising prospects of continual culture of pluripotent cells containing plasmids with selectable properties. Unfortunately, there are major constraints, starting with the need to identify true ES cells in the bird, extending to the problem of maintaining them in a nondifferentiating culture and culminating in the difficulty of ensuring efficient germ-line insertion. 3. Manipulation of PGCs: In many ways this is a compromise approach in that it initially was thought to depend on the use of committed cells that would enter the germ-line. Because of this commitment there are significant problems in inducing cell division in culture and ensuring that the PGCs are still capable of completing their migration to the gonad. There are major barriers to under-

130

K. Simkiss

standing the membrane receptors that will control migration, proliferation and differentiation in the gonad, and the significance of this has recently been increased by the observation that pluripotent ES cells can be derived from PGCs (Matsui et al. 1992) and reinserted into the germ-line (Stewart et al. 1994). In all three types of transgenesis the chimera approach depends on raising the embryo to a mature adult before it can be screened to see if the foreign DNA has entered the germ-line. Transgenics do not exist until the Fl generation, making the experiments expensive in time and resources. The probability of producing transgenic offspring from chimeric adults depends on the efficiency of introducing foreign DNA into the embryonic gonad, which in turn depends on the following: 4. Manipulation of the cell kinetics of the embryo in favour of the introduced transfected cells: In practice this involves disrupting either (a) the development of the whole embryo so that introduced cells are at an advantage in establishing themselves or (b) specific disruption (sterilization) of the germ cells so that the gonad is populated by transfected cells. This can be done physically (e.g., X rays, ultraviolet light, lasers) or chemically (e.g., chemosterilants). There is, however, a basic lack of information about the production and survival of germ cell clones in the vertebrate gonad. At the present time (1992) the only vectors that have been successfully used to produce avian transgenics are retroviruses. There is a certain historical satisfaction about this since retroviruses were first discovered in chickens, but their instability and public image mean that plasmids may be more useful in the long term. Plasmid vectors have not been introduced into blastoderm cells in ovo. Thus, the most successful route for producing avian transgenics is not easily accessible to plasmid vectors without going down the cell chimera route. Two other points are worth noting. First, relatively few genes have been isolated and cloned in the bird, so that current opportunities are limited (Bulfield 1990). Second, where avian genes have been successfully introduced into birds, as with cGH, the expected results were not obtained. The continued secretion of growth hormone downregulates membrane receptors, and in order to get its normal physiological responses the hormone has to be pulsed (Johnson 1988; Bacon et al. 1989). It is somewhat paradoxical that the first efforts to produce

Transgenie birds

131

avian transgenics turned out to be disappointing because the physiological properties of the protein that was produced were not fully appreciated. Clearly, the regulation of foreign DNA at the organ level will be an important priority in all transgenic work in the immediate future.

References Aige-Gil, V. & Simkiss, K. (1991). Sterilisation of avian embryos with busulphan. Res. Vet. Sci. 50, 139-44. Al-Thani, R. & Simkiss, K. (1991). Effects of an acute in vivo application of concanavalin A on the migration of avian primordial germ cells. Protoplasma 161, 52-7. Bachiller, D., Schellander, K., Peli, J. & Ruther, U. (1991). Liposome-mediated DNA uptake by sperm cells. Mol. Reprod. Dev. 30, 194-200. Bacon, W. L., Vasilatos-Younken, R., Nestor, K. E., Anderson, B. L. & Long, D. W. (1989). Pulsatile patterns of plasma growth hormone in turkeys: Effects of growth rate, age and sex. Gen. Comp. Endocrinol., 75, 417-26. Bellin, M. E., Gale, J. A. & Wentworth, B. C. (1985). Avian PGC growth in response to components of the extracellular matrix in vitro. Biol. Reprod. 32(suppl.), 72 (abstract). Bishop, J. B. & Wassom, J. S. (1986). Toxicological review of busulphan. Mutation Res. 168, 15-45. Bosselman, R. A., Hsu, R.-Y., Boggs, T., Hu, S., Bruszewski, J., Ou, S., Kozar, L., Martin, F., Green, C , Jacobsen, F., Nicholson, M., Schultz, J. A., Semon, K. M., Rishell, W. & Stewart, R. G. (1989a). Germ line transmission of exogenous genes in the chicken. Science 243, 533-5. Bosselman, R. A., Hsu, R.-Y., Boggs, T., Hu, S., Bruszewski, J., Ou, S., Souza, S., Kozar, L., Martin, F., Nicholson, M., Rishell, W., Schultz, J. A., Semon, K. M. & Stewart, R. G. (1989b). Replication-defective vectors of reticuloendotheliosis virus transduce exogenous genes into somatic stem cells of the unincubated chicken embryo. J. Virol. 63, 2680-9. Bosselman, R. A., Hsu, R.-Y., Briskin, M. J., Boggs, T., Hu, S., Nicholson, M., Souza, L. M., Schultz, J. A., Rishell, W. & Stewart, R. G. (1990). Transmission of exogenous genes into the chicken. J. Reprod. Fert. 41,(suppl.), 183-95. Brazolot, C. L., Petitte, J. N., Clark, M. E. & Etches, R. J. (1990). Introduction of lipofected chicken blastodermal cells into the early chicken embryo. J. Cell Biochem. Suppl. 15E, 200 (abstract). Brinster, R. L., Sandgren, E. P., Behringer, R. R. & Palmiter, R. D. (1989). No simple solution for making transgenic mice. Cell 59, 239-41. Buerstedde, J.-M. & Takeda, S. (1991). Increased ratio of targeted to random integration after transfection of chicken j8 cell lines. Cell 67, 179-88. Bulfield, G. (1990). Molecular genetics. In Poultry Breeding and Genetics, ed. R. D. Crawford, pp. 543-84. Amsterdam: Elsevier. Bumstead, N. & Palyga, J. (1992). A preliminary linkage map of the chicken genome. Genomics 13, 690-7. Bumstead, N., Messer, L. L., Freeman, B. M. & Manning, A. C. C. (1987).

132

K. Simkiss

Genetic transformation of chickens using irradiated male gametes. Heredity 58, 25-30. Carscience, R. S., Clark, M. E. & Etches, R. J. (1992). Manipulation of the donor cell contribution to chimaeric chickens. Poultry Sci. 71(suppl. 1), 72 (abstract). Chen, H. Y., Garber, E. A., Mills, E., Smith, J., Kopchick, J. J., Dilella, A. G. & Smith, R. G. (1990). Vectors, promoters & expression of genes in chick embryos. /. Reprod. Fert. 41 (Suppl.), 173-82. Climent, P. S., Gotzens, V. J. & Dominguez, L. (1979). Nuage in the primordial germ cells of the chick embryo. Z. Vet. Med. C. Anat. Histol. Embryol. 8, 200-4. Crittenden, L. B. & Salter, D. W. (1990). Expression of retroviral genes in transgenic chickens. /. Reprod. Fert. 40 (suppl.) 163-71. Deeming, D. C , Rowlett, K. & Simkiss, K. (1987). Physical influences on embryo development. /. Exp. Zool. 1 (suppl.), 341-5. Dieterlen-Lievre, F., Hajji, K. & Martin, C. (1985). Is the time of meiosis initiation programmed in the germ cells or the ovarian stroma? A study in avian chimaeras. Arch. Anat. Microscop. Morphol. Exp. 74, 20-2. Donovan, P. J., Stott, D., Cairns, L. A., Heasman, J. & Wylie, C. C. (1986). Migratory and post migratory mouse primordial germ cells behave differently in culture. Cell 44, 831-8. Elbrecht, A. & Smith, R. G. (1992). Aromatase enzyme activity and sex determination in chickens. Science 255, 467-70. Ephrussi, A. & Lehman, R. (1992). Induction of germ cell formation by oskar. Nature 358, 387-92. Etkins, L. D. & Pearman, B. (1987). Distribution, expression and germ line transmission of exogenous DNA sequences following microinjection into Xenopus laevis eggs. Development 26, 194-202. Evans, M. J. & Kaufman, M. H. (1981). Establishment in culture of pleuripotential cells from mouse embryos. Nature 292, 154-6. Eyal-Giladi, H. (1984). The gradual establishment of cell commitments during the early stages of chick development. Cell. Diff. 14, 245-55. Eyal-Giladi, H. & Kochav, S. (1976). From cleavage to primitive streak formation: A complementary normal table and a new look at the first stages of the development of the chick. 1. General Morphology. Dev. Biol. 49, 321-37. Eyal-Giladi, H., Debby, A. & Harel, N. (1992). The posterior section of the chick's area pellucida and its involvement in hypoblast and primitive streak formation. Development 116, 819-30. Fargeix, N. (1969). Les cellules germinales du canard chez des embryons normaux et des embryons de regulation: Etude des jeunes stades du developpement. J. Embryol. Exp. Morphol. 22, 477-503. Fargeix, N. (1986). Regulation du nombre des genocytes dans les ebauches gonadique de Tembryon de canard apres destruction partielle du stock initial des cellules germinales. C.R. Acad. Sci. Paris. 282D, 305-8. Ferhane-Tachinante, F. (1975). Experimental study of exchanges of germ cells in birds and mammals by culture and graft techniques. Bull. Soc. Sci. Nat. Phys. Maroc. 55, 21-90. Gardner, R. L. & Papaioannou, V. (1975). Differentiation in the trophectoderm and inner cell mass. In The Early Development of Mammals, ed. M. Balls & A. E. Wild, pp. 107-32. Cambridge University Press. Gavora, J. S., Benkel, B., Sasada, H., Cantwell, W. J., Fiser, P., Teather,

Transgenie birds

133

R. M., Nagai, J. & Sabour, M. P. (1991). An attempt at sperm-mediated gene transfer in mice and chickens. Can. J. Anim. Sci. 71, 287-91. Ginsburg, M. & Eyal-Giladi, H. (1987). Primordial germ cells of the young chick blastoderm originate from the central zone of the area pellucida irrespective of the embryo-forming process. Development 101, 209-19. Ginsburg, M. & Eyal-Giladi, H. (1989). Primordial germ cell development in cultures of dispersed central discs of stage X chick blastoderms. Gamete Res. 23, 421-8. Gruenbaum, Y., Revel, E., Yams, S. & Fainsod, A. (1991). Sperm cells as vectors for the generation of transgenic chickens. /. CellBiochem. Suppl. 15E, 194 (abstract). Haffen, K. (1975). Sex determination of avian gonads in vitro. Am. Zool. 15, 257-72. Hajji, K., Martin, C , Perramon, A. & Dieterlen-Lievre, F. (1988). Sexual phenotype of avian chimaeric gonads with germinal and stromal cells of opposite genetic sexes. Biol. Struct. Morphol. 1, 107-16. Hallett, J. S. & Wentworth, B. C. (1991). The effect of busulphan on gonadal differentiation in Japanese quail (Coturnix coturnix japonica). Poultry Sci. 70, 1618-23. Hamburger, V. & Hamilton, H. L. (1951). A series of normal stages in the development of the chick embryo. J. Morphol. 88, 49-92. Hemsworth, B. N. & Jackson, H. (1962). Effect of busulphan on the foetal gonad. Nature, Lond., 195, 816-17. Howarth, B. (1971). An examination of sperm capacitation in the fowl. Biol. Reprod. 3, 338-41. Hughes, G. C. (1963). The population of germ cells in the developing female chick. J. Embryol. Exp. Morphol. 11, 513-36. Hughes, G. C. (1964). The radiosensitivity of oogonia and oocytes in the chick embryo. J. Embryol. Exp. Morphol. 12, 273-87. Johnson, R. J. (1988). Diminution of pulsatile growth hormone secretion in the domestic fowl (Gallus domesticus): Evidence of sexual dimorphism. J. Endocrinol. 119, 101-9. Kopchick, J. J., Mills, E., Rosenblum, C , Taylor, J., Macken, F., Leung, F., Smith, J. & Chen, H. (1991). Methods for the introduction of recombinant DNA into chicken embryos. In Transgenic Animals, ed. N. L. First & F. P. Haseltine, pp. 275-93. Boston: Butterworth-Heinemann. Kuwana, T., Maeda-Suga, H. & Fujimoto, T. (1986). Attraction of chick primordial germ cells by gonadal anlage in vitro. Anat. Rec. 215, 403-6. Lake, P. E. (1971). The male in reproduction. In Physiology and Biochemistry of the Domestic Fowl, ed. D. J. Bell, & B. M. Freeman, Vol. 3, pp. 1411-47. London: Academic Press. Lavitrano, M., Camaioni, A., Fazio, V. M., Dolci, S., Farace, M. G. & Spadafora, C. (1989). Sperm cells as vectors of introducing foreign DNA into eggs: Genetic transformation of mice. Cell 57, 717-23. Le Douarin, N. & McLaren, A., eds. (1984). Chimaeras in Developmental Biology. London: Academic Press. Lee, H., Karasany, N. & Nagele, R. G. (1978). The role of cell surface in the migration of primordial germ cells in early chick embryos: Effects of concanavalin A. J. Embryol. Exp. Morphol. 46, 5-20. Lee, M.-R. & Shuman, R. (1990). Transgenic quail produced by retrovirus vector infection transmit and express a foreign gene marker. Proc. 4th World Congr. Genet. Appl. Livestock Prod. 16, 107-10.

134

K. Simkiss

Lillie, F. R. (1919). The Development of the Chick: An Introduction to Embryology, 2d ed. New York: Holt. Loveless, W., Bellairs, R., Thorpe, S. J., Page, M. & Feizi, T. (1990). Developmental patterning of the carbohydrate antigen FC102 during early embryogenesis in the chick. Development 108, 97-106. Marzullo, G. (1970). Production of chick chimaeras. Nature, Lond. 225, 72-3. Matsui, Y., Zsebo, K. & Hogan B. L. M. (1992). Derivation of pluripotential embryonic stem cells from murine primordial germ cells in culture. Cell 70, 841-7. McCarrey J. R. & Abbott, U. K. (1982). Functional differentiation of chick gonads following depletion of primordial germ cells. J. Embryol. Exp. Morphol. 68, 161-74. McLaren, A. (1992). The quest for immortality. Nature, Lond. 359, 482-3. Meyer, D. B. (1964). The migration of primordial germ cells in the chick embryo. Dev. Biol. 10, 154-90. Meyers, N. L., Booth, S. C , Bumstead, N., Vick, L. & Simkiss, K. (1991). Development and testing of a packaging cell line for avian retroviral vectors. Arch. Virol. 119, 257-64. Mimms, M. F. & McKinnell, R. G. (1971). Laser irradiation of the chick embryo germinal crescent. /. Embryol. Exp. Morphol. 26, 31-6. Mongin, P. & Sauveur, B. (1970). Composition du fluide uterin et de l'albumen durant le sejour de l'oeuf dans l'uterus chez la poule domestique. C. R. Acad. Sci. Paris 270D, 1715-18. Motro, B., Kooy, D. Vander., Rossant, J., Reith, A. & Bernstein, A. (1991). Continuous patterns of ckit and steel expression: Analysis of mutation at the W & SI loci. Development 173, 1207-21. Muniesa, P. & Dominguez, L. (1990). A morphological study of primordial germ cells at pregastrular stages in the chick embryo. Cell Diff. Dev. 31, 105-17. Naito, M., Nirasawa, K. & Oishi, T. (1990). Development in culture of the chick embryo from fertilized ovum to hatch. /. Exp. Zool. 254, 322-6. Naito, M., Agata, K., Otsuka, K., Kino, K., Ohta, M., Hirose K., Perry, M. M. & Eguchi, G. (1991a). Embryonic expression of p actin-lac Z hybrid gene injected into the fertilized ovum of the domestic fowl. Int. J. Dev. Biol. 35, 69-75. Naito, M., Watanabe, M., Kinutani, M., Nirasawa, K. & Oishi, T. (1991b). Production of quail-chick chimaeras by blastoderm cell transfer. Br. Poultry Sci. 32, 79-86. Nichols, J., Evans, E. P. & Smith, A. G. (1990). Establishment of germ-line competent embryonic stem (ES) cells using differentiating inhibiting factor. Development 110, 1341-8. Nieuwkoop, P. D. & Satasurya, L. A. (1979). Primordial Germ Cells in the Chordates. Cambridge University Press. Ono, T. & Wakasugi, N. (1984). Mineral content of quail embryos cultured in mineral rich and mineral free conditions. Poultry Sci. 63, 159-66. Page, N., Savva, D., Luke, G. & Simkiss, K. (1991). Transfection of chick cells by non-retroviral DNA. Biochem. Soc. Trans. 19, 328. Palmiter, R. D. & Brinster, R. L. (1986). Germ-line transformation of mice. Annu. Rev. Genet. 20, 465-99. Pandey, K. K. & Patchell, M. R. (1982). Genetic transformation in chicken by the use of irradiated male gametes. Mol. Gen. Genet. 186, 305-8. Pardanaud, L., Altmann, C , Kitos, P., Dieterlen-Lievre, F. & Buck, C.

Transgenie birds

135

(1987a). Vasculogenesis in the early quail blastodisc as studied with a monoclonal antibody recognizing endothelial cells. Development 22, 4760. Pardanaud, L., Buck, C. & Dieterlen-Lievre, F. (1987b). Early germ cell segregation and distribution in the quail blastoderm. Cell Diff. 22, 47-60. Perry, M. M. (1987). Nuclear events from fertilization to the early cleavage stages in the domestic fowl (G. domesticus). J. Anat. 150, 99-109. Perry, M. M. (1988). A complete culture system for the chick embryo. Nature, Lond. 331, 70-2. Perry, M. M., Morrice, D., Hettle, S. & Sang, H. (1991). Expression of exogenous DNA during the early development of the chick embryo. Roux's Arch. Dev. Biol. 200, 312-19. Petitte, J. N., Clark, M. E., Luv, G., Gibbins, A. M. V. & Etches, R. J. (1990). Production of somatic and germ line chimaeras in the chicken by transfer of early blastodermal cells. Development 108, 185-95. Petitte, J. N., Clark, M. E. & Etches, R. J. (1991). Assessment of functional gametes in chickens after transfer of primordial germ cells. / . Reprod. Fert. 92, 225-9. Resnick, J. L., Bixler, L. S., Cheng, L. & Donovan, P. J. (1992). Long term proliferation of mouse primordial germ cells in culture. Nature, Lond. 359, 550-1. Reynaud, G. (1969). Transfert de cellules germinales primordiales de Dindon a l'embryon de Poulet par injection iuntravasculaire. /. Embryol. Exp. Morphol. 21, 485-507. Reynaud, G. (1976). Capacites reproductrices et descendance de poulets ayant subi un transfert de cellules germinales primordiales durant la vie embryonnaire. Wilhelm Roux's Arch. Dev. Biol. 179, 85-110. Reynaud, G. (1981). The effect of busulfan on the germ cell line of the quail embryo. Arch. Anat. Microscop. Morphol. Exp. 70, 251-8. Robertson, E. J. (1986). Pluripotential stem cell lines as a route into the mouse germ line. Trends Genet. 2, 9-13. Rogulska, T., Ozdzenski, W. & Komar, A. (1971). Behaviour of mouse primordial germ cells in the chick embryo. / . Embryol. Exp. Morphol. 25, 155-64. Romanoff, A. L. (1972). Pathogenesis of the Avian Embryo. New York: Wiley-Interscience. Rosenblum, C. I. & Chen, H. Y. (1991). Liposome-mediated gene transfer into chicken embryos in ovo. J. Cell. Biochem. Suppl. 15E, 208 (abstract). Rowlett, K. (1991). Embryo growth and development in culture. In Avian Incubation, ed. S. G. Tullett, pp. 107-24. Boston: ButterworthHeinemann. Rowlett, K. A. & Simkiss, K. (1985). The surrogate egg. New Scientist, 1469, 42-4. Rowlett, K. & Simkiss, K. (1987). Explanted embryo culture: In vitro and in ovo techniques for domestic fowl. Br. Poultry Sci. 28, 91-101. Salter, D. W., Smith, E. J., Hughes, S. H., Wright, S. E., Fadly, A. M., Witter, R. L. & Crittenden, L. B. (1986). Gene insertion into the chicken germ line by retroviruses. Poultry Sci. 65, 1445-58. Salter, D. W., Smith, E. J., Hughes, S. H., Wright, S. E. & Crittenden, L. B. (1987). Transgenic chickens: Insertion of retroviral genes into the chicken germ line. Virology, 157, 236-40.

136

K. Simkiss

Sang, H. & Perry, M. M. (1989). Episomal replication of cloned DNA injected into the fertilized ovum of the hen Gallus domesticus. Mol. Reprod. Dev. 1, 98-106. Savva, D., Page, N., Vick, J. & Simkiss, K. (1991). Detection of foreign DNA in transgenic chicken embryos using the polymerase chain reaction. Res. Vet. Sci. 50, 131-3. Shih, C.-C, Stoye, J. P. & Coffin, J. M. (1988). Highly preferred targets for retroviral integration. Cell 53, 531-7. Shoffner, R. N., Otis, J. S. & Guise, K. (1987). Improbability of transfection by irradiated spermatozoa in the chicken. Poultry Sci. 66(suppl.) 176 (abstract). Shuman, R. M. (1986). Gene transfer in avian species using retrovirus vectors, 7th European Poultry Congress, pp. 24-37. Shuman, R. M. (1990). Genetic engineering. In Poultry Breeding and Genetics, ed. R. D. Crawford, pp. 585-98. Amsterdam: Elsevier. Shuman, R. M. (1991). Production of transgenic birds. Experientia 47, 897905. Simkiss, K. (1991). Primordial germ cells and the scope for genetic manipulation using embryos. In Avian incubation, ed. S. Tullett, pp. 125-36. London: Butterworth-Heinemann. Simkiss, K. (1993). Surrogate eggs, chimaeric embryos & transgenic birds. Comp. Biochem. Physiol. 104A, 411-17. Simkiss, K., Rowlett, K., Bumstead, N. & Freeman, B. M. (1989). Transfer of primordial germ cell DNA between embryos. Protoplasma 151, 164-6. Simkiss, K., Vick, L., Luke, G., Page, N. & Savva, D. (1990). Infection of primordial germ cells with defective retrovirus and their transfer to the developing embryo. 4th World Congr. Genet. Appl. Livestock Prod. 16, 111-14. Simkiss, K., Li, Y., Luke, G., Page, N. & Vick. L. (1993). The production of transgenic birds from primordial germ cell manipulated embryos. In Avian Endocrinology, ed. P. J. Sharp, pp. 367-80. Bristol: Journal of Endocrinology. Simon, D. (1960). Contribution a Tetude de la circulation et du transport des gonocytes primaires dans les blastodermes d'oiseau cultives in vitro. Arch. Anat. Microsc. Morphol. Exp. 49, 93-176. Spencer, J. L., Gavora, J. S. & Gowe, R. S. (1980). Lymphoid leukosis virus: Natural transmission and non-neoplastic effects. Proc. Cold Spring Harbor Conf. CellProlif. 7, 553-64. Spratt, N. T. & Haas, H. (1960). Morphogenetic movements in the lower surface of the incubated and early chick blastoderm. J. Exp. Zool., 144, 139-57. Stern, C D. (1990). The marginal zone and its contribution to the hypoblast and primitive streak of the chick embryo. Development 109, 667-82. Stewart, C. L., Gadi, I. & Bhatt, X. (1994). Stem cells from primordial germ cells can reenter the germ line. Dev. Biol. 161, 626-8. Temin, H. M. (1989). Retrovirus vectors, promise and reality. Science 246, 983. Thomas, K. R. & Capecchi, M. R. (1987). Site-directed mutagenesis by gene targeting in mouse embryo derived stem cells. Cell 51, 503-12. Tomita, T., Yamamoto, N., Otsuka, K., Ohta, M. & Hirose, K. (1988). Genetic transformation of eggshell colour in chickens by the use of irradi-

Transgenie birds

137

ated sperm. In Proceedings of the 18th World Poultry Congress, Nagoya, Japan, pp. 515-16. Ukeshima, A. & Fujimoto, T. (1984). Ultrastructure of primordial germ cells in the early chick embryo. In Ultrastructure of Reproduction, ed. J. Van Blerkom & P. M. Motta, pp. 12-18. Boston: Martinus Nijhoff. Urven, L. E., Erickson, C. A., Abbott, U. K. & McCarrey, J. R. (1988). Analysis of germ line development in the chick embryo using an antimouse EC cell antibody. Development 103, 299-304. Varmus, H. (1988). Retroviruses. Science. 240, 1427-35. Vick, L., Li, Y. & Simkiss, K. (1993). Transgenic birds from transformed PGCs. Proc Roy. Soc. Lond. 251, 179-82. Watanabe, S. & Temin, H. M. (1983). Construction of a helper cell line for avian reticulo endotheliosis virus cloning vectors. Mol. CellBiol. 3, 2241-9. Wentworth, A. L., Frank, A. & Wentworth, B. C. (1992). Transfection of Japanese quail primordial germ cells with DNA expressing gene product lac Z. Poultry Sci., 71(Suppl. 1), 73 (abstract). Wentworth, B. C. & Wentworth, A. (1991). Primordial germ cell characteristics in the Japanese quail embryo. Poultry Sci., 70(Suppl. 1), 129 (abstract). Wentworth, B. C , Tsai, H., Hallett, J. H., Gonzales, D. S. & Rajcic-Spasojevic, G. (1989). Manipulation of avian primordial germ cells and gonadal differentiation. Poultry Sci. 68, 999-1010. Witte, O. N. (1990). Steel locus defines new multipotent growth factor. Cell. 63, 5-6. Yoshinaga, K., Fujimoto, T., Nakamura, M. & Terakura, H. (1992). Selective lectin-building sites of primordial germ cells in chick and quail embryos. Anat. Rec. 233, 625-32.

5 Transgenic rodents MARTIN J. EVANS, DARREN T. GILMOUR AND WILLIAM H. COLLEDGE

5.1 Introduction

The use of transgenic rodents, in particular transgenic mice, to address a number of diverse biological problems has been so vast that this review chapter cannot cover every facet of transgenesis. We hope that by explaining the technology associated with the generation of transgenic mice and by focusing on a few current areas of research where transgenic animals are proving invaluable, we will help the reader understand the use of transgenic animals in research, as well as the power and limitations of this approach. In this review a transgenic animal is defined as an animal that carries a foreign piece of DNA stably integrated into the genome of every cell in that animal. Thus, all somatic cells and the germ-cells of a transgenic animal carry the same DNA fragment at the same chromosomal location, and this DNA fragment can be transmitted in a Mendelian fashion to offspring. This DNA element is termed a trans gene. 5.2 Generation of transgenic mice There are essentially three ways of generating animals with the capacity to transmit a genetic element through the germ-line to their offspring. These are (a) retroviral integration into an early-developing embryo, (b) injection of DNA into the pronucleus of a newly fertilized egg or (c) genetic manipulation of embryonic stem cells (Fig. 5.1). All three routes have been employed to generate transgenic mice, while a few transgenic rat strains have been established chiefly through the pronuclear injection route. 138

1. DNA microinjection transgenesis Sperm Zygote

Morula

Blastocyst

Transgenic

Intercross or backcross to homozygosity 3. Retroviral Infection

Fig. 5.1. Diagram illustrating the three major routes to transgenesis ('ES cells' denotes embryonic stem cells).

140

Martin J. Evans et al.

5.2.1 Retroviral infection If preimplantation embryos are exposed to retrovirus, a proportion of the embryonic cells will stably integrate proviral sequences into their genome, usually as one copy per cell. Jaenisch (1976) showed that adult mice derived after Moloney murine leukaemia virus (M-MuLV) infection could transmit integrated proviral sequences through the germ-line. Unfortunately, preimplantation mouse embryos are not permissive for M-MuLV expression from the long terminal repeat promoter, so that genes driven by this promoter are not expressed in the embryo. Upon integration, the provirus is subject to de novo methylation (Jahner et al. 1982), effectively shutting off proviral transgene expression even in cell lineages derived from the original infected cells. This problem has been circumvented somewhat by the use of an internal promoter to give expression of a transgene. Such a promoter can be aimed at providing either ubiquitous expression of a transgene (e.g., herpes simplex virus thymidine kinase promoter; Stewart et al. 1987) or cell-specific expression (e.g., /3-globin promoter giving expression in haematopoietic tissues; Soriano et al. 1986). One of the limitations of using retroviruses to generate transgenic mice is the packaging limits of the virus, which restrict the insert size to approximately 9 kb. If the insert is much larger than this, the viral RNA cannot be packaged into the viral capsid. It should be noted that depending on the developmental stage at which the retroviral infection occurs and the number of embryonic cells infected, the resulting offspring are mosaic for the proviral sequences. That is, not all somatic cells will contain the proviral insert, and indeed different somatic cells might contain pro viruses integrated at different genomic locations. It is not until the provirus is transmitted through the germ-line that an animal can be truly referred to as transgenic.

5.2.2 Pronuclear injection Until recently, the major route for generating transgenic animals was through microinjection of DNA into the pronuclei of fertilized eggs (Gordon et al. 1980). The egg is held securely with a glass holding pipette, and the DNA solution is injected into a pronucleus by insertion of a fine glass injection needle. Usually the male pronucleus is injected, as it is normally larger than the female pronucleus. The injected DNA is integrated into the genome of 10-40% of surviving

Transgenic rodents

141

embryos. The injected zygotes are then transferred into foster mothers and allowed to develop to term. Carrier transgenic animals can pass the transgene through the germ-line as stable genetic information. In this way transgenic mice, rats, pigs and other animals have been generated. The advantage of this method is that large fragments of DNA can be injected into the pronucleus, thus allowing complete genes with associated regulatory regions to integrate into the genome. In addition, approximately 80% of the founder animals derived from the injected egg are true transgenics; every cell in the animal contains the integrated transgene. This indicates that integration of the injected DNA usually occurs at the one-cell stage. In approximately 20% of transgenic mice the exogenous DNA is integrated at a later stage to give mosaic animals similar to those generated by retroviral infection of preimplantation embryos. The transgene is often found in a multicopy, head-to-tail concatomeric array thought to result from homologous recombination between injected DNA molecules before integration (Constantini & Lacy 1981). The site of integration in the genome can be associated with rearrangements and duplications, making it difficult to clone flanking sequences adjacent to a transgene. The site of integration in the genome can affect the expression of the transgene, so that different founder animals with different integration sites often demonstrate widely different levels of expression. Nevertheless, appropriate temporal and spatial expression of transgenes has been observed in many instances, irrespective of chromosomal location. For some genes a relatively small region of flanking sequence is necessary to provide tissue-specific expression - for example, the ovine /3-lactoglobulin gene, which requires only 146 bp of 5' promoter sequence to be expressed specifically in the mammary gland of transgenic mice (Whitelaw et al. 1992). For a few genes, however, appropriate temporal and tissue-specific expression requires larger segments of flanking DNA to be introduced which typically contain DNase I hypersensitive sites. For the /3-globin gene, a 38-kb DNA segment containing DNase I hypersensitive sites, both 5' and 3' of the gene, is required to obtain transgene expression specific to erythroid lineages (Grosveld et al. 1987). Flanking sequences that control appropriate tissue-specific expression have been termed locus control regions (LCRs). The demonstration that yeast artificial chromosomes (YACs) can be used to generate transgenic mice now allows sufficiently large chunks (up to 250 kb) of DNA to be introduced into

142

Martin J. Evans et al.

the germ-line to guarantee the fidelity of temporal and spatial expression (Jakobovits et al 1993; Schedl et al 1993; Strauss et al 1993). Other factors have been shown to affect transgene expression, including the strain of mouse (Brinster et al 1985). Brinster and colleagues found that C57/BL6 x SJL hybrid mice were more likely than C57/BL6 inbred mice to express a growth hormone transgene. In addition, the presence of heterologous introns within the coding region can enhance the expression of some transgenes in mice (Palmiter et al 1991). 5.2.3 Embryonic stem cells Embryonic stem (ES) cells are derived from a preimplantation-stage embryo, usually at the 3.5-day blastocyst stage. Early embryos are flushed from the uterine horns and maintained in cell culture medium in order to hatch from the zona pellucida and attach to the substratum. Originally, ES cell lines were derived and grown on mitotically inactivated feeder layers which provided the necessary factors to prevent the differentiation and death of the ES cell lines (Evans & Kaufman 1981; Martin 1981). Under these conditions the cells of the attached blastocysts spread over the feeder layer. Cell lines can be isolated from the epiblast component and can be maintained in culture through multiple passages. Under appropriate culture conditions, these cell lines will maintain a normal karyotype and possess the ability to differentiate into a wide spectrum of cell types in vitro, including neurons, glia, skeletal, cardiac and smooth muscle, endoderm and haematopoietic precursors. A factor has been described, variously termed differentiating inhibitory activity (DIA), leukaemia inhibitory factor (LIF) or human interleukin for DA cells (HILDA), that is produced by many cell lines and prevents the spontaneous differentiation of ES cells in culture, thus circumventing the requirement of a feeder layer in their maintenance (Hilton & Gough 1991). It remains to be seen, however, whether the feeder layers produce other, as yet uncharacterized factors that help maintain the normal karyotype of ES cells. If ES cells are microinjected into the blastocoel cavity of a host blastocyst, they will combine with the inner cell mass component and contribute to the developing embryo (Bradley et al 1984). Offspring have somatic tissue composed of both ES cell-derived cells and host blastocyst-derived cells. These animals are termed chimeras and dem-

Transgenic rodents

143

onstrate the true pluripotentiality of the ES cells. Various markers can be used to demonstrate chimerism. If the ES cell line is derived from a strain of mouse with coat colour markers that are different from those of the strain used to obtain the host blastocyst, hair pigmentation can be easily used to assess contribution. To determine ES cell contribution to somatic tissues and organs, however, a useful marker is glucose-phosphate isomerase (GPI), which occurs in three isoforms in laboratory mice. The GPI isozymes can be readily and sensitively distinguished by electrophoresis. By utilizing an ES cell line with a GPI isotype that is different from the host blastocyst, one can assess ES cell contribution in somatic tissues removed from the animal. Most important, however, the chimeras contain not only ES cellderived somatic cell lineages but also functional ES cell-derived germ cells (Bradley et al. 1984). In some circumstances the germ-cells are derived solely from the ES cells, giving a chimera that will transmit only ES cell-derived genetic material to its offspring. This happens when a male ES cell (XY) is injected into a female (XX) blastocyst and the contribution of the ES cells in the developing genital ridge of the embryo diverts the normal development of the female to give rise to a male with only ES cell-derived sperm (Robertson 1986). The ability of ES cells to colonize the germ-line of chimeric animals and the fact that they can be maintained in culture for indefinite periods provides an alternative route to the generation of transgenic mice. Any genetic alteration (i.e., a transgene) introduced into the ES cells in culture can be incorporated into a chimera and subsequently passed to its offspring. This route to transgenesis has not been widely used, since it involves an additional breeding step to generate true transgenics, whereas in the pronuclear injection route founder transgenics are obtained in the Fl litters. In addition, retention of the ability of the ES cells to colonize the germ-line can sometimes present problems, particularly if the ES cells have been cultured under suboptimal conditions. Nevertheless, the ES cell route to transgenics does have several advantages over the pronuclear injection route. First, ES cell clones carrying an exogenously introduced gene can be characterized for gene expression and copy number before the clone with which to generate chimeras is chosen. Second, dominant lethal mutations can be analysed if not too detrimental to the chimeras. Dominant lethal mutants in transgenic mice generated by pronuclear injection can be maintained only if they have conditional expression. Since an ES cell

144

Martin J. Evans et al.

clone can be used again and again to generate chimeras identical in every way apart from levels of ES cell contribution, it is also possible to analyse developmental perturbations at the chimera level (Boulter et al. 1991). This is not possible through pronuclear injection, since every transgenic animal carries a different transgene integration site in the Fl founders. So far, embryonic stem cells have been isolated from only two rodent species: mouse and hamster (Doetschman et al. 1988). Little work has been carried out with hamster ES cell lines because of their unproven potential to colonize the germ-line. Several groups are trying to derive rat ES cells, since this is the animal of choice for many pharmacological, neurological and behavioural studies. It is anticipated that rat ES cells will be obtained in the very near future, thus making it possible to generate transgenic rats by a route other than pronuclear injection. 5.2A Gene targeting Gene targeting is defined as the genetic modification of an endogenous DNA sequence within a cell by homologous recombination with a DNA segment introduced into the cell. Gene targeting was first demonstrated in higher eukaryotic cells by Oliver Smithies and colleagues, who identified clones that had integrated plasmid sequences into the human j8-globin locus by homologous recombination. The frequency of homologous recombination was low (1 in 103 of all transformed cells), and Smithies et al. (1985) identified targeted clones by a sensitive screening assay that involved cloning the integrated DNA in bacteriophage A and screening plaques for /3-globin sequences. The first gene to be targeted in ES cells was the hypoxanthine phosphoribosyltransferase (hprt) gene. The hprt gene is located on the X chromosome and therefore is present as only one copy in male ES cells. Cells that have lost a functional hprt gene can be selected in 6thioguanine, a toxic purine analogue that kills hprt+ cells. Conversely, cells with a functional hprt gene can be selected in HAT (hypoxanthine— aminopterin—thymidine) medium. The ability to select readily for hprt~ ES cells and the presence of only one hprt gene in male ES cells allowed the identification of ES clones that had undergone inactivation of the hprt gene (Hooper et al. 1987; Kuehn et al. 1987). The demonstration that one of these hprt~ ES cell lines could be repaired through homologous recombination and still colonize the

Transgenic rodents

145

germ-line of chimeras illustrated that gene targeting could be used to generate mice theoretically carrying a specific mutation at any genetic locus (Thompson et al. 1989). An additional problem posed by targeted disruption of most genes other than hprt, however, is that no direct selection can be made to identify targeted clones. The targeting DNA must therefore contain a selectable marker so that cells which have taken up the DNA and stably incorporated it into their genome can be selected from the cells which did not take up or integrate that DNA. In addition, in only a small proportion of the transformed cells has homologous recombination occurred. Thus, various strategies were developed to enrich the population for targeted clones. The most widely used is that developed by Mansour et al. (1988), termed the positive-negative selection system (Fig. 5.2). This depends on using a replacement construct in which a region of nonhomology to be introduced into the locus is flanked by two regions homologous to the endogenous DNA. Notwithstanding details of the mechanism of homologous recombination, the final structure is equivalent to cross-overs having taken place in both homologous regions. Thus, sequences in the original construct distal to these will have to be lost. In contrast there is no a priori reason for their loss on random integration. Selection both for a positive (internal) marker and against a negative (terminal) marker will enrich for homologous as opposed to random integrations. The general experience with this method is that the enrichment is only about 10-fold. This is probably due to the fact that terminal sequences are relatively readily lost in random as well as in homologous integration events. Various genes can be used to provide the positive and negative selection in ES cells. A positively selectable marker (e.g., neo selected with G418) and a negative one [(e.g., herpes simplex virus (HSV) thymidine kinase (tk) gene selected against with gancyclovir, or l(2-deoxy-2-fluoro-/3-Darbinofuranosyl)salodouracil (FIAU)] are commonly used. These are listed in Table 5.1 along with the compounds used in the selection. For transfected cells to survive in the presence of G418 and gancyclovir, they must acquire the neo gene but not the HSV-fA: gene. Thus, if a construct is engineered in which the negative selectable marker can be lost only through homologous recombination with the target locus, a negative selection procedure should enrich for the appropriate targeting event (Fig. 5.2). An alternative approach to enriching the population of targeted clones is to use a targeting construct in which a promoterless se-

Martin J. Evans et al.

146 A) Gene Targeting

HSV-tk Gene X targeting vector

nr

3 " Gene X

neo r

_L X", neo r , HSV-tk" (G418 r , GANC r )

B) Random Integration

HSV-tk

Y/////A

I

I

Jf

neo r

LocusY

HSV-tk

X + n e o r HSV-tk + (G418 r , GANC S )

Fig. 5.2. Gene targeting using the positive-negative selection strategy to enrich the population of ES cells containing a targeted disruption of gene X. The gene X targeting vector contains a positive selectable marker, neo, and a counterselectable marker, HSY-tk, each complete with active promoter, polyadenylation site and transcriptional start and stop sites. The neo gene is inserted into an exon of the sequence homologous to the target gene (gene X), and the HSV-M: gene is placed outside the region of homology. Homologous recombination with the target gene (A) results in the disruption of one copy of gene X and expression of neo but loss of the HSV-f& gene. However, heterologous integration (B) usually occurs through the ends of the construct, resulting in incorporation of the HSV-fA; gene into the chromosome. Expression of

Transgenic rodents

147

Table 5.1. Marker genes for selection or counterselection of transformed ES cells Genotype

Selection agent

Counterselection

Neo +

G418

None

Hyg+ Hprt+ Hprt~

Hygromycin B HAT 6TG

6TG HAT

Gpt+

HAT

Tk

+

HAT

DT-A (diphtheria toxin A-fragment)

Possibly 6TG GANC or FIAU or acyclovir

—

No selective agent required

None

Reference Southern and Berg (1982) te Riele et al. (1990) Hooper (1987) Hooper (1987) Mulligan and Berg (1981) Borelli et al. (1988) Yagi et al. (1990)

lectable marker gene is fused in frame with the target gene. Random integration of this construct in the genome would not usually be associated with expression of the selectable marker gene. Upon gene targeting, however, the selectable marker gene is placed under the control of the target locus promoter, thus conferring resistance to the selection drug. This approach generally gives enrichments of up to 100-fold but can be used to target only genes that are expressed in ES cells. Such enrichment strategies are necessary because for many target loci the frequency of random integration of the targeting construct is much higher than the frequency of homologous recombination. Many factors affect the frequency of homologous recombination at a chromosomal location, including the length of homology (Thomas & Capecchi 1987; Shulman et al. 1990; Hasty et al. 1991b; Deng & Capecchi 1992) and the topology of the plasmid DNA (Mansour et al. 1988; Hasty et al. 1991; Deng & Capecchi 1992). The presence of a ZCaption to Fig. 5.2 (cont.). HSW-tk renders cells sensitive to gancyclovir (GANC), so that selection of G418 plus GANC enriches for cells that have undergone type (A) events at the expense of those that have undergone events of type (B). Open boxes denote introns or flanking DNA sequences; chequered boxes, gene X exons; hatched boxes, neox genes; stippled boxes, HSVtk genes;filledboxes, sequences of a heterologous gene designated locus Y. (Reproduced from Mansour et al. 1988.)

148

Martin J. Evans et al.

DNA motif promotes homologous recombination between plasmids in cultured mammalian cells (Wahls et al. 1990), as does the addition of dideoxy nucleotides to the ends of DNA fragments to prevent end joining (Chang & Wilson 1987). The most dramatic improvement in the frequency of gene targeting has been found by using isogenic DNA. Initially, many gene-targeting experiments were performed using ES cells derived from a 129 strain of mouse and targeting constructs usually carrying DNA isolated from a BALB/c genomic library. Te Riele et al. (1992) demonstrated that homologous recombination at the retinoblastoma susceptibility gene (Rb) was 20-fold more efficient with a 129-derived targeting construct than with a BALB/c-derived construct when 129-derived ES cells were used. Slight sequence divergencies between the 129 and BALB/c DNA, including base-pair substitutions, small deletions/insertions and a polymorphic CA repeat are thought to be responsible. Similarly, Miller et al. (1992) found that a targeting vector containing sequences from the renin gene Ren-ID preferentially targeted the Ren-ID gene and not the Ren-2 gene, although the two genes differ by only 5%. Deng and Capecchi (1992) have shown that isogenic DNA targets the Hprt locus of ES cells four to five times more efficiently than nonisogenic vectors. The genetic manipulation of genes within their endogenous chromosomal environment and the transmission of these genetic alterations into mice are powerful means for addressing many biological questions. It is therefore not surprising that gene targeting in ES cells has become a technique widely used by developmental biologists to assess the function of genes in embryogenesis. The catalogue of genes that have been disrupted and for which mutant animals have been generated is ever increasing. 5.3 Applications 5.3.7 Assessment of a gene's normal function Inappropriate, deregulated expression of a transgene in a wholeanimal environment has proved a useful tool with which to try to elucidate the normal function of a gene product. For example, deregulated c-fos expression under the control of the metallothionein I promoter interfered with normal bone development, suggesting a role for this protein in bone modelling (Ruther et al. 1987). Under the control of the H-2Kb promoter, expression of c-fos specifically interfered with thymus development even though the transgene was also expressed in

Transgenic rodents

149

other tissues (Ruther et al. 1988). Transgenic mice with inappropriate expression of the mos proto-oncogene develop neuropathological changes in the brain, suffer progressive limb paralysis and show aberrant eye lens fibre differentiation (Khillan et al. 1987; Propst et al. 1992). If a transgene is expressed in many tissues, including those where the endogenous version of the gene is not normally expressed, it is questionable whether abnormalities in growth and development are directly related to the normal function of the gene. Clearly, certain tissues are more susceptible than others to particular transgene expression, but this may be more biologically relevant if those tissues normally express the endogenous copy of the transgene at some point in their development. 5.3.2 Oncogenicity studies Transgenic mice carrying the SV40 T-antigen develop a wide variety of tumour types, including those of the pancreas (both exocrine and endocrine), lens, choroid plexus, thymus and pituitary gland. The tumour type depends primarily on the tissue specificity of expression of the T-antigen transgene rather than affecting one particular cell type. This indicates that almost any cell type is susceptible to transformation by T-antigen. T-antigen transgenic mice have been very useful in generating immortal cell lines from tissues that are refractory to establishing lines in standard ways. Similarly, an 'Immorto Rat' is currently being produced by pronuclear injection which carries a temperature-sensitive SV40 large T-antigen and can therefore be used to immortalize rat cell types in a temperature-controlled manner. Numerous transgenic animals have been generated expressing various oncogenes (e.g., ras, myc, abl). One of the first, the so-called OncoMouse, was created by Leder et al. (1986). This mouse strain carries the viral Harvey ras (y-Ha-ras) oncogene controlled by the mouse mammary tumour virus (MMTV) promoter. Thus, expression of the activated ras oncogene is under hormonal control, and females typically develop breast cancer upon sexual maturation. Other tumour types arise, however, with predictable kinetics in both sexes, including parotid adenocarcinomas and Haderian hyperplasias (Leder et al. 1986). Tumour development requires long latency periods consistent with oncogenesis as a multistage process requiring several activation steps (Weinberg 1985; Bishop 1987). Indeed, cross-breeding of myc

150

Martin J. Evans et al.

and ras transgenics to generate myc-ras compound transgenic mice resulted in accelerated tumour formation (Sinn et al. 1987). A noteworthy conclusion from these experiments, however, is that some cell lines are more susceptible to tumour formation initiated by expression of a particular oncogene than others. Thus, the ras oncogene can give rise to pancreatic acinar tumours, while the myc oncogene alone does not (Quaife et al. 1987). For some oncogenicity studies, particularly the evaluation of a compound's capacity to act as a carcinogen or teratogen, the rat is the animal of choice, as it has been used for many years by pharmaceutical companies as the bench-mark in such tests. It may not be long before a so-called OncoRat is developed much as the OncoMouse was. Animals expressing activated oncogenes that predictably develop specific tumour types are very useful in many fields of cancer research. These are described in the following subsections. 5.3.2.1 Studying the pathogenesis of cancer The mechanism(s) by which oncogenes influence tumour development can be assessed, and biochemical and immunological changes in the animal studied. 5.3.2.2 Testing the carcinogenicity of compounds It would be expected that transgenic animals carrying an activated oncogene would be more susceptible to tumour induction in response to exposure to a carcinogenic compound. The latency period for tumour development should be reduced. The effect of a carcinogenic compound on a specific organ (e.g., liver) can be studied if a transgenic animal is generated with tissue-specific expression of the oncogene only in that organ. 5.3.2.3 Screening anticancer compounds Since transgenic mice expressing activated oncogenes develop specific tumour types with predictable kinetics, they can be used in an initial screen for anticancer agents. Many compounds can be rapidly screened to identify those with antineoplastic effects. Since these studies are carried out on live animals, any toxic side effects of such compounds are identified at an early stage.

Trans genic rodents

151

5.3.3 Assignment of a gene to a particular genetic defect If a naturally occurring mouse mutant already exists that displays a developmental defect, it is sometimes possible to verify the gene involved by correcting the defect by expression of a candidate transgene in the mutant animals. For example, introduction of a transgene encoding the myelin basic protein corrected the tonic seizures and early death associated with 'shiverer' mutant mice (Readhead et al. 1987). A similar approach using YACs to introduce large genomic segments from a chromosomal region identified as containing a mutant locus should facilitate cloning of these genes. 5.3.4 Models of human diseases and disorders 5.3.4.1 Via the pronuclear injection route Transgenic animals have been generated that show some of the physiological and pathological changes associated with human genetic disorders or diseases. The demonstration by Behringer et al. (1989) that a- and /3-globin genes could be correctly co-expressed in erythroid tissues of transgenic mice laid the foundation for trying to generate mouse models of various human haemoglobinopathies such as sickle cell disease and thalassaemias. Expression of a human sickle haemoglobin transgene in /3-thalassaemic mice gave partially anaemic animals whose erythrocytes became sickle-shaped when subjected to low oxygen tension (Ryan et al. 1990). Similarly, Greaves et al. (1990) produced a transgenic mouse model of the sickle cell disorder by coordinately expressing the human a-globin and sickle cell /3-globin genes in transgenic mice. Transgenic mice models of diabetes mellitus have also been made. Allison et al. (1988) overexpressed a class I histocompatibility antigen (H-2Kb) in the pancreatic /3-cells of transgenic mice which subsequently developed insulin-dependent diabetes without an autoimmune response. In a separate study, overexpression of calmodulin, a Ca 2+ binding protein involved in signal transduction, was found to induce very early onset diabetes, within hours after birth (Epstein et al. 1989). Several mouse models of human neoplasias have also been developed. For example, Windle et al. (1990) found that transgenic mice expressing the SV40 T-antigen in the retina develop heritable ocular tumours with histological, ultrastructural and immunohistochemical features identical to those of human retinoblastoma, an autosomal

152

Martin J. Evans et al.

recessive eye malignancy occurring in young children. Similarly, transgenic models of hepatocarcinogenesis have been generated by directing the expression of different oncogenes (SV40 T-antigen, c-Ha-ras or c-myc) to the liver of transgenic mice using the albumin enhancer/promoter (Sandgren et al. 1989). Transgenic mice expressing an MMTV-human TGFa fusion gene develop mammary gland hyperplasia culminating in adenocarcinoma development (Halter et al. 1992). One transgenic MMTV-TGFa line was found also to develop sebaceous gland hyperplasia of the skin. Transgenic mice overexpressing TGFa in the stomach induced severe adenomatous hyperplasia and other structural lesions that were similar to Menetrier's disease in humans (Takagi et al. 1992). On the basis of these studies it appears that TGFa is very effective in transforming epithelial cells and may therefore be involved in the development of organs containing epithelial cells such as kidney, lung and intestine. When the TGFa locus was inactivated by gene targeting (Luetteke et al. 1993; Mann et al. 1993), there were only minor effects on hair and whisker morphology and on the cornea. This phenotype identified an already known spontaneous mutant, waved-1, which indeed proved to be a loss-of-function allele at the TGFa locus. These observations re-emphasize that the normal function of a locus may not be revealed by transgenic overexpression experiments. An area of intense research is concerned with generating an animal with symptoms of the neurological degenerative disorder Alzheimer's disease (AD). This disorder is characterized by the formation of senile plaques and neurofibrillary tangles throughout the central cortex and hippocampus regions of the brain. Two proteins have been identified as major components of senile plaques: amyloid /3-protein and the serine protease inhibitor antichymotrypsin. The /3-protein is generated by proteolytic cleavage of a larger, cell-surface-associated glycoprotein termed amyloid precursor protein (APP). Three forms of APP exist; the larger variant (APP-751) contains a region homologous to the Kunitz-type protease inhibitors which is not present in the major smaller precursor (APP-695). The amyloid protein has been directly implicated in the development of the neuropathology of AD; there have been reports that in a few families with hereditary predisposition to the disorder there is a missense mutation close to the C-terminal end of /3-amyloid segregates with the phenotype (Chartier-Harlin et al. 1991; Goate et al. 1991). Indeed, j8-amyloid is directly toxic to neurons in vitro (Yankner et al. 1990). These observations have led to the

Transgenie rodents

153

hypothesis that the accumulation of /3-amyloid itself is responsible for the neurodegeneration in AD. Unfortunately, while transgenic mice expressing APP-751 show amyloid deposits in the cortex and hippocampus, they do not develop other pathology, which limits their use as a model system for AD (Quon et al. 1991). It remains to be seen whether overexpression of the shorter isoforms of APP will generate a transgenic mouse with AD pathology. The number of transgenic rats generated via pronuclear injection is considerably lower than the number of transgenic mice. Nevertheless, some transgenic rats have been produced for modelling human disorders. For example, rats that mimic human hypertension have been made by the introduction of renin transgenes (Ganten et al. 1992; Yamaguchi et al. 1992). Transgenic rats that express the human apolipoprotein A-I gene show a raised total serum high-density lipoprotein (HDL) cholesterol level and may be useful as models for understanding lipoprotein metabolism (Swanson et al. 1992). Transgenic rats expressing both a class I MHC molecule (HLA-B27) and a human j82microglobulin transgene are susceptible to inflammatory disease when expression of B27 is above a critical threshold (Taurog et al. 1993). These animals may be useful in understanding the pathogenesis of arthritis development.

5.3.4.2 Via ES gene targeting Not only can gene targeting be used to generate null mutant animals for elucidating the normal function of a gene product, it can also be used to generate mouse models of human genetic disorders. Once a gene involved in a human genetic syndrome is mapped and the causative mutation identified, it becomes possible to engineer an analogous mutation into the murine homologue of the gene. The first mouse locus to be mutated as the Hprt gene in an attempt to create an animal model for Lesch-Nyhan syndrome, a sex-linked recessive disease causing neurological and behavioural problems (Hoper et al. 1987; Kuehn et al. 1987). Hprt~ mice are relatively normal and show no major metabolic or neurological characteristics associated with the syndrome. Lack of a phenotype in these mutant mice appears to be caused by purine metabolism differences between rodents and humans. HPRT is the key enzyme for purine salvage in humans, while adenine phosphoribosyltransferase (APRT) is more important in rodents. The administration of an APRT inhibitor to HPRT-deficient mice induces persis-

154

Martin J. Evans et al.

tent self-injurious behaviour, which is one of the behavioural alterations associated with the Lesch-Nyhan syndrome (Wu & Melton 1993). Other mouse models of human genetic disorders have been generated. Tybulewicz et al. (1992) disrupted the glucocerebrosidase gene in ES cells to generate a model of Gaucher's disease, a lysosomal storage disorder. Mice homozygous for the mutation have less than 4% of normal glucocerebrosidase activity and fail to degrade the sphingolipid glucocerebroside. These animals die within 24 hr of birth. To generate a mouse model for spontaneous atherosclerosis, Maeda and colleagues have been engaged in inactivating the genes involved in lipid metabolism. Apolipoprotein A-I (apoA-I) is the major protein complexed with HDL in mammals, and it also participates in cholesterol metabolism. In humans, mutations of the gene coding for apoA-I are correlated with a predisposition to atherosclerosis. Mice lacking apoA-I protein show a marked reduction of plasma HDL cholesterol (Williamson et al. 1992). A reduction in plasma HDL levels in humans is associated with an increased risk of atherogenesis, so it is hoped that these animals may develop atherosclerotic plaques with age. Another important gene involved in lipid metabolism is that for apolipoprotein E (apoE). ApoE is a glycoprotein that forms a structural component of all lipoprotein aggregates apart from low-density lipoprotein (LDL). It functions mainly as a ligand for specific receptor molecules present on the surface of hepatocytes, thus allowing apoEcontaining particles to be removed from the circulatory system by the liver for processing. Mice lacking apoE show elevated plasma cholesterol levels and develop spontaneous arterial lesions that gradually occlude the coronary and pulmonary arteries (Plump et al. 1992; Zhang et al. 1992). A third gene involved in circulatory lipid metabolism is the apob gene, which encodes a major structural component of LDL molecules. In humans, mutations in the APOB gene cause familial hypobetalipoproteinaemia (HBL), a condition characterized by a reduction in circulating apoB, /3-lipoproteins and total cholesterol. Affected individuals show a reduced rate of coronary vascular disease. Targeted modification of the apoB gene generates mice showing HBL (Homanics et al 1993). The apoA-I-, apoE- and a/?a#-deficient mice will be valuable for investigating the genetic and environmental factors that modulate atherogenesis. Compound mutants bearing both apoA-I and apoE defective genes would be expected to develop atherosclerosis at an elevated

Transgenic rodents

155

rate. Crossing a/?oE-deficient mice with HBL animals might reduce the rate of atherosclerotic development. Cystic fibrosis (CF) is the most common autosomal recessive genetic disorder among Caucasians, affecting approximately 1 in 2,000 live births. The disorder is characterized by a panoply of symptoms, including elevated salt levels in sweat, hyperaccumulation of mucus in the airways and gastrointestinal tract, pancreatic enzyme insufficiency, malabsorption of intestinal contents, intestinal obstructions and male sterility. The mucus accumulation in the airways promotes microbial colonization, which gradually destroys the lung tissue, resulting in death. All the organs affected in CF are associated with epithelial cells. The gene responsible for CF, the cystic fibrosis transmembrane conductance regulator (CFTR) gene, encodes a Cl" channel that is regulated by intracellular cAMP levels and nucleotide triphosphate binding and located within the apical membrane of epithelial cells. Identification of the most common CFTR mutation, present in about 75% of CF patients, as the deletion of phenylalanine 508 (AF508) started a race by several laboratories to generate a mouse model for CF. Almost simultaneously, three groups reported the successful generation of mice carrying a disrupted cftr gene. Snouwaert et al. (1992) and Ratcliff et al. (1992) used replacement-type targeting constructs to disrupt exon 10 of the cftr gene in ES cells. Dorin et al. (1992) used an insertion vector to disrupt exon 10, but this type of targeting construct also duplicated the cftr sequence around exon 10. When the carrier CF animals were crossed to generate homozygous mutants, the phenotypes of the animals were not identical among the three groups. Mice that carried the replacement arrangement had the more severe phenotype, and many died soon after birth from intestinal blockages akin to neonatal intestinal obstructions found in human CF patients (Colledge et al. 1992; Snouwaert et al. 1992). In addition, animals that survived the neonatal period failed to thrive, suffered from distal intestinal obstructions, showed pancreatic and lacrymal gland pathology and did not possess a cAMP-stimulated Cl ~ channel activity (Clarke et al. 1992; Snouwaert et al. 1992; Ratcliff e/ al. 1993). The CF animals generated by Dorin et al. (1992) showed no overt clinical disease and developed only minor pathological changes. The differences in phenotype between these animal models can be ascribed to the type of genetic alteration (i.e., replacement vs. insertion) engineered into the cftr gene. It is likely that the insertion muta-

156

Martin J. Evans et al.

tion can undergo transcriptional readthrough, and because of the gene duplication these transcripts may be alternatively spliced to generate wild-type message. There is a precedent for alternative splicing in an insertion mutation giving rise to mature wild-type mRNA. Moens et al. (1992) found that an insertion within an N-myc coding exon could still give rise to full-length functional mRNA at a low level by exon splicing. None of the CF animal models yet described has developed the full range of pathological symptoms observed in CF patients. In particular, the CF mice do not display severe lung pathology or recurrent lung infections, which are the major cause of death in humans. Nevertheless, the CF animals are useful for gene therapy studies because they do not possess a cAMP-activatable Cl ~ channel. They can be used as a very sensitive assay for testing the effectiveness of gene therapy delivery techniques and expression of an exogenous cftr gene. Other knock-out mice have been generated recently that may prove useful as models of different human disorders. In general, the genes disrupted in these animals are not likely to be the primary cause of the disorder in humans, but some of the symptoms observed in these animals are sufficiently similar to those in humans to provide a pseudomodel system. For example, both c-fos- and c-src-deficient mice develop osteopetrosis and may provide models for studying bone remodelling, a process accelerated after bone fractures (Soriano et al. 1991; Johnson et al. 1992; Wang et al. 1992). Mice lacking Po, a major protein of the peripheral nervous system myelin, are subject to hypomyelination and degeneration of nerve axons and may provide a model for peripheral neuropathies such as the Dejerine-Sottas type (Giese et al. 1992). TGF-fil-deficient mice develop multifocal inflammatory disease and may be a model for human immune and inflammatory disorders (Shull et al. 1992). Targeted deletion of the a-inhibin gene generates mice that develop gonadal stromal tumours and can be used to study antitumour drugs (Matzuk et al. 1992). Mice deficient for the p53 tumour-suppressor protein develop spontaneous tumours in a variety of tissues and provide an animal model for the development of neoplastic disease (Donehower et al. 1992). As the number of identified human disease genes increases, the number of animal models generated through gene targeting in ES cells will also increase. It is anticipated that in the near future mouse models for a variety of human disorders will be generated. Some of these are illustrated in Table 5.2.

Table 5.2. Human genetic disorders for which animal models may be generated by gene targeting Gene

Disorder Huntington's disease

GT24/IT15

Kennedy disease

Androgen receptor

Fragile X syndrome

FMR-1

Myotonic dystrophy

MT-PK

Neurofibromatosis type 1 Bloom's syndrome

Neurofibromin

Adrenoleukodystrophy

Peroxisomal membrane transport protein

DNA ligase I

Phenotype Autosomal; lethal, late onset; dominant neurodegeneration X-linked recessive; progressive muscular weakness of upper and lower extremities; neural degeneration X-linked recessive; incomplete penetrance; mental retardation; macro-orchidism Autosomal, dominant; myotona, cardiac arrhythmia, cataracts, male infertility Benign and malignant; neuralcrest-cell-derived tumours Autosomal recessive; increased cancer frequency X-linked; lethal; nervous system demyelination

Reference Goldberg et al. (1993) La Spada et al. (1991)

Fu et al. (1991); Verkerk et al. (1991); Kremer et al. (1991) Brook et al (1992); Fu et al. (1992) Xu et al. (1990) Chan et al. (1987); Willis and Lindahl (1987) Mosser et al. (1993)

158

Martin J. Evans et al.

5.3.5 Genomic imprinting Genomic imprinting is the phenomenon whereby allele-specific expression in a zygote is affected by the maternal or paternal origin of the allele. Several transgenic lines have been described that show parental imprinting of autosomal transgenes (Swain et al. 1987; Reik et al. 1990). Recently several endogenous mouse loci have also been shown to undergo imprinting. For example, the mouse insulin-like growth factor II gene (/g/2), which stimulates embryonic growth in utero, is subject to tissue-specific parental imprinting such that the paternal allele is expressed in embryos while the maternal allele is silent (DeChiara et al. 1991). The Igf2 receptor gene (Igf2r) shows an opposite imprinting pattern; that is, the maternal allele is expressed and the paternal allele repressed in embryos (Barlow et al. 1991). The molecular basis of genomic imprinting is still unclear, but the availability of mice in which a transgene shows imprinting has enabled researchers to study the possible mechanisms. A correlation has been observed between inactivation of a transgene and hypermethylation of CpG islands in the promoter region. Transcriptionally active transgenes are associated with hypomethylation, and methylation patterns are acquired or eliminated in a male/female-specific manner (for a review, see Sasaki et al. 1992). While methylation patterns may play a role in the imprinting mechanism for a transgene, it is unclear whether a similar mechanism is involved in the imprinting pattern of endogenous loci. The proximal promoter region of the Igf2 gene does not exhibit parent-specific methylation differences, suggesting that perhaps other epigenetic modifications are involved in imprinting of this locus (Sasaki et al. 1992). In contrast, however, the H19 gene, which produces an abundant transcript of unknown function in embryos, shows methylation only of the paternal, repressed allele (Ferguson-Smith et al. in press). Targeted knock-out in ES cells of an imprinted gene may produce an unexpected phenotype in the germ-line offspring of an ES cellderived chimera. Usually, a male germ-line chimera will transmit a disrupted gene to half of its offspring to give carrier animals that have no obvious phenotype. If the targeted gene is maternally imprinted, however, the offspring that inherit the disrupted gene will also inherit an inactive imprinted gene from the female to produce phenotypically null mutants in the Fl generation (Fig. 5.3). Thus, any phenotype associated with homozygosity is manifest one generation earlier than

Trans genic rodents

x

159

PATERNAL IMPRINTING

MATERNAL IMPRINTING

CHIMAERA

CHIMAERA

1 :

+/-

NO LETHALITY CARRIER IDENTIFIED

Fi

+/+1

:

+V-

IF LETHAL NO CARRIER IDENTIFIED

Fig. 5.3. Transmission of a disrupted imprinted gene from a male ES cellderived chimera. The wild-type gene is denoted + , the disrupted gene — and the imprinted wild-type gene (which is phenotypically —) + ' \

expected. This was observed for Igf2 knock-out mice where heterozygous progeny derived from a male chimera were growth deficient (DeChiara et al. 1991). If homozygosity is lethal, no viable offspring will be obtained that carry the targeted gene, although ES cell-derived offspring will be born (Fig. 5.3). It also follows that a paternally imprinted gene can be successfully transmitted by a male chimera even if lethal in the homozygous state (Fig. 5.3). 5.3.6 Random insertional mutagenesis The techniques described in this chapter so far are limited to the functional analysis of sequences which have previously been cloned. However, if transgenic mice are to play a major role in the field of embryology, it is important that they can be used for the mutational analysis and cloning of novel genes involved in this process. Although many mutant mice have been described in the past (Green 1989), the molecular basis of most of these mutant phenotypes remains undetermined. The cloning of the genes involved is a very difficult task, since most of the mutants have either risen spontaneously or through chemical or radiation mutagenesis studies, which often do not leave obvious lesions. In fact, the greatest success in identifying these genes has occurred when cloned genes with suitable biochemical functions have been found to map to the correct chromosomal position (reviewed by Reith & Bernstein 1991).

160

Martin J. Evans et al.

Genes which have been mutated by the integration of exogenous DNA fragments, such as transgenes or viruses, should be considerably easier to characterize. This is because the exogenous DNA fragment which has mutated the gene is of known sequence or structure and can therefore act as a 'molecular tag'. These so-called insertional mutants have been generated by the three routes to transgenesis discussed earlier, and probing the mutant genome with the exogenous sequence has facilitated the cloning of the gene involved in each case. 5.3.6.1 Insertional mutagenesis by retroviral infection of embryos Early studies by Jaenisch showed that the infection of preimplantation embryos with recombinant retroviruses could lead to the disruption of genes important in embryogenesis. One of the embryonic lethal mutants resulting from this study, Mov-13, turned out to be due to the integration of a provirus in the first intron of the a 1(1) collagen gene (Schnieke et al. 1983). The cloning of genes disrupted in this way is facilitated by the fact that retroviruses usually give clean singlecopy integrations. 5.3.6.2 Insertional mutagenesis by pronuclear injection Traditionally, the technique of pronuclear injection was developed and employed for means other than a mutagenic analysis of development. However, many interesting mutants have been generated by the disruption of developmentally important loci by transgene constructs. The most famous example is the limb-deformity phenotype seen by Leder's group (Woychik et al. 1985). The major problem with this approach is that the transgenes integrate as large tandem arrays which often induce genomic rearrangements. This makes cloning of the sequences flanking the transgene very difficult, with the result that most of the genes affected in these mutants remain uncharacterized. However, examples in which the disrupted genes have been isolated include the formins gene, responsible for the limb-deformity phenotype just described (Woychik et al. 1991), and H/358, which is important in early postimplantation development (Radice et al. 1991). Significantly, in both of these cases the transgene was integrated at a low copy number with little genomic rearrangement.

Transgenie rodents

161

5.3.6.3 Insertional mutagenesis in ES cells ES cells can be cultured in a suspension of recombinant retroviruses to derive colonies with several viral integrations (Robertson 1986; Kuehn et al. 1987). Mice generated from these ES cell colonies, using the techniques described previously, can be bred to homozygosity for different viral integrations. This strategy has resulted in the characterization of the gene nodal (Zhou et al. 1993), originally named 413.d by our group (Conlon et al. 1991), which encodes a TGF /3-like molecule that is important during gastrulation. Genes from other interesting mutants should be cloned in the near future. 5.3.7 Towards saturation mutagenesis in the mouse Large-scale mutagenic screens have played a large role in our understanding of the embryonic development of invertebrate organisms such as Caenorhabditis elegans and Drosophila melanogaster (Nusslein-Volhard & Wieschaus 1980). Many feel it would be desirable to take this type of approach in studying mammalian development (see review by Rossant & Hopkins 1992). However, because mammals have fewer offspring, have a considerably longer gestation time and are obviously much larger, in both physical and genomic terms, than either worms or flies, it would be impractical to undertake identical screens in mice. Therefore, before such a strategy can be considered, techniques have to be developed to make the study less time consuming and more cost effective. Some of the most important recent developments will now be looked at in more detail. 5.3.7.1 Gene trapping in ES cells A large-scale insertional mutagenesis approach in mice obviously requires the generation of many thousands of mutant embryos. This is complicated by the fact that only an estimated 5-10% of integration events are likely to be mutagenic (reviewed by Meisler 1992). The 'enhancer trap' strategy developed in Drosophila has been very useful in identifying and isolating developmentally interesting genes (O'Kane & Gehring 1987). These constructs use a reporter gene, usually Escherichia coli /3-galactosidase {lac Z), driven by a weak promoter, whose temporal and spatial aspects of expression can be con-

162

Martin J. Evans et al.

trolled by the enhancers of nearby genes. Cells which express lac Z within the developing embryo can be easily identified using a chromogenic substrate, X-Gal, which stains them blue. A similar approach was taken in mice by Kothary et al. (1988) and Allen et al. (1988), who generated several transgenic mouse lines with unique patterns of lac Z expression during development. However, as enhancers can stimulate transcription over large distances, characterization of the genes trapped by these constructs is difficult, and this also means that these integrations are not necessarily mutagenic. In addition, these transgenic lines were generated using the pronuclear injection technique, which can cause large rearrangements at the integration site, making cloning the integration site difficult. To overcome these problems, Gossler and colleagues (1989) developed a system which they called a 'gene trap'. Here the lac Z reporter is promoterless and therefore has to be integrated within an actively transcribed gene before it can be expressed. When this system is used in ES cells, it is possible to prescreen for interesting integrations before generating transgenic animals. This greatly reduces the number of animals required for a large-scale mutagenesis. Also, as the pattern of lac Z expression is observed in mice hemizygous for the transgene, this gives spatial and temporal information about the trapped gene without the need for costly intercrossing. Any integrations considered 'uninteresting', by whatever criteria, can be disregarded at this level. These constructs contained a dominant marker for DNA integration, neo, which confers resistance to the antibiotic G418. In practice, the prescreen might take the form of checking G418resistant ES cells for lac Z expression, before and after in vitro differentiation, or looking for interesting expression patterns in transient chimeras. 5.3.7.2 Improvements to trap vector design It was found that placing a splice acceptor sequence before the promoterless lac Z sequence increased the number of lac Z-expressing clones 12-fold. This suggests that the majority of active insertions occur within the introns of genes, producing lac Z fusion proteins (Gossler et al. 1989). When an ATG translational initiation codon is placed before a promoterless lac Z sequence, there is an increase in the number of cells

Trans genie rodents

163

expressing the protein (Brenner et al. 1989). Presumably, this allows activation of traps which integrate in 5' untranslated regions. To allow for direct selection of promoter-trap events, Friedrich and Soriano (1991) developed a promoterless marker with both lac Z and neo activities, in the form of a fusion protein, which they named j8-geo (Fig. 5.4). In order for this bifunctional fusion protein to be expressed, allowing growth in the presence of G418, the promoterless marker sequence must disrupt a transcriptionally active region. This is confirmed by the fact that 95% of G418-resistant colonies stain blue with X-Gal, and Northern blot analysis carried out on several G418-resistant colonies showed that transcription of the j8-geo sequence started from an upstream promoter in every case. In order to take advantage of the fact that retroviruses integrate cleanly without much rearrangement, Friedrich and Soriano (1991) inserted their splice-accepting markers, SA j8-gal and SA /3-geo, within the self-activating retroviral vector pGen". These were inserted in the reverse orientation to viral transcription to prevent the splicing-out of the viral packaging sequence (#) from an upstream viral splice acceptor (Fig. 5.4). These constructs were given the name ROSA (reverse orientation splice acceptor), and analysis of the integration sites showed that these constructs do indeed give integrations without genomic rearrangement. Using the strategies described here together with improved techniques in molecular biology and cell culture, it should be possible to mutate and characterize many, if not most, of the genes involved in mammalian development. With this strategy of gene trapping, together with targeted mutagenesis and overexpression studies, work on the transgenic mouse will continue to increase our understanding of the molecular basis of mammalian development. 5.4 Prospects 5.4.1 Gene targeting

Gene knock-out experiments will no doubt continue at the same incredible rate. However, it is likely that the future for gene targeting will lie in the introduction of more subtle mutations into chromosomal sequences. This will allow the generation of a series of alleles for a particular gene which will yield a greater understanding of its function. Direct functional analysis of specific gene regulatory elements and

Martin J. Evans et al.

164 SA p-geo

II SCO

Intron sequence spliced 'fromniRNA

Integration of SAfi-geo into a gene

1 rj f—'

A

(n)

E&-^

-g£0 fused to N-terminus of endogenous protein

Premature termination of endogenous protein translation

Fig. 5.4. Gene disruption and tagging by SA /3-geo vectors. When the gene trap vector integrates within the intron of an actively transcribed gene, its highly effective splice acceptor ensures that it is spliced to and is thus included in the final transcript from the endogenous locus. On translation this will produce a fusion protein which terminates at the end of the p-geo coding sequence. This integration event therefore mutates the endogenous gene and places expression of the p-gal-neo fusion (/3-geo) under the control of the endogenous promoter.

protein engineering in vivo should also be possible. Several strategies have been developed for introducing single base-pair changes into endogenous genes (Zimmer & Gruss 1989; Hasty et al. 1991a; Valancius & Smithies 1991; Davis et al 1992; Deng et al. 1993; Askew et al. 1993), and these will undoubtedly be improved. A recent paper has described the introduction of two mutations in the Hox b-4 (hox-2.6) gene (Ramirez-Solis et al. 1993). The phenotypes caused by these mutations differed in severity. It would also be desirable to swap exons or regulatory elements for those found in other genes; no doubt this will be done in the near future. Many genes of related or overlapping functions have been disrupted by gene targeting. A good example of this is the gene knock-outs carried out on several Hox genes (Chisaka & Capecchi, 1991; Lufkin et al. 1991; Chisaka et al. 1992; Le Mouellic et al. 1992; Ramirez-Solis et al. 1993), which are the mammalian homologues of the Drosophila homeotic genes. It is likely that much will be learned about the interactions between such genes by crossing newly generated mutants to

Transgenic rodents

165

make mice deficient in several related gene products. This, it is hoped, will also allow the analysis of genes that appear to be functionally 'redundant' when on a wild-type background. For example, the extracellular glycoprotein tenascin is thought to play an important role in development (Chiquet-Ehrismann et al. 1986). However, when tenascin gene expression is completely disrupted by gene targeting, the mutant mice appear to develop normally (Saga et al. 1992). Crossing this mutant to mice with targeted mutations in functionally related genes could offer insights into its in vivo function. Most gene knock-out animals are initially obtained on an outbred genetic background. It is also important to derive the mutation on a completely inbred (usually 129-strain) background to overcome any genetic heterogeneity effects of modifier genes. For example, Hox b4-null mice show homeotic transformation of a cervical vertebra and defects in the closure of the sternal rudiments with variable expressivity and incomplete penetrance on an outbred background, but the sternum defect is completely penetrant in the inbred background (Ramirez-Solis et al. 1993). It will be interesting to ascertain if the phenotypes of knock-out animals when maintained on an outbred background are different on an inbred background. At its most extreme it is possible that a null mutation may have no phenotype on a heterogeneous genetic background but may elicit a phenotype on an inbred background. 5.4.2 Inducible gene expression in mice 5.4.2.1 Early studies Ideally, the expression of manipulated genes within transgenic mice is completely under the researcher's control. This is especially important in the study of genes, such as oncogenes, whose expression can have a severe effect on the viability of the animal. Some of the earliest transgenics studies used constructs whose expression could be regulated to various degrees (Brinster et al. 1981; Palmiter et al. 1982a; Le Meur et al. 1985; Leder et al. 1986). For example, when Palmiter and Brinster and colleagues made the so-called Supermouse (Palmiter et al. 1982b), they fused both rat and human growth hormone genes to the mouse metallothionein (MT-1) gene promoter. Studies in cultured cells and transgenic mice had previously shown that transcription from this promoter could be increased by the addition of glucocorticoids and/or heavy metals (Brinster et al. 1981). Indeed, when mice carrying

166

Martin J. Evans et al.

the MT-l-rGH fusion genes were given ZnSO4 in their diet their body weight increased dramatically. However, it appears there is considerable basal level expression from these fusion constructs, and to date no satisfactory way of tightly regulating gene expression in mice has been developed. It is likely that the future will see the use of prokaryotic inducible systems, such as the lac operon, to control gene expression tightly in transgenic mice. 5.4.2.2 Binary systems Recently a binary system for regulating gene expression in mice has been described (Ornitz et al. 1991). This system depends on the capacity of the yeast transcription factor GAL4 to regulate the expression of genes fused to its recognition sequence, which is known as an upstream activation sequence (UAS). When transgenic mice expressing Gal4 ('transactivator mice') are crossed to mice carrying the int-2 gene fused to the UAS sequence ('target mice'), there is ectopic expression of the int-2 in the bigenic progeny which is not seen in either monogenic parent. The tissue specificity of int-2 expression in the bigenic mice appears to be identical to the Gal4 expression pattern in the monogenic parent. Therefore, it seems as though GAL4 protein is driving expression of the UAS-int-2 construct in specific tissues. One transactivator strain can be crossed to several UAS-target mice and thus allow ectopic expression of many different genes by simple matings. By placing the Gal4 in an enhancer-trap construct, Brand and Perrimon (1993) have generated lines of Drosophila which express GAL4 in a diverse range of tissues. By crossing these transactivator flies to UAS-reporter lines, they have been able to express ectopically a range of genes thought to be important in determining cell fate. It is possible that the GAL4 system could be used for similar studies in the mouse. 5.4.2.3 'Flipping' switches by site-specific recombination Site-specific recombination systems from bacteriophage and yeast have been used to manipulate the genomes of higher organisms with remarkable efficiency and specificity (for a recent review see Kilby et al. in press). The most widely used enzymes are the Cre and FLP recombinases from bacteriophage PI and Saccharomyces cerevisiae, respectively, which function by catalysing recombination at distinct

Transgenic rodents

167

34-bp recognition sequences, known as loxP in the case of Cre and FRT in the case of FLP. When these recognition sequences are found on two otherwise heterologous DNA molecules, there is highly efficient recombination in the presence of the suitable site-specific recombinase. Depending on the positioning and orientation of the recognition sequences, these recombinases can be used to engineer any desired genomic alteration such as inversion, deletion or insertion. For example, when two directly repeated loxP/FRT are positioned flanking a DNA sequence of interest, addition of the recombinase results in the deletion of the intervening sequence, leaving one loxP/ FRT site behind. Conversely, when an incoming DNA molecule and its target sequence each contain a loxP/FRT site, the recombinase inserts the incoming DNA with high efficiency and is now flanked by the recognition sites. This reversible insertion-deletion reaction is shown in Fig. 5.5. An important point is that even when the recognition sequences have been positioned in the genome, the site-specific recombination will not occur until the recombinase is expressed. This means that the timing and tissue specificity of the reaction can be controlled by placing recombinase expression under the control of the appropriate promoter. This has been shown by two groups who have used the Cre/ loxP system to give tissue-specific genome manipulation. Orban et al. (1992) placed Cre recombinase under the thymocyte-specific Ick promoter and crossed mice expressing Cre in their lymphocytes to mice carrying a lac Z expression cassette flanked by loxP sites. Unfortunately, the lac Z construct in these mice was not expressed at levels high enough to allow protein detection; however, the recombination reaction could be followed by probing the thymocyte DNA of doubly transgenic offspring for the lac Z construct. This revealed that deletion of greater than 95% of the lac Z construct had occurred in thymocytes of doubly transgenic mice. This estimated deletion figure is even more impressive when one considers the fact that the lac Z construct was originally present as a 'head-to-tail' transgene array estimated to be around 70 kb. Therefore, it appears that not only is Cre protein catalysing deletion of most of this large construct, it is doing it in the vast majority of expressing cells. Lakso et al. (1992) placed a disruptive insert, which was flanked by loxP sites, between the lens-specific aAcrystallin promoter and the coding sequence of the SV40 large Tantigen. This construct is not active, as trangenic mice carrying this sequence are free of lens tumours. However, when these mice are

168

Martin J. Evans et al. Deletion / Co-integration loxP

maA

disruption

LOXPSN/40 TAg

It

insert

Fig. 5.5. Cre recombinase-mediated deletion-integration. When loxP sites are positioned in direct repeat flanking a region of DNA, the addition of Cre results in the deletion of the intervening region. This reaction results in a single loxP site being left in the genome, the other being present on the excised fragment. Note that the insertion of exogenous DNA is this reaction in reverse (maA is the mouse a-crystallin promoter). (From Lakso et al. 1992.)

crossed to mice expressing Cre enzyme, the disruptive insert is efficiently removed, activating the construct, with the result that every doubly transgenic mouse develops lens tumours. However, it is difficult to estimate the efficiency of the deletion in this case as tumours can arise from single transformed cells. As discussed previously in this chapter, one of the problems with introducing subtle mutations into ES cells is that there is the requirement for a positive marker to select for cells which have stably integrated the construct. Although this marker can then be removed by either intrachromosomal recombination, as with the 'hit-and-run' strategy (Hasty et al. 1991a), or by a second round of homologous recombination, as with the 'tag and exchange' strategy, these routes are still relatively inefficient. However, a recent paper describes the

Trans genie rodents

169

use of the Cre/loxP system to overcome this problem (Gu et al. 1993). The first round of homologous recombination introduces the marker gene flanked by loxP sites into the target gene. This region can then be deleted by transient expression of Cre from an episomal plasmid, leaving behind a single loxP site. Gu et al. (1993) observed deletion of the JH segments and intron enhancer region of the IgH locus in 4080% of the cells expressing Cre. Cre will also mediate site-specific recombination when introduced into cells by lipofection as a purified protein (Baubonis & Sauer 1993). It is highly likely that there will be many further examples of the use of site-specific recombinases in transgenic mice to answer biological problems. In the field of developmental biology, these systems will make closer examination of gene function possible by allowing controlled tissue-specific gene mutation. By coupling these recombinase systems to improved inducible gene expression systems, it should be possible to ablate gene function in adult mice. 5.4.3 Lineage tracing studies It is anticipated that transgenic mice will be increasingly used to study the establishment of cell lineages and the interactions between different lineages in developmental processes. These analyses will be helped by specific cell lineage ablation in transgenic mice using tissuespecific expression of a toxin gene - for example, diphtheria toxin A (Palmiter et al. 1987). An improvement on this approach would be to express thymidine kinase in a tissue-specific fashion and kill these cells at a specific developmental time by administration of gancyclovir or FIAU. Gene disruption in ES cells using a lac Z gene provides a lineage marker with which to monitor the expression pattern of the targeted gene and also to study the fate of the cells that no longer express the gene (Le Mouellic et al. 1990, 1992). Thus, the consequences of gene inactivation can be more clearly understood by allowing the appearance of a phenotype to be followed at the cellular level. More gene targeting strategies will probably utilize this approach in the future. Acknowledgements We thank Mark Carlton and Jim Murray for help with the figures and Linda Millett for expert secretarial assistance.

170

Martin J. Evans et al.

References Allen, N. D., Cran, D. G., Barton, S. C , Hettle, S., Reik, W. & Surani, M. A. H. (1988). Transgenes as probes for active chromosomal domains in mouse development. Nature 333, 852-5. Allison, J., Campbell, I. L., Morahan, G., Mandel, T. E., Harrison, L. C. & Miller, J. F. (1988). Diabetes in transgenic mice resulting from overexpression of class I histocompatibility molecules in pancreatic beta cells. Nature 333, 529-33. Askew, G. R., Doetschman, T. & Lingerl, J. B. (1993). Site-directed mutations in embryonic stem cells: A gene-targeting tag-and-exchange strategy. Mol. Cell. Biol. 13, 4115-24. Barlow, D. P., Stoger, R., Herrmann, B. G., Saito, K. & Schweifer, N. (1991). The mouse insulin-like growth factor type-2 receptor is imprinted and closely linked to the Tme locus. Nature 349, 84-7. Baubonis, W. & Sauer, B. (1993). Genomic targeting with purified ere recombinase. Nucl Acids Res. 9, 2025-9. Behringer, R. R., Ryan, T. M., Reilly, M. P., Asakura, T., Palmiter, R. D., Brinster, R. L. & Townes, T. M. (1989). Synthesis of functional human haemoglobin in transgenic mice. Science 247, 566-8. Bishop, J. M. (1987). The molecular genetics of cancer. Science 235, 305-11. Borelli, E., Heyman, R., Hsi, M. & Evans, R. M. (1988). Targeting of an inducible toxic phenotype in animal cells. Proc. Natl. Acad. Sci. USA, 85, 7572-6. Boulter, C. A., Aguzzi, A., Williams, R. L., Wagner, E. F., Evans, M. J. & Beddington, R. (1991). Expression of v-src induces aberrant development and twinning in chimaeric mice. Development 111, 357-66. Bradley, A., Evans, M., Kaufman, M. H. & Robertson, E. (1984). Formation of germline chimaeras from embryo-derived teratocarcinoma cell lines. Nature 309, 255-6. Brand, A. H. & Perrimon, N. (1993). Targeted gene expression as a means of altering cell fates and generating dominant phenotypes. Development 118, 401-15. Brenner, D. G., Lin-Chao, S. & Cohen, S. N. (1989). Analysis of mammalian cell genetic regulation in situ by using retrovirus-derived 'portable exons' carrying Escherichia coli lac Z gene. Proc. Natl. Acad. Sci. USA 86, 5517-21. Brinster, R. L., Chen, H. Y., Trumbauer, M., Senear, A. W., Warren, R. & Palmiter, R. D. (1981). Somatic expression of herpes thymidine kinase in mice following injection of fusion gene into eggs. Cell 27, 223-31. Brinster, R. L., Chen, H. Y., Trumbauer, M. E., Yagle, M. K. & Palmiter, R. D. (1985). Factors affecting the efficiency of introducing foreign DNA into mice by microinfecting eggs. Proc. Natl. Acad. Sci. USA 82, 443842. Brook, J. D., et al. (1992). Molecular basis of myotonic dystrophy: Expansion of a trinucleotide (CTG) repeat at the 3' end of a transcript encoding a protein kinase family member. Cell 69(2), 385. Chan, J. Y. H., Becker, F. F., German, J. & Ray, J. H. (1987). Altered DNA ligase I activity in Bloom's syndrome cells. Nature 325, 357-9. Chang, X. B. & Wilson, J. H. (1987). Modification of DNA ends can decrease

Transgenie rodents

171

end joining relative to homologous recombination in mammalian cells. Proc. Natl. Acad. Sci. USA 84, 4959-63. Chartier-Harlin, M. C , et al. (1991). Early onset Alzheimer's disease caused by mutations at codon 717 of the beta-amyloid precursor protein gene. Nature 353, 844-6. Chiquet-Ehrismann, R., Mackie, E. J., Pearson, C. A. & SakakuraT. (1986). Tenascin: An extracellular matrix protein involved in tissue interactions during fetal development and oncogenesis. Cell 47, 131-9. Chisaka, O. & Capecchi, M. (1991). Regionally restricted developmental defects resulting from targeted disruption of the mouse homeobox gene Hox-1.5. Nature 350, 473-9. Chisaka, O., Muski, T. & Capecchi, M. (1992). Developmental defects of the ear, cranial nerves and hindbrain resulting from targeted disruption of the mouse homeobox gene Hox-1.6. Nature 355, 516-20. Clarke, L. L., Grubb, B. R., Gabriel, S. E., Smithies, O., Koller, B. H. & Bucher, R. C. (1992). Defective epithelial chloride transport in a gene targeted mouse model of cystic fibrosis. Science 257, 1125-8. Colledge, W. H., Ratcliff, R., Foster, D., Williamson, R. & Evans, M. J. (1992). Cystic fibrosis mouse with intestinal obstruction. Lancet 340, 680. Conlon, F. L., Barth, K. S. & Robertson, E. J. (1991). A novel retrovirally induced embryonic lethal mutation in the mouse: Assessment of the developmental fate of embryonic stem cells homozygous for the 413.d proviral integration. Development 111, 969-81. Constantini, F. & Lacy, E. (1981). Introduction of a rabbit j8-globin gene into the mouse germ line. Nature 294, 92-4. Davis, A. C , Wims, M. & Bradley, A. (1992). Investigation of coelectroporation as a method for introducing small mutations into embryonic stem cells. Mol. Cell. Biol. 12, 2769-76. DeChiara, T. M., Robertson, E. J. & Efstratiadis, A. (1991). Parental imprinting of the mouse insulin-like growth factor II gene. Cell 64, 849-59. Deng, C. & Capecchi, M. R. (1992). Re-examination of gene targeting frequency as a function of the extent of homology between the targeting vector and the target locus. Mol. Cell Biol. 12(8), 3365-71. Deng, C. X., Thomas, K. R. & Capecchi, M. (1993). Location of cross-overs during gene targeting with insertion and replacement vectors. Mol. Cell Biol. 13, 2134-40. Doetschman, T., Williams, P. & Maeda, N. (1988). Establishment of hamster blastocyst-derived embryonic stem (ES) cells. Dev. Biol. 127, 22-227. Donehower, L. A., Harvey, M., Slagle, B. L., McArthur, M. J., Montgomery, C. A., Jr., Butel, J. S. & Bradley, A. (1992). Mice deficient for p53 are developmentally normal but susceptible to spontaneous tumours. Nature 356, 215-21. Dorin, J. R. et al. (1992). Cystic fibrosis in the mouse by targeted insertional mutagenesis. Nature 359, 211-15. Epstein, P. N., Overbeek, P. A. & Means, A. R. (1989). Calmodulin-induced early onset diabetes in transgenic mice. Cell 58, 1067-73. Evans, M. J. & Kaufman, M. H. (1981). Establishment in culture of pluripotential cells from mouse embryos. Nature 292, 154-6. Ferguson-Smith, A. C , Hiroyuki, S., Cattanach, B. M. & Surani, M. A. (in press). Parental-origin-specific epigenetic modification of the mouse H19 gene. Nature.

172

Martin J. Evans et al.

Friedrich, G. & Soriano, P. (1991). Promoter traps in embryonic stem cells: A genetic screen to identify and mutate developmental genes in mice. Genes Dew. 5, 1513-23. Fu, Y. H. et al. (1991). Variation of the CGG repeat at the fragile X site results in genetic instability: Resolution of the Sherman paradox. Cell 67(6), 1047-58. Fu, Y. H. et al. (1992). An unstable triplet repeat in a gene related to myotonic muscular dystrophy. Science 255(5049), 1256-8. Ganten, D. et al. (1992). Species specificity of renin kinetics in transgenic rats harboring the human renin and angiotensinogen genes. Proc. Natl. Acad. Sci. USA 89, 7806-10. Giese, K. P., Martini, R., Lemke, G., Soriano, P. & Schachner, M. (1992). Mouse Po gene disruption leads to hypomyelination, abnormal expression of recognition molecules and degeneration of myelin and axons. Cell 71, 565-76. Goate, A. et al. (1991). Segregation of a missense mutation in the amyloid precursor protein gene with familial Alzheimer's disease. Nature 349, 704-6. Goldberg, Y. P., Rommens, J. M., Andrew, S. E., Hutchinson, G. B., Lin, B., Theilmann, J., Graham, R., Glaves, M. L., Starr E., McDonald, H., Nasir, J., Schappert, K., Kalchman, M. A., Clarke, L. A. & Hayden, M. R. (1993). Identification of an Alu retrotransposition event in close proximity to a strong candidate gene for Huntington's disease. Nature 362, 370-3. Gordon, J. W., Scangos, G. A., Plotkin, D. J., Barbosa, J. A. & Ruddle, F. H. (1980). Genetic transformation of mouse embryos by microinjection of purified DNA. Proc. Natl. Acad. Sci. USA 77, 380-4. Gossler, A., Joyner, A. L., Rossant, J. & Skarnes, W. C. (1989). Mouse embryonic stem cells and reporter constructs to detect developmentally regulated genes. Science 244, 463-5. Greaves, D. R., Fraser, P., Vidal, M. A., Hedges, M. J., Ropers, D., Luzzatto, L. & Grosveld, F. (1990). A transgenic mouse model of sickle cell disorder. Nature 343, 183-5. Green M. C. (1989). Catalog of mutant genes and polymorphic loci. In Genetic Variants and Strains of the Laboratory Mouse (2d ed.), ed. M. F. Lyon & A. G. Searle, pp. 12-403. Grosveld, F., Van Assendelft, G. B., Greaves, D. R. & Kollia, S. G. (1987). Position independent, high level expression of the human /3-globin gene in transgenic mice. Cell 51, 975-85. Gu, H., Zou, Y. & Rajewsky, K. (1993). Independent control of immunoglobulin switch recombination at individual switch regions evidenced through cre-lox-P mediated gene targeting. Cell 73, 1155-64. Halter, S. A., Dempsey, P., Matsui, Y., Stokes, M. K., Graves-Deal, R., Hogan, B. L. & Coffey, R. J. (1992). Distinctive patterns of hyperplasia in transgenic mice with mouse mammary tumor virus transforming growth factor-alpha. Characterization of mammary gland and skin proliferations. Am. J. Pathol. 140, 1131-46. Hasty, P., Ramirez-Solis, R., Krumlauf, R. & Bradley, A. (1991a). Introduction of a subtle mutation in the Hox-2.6 locus in embryonic stem cells. Nature 350, 243-6. Hasty, P., Rivera-Perez, J., Chang, C. & Bradley, A. (1991b). Target frequency and integration pattern for insertion and replacement vectors in embryonic stem cells. Mol. Cell. Biol. 11(9), 4509-17.

Transgenic rodents

173

Hilton, D. J. & Gough, N. M. (1991). Leukaemia inhibitory factor: A biological perspective. / . Cell. Biochem. 46, 21-6. Homanics, G. E., Smith, T. J., Zhang, S. H., Lee, D., Young, S. G. & Maeda, N. (1993). Targeted modification of the apolipoprotein B gene results in hypobetalipoproteinaemia and developmental abnormalities in mice. Proc. Natl. Acad. Sci. USA 90, 2389-93. Hooper, M. L. (1987). Isolation of genetic variants and fusion hybrids from embryonal carcinoma cell lines. In Teratocarcinomas and Embryonic Stem Cells: A Practical Approach, pp. 51-70. Washington, DC: IRL Press. Hooper, M. L., Hardy, K., Handyside, A., Hunter, S. & Monk, M. (1987). HPRT-deficient (Lesch-Nyhan) mouse embryos derived from germline colonization by cultured cells. Nature 326, 292-5. Jaenisch, R. (1976). Germline integration and Mendelian transmission of the exogenous Moloney leukemia virus. Proc. Natl. Acad. Sci. USA 73, 1260-4. Jahner, D., Stuhlmann, H., Stewart, C. L., Harbers, K., Lohler, J., Simon, L. & Jaenish, R. De novo methylation and expression of retroviral genomes during mouse embryogenesis. Nature 298, 623-6. Jakobovits, A., Moore, A. L., Green, L. L., Vergara, G. J., Maynard-Currie, C. E., Austrin, H. A. & Klapholz, S. (1993). Germ-line transmission and expression of a human-derived yeast artificial chromosome. Nature 362, 255-8. Johnson, R. S., Spiegelman, B. M. & Papaioannou, V. (1992). Pleiotropic effects of a null mutation in the c-fos proto-oncogene. Cell 71, 577-86. Khillan, J. S., Oskarsson, M. K., Propst, F., Kuwabara, T., Vande-Woude, G. F. & Westphal, H. (1987). Defects in lens fiber differentiation are linked to c-mos overexpression in transgenic mice. Genes Dev. 1, 1327— 35. Kilby, N. J., Snaith, M. R. & Murray, J. A. H. (in press). Site specific recombinases: Tools for genome engineering. Trends Genet. Kothary, R., Clapoff, S., Brown, A., Campbell, R., Peterson, A. & Rossant, J. (1988). A transgene containing lac Z inserted into the dystonia locus is expressed in the neural tube. Nature 335,435-7. Kremer, E. J., Pritchard, M., Lynch, M., Yu, S., Holman, K., Baker, E., Warren, S. T., Schlessinger, D., Sutherland, G. R. & Richards, R. I. (1991). Mapping of DNA instability at the fragile X to a trinucleotide repeat sequence p(CCG)n. Science 252(5013), 1711-14. Kuehn, M. R., Bradley, A., Robertson, E. J. & Evans, M. J. (1987). A potential animal model for Lesch-Nyhan syndrome through introduction of HPRT mutations into mice. Nature 326, 295-8. La Spada, A. R., Wilson, E. M., Lubahn, D. B., Harding, A. E. & Fischbeck, K. H. (1991). Androgen receptor gene mutations in X-linked spinal and bulbar muscular atrophy. Nature 352, 77-9. Lakso, M., Sauer, B., Mosinger, B., Lee, E. J., Manning, R.W., Yu, S. H., Mulder, K. L. & Westphal, H. (1993). Targeted oncogene activation by site-specific recombination in transgenic mice. Proc. Natl. Acad. Sci. USA 89, 6232-6. Le Meur, M., Gerlinger, P., Benoist, C. & Mathis, D. (1985). Correcting an immune-response deficiency by creating Ea gene transgenic mice. Nature 316, 38-42. Le Mouellic, H., Lallemand, Y. & Brulet, P. (1990). Targeted replacement of

174

Martin J. Evans et al.

the homeobox gene HOX-3.1 by the Escherichia coli lac Z in mouse chimeric embyros. Proc. Natl. Acad. Sci. USA 87, 4712-16. Le Mouellic, H., Lallemand, Y. & Brulet, P. (1992). Homeosis in the mouse induced by a null mutation in the Hox-3.1 gene. Cell 69, 251-64. Leder, A., Pattengate, P. K., Kuo, A., Stewart, T. A. & Leder, P. (1986). Consequences of widespread deregulation of the c-myc gene in transgenic mice: Multiple neoplasms and normal development. Cell 45, 485-95. Luetteke, N. C , Qui, T. H., Peiffer, R. L., Oliver, P., Smithies, O. & Lee, D. C. (1993). TGF alpha deficiency results in hair follicle and eye abnormalities in targeted and waved-1 mice. Cell 73, 262-78. Lufkin, T. A., Dierich, M., LeMeur, M. & Chambon, P. (1991). Disruption of the Hox 1.6 homeobox gene results in defects in a region corresponding to its rostral domain of expression. Cell 66, 1105-19. Mann, G. G., Fowler, K. J., Gabriel, A., Nice, E. C , Williams, R. L. & Dunn, A. R. (1993). Mice with a null mutation of the TGF alpha gene have abnormal skin architecture, wavy hair, and curly whiskers and often develop corneal inflammation. Cell73, 249-61. Mansour, S. L., Thomas, K. R. & Capecchi, R. (1988). Disruption of the protooncogene int-2 in mouse embryo-derived stem cells: A general strategy for targeting mutations to non-selectable genes. Nature 336, 348-52. Martin, G. (1981). Isolation of a pluripotent cell line from early mouse embryos cultured in medium conditioned by teratocarcinoma stem cells. Proc. Natl. Acad. Sci. USA 78, 7634-8. Matzuk, M. M., Finegold, M. J., Su, J. J., Hsueh, A. J. W. & Bradley, A. (1992). A-Inhibin is a tumour-suppressor gene with gonadal specificity in mice. Nature 360, 313-19. Meisler, M. H. (1992). Insertional mutation of 'classical' and novel genes in transgenic mice. Trends Genet. 8, 341-4. Miller, C. J., McPheat, J. C. & Potts, W. J. (1992). Targeted integration of the REN-ID locus in mouse embryonic stem cells. Proc. Natl. Acad. Sci. USA 89(5020), 5024. Moens, C. B., Auerbach, A. B., Conlon, R. A., Joyner, A. L. & Rossant, J. (1992). A targeted mutation reveals a role of N-myc in branching morphogenesis in the embryonic mouse lung. Genes Dev. 6, 691-704. Mosser, J., Douar, A. M., Sarde, C. O., Kioschis, P., Feil, R., Moser, H., Poustka, A. M., Mandel, J. L. & Aubourg, P. (1993). Putative X-linked adrenoleukodystrophy gene shares unexpected homology with ABC transporters. Nature 361, 726-30. Mulligan, R. C. & Berg, P. (1981). Selection for animal cells that express the Escherichia coli gene coding for xanthine-guanine phosphoribosyltransferase. Proc. Natl. Acad. Sci. USA 78, 2027-6. Nusslein-Volhard, C. & Wieschaus, E. (1980). Mutations affecting segment number and polarity. Nature 287, 795-801. O'Kane, C. & Gehring, W. J. (1987). Detection in situ of genomic regulatory elements in Drosophila. Proc. Natl. Acad. Sci. USA 84, 9123-7. Orban, P. C , Chui, D. & Marth, J. D. (1992). Tissue and site specific DNA recombination in transgenic mice. Proc. Natl. Acad. Sci. USA 89, 8681-5. Ornitz, D. M., Moreadith, R. W. & P. Leder. (1991). Binary system for regulating transgene expression in mice: Targeting int-2 gene expression with yeast GAL4/UAS control elements. Proc. Natl. Acad. Sci. USA 88, 698702.

Transgenic rodents

175

Palmiter, R. D., Chen, H. Y. & Brinster, R. L. (1982a). Differential regulation of metallothionein-thymidine kinase fusion genes in transgenic mice and their offspring. Cell 29, 701-10. Palmiter, R. D., Brinster, R. L., Hammer, R. E., Trumbauer, M. E., Rosenfeld, M. G., Birnberg, N. C. & Evans, R. M. (1982b). Dramatic growth of mice that develop from eggs microinjected with metallothioneingrowth hormone fusion genes. Nature 300, 611-15. Palmiter, R. D., Behringer, R. R., Quaife, C. J., Maxwell, F., Maxwell, I. H. & Brinster, R. L. (1987). Cell lineage ablation in transgenic mice by cellspecific expression of a toxin gene. Cell 50, 435—43. Palmiter, R. D., Sandgren, E. P., Avarbock, M. R., Allen, D. D. & Brinster, R. L. (1991). Heterologous introns can enhance expression of transgenes in mice. Proc. Natl. Acad. Sci. USA 88, 478-82. Plump, A. S., Smith, J. D., Hayex, T., Aalto-Setala, K., Walsh, A., Verstuyft, J. G., Propst, F., Cork, L. C , Kovatch, R. M., Kasenally, A. B., Wallace, R. & Rosenberg, M. P. (1992). Progressive hind limb paralysis in mice carrying a v-mos transgene. / . Neuropath. Exp. Neurol. 51, 499505. Propst, F., Cork, L. C, Kovatch, R. M., Kasenally, A. B., Wallace, R. & Rosenberg, M. P. (1992). Progressive hind limb paralysis in mice carrying a v-Mos transgene. J. Neuropathol. Exp. Neurol. 51(5), 499-505. Quaife, C. J., Pinkert, C. A., Ornitz, D. M., Palmiter, R. D. & Brinster, R. L. (1987). Pancreatic neoplasia induced by ras expression in arcinar cells of transgenic mice. Cell 48, 1023. Quon, D., Wang, Y., Catalano, R., Scardina, J. M., Murakami, K. & Cordell, B. (1991). Formation of beta-amyloid protein deposits in brains of transgenic mice. Nature 352, 239-41. Radice, G., Lee, J. J. & Constantini, F. (1991). H/358, an insertional mutation affecting early postimplantation development of the mouse embryo. Development 111, 801-11. Ramirez-Solis, R., Zheng, H., Whiting, J., Krumlauf, R. & Bradley, A. (1993). Hox B-4 (Hox-2.6) mutant mice show homeotic transformation of a cervical vertebra and defects in the closure of the sternal rudiments. Cell 73, 279-84. Ratcliff, R., Evans, M. J., Doran, J., Wainwright, B. J., Williamson, R. & Colledge, W. H. (1992). Disruption of the cystic fibrosis transmembrane conductance regulator gene in embryonic stem cells by gene targeting. Transgen. Res. 1, 177-81. Ratcliff, R., Evans, M. J., Cuthbert, A. W., MacVinish, L. J., Foster, D., Anderson, J. R. & Colledge, W. H. (1993). Production of a severe cystic fibrosis mutation in mice by gene targeting. Nature Genet. 4, 35-41. Readhead, C , Popko, B., Takahashi, N., Shine, H. D., Saavedra, R. A., Sidman, R. L. & Hod, L. (1987). Expression of a myelin basic protein gene in transgenic shiverer mice: Correction of the dysmyelinating phenotype. Cell 48, 703-12. Reik, W., Howlett, S. K. & Surani, M. A. (1990). Imprinting by DNA methylation: From transgene to endogenous gene sequences. Dev. Suppl. 99, 106. Reith, A. D. & Bernstein, A. (1991). Molecular basis of mouse developmental mutants. Genes Dev. 5, 1115-23. Robertson, E. J. (1986). Pluripotential stem cell lines as a route into the mouse germ line. Trends Genet. (Jan.), 9-13.

176

Martin J. Evans et al.

Rosenberg, M. P. (1992). Progressive hind limb paralysis in mice carrying avMos transgene. J. Neuropathol. Exp. Neurol. 51, 499-505. Rossant, J. & Hopkins, N. (1992). Of fin and fur: Mutational analysis of vertebrate embryonic development. Genes Dev. 6, 1-13. Rubin, E. M. & Breslow, J. L. (1992). Severe hypercholesterolemia and athersclerosis in apolipoprotein E-deficient mice created by homologous recombination in ES cells. Cell71, 343-53. Ruther, U., Garber, C , Komiroswki, D., Muller, W. & Wagner, E. F. (1987). Deregulated c-fos expression interferes with normal bone development in transgenic mice. Nature 325, 412-16. Ruther, U., Muller, W., Sumida, T., Tokuhisa, T., Rajewsky, K. & Wagner, E. F. (1988). c-fos expression interferes with thymus development in transgenic mice. Cell 53, 847-56. Ryan, T. M., Townes, R. M., Reilly, M. P., Asakura, T., Palmiter, R. D., Brinster, R. L. & Behringer, R. R. (1990). Human sickle hemoglobin in transgenic mice. Science 2A7, 566-8. Saga, Y., Yagi, T., Ikawa, Y., Sakakura, T. & Aizawa, S. (1992). Mice develop normally without tenascin. Genes Dev. 6(10), 1821-31. Sandgren, E. P., Quaife, C. J., Pinkert, C. A., Palmiter, R. D. & Brinster, R. L. (1989). Oncogene-induced liver neoplasia in transgenic mice. Oncogene 4, 715-24. Sasaki, H., Jones, P. A., Chaillet, J. R., Ferguson-Smith, A. C , Barton, S. C , Reik, W. & Surani, M. A. (1992). Parental imprinting: Potentially active chromatin of the repressed maternal allele of the mouse insulin-like growth factor II (Igf2) gene. Genes Dev. 6, 1843-56. Schedl, A., Montoliu, L., Kelsey, G. & Schutz, G. (1993). A yeast artificial chromosome covering the tyrosinase gene confers copy numberdependent expression in transgenic mice. Nature 362, 258-61. Schnieke, A., Harbers, K. & Jaenisch, R. (1983). Embryonic lethal mutation in mice induced by retro virus insertion inton the a 1(1) collagen gene. Nature 304, 315-20. Shull, M. M. et al. (1992). Targeted disruption of the mouse transforming growth factor-Bl gene results in multifocal inflammatory disease. Nature 359, 693-9. Shulman, M. J., Nissen, L. & Collins, C. (1990). Homologous recombination in hybridoma cells: Dependence on time and fragment length. Mol. Cell Biol. 10, 4466-72. Sinn, E., Muller, W., Pattengale, P., Tepler, I. W., Wallace, R. & Leder, P. (1987). Co-expression of MMTV/v-Ha ras and MMTV/C-myc genes in transgenic mice: Synergistic action of oncogenes in vivo. Cell 49, 465-75. Smithies, O., Greeg, R. G., Boggs, S. S., Koralewski, M. A. & Kucherlapati, R. S. (1985). Insertion of DNA sequences into the human chromosomal beta-globin locus by homologous recombination. Nature 317, 230-4. Snouwaert, J.N., Brigman, K. K., Latour, A. M., Malouf, N. N., Boucher, R. C , Smithies, O. & Killer, B. H. (1992). An animal model for cystic fibrosis made by gene targeting. Science 257, 1083-8. Soriano, P., Cone, R. D., Mulligan, R. C. & Jaenisch, R. (1986). Expression of retroviral vectors in transgenic mice obtained by embryo infection. EMBO J. 6, 383-8. Soriano, P., Montgomery, C , Geske, R. & Bradley, A. (1991). Targeted disruption of the c-src proto-oncogene leads to osteopetrosis in mice. Cell 64, 693-702.

Trans genic rodents

177

Southern, P. J. & Berg, P. (1982). Transformation of mammalian cells to antibiotic resistance with a bacterial gene under control of the S V40 early region promotor. / . Mol. Appl. Genet. 1, 327-41. Stewart, C. L., Schuetze, S., Vanek, M. & Wagner, E. F. (1987). Expression of retroviral vectors in transgenic mice obtained by embryo infection. EMBO J. 6, 383-8. Strauss, W. M., Dausman, J., Beard, C , Johnson, C , Lawrence, J. B. & Jaenisch, R. (1993). Germ line transmission of a yeast artificial chromosome spanning the murine a-1(1) collagen locus. Science 259, 1904-7. Swain, J. L., Stewart, T. A. & Leder, P. (1987). Parental legacy determines methylation and expression of an autosomal transgene: A molecular mechanism for parental imprinting. Cell 50, 719-27'. Swanson, M. E., Hughes, T. E., Denny, I. S., France, D. S., Paterniti, J. R., Jr., Tapparelli, C , Gfeller, P. & Burki, K. (1992). High level expression of human apolipoprotein A-I in transgenic rats raises total serum high density lipoprotein cholesterol and lowers rat apolipoprotein A-I. Trans. Res. 1(3), 142-7. Takagi, H., Jhappan, C , Sharp, R. & Merlino, G. (1992). Hypertrophic gastropathy resembling Merrtrier's disease in transgenic mice overexpressing transforming growth factor alpha in the stomach. /. Clin. Invest. 90, 1161-7. Taurog, J. D., Maika, S. D., Simmons, W. A., Breban, M. & Hammer, R. E. (1993). Susceptibility to inflammatory disease in HLA-B27 transgenic rat lines correlates with the level of B27 expression. /. Immunol. 150(9), 4168-78. te Riele, H., Maandag, E. R., Clarke, A., Hooper, M. & Berns, A. (1990). Consecutive inactivation of both alleles of the pim-1 proto-oncogene by homologous recombination in embryonic stem cells. Nature 348, 649-51. te Riele, H., Maandag, E. R. & Berns, A. (1992). Highly efficient gene targeting in embryonic stem cells through homologous recombination with isogenic DNA constructs. Proc. Natl. Acad. Sci. USA 89, 5128-32. Thomas, K. R. & Capecchi, M. R. (1987). Site-directed mutagenesis by gene targeting in mouse embryo-derived stem cells. Cell 51, 503-12. Thompson, S., Clarke, A. R., Pow, A. M., Hooper, M. L. & Melton, D. W. (1989). Germ line transmission and expression of a corrected HPRT gene produced by gene targeting in embryonic stem cells. Cell 56, 313-21. Tybulewicz, V. L. J. et al. (1992). Animal model of Gaucher's disease from targeted disruption of the mouse glucocerebrosidase gene. Nature 357, 407-9. Valancius, V. & Smithies, O. (1991). Testing an 'in-out' targeting procedure for making subtle genomic modifications in mouse embryonic stem cells Mol. Cell Biol. 11, 1402-8. Verkerk, A. J. et al. (1991). Identification of a gene (FMR-1) containing a CGG repeat coincident with a breakpoint cluster region exhibiting length variation in fragile X syndrome. Cell 65(5), 905-14. Wahls, W. P., Wallace, L. J. & Moore, P. D. (1990). The Z-DNA Motif D (TG) 30 promotes reception of information during gene conversion events while stimulating homologous recombination in human cells in culture. Mol. Cell Biol. 10(2), 785-93. Wang, Z. Q., Ovitt, C , Grigoriadis, A. E., Mohle-Steinlein, U., Ruther, R. & Wagner, E. F. (1992). Bone and haematopoietic defects in mice lacking c-fos. Nature 360, 741-53.

178

Martin J. Evans et al.

Weinberg, R. A. (1985). The action of oncogenes in the cytoplasm and the nucleus. Science 230, 770-6. Whitelaw, C. B., Harris, S., McClenaghan, M., Simons, J. P. & Clark, A. J. (1992). Position-independent expression of the ovine j8 lactoglobulin gene in transgenic mice. Biochem. J. 286, 31-9. Williamson, R., Lee, D., Hagaman, J. & Maeda, N. (1992). Marked reduction of high density lipoprotein cholesterol in mice genetically modified to lack apolipoprotein A-I. Proc. Natl. Acad. Sci. USA 89, 7134-8. Willis, A. E. & Lindahl, T. (1987). DNA ligase I deficiency in Bloom's syndrome. Nature 325, 355-7. Windle, J. J., Albert, D. M., O'Brien, J. M., Marcus, D. M., Disteche, C. M., Bernards, R. & Mellon, P. L. (1990). Retinoblastoma in transgenic mice. Nature 343, 665-9. Woychik, R. P., Stewart, T. A., Davis, L., D'Eustachio, G. P. & Leder, P. (1985). An inherited limb deformity created by insertional mutagenesis in a transgenic mouse. Nature 318, 36-40. Woychik, R. P., Maas, R. L., Zeller, R., Vogt, T. F. & Leder, P. (1991). T o r mins': Proteins deduced from the alternative transcripts of the limb deformity gene. Nature 346, 850—3. Wu, C. L. & Melton, D. W. (1993). Production of a model for Lesch-Nyhan syndrome in hypoxanthine phosphoribosyltransferase-deficient mice. Nature Genet. 3, 235-40. Xu, G. et al. (1990). Deletions and a translocation interrupt a cloned gene at the neurofibromatosis type 1 locus. Cell 62, 599-608. Yagi, T., Ikawa, Y., Yoshida, K., Shigetani, Y., Takeda, N., Mabuchi, I., Yamamoto, T. & Aizawa, S. (1990). Homologous recombination at c-fyn locus of mouse embryonic stem cells with the use of diphtheria toxin Afragment gene in negative selection. Proc. Natl. Acad. Sci. USA 87, 9918-22. Yamaguchi, T., Tokita, Y., Franco-Saenz, R., Mulrow, P. J., Peters, J. & Ganten, D. (1992). Zonal distribution and regulation of adrenal renin in a transgenic model of hypertension in the rat. Endocrinology 131, 1955-62. Yankner, B. A., Duffy, L. K. & Kirschner, D. A. (1990). Neurotrophic and neurotoxic effects of amyloid beta protein: Reversal by tachykinin neuropeptides. Science 250, 279-82. Zhang, S. H., Reddick, R. L., Piedrahita, J. A. & Maeda, N. (1992). Spontaneous hypercholesterolemion and arterial lesions in mice lacking apolipoprotein E. Science 258, 468-71. Zhou, X., Sasaki, H., Lowe, L. B., Hogan, L. M. & Kuehn, M. R. (1993). Nodal is a novel TGF-/3-like gene expressed in the mouse node during gastrulation. Nature 361, 543-7. Zimmer, A. & Gruss, P. (1989). Production of chimaeric mice containing embryonic stem (ES) cells carrying a homeobox hoxl.l allele mutated by homologous recombination. Nature 338, 150-3.

6 Large transgenic mammals G. BREM AND M. MULLER

6.1 Introduction

The possibility of expressing foreign genes in mammals by gene transfer has opened new dimensions in the genetic manipulation of animals. The basic techniques of gene transfer were developed in mice, which have been most extensively used in such experiments because they are ideal for studying gene expression during development and for establishing animal models of carcinogenesis and other diseases. Additional applications include analysis of mutations and the use of the transgene as a genetic marker (Jaenisch 1988). Only a few years after the first successful gene transfer into mice, the new technique was used with farm animals, offering the prospect of completely new breeding strategies and other novel applications. However, despite a decade of gene transfer experiments in farm animals, only a few applications have achieved fruition (reviewed in Wall et al. 1992). This is mainly because of fundamental experimental difficulties with these species. In comparison with mice, farm animals have very long generation intervals and the time scale of a transgene project is thus extremely prolonged (Brem 1988, 1989). Naturally this problem is also experienced in conventional breeding programmes, but the major advantage of gene transfer is that significant advances can be made in one generation, whereas conventional breeding techniques require several. Among the theoretically possible techniques of transferring genes, the only one successfully applied to farm animals so far is microinjection of DNA. Researchers have been rather reluctant to perform gene transfer experiments using retroviral vectors due to the slight risk of recombination with wild-type viruses. The feasibility of the simple 179

180

G. Brem and M. Muller

method of using sperm cells as carriers for foreign gene constructs cannot yet be determined (see Section 6.2.4). Without doubt, the most exciting development has been the recent establishment of embryonic stem cells of farm animals and their subsequent use in cloning experiments (Sims & First 1993). Optimization of this technique will give new impetus to gene transfer in farm animals, because it will not only offer the possibility of additive gene transfer and homologous recombination but will also notably reduce such problems as low efficiency, insertional mutations and generation of mosaicism (see Section 6.2.2). Applications of gene transfer into farm animals fall into three groups: 1. The improvement of production efficiency and quality of animal products 2. The production of new proteins of high value 3. The creation of animal models for human diseases and organs for xenotransplantation The most obvious application is the optimization of efficient animal production - not only for economic reasons but, more important, for satisfying the ever-growing requirement for food as the world's population increases exponentially. In future, human nutrition will probably include more plant products; therefore, as the amount of agricultural land declines owing to the increase in population, an obvious challenge for gene transfer will be to improve plant productivity. Certainly by the twenty-first century animal production will be confronted by a dramatic increase in demands for both quality and quantity. These can be met only by intensive efforts in breeding, because the effects of optimizing animal husbandry will presumably soon reach a plateau. Therefore, it is of great importance to improve molecular genetics - for example, gene mapping and gene diagnosis of farm animals. Gene transfer plays an important role in molecular genetics, apart from analytical techniques, because it makes possible the alteration of genotypes. All areas of animal production (e.g., reproduction, health, quantity and quality of products) can be the subject of gene transfer programmes (Table 6.1). So far experiments attempting to fulfil these applications have not been overwhelmingly successful. Nevertheless, it should be possible to use new strategies to improve the efficiency and quality of animal production in the near future, considering the increasing knowledge in genome structure, gene expression and genetic cross-talk.

Large trans genie mammals

181

Table 6.1. Applications of gene transfer in farm animals Improving the efficiency and quality of animal production Production performance (meat, milk, wool, utilization of nutrients, etc.) Reproduction performance (pregnancy rate, ratio of new-born to surviving animals) Health (resistance and susceptibility to diseases, immune response) Quality of animal products (taste, biological value, etc.) Processing of animal products (storage, expiring, structure, etc.) Gene farming Pharmaceuticals for human and veterinary medicine (vaccines, growth factors, blood coagulating factors, antibodies, etc.) Raw materials (proteins for further industrial processing) Enzymes Nutrients (baby food, diets, geriatric diets) Animal models and organs Models of human diseases (high blood pressure, atherosclerosis, cancer, etc.) Organ donors for xenotransplantation

Gene transfer offers a completely new use for large mammals, that is 'gene farming' (see Section 6.4.2.3). Farm animals can be used to synthesize proteins the production of which has been impossible or difficult to obtain at high purity, as raw materials for industrial processing, Pharmaceuticals, enzymes or nutrients. Another aspect that is potentially of great importance for medicine is the genetic alteration of animal organs so that they can be used for transplantation into humans without being rejected by the recipient (xenotransplantation). This would not only supply an almost infinite number of organs but would also have great significance for transplantation surgery in that the toxic permanent immunosuppression of patients receiving human organs might become obsolete. Furthermore, farm animals are useful in establishing models for some human diseases; both the causes and potential means of therapy can be investigated in great detail. Of course, this applies only to diseases where animal models with small mammals are adequate (e.g., heart and circulatory diseases, metabolic diseases, transplantation medicine). In the near future, somatic gene transfer will gain in significance for farm animals. Novel methods for gene transfer into somatic cells (e.g., microbombarding or jet DNA injection of tissue with DNA particles or solutions) promise to be highly efficient. They will provide the possibility of direct genetic immunization (Tang et al. 1992) or other

182

G. Brem and M. Muller

Table 6.2. Assessment ofDNA microinjection for generating transgenic farm animals Advantages Safe and reliable method of transfer No risk of recombination events with wild-type viruses by avoiding viral vectors Direct transfer of DNA (cloned gene constructs, YAC or genomic fragments, chromosomes) into (pro-)nuclei Unlimited size of the DNA transferred DNA integration usually at a single chromosomal locus No need for occasionally complicated culture of early embryonic stages Disadvantages Technically complicated procedure requiring a range of expensive equipment Low efficiency combined with high costs Need for experimenters with high technical skills and extensive experience Random DNA integration in terms of copy number and chromosomal loci Necessity of surgical embryo isolation and transfer in certain species

immunomodulations. Because of the thematic frame of this book, this chapter will concentrate on the generation of transgenic farm animals by germ-line integration. A summary of the methods for gene transfer into different species of farm animals will be followed by a description of transgenic species generated so far and their applications. 6.2 Methods for gene transfer into farm animals 6.2.1 DNA microinjection

At present, direct microinjection of DNA into the pronuclei of zygotes is the method of choice for the generation of transgenic livestock (Table 6.2). In principle the procedure does not differ from that used for mice; however, some features of livestock based on distinctions in morphology of the zygotes and in the embryogenesis should be considered. The efficiency of gene transfer is usually notably lower in farm animals than in mice (Table 6.3); this is because of difficulties in isolation and/or transfer of zygotes, microinjection of DNA and insufficient knowledge of the embryology and reproduction of these species. Since the first reports on the generation of transgenic rabbits, pigs and sheep (Brem et al. 1985; Hammer et al. 1985) gene transfer into livestock has become a reliable and reasonably efficient procedure.

Large trans genie mammals

183

Table 6.3. Efficiencies of gene transfer into mammals

Injectable embryos per donor (n) Donors per recipient (n) Pregnancy rate (%) Born animals/injected embryos transferred (%) Integration frequency (%) Efficiency (transgenics/ injected embryos transferred) (%)

Mouse

Pig

Sheep

Goat

Cattle

15 2 60

20 2 50

5 1 60

4 1 50

7 1 30

20 20

18 10

25 15

15 7

10 7

3

1.5

1

0.5

2.5

Although it is not yet a routine breeding technique, it is used in a number of laboratories. To guarantee the transmission of the foreign gene construct - that is, the integration into somatic cells as well as into the germ-line - the gene transfer has to be carried out as early as possible in the development of the recipient. This requires suitable techniques for the isolation, manipulation and culture of embryonic cells of the corresponding animal species. A programme for the production of transgenic farm animals consists of six main phases: 1. Cloning of an appropriate gene construct and preparation of DNA solution for microinjection 2. Isolation of embryos, that is, superovulation and insemination of donor animals, and synchronization of recipients 3. Microinjection of DNA solution into pronuclei of zygotes 4. Transfer of injected embryos into recipients 5. Screening of new-born animals for integration and expression of the transferred gene 6. Establishment of (homozygous) transgenic lines by the means of conventional breeding The different steps required for the production of transgenic livestock and the subsequent breeding tests with the transgenic founder animals are shown in Fig. 6.1. A typical gene construct should ideally provide all elements controlling specific transcription. The coding portion (structural gene or cDNA) is linked to proximal cis-control elements, the promoter, a

G. Brem and M. Mailer

184 Superovulation of donors

i

Cloning of structural genes and regulatory sequences

In vitro production of zygotes

Collection of embryos

Recombination of gene construct

Visualisation of pronuclei (centrifugation, phase contrast optics)

Preparation of DNA-microinjection solution

DNA-Microinjection into (pro) nuclei In vitro or in vivo temporary culture Synchronisation of recipients

Embryo transfer (surgical, non surgical)

Parturition of offspring after gene transfer

Integration analyses (Southern-, Dot- Blot, PCR)

Transgenic founder Investigation about biological activity (health)

Expression analyses (Northern-, Western-Blot etc)

Mating for Production of F1-generation

Test for homozygous integration (mutation?)

Breeding test of the lines (production, reproduction)

Establishment of a transgenic population

Fig. 6.1. Steps required for the production of transgenic livestock and breeding tests with transgenic animals.

region surrounding the transcription start site specifying quantity, accuracy of initiation and polarity of transcription (for reviews see Mitchell & Tjian 1989; Kozak 1991). More distant cis-elements (Maniatis et al. 1987), enhancers (Miiller & Schaffner 1990 and references

Large transgenie mammals

185

included) and/or silencers (Goodbourn 1989; Baniahmad et al. 1990) control, in co-operation with the promoter, the transcriptional activity and the tissue specificity. Additional remote control elements, referred to as dominant control region or locus control region (LCR), regulate the overall on-off state of the chromatin containing the transcribed unit (reviewed in Eissenberg & Elgin 1991). The structure of gene constructs depends on the experiment they are designed for, the availability of coding sequences as well as promoter/enhancer elements and the capacity of cloning vectors. Early constructs often originated from vectors designed for transfections and expression in mammalian cells and therefore consisted of a simple transcription unit providing basic transcriptional initiation signals that is, constitutive promoters and coding sequences (mostly cDNAs) including appropriate transcriptional termination signals. Besides the choice of appropriate promoters, improved knowledge of active chromatin structure and increased experience in generating transgenic animals have helped to define the following crucial properties of gene constructs required for transgene expression (reviewed in Palmiter & Brinster 1986; Rusconi 1991): 1. Removal of all prokaryotic replication vector sequences for avoiding possible inhibitory effects. 2. Inclusion of intron/exon structures if the coding region consists of cDNA. mRNA processing increases the transcriptional activity of transgenes (Brinster et al. 1988; Choi et al. 1991; Palmiter et al. 1991). 3. Addition of an LCR to the gene construct. LCRs of different genes result in high-level transgene expression which is copy number dependent, integration locus independent and tissue specific (Grosveld et al. 1987; Bonifer et al. 1990; Chamberlain et al. 1991; Schedl et al. 1993). There is still a great lack of inducible transcription systems that will permit tight control of gene expression. None of the existing systems guarantees sufficiently low basal levels and a satisfactory induced expression. Moreover, it is difficult to find systems which would not overlap with normal physiological functions. Cloning of additional promoters and further insight into their regulation should help to overcome these problems. A conceivable alternative for controlling transgene expression is to use a heterologous activator system. Two transgenic lines are produced, one harbouring the target transgene

186

G. Brem and M. Muller

(transresponder), the other an activator transgene (transactivator). The advantage of this multiplex gene regulatory system is that the transgene expression can be activated by mating homozygous lines of transactivators and transresponders (Byrne & Ruddle 1989). Usually, gene constructs for microinjection are cloned in plasmids, cosmids or A-phages. These vectors limit the length of the constructs to 20 or 40 kb, respectively. If the transfer of larger fragments is required, different strategies have to be used. The simplest method is co-injection of two or more fragments based on the fact that DNA integration normally occurs at one chromosomal location per genome. The precise mechanism of DNA integration is unknown, although the observed integration pattern indicates the involvement of ligation and/ or recombination between the injected fragments. Microinjection of overlapping fragments which undergo homologous recombination during the integration provides functional transgenes of more than 40 kb length. Pieper et al. (1992) injected three fragments spanning the human serum albumin gene (hSA) into mouse zygotes. The fragments were 17, 13.1 and 6.7 kb long and had overlaps of 2.5 and 1.85 kb. Twenty (74%) of 27 transgenic mice (DNA integration frequency 25%) had integrated the complete hSA locus; thirteen of those mice expressed hSA. Therefore, nearly 50% of the transgenic mice showed a functional hSA locus resulting from homologous recombination of the injected fragments. A further possibility for transfer of gene constructs of more than 40 kb is the application of yeast artificial chromosomes (YACs) as cloning vectors. The feasibility of this procedure was demonstrated by Schedl et al. (1992). A YAC vector containing the tyrosinase gene (35 kb) was used to generate transgenic mice. The transfer of 147 injected zygotes resulted in 10 transgenic mice (integration frequency 28%). Five mice expressed the expected phenotype. Recently, the same workers successfully transferred even bigger YAC fragments. Schedl et al. (1993) have reported the transfer of a 250-kb YAC construct by pronuclear injection of gel-purified YAC DNA. The YAC construct was inserted without major rearrangements, and the transgene expression was integration site independent. Microinjection into all mammals requires the rather complex equipment listed in Table 6.4. The integration frequency of gene constructs can be positively influenced by the properties of the DNA microinjection solution such as purity and osmolarity (sterile filtered, isotonic), conformation of the gene construct (linearized), as well as its termini (protruding) and

Large transgenic mammals

187

Table 6.4. Equipment necessary for microinjection Equipment Stereodissecting microscope with transmitted illumination Stable working bench for micromanipulation Inverse microscope (32 x , 400 x), Nomarski's optic (and video equipment) Mechanically movable micromanipulator Smooth movable micromanipulator Microinjection device Injection chamber (covered with paraffin) and criss-cross table (temperature-controlled) Microforge Mechanical pipette puller Centrifuge (15,000 g)

Use Embryo collection and preparation for transfer Construction of microinjection unit Visualization of (pro-)nuclei and microinjection Holding of fixing pipette Moving of injection pipette DNA flow in injection pipette Keeping and moving of embryos Production of pipettes Production of injection pipettes Removal of particles from the DNA microinjection solution by centrifugation; centrifugation of porcine and bovine oocytes for visualization of the pronuclei

DNA concentration (100 to several 1,000 copies per picolitre). The protocols for the preparation of DNA microinjection solutions used for farm animals are identical to those developed for mice (Hogan et al. 1986). Because gene transfer into farm animals is an extremely complicated procedure, it is necessary to optimize all factors influencing its efficiency (Table 6.5) (Grosschedl et al. 1984; Brinster et al 1985; De Pamphelis et al. 1988; Jaenisch 1988). The survival rates of injected embryos and the integration rates achieved in published gene transfer programmes differ among experiments. Success rates in generating transgenic livestock are therefore unpredictable, and in most cases it is impossible to identify the factors that have influenced the gene transfer efficiency. Our knowledge of events during microinjection and DNA integration and of the subsequent development of injected embryos is still too scanty to apply systematic optimization.

188

G. Brem and M. Muller

Table 6.5. Factors influencing the efficiency of gene transfer by DNA microinjection Factor

Influence

DNA concentration

Optimum integration frequency at concentration of >1 /xg/ml, which usually equals several 100 copies/ pi; concentrations > 10 /xg/ml might decrease the embryonic survival rate Linearized gene constructs integrate with higher frequency than circular DNAs DNA fragments with protruding ends integrate better than blunt-ending DNA DNA solutions are kept in 5-10 mM Tris/HCl (pH 7.4)/0.1-0.25 mM EDTA; higher EDTA solutions increase embryonic death; alternatively, physiological buffer can be used (48 mM K2HPO4/4.5 mM KH2PO4/14 mM NaH2PO4) Requirement for any particles or components that might obstruct the injection pipette or be embryotoxic Transfer of prokaryotic vector sequences might negatively influence the transgene expression level

DNA conformation Ends of the DNA fragments Injection buffer

Purity of the DNA solution Vector sequences

6.2.2 Embryonic stem cells Nowadays embryonic stem cells (ES cells) of mice are routinely isolated, cultured, genetically transformed and used for the generation of chimeras in many laboratories (see Chapter 5). So far this method has failed with farm animals due to the lack of appropriate stem cell lines. Many laboratories are devoted to the production of ES cell lines of farm animals. (For cattle see Stringfellow et al. 1987, 1991; Schellander et al. 1989; Evans et al. 1990; Hassan-Hauser et al. 1990; Strelchenko et al. 1991; Strojek-Baunack et al. 1991; Anderson 1992; Saito et al. 1992. Swine: Piedrahita et al. 1988, 1990a, b; Evans et al. 1990; Notarianni et al. 1990a, b, 1991; Strojek et al. 1990. Sheep: Butler et al. 1987; Handyside et al. 1987; Rexroad 1990; MeineckeTillmann & Meinecke 1991; Notarianni et al. 1991. Goat: MeineckeTillmann & Meinecke 1991. Rabbit: Giles et al. 1993; N. Strelchenko, pers. commun.) The problems of establishing ES cells of livestock are caused mainly by the limited availability of embryos with a defined genetic background (i.e., there are no comparable inbred strains), and furthermore, the embryonic development of large mammals is not as well under-

Large transgenie mammals

189

stood as that of mice. It is not known at present which developmental stage is optimal for the isolation of stem cells. Cell lines derived from the inner cell mass (ICM) of blastocysts, first, have an extremely slow cell division rate, and, second, tend to differentiate, resulting in cell death after a few passages. There are several ways to increase the efficiency of establishing ES cells in farm animals. One is to use homologous leukaemia inhibitory factor (LIF) for the inhibition of cellular differentiation. The derivation of ES cells from mouse primordial germ cells has demonstrated the necessity of LIF as well as steel factor (SF) and fibroblast growth factor (Matsui et al. 1992). These growth factors are at present being tested on stem cell cultures of farm animals. Other growth factors, specialized media and feeder cell lines are being investigated with the aim of improving the culture conditions. Media include Ham's F20, CMjS (buffalo red liver cell-conditioned), DMEM, BME/Ham's F10 (1:1), TCM-199 and a-MEM (5637-conditioned). The various feeder cell lines as well as mixed cultures include STO fibroblasts, primary embryonic fibroblasts of mice, sheep or swine, bovine foetal fibroblasts derived from liver, kidney, testis and uterus, epithelial cells of oviduct and uterus, granulosa cells and buffalo red liver cells. Growth factors tested besides LIF and SF include ciliary neurotropic factor, insulin, epidermal growth factor, and transforming growth factor beta. Cell lines similar to stem cells have been isolated from hamster and porcine embryos by Doetschmann et al. (1988) and Piedrahita et al. (1988), respectively. Evans et al. (1990) and Notarianni et al. (1990a, b, 1991) reported the in vitro culture of cells derived from porcine and bovine embryos for a period of 9 months. These cells are phenotypically ES cells. However, the main property of ES cells (i.e., pluripotency) has yet to be shown. Recently, two groups have succeeded in establishing bovine totipotential cell lines (Sims & First 1993; N. Strelchenko, pers. commun.). N. Strelchenko (pers. commun.) harvested in v/fro-generated blastocysts on days 8-11 and cultured them on mitosis-inactivated primary mouse embryo fibroblasts in a-MEM medium supplemented with 15% foetal calf serum (FCS). Cells were passaged every 8-12 days. Fifty to sixty per cent of all blastocysts resulted in established cell lines that have so far been cultured for more than 16 passages. These cell lines were used for cloning experiments by means of nuclear transfer into enucleated oocytes. One per cent of the fusion products developed to blastocysts. The cloned embryos were transferred into appropriate recipients (10% pregnancy

190

G. Brem and M. Muller

rate). Unfortunately, so far all pregnancies have been aborted between days 34 and 45. Sims and First (1993) generated cell lines from in v/fro-produced embryos of 9-10 days of age (Fig. 6.2). The immunosurgical isolation of the ICM of up to three embryos resulted in stable cell lines with a success rate of 30-50%. The cells were cultured in low-density suspension and kept in CR1 medium supplemented with 5% FCS, additional amino acids, selenium, insulin and transferrin. These ES-resembling cells were used in 659 nuclear transfers into enucleated oocytes (fusion with polyethylene glycol). Seventy per cent of the fused cells underwent cell divisions and 24% of the surviving clones developed to blastocysts. The transfer of 34 blastocysts into 27 cows gave a pregnancy rate of 49%, resulting in the live birth of four healthy calves. This experiment demonstrated for the first time the feasibility of deriving ES cell lines from embryos, culturing them to totipotency in vitro and generating calves via nuclear transfers. Although this procedure has to be repeated and optimized for practical application, it nevertheless marks the beginning of a new era in the generation of transgenic livestock and animal breeding. The use of ES cells eliminates the majority of problems that have previously occurred in gene transfer experiments: 1. The integration of the transgene can be readily tested in ES cells before further development. Appropriate techniques even make it possible to study transgene expression. 2. Chimeras in cattle can be generated from morulae or blastocysts, which are isolated nonsurgically or produced in vitro. After being manipulated, the embryos are transferred by conventional nonsurgical techniques. 3. One hundred per cent of all animals born have integrated the gene construct and have the desired sex. 4. Cloning of ES cells does not result in mosaics. All animals born are genetically identical. 5. Homologous recombination permits the removal of certain existing alleles as well as their replacement with others. This considerably extends the spectrum of gene transfer applications in farm animals. ES cell-mediated gene transfer will certainly replace DNA microinjection as soon as appropriate cell lines are available for farm animals.

culture in CR1-medium spontaneous disaggregation + AA + 5% FCS + SIT of the iCM and proliferation of cells (up to 2000/ICM)

IVP blastocysts (day 9-10) isolation of the ICM (immunosurgery)

fusion (PEG or electrofusion)

oocyte with cumulus

maturation

enucleation

enucleated oocyte

Fig. 6.2. Generation of embryonic stem cells from bovine embryos and cloning of embryonic stem cells. AA, Amino acids; FCS, fetal calf serum; ICM, inner cell mass; PEG, polyethylene glycol; SIT, selene-insulin-transferrin. (After Sims & First 1993.)

192

G. Brem and M. Mtiller

6.2.3 Retroviral vectors Viral DNA was first used for gene transfer in 1974; transgenic mice were generated following injection of SV40 DNA into preimplantation mouse blastocysts (Jaenisch 1974; Jaenisch & Mintz 1974). Subsequent experiments were carried out with Moloney murine leukaemia virus. Infection of preimplantation mouse embryos resulted in the integration of viral DNA and its transmission to progeny (Jaenisch et al. 1975; Jaenisch 1976). These early findings resulted in the development of two principal types of vector systems for the expression of foreign genes. The first utilizes replication-competent vectors containing all genes necessary for viral replication. Infection of an early embryo or cell line is followed by reverse transcription of the viral RNA. The newly synthesized proviral DNA containing the transgene is integrated into the host genome. These vectors have the major disadvantage of generating transgenic animals that could potentially produce infectious retrovirus. In addition, they have a limited cloning capacity of only approximately 2 kb. The second system is based on replication-defective retroviral vectors depending on a helper virus cell system for infectivity (Mann et al. 1983). Vectors of this kind lack the viral gag, pol, and env genes, but contain the information for encapsulation (ijs+). Transfection of vector DNA into a helper cell line will produce 'oneround' infectious retro virus, which inserts the transgene into the genome of infected cells but is subsequently unable to replicate. These vectors have a cloning capacity of approximately 8-10 kb. Retroviral-mediated gene transfer has been used mainly in mice (reviewed in McLachlin et al. 1990a). Since this chapter is concerned predominantly with gene transfer into large mammals, the following is a list of early publications describing gene expression in rodents following retroviral-mediated gene transfer: mutant human dihydrofolate reductase (van der Putten et al. 1985); bacterial neomycin gene (Huszar et al. 1985; Rubenstein et al. 1986); human j8-globin gene and neomycin (Soriano et al. 1986); v-myc and human adenosine deaminase gene and f&-neomycin (Stewart et al. 1987). So far, few publications have described the use of retroviral vectors in domestic animals. However, Salter et al. (1986, 1987; Salter & Crittenden 1989) achieved the insertion of wild-type and recombinant avian leukaemia virus into the germ-line of chicken by injection of retrovirus into eggs before incubation. The following initial attempts have been made to use retroviral

Large trans genie mammals

193

vectors for gene transfer in sheep and pigs. Harvey et al. (1990) injected a high-titer solution of feline leukaemia virus (FeLV) under the zona pellucida of two-to-four-cell ovine embryos. After transfer of the embryos into recipients and 59 days of pregnancy, the foetuses were analysed. Of the 17 foetuses, 2 (12%) showed integrated FeLVspecific sequences. Recent efforts have concentrated on the development of a packaging cell line for FeLV. Petters et al. (1989) established an avian retrovirus (spleen necrosis virus, SNV) in a canine cell line. Approximately 100 cells were injected into 122 porcine blastocysts and transferred to 12 recipients. From the 4 resulting pregnancies 21 normally developed foetuses were isolated after 6 weeks. Polymerase chain reaction (PCR) analysis showed the integration of SNV DNA in the organs of 17 foetuses (80%). There are some obvious advantages of using retroviral vectors, including the fact that infection generally results in collinear integration of few gene copies, many cells can be handled simultaneously and the host genome is usually unaffected by the integration of the retroviral sequences. However, some major disadvantages have become apparent. Expression studies in transgenic mice generated by retroviral vectors revealed a frequent lack of transgene transcription controlled by retroviral enhancer/promoters. Apparently there are mechanisms in early embryos or ES cells that abolish the expression of these vectors during development and occasionally permit weak expression in adults. Although DNA methylation was initially thought to be responsible, it does not seem to be the only cause, as the degree of methylation is not sufficiently correlated to the expression data. However, this problem can be avoided by the use of appropriate 'external' promoter elements instead of the retroviral expression cassette. As mentioned earlier, retroviral vectors have a limited cloning capacity (maximum of 8-10 kb). Safety considerations are very important with the use of viruses as gene transfer agents (Goff & Shenk 1993). The potential of spreading infectious vector virus is extremely unlikely with replication-defective retroviral vectors. These 'one-round' infectious vectors are generally used in gene transfer experiments. However, even replicationdefective retrovirus might cause the activation of cellular retroviral sequences by recombination, though the occurrence of such an event is highly unlikely. Nevertheless, mouse models should be used in long-term experiments to transmit retroviral transgenes through many generations with the aim of satisfying safety demands. Perpetual improvement of retroviral vector systems will guarantee high safety stan-

194

G. Brem and M. Muller

dards that should rule out any hypothetical risks (Dougherty & Temin 1987; Salmons et al. 1991; reviewed in Boris-Lawrie & Temin 1993). Our laboratory constructed a retroviral vector system lacking all infectious properties (Janka et al. 1993), so as to avoid the potential risk of retroviral recombination and yet retain the advantages of the retroviral integration characteristics. The system is based on a vector containing the packaging signal (BAG vector) and a cell line (GAPp BAG) providing only the intracellular predecessors of virus particles that is, core particles. GAPpBAG lacks the env gene and therefore is unable to produce infectious retrovirus. Core particles can be isolated either from cytoplasmic extracts or from medium supernatants and subsequently injected into nuclei of embryos. 6.2.4 Alternative gene transfer techniques Alternative methods of gene transfer are provided by techniques originally developed for the transfection of eukaryotic cells. However, DNA transfection procedures using calcium phosphate, DEAE-dextran or electroporation are inappropriate because of their relative inefficiency and problems with unstable integration and rearrangements. The use of liposome-mediated gene transfer is combined with blastocyst manipulations. Rottmann et al. (1985) microinjected liposomes containing gene constructs into 89 mouse embryos. Five of the resulting 24 animals were transgenic. Reed et al. (1988) performed a similar experiment with bovine blastocysts, though the results have not been published. A stunningly simple procedure for generating transgenic animals has been published by Lavitrano et al. (1989). DNA (pSV-CAT gene construct) was mixed with spermatozoa before in vitro fertilization of mice. Nearly 30% of the new-born mice carried the foreign DNA and transmitted it to their progeny. Expression of the pSV-CAT construct could be monitored. The same gene construct was used for spermatozoa-mediated gene transfer into pigs. The surgical insemination of 22 sows resulted in 16 pregnancies. Ten piglets (21% of newborn piglets) were transgenic and expressed the SV-CAT construct (Gandolfi et al 1989). Brackett et al. had already demonstrated in 1971 that sperm cells could function as DNA transport vehicles into rabbit oocytes. However, the SV40 DNA used was not shown to be integrated into the host genome.

Large transgenie mammals

a

195

a

• •

Fig. 6.3. Polymerase chain reaction (PCR) and Southern blot analysis of murine embryos and foetuses generated by in vitro fertilization with sperm cells incubated with foreign DNA. The great potential of spermatozoa-mediated gene transfer led to immediate efforts by various laboratories to repeat the experiments just described. So far, no successful genetic transformation by this method has been reported (Brinster et al. 1989). PCR analysis has been used to trace exogenous DNA being carried into oocytes by sperm cells. Hochi et al. (1990) found that the last embryonic stage at which they could detect the foreign DNA was the blastocysts. Our own experiments showed specific amplification of a foreign gene construct (MT-GHRH) in early mouse embryos, but no exogenous DNA was detectable after 10 days of development (Fig. 6.3). The mechanisms of spermatozoa-mediated gene transfer have to be explored further, before its use becomes a practical possibility. 6.3 Gene transfer into pigs 6.3.1 Treatment of donors and recipients

Most gene transfer experiments in livestock have been performed with swine. This is due in part to the economical importance of pig production in many countries and in part to the reproductive nature of pigs. Superovulation of sows usually results in more than 20 embryos for microinjection per donor (Table 6.6). Embryos can be collected not

Table 6.6. Gene transfer programmes in pigs: treatment of donors and isolation of embryos

Breed

Age/ body weight

German Landrace, Prepuberal Pietrain, crossgilts/60-90 breeds kg Mature gilts

Mature gilts

Large White

Multiparous

Yorkshire

—

Cross-breeds, Sows Landrace, Yorkshire, Hampshire 8.5-9 months/ 115-25 kg Large White, Lan- 230 days drace, crosses >7 months

Synchronization

InsemiInduction of ovulation nation, (HCG) [hr after hr after HCG PMSG] Superovulation (PMSG)

Reference

Not necessary

1,250 IU

750 IU i.m. [72 hr]

24 + 36hr

Brem et al. (1985)

15 mg Altrogenost for 5-9 days starting at 12th16th day of cyclus

1,500-2,000 IU 24-30 hr after last admin, of Altrogenost

500 IU [72 hr]

18 + 36hr

120 mg Methallibure over period of 6 days, starting at 10th— 15th day of cyclus

1,500 I U s . c , 24 hr after last admin, of Methallibure

500 IU i.m. [72 hr]

16 + 36hr

Hammer et al. (1985); Miller et al (1989) Pursel et al. (1988)

750 IU

500 IU [44 hr]

Mating

2,000 I U s . c , 24 hr after last admin, of Altrogenost

1,000 IU [78 hr]

24, 36 + 48 hr Ebert et al. (1988)

400 IU i.m.

200 IU

Wieghart et al. (1990)

5 g SuisynchronPremix over a period of 15 days Cloprostenol

1,500 IU

500 IU [79 hr]

24 + 30hr after begining of oestrus 24 + 39 hr

1,000 IU

32 hr

Allyl-Trenbolone over a period of 12-14 days, 15 mg/day oral, on last day 2 i.m. admin, of 10 mg PGF 2a in an 8hr interval

1,500 IU i.m. after 24 hr

500 IU [72 hr] 750 IU [80 hr]

French (1991) Swanson et al (1992)

15 mg Altrogenost for 9 days, starting at 12th16th day of cyclus

12 + 24hr after beginning of oestrus

Vize et al. (1988)

Briissow et al. (1990, 1991)

198

G. Brem and M. Muller

only from living donors by surgical flushing of oviducts but also from slaughtered animals. This greatly increases the number of potential donors and therefore the number of available embryos. While both quantity and quality of embryos isolated from prepuberal donors are higher than from older animals, multiparous sows offer notable advantages as recipients in terms of pregnancy rates and embryonic survival. The use of older sows requires the synchronization of the cyclic hormonal activity by medications that are unlicensed in some countries (e.g., Federal Republic of Germany). Our experimental procedure, described in the following paragraphs, is therefore designed for prepubescent gilts. Superovulation in donor animals is stimulated on day 0 by administration of pregnant mare's serum gonadotropin (PMSG) (1.250 IU, Intergonan, VeMie). Ovulation is induced by human chorionic gonadotropin (HCG) (750 IU, Ekluton, VeMie) on day 2 followed by two consecutive inseminations after 24 and 36 hr. The fertilized oocytes are collected 24 hr later (Table 6.7). The recipients are treated likewise, except with a 12-hr delay and a reduction in the dosage of PMSG treatment to 750 IU. Manipulated embryos are transferred to the recipients 60-3 hr after induction of ovulation. Embryos are isolated surgically from anaesthetized donors (160 mg azaperon, Stresnil, Janssen per 400 mg metomidat hypochloride, Hypnodil, Janssen). The operation is carried out on a mobile operating table with the animal lying on its back. After the usual hygienic precautions, the abdomen is opened in the linea alba, and uterus, oviduct and ovary are retracted from the abdomen. The oviduct is reached via perforation of the uterus at the uterotubular border with a curved glass canula and is flushed with 50 ml phosphate-buffered Dulbecco's medium (PBS) (Table 6.7), which is collected in a sterile petri dish (Fig. 6.4). The zygotes are isolated, cultured in fresh medium and classified morphologically with the help of a stereomicroscope. The number of isolated zygotes suitable for microinjection was positively correlated with the length of the seasonal signs shown by the donors as well as by their body weight (Table 6.8). In addition, we observed an increase in fertilized oocytes by using different boars for the insemination of the donor animals. The influence of stress factors (e.g., transport, centrifugation and microinjection) on the further development of transferred zygotes is summarized in Table 6.9.

Large transgenic mammals

199

Table 6.7. Isolation of embryos from superovulated sows Medium

Hr after HCG 60-3

59-66

50 ml PBS (Dulbecco's) per oviduct PBS + 0.4%BSA — 10-15 ml Kreb's Ringerbiscarbonate solution — BMOC-3 buffered with Hepes —

65-8

PBS

57-60 10/ovul.

52-4

Centrifugation

Reference

15,000 g, 3 min

Brem et al. (1985)

7,000 g, 3 min 13,000 #

Pursel etal. (1988) Vize et al. (1988) Ebert et al. (1988)

— 10,000 g 5,000 g, 6 min 9,800 g, 6-10 min

Polge et al. (1989) Wieghart et al. (1990) Swanson et al. (1992) Brussow et al. (1990, 1991)

Abbreviations: PBS, phosphate-buffered saline; BSA, bovine serum albumin; BMOC ~3 , modified Brinsters mice ova culture medium 3.

Fig. 6.4. Surgical embryo collection by flushing of a porcine oviduct.

G. Brem and M. Muller

200

Table 6.8. Influence of donor's body weight on number of collected and microinjected porcine oocytes Oocytes per donor animal

Body weight (kg)

Donor (i)

Microinjected

Isolated0 (n)

n

%

81

221 182 51

35

30.7 34.6 34.2 38.1

20.1 20.0 18.6 17.2

66 58 54 45

Total

489

34.5

19.2

56

Note: 509 stimulations; 489 oviduct flushings (96%). a Embryo collection rate (collected oocytes per corpora lutea), 95%.

Table 6.9. Influence of centrifugation and microinjection on the developmental rate of porcine zygotes kept in oviduct in vivo culture for 4 days

Manipulation of zygotes No transport, no microinjection Centrifugation Centrifugation and transport Centrifugation and microinjection Centrifugation, microinjection and transport

Transfers in recipients («)

Embryos transferred into oviduct in)

Collection rate (%)

Developmental rate (%)

14 12

655 639

73 66

52 44

4

185

54

30

25

1,147

68

14

13

554

64

10

The pronuclei of porcine zygotes are obscured microscopically, because their cytoplasm is extremely granular (Fig. 6.5). Therefore, before DNA microinjection, the pronuclei have to be visualized by centrifugation of the zygotes (e.g., 15,000 gy 3 min; Table 6.7) (Wall et

Large transgenie mammals

201

Fig. 6.5. DNA microinjection into the pronucleus of a porcine zygote after centrifugation. al. 1985). After culturing of the microinjected cells for a couple of hours, morphologically damaged cells are removed. Embryos were transferred into the oviduct of 12-hr asynchronous recipients. Surgical preparation of these animals was performed in the same manner as for the donors. Migration of embryos ('spacing') results in their equal distribution in both uterine horns. Under appropriate conditions, porcine zygotes can be cultured in vitro to a developmental stage that enables transfer directly into the uterus. This procedure is obligate for nonsurgical embryo transfer, first applied by transcervical implantation (Polge & Day 1968; Sims & First 1987). Our experiments applying nonsurgical transfer in swine resulted in pregnancy rates of 10% (Reichenbach et al. 1993) and 50% respectively (J. Modi et al. unpublished data). However, the embryo survival rate in these experiments was considerably lower than after surgical transfer. Pregnancy and embryo survival rates (embryo transfer efficiency) are positively influenced by the intensity of oestrus and the recipients having an optimum body weight (70-80 kg) (Table 6.10). Pursel et al. (1987) and Wei et al. (1993) noticed that efficiency was increased from 0.83% to 1.04% by co-transferring control embryos alongside the injected embryos. Similar results were obtained in our experiments (Table 6.11). Furthermore, a modified embryo transfer programme

202

G. Brem and M. Muller

Table 6.10. Influence of body weight of prepuberal recipients on the pregnancy rate and number of piglets Piglets per Body weight (kg) 81 Total

No. of transfers

Pregnancy rate (%)

Litter (")

Transfer (n)

48 92 57 197

31 48 30 37

3.4 4.0 5.4 4.3

1.1 1.9 1.7 1.6

Table 6.11. Influence of transferring control embryos on the gene transfer efficiency in pigs Transfer of microinjected zygotes Without control embryos No. of transfers Microinjected embryos per recipient (n) Control embryos per recipient (n) Pregnancy rate (%) Total no. of new-born piglets Embryonic survival rate (%) Transgenic piglets Integration rate (%) Efficiency {%) (transgenic piglets per embryos transferred

With control embryos

20

11

38

28

— 40

8 64

38

37

5 4 10.5

9.3 3 8.1

0.5

1.0

has been developed to reduce the number of animals and surgical manipulations required (Table 6.12). Although all criteria affecting the integration and developmental rate of transferred embryos have not been fully clarified, the data collected in other laboratories indicate tendencies similar to those mentioned here (Hammer et al. 1986; Pursel et al. 1987, 1988; Rexroad et al. 1987).

Table 6.12. Usage of donor animals as recipients during the same embryo transfer experiment

Treatment Insemination of donors during the following oestrus after embryo collection Transfer of microinjected embryos into synchronized recipients (conventional) Transfer of unmanipulated embryos into oviducts of donors Transfer of microinjected embryos into oviducts of donors

Donors

Embryos per animal

14

Foetuses and/or piglets per Pregnancy rate

Litter

Donor/recipient

43

10.7

4.6

Survival rate

42

36

40

4.8

1.9

5.3

30

24

53

6.3

3.3

14.2

45

35

40

3.9

1.6

4.5

204

G. Brem and M. Muller

6.3.2 Transgenic pigs The use of 21 different constructs in gene transfer experiments revealed the following data: embryonic survival rate, 7.8%; integration rate of the gene construct, 8.7%; expression rate among transgenics, 42%; and total efficiency (transgenic piglets per transferred microinjected embryos), 0.67% on average (Table 6.13). Most gene transfer experiments have been carried out for the purpose of altering growth performance and/or carcass composition. In addition, some transfer programmes have concentrated on the improvement of disease resistance, expression of foreign proteins in the mammary gland and production of functional human haemoglobin.

6.3.2.1 Gene constructs altering growth-related functions Experimenters attempting to change growth performance by transgenic means have to be aware that growth is a complex process depending on both genetic and environmental factors. Briefly, the growth hormone cascade consists of polypeptides synthesized in the hypothalamus, that is, growth hormone-releasing hormone (GHRH) and its antagonist, controlling the production of growth hormone (GH) in the pituitary gland. Association of GH with its receptor induces synthesis of insulin-like growth factor I (IGF-1). All of these hormones are tightly controlled by positive and negative feedback involving both themselves and others. Despite positive effects of GH administration in livestock and in contrast to the findings in transgenic mice, GH-transgenic pigs did not generally demonstrate increased growth performance (Hammer et al. 1985; Ebert et al. 1988; Viz et al. 1988). This is caused by a 20% reduction in their food uptake combined with an increased utilization of nutrients. Only after the application of a protein-enriched diet supplemented with lysine, minerals and vitamins did GH-transgenic pigs attain a 15% higher daily weight gain than control animals (Pursel et al. 1988, 1989a). In terms of production-relevant properties, the transgenic pigs showed a massive reduction of the back fat thickness (40%) (Hammer et al. 1986; Pursel et al. 1989a, 1990). However, overexpression of GH by use of strong promoters (i.e., metallothionein promoter or M-MLV promoter/enhancer) causes a variety of pathological side effects in pigs, such as gastric ulcers, severe synovitis, dermatitis, nephritis, cardiomegaly, pneumonia, in-

Table 6.13. Results of transferring gene constructs into pigs

Construct mMT-I-hGH mMT-I-hGH mMT-I-hGH mMT-I-bGH hMT-pGH MLV-rGH bPRL-bGH WAP-2-hGH mMT-I-hGRH niMT-I-hGRH MT-hGRF Alb-hGRF mMT-IhlGFI mMT-Mx SV-Mx mMx-Mx PEPCK-bGH MLV-pGH CMV-pGHSV40 MLV-pGH-SV40 k-y-l-mAB WAP-WAP IgA LTRcSKI LCR-aap Total

Embryos injected and transferred (n) 2,035 268 1,014 2,193 423 170 289 1,028 1,041 2,236 2,627 968 387 1,083 809 1,629 1,057 410 372 312 — 850 542 1,091 709 26,602

Offspring (incl. foetuses) n

%

192 9.4 5.6 15 2.1 21 6.7 149 4.0 17 8.8 15 6.9 20 51 5.0 5.4 54 177 7.9 238 8.9 108 11.2 34 8.8 22 2.0 26 3.2 77 4.7 124 11.7 59 14 32 9 33 11 — — 189 22.2 302 112 2,067

28 15.8 7.8

Integration n

%

20 1 4 11 6 1 5 7 6 7 8 5 4 6 1 8 7 6 15 10 2 5 2 29 3 179

11.0 6.7 19 7.4 35.3 6.7 25.0 13.7 0.6 4.0 3.4 3.8 11.8 27.3 3.8 10.4 5.6 10 47 30 — 2.6 9.6 2.6 8.7

Efficiency (%) 0.98 0.4 0.4 0.5 1.42 0.59 1.73 0.7 0.6 0.3 0.30 0.52 1.03 0.6 0.1 0.5 0.66 1.5 4.0 3.2 — 0.59 0.4 2.7 0.42 0.67

Expression n

%

Reference

11/18 — 2/4 8/11 1/6 1/1 3/3 1/3 1/4 2/7 2/8 3/3 1/4 0 — 2/5 5/7 1/6 2/23 0/10 1/1 — 2/2 10/29 3/3 62/148

61 — 50 73 17 100 100 33 25 29 25 100 25 0 — 40 71 17 15 0 100 — 100 34 100 42

Hammer et al. (1985) Brem et al. (1985) Brem et al. (1988b) Rexroad and Wall (1987) Vize et al. (1988) Ebert et al. (1988) Polge et al. (1989) Brem et al. (1989) Brem et al. (1988a,b) Pursel e/a/. (1989b) Pursel e/a/. (1989b) Pinkert et al. (1987), Pursel et al. (1989a) Pursel ef a/. (1989a) Brem et al. (1988a, b), Miiller et al. (1992) Brem et al. (1988b), Muller et al. (1992) Brem et al. (1989), Muller et al. (1992) Wieghart et al. (1990) Ebert et al. (1990) Ebert et al. (1990) Ebert et al. (1990) Weidle et al. (1991) Wall et al. (1991) Lo etal. (1991) Pursel et al. (1992) Swanson et al. (1992)

206

G. Brem and M. Muller

sulin resistance and reduced fertility (Pursel et al. 1989a; Ebert et al. 1990). Gene constructs with different regulatory elements providing lower constitutive (PEPCK promoter) or artificially inducible levels of GH expression (prolactin promoter) resulted in the desired increase in carcass leanness and reduction in detrimental side effects (Polge et al. 1989; Wieghart et al. 1990). Pursel et al. (1992) transferred the avian c-ski gene controlled by the mouse sarcoma virus promoter/enhancer for the purpose of improving muscle development. The proto-oncogene c-ski induces myogenic differentiation. As shown in mice (Sutrave et al. 1990), the c-ski gene construct was predominantly expressed in skeletal muscles, causing selective hypotrophy of type II fast fibers. The extent of this was highly dependent on the transgene expression level. Five transgenic pigs showed muscle hypertrophy at the age of 3-7 months. However, five other pigs developed muscle atony between shortly after birth and 3 months. Histological examination demonstrated a high degree of vacuolic degeneration of the muscle tissue. Altering growth-related traits by gene transfers requires gene constructs that ideally show gene expression correlated to specific metabolic procedures or inducible by administered inducers. This prevents detrimental overexpression and guarantees transgene activity limited to growth-performance-related periods. 6.3.2.2 Gene constructs for improving disease resistance An important and challenging aspect of gene transfer in farm animals is the introduction of beneficial genes to improve health and disease resistance (Muller & Brem 1991; Staeheli 1991). Such attempts in pigs include the use of immunoglobulin gene constructs ('genetic immunization') (Lo et al. 1991; Weidle et al. 1991) and the transfer of a murine gene (Mxl) conferring specific resistance to influenza viruses (Muller et al. 1992). Expression of a transgenic immunoglobulin specific for a common pathogen could provide an animal with congenital immunity for that pathogen. As shown by many investigations, cloned genes coding for monoclonal antibodies can be expressed in large amounts in transgenic mice. These mice produce antibodies against specific antigens without prior contact or immunization (reviewed in Storb 1987; Bluethmann 1991; Iglesias 1991). To evaluate whether antibodies of diagnostic or

Large transgenic mammals

207

therapeutic interest could be produced in farm animals, Weidle et al. (1991) introduced the genes for the light and heavy chain of a mouse monoclonal antibody directed against 4-hydroxy-3-nitrophenylate into the germ-line of pigs. One pig expressed the transgene. A level of 1,000 /mg monoclonal antibody per millilitre was measured in the serum. Isoelectric focusing experiments revealed that in the transgenic pig only a minority of the bands matched those of the purified mouse antibody. This finding can be explained by assuming tissue-specific post-translational modifications and heterologous immunoglobulin chain associations. Since the production of antibody to various polysaccharide antigens can be protective against pathogenic bacteria, Lo et al. (1991) generated transgenic pigs carrying genes coding for the mouse a- and K-chains for antibodies against phosphorylcholine. Two transgenic pig lines were established; in both founder animals, only the mouse IgA transgene was integrated. Despite the absence of any mouse L-chain, high titers (600 to more than 1,000 jtig/ml) of mouse IgA were found in the serum, indicating that it was able to form complexes with the endogenous pig L-chain. Little if any of the mouse IgA in the porcine serum was capable of binding phosphorylcholine specifically, but it is suggested that the mouse IgA may be able to function as an antigen receptor in pig B-cells and therefore contribute to the antibody repertoire of the transgenic animal. Despite some unexpected findings, both experiments illustrate the potential of introduction of beneficial traits such as germ-line-encoded immunity into pigs. In animals only a few instances of a single genetic locus responsible for disease resistance are known. A well-examined example is the Mxl gene product of certain mouse strains. The synthesis of mouse Mxl protein in various cell lines and transgenic mice demonstrated that it is both necessary and sufficient to promote resistance to influenza viruses in previously susceptible cells and animals (Staeheli et al. 1986; Arnheiter et al. 1990; Kolb et al. 1992). With the cloning and functional characterization of this specific disease-resistance gene, it was possible to undertake a gene transfer programme examining whether Mxl transgenic pigs would show reduced susceptibility to influenza infections (Muller et al. 1992). The gene construct used consisted of the Mxl cDNA controlled by the Mxl specific regulatory elements, which is inducible by interferons. Five transgenic pig lines were established, of which two showed interferon-inducible expres-

208

G. Brem and M. Muller Control animal

Mxl transgenic pig +

—

Interferon

28S

18 S

Fig. 6.6. Northern blot analysis of Mxl transgenic pigs.

sion of transgene-specific mRNA (Fig. 6.6). Despite extensive protein analysis, no mouse Mxl protein could be detected in the transgenic animals. The most likely explanation for this result is based on the findings that permanent high-level expression of Mxl protein, particularly its presence during embryogenesis, is not tolerated by the organism (Arnheiter et al. 1990; Kolb et al. 1992; Muller et al. 1992). Although the gene construct was shown to be tightly regulated in mouse cell lines, it showed a basal low-level transcriptional activity in tissues of the transgenic animals. Considering the deleterious effect of permanent Mx synthesis, the observed leakiness of the promoter used may have allowed the embryonic development of only those animals with abolished translation of the gene construct. The gene transfer experiments in pigs demonstrated that the choice of the regulatory elements controlling Mxl transgene expression is crucial. Therefore, future gene constructs should guarantee tight transgene regulation during embryonic development and highly inducible Mxl synthesis after exogenous stimuli (Staeheli 1991; Muller et al 1992).

Large transgenic mammals

209

6.3.2.3 Trans gene expression in the mammary gland It is obviously more efficient to obtain milk from a ruminant than from a pig; hence, the conversion of mammary glands of transgenic animals into bioreactors is more frequently attempted in these species (reviewed in Hennighausen, 1992; see Sections 6.4, 6.5 and 6.6). Nevertheless, due to greater experience in generating transgenic pigs, they have been used to establish a transgenic domestic animal system for targeting the synthesis of foreign protein to the mammary gland. Two laboratories reported the high-level synthesis of heterologous protein in the mammary gland of swine (Wall et al. 1991; Velander et al. 1992). In both cases the regulatory sequences controlling the gene of interest were derived from the mouse whey acidic protein (WAP) (Campbell et al. 1984). WAP gene constructs have been used successfully to direct synthesis of foreign proteins to the mammary glands of mice (Andres et al. 1988; Pittius et al. 1988). Wall et al. (1991) generated three transgenic pig lines expressing mouse WAP. Transgene expression was found only in the mammary glands up to a level of 1 g/1. However, three further transgenic lines showed dramatically reduced lactation caused by undifferentiated mammary gland tissue. The phenomenon of 'milklessness' has been previously observed in WAP transgenic mice (Burdon et al. 1991). There is evidence that early WAP expression in virgin pigs inhibits the functional development of the mammary gland (Shamay et al. 1992). In a second experiment a fusion gene consisting of the cDNA for human protein C (hPC) inserted in the first exon of the mouse WAP gene was inserted into the germ-line of pigs (Velander et al. 1992). hPC is a regulator of hemostasis, suggesting its potential therapeutic use for many disease states. Two transgenic pig lines showed highlevel expression of hPC (highest 1 g/1). Protein C produced in transgenic pig milk possessed anticoagulant activity equivalent to that of hPC derived from human plasma. 6.3.2.4 Production of human haemoglobin Transfusion-related diseases, combined with shortages and difficulties in long-term storage of blood, make an alternative to human erythrocytes as a source of haemoglobin desirable. The synthesis of human haemoglobin in the serum of transgenic pigs is an approach yielding

210

G. Brem and M. Muller

appropriately modified cell-free solutions of haemoglobin that can be used as a source of an oxygen-carrying erythrocyte substitute. Swanson et al. (1992) used a construct consisting of the LCR from the human /3-globin locus and two copies of the human a r and one copy of the human /3A-gene to obtain three transgenic pigs. All transgenic animals expressed the human genes at a copy-number-dependent level. They were healthy and nonanaemic, and grew at rates similar to those of nontransgenic control animals. Purification of the human haemoglobin from porcine proteins including porcine haemoglobin was accomplished by ion-exchange chromatography. The human haemoglobin derived from transgenic pigs and that from human serum exhibited similar oxygen equilibrium curves. 6.4 Gene transfer into sheep

Most gene transfer experiments in ruminants have been carried out with sheep. This is due mainly to the relatively short generation intervals of this species and to financial considerations. 6.4.1 Treatment of donors and recipients

The isolation of a sufficient number of embryos for microinjection requires the precise control of oestrus and ovulation of the donor animals. A variety of protocols are used to collect eggs from sheep, although no consensus has emerged on an optimal procedure. Sheep are seasonally anoestrus (as are goats); therefore, the effectiveness of superovulation varies with season. Synchronization is usually achieved by the intravaginal administration of progestagen depots (i.e., sponges) for 10-16 days. Superovulation is induced by administration of follicle cell-stimulating hormone (FSH) or PMSG during the late luteal phase of the oestrus (Table 6.14), followed by artificial or natural insemination of sheep 16-24 hr later. After a further 16-26 hr, the zygotes are isolated by obligatory surgical techniques. On average, superovulation of sheep yielded 7.2-10.5 zygotes per donor (Hammer et al. 1986; Rexroad & Pursel 1988; Halter et al 1993). Approximately 50% of the isolated embryos could be used for microinjection ofDNA. The pronuclei of fertilized ovine oocytes are visible under interference-contrast microscopy (Fig. 6.7) without prior centrifugation (Hammer et al. 1985), but because of the morphology and size of

Table 6.14. Collection of sheep zygotes: treatment of donors Breed Rambouillet

Merino Welsh Mountain, Scottish Blackface, Greyface, Chevriot Merino

Merino Blackface

Synchronization

Superovulation

Ovulation"

Insemination"

Embryo recovery6

Reference

Day 10 of oestrus cy- Twice daily 2.5 mg Onset of oestrus Mated or in utero 72 hr after sponge recle, progestapFSH; starting 3 with 0.2 ml tested moval, Ham's gen-impregnated days before sponge washed ram seF'°+10%FCS vag. sponges; 12 removal until 1 day men per horn days 6a-methylafter 17a-acetoxy progesterone, 60 mg Progestagen sponges 1,200-1,500 IUPMSG 50 /Lig GnRH 24 hr Intrauterine, 16 hr 24 hr later after sponge relater (endo33 hr after sponge moval scopic) removal 12-16 days progesta- Two equal doses of Onset of oestrus Served by at least Ovum culture medium gen sponges tested two rams equine 30 hr before sponge removal; FSH (1.75 or 2.15 mg)

Hammer et al. (1985)

14-day treatment 46 hr before sponge re- 24 hr after sponge 50/xl of diluted sewith progesteroneremoval, GnRH men into each moval; 12-hr interimpregnated intraapplication uterine horn, 24 vals 6.4/4.2/2.2 mg vag. sponges hr after GnRH FSH injection 12-14 days, intra24 hr after sponge Mated twice, 141500IU PMSG or 17 vag. sponges with removal, GnRH mg FSH; 6 injec15 hr after 40 mg fluogestone application tions every 12 hr, GnRH acetate decreasing amounts

16 hr after fertilization, DBS+5% sheep serum

Murray et al. (1989)

24-6 hr after mating, PBS + 1% NBCS

Halter et al. (1993)

"GnRH, gonadotropin-releasing hormone.

*NBCS, new-born calf serum.

Nancarrow etal. (1987) Simons et al. (1988)

212

G. Brem and M. Muller

Fig. 6.7. DNA microinjection into the pronucleus of an ovine zygote. the zygotes it is very difficult to control the success of a microinjection by observing the swelling of these pronuclei. Transfers of manipulated but noninjected zygotes showed that only 26% of the isolates developed in vitro to blastocysts and could be transferred to recipients. Fifty-two per cent of the transferred embryos resulted in new-born lambs. Microinjection of DNA dramatically reduced the embryonic development rate to 10% (Hammer et al. 1986). Rexroad and Wall (1987) found a 35% developmental rate of manipulated embryos being reduced by microinjection to 18.4%. Growing experience in handling ovine embryos and the use of more narrow injection pipettes increased the survival rate of injected zygotes to 5371% (Walton et al. 1987; Murray et al 1989). Rexroad et al. (1990) used co-culture with ovine oviduct cells to select the embryos surviving after microinjection. The developmental rate of control embryos was 96%, that of microinjected embryos 75%. The transfer to recipients resulted in normal development of 11.3% for the control embryos and 7.3% for the microinjected embryos with pregnancy rates of 50% and 41%, respectively. Walker et al. (1990) investigated the survival and developmental rate of microinjected zygotes in synthetic oviduct fluid medium. Despite a slightly lower survival rate there were no significant differences in the further development of 1- to 3-day-old embryos cultured in the synthetic medium or under in vivo culture conditions.

Large transgenie mammals

213

At first gene transfer experiments resulted in low embryonic survival and transgene integration rates with total efficiencies of 0.2%. Optimization of embryo manipulation and microinjection yielded, in total, efficiencies of up to 4% (Table 6.15) and integration rates of 26% (Rexroad et al. 1991). 6.4.2 Transgenic sheep 6.4.2.1 Gene constructs of the growth hormone cascade At least nine different constructs containing genes of the GH cascade have been used to generate transgenic sheep (Table 6.15). So far all experiments indicate that permanent synthesis of hormones altering growth performance is not tolerated by the organism. Expression of GH or GHRF gene controlled by different promoters (mouse or ovine metallothionein I promoters, mouse albumin promoter, mouse transferrin enhancer/promoter) resulted in increased serum levels of GH and IGF-1 (Rexroad et al. 1989, 1990, 1991; Rexroad & Pursel 1988; Nancarrow et al. 1988, 1991). However, none of the transgenic sheep showed a growth performance different from that of control animals. Moreover, most transgenic animals had severe health problems such as visceromegalia (Nancarrow et al. 1988, 1991), diabetes (Rexroad et al. 1991) and pneumonia and therefore died at an early age. These results suggest that alternative strategies for expressing growth-related genes are required to modulate growth in sheep. 6.4.2.2 Wool production and metabolic pathways Gene transfer experiments attempting to influence wool production utilize genes that increase cysteine biosynthesis in sheep (for review see Rogers 1990; Ward & Nancarrow 1991). It is well established that cysteine is a limiting factor in wool synthesis. It is impossible, however, to supply extra dietary cysteine. Cys E (serine transacetylase) and cys Klcys M (Oacetylserine sulfhydrylase) are two bacterial genes coding for enzymes that catalyse the synthesis of cysteine from sulfide and serine. Ward et al. (1991) successfully transferred and expressed these prokaryotic genes in mice. Rogers et al. (1991) introduced a gene construct consisting of both genes arranged in a tandem array into the germ-line of sheep and observed constitutive expression of both enzymes. The effects on wool production in these animals have not yet been published.

Table 6.15. Gene transfer into sheep

Construct mMT-I-bGH 0MT-0GH mMT-I-bGH MT-hGRF pMK BLG-FIX BLG-a^AT MtsGH5 MtsGH 9 MtsGH 10 BLG-AAT Trf-GH Alb-GRF cysE—cysM pPBLac-FVIII pBpLac-FVIIIMtl pPMt-FVIII mMT-I-hGH Total

Embryos injected and transferred (n)

Offspring (incl. foetuses)

Integration

Efficiency

Expression

AJ111V1V11V T

n

%

n

%

2 1 2 9 1 4 1 4 4 1 5 11 4 33 2 — 1 1 86

5.2 3.7 4.2 14.3 2.6 7.0 7.1 4.9 17.4

0.28 0.23 0.24 2.07 0.65 1.3 2.04 3.7 1.0

4.5 26.2 25.0 17.7 1.0 — 4.8 1.3 8.1

1.1 4.5 2.3 4.1 0.42 — 1.7 0.10 1.1

711 436 842 435 155 307 49 1,079 409

38 27 47 63 38 57 14 81 23

5.3 6.2 5.6 14.5 24.5 18.6 28.6 7.5 5.6

439 247 171 803 475 54 57 1,032 7,701

113 42 16 186 192 17 21 73 1,048

25.7 17 9.4 23.2 40.4 31.5 36.8 7.1 13.6

(%)

n

%

Reference

2/2 0/1 2/2 1/7 — 2/2 — 0/4

100 — 100 14 — 100 — 0

0/1 3/3 3/11 2/4 — — — — 0/1 15/40

0 100 27 50 — — — — — 38

Pursel etal. (1987) Nancarrow et al. (1987) Rexroad and Pursel (1988) — Simons et al. (1988) Clark et al. (1989a,b) — Murray et al. (1989) Murray et al. (1989) Murray et al. (1989) Wright et al. (1991) Rexroad et al. (1991) Rexroad et al. (1991) Rogers et al. (1991) Halter et al. (1993) Halter et al. (1993) Halter et al. (1993) —

Large transgenic mammals

215

Experiments aiming to change the susceptibility of sheep to ketonuria by the establishment of novel metabolic pathways by transgenic means (for review see Ward & Nancarrow 1991) were carried out by Ward et al. (1991). Excess acetate in the rumen cannot be utilized for gluconeogenesis and gets converted to ketone. A biochemical pathway capable of using acetone for glucose synthesis is the glyoxylate cycle via the enzymes malate synthase and isocitrate lyase. In a model experiment, transgenic mice have been shown to express both enzymes functionally in liver, kidney and intestinal tissue. The next step will be to test these gene constructs in transgenic sheep. Other bacterial biochemical pathways considered for establishment in transgenic farm animals include the synthesis of essential amino acids. Before these transgene experiments become a reality, a variety of technical difficulties have to be solved. The transgene expression has to be tightly controlled, so that the transferred genes do not interfere with any physiological processes. Therefore, constitutive gene expression is not practicable for most purposes. Most metabolic pathways are composed of multiple components, so that their establishment in transgenic animals requires the transfer of several gene constructs. The present state of the art does not allow the transfer of such complex gene constructs. A step towards such experiments would require greater availability of ES cells and improved techniques for handling large DNA fragments. 6.4.2.3 Gene farming A variety of therapeutic or diagnostic proteins cannot be produced in sufficient quantity or quality by conventional methods, including prokaryotic or eukaryotic expression systems and extraction from mammalian sources. High-level expression of recombinant protein in micro-organisms has been successfully used in many cases. However, some proteins cannot be synthesized in prokaryotic expression systems due to the lack of appropriate purification procedures and more often to a deficiency of the post-translational modifications required for protein stability and/or function. Large-scale eukaryotic expression systems are technically complex, elaborate and cost intensive. As an alternative to cell culture systems, the production of large quantities of proteins in transgenic animals is appealing because of safety considerations, the high production capability of the transgenic organism, the comparatively low operating costs and the potentially un-

216

G. Brem and M. Muller

limited expansion of the producer animal by conventional breeding (Lovell-Badge 1985; Lathe et al. 1986; Mercier 1986; Church 1987; Clark et al. 1987; Hennighausen 1990; Wilmut et al. 1991; Brem et al. 1993). The most appropriate organ for the expression of large amounts of foreign protein is the mammary gland because of easy access to the synthesized protein. In addition, the mammary gland has an enormous physiological potential for synthesizing proteins. The production of certain recombinant proteins might require their expression in cell types and organs other than the mammary gland because of their posttranslational processing being restricted to a certain cell type. An example is the synthesis of modified antibodies in B-lymphocytes and the subsequent purification from blood of the transgenic animal (Lo et al. 1991; Weidle et a/. 1991). The feasibility of mammary-gland-specific transgene expression has been demonstrated in numerous experiments in several species (reviewed in Hennighausen 1992). Eighty per cent of farm animals' milk consists of six main proteins (four caseins, /3-lactoglobulin and aalbumin) being secreted from epithelial cells of the mammary gland under the control of several hormones (Clark et al. 1992). The main milk protein of sheep is /3-lactoglobulin (BLG) with an average concentration of 3-5 g/1. The BLG mRNA represents approximately 5% of total mRNA of mammary gland cells (Clark et al. 1989a). Ovine BLG regulatory sequences promote tissue-specific high-level transgene expression in mouse models. The transgene expression with BLG promoter constructs was tightly controlled and resembled the endogenous mouse /3-casein gene expression profile (Simons et al. 1987; Harris et al 1991). Human arantitrypsin (hajAT) is a glycoprotein which is normally present at 2 g/1 in plasma. Genetic deficiencies in circulating concentrations of ha j AT are a common lethal hereditary disorder affecting males and leading to life-threatening emphysema. The annual demand for replacement therapy using human-plasma-derived axAT is more than 4,000 kg. Therefore, recombinant-DNA-derived sources are highly desirable. Archibald et al. (1990) reported the production of biologically active ha,AT at yields of up to 7 mg/ml in the milk of transgenic mice expressing a gene construct containing the sheep BLG promoter fused to ha]AT coding sequences. To form the basis for a manufacturing process by transgenic means, a BLG-hajAT gene construct was used to generate five transgenic

Large trans genie mammals

217

sheep (four female, one male) (Simons et al. 1988; Wright et al. 1991). Three of the transgenic females expressed the human protein, all at levels greater than 1 g/1. In one case, initial levels exceeded 60 g/1 and stabilized at 35 g/1 during the lactation period. hajTA purified from sheep milk was fully glycosylated and exhibited biological activity indistinguishable from human-plasma-derived protein. In the highlevel expressing animal, the recombinant protein represented nearly 50% of total milk protein, considerably exceeding the production levels of ajAT obtained in other expression systems. In a second experiment Simons et al. (1988) generated five transgenic sheep harbouring gene constructs consisting of the ovine BLG promoter fused to the human blood coagulating factor IX. Two female transgenic animals secreted the foreign protein into their milk. However, the transgene transcription was approximately 1,000-fold lower than endogenous BLG expression (Clark et al. 1989b). 6.5 Gene transfer into goats

The reproductive and milk production data for goats are very similar to those for sheep. Therefore, it is surprising that the first gene transfer experiments in the goat have been initiated only recently (Ebert et al. 1991; Ebert & Schindler 1993). This is probably because in most countries investigating gene transfer into farm animals, goat production is traditionally not of great importance. 6.5.1 Treatment of donors and recipients

Several protocols are currently used to isolate oocytes from goats. Since goats are seasonally anoestrus, the efficiency of collecting and transferring embryos varies with season. The oestrus of donors is synchronized by administration of Norgestomet ear implants (Pendleton et al. 1986; Selgrath et al. 1990; Ebert et al. 1991). Alternatively, progesterone-containing vaginal sponges or progesterone injection is used for donor synchronization (Armstrong et al. 1987; Nuti et al. 1987; Cameron et al. 1988). The progesterone treatment is carried out for 11-18 days. Superovulation is induced by either PMSG or FSH, ideally 10-48 hr after the progestagen treatment has ended. The induction scheme follows mainly that of sheep (see Table 6.14). Luteolysis can be increased by treatment with progesterone F 2a 36 hr before induction of superovulation. Further optimization of synchronization

218

G. Brem and M. Mtiller Table 6.16. Embryo collection and transfer in a transgenic goat programme n Donors Ovulations Recovered Fertilized Recipients Embryos transferred Pregnant recipients Live young Transgenic kids

269

2,356 1,429 1,058

190 782 107 169 12

%

X

8.8 5.3 3.9 4.1*

1.6 1.6

60.6 74.4(44.9)* 56 c 22 "

l.\

a

Percentage of embryos collected per ovulation. * Embryos per recipient. c Pregnancy rate (oviduct and uterus transfer). d Survival rate of embryos transferred. e Integration rate (transgenic per live young). Source: Ebert and Schindler (1993).

and ovulation can be achieved by administration of gonadotropinreleasing hormone (GnRH) 24-6 hr after the prostaglandin treatment is ended. When signs of oestrus are observed, the donor animals are mated several times to fertile males or inseminated artificially over a 2-day period. Embryos are recovered surgically 48-72 hr after the progesterone treatment is finished. On average, superovulation yielded 7.7-19.4 zygotes per goat (Armstrong et al. 1987; Fabricant et al. 1987; Selgrath et al. 1990). The best results were obtained using FSH. The pronuclei of fertilized oocytes are visualized either under Nomarski's interference-contrast microscopy or after centrifugation (7,000 g, 2 min). Microscopy is sometimes difficult owing to large lipoid granula in the cytoplasm. The preparation of recipients is similar to that of donor animals. Embryos are transferred surgically into one oviduct of synchronized recipients. Table 6.16 summarises the results of the embryo transfer programme of Ebert and Schindler (1993). Embryo and gene transfer experiments in goats resulted in overall pregnancy rates of 56% and a total gene transfer efficiency (transgenic goats per microinjected embryos transferred) of 1.5% (Table 6.17).

Table 6.17. Gene transfer into goats

Construct mMTI-oGH WAP-LAtPA

Offspring (inci. foetuses)

Zygotes and embryos injected and transferred (n)

n

%

n

153 203

9 29

5.9 14

0 2

Integration

Expression

%

Efficiency (%)

n

%

Reference

0.0 6.9

1.0

1/2

50

Fabricant et al. (1987) Ebert et al. (1991)

220

G. Brem and M. Mtiller

6.5.2 Transgenic goats So far only a few gene transfer experiments involving goats have been carried out. The experiments of Fabricant et al. (1987) using GH gene constructs failed to generate transgenic goats. The first successful production of transgenic goats expressing a heterologous protein in their milk was reported by Ebert et al. (1991). The goat was chosen to establish a commercial prototype for the large-scale manufacture of high-value proteins in the transgenic mammary gland for several reasons: dairy goats produce large quantities of milk (on average 4 I/day); goats have gestation and development periods of moderate length (5 and 8 months, respectively); and the biochemistry of goat milk has been well described. Tissue plasminogen activator (tPA) plays a key regulatory role in the orderly progression of wound-healing processes and has a potential clinical value in the treatment of coronary thromboses. A glycosylation variant of human tissue plasminogen activator (longer-acting tPA or LAtPA), controlled by the murine WAP promoter was expressed in the mammary gland of a transgenic dairy goat. Milk was obtained upon parturition of the goat and contained enzymatically active LAtPA at a concentration of 3 )ug/ml. On average the transgenic goat produced 10-15 mg foreign protein per day during the lactation period. Denman et al. (1991) developed a purification protocol for LAtPA from the milk of the transgenic goat by combining acid fractionation, hydrophobic interaction chromatography and immunoaffinity chromatography. The procedure resulted in 8,700-fold purification with a cumulative recovery of 25%. The specific activity of the purified transgenic protein was 6.1 x 105 U/mg. This was approximately 84% of the value observed for the recombinant protein synthesized in an in vitro expression system based on mouse cells. Although the transgenic protein was biologically active, there were significant differences in the oligosaccharide structures between in vitro expression-system-derived and transgenic LAtPA. As with any recombinant protein synthesized in different hosts, the consequences of structural differences on therapeutic utility require evaluation on a case-by-case basis. Another gene construct consisting of LAtPA regulated by a /3-casein promoter was used to produce a transgenic goat expressing the protein at 2-3 mg/ml (Ebert & Schindler 1993). Unfortunately, the lactation of this goat stopped for unknown reasons.

Large trans genie mammals

221

6.6 Gene transfer into cattle

Although cattle are the most important species in agricultural production, they have rarely been used for gene transfer experiments. This is due to the nature of their reproduction (i.e., a long generation interval of 2-3 years with normally only one offspring per gestation), to difficulties in collecting and transferring early embryo stages and to the generally high cost of cattle husbandry. The development of procedures for in vitro production and in vitro culture of bovine embryos has dramatically reduced the technical problems.

6.6.1 Treatment of donors and recipients

Preparation of donor and recipient animals in gene transfer programmes is carried out according to procedures developed in commercial embryo transfer programmes. If necessary the oestrus cycles of donors and recipients are synchronised by PGF2a administration. The gonadotropic substances used for the induction of superovulation include PMSG, FSH or a combination of FSH and luteinizing hormone. PMSG has the advantage of a long half-life (40-120 hr) and hence need be administered only once between days 11 and 14 after the onset of oestrus (dosage of 2,000-3,000 IU). Ovulation is induced by administration of prostaglandin 2 days after the PMSG injection. After 48-64 hr, the donors are inseminated. The most convenient time for embryo collection is 75-96 hr after the prostaglandin treatment. Embryo isolation from oviducts is usually achieved by one of several procedures: upon slaughtering (Biery et al. 1988), surgical flushing of the oviducts (Loskutoff et al. 1986; McEvoy & Greenan 1990) or removal of the ovary and oviduct by castration (Roschlau et al. 1989). The collection of embryos upon slaughtering is usually the method of choice, since the other procedures are too laborious and cost intensive. On average, 12 oocytes can be collected per donor animal, of which half are suitable for microinjection. The pronuclei of bovine embryos have to be visualized by centrifugation. Controlling the microinjection is frequently difficult because the pronuclei are often blurred even after centrifugation. Lohse et al. (1985) injected a thymidine kinase construct into bovine embryos to estimate the success rate of the microinjection. Approximately 30% of the injected embryos showed transient expression of thymidine kinase activity and therefore demonstrated successful microinjection.

222

G. Brem and M. Muller

The effects of manipulation of bovine embryos (i.e., centrifugation, microinjection and culturing) on their survival rates and further development have been investigated by several laboratories (McEvoy et al. 1987; Hawk et al. 1989; Peura et al. 1993). They found that the gene injection alone seemed responsible for decreased embryonic development, whereas treatments required for microinjection (centrifugation and microscopic evaluation) and in vivo culture (e.g., in rabbit oviducts) did not have detrimental effects. Reichenbach et al. (1991) have tested various in vivo culture systems (temporary culture in oviducts) as well as in vitro cultures (see Section 6.6.2; Table 6.18). On average 20% of the microinjected embryos showed further development after in vivo culture in bovine, porcine or rabbit oviducts and subsequent transfer into recipients. In vitro culture with bovine epithelial cells resulted in a developmental rate of 15%. Jura et al. (1992) achieved in separate experiments with in vivo and in vitro culture systems 11 and 6 pregnancies by transferring 63 and 39 embryos, respectively. 6.6.2 In vitro production of bovine zygotes A scheme for the in vitro production of zygotes suitable for microinjection from ovaries of slaughtered cows has been developed in our laboratory and is depicted in Fig. 6.8. A similar procedure has been described by other laboratories as well. On average 15 intact cumulusoocyte complexes per ovary were collected and subsequently matured in modified tissue culture medium 199 (TCM 199) (Pavlok et al. 1988) supplemented with heat-inactivated serum of cows in oestrus. The in vitro fertilization was performed with cryo-conserved sperm (106 sperm cells per millilitre) (Parrish et al. 1986) in fertilization medium (Ball et al. 1983). The fertilized oocytes were co-cultured with cumulus-derived granulosa cells in modified TCM 199 for a further 90 hr (Berg & Brem 1989). Thirty-two per cent of the oocytes matured and fertilized in vitro (fertilization rate of 68%) showed further development to the morulae/blastocysts and therefore reached the developmental stage, providing the possibility of nonsurgical transfer into recipients (Reichenbach et al. 1992). In total, 74% of the cultured embryos overcame the developmental block at the 8- to 16-cell stage. Initial experiments used 311 m vitro matured and fertilized oocytes for microinjection (Fig. 6.9). Further development was shown by 45 injected zygotes, which were transferred nonsurgically into 29 recipients. This resulted in 9 pregnancies (pregnancy rate of 31%) (G. Brem

Table 6.18. In vivo and in vitro development of microinjected bovine zygotes Zygotes and 2-cell embryos transferred (Dl D7)

Recipients Culture system In vivo Bovine Rabbit Swine Total

n 8 3 4 2 3 1 21

In vitro —

Embryos recovered (D5-D1)

Embryos developing > 1 cell division

Embryos developing i norulae and blastocysts

n (a)

n (b)

(b/a)

/o

n (c)

(c/b)

/o

n (d)

(d/b)

MI C MI C MI C

MI/C

183 66 82 41 79 29 480

93 48 46 31 22 12 252

51 73 56 76 28 41 53

24 27 28 21 7 5 112

26 56 61 68 32 42 44

17 14 13 14 3 2 63

18 29 28 45 14 17 25

MI C

123 71

123 71

100 100

63 43

51 61

18 21

15 30

type

Abbreviations: D, day; MI, microinjected embryos; C, control group (not microinjected). Source: Reichenbach et al. (1991).

/o

Slaughtered heifers and cows

Recovery of ovaries

Transport of ovaries in a thermos flask

( 5% CO;, 100% RH 39°C 24 h

I

39°C Straw with frozen sperm

Puncture of follicles (0 2-6 mm) and recovery of oocytes by suction

^

J

Maturation of oocytes , in an incubator

Swim-up treatment of sperm

Oocytes and sperm in fertilization medium

Selection and washing of oocytes

Transfer of oocytes to tissue culture medium (supplemented with serum) under paraffin

— Motile sperm after swim-up

Maturated oocyte

Cumulus cells In vitro fertilization (24 h)

Oocyte

5% C02 5% 02 100% RH 39°C 90 h In vitro culture

non suitable

degenerated embryos Blastocyst Remove cumulus cells from embryos

In vitro culture Selection of embryos for transfer 7 days after recovery of oocytes Non-surgical transfer of embryos to recipients and birth of calves

Fig. 6.8. Scheme for the in vitro production of bovine zygotes.

226

G. Brem and M. Miiller

Fig. 6.9. DNA microinjection into the pronucleus of a bovine zygote. et al. unpublished data). Meanwhile, similar results (Table 6.19) have been reported by Massey (1990), Krimpenfort et al. (1991) and Hill et al. (1992). In vitro production of bovine embryos provides a cheap and simple source of oocytes and early embryonic stages for gene transfer experiments. In addition, embryos produced in vitro show less variability in the time schedule of their developmental stages than embryos collected from superovulated donors and are therefore notably easier to manipulate successfully. 6.6.3 Transgenic cattle So far only a few experiments resulting in the generation of transgenic cattle have been published. As yet there are no data available concerning the expression of transferred genes in these animals. In an early experiment Biery et al. (1988) injected the bacterial chloramphenicol transferase gene controlled by the Rous sarcoma virus promoter (RSV-CAT) into the pronuclei of embryos and subsequently cultured them in vivo in ovine oviducts. Twenty-one per cent of the injected embryos developed to the morula/blastocyst stage and were nonsurgically transferred into synchronized recipients. The analysis of 79 foetuses on day 60 of the pregnancy resulted in 4 transgenic foetuses. Roschlau et al. (1989) reported the generation of a transgenic calf by

Table 6.19. Gene transfer into cattle Offspring (incl. foetuses)

Zygotes and embryos Construct a-Fetoprotein RSV-CAT pMMTV-bGH ASK-HER asX cas-hLF ASKHER 202 aASK-IGF-I MMTV-IGF-I 733 ASK-GF-I Total

iniected find C H l

iectedandIllJvvlvU

Integration

transferred (n)

n

%

n

%

852 819 201 1,704 1,154 4,150 7,810 1,336 6,070 24,096

79 14 79 21 53 67 20 53 386

9.6 7.0 4.6 1.8 1.5 0.9 1.5 0.9 1.7

4 4 1 1 2 1 3 2 1 19

5.1 7.1 6.3 9.5 1.9 4.5 10 1.9 3.9

Pfficiencv (%)

Reference

0.5 0.49 0.5 0.3 0.2 0.02 0.04 0.01 0.02 0.08

Churchetal. (1986) Biery et al (1988) Roschlau et al. (1989) Massey (1990) Krimpenfort et al. (1991) Hill etal. (1992) Hill et al (1992) Hill et al (1992) Hill et al (1992)

228

G. Brem and M. Muller

O