v
Preface
What is Biotechnology Today? What is biotechnology today? More specifically, as one invites chapters for an Annual Review of Biotechnology, what should be the range of topics? This is quite a complicated question. The world of biotechnology has changed profoundly with the advent of the Human Genome Project (HGP). The HGP was the first example of global discovery science – the attempt to take a complex biological object, the genome, and define all of its elements – the DNA sequences of the 24 different human chromosomes. For the first time, biologists had a genetics parts list for the human, namely an enumeration of all (most) of our genes (and by translation, our proteins). This raised the possibility for global studies in which all (most) of the elements of a system could be studied quantitatively or in terms of their interactions. Accordingly, genomic discovery science led quite naturally to two other applications of discovery science – the transcriptome, a quantitative measure of all (many) the mRNAs present in a particular cell, organ or organism and the proteome, a similar quantitative measure of all (many) the proteins present in a particular biological system. In a similar vein, others are studying metabolites, the metabolome, phenotypes, the phenome, etc. This ability to carry out global discovery science for the multiplicity of different types of biological information raised the possibility this information could be used to actually understand biology through an approach termed systems biology. What is systems biology? In its simplest terms, systems biology is the identification of the elements in a system and an analysis of their interactions while the system is functioning so as to understand the systems or emergent properties of the system. Systems biology probably started with the integrative physiologists of the early 20th century who were interested in homeostasis, and the effects of environmental perturbations such as hormones on homeostasis. The systems biology of today is quite different in that it can interrogate many different types of biological information in a global manner. Systems biology has just recently been employed in this manner and it has a number of interesting features. Systems biology is global, quantitative, hypothesis-driven, iterative, integrative (different types of biological information must be integrated together to understand systems), dynamic (systems must be studied as they execute their developmental or physiological functions), and multi-scale (systems extend from a few molecules carrying out a particular function, such as the metabolism of a sugar, to biological networks in cells to cells and organs and organisms; thus, the physical scale across which biological systems operate is enormous). It is important to point out that high-throughput global technologies must be invented and improved to capture the many different types of information. Likewise, it is essential to develop the computational, mathematical, and statistical tools for capturing, storing, analyzing, integrating, modeling, and finally dispersing biological information.
vi What is the context from which this global systems biology emerged? It has all happened in the last 5–10 years. Several factors have been important. (1) The completion of the HGP led to the genetics parts list, discovery science, and it initiated the development of high-throughput technical platforms such as automated DNA sequencing. These high-throughput platforms are, of course, essential to the capture of large and global data sets, which constitute the foundation of systems biology. (2) Biology has come to recognize the power of cross-disciplinary biology because of the technologies and computational tools that remain to be developed. Thus, computer scientists, mathematicians, and statisticians, as well as physicists, must develop the computational and modeling tools; the engineers, physicists, chemists, and biologists must develop the new technologies for high-throughput biology. (3) The internet has given us the capacity to store and dispense large amounts of information. (4) Finally, the idea that biology is an informational science has emerged – and this is the foundation for thinking about systems biology. Let us consider biological information in this context. Biological information falls into two categories – the digital information encoded in the genome, and the environmental signals that impinge of the digital genome to initiate developmental and physiological responses. The digital genome has two major types of information – the genes encode proteins, the molecular machines of life, and the cis-control elements that specify, in conjunction with their cognate transcription factors, the behavior of individual genes (when, where, and how much mRNA is expressed) and, the cis elements establish the linkages and architectures of the gene regulatory networks – those networks of transcription factor genes which control the peripheral batteries of genes executing developmental and physiological functions on the one hand, and which are triggered by the protein networks of signal transduction on the other hand. Thus, the major challenge of systems biology is defining and understanding the interactions of the protein and gene regulatory networks. Finally, biological information is hierarchal in its expression starting at the DNA and moving outward to ecological systems – that is, it goes from DNA to mRNA to protein to protein interactions and biomodules to protein networks to cells to organs to organisms to populations of organisms and finally to ecologies. Environmental signals modify the initial digital input of information at each of these successive informational levels – hence, global data sets must be gathered on as many of these informational levels as possible and, ultimately, these data sets must be integrated to begin understanding how the corresponding system works. Thus, systems biology is global and all encompassing with regard to its need to gather biological information. In this context, the contents of Volume 10 encompass one or more of the various types of biological information that constitute the foundations of systems biology. Indeed, one chapter is on a systems approach to human health. Several are on DNA mediated technologies and five are on varying aspects of proteomics. The remainders of the papers are on several of the enormous ranges of policy questions emerging from modern biology. We can begin to glimpse from these chapters the enormous challenges facing modern biology in general and fascinating promises (poised against a very challenging reality) that systems biology has made for providing an integrated picture of biological complexity. The fascinating question is how long will it take
vii systems biology to begin filling its enormous promise to understand biological complexity, for in time it surely will begin to comprehend and even reengineer this complexity.
Leroy Hood, M.D., Ph.D. President The Institute for Systems Biology 1441 North 34th Street, Seattle, WA 98103-8904 Phone: 1-(206) 732-1201; Fax: 1-(206) 732-1299
[email protected] http://www.systemsbiology.org/
ix
EDITORIAL BOARD FOR VOLUME 10 CHIEF EDITOR Dr. M. Raafat El-Gewely Department of Molecular Biotechnology Institute of Medical Biology University of Tromsø 9037 Tromsø, Norway Phone: þ 47-77 64 46 54 Fax: þ 47-77 64 53 50 E-mail:
[email protected] EDITORS Dr. MaryAnn Foote Associate Director Medical Writing Department Amgen, Thousand Oaks, CA 91320-1879, USA Phone: þ 1-805-447-4925 Fax: þ 1-805-498-5593 E-mail:
[email protected] Dr. Guido Krupp Director & Founder artus GmbH Koenigstr. 4a D-22767 Hamburg, Germany Phone: þ 49-40-41 364 783 Fax: þ 49-40-41 364 720 E-mail:
[email protected] website: http://www.artus-biotech.com ASSOCIATE EDITORS Dr. Marin Berovic Department of Chemical and Biochemical Engineering University of Ljubljana Hajdrihova 19, Ljubljana Slovenia E-mail:
[email protected] Dr. Thomas M.S. Chang Artificial Cells & Organs Research Centre McGill University
3655 Drummond St., Room 1005 Montreal, Quebec, Canada H3G 1Y6 Phone: þ 1-514-398-3512 Fax: þ 1-514-398-4983 E-mail:
[email protected] Dr. Thomas T. Chen Department Molecular and Cellular Biology University of Connecticut 91 North Eagleville Rd, Unit 3125 Storrs, Connecticut 06269-3149, USA Phone: 1-860-486-5481 Fax: 1-860-486-5005 E-mail:
[email protected] Dr. Frank Desiere Nestle´ Research Centre, P.O. Box 44, CH-1000 Lausanne 26 Switzerland E-mail:
[email protected] Prof. Franco Felici Dipartimento di Scienze Microbiologiche, Genetiche e Molecolari Universita` di Messina Salita Sperone, 31 98166 Messina, Italy Phone: þ 39 090 6765197 Fax: þ 39 090 392733 E-mail:
[email protected] Dr. Leodevico L. Ilag Xerion Pharmaceuticals AG Fraunhoferstrasse 9 82152 Martinsried Germany Phone: þ 49 89 86307 201 Fax: þ 49 89 86307 222 E-mail:
[email protected] x Dr. Kuniyo Inouye Laboratory of Enzyme Chemistry Division of Applied Life Sciences Graduate School of Agriculture Kyoto University Sakyo-ku, Kyoto 606-8502, Japan Phone: þ 81-75-753 6266 Fax: þ 81-75-753 6265 E-mail:
[email protected] Dr. Alfons Lawen Senior Lecturer Monash University, Clayton Campus Department of Biochemistry and Molecular Biology Room 312, Building 13D Clayton, Victoria 3800 Phone: þ 61-3-9905 3711 Fax: þ 61-3-9905 4655 E-mail:
[email protected] Dr. Jocelyn H. Ng Hirsch-Gereuth-strasse 56 81369 Munich Germany Phone: þ 49 89 78018945 E-mail:
[email protected] Dr. Eric Olson Program Executive, Strategic Development Vertex Pharmaceuticals, Inc 130 Waverly Street Cambridge, MA 02139, USA Phone: þ 1-617-444-6917 E-mail:
[email protected] Dr. Steffen B. Petersen Biostructure and Protein Engineering Laboratory Department of Biotechnology University of Aalborg Sohngaardsholmsvej 57 DK-9000 Aalborg Denmark Phone: 45-9-635 8469 Fax: 45-9-814 2555 E-mail: steff
[email protected] Prof. Vincenzo Romano-Spica Professor of Hygiene University Institute of Motor Science, IUSM Piazza Lauro e Bosis 15, 00194 Rome, Italy Phone/Fax: þ 39-06-36733247 E-mail:
[email protected] xv
Contents Preface Editorial Board List of contributors
v ix xi
Rapid translation system: A novel cell-free way from gene to protein Michael Hoffmann, Cordula Nemetz, Kairat Madin and Bernd Buchberger
1
Protein expression and refolding – A practical guide to getting the most out of inclusion bodies Lisa D. Cabrita and Steve Bottomley Towards a systems biology understanding of human health: Interplay between genotype, environment and nutrition Frank Desiere Public health issues related with the consumption of food obtained from genetically modified organisms Andrea Paparini and Vincenzo Romano-Spica p75 Neurotrophin receptor signaling in the nervous system Yuiko Hasegawa, Satoru Yamagishi, Masashi Fujitani and Toshihide Yamashita
31
51
85
123
Phage display for epitope determination: A paradigm for identifying receptor–ligand interactions Merrill J. Rowley, Karen O’Connor and Lakshmi Wijeyewickrema 151 DNA vaccines and their application against parasites – promise, limitations and potential solutions Peter M. Smooker, Adam Rainczuk, Nicholas Kennedy and Terry W. Spithill 189 Drug-induced and antibody-mediated pure red cell aplasia: A review of literature and current knowledge Ralph Smalling, MaryAnn Foote, Graham Molineux, Steven J. Swanson and Steve Elliott
237
Using the biologic license application or new drug application as a basis for the common technical document MaryAnn Foote Guidelines and policies for medical writers in the biotech industry: An update on the controversy MaryAnn Foote
259
Radioimmunotherapy of non-Hodgkin’s lymphoma: Clinical development of the Zevalin regimen Charles P. Theuer, Bryan R. Leigh, Pratik S. Multani, Roberta S. Allen and Bertrand C. Liang
265
251
xvi Biosimulation software is changing research Richard L.X. Ho and Lenore Teresa Bartsell
297
Index of authors
303
Keyword index
305
1
Rapid translation system: A novel cell-free way from gene to protein Michael Hoffmann*, Cordula Nemetz, Kairat Madin, and Bernd Buchberger Roche Diagnostics GmbH, Nonnenwald 2, 82372 Penzberg, Germany Abstract. Proteome research has recently been stimulated by important technological advances in the field of recombinant protein expression. One major breakthrough was the development of a new generation of cell-free transcription/translation systems. The open and flexible character of these systems allows direct control over expression conditions via the addition of supplements to the expression reaction. The possibility of working with linear expression templates instead of cloned plasmids and the ease of downstream processing, circumventing the need for cell-lysis, makes them ideally suited for high-throughput screening applications. Among these novel cell-free systems, the Rapid Translation System (RTS) developed by Roche is the first one that is scalable from micrograms to milligrams of protein. This review describes the basic principles of RTS which differentiate it from traditional in vitro expression technologies, starting from template generation to high-end applications like labeling for structural biology research. Recent results obtained by RTS users from different institutions are presented to illustrate each step of a novel cellfree protein expression workflow and its benefits compared to traditional cell-based expression. Keywords: in vitro translation, cell-free expression, template optimization, solubility, protein– protein interactions, high-throughput screening, scale-up, wheat germ, E. coli lysate.
1. Introduction to scalable cell-free protein expression Most of our knowledge about the molecular basis of protein synthesis including the genetic code, mRNA, ribosome functions, protein factors involved in translation, initiation and control of translation, co-translational protein folding, etc. was originally derived from in vitro studies carried out with cell-free expression systems. Now, in addition to their contributions to basic science, such systems offer a broad range of new technological applications. Since the first experiments in the early 1960s, cell-free protein biosynthesis has been improved continuously. An important step in this was the method of preparing S30 bacterial extracts and the introduction of exogenous messages into the system by Nirenberg and Matthaei [1]. Later, the coupled in vitro transcription–translation was developed by adding exogenous DNA as a template for the expression reaction by Zubay [2]. Owing to the simplicity of preparation, the Zubay bacterial extracts became very popular. However, a major shortcoming of most test-tube translation and transcription–translation systems remained their limited lifetime and, as a consequence, their low yield of expressed proteins. A major experimental breakthrough addressing this insufficiency was the invention of the continuous exchange cell-free (CECF) translation and coupled transcription–translation systems by Spirin in 1988 [3]. *Corresponding author: E-mail: michael.hoff
[email protected] BIOTECHNOLOGY ANNUAL REVIEW VOLUME 10 ISSN: 1387-2656 DOI: 10.1016/S1387-2656(04)10001-X
ß 2004 ELSEVIER B.V. ALL RIGHTS RESERVED
2 The principle of CECF relies on passive diffusion via a semipermeable membrane. Instead of incubating the reaction mixture in a fixed volume of a test-tube, the reaction is performed under conditions where the reaction products (Pi, NMPs and polypeptides) are continuously removed from, and the consumable substrates (amino acids, NTPs and energy-regenerating compounds) are permanently supplied to the reaction chamber. This setup allows protein expression to continue for up to 24 h with maximized protein yields, because the accumulation of inhibitory reaction by-products or the exhaustion of substrates is avoided. This contrasts the situation in in vitro protein synthesis reactions performed in batch mode (without continuous removal and supply), where a plateau is reached after approximately 2 to 4 h. The CECF-principle can be implemented by using a device containing two chambers that are separated by a semi-permeable membrane, allowing continuous supply of substrates and removal of inhibitory by-products from the reaction chamber (Fig. 1). Only the enzymatic machinery required for coupled transcription and translation and the expressed protein remain in the reaction compartment permanently. Using the CECF format, protein yields of several milligrams per milliliter have been reached [4]. An optimization of the energy regeneration system and an enhanced bacterial lysate helped further to increase the productivity [5]. Roche Applied Science has introduced commercial versions (RTS 500 E. coli HY Kit, RTS 9000 E. coli HY Kit and RTS 500 ProteoMaster E. coli HY Kit) of the CECF format based on E. coli extracts. In their RTS 500 and RTS 9000 reaction devices, up to 5 mg of an individual protein can be synthesized per ml in 24 h. All RTS reactions can be conveniently carried out in
protein DNA template
membrane
supply
supply
reaction change inhibitory by-products
feeding chamber
feeding chamber
membrane
Fig. 1. Schematic illustration of the continuous-exchange cell-free (CECF) reaction principle.
3 the RTS ProteoMaster Instrument under reproducible conditions, ensuring that the CECF process is optimally supported by shaking and temperature control. The workflow leading from a gene of interest to its preparative expression in the Rapid Translation System is outlined in the next chapter and illustrated with examples. 2. Generation of optimized expression templates Similar to any other expression system, successful cell-free expression of a DNA encoding a specific protein of interest depends on a correct starting material, i.e., a high-quality cDNA library or an isolated cDNA clone, sequenced and characterized and thereby shown to be free of frameshifts or stop codons. This basic condition met, one has to additionally keep in mind that (1) it often makes sense to specifically adapt the coding sequence to the expression system in question, and (2) regulatory elements for transcription and translation initiation and termination have to be added up- and downstream of the coding sequence. 2.1 Rational cDNA design While it is obvious that the addition of promoter and terminator sequences as well as a ribosomal binding site are absolutely essential to start and stop an E. coli-based expression reaction, sequence optimization is often considered as a recommended ‘‘nice-to-do’’, but maybe a dispensable option. However, it is worth spending some effort on sequence optimization, and confirm that it works under the best conditions from the very beginning instead of going back to the start at a later stage of the expression project because suboptimal sequences were used. The Rapid Translation System uniquely offers the possibility to use the tailor-made bioinformatics tool ProteoExpert to reliably perform this task of sequence optimization (see www.proteoexpert.com for further details). The ProteoExpert service runs on a Roche-independent server at Biomax Inc. in Martinsried/Germany and can be accessed via the internet using SSLconnections. The algorithm included in ProteoExpert does not calculate RNA secondary structures, although the changes it suggests finally result in a modified and therefore optimized RNA structure. Instead of trying to model the existing mRNA structure, any wild-type sequence submitted is compared to a set of data derived from several hundred genes whose expression yields have been very thoroughly determined experimentally and fed into a database. Since the correlation between the yields and certain biophysical parameters derived from the sequences of these previously expressed proteins is known, these same parameters can be optimized for each newly investigated target sequence as well, resulting in suggestions for yield-improved sequence variants. The mutations
4 A 1 2 3 4 5 6 7 8 9 10 wt
B 1 2 3 4 5 6 7 8 9 10 wt
Fig. 2. Examples of enhanced yields obtained after template optimization with ProteoExpert. 1–10: Variants suggested by ProteoExpert; wt: Wild-type sequence used as template. (A) Antibody fragment (55 kD, hybrid). (B) Human p58 (unknown function).
proposed by ProteoExpert can be incorporated into the coding region of the target gene very easily by PCR. The output of each ProteoExpert calculation contains a list of primer sequences that can be used directly in PCRs for template generation. Bioinformatic algorithms in general are not precise enough to determine exactly which sequence variant will finally give the highest yield in an actual expression reaction. Therefore, ProteoExpert suggests that 10 different variants be derived from the wild-type sequence, each of them with a very high probability to have a higher expression yield than the wild-type sequence. Testing 5 of these 10 variants was sufficient for the identification of the maximumyield sequence variant in about 75% of all optimization calculations carried out so far (see Fig. 2 for examples). Up to now, ProteoExpert has been able to find sequence variants with significantly increased yields in about 80% of all sequences that were initially expressed at very low levels or not expressed at all. When compared to ProteoExpert, other approaches like direct modeling of the mRNA secondary structure were shown to mostly fail because they still require an empiric decision about which of the several alternative structures should be chosen for sequence variant generation. Also, most other tools are not based on biochemical experiments and are not tailor-made for a specific expression system. All the base substitutions proposed by ProteoExpert are translationally silent. Following calculation of optimized DNA templates by ProteoExpert, these DNAs can be generated using appropriate primers and the RTS E. coli Linear Template Generation Set (LTGS). The resulting linear DNA can then be expressed directly using Roche’s RTS 100 E. coli HY system. Researchers at Roche now use ProteoExpert routinely at the beginning of each cell-free expression project. For example, in a case where the yield obtained for an RTS-expressed SH3 domain was too low and therefore a bottleneck to continue with an expression project finally aiming at X-ray or NMR studies, the use of ProteoExpert at an early stage in the whole project allowed to work with an optimized sequence that gave several milligrams of protein, as required for the subsequent experimental steps (cf. Section 6.2).
5 wt
m1 m2 m3 m4 m5 m6 m7 m8 m9 m10
Fig. 3. Analysis of sequence optimizations suggested by the ProteoExpert software. Shown is an anti-His6-tag Western blot of a wild-type SH3 domain (wt) and 10 mutants (m1–m10). His6-tagged expression templates were generated by PCR and expressed in 50 ml batch reactions.
2.2 Generation of templates for small-scale expression reactions and scale-up A range of different T7-driven plasmids (called pIVEX, plasmids for in vitro expression) for RTS in vitro expression have been designed. They include vectors with His6-, HA-, Avi-tags, or vectors for the reversible fusion of the target protein to MBP or GST to increase its solubility and/or facilitate purification. Standard cloning procedures via PCR and restriction enzymes can be used to insert target cDNAs into these vectors. One template generation method, much more convenient than restriction site cloning into expression plasmids, is used to linear templates generated by PCR (Fig. 3). The up- and downstream regions (T7 promoter/ribosomal binding site/T7 terminator) can be added to the product of a first gene-specific PCR using the commercially available RTS Linear Template Generation Sets. DNA fragments included in these kits carry the necessary regulatory regions. The ends of these DNA fragments overlap with the ends of the first gene-specific PCR product and can be linked to them in a second (so called ‘‘overlap-extension’’) PCR, using primers provided in the kit (Fig. 4). As for pIVEX vectors, different options for protein tagging are provided (His6, Avi-tag, HA or MBP fusion). Linear templates generated using this method can be used directly in smallscale expression reactions producing up to 20 mg protein/50 ml. These reactions are carried out in 96-well-plates or PCR tubes without prior PCR product purification or cloning. If a certain number of sequence variants (e.g., silent – for yield optimization, or mutational – to study the effect of amino acid point mutations on protein function) are to be studied in parallel, this process can be easily automated (see Section 4.2). If plasmids instead of linear templates are preferred or finally required for large scale expressions, molecules generated during the two-step PCR as shown in Fig. 4 must be linked to a plasmid backbone. When individual molecules are cloned out of a pool of linear expression cassettes and inserted into PCR cloning vectors, one cannot exclude the risk of picking clones that carry mutations introduced by previous high-cycle-number PCRs. To circumvent this problem, the BD In-FusionTM PCR Cloning Technology (BD Biosciences Clontech, Palo Alto, CA), provides a powerful alternative to standard cloning: the product of the first, less error-prone gene-specific PCR reaction can be taken instead of the complete linear expression cassette and inserted (‘‘in-fused’’) into
6 First PCR: Addition of overlap regions Gene-specific primers (customer designed)
Second PCR: Addition of regulatory elements and a tag
T7 Promoter Primer
Downstream regulatory element encoding DNA
Upstream regulatory element encoding DNA
T7 Terminator Primer tag T7 Promoter
T7 Terminator
Fig. 4. Application of the overlap extension PCR technique. Linear expression fragments are generated by two PCR steps. The overlap regions added to the gene of interest during the first PCR hybridize to DNA sequences that carry the regulatory elements. In the early cycles of the second PCR, the 30 ends are extended. The full-length fragment is finally amplified via short external oligonucleotides.
a linearized and purified pIVEX vector which already contains the regulatory elements and tags. It has recently been demonstrated that this method, which does not depend on restriction digestion or T/A-overhangs but only on the activity of the BD In-FusionTM enzyme, also works in a dry-down, microtiter plate-based format, making it easy to adapt high-throughput applications. Figure 5 provides a summary of methods recommended for the generation of RTS expression templates. 3. Small-scale expression and optimization Cell-free synthesis of proteins has the advantage of allowing direct access to the reaction conditions, e.g., by offering the possibility of adding co-factors, chaperones or other supplements, and to permit synthesis of cytotoxic proteins. Optimal conditions (e.g., reaction time, temperature) for best expression results can be quickly identified and easily reproduced. Some examples are discussed in the following sections. 3.1 Influence of temperature The influence of temperature on expression of the MIA (Melanoma Inhibitory Activity) protein, a secreted protein from human melanoma cell lines whose
7 RTS E. coli Workflow wt sequence
ProteoExpert E. coli cDNA primer suggestions for sequence optimization
LinTempGenSet E. coli
1st PCR *
PCR product 1 2nd PCR
pIVEX E. coli In-Fusion cloning
circular template E. coli
PCR product 2: linear template E. coli RTS 100 E. coli HY
µgs of protein
sequence optimization expression screening expression scale-up
RTS 500/9000 HY
mgs of protein * 1 to 10 reactions, depending on how many variables are compared.
Fig. 5. Methods for template generation in RTS and use of these templates for expression on different scales.
expression levels are closely correlated with the capability of melanoma cells to form metastasis, was investigated [6]. As recombinant synthesis of functionally active MIA is a time-consuming and difficult procedure and refolding is problematic due to the importance of disulfide formation, expression of mutant MIA proteins was performed via in vitro protein transcription/translation. Correct refolding of the molecule, known to be critical for function, had only been achieved by high effort procedures before. For the screening of mutants, RTS proved to be able to produce the amounts of protein needed for initial testing. The highest amount of correctly folded MIA was achieved using the RTS 500 E. coli HY Kit at 30 C, whereas the highest percentage of correctly folded protein was obtained at 25 C. Functional testing proved the correct folding and activity of MIA. Downstream analysis of RTS-expressed mutant MIA allowed to derive information regarding the importance of defined amino acids for protein structure and function, and to identify individual amino acid residues important for folding (Fig. 6). 3.2 Influence of ligand addition on solubility In another study on the solubility of recombinant proteins, the ligand binding domain of a human nuclear steroid hormone receptor was investigated. Due to being detached from the native protein structure, the single domain was insoluble when expressed both in vivo and in vitro. The addition to the RTS reaction of a receptor agonist known to bind to the single domain had a strong solubilizing effect. Western blot analysis revealed a shift of the soluble protein
8
100 80 60 40
MIA_G61R MIA_Y69H
MIA_G56D
MIA_D29G MIA_D34A MIA_V48I MIA_L52Q
0
MIAdel82 MIAdel79 MIAdel73 MIAdel66
20 reaction buffer MIAwt
Invasion [per cent]
120
Fig. 6. Functional activity of MIA wild-type protein and various mutants tested in invasion assays. The RTS samples were used in a Boyden Chamber with approximately equal amounts of protein in each sample. An RTS sample of the empty vector was used as negative control (reaction buffer). Invasion of melanoma cells was set as 100%. Invasion is inhibited by MIAwt and by several MIA mutants, whereas MIAdel73, MIAdel66, MIA_D34A, MIA_V48I, MIA_L52Q and MIA_G61R lost functional activity. For experimental details see Ref. [7].
M
S
w/o P
1x S
10x P
S
P
S
50x P
Addition of agonist
S
supernatant
P
pellet
Fig. 7. Ligand-dependent solubility of a human nuclear receptor expressed in RTS 100. A ligand known to bind to the expressed receptor was added to the expression reaction in different amounts. Reaction pellets and supernatants were analyzed by anti-His6-tag Western blotting.
fraction from 30% to more than 90% in the presence of the ligand (Fig. 7) (unpublished results). It is only in cell-free expression systems that the addition of a ligand for solubilization is practical. Consequently, RTS displays a wide range of possibilities for screening approaches if insoluble proteins are investigated. 3.3 Co-expression and subunit complementation Protein–protein interactions play a critical role in nearly all cellular processes. Thus, a practical strategy for studying the function of a particular protein
9 of interest is to identify other proteins that interact with it. This approach may lead to the isolation of new components participating in the same pathway, or the identification of previously characterized factors that can help elucidate the function of a protein under study. Moreover, for hetero-oligomeric proteins which assemble as intermediates or metastable species during folding, co-expression of the different subunits is often essential to build a functional assembly. The most widely used technique for identifying protein–protein interactions in genome-wide proteomics studies is the two-hybrid system. However, it has been noted that this system suffers from technical drawbacks inherent in the cellular approach: non-specific interactions, for example, can generate a high proportion of false positive results. It is therefore important to obtain biochemical evidence for specific protein–protein interactions using a second independent method. Combining the genetic technique with a simple in vitro assay to detect physical protein–protein interactions allows validation of twohybrid results. In vitro, proteins can be co-expressed easily and the influence of binding partners or ligands can be analyzed. RTS is an open system and is very well suited for co-expression experiments. Lorenz and Thiesen, for example, explored the possibilities of synthesizing Kox1, a member of the KRAB zinc finger protein family, and its interaction partner TIF1b in RTS [8]. They took advantage of the known interaction of these two proteins as an indicator of proper folding. cDNAs were subcloned in different vectors suitable for T7 polymerasedriven transcription and their potential to give rise to the respective proteins was analyzed by Western blotting. Reconstitution experiments of Kox1/ TIF1b complexes followed by immunoprecipitation showed that complex formation only occurred if the two binding partners were co-translated (Fig. 8). The data led to the hypothesis that TIF1b protein helped the Kox1 proteins to adopt a proper conformation for association. In contrast to this, when working with E. coli, researchers had experienced inclusion body formation and consequently denatured protein, whereas expression of Kox1 in other systems like yeast or baculovirus appeared to be toxic for the host cells and was not possible at all. In another interaction study, two fragments of the E. coli protein MalE, an exported periplasmic receptor for high-affinity transport of maltodextrins, were expressed separately under standard RTS 500 conditions and their reassembly was analyzed [9]. Both fragments were expressed at comparable steady state levels, whereas in vivo, the production of any MalE fragment smaller than 30 kDa had not been detectable previously because of excessive degradation. This suggested that proteolytic activity of RTS lysates was lower than in bacterial cells. A short N-terminal fragment was correctly produced, whereas a C-terminal fragment was mainly found in an aggregated and insoluble form. The complementation of both fragments was assessed by their ability to bind cross-linked amylose, providing a rapid and simple assay for probing native
10 A 1 2 3 4 5 6
TIF1β
kD
94 67
Kox 1 43 30 IP anti TIF1β + + − + − −
Fig. 8. Analysis of possible interactions between the Kox1 protein and TIF1b after individual vs. co-expression in RTS100. Immunoprecipitation with anti-TIF1b antibodies as a means to show association is indicated at the bottom of each lane. Positions of TIF1b and Kox1 as well as of molecular weight standards are indicated. Kox1 and TIF1b were both expressed from T7-driven pRSET vectors. Immunoprecipitation after 20 ml aliquots of His-Kox1 and His-TIF1b RTS reactions had been added together (lane 1); immunoprecipitation after 20 ml Kox1 RTS reaction had been added to a HeLa cell extract representing 500 mg total protein (lane 2); 10 ml aliquot of His-Kox1 RTS reaction (lane 3); immunoprecipitation of in-vitro cotranslated His-Kox1 and His-TIF1b (lane 4); 10 ml aliquot of in-vitro cotranslated His-Kox1 and His-TIF1b RTS reaction (lane 5); and 10 ml aliquot of His-TIF1b RTS reaction (lane 6). The upper part of the blot has been stained with monoclonal anti-TIF1b and anti-His tag antibodies, the lower part with rabbit anti-His tag antibodies. For details see Ref. [8]. Picture with courtesy of Springer and P. Lorenz, Universita¨t Rostock.
MalE conformation [10]. Independently produced MalE fragments were unable to bind amylose, as expected, presumably because residues involved in maltose binding are equally distributed between both fragments. In contrast, the simultaneous expression of both fragments in a single RTS reaction resulted in an active complex that could be purified with a yield of approximately 200 mg/ml. The N- and C-terminal fragments were detected after SDSpolyacrylamide gel electrophoresis and Coomassie staining in stoichiometric amounts, indicating that they had been present on the column as a stable and active complex (Fig. 9). Turbidity measurements were then used to determine the solubility of both fragments. Since precipitation occurred when the C-terminal fragment was expressed individually, misfolding in absence of the complementary N-terminal fragment was concluded. In contrast, the smaller N-terminal fragment remained soluble in the expression solution. Interestingly, when both fragments were simultaneously expressed in the RTS reaction, turbidity decreased to a value comparable to the sample containing only the N-terminal fragment. These results were confirmed by analyzing the protein contents of supernatants and pellets after centrifugation of RTS extracts (Fig. 10). In summary, it could be concluded that the presence of the N-terminal fragment prevented aggregation of the C-terminal fragment and mediated the final assembly of active MalE. Therefore, the RTS strategy for co-expressing protein fragments led to identification of an intramolecular chaperone-like function for the N-terminal
11 A
whole extracts 1
2
3
4
B
eluates 5
6
7
8
1
kDa
2
−175 −83 −62 −48 −33 C-fragment
−25 −17
N-fragment
−7
Fig. 9. (A,B) RTS production of MalE fragments. (A) Production and purification of MalE fragments were analyzed by SDS-polyacrylamide gel electrophoresis and stained by Coomassie blue. Lanes 1–4 whole RTS extracts after 18 h at 30 C; 1 empty vector pIVEX2.4a; 2 pIVME-N; 3 pIVME-C; 4 pIVME-N þ pIVME-C; lanes 5–7 maltose eluates from cross-linked amylose columns; 5 pIVME-N; 6 pIVME-C; 7 pIVME-N þ pIVME-C; lane 8 molecular weight standards. Arrows indicate the position of MalE fragments. (B) Native gel stained by Coomassie blue. Lane 1 wild-type MalE; lane 2 purified complex from coexpression of MalE fragments.
A
B soluble fraction insoluble fraction
1
Turbidity (Abs550nm)
2.0 1.5
2
3
4
5
6
7
kDa −175 −83 −62 −48
1.0
−33
0.5
−25
0
−17 1
2
3
Fig. 10. (A,B) Solubility assays. (A) Turbidity of RTS extracts after 18 h at 30 C; 1 pIVME-N; 2 pIVME-C; 3 pIVME-N þ pIVME-C. (B) Soluble/insoluble production of MalE fragments was analyzed by SDS-polyacrylamide gel electrophoresis and stained by Coomassie blue. Lanes 1–3 RTS supernatants; 1 pIVME-N; 2 pIVME-C; 3 pIVME-N þ pIVME-C; lanes 4–6, RTS pellets; 4 pIVME-N; 5 pIVME-C; 6 pIVME-N þ pIVME-C; lane 7 molecular weight standards. Arrows indicate the position of MalE fragments. For details see Ref. [9]. Picture with courtesy of Springer and J-M Betton, Institut Pasteur, Paris.
fragment of MalE. Interestingly, such a critical role in protein folding for N-terminal fragments has been recently hypothesized for several other proteins as well [11]. The results of these studies validate co-expression of proteins or single domains as a successful application of the RTS system and demonstrate its
12 suitability to detect protein–protein interactions or to study the role of binding partners for proper folding. In contrast to in vivo expression systems, the in vitro environment of RTS offers considerable flexibility for co-expressing two or more proteins without the requirement of constructing plasmids carrying compatible origins of replication.
4. High-throughput protein expression and analysis 4.1 Introduction The analysis of all proteins encoded by the enormous sequence data collected from the various genome projects is a major challenge and requires the synthesis of many target proteins in parallel, often combined with PCR-based mutagenesis. Common approaches to evaluate protein function include the introduction of point mutations, deletions or insertions, and techniques for domain fusion. Since linear expression constructs can directly be expressed in vitro, it is easy to combine Expression-PCR (E-PCR) with other wellestablished PCR mutagenesis methods. In cases where research focuses on the rapid production of pharmaceutically relevant proteins and their functional and structural analysis for the development of inhibitory or activating drugs, in vitro expression displays considerable advantages. As the proteins are not produced in vivo, direct functional analysis, e.g., by Surface Plasmon Resonance (SPR), without the otherwise necessary purification from a complex E. coli cell is possible. Using in vitro expression PCR, the time required for protein engineering is dramatically shortened. The whole process of two PCR steps followed by in vitro expression is feasible in less than 16 h. The proteins can subsequently be subjected to activity assays as demonstrated for example, by studies on the fluorescence of green fluorescent protein (Fig. 11). Random mutagenesis, PCR misincorporation procedures and recombination strategies may be combined with linear template generation by PCR in a similar way. After mutation or shuffling of the gene of interest, an expression fragment is constructed via overlap extension methods. In contrast to the introduction of single point mutations, the PCR products resulting from random mutagenesis are not homogeneous but represent a mixture of different variants. The constructs can be ligated into RTS pIVEX vectors to divide the pool into single species that can be screened. Small-scale expression reactions in 50 ml volumes can be carried out in microtiter plate-based 96 well formats, and the whole process can be easily adapted to high-throughput screening (HTS) or automated liquid handling. Plasmid preparation is the only step in the whole workflow where transformation and growth of E. coli cells is required, while the steps of growth and lysis of large-volume cultures of more than 100 ml during downstream processing can be avoided.
13 A kD
B
GFP GFP T203Y
30
20
fluorescence activity
45
fluorescence activity
75
10
502 nm GFP T203Y
502 nm GFP
Fig. 11. GFP and mutated GFP (T203Y) were expressed in RTS100 E. coli HY from linear templates. (A) 0.5 ml of each reaction solution were separated by SDS-PAGE, blotted and detected with Anti-His6 Peroxidase conjugate. (B) Fluorescence activity was measured by exciting the protein at 395 nm and monitoring the emission at 430–580 nm. The reaction solutions were diluted 1:80.
4.2 Example: Screening of single-point muteins Researchers at Roche Pharmaceuticals developed a production platform which enables them to generate, express and purify recombinant proteins automatically in a small-scale format for initial construct evaluation [12]. All reaction steps were performed in microtiter plates allowing the automated and parallel processing of up to 96 protein variants through all steps including mutations, truncations and species variations. The expression cassettes used in this study comprised sequences for C-terminal affinity purification via an N-terminal hexahistidine-tag and for site-directed biotin labeling. Selective biotin-labeling of the fusion proteins was performed by using a biotin-accepting peptide (BAP) sequence called Avi-tag, which is enzymatically biotinylated by E. coli biotin ligase during the in vitro translation [13]. After a robotically performed Ni-NTA purification procedure, a fraction of each purified protein was immobilized on streptavidin-coated microchips via the mono-biotinylated BAP sequence and analyzed by protein–protein interaction measurements using SPR technology. As a model protein, the authors chose the N-terminal binding domain of the human Insulin-like Growth Factor-Binding Protein-4 (IGFBP-4). IGFBP-4 is a 24 kDa protein that binds Insulin-like Growth Factor 1 (IGF-I) and 2 with high affinity, presumably via its N-terminal region (mini-BP4). To characterize the interaction of mini-BP4 and IGF-I and to identify the amino acids that are involved, site directed mutagenesis was performed by PCR with the mini-BP4 gene. Altogether 30 different single-point muteins were synthesized (Fig. 12).
A
m
miniBP4 bridged LEE wt V49L V49I V49M V49F V49Y V49W Y50R Y50C R53Y R53M R53F R53H Y61W Y61W Y61F K68Q L70Y L70W L70M L70I L70F L73I L73W L73M L73F M74Y M74W M74I M74F H75D #
14
B
1
2
3
4
5
A B C D E F G H
m
Western blot 1234 BCCP mini-BP4
Slot-Blot (native protein)
Fig. 12. (A) Ethidium bromide-stained, 2% agarose-gel showing mini-BP4 constructs generated by PCR. mini-BP4 Wild-type mini-BP4 after the PCR synthesis reaction. Bridged:mini-BP4 elongated with the bridging primers bridgeF1 and bridgeR1 to form overlaps with the DNA modules; LEE wt the His-mini BP4-Avi wild-type construct; V49L-H75D 31 different mini-BP4 constructs. LEE Y61W was produced twice; þ positive control DNA; m: marker. (B) Nondenaturating slot-blotting as a fast and highly specific detection method. In the slot blot ( positions A1–G4) and in the Western blot (lanes 3, 4) His-mini-BP4-Avi muteins were detected via the biotinylated AviTag with SA-HRP conjugate. The positive control in position C5 and in lane 1 is monobiotinylated PEX2 protein. Positions H4 and A5 correspond to lane 2 and are negative controls of incubated RTS 100 HY lysate containing no Linear Expression Element. BCCP provides no background signal in non-denatured slot blotting (H4, A5) but was detected after denaturing Western blotting (lanes 2–4). The arrows indicate the same samples detected by slot and Western blotting. Reproduced from Ref. [12] with courtesy of Schra¨ml, Roche Pharmaceuticals, Penzberg, Germany and Springer.
The influence of the site directed mutations on the binding affinity of the mini-BP4 protein to its 7.6 kDa IGF-I protein binding partner was determined and the effect of each specific mutation on binding was demonstrated (Fig. 13). Thus, the sequence-specific, co-translational biotinylation of fusion proteins containing an Avi-tag allowed their specific detection, quantification and site-directed immobilization on surfaces coated with streptavidin. The combination of fast template-generation, in vitro expression, robot-assisted protein refolding and purification as well as the automated interaction analysis turned out to be a valuable system for the production and analysis of protein variants. 5. Expression scale-up and protein purification For studies on protein structure (e.g., NMR), the recombinant protein is often required in milligram amounts. Finding an appropriate expression system or optimizing the yield of a given one is often a question of trial-and-error methods. In RTS, a more rational approach exists: as a first step, cDNA templates for optimized expression yields can be calculated via ProteoExpert.
15
RU
IGF-1
Response units
160
biot. mini-BP4
140 120
A
100 80 60
B
V49F mini-BP4 Ligand-Activity
40
C
20 0
IGFl inject 350
SA-biosensor
400
450
500
550
600
Time (s) wt-miniBP4
K68Q
RU
RU
Time (s)
Time (s)
V49I RU
Y61F RU
Time (s)
Time (s)
Fig. 13. Thirty-one mini-BP4 constructs were analyzed in SPR-interaction analysis (Biacore 3000) with IGF-I. The mini-BP4 muteins were site-directed immobilized on streptavidin-coated sensor chips via the biotin-labeled AviTag; 26 muteins were functionally active, 6 muteins were inactive and 11 muteins revealed a complex binding behavior (4 exemplary sensorgrams are shown). To determine the refolding efficacy of the three different buffers, equal amounts of Mini-BP4 mutein were immobilized (RU loaded) on the flow cells of a BIAcore SA chip. The chip was saturated (RMAX) by an 800-nM IGF-I injection and the ligand-binding activities of the immobilized mutein were calculated by the ratio RMAX/RU loaded. Refolding of the constructs revealed highest ligand-binding activity (20%) with a redox/arginine buffer system (A). A moderate ligand-binding activity (3%) was reached with buffer I (B). As expected, no activity was observed with the reducing buffer system (C). The curve plot of the differently treated V49F mutein is given here, as example. Reproduced from Ref. [12] with courtesy of M. Schra¨ml, Roche Pharmaceuticals, Penzberg, Germany and Springer.
16 Using the modified primer sequences proposed by this program, a number of different linear expression constructs can be produced rapidly in parallel PCRs and fused to the necessary regulatory regions via the RTS E. coli Linear Template Generation Sets in a second PCR step (see Section 2.1). Next, a smallscale expression of 2 h in RTS 100 E. coli HY reactions allows to identify the template with the highest productivity. This template can then be preserved e.g., by BD In-FusionTM cloning. After transformation and plasmid amplification, the circular expression template is ready for expression scale-up in RTS 500 ProteoMaster or RTS 9000 E. coli HY reactions. The use of the RTS ProteoMaster Instrument thereby guarantees reproducible expression conditions for the scale-up. The complete RTS workflow starting from template optimization by ProteoExpert via expression PCR and cloning of the best DNA variant to high-level expression in RTS 500 or RTS 9000 has been carried out successfully with different proteins. For structural analysis, the SH3 domain of a human kinase was investigated following this procedure, resulting in an expression level of 3 mg/ml in RTS 500 E. coli HY reactions [14]. This protein amount was enough for homogenous purification via an His6-tag (Fig. 14). The RTS system proved also to be very effective for synthesis and purification of chloramphenicol acetyl transferase (CAT) for interaction studies [15]. Here, a ProteoExpert-optimized sequence coding for CAT was cloned into a pIVEX construct that contained a sequence for an N-terminal Avi-tag. By adding
M: Multicolor marker S: Starting material F1, F2: Flow through W: Wash E1, E2: Elution
M
S
F
W
E1
E2
Fig. 14. His6-tag purification of an SH3 domain (9 KD) after sequence optimization and expression in RTS 500 E. coli HY.
17 RU 1400
Dissociation
Binding of Bio-CAT on SA-chip
1200 RU
pIVEX 2.8dWT CAT/pIVEX 2.8d CAT RTS 500 E. coli HY
1000 Association .DIM
1 2
800 600 400
CAT
200 0 −200 200 250 300 350 400 450 500 550 600 650 Time s
Fig. 15. Optimized expression of CAT (chloramphenicol acetyltransferase) in RTS500 E. coli HY followed by functional studies. Left: Concomitant biotinalytion and expression reactions of pIVEX 2.87d vT (wild-type) CAT (lane 1) and pIVEX 2.8d CAT (lane 2, codon optimized by ProteoExpert algorithm) analyzed by Coomassie staining. Right: SPR Analysis of Avi-Tag CATbio immobilized on Biacore SA chip from 1, 10 and 100 nM solutions, binding/dissociation of polyclonal anti CAT antibody is shown. Insert Immobilization of Avi-Tag CATbio onto Biacore SA Chip from 1-, 10- and 100-nM solution.
biotin, biotin protein ligase and ATP, specific monobiotinylation was accomplished during expression reactions. The successful scale-up from RTS 100 to RTS 500 E. coli HY was finally followed by an SPR analysis to measure the interaction with an anti-CAT antibody (Fig. 15). 6. Protein labeling In the last decade, it became clear that the set of 20 canonical amino acids prescribed by the universal genetic code needs to be enlarged to span all dimensions of chemical variability that could be potentially advantageous for functional diversification of proteins. Canonical amino acids attached to cognate tRNAs can be chemically or enzymatically modified or even loaded onto desired tRNAs before they enter the ribosome. Specific protein labeling with modified amino acids is used in a wide variety of experimental setups to obtain information on conformational changes, protein folding, ligand and co-factor binding, or to monitor protein thermal stability [16], and to simplify spectra for structure elucidation [17]. Recent examples of artificial (i.e., tailor-made) proteins include novel classes of functionally designed protein pH-sensors, variants of green fluorescent proteins or luminescent proteins with enhanced stability. 6.1 Labeling with fluorescence-enhanced amino acids Specific incorporation of spectrally enhanced Trps (e.g., 5-fluorotryptophan, 5-FW) is a widely used approach to study protein–protein interactions.
18 Conformational changes in proteins, e.g., due to denaturation or ligand and substrate binding, can be analyzed by fluorescence spectroscopy. Trp fluorescence is sensitive to solvent effects, and its emission spectrum and quantum yield strongly depend on the protein structure and local microenvironment. However, since the intrinsic fluorescence of different proteins resulting from naturally occurring Trp residues overlap, it is often impossible to assign and interpret fluorescence changes that result from intermolecular associations. Replacement of natural Trp residues in proteins by appropriate analogues is therefore an important option. It is usually achieved by the tightly controlled overproduction of the protein in a Trp auxotrophic E. coli strain growing in minimal medium containing the desired Trp analogue [18]. High level cell-free expression techniques like RTS offer an excellent alternative possibility for the efficient label incorporation into recombinant proteins, allowing the convenient and uniform labeling of virtually any amino acid. In addition, common problems of standard in vivo labeling protocols associated with toxic effects of the label precursors, reduced protein yields or low label incorporation into protein samples can be eliminated. In an example underlying this principle, Sengupta et al. studied the homoand heterodimer formation of the E. coli RcsB protein, a key regulator in enteric and plant pathogenicity [19]. RcsB is known to interact with the coactivator RcsA during transcriptional regulation of the expression of bacterial capsules. The researchers showed replacement of the natural Trp residue in RcsB by various Trp analogues as well as high level production of the modified RcsB derivatives using cell-free expression. The isolated RcsB alloproteins proved to be suitable for protein interaction studies by fluorescence spectroscopy and evidence was obtained that RcsB also oligomerizes due to molecular association of the C-terminal effector domains. No negative effects due to the utilized amino acid analogues on the kinetics or efficiency of cell-free protein production were observed. As demonstrated by fluorescence spectroscopy, cell-free production of RcsB alloproteins in RTS affected neither protein–protein interactions (homo- or heterodimer formation) nor DNA binding activities. The incorporation of electron-rich selenium-containing noncanonical counterparts of Trp and Met into proteins, often used due to the rare occurrence of these amino acids in sequences, via in vitro translation system, represents a useful route for solving the phase problem in protein X-ray crystallography. In a study carried out by Budisa et al. [20], b-Selenolo[3,2-b]pyrrolyl-L-alanine ([3,2]Sep), a surrogate of tryptophan (Trp), and selenomethionine (SeMet), an analogue of methionine (Met) were incorporated into GFP (green fluorescent protein). GFP was used as the model protein for the replacement experiments since it has a known three-dimensional structure and well-established biochemical, biophysical and genetic properties. The seleno-proteins were obtained at yields comparable to those of the wild-type protein and also crystallized under similar conditions (Fig. 16).
19 A
B C Se
S
H2N
H2N
O
HO Methionine (Met)
O HO Selenomethionine (SeMet) O
D d c
N H
Tryptophan (Trp) τM = 5580 M−1cm−1 (at 280 nm)
a 1
O OH
OH NH2
b
NH2 N H β-Selenolo[3,2-b]pyrrolyl-alanine; [3,2]Sep or Selenolo-Tryptophan τM = 12880 M−1cm−1 (at 267 nm)
2
Fig. 16. Green fluorescent protein (GFP) engineering by in vitro substitution of its Trp and Met residues with indicated amino acid analogues. (A) Structures of GFP with marked chromophore and five Met residues (upper plot) and only one Trp residue (lower plot) based on its PDB coordinates and represented as a ribbon plot. (B) Structural representations of side chains of canonical amino acids Met and Trp and their analogues and surrogates selenomethionine (SeMet) and b-Selenolo[3,2-b]pyrrolyl-L-alanine ([3,2]Sep), respectively. (C) Analysis of the expression profiles in RTS HY lysates after overnight reaction: a wt-GFP; b SeMet-GFP; c [3,2]Sep-GFP; d control (i.e., reaction without plasmid). Arrows indicate the position of overexpressed native and substituted proteins. Lysates were separated by 12% SDS-polyacrylamide gel and stained with Coomassie blue. (D) (box): Photographs of crystals (1: [3,2]Sep-GFP; 2: SeMet-GFP). Reproduced from [20] with courtesy of N. Budisa, Max-Planck-Institut fu¨r Biochemie, Martinsried, Germany and Springer.
6.2 Labeling with isotopes for NMR and X-ray NMR spectroscopy is uniquely capable of providing information on the structure, function, and dynamics of proteins and other biomolecules in solution. Recent advances in NMR spectroscopy promise to revolutionize the sensitivity of the technique as well as the molecular weight range amenable to investigation. However, milligram quantities of proteins are typically required for such studies. The assignment and structure determination of peptides and proteins with less than approximately 100 amino acids can usually be accomplished utilizing two-dimensional experiments such as the NOESY, COSY, and TOCSY that rely on 1H signals [21]. With increasing molecular weight and the concomitant increase in the number of 1H resonances, multidimensional (two, three, and four-dimensional) spectra resulting from labeling with other isotopes (usually 15N and 13C) must be used to solve the 1H overlap problem. Most of the work carried out nowadays therefore relies on uniform 15N, 13C and 2H labeling of the protein. Information about specific residues and specific sites can also be obtained by selective incorporation of only one or a few amino acids labeled with 15N, 13C, and possibly 2H. This selective labeling strategy can dramatically simplify the NMR spectrum. It is routinely possible to determine the structures of proteins up to 25 kDa using a combination of 15N and 13C labeling.
20 Recent experimental advances have increased the molecular weight range accessible to such studies to complexes as large as 800 kDa [22]. In vitro protein synthesis shows a number of important advances for the production of labeled NMR samples. Although different cell-based methods exist for the production of uniformly 15N and 13C-labeled proteins [23], there are several problems that occur when expressing proteins in in vivo systems. The proteins must be purified from a range of other cellular proteins, a process that can be very tedious. Additional complications arise when trying to express selectively labeled proteins using a cell-based approach. One major problem results from metabolism of the amino acids once they are taken up by the cell. For example, aspartate can be converted into glutamate, asparagine, and lysine. Thus, addition of labeled aspartate can result in the inadvertent labeling of other amino acids as well. Similar metabolic interconversion pathways exist in E. coli for many amino acids. Adding a complement of unlabeled amino acids helps to suppress endogenous amino acid synthesis. However, even in this case, scrambling of the label has been shown to occur, albeit at a reduced level [24]. Another problem is that significant amounts of the often expensive labeled amino acids are either thrown away with the medium after the cells have been harvested, or incorporated into other cellular proteins. Cell-free protein expression has the potential to overcome this limitation since the labeled amino acids can be just added by replacing the unlabeled ones, the expressed protein is labeled exclusively and label scrambling hardly occurs due to the absence of cellular metabolism. In the future, with continuing research into cell-free technology, in particular with regard to protein yields and the inhibition of amino acid metabolic pathways, it is expected that most specifically labeled proteins for NMR studies will be produced by in vitro methods. As an example, Fernholz et al. tested the labeling efficiency of the Rapid Translation System with an SH3 domain as a model protein [25]. The expression yield was substantially increased (more than 6-fold) by optimizing the sequence of the gene by silent mutations (cf. Section 2.1). Using the improved expression construct, the incorporation of 15N-labeled derivatives of each canonical amino acid was evaluated and analyzed via NMR (Fig. 17). Eighteen of the 20 amino acids (all except Glu and Gln) could be specifically incorporated. Scrambling was observed for only three amino acids (Ser, Asp and Asn). The yields in all expression reactions were comparable to the reaction using unlabeled amino acids ( 3 mg of soluble protein). Isotope effects were not detected, even when deuterated amino acids had been used. These results demonstrate that cell-free protein expression has unique advantages over cellular expression regarding selective labeling of proteins. Due to a dramatic reduction of the time needed for cross-peak assignments of 15N-HSQC spectra, it can be considered as a valuable tool to solve protein structures in significantly shorter time.
21 A
B
Pro Val Gly Ala Leu
Arg His Lys
C
D
Met Cys Thr Ser
Trp Phe Tyr
10.5 10.0 9.5 9.0 8.5 8.0 7.5 7.0 6.5
104 106 108 110 112 114 116 118 120 122 124 126 128 130 104 106 108 110 112 114 116 118 120 122 124 126 128 130
10.5 10.0 9.5 9.0 8.5 8.0 7.5 7.0 6.5
15
Fig. 17. N-HSQC spectra of differently labeled SH3 domain. (A) Aliphatic residues. (B) Basic residues. (C) Aromatic residues. (D) Heteroaliphatic residues.
7. Advanced applications and limitations 7.1 Expression of disulfide-bonded proteins Expression of recombinant proteins in E. coli cells often results in intracellular aggregation of the produced proteins. This is also observed in cell-free synthesis, especially for large, complex mammalian proteins. However, in cell-free synthesis, one can control and influence the reaction. In order to be biologically active, recombinant polypeptide chains have to fold into their native threedimensional structure, e.g., by interacting with GroE or DnaK chaperones. The issue of protein folding becomes even more complicated for mammalian proteins, especially extracellular molecules like antibodies or receptors harboring multiple disulfide bonds. Modifications of the conventional cell-extract made it possible to obtain such target proteins in a functional form. It has been demonstrated that E. coli cell-free protein synthesis reactions allow the formation of disulfide bonds when the free sulfhydryl groups of endogenous proteins contained in the lysate are blocked by sulfhydryl-specific alkylating agents [26]. Treating the cell-extract with iodoacetamide (IAM) completely eliminated the reducing activity without a severe reduction in protein synthesis capability. IAM was chosen as the thiol modifying agent since it efficiently blocks thiolmediated catalysis without changing the ionic charge on proteins. In addition, the use of chaperone-enriched extract further enhanced the solubility of expressed proteins. The use of IAM-treated extract along with the addition of DsbC disulfide isomerase and an oxidizing glutathione mixture enabled the expression of an enzymatically active murine urokinase, a serine protease,
22 1000
Active rUK (ug/ml)
800 600
400 200 0 0
10
20
30
Incubation time (hr)
Fig. 18. Expression of the protease domain of murine urokinase (rUK). Plasmid pIVEX2.3rUK was incubated in RTS500 HY systems with normal (open circles) or IAM-treated ( full circles) extract in the presence of glutathione mixture (4 mM GSSG and 1 mM GSH). 40-ml samples were taken at the indicated time points and assayed for enzymatic activity. Stn'd extract S
P
IAM - extract S
P rPA(9)
rUK(6)
sFcgR 1(4)
scFv a-Hag (2)
CD40Fc (10)
Fig. 19. Enhanced solubility of disulfide bonded proteins due to expression in chaperone-enriched, IAM-treated extract. Various proteins with different numbers of disulfide bonds were expressed with standard (Stn’d) extract or GroE-enriched extract after IAM treatment. S Soluble fraction; P pellet. Numbers of disulfide bonds are indicated in parentheses.
demonstrating the feasibility of using cell-free protein synthesis for the rapid and general expression of bioactive eukaryotic proteins (Fig. 18). In terms of protein solubility, it was found that IAM-treatment alone is often not sufficient to allow production of protein in a soluble and functionally active form. The inclusion of chaperones like GroE, present as overexpressed components of the IAM-treated E. coli lysate, helped in many cases to overcome this problem (Fig. 19).
23 Another important example is the expression of functional single-chainantibody molecules and their derivatives in RTS. The relatively small number of disulfide bonds present in scFv molecules and their conserved position makes them ideally suited as targets for cell-free expression. Basically the same conditions and additives (e.g., chaperones) can be used for the cell-free expression of all molecules belonging to this class, irrespective of their specificities. In addition to this, the presence of highly conserved framework regions flanking the variable parts on the cDNA level makes scFv-sequences good candidates for high-throughput template generation, because a very limited set of primers is sufficient to generate linear expression templates encoding many different specificities. In an unpublished study by Ylera and LeGall, a diabody (bispecific molecule composed of two scFv fragments connected via a stretch of amino acids as linker) specific for CD3 and CD19 was found to specifically bind the antigen-expressing target cells as monitored by FACS analysis, provided that diabody was expressed in an IAM-treated lysate, but not when produced in a conventional, unmodified lysate. Interestingly, a side-product derived from the main target protein by proteolytic cleavage was present in high amounts in the conventional but not the IAM-treated lysate, presumably because the IAM treatment blocked the functional sites of the proteases. 7.2 Expression of toxic proteins Cell-free expression systems are especially useful for the expression of toxic proteins. As an example, D-amino acid oxidase (DAO) from T. variabilis was expressed for the first time using RTS 500 [27]. DAO is the prototype of FAD-containing oxidases, and in vivo expression in E. coli is FAD-dependent. However, higher expression levels seem to interfere with growth of E. coli cells, presumably by disturbance of cell wall metabolism and/or toxic effects of enzymatically produced H2O2. By addition of the essential cofactor FAD to the cell-free expression reaction the yield of soluble and functional active DAO was increased dramatically (Fig. 20). It can be assumed that by identifying and supplementing limiting cofactors such an approach can also be used for the expression of other critical proteins to increase their solubility and functionality. This shows again that cell-free expression systems are particularly advantageous as reaction conditions and concentrations of added components can be easily modified. 7.3 Expression of peptides Expression of short polypeptide chains in RTS is challenging because small molecules are more easily attacked by proteases than larger protein complexes. High concentrations are difficult to obtain because they depend on the use of the CECF technology which in turn is based on the use of a semipermeable
24 1000
Volume Activity [U/ml]
800 1 2
34
600
400
200
0 0
500
1000
1500
2000
2500
FAD Conc. [mM]
Fig. 20. Dependency of DAO volume activity on the FAD supplement concentration. The inserts show the SDS-PAGE analysis of DAO expression mixture at 0 mM FAD (lane 1 supernatant; lane 2 pellet fraction) and at 2000 mM FAD (lane 3 supernatant; lane 4 pellet fraction). Reproduced from [27] with courtesy of Frank Wedekind, Roche Applied Science, Penzberg, Germany and Springer.
membrane of a certain cut-off (usually 10 kDa). Since smaller peptides are not retained in the reaction compartment of the CECF devices, only batch reactions excluding the dialysis effect normally make sense for peptide synthesis, or reaction products must be isolated from both CECF reaction compartments. Researchers at Phylos GmbH (Zwingenberg, Germany) showed recently that expression of peptides in RTS 100 E. coli HY batch reactions is possible from a circular template (unpublished results). Since they failed with the detection of the small peptides when expressing them alone, they introduced an additional tag, thereby increasing the size of the peptides and allowing their detection by radioactive labeling and fluorescence imaging. Product functionality, especially the possibility to immobilize the peptides on beads, will be the subject of further studies. 7.4 Expression of transmembrane proteins The synthesis of eukaryotic membrane proteins in a preparative scale in prokaryotic systems is another challenging task. Membrane proteins have been expressed in RTS 100 and 500 at high yields, however, in insoluble form [28]. Four types of membrane proteins were selected that biosynthetically become integrated into the membrane of the endoplasmic reticulum and are resident proteins of the endoplasmic reticulum and the plasma membrane, respectively: a type I membrane protein (Mtj1p), a type II membrane protein
25
+ CHAPS (8.5 mM) + additional salt
+ CHAPS (8.5 mM)
+ Deoxycholic acid (0.5 mM)
+ Mega10 (6 mM)
+ Mega8 (10 mM)
+ betain (500 mM)
silent mutations (PM22)
M
+additional salt
kDA
without addition
Coomassie-stained gel with pellets
200 120 80 50
Mtj1p
30 20
reaction
1
2
3
4
5
6
7
8
9
Fig. 21. Optimization of the synthesis of Mtj1p in RTS 500 E. coli HY (reaction time: 24 h). Protein synthesis is shown in the presence of higher salt concentrations, silent mutations, betain and different detergents. The figure shows the sediments corresponding to 5 ml of the reaction mixtures on a Coomassie-stained gel.
(Sec62p), two multispanning membrane proteins (Trp4/TRPC4, Trp8/CaT-L/ TRPV6) and a tail-anchored membrane protein (Ubc6). Out of five model membrane proteins four were produced in milligram amounts in RTS 500. The non-native proteins could be used for immunization and as protein standards for quantification of the respective proteins in biological material by semiquantitative Western-blotting. In addition to producing membrane proteins in high yields, it was observed that detergents can play a significant role in the efficiency of translation: as shown for Mtj1p in Fig. 21, the presence of detergents in RTS 500 E. coli HY reactions affected the yield of some of the expressed proteins but did not give rise to more soluble product. Based on these findings the investigators proposed that in the absence of detergent certain newly synthesized membrane proteins may accumulate on the surface of ribosomes via hydrophobic interactions and that this may lead to inactivation of ribosomes. 8. Switching between and direct comparison of cellular vs. cell-free expression Since constitutive T7-driven protein expression is characteristic not only for RTS but for a number of other vectors for cellular expression as well (e.g., pRSET vectors or the pDEST vector series used in the context of the Gateway platform), many of these vectors can be directly used for cell-free expression. If, for example, expression of a panel of eukaryotic proteins fails because of toxicity, low yield or low solubility, one can quickly get an idea about expression efficiency in RTS simply by running a small-scale pilot expression experiment using already existing vectors in batch mode. Control of gene expression by the
26 strong phage T7 promoter that yields constitutive expression induced by endogenous T7 RNA polymerase expression is advantageous for protein production in terms of yield, but since many heterologous proteins are toxic to E. coli cells, basal expression of the genes may lead to reduced cell growth, increased cell death, and an overall failure of protein synthesis. In a cell-free context, this toxicity effect is much more unlikely as shown by the following examples. In a systematic study of optimization of protein expression the Rapid Translation System was used for proteins that had been expressed in T7-driven Gateway vectors before or in parallel [29]. When expressing a total of 21 ORFs from pDEST17 vectors either in the E. coli strain BL21(DE3) pLysS or directly in an RTS 100 E. coli HY batch reaction, about 50% of all tested proteins were expressed successfully in each case. When the software tool ProteoExpert was used in conjunction with the RTS E. coli Linear Template Generation Set (LTGS) to optimize a subset of low-yield sequences, it proved to be a very important part of the cell-free workflow not available in the cellular context, improving dramatically the success rate and yield of the cellfree reactions. In summary, a total of 21 ORFs were tested. Of these, 10 were successfully expressed in BL21pLysS E. coli (as detected by Coomassie stain), and 12 in the RTS 100 E. coli HY Kit. All 21 ORFs were submitted to ProteoExpert for optimization calculations, linear expression constructs were generated with RTS E. coli Linear Template Generation Sets and finally expressed in RTS 100 E. coli HY. Eighteen samples had higher yields of protein when expressed in vitro. Remarkably, removal of the att (attachment) sites included in the Gateway pDest vectors, and a change of the His6-tag position from N- to C-terminus led to a significant increase in expressability and cell-free protein yield. In another case, the expression of chemokine-like factor 1 (CKLF1) from the T7-driven vector pRSET in RTS was compared with results obtained in E. coli BL21(DE3)pLysS [30]. SDS-PAGE and Western blotting showed that the recombinant protein was synthesized in the cell-free system but not in E. coli cells, showing the advantages of a cell free protein expression system over in vivo expression for recombinant CKLF1 (Fig. 22). The fact that cellular expression of this protein had also failed utilizing the Glutathione S-transferase (GST) Gene Fusion System from Amersham Pharmacia Biotech made these researchers conclude that CKLF1 was possibly toxic to E. coli, and therefore unable to be expressed in living cells. This again demonstrates that cellfree biosynthesis systems provide a good alternative method for expressing proteins such as CKLF1 that cannot be expressed using traditional methods. 9. RTS – the next generation: Eukaryotic cell-free expression in wheat germ lysates Taking all these examples together, it is clear that the expression of proteins has many advantages in a cell-free environment but also that – in spite of
La ne
2 La ne
2 La ne
La ne
1
La ne
1
3
27
A
B 14 kDa
Fig. 22. Expression of His-tagged CKLF1 in the Rapid Translation System. Left: Comparison between expression in vitro and in vivo using the same expression vector as template. Western blotting analysis probed with (A) anti-CKLF1 and (B) anti-His; lane 1 CKLF1 synthesized with RTS 500; lane 2 CKLF1 expressed in E. coli. Right: SDS-PAGE analysis of cell-free protein synthesis. One ml of the reaction solution diluted 10-fold with PBS was applied to denaturing SDSPAGE. Lane 1 Protein marker; lane 2 pRSET C-CKLF1 utilized as template. A notable band with apparent molecular mass of 14 kDa was observed; lane 3 pRSET C without gene of interest used as template (negative control).
sequence optimization and fast screening methods – eukaryotic target proteins remain challenging in a system based on E. coli lysates. CECF technology has therefore recently been transferred to a eukaryotic environment and adapted to a lysate based on wheat germ extracts [31,32]. Such lysates are highly advantageous in terms of success rate. Even non-optimized wild-type sequences can be expressed at levels detectable non-radioactively by Western blotting. This has been shown for many eukaryotic proteins that are not expressed at all in E. coli lysates, even when optimized mutants are generated. Higher stability and better compatibility of eukaryotic mRNAs with the eukaryotic translational machinery than with E. coli ribosomes may explain this finding. From those proteins successfully expressed in both systems, solubility (and presumably also function) have been found to be significantly higher, possibly due to the presence of eukaryotic chaperones and lower protease activity of the wheat germ lysate compared to E. coli extracts. The RTS Wheat Germ system (visit www.proteinexpression.com to check for commercial availability) differs from other available eukaryotic cell-based systems in two important aspects. Including CECF technology with running times of up to 24 h (instead of 2–4 h batch reactions) also, for small-scale reactions of 50 ml, it aims at the expression of proteins in a range that allows detection by Coomassie stain or Western blotting instead of radioactive labeling. With an also upcoming RTS 500 Wheat Germ platform, the Rapid
28 wt sequence
ProteoExpert E. coli
primer suggestions for sequence optimization LinTempGenSet E. coli
cDNA
sequence optimization expression screening expression scale-up
1st PCR
PCR product 1* 2nd PCR
LinTempGenSet Wheat Germ 2nd PCR
In-Fusion cloning PCR product 2: linear template wheat germ
PCR product 2: linear template E. coli
RTS 100 E. coli HY
pIVEX WG pIVEX E. coli e.g. Ncol/Smal RTS 500/9000 E. coli HY
RTS 100 WG CECF
RTS 100/500 WG CECF protein
protein protein
protein
*same product can be used for E. coli and WG 2nd PCR
Fig. 23. Overall summary of expression workflow for template generation and expression in RTS E. coli and wheat germ.
Translation System will become the first scalable eukaryotic cell-free protein expression system, with amounts allowing to easily link up expression reactions to downstream applications like functional assay, immobilization on chips etc. A consequence of this focus on yield, however, is that posttranslational modifications, e.g., microsome-dependent glycosylations, cannot be carried out in RTS, since – unlike ribosomes – microsomes work stoichiometrically instead of catalytically and cannot carry out their function throughout the running time of a preparative-scale expression reaction. The approaches for expressing a given target protein in RTS E. coli vs. wheat germ are very similar, except for the fact that a template optimization for a eukaryotic protein expressed in the eukaryotic wheat germ lysate is not required. Template generation strategies like overlap extension PCR and BD In-FusionTM cloning work similarly in both systems based on the same genespecific primers. As restriction sites between expression plasmids are also compatible switching from E. coli to wheat germ or vice versa is easy and straightforward (Fig. 23). 10. Summary and conclusions Cell-free protein expression is a promising technology for keeping pace with the exponentially increasing amount of genetic sequence information and for empowering the exciting field of protein and pathway evolution. The consensus in the scientific community is that the 21st Century will be the era of proteome
29 research, and cell-free systems are key technologies that will support this expanding research area. The Rapid Translation System (RTS) from Roche Applied Science combines a unique set of innovative technologies to a powerful new protein expression approach. It offers unique possibilities such as rational template design, expression of toxic proteins, co-expression of multiple polypeptides, efficient labeling for NMR and X-ray studies, and rapid production of engineered protein variants. RTS is a rapidly evolving platform overcoming the limitations of traditional cell-based methods in terms of speed, throughput and flexibility. Based on an isolated translational machinery under optimized reaction conditions, it enables scientists to expand their research projects into multiplexed and automated formats.
References 1. Nirenberg MW and Matthaei JH. The dependence of cell-free protein synthesis in E. coli upon naturally occurring or synthetic polynucleotides. Proc Natl Acad Sci USA 1961;47: 1588–1602. 2. DeVries JK and Zubay G. DNA-directed peptide synthesis. II. The synthesis of the a-fragment of the enzyme beta-galactosidase. Proc Natl Acad Sci USA 1967;57:1010–1012. 3. Spirin AS, Baranov VI, Ryabova LA, Ovodov SY and Alakhov BY. A continuous cell-free translation system capable of producing polypeptides in high yield. Science 1988;242: 1162–1164. 4. Kigawa T, Yabuki T, Yoshida Y, Tsutsui M, Ito Y, Shibata T and Yokoyama S. Cell-free production and stable-isotope labelling of milligram quantities of proteins. FEBS Lett 1999; 442:15–19. 5. Kim D-M and Swartz JR. Prolonging cell-free protein synthesis with a novel ATP regeneration system. Biotechnol Bioeng 1999;66:180–188. 6. Guba M, Bosserhoff AK, Steinbauer M, Abels C, Anthuber M, Buettner R and Jauch KW. Overexpression of Melanoma Inhibitory Activity (MIA) enhances extravasation and metastasis of A-mel 3 melanoma cells in vivo. Br J Cancer 2000;83:1216–1222. 7. Bosserhoff AK. Recombinant expression of functional active MIA (Melanoma Inhibitory Activity) protein for mutation analysis using the RTS system. In: Cell-Free Protein Expression, Swartz James R (ed), Springer, 2003, pp. 173–179. 8. Lorenz P and Thiesen HJ. In-vitro translation of KRAB zinc finger transcriptional repressor proteins and their interaction with their TIF1b co-repressor. In: Cell-Free Protein Expression, Swartz James R (ed), Springer, 2003, pp. 151–157. 9. Betton J-M. Using maltose-binding protein fragment complementation to probe protein– protein interactions by co-expression in the RTS system. In: Cell-Free Protein Expression, Swartz James R (ed), Springer, 2003, pp. 143–149. 10. Betton JM and Hofnung M. In vivo assembly of active maltose binding protein from independently exported protein fragments. EMBO J 1994;13:1226–1234. 11. Ma B, Tsai CJ and Nussinov R. Binding and folding: in search of intramolecular chaperonelike building block fragments. Protein Eng 2000;13:617–627. 12. Nemetz C, Wessner S, Schweitzer R, Graentzdoerffer A and Buchberger B. Rapid protein engineering by expression-PCR. In: Cell-Free Protein Expression, Swartz James R (ed), Springer, 2003.
30 13. Schra¨ml et al. Rapid generation of protein variants and subsequent analysis by surface plasmon resonace. In: Cell-Free Protein Expression, Swartz James R (ed), Springer, 2003, pp. 69–79. 14. Cronan JE. Biotination of proteins in vivo. J Biol Chem 1990;265,18(Issue of June 25): 10327–10333. 15. Fernholz, et al. Production of a specifically labeled protein in mg quantities for NMR analysis. In: Cell-Free Protein Expression, Swartz James R (ed), Springer, 2003, pp. 55–60. 16. Schlo¨ssmann T, et al. In situ mono-biotinylation of cell-free expressed proteins using the AviTag technology. In: Cell-Free Protein Expression, Swartz James R (ed), Springer, 2003, pp. 61–67. 17. Biekofsky RR, Martin SR, McCormick JE, Masino L, Fefeu S, Bayley PM and Fenney J. Thermal stability of calmodulin and mutants studied by 1H–15N HSQC NMR measurements of selectively labeled 15N-Ile proteins. Biochemistry 2002;28:6850–6859. 18. Kelly MJ, Ball LJ, Krieger C, Yu Y, Fischer M, Schiffmann S, Schmieder P, Kuhne R, Bermel W, Bacher A, Richter G and Oschkinat H. The NMR structure of the 47-kDa dimeric enzyme 3,4-dihydroxy-2-butanone-4-phosphate synthase and ligand binding studies reveal the location of the active site. Proc Natl Acad Sci USA 2001;6:13025–13030. 19. Ross JBA, Szabo AG and Hogue CWV. Enhancement of protein spectra with tryptophan analogs: fluorescence spectroscopy of protein–protein and protein–nucleic acid interactions. Methods Enzymol 1997;278:151–190. 20. Sengupta K, et al. Incorporation of fluorescence labels into cell-free produced proteins. In: Cell-Free Protein Expression, Swartz James R (ed), Springer, 2003, pp. 81–88. 21. Budisa N, et al. Expression of ‘Tailor-Made’ proteins via incorporation of synthetic amino acids by using cell-free protein synthesis. In: Cell-Free Protein Expression, Swartz James R (ed), Springer, 2003, pp. 89–98. 22. Skelton NJ and Chazin WJ. Solution structure determination of proteins by nuclear magnetic resonance spectroscopy. Drugs Pharm Sci 2000;101:683–726. 23. Riek R, Flaux J, Bertelsen EB, Horwich AL and Wuthrich K. Solution NMR techniques for large molecular and supramolecular structures. J Am Chem Soc 2002;124:12144–12153. 24. McIntosh LP and Dahlquist FW. Biosynthetic incorporation of 15N and 13C for assignment and interpretation of nuclear magnetic resonance spectra of proteins. Quart Rev Biophys 1990; 23:1–38. 25. Lian LY and Middleton DA. Labeling approaches for protein structural studies by solution state and solid-state NMR. Prog Nucl Mag Reson Spect 2001;39:171–190. 26. Kim D-Y, et al. Cell-free expression of proteins containing multiple disulfide bonds. In: CellFree Protein Expression, Swartz James R (ed), Springer, 2003, pp. 125–131. 27. Lehmann D and Wedekind F. Optimization of cell-free expression of FAD-dependent D-amino acid oxidase. In: Cell-Free Protein Expression, Swartz James R (ed), Springer, 2003. 28. Maurer P, et al. Cell-free synthesis of membrane proteins on a preparative scale. In: Cell-Free Protein Expression, Swartz James R (ed), Springer, 2003. 29. Langlais C, et al. The linear template generation set: Optimization of protein expression in the RTS 100 HY. BIOCHEMICA, 03/2003. 30. Liu P and Ma D. Expression of recombinant chemokine-like factor 1 with a cell-free protein biosynthesis system. In: Cell-Free Protein Expression, Swartz James R (ed), Springer, 2003, pp. 165–171. 31. Madin K, Sawasaki T, Ogasawara T and Endo Y. A highly efficient and robust cell-free protein synthesis system prepared from wheat embryos: Plants apparently contain a suicide system directed at ribosomes. Proc Natl Acad Sci USA 2000;97:559–564. 32. Akbergenov RZ, et al. Complementary interaction between the central domain of 18S rRNA and the 50 untranslated region of mRNA enhances translation efficiency in plants. In: Cell-Free Protein Expression, Swartz James R (ed), Springer, 2003, pp. 199–208.
31
Protein expression and refolding – A practical guide to getting the most out of inclusion bodies Lisa D. Cabrita and Stephen P. Bottomley* Monash University, Department of Biochemistry and Molecular Biology, School of Biomedical Sciences, P.O. Box 13D, Melbourne, Victoria 3800, Australia Abstract. The release of sequence data, particularly from a number of medically and biotechnologically important genomes, is increasing in an exponential fashion. In light of this, elucidating the structure and function of proteins, particularly in a ‘‘high throughput’’ manner, is an important quest. The production of recombinant proteins however is not always straightforward, with a number of proteins falling prey to low expression problems, a high susceptibility to proteolysis and the often despised production of inclusion bodies. Whilst expression as inclusion bodies can often be advantageous, their solubilization and renaturation is often a time consuming and empirical process. In this review, we aim to outline some of the more common approaches that have been applied to a variety of proteins and address issues associated with their handling. Keywords: inclusion body, protein folding, protein aggregation, protein engineering, protein expression, refolding, protein purification
Introduction One of the areas of great importance in this post-genomic era is the ability to rapidly express and purify a protein of interest. With the vast amount of sequence data now available, numerous sites around the world are now attempting structural genomic projects in which a vast array of proteins are being expressed, purified and structurally defined. In addition to these highthroughput approaches almost all biomedical labs, academic or industrial, are expressing proteins of interest for structural, functional and therapeutic investigations. The general scheme in this research involves rapid cloning of the genes of interest and then expression of the protein, usually in the host E. coli. At present, E. coli remains the ‘‘king’’ of expression hosts, due to its advantages such as ease of handling, cost effectiveness and high success rates. However, it has some associated disadvantages such as the inability to perform many post-translational modifications and expression of the protein product often leads to its deposition as inclusion bodies. Inclusion body formation is often dreaded because there are a bewildering number of approaches which a researcher can utilize to obtain soluble proteins. However, in the past few years, there has been an increase in products and protocols which are generally applicable to the inclusion body problem. New strains of E. coli and fusion partners are now available, which offer a range of options to minimize, or *Corresponding author: E-mail:
[email protected] BIOTECHNOLOGY ANNUAL REVIEW VOLUME 10 ISSN: 1387-2656 DOI: 10.1016/S1387-2656(04)10002-1
ß 2004 ELSEVIER B.V. ALL RIGHTS RESERVED
32 maximize, inclusion body formation. In addition, a more rational approach to protein refolding has become established, which allows the user to establish very quickly, whether refolding is a viable option. In this review, we will initially outline the theoretical background to protein folding and then we will focus on firstly, the ways in which inclusion body formation can be avoided during the growth of E. coli and then techniques that can be used to refold any inclusion bodies that are formed. Mechanisms of protein folding and misfolding The classic work of Anfinsen demonstrated that it is possible to reversibly unfold and refold proteins, indicating that the amino acid sequence of a protein contains all the information required for successful folding [1]. Moreover, these experiments also revealed that a protein is able to form its native conformation whilst avoiding the vast array of other structures accessible to the polypeptide chain as it folds. These alternative structures are often stable, non-native conformations that self-associate and result in the formation of aggregated material. The process of folding/misfolding therefore, is of great biomedical importance, as protein aggregation forms the basis for an increasing number of diseases such as cystic fibrosis, liver cirrhosis and many neurodegenerative diseases [2]. In the in vitro context, protein aggregation is a very common feature that occurs during recombinant protein expression in E. coli, which dramatically reduces the yield of soluble, useful protein. The folding pathway of a protein is extremely complex. Pioneering work by Alan Fersht and colleagues has shown that many small proteins, of less than 150 amino acid residues, fold via a two-state pathway in which only the native and unfolded ensembles are significantly populated [3]. A combination of protein engineering and theoretical techniques have revealed that a folding protein follows multiple paths to a native-like transition state ensemble which collapses to form the native conformation. In contrast to small proteins, proteins consisting of more than 150 amino acid residues generally fold via multi-state pathways in which intermediate species are populated (Fig. 1). Some of these intermediates (Io) are obligatory and lie on the productive folding pathway, whereas others (IA) lie on non-productive pathways which culminate in the formation of aggregated material. There is a plethora of examples demonstrating that factors such as mutation, temperature and protein concentration can disrupt the energetics of the folding reaction, making misfolding and aggregation a more likely (though usually undesirable) endpoint. During recombinant protein expression where inclusion body formation can result, many of these factors as well as the strength of the promoter and cell type play a significant role, as discussed below. Recent experimental data has lead to the hypothesis that all proteins have the potential to form aggregates [4]; in addition the converse also holds that under specific solution conditions, a denatured protein should be able to successfully refold to its native state.
33 U
Io
N
Factors that disrupt the productive folding
landscape:
IA + IA
temperature, pH, salt concentration, mutation
A Fig. 1. Possible routes for the folding of a protein. From the unfolded state (U), the protein can proceed through an obligatory intermediate (Io) which leads to successful adoption of the native state (N). Alternatively misfolding can occur through an off-pathway reaction in which an aggregation prone intermediate (IA) is formed, either directly from U or through structural changes in Io, which can self-associate and form aggregates (A).
The problem faced, however, is identification of these specific conditions, which is a bottleneck in dealing with refolding of solubilized inclusion bodies. In the rest of this review we will discuss many of the options available to firstly avoid inclusion body formation and secondly refold proteins that are present within inclusion bodies. Recombinant protein expression in E. coli It has been well described in several excellent reviews, that the expression of recombinant proteins within E. coli is affected by several factors. These include, but are not limited to: plasmid copy number, mRNA stability, upstream elements, temperature and codon usage and these are covered in detail elsewhere [5–7]. It appears therefore, that successful production of protein is a combination of these factors and to a degree, some might argue, a stroke of luck. So while the expression of an uncharacterized protein is at best, difficult to predict, numerous advances have been made to improve both expression and solubility. This has been made possible through the development of novel tags, fusion partners, and vector systems, such that the choices for recombinant expression in E. coli are forever expanding. E. coli strains, promoters and fusion partners In regulating protein expression, the choice of promoter/vector system is important, as one system may be more suited to a target than another. There are several promoters that are available (Table 1), perhaps the most well known variety being T7-derived, as found in the pET vectors (Novagen). These IPTG inducible promoters have in the past been associated with ‘‘leaky’’ expression prior to induction, which is a pitfall for proteins that are toxic to cells. This however, has been addressed with the co-transformed pLysS and pLysE plasmids that are able to act as strong repressors. This issue aside, the
34 Table 1. Examples of promoters that are used in E. coli expression. Promoter
Example vector
Inducer
T7 T5 trc tac lac trp araB PL (l) phoA PLtetO-1
pET (Novagen) pQE (Qiagen) pTrcHis (Invitrogen) pMAL (New England Biolabs) pTriplEx2 (Clontech) pLEX (Invitrogen) PBAD (Invitrogen) pKC30 pBKIGF2B-A pLP-PROTet-6xHN (Clontech)
IPTG IPTG IPTG IPTG IPTG tryptophan L-arabinose temperature shift (42 C) phosphate tetracycline
T7 is a strong promoter which allows high level expression of target proteins. Unfortunately, it is overexpression which can also lead to the formation of intracellular inclusion bodies, sometimes in far greater yields than the soluble version. Again, this can sometimes be regulated, by altering the expression time, induction temperature and IPTG levels [5]. The T7-based vectors have previously included affinity tags for protein detection and purification, though they are also now designed with other features. Of interest are the ‘‘fusion partners,’’ that in general, ‘‘fuse’’ the protein of interest with another, to circumvent a solubility issue and simultaneously aid in purification. The classic GST fusion system has been successful for a large number of proteins [8–11], however more recently, there has been the development of alternatives (Table 2 and [12]). One of these is the 42 kDa maltose binding protein (MBP), which has been successfully used to aid in increasing the solubility of a range targets [13–15] and has been shown recently to have little effect during crystallization [16] – an advantage, as it is generally accepted that ‘‘tags’’ need to be removed to aid in crystallography. A similar tag is Nus A (Novagen), a 54 kDa protein, which was identified as being the most soluble protein out of a pool of almost 4000 E. coli proteins [17]. In addition to aiding in folding, some tags such as the S-tag (Novagen), for instance, which comprises of 15 amino acids, can be used to detect, quantitate and purify soluble protein [18]. Poor expression of a target protein can amongst many things, be associated with codon bias – that is, codons used infrequently in the prokaryotic system. The Rosetta (Novagen) and BL21-CodonPlus strains (Stratagene) of E. coli have the advantage of co-expressing plasmids that code for the rare tRNAs and have been successfully used recently for a number of proteins (cellobiose phosphorylase [19], phosphoenolpyruvate carboxylase [20], Dictyostelium 5NT [21]). Moreover, the development of E. coli that accommodates disulfide bond formation also can improve the yields of certain targets. The AD494(DE3) strain (Novagen) has a single mutation in the thioredoxin reductase gene,
35 Table 2. Examples of fusion partners and tags which aid solubility and tracking protein expression. Tag
Size
Location
Refs.
Glutathione S-transferase (GST) Maltose Binding Protein (MBP) S-Tag Chloramphenicol acetyltransferase NusA Ubiquitin Thioredoxin Z-domain (derived from Protein A)
26 40 15 24 54 76 11 58
N N, C N, C, internal N N N N N
[11] [15] [18] [90] [17] [91] [92] [93]
kDa kDa aa kDa kDa aa kDa aa
and has enabled the production of proteins including C1-inhibitor (containing two disulfide bonds) [22], chitinases (one disulfide) [23] and a superoxide dismutase (two disulfides) [24]. The Origami strain (Novagen), however, combines a double mutation in both the thioredoxin reductase and the glutathione reductase gene, hence providing a more favorable oxidizing environment within the cytoplasm. In addition, the recent Rosetta-Gami strain (Novagen) is a powerful combination of the Rosetta and Origami strains, encompassing rare codon usage and disulfide bond formation. Such a strain may be particularly useful for eukaryotic or intracellular proteins, which typically display alternative codon usage and require disulfide bond formation, respectively. Cell-free expression Cell-free expression (‘‘in vitro transcription–translation’’) has been an invaluable tool for many years, exploiting the E. coli, wheat germ and rabbit reticulocyte systems to generate protein. In more recent times, the demand for cell-free expression has increased, particularly as it is seen as a viable alternative to other expression systems for recalcitrant proteins. This technology was developed some 30 years ago, with the production of rat growth hormone, using an S30 extract derived from mice [25]. It has since been brought to the forefront by some ground breaking work by Spirin and coworkers [26] and also by Kim and Choi [27]. Spirin developed a ‘‘continuous-flow’’ system whereby amino acids and energy sources are supplied into a reaction chamber, while synthesized proteins and used substrates are removed using an ultrafiltration membrane. In contrast, Kim and Choi probed a ‘‘semicontinuous’’ method, involving the use of two chambers separated by a dialysis membrane, allowing for the replenishment of substrates and removal of by-products. With this system they were able to produce 1.2 mg/ml of chloramphenicol o-acetyltransferase, in 14 h and since then, have increased its efficiency by over 70% to produce 0.3 mg of protein per hour [28].
36 Roche also now offers the in vitro transcription–translation ‘‘Rapid Translation System,’’ which has been successful for a number of targets and whose capabilities are expanding, particularly with the incorporation of molecular chaperones, detergents and other additives to assist in folding. Although not seen to be as efficient as the E. coli system at present, the cell free system based upon wheatgerm has also been used [29–31]. Being a eukaryotic system, it is being explored as an alternative, particularly for the production of some proteins that will not otherwise express in E. coli. The cell-free approach to protein expression has suffered in the past due to concerns relating to its low efficiency, membrane clogging and as an effect, questionable reproducibility. At present however, many advancements are being made, such that it may encourage more ‘‘mainstream’’ use and with continued improvements, it may perhaps be a competitive rival to the traditional E. coli expression system in the future.
Preparation of inclusion bodies Whilst a researcher may attempt all of the tried and proven methods of obtaining soluble expression, it is also to be noted that the inclusion bodies persist and that rather than work against them, it is best to try and work with them (Fig. 2). Their formation can at times be advantageous, particularly if the protein is toxic to the cell, or is likely to be the subject of proteolytic attack. The inclusion body itself is a dense amorphous aggregate of misfolded protein and is generally formed if the protein has a high propensity to misfold and aggregate, or if the cellular protein production machinery itself is overwhelmed and hence unable to operate efficiently. It has been predicted that as much as 20–40% of human gene constructs will express as inclusion bodies in E. coli [32]. Being in the post-genome era, the view towards ‘‘high throughput’’ cloning/expression/purification allows for the rapid isolation of numerous proteins. However ‘‘difficult targets,’’ such as those that persist to express as inclusion bodies, may be left behind in such a pursuit. This is of course understandable, considering that the preparation, solubilization and renaturation of inclusion bodies is indeed laborious and varies greatly for any individual protein. Providing set ‘‘rules’’ therefore, is virtually impossible and finding conditions that enable a protein to refold from inclusion bodies can be likened to the factorial screens used in crystallography – it may require trial and error and many attempts. Inclusion bodies are easily identifiable, with a morphology similar to strings or clusters [33] and their isolation from other cellular components can be easily achieved. They are contained within the insoluble fraction that is obtained after cell lysis and subsequent centrifugation. While it is generally accepted that inclusion bodies can comprise 40–90% of the target protein, successful renaturation is dependent upon their level of purity. Studies with lysozyme have
37 Lyse cells (sonication, french press) Centrifuge, retain pellet (insoluble fraction)
Purify inclusion bodies Chromatography (denaturing conditions)
Centrifugation
Gel Filtration, Ni-NTA
Triton X-100, high salt, EDTA
Solubilization 4−6 M GdnHCl
Dialysis (step)
Activity
6−8 M Urea
Refolding On-column refolding (gradient, stepwise)
0.5-2% SDS
Dilution (step, dropwise)
Folded or Aggregated? Circular Dichroism
Non-denaturing PAGE
Size exclusion chromatography
Light scattering
Ultracentrifugation
Fig. 2. A schematic outlining of a procedure for the purification from inclusion bodies.
shown that the presence of proteinaceous contaminants, during renaturation, decreases the efficiency of refolding, presumably as protein contaminants can promote co-aggregation [34]. Inclusion bodies are therefore washed several times with detergents such as Triton X-100 (0.1 ! 4% (v/v)), sodium deoxycholate (2% (w/v)), sarkosyl and even low molar concentrations (0.5 ! 1 M) of denaturants such as guanidine hydrochloride or urea. They essentially remove the cellular contaminants, that adsorb onto the hydrophobic inclusion bodies. After washing, the inclusion bodies are solubilized, usually with the use of 4–6 M GdnHCl or 8 M urea. GdnHCl is the stronger denaturant of the two, because during prolonged incubations at alkaline pH, urea can suffer from the formation of isocyanate ions, which can modify amino acid side chains [35,36]. Aside from denaturants, solubilizing alternatives also include detergents such as sarkosyl [37,38], SDS [39] and also alkaline pH [40,41]. It is important to note that during unfolding, there must be reduction of any disulfide bonds present, which is usually achieved using b-mercaptoethanol or DTT (5 mM ! 100 mM).
38 This is in light of a pivotal study which demonstrated that disulfides persisted in high denaturant concentrations in the absence of reducing agents [42]. Normally, solubilization can be performed in any buffer that is compatible with the protein of interest (Tris/Hepes/Phosphate), generally a neutral pH [7–8] is suitable for solubilization. Upon solubilizing the inclusion bodies, ample incubation time should be incorporated to allow for complete unfolding. Incubation at either room temperature or 30 C for 1–4 h is an acceptable time frame, while some have opted for 16–24 h at 4 C. At times, inclusion bodies may be difficult to solubilize, hence the use of some form of agitation and increase in temperature may be necessary. Some researchers also choose to purify the inclusion bodies further, by incorporating column chromatography (conducted under denaturing conditions) such as ion exchange, size exclusion or metal affinity [43–45]. Again this has been seen to enrich the proportion of monodispersed protein which can increase the yield of the refolded target protein [46]. More recently, Gu and colleagues took the approach of purifying inclusion bodies exclusively by gel filtration using a macroporous medium (e.g., 4% agarose). By coupling with the use of a French press to reduce cell debris, the column matrix excluded only the inclusion bodies and hence allowing their separation [47]. This was used as an alternative to the standard centrifugation steps mentioned earlier. Refolding solubilized inclusion bodies After solubilization, the renaturation process aims to effectively remove the denaturant and thiol reagents and allow the protein to refold. Whilst this appears trivial, the refolding process is a competing reaction with misfolding and aggregation events and as such, numerous factors impact on its success. Studies on proteins, especially those larger than 150 amino acids have shown that for many, folding involves the formation of an intermediate species, often resembling a ‘‘molten globule.’’ Such a species represents a branch point within folding, where it may also lead to misfolding and subsequent protein aggregation. With the molten globule containing some secondary structure, but little tertiary structure, hydrophobic patches normally buried within the protein are exposed to solvent [48]. Under appropriate conditions, therefore, these regions can promote inappropriate interactions that may lead to aggregation. Several issues are therefore of importance in aiming to minimize the aggregation reaction: final concentration of protein to be refolded, the components of the refolding buffer and method of refolding. Importance of protein concentration during refolding The amount of protein refolded has an impact on the yields that will be obtained. As folding competes with aggregation, it is generally acknowledged
39 that refolding at low protein concentrations (10–100 mg/ml) is the most successful approach. There have been instances where proteins have refolded into high concentrations of up to 5 mg/ml, prochymosin [49], lysozyme [50] carbonic anhydrase [51], albeit in low concentrations of denaturant. On the most part, however, the lower the final protein concentration attained during refolding, the greater the efficiency. Components of the refolding buffer The components of a refolding buffer vary widely, depending on the protein of interest with pH, ionic strength, redox conditions and ligands all influencing the outcome (Tables 3 and 4). Most commonly Tris or Hepes based buffers at neutral pH with NaCl (between 50 mM and 500 mM) are used, however again this depends on the protein of interest. There are also numerous additives which can be included, that have had varying success with numerous targets. They include detergents, polar additives, weak chaotrophs, osmolytes and cations (Table 3). These additives to one degree or another, either act as stabilizers (stabilizing the native state/solubilizing intermediates) or promote correct folding by preventing aggregation. Another common approach is to include proteinase inhibitors (PMSF, aprotinin, leupeptin, etc.) within the refolding buffer, if the protein of interest is prone to proteolysis.
Table 3. Additives that have been successfully used for refolding. Additive
Concentration
Effect
L-Arginine
0.4–0.5 M 10–50% 0.4 M
stabilizer stabilizer stabilizer
up to 1 M 0.1–2 M 0.1–1 M 0.1–1% 0.01% 0.3 mM 30 mM 0.1% up to 4 M up to 0.05% (w/v) 1–3 M
solubilizer chaotroph chaotroph detergent detergent detergent detergent detergent detergent osmolyte osmolyte salt salt cation chelator buffer
Glycerol Sucrose/glucose Non-detergent sulfobetaine (NDSB) 256/201 Urea Guanidine HCl Triton X-100 Tween-80 Lauryl maltoside CHAPS SDS Lauroylsarcosine PEG 3350 TMAO Sodium citrate/Sulfate NaCl/Ammonium sulfate MgCl2/CaCl2 EDTA Tris
0.2–0.5 2 mM–10 mM 20 mM 0.4–1 M
40
Table 4. Examples of proteins refolded from inclusion bodies. Protein
IB purification
Solubilization
Refolding
Technique
Ref.
Human IL-15
Ni-NTA
[94]
2 M urea, 5 mM BME
Ni-NTA
[95]
Plasminogen activator-1 Resistin
0.05% Tween 80 Ni-NTA
8 M Urea 1 mM DTT, 10% glycerol 1% SDS 20 mM BME pH 9 4 M GdnHCl 6 M GdnHCl
Gradient
Clostridium difficile
2 M Urea 2% Triton X-100 –
Dilution Step dialysis
[96] [56]
Alpha lytic protease
S sepharose
5 M Urea
Pulse dilution
[45]
ACC synthase Glutamyl-tRNA reductase
0.1% Triton X-100 2 M Urea 0.2% Triton X-100 0.5% NP-40/3 M urea/ Gel filtration (superdex 200)
6 6 5 6
1 M NaCl, 0.01% Tween 80 1 mM DTT, 0.1% mannitol 2 M urea/1 M urea, pH 11 5–20% sucrose/glucose 200 mM methionine 1% glycerol, 33 mM Chaps 20% glycerol
Dilution Ni-NTA
[58] [64]
Gel filtration
[65]
Transposase Tc1A
M urea, 10 mM DTT M GdnHCl mM DTT M GdnHCl
10% glycerol, 5 mM MgCl2, 1 mM DTT
41 Disulfide bond formation For proteins with native disulfide bonds a redox system during renaturation is required, for their correct formation. The combination of reduced and oxidized forms in molar ratios ranging from 1:1 up to 10:1 are generally used to form disulfides correctly. Glutathione (GSH/GSSG) is a common reagent, however the combination of cysteine and cystine or DTT/oxidized glutathione are alternatives. While air/molecular oxygen is suitable to promote disulfide bond formation, a redox system accelerates the shuffling of disulfide bonds. It is generally catalyzed by trace amounts of metal ions and also by slightly alkaline pH (8–9) [52]. On the contrary, to minimize the effects of oxygen oxidizing thiols, the presence of EDTA can be beneficial [53]. The time taken for the ‘‘shuffling’’ to take place varies between proteins, anywhere from 2 h to 150 h, though 16–48 h is common place for efficient disulfide shuffling, although this must be determined empirically. A recent report using controlled air oxidation demonstrated the successful production of prochymosin. Menzella and colleagues carried out oxidation by introducing air at a flowrate of 0.1 L/min, in the presence of Cu2 þ as a catalyst. With this and the use of additives (L-arginine), they were able to recover up to 67% of active protein from inclusion bodies [49]. There are however, conflicting reports on the impact of inappropriate disulfide bonding, with suggestions that it may not necessarily lead to aggregation, considering a carboxymethylated version of lysozyme (cysteines blocked) was still found to be aggregation-prone [54]. Methods of refolding There are several methods to refold proteins including: dialysis, dilution and use of column chromatography techniques. The method which is used depends on the propensity of the protein to aggregate and the kinetics of refolding. The temperatures at which refolding takes place can vary, though to minimize aggregation, 4 C is best. Dialysis Here, the concentrated denatured protein is dialysed against a refolding buffer, such that the concentration of the denaturant decreases as it is bufferexchanged. It is this slow removal process which allows for the refolding to take place [15,55]. Unfortunately, the slow removal of the denaturant often results in the formation of the exposure of long lived intermediate species over a long period of time and hence there may be increased propensity for the protein to aggregate. A variation of a one-step dialysis as previously described, is the use of refolding buffers with decreasing denaturant concentrations [56]. By dialysing in
42 a step-wise fashion (usually 1 to 3 progressively lower denaturant concentrations), this allows for an equilibrium to be established. Again, whilst the refolding is more controlled in this environment, long-lived intermediate species can still present a challenge. Dilution The dilution method can be described as ‘‘rapid’’ or ‘‘slow.’’ In the rapid dilution method, the denatured protein is delivered to a refolding buffer, such that in a very short period of time, the concentrations of both the protein and denaturant decrease rapidly. For instance, if the protein were denatured in 8 M urea and diluted 10-fold into buffer, the final concentration of the denaturant is 0.08 M. Like dialysis, the aim of dilution is to remove the denaturant, so that its final concentration would be low. There are instances though, where refolding into low molar concentrations (1–2 M) aids in the solubility of the protein. The period of time used in ‘‘rapid’’ dilution may be detrimental to the proteins, particularly if they refold over a time period of minutes to hours. By forcing the protein to adopt its conformation in a limited time frame, it may increase its chances of misfolding and hence aggregating, however this approach has been successful for a range of proteins (monogalactosyldiacylglycerol [57], 1-aminocyclopropane-1-carboxylate synthase [58], antichymotrypsin [59], transglutaminase [60]). Slow dilution is an alternative, as it is a more gentle approach, by dramatically decreasing the effective concentration of the refolding protein. The ‘‘dropwise’’ or ‘‘pulsed dilution’’ method, involves the solubilized inclusion bodies being delivered very slowly to the refolding buffer (using a pump) and has been used successfully for proteins such as a1-antitrypsin [61] and also a-lytic protease [45]. On-column refolding This approach has been used for several proteins with success and offers an alternative where other methods may not be applicable. If a protein is tagged with a hexa-histadine, it can be immobilized on a Ni2 þ affinity column. By applying buffers with a decreasing concentration of denaturant, either stepwise or by using a gradient, the protein can be refolded and then eluted [62,63]. Gel filtration is an alternative, whereby the denatured protein is loaded onto the column and refolds as it is passed through with buffer [64,65]. A variation of this is to equilibrate the column with a linear gradient, with decreasing concentrations of denaturant, so the protein refolds gradually [66]. Flow rates required for successful on-column refolding appear to vary with slower flow rates improving recovery in some cases [47] whilst in others, faster flow rates were best [67].
43 Recently, the on-column refolding technique has been improved by immobilizing common ‘‘foldases’’ onto the column matrix and exploiting them as a folding ‘‘platform.’’ This has been used successfully with GroEL [68,69] and DsbA/DsbC [70]. One associated concern with on-column refolding is the clogging of filters by aggregated protein, however, if performed with samples that have been carefully filtered/centrifuged and with a sensible column matrix, may minimize these problems. Other refolding techniques One technique which is a departure from the aforementioned ‘‘classic’’ techniques involves the use of reverse micelles, which has been explored recently. Vinogradov and colleagues trapped their enzyme in a water–sodium bis2-ethylhexyl sulfosuccinate-isooctane reverse micellar system and reported the recovery of monomeric protein, where other techniques had failed. By varying the size of the micelles, they were able to manipulate the degree of oligomerization of the protein [71]. Other folding aids There are several molecular chaperones that can be incorporated either in vivo or in vitro to aid in folding. The most well known E. coli chaperones include GroEL-GroES, DnaK-DnaJ-GrpE (Hsp70) and also ClpA/ClpB (Hsp100) which have been used successfully in renaturation studies [72–75]. While molecular chaperones can promote correct folding, foldases can accelerate the process. The three types of foldases include: peptidyl prolyl cis/ trans isomerases (PPI’s) (arranges prolines into correct conformation, an otherwise a lengthy process) [76], disulfide oxidoreductase (DsbA) and disulfide isomerase (DsbC) (which promote disulfide bonds, found in E. coli) [77,78] and protein disulfide isomerase (PDI) (a eukaryotic protein catalyzing oxidation and isomerization) [79]. Folding screens As trying to refold a protein can be a time-consuming process, several commercial screens are now available to focus on potential conditions that return the best yield of monomeric protein. One such kit is Hampton Research’s FoldIt screen, which uses a variety of additives, redox conditions and pH and similar kits are available through Novagen and Sigma-Aldrich. Such kits are the result of sparse screens that have been successful in the past [80–82] to refold a number of proteins. With the commercially prepared kits, it enables a number of refolding conditions to be sampled simultaneously on a small scale, the aim being to find a suitable condition that can be up-scaled for purification.
44 An elegant example of using factorial screens was with work done on procathepsin S and cathepsin S, which could not be previously refolded, in vitro. An initial screen was used to identify conditions for folding, targeting both L-arginine and pH as two important factors. While pH was important for both, arginine was more beneficial for procathepsin only. Once these conditions were established, it was then followed up, using a second screen to improve yields and hence produce the proteins on a larger scale. It was concluded that procathepsin required detergent, arginine and a redox system, while cathepsin only required glycerol and the redox couple [83]. This in itself illustrated the powerful nature of the factorial screen and how seemingly similar proteins may have very different requirements for refolding.
Correctly folded or aggregated material? Once refolding is complete, one must determine whether the procedure has yielded folded or aggregated protein and whether disulfide bonds (if any) have been formed correctly. Perhaps the simplest means of assessing the quality of the protein is by using a known activity assay. While this is possible for some, the need to determine the nature of the protein using other techniques is important, considering the high throughput production of proteins from different genomes will undoubtedly yield targets with no known function or structure. The use of size-exclusion chromatography and non-denaturing PAGE can be employed, as is circular dichroism (CD) which reports on the secondary structure. There are distinct features in a typical CD scan that report the presence of a helical, b sheet or random coil structure (Fig. 3). While proteins will generally be a combination of a helical and b sheet content, for aggregated
α helix
20 15
β sheet
ellipticity
10
turn
5 0 −5
−10
random coil 180
200
220
240
260
wavelength (nm) Fig. 3. A typical far-UV CD spectrum of classical secondary structure elements.
45 material, a random coil may predominate. Fluorescence techniques such as dynamic light scattering [84], which determines the average diameter of particles in solution and lateral turbidimetry [13] can also report the presence of aggregates [85], as can ultracentrifugation [86] and also electron microscopy [87]. Assessing the nature of disulfide bonds can be defined by a number of methods. Perhaps the most straightforward is the use of reducing and nonreducing gel electrophoresis. In the absence of reducing agent, there should be a defined ‘‘band shift,’’ accompanied with any additional bands that were record of linked domains. For a single domain protein, a ‘‘laddering’’ effect is a telltale sign of intermolecular disulfide bond formation. The use of thiol specific dyes (DTNB (Ellman’s reagent)) [88], mass spectrometry [89] and gel electrophoresis (‘‘cysteine counting’’) [52] are all common techniques that are also employed. Conclusion Handling inclusion bodies is time consuming and often a frustrating trial-anderror process. There are also no clear rules that apply to all proteins, as where an approach suitable to one may also be just as detrimental to another. Despite this, working with and conquering the inclusion body may be beneficial, particularly when it is impossible to express soluble protein in reasonable amounts. As more proteins are successfully refolded the rules may become clearer. This will aid biotechnology and may also aid medicine as the tricks which are used to refold proteins may be useful in preventing the array of diseases in which misfolding plays a central role. Abbreviations IPTG GST MBP CD DTT
isoproyl-thiogalactoside, glutathione-S-transferase, maltose binding protein, circular dichroism, dithiothreitol
Acknowledgments The authors would like to thanks Michelle Chow for her help with the manuscript. This work was in part funded by both the Australian Research Council and the National Health and Medical Research Councils of Australia. SPB is a Monash University senior Logan research fellow and an RD Wright Fellow of the NHMRC.
46 References 1. Anfinsen CB. Principles that govern the folding of protein chains. Science 1973;181: 223–230. 2. Horwich A. Protein aggregation in disease: a role for folding intermediates forming specific multimeric interactions. J Clin Invest 2002;110:1221–1232. 3. Daggett V and Fersht AR. Is there a unifying mechanism for protein folding? Trends Biochem Sci 2003;28:18–25. 4. Dobson CM. Protein misfolding, evolution and disease. Trends Biochem Sci 1999;24: 329–332. 5. Baneyx F. Recombinant protein expression in Escherichia coli. Curr Opin Biotechnol 1999;10: 411–421. 6. Jonasson P, Liljeqvist S, Nygren PA and Stahl S. Genetic design for facilitated production and recovery of recombinant proteins in Escherichia coli. Biotechnol Appl Biochem 2002;35: 91–105. 7. Swartz JR. Advances in Escherichia coli production of therapeutic proteins. Curr Opin Biotechnol 2001;12:195–201. 8. Heidari M, Rice KL, Kees UR and Greene WK. Expression and purification of the human homeodomain oncoprotein HOX11. Protein Expr Purif 2002;25:313–318. 9. Park SM, Jung HY, Chung KC, Rhim H, Park JH and Kim J. Stress-induced aggregation profiles of GST-alpha-synuclein fusion proteins: role of the C-terminal acidic tail of alphasynuclein in protein thermosolubility and stability. Biochemistry 2002;41:4137–4146. 10. Lipari F, McGibbon GA, Wardrop E and Cordingley MG. Purification and biophysical characterization of a minimal functional domain and of an N-terminal Zn2 þ -binding fragment from the human papillomavirus type 16 E6 protein. Biochemistry 2001;40: 1196–1204. 11. Smith DB and Johnson KS. Single-step purification of polypeptides expressed in Escherichia coli as fusions with glutathione S-transferase. Gene 1988;67:31–40. 12. Terpe K. Overview of tag protein fusions: from molecular and biochemical fundamentals to commercial systems. Appl Microbiol Biotechnol 2003;60:523–533. 13. Nomine Y, Ristriani T, Laurent C, Lefevre JF, Weiss E and Trave G. Formation of soluble inclusion bodies by hpv e6 oncoprotein fused to maltose-binding protein. Protein Expr Purif 2001;23:22–32. 14. di Guan C, Li P, Riggs PD and Inouye H. Vectors that facilitate the expression and purification of foreign peptides in Escherichia coli by fusion to maltose-binding protein. Gene 1988;67:21–30. 15. Sachdev D and Chirgwin JM. Solubility of proteins isolated from inclusion bodies is enhanced by fusion to maltose-binding protein or thioredoxin. Protein Expr Purif 1998;12: 122–132. 16. Smyth DR, Mrozkiewicz MK, McGrath WJ, Listwan P and Kobe B. Crystal structures of fusion proteins with large-affinity tags. Protein Sci 2003;12:1313–1322. 17. Davis GD, Elisee C, Newham DM and Harrison RG. New fusion protein systems designed to give soluble expression in Escherichia coli. Biotechnol Bioeng 1999;65:382–388. 18. Kim JS and Raines RT. Ribonuclease S-peptide as a carrier in fusion proteins. Protein Sci 1993;2:348–356. 19. Rajashekhara E, Kitaoka M, Kim YK and Hayashi K. Characterization of a cellobiose phosphorylase from a hyperthermophilic eubacterium, Thermotoga maritima MSB8. Biosci Biotechnol Biochem 2002;66:2578–2586. 20. Chen LM, Omiya T, Hata S and Izui K. Molecular characterization of a phosphoenolpyruvate carboxylase from a thermophilic cyanobacterium, Synechococcus vulcanus with unusual allosteric properties. Plant Cell Physiol 2002;43:159–169.
47 21. Ubeidat M and Rutherford CL. Expression and one-step purification of a developmentally regulated protein from Dictyostelium discoideum. Protein Expr Purif 2002;25: 472–480. 22. Simonovic I and Patston PA. The native metastable fold of C1-inhibitor is stabilized by disulfide bonds. Biochim Biophys Acta 2000;1481:97–102. 23. Vinetz JM, Valenzuela JG, Specht CA, Aravind L, Langer RC, Ribeiro JM and Kaslow DC. Chitinases of the avian malaria parasite Plasmodium gallinaceum, a class of enzymes necessary for parasite invasion of the mosquito midgut. J Biol Chem 2000;275:10331–10341. 24. Lin MT, Kuo TJ and Lin CT. Molecular cloning of a cDNA encoding copper/zinc superoxide dismutase from papaya fruit and overexpression in Escherichia coli. J Agric Food Chem 1998;46:344–348. 25. Bancroft FC, Wu GJ and Zubay G. Cell-free synthesis of rat growth hormone. Proc Natl Acad Sci USA 1973;70:3646–3649. 26. Spirin AS, Baranov VI, Ryabova LA, Ovodov SY and Alakhov YB. A continuous cell-free translation system capable of producing polypeptides in high yield. Science 1988;242: 1162–1164. 27. Kim DM and Choi CY. A semicontinuous prokaryotic coupled transcription/translation system using a dialysis membrane. Biotechnol Prog 1996;12:645–649. 28. Kim RG and Choi CY. Expression-independent consumption of substrates in cell-free expression system from Escherichia coli. J Biotechnol 2001;84:27–32. 29. Morita EH, Sawasaki T, Tanaka R, Endo Y and Kohno T. A wheat germ cell-free system is a novel way to screen protein folding and function. Protein Sci 2003;12:1216–1221. 30. Madin K, Sawasaki T, Ogasawara T and Endo Y. A highly efficient and robust cell-free protein synthesis system prepared from wheat embryos: plants apparently contain a suicide system directed at ribosomes. Proc Natl Acad Sci USA 2000;97:559–564. 31. Kawarasaki Y, Nakano H and Yamane T. Prolonged cell-free protein synthesis in a batch system using wheat germ extract. Biosci Biotechnol Biochem 1994;58:1911–1913. 32. Stevens RC. Design of high-throughput methods of protein production for structural biology. Structure Fold Des 2000;8:R177–185. 33. Misawa S and Kumagai I. Refolding of therapeutic proteins produced in Escherichia coli as inclusion bodies. Biopolymers 1999;51:297–307. 34. Maachupalli-Reddy J, Kelley BD and De Bernardez Clark E. Effect of inclusion body contaminants on the oxidative renaturation of hen egg white lysozyme. Biotechnol Prog 1997; 13:144–150. 35. Hagel P, Gerding JJ, Fieggen W and Bloemendal H. Cyanate formation in solutions of urea. I. Calculation of cyanate concentrations at different temperature and pH. Biochim Biophys Acta 1971;243:366–373. 36. Gerding JJ, Koppers A, Hagel P and Bloemendal H. Cyanate formation in solutions of urea. II. Effect of urea on the eye lens protein–crystallin. Biochim Biophys Acta 1971;243: 375–379. 37. Jekabsons MB, Echtay KS and Brand MD. Nucleotide binding to human uncoupling protein-2 refolded from bacterial inclusion bodies. Biochem J 2002;366:565–571. 38. Hanagan A, Meyer JD, Johnson L, Manning MC and Catalano CE. The phage lambda terminase enzyme: 2. Refolding of the gpNu1 subunit from the detergent-denatured and guanidinium hydrochloride-denatured state yields different oligomerization states and altered protein stabilities. Int J Biol Macromol 1998;23:37–48. 39. Puri NK, Cardamone M, Crivelli E and Traeger JC. Characterization of a truncated form of recombinant porcine growth hormone generated in vitro during solubilization of inclusion bodies. Protein Expr Purif 1993;4:164–175. 40. Patra AK, Gahlay GK, Reddy BV, Gupta SK and Panda AK. Refolding, structural transition and spermatozoa-binding of recombinant bonnet monkey (Macaca radiata) zona pellucida glycoprotein-C expressed in Escherichia coli. Eur Biochem 2000;267:7075–7081.
48 41. Khan RH, Rao KB, Eshwari AN, Totey SM and Panda AK. Solubilization of recombinant ovine growth hormone with retention of native-like secondary structure and its refolding from the inclusion bodies of Escherichia coli. Biotechnol Prog 1998;14:722–728. 42. Schoemaker JM, Brasnett AH and Marston FA. Examination of calf prochymosin accumulation in Escherichia coli: disulphide linkages are a structural component of prochymosin-containing inclusion bodies. Embo J 1985;4:775–780. 43. Ferre H, Ruffet E, Blicher T, Sylvester-Hvid C, Nielsen LL, Hobley TJ, Thomas OR and Buus S. Purification of correctly oxidized MHC class I heavy-chain molecules under denaturing conditions: A novel strategy exploiting disulfide assisted protein folding. Protein Sci 2003;12:551–559. 44. Tang L, Morales T, Boroughs KL, Cailles Lo-Keiser K, Sellins K, Stedman K, McCall C and McDermott MJ. Expression and characterization of recombinant canine IL-13 receptor alpha2 protein and its biological activity in vitro. Mol Immunol 2003;39:719–727. 45. Anderson DE, Peters RJ, Wilk B and Agard DA. Alpha-lytic protease precursor: characterization of a structured folding intermediate. Biochemistry 1999;38:4728–4735. 46. Ouellette T, Destrau S, Zhu J, Roach JM, Coffman JD, Hecht T, Lynch JE and Giardina SL. Production and purification of refolded recombinant human IL-7 from inclusion bodies. Protein Expr Purif 2003;30:156–166. 47. Gu Z, Weidenhaupt M, Ivanova N, Pavlov M, Xu B, Su ZG and Janson JC. Chromatographic methods for the isolation of, and refolding of proteins from, Escherichia coli inclusion bodies. Protein Expr Purif 2002;25:174–179. 48. Ptitsyn OB, Pain RH, Semisotnov GV, Zerovnik E and Razgulyaev OI. Evidence for a molten globule state as a general intermediate in protein folding. FEBS Lett 1990;262:20–24. 49. Menzella HG, Gramajo HC and Ceccarelli EA. High recovery of prochymosin from inclusion bodies using controlled air oxidation. Protein Expr Purif 2002;25:248–255. 50. De Bernardez Clark E, Hevehan D, Szela S and Maachupalli-Reddy J. Oxidative renaturation of hen egg-white lysozyme. Folding vs aggregation. Biotechnol Prog 1998;14:47–54. 51. Xie Y and Wetlaufer DB. Control of aggregation in protein refolding: the temperature-leap tactic. Protein Sci 1996;5:517–523. 52. Creighton TE. Disulfide bonds between cysteine residues. In: Protein Structure: A Practical Approach. Creighton TE (ed), Oxford, IRL Press), pp. 155–168, 1990. 53. Jaenicke R and Rudolph R. Folding proteins. In: Protein Structure: A Practical Approach. Creighton TE (ed), Oxford, IRL Press), pp. 191–224, 1990. 54. Goldberg ME, Rudolph R and Jaenicke R. A kinetic study of the competition between renaturation and aggregation during the refolding of denatured-reduced egg white lysozyme. Biochemistry 1991;30:2790–2797. 55. Nieuwenhuizen WF, van Leeuwen S, Jack RW, Egmond MR and Gotz F. Molecular cloning and characterization of the alkaline ceramidase from Pseudomonas aeruginosa PA01. Protein Expr Purif 2003;30:94–104. 56. Juan CC, Kan LS, Huang CC, Chen SS, Ho LT and Au LC. Production and characterization of bioactive recombinant resistin in Escherichia coli. J Biotechnol 2003;103:113–117. 57. Nishiyama Y, Hardre-Lienard H, Miras S, Miege C, Block MA, Revah F, Joyard J and Marechal E. Refolding from denatured inclusion bodies, purification to homogeneity and simplified assay of MGDG synthases from land plants. Protein Expr Purif 2003;31:79–87. 58. Huxtable S, Zhou H, Wong S and Li N. Renaturation of 1-aminocyclopropane-1-carboxylate synthase expressed in Escherichia coli in the form of inclusion bodies into a dimeric and catalytically active enzyme. Protein Expr Purif 1998;12:305–314. 59. Im H and Yu MH. Role of Lys335 in the metastability and function of inhibitory serpins. Protein Sci 2000;9:934–941. 60. Yokoyama K, Ejima D, Kita Y, Philo JS and Arakawa T. Structure of folding intermediates at pH 4.0 and native state of microbial transglutaminase. Biosci Biotechnol Biochem 2003;67: 291–294.
49 61. Bottomley SP and Stone SR. Protein engineering of chimeric Serpins: an investigation into effects of the serpin scaffold and reactive centre loop length. Protein Eng 1998;11: 1243–1247. 62. D’’Alatri L, Di Massimo AM, Anastasi AM, Pacilli A, Novelli S, Saccinto MP, De Santis R, Mele A and Parente D. Production and characterisation of a recombinant single-chain anti ErbB2-clavin immunotoxin. Anticancer Res 1998;18:3369–3373. 63. Rogl H, Kosemund K, Kuhlbrandt W and Collinson I. Refolding of Escherichia coli produced membrane protein inclusion bodies immobilised by nickel chelating chromatography. FEBS Lett 1998;432:21–26. 64. Schauer S. Large scale production of biologically active Eschericia coli glutamyl-tRNA reductase from inclusion bodies. Protein Expr Purif, 2003;31:276–285. 65. Garcia-Saez I and Plasterk RH. Purification of the Caenorhabditis elegans transposase Tc1A refolded during gel filtration chromatography. Protein Expr Purif 2000;19: 355–361. 66. Gu Z, Su Z and Janson JC. Urea gradient size-exclusion chromatography enhanced the yield of lysozyme refolding. J Chromatogr A 2001;918:311–318. 67. Harrowing SR and Chaudhuri JB. Effect of column dimensions and flow rates on sizeexclusion refolding of beta-lactamase. J Biochem Biophys Methods 2003;56:177–188. 68. Altamirano MM, Golbik R, Zahn R, Buckle AM and Fersht AR. Refolding chromatography with immobilized mini-chaperones. Proc Natl Acad Sci USA 1997;94: 3576–3578. 69. Preston NS, Baker DJ, Bottomley SP and Gore MG. The production and characterisation of an immobilised chaperonin system. Biochim Biophys Acta 1999;1426:99–109. 70. Tsumoto K, Umetsu M, Yamada H, Ito T, Misawa S and Kumagai I. Immobilized oxidoreductase as an additive for refolding inclusion bodies: application to antibody fragments. Protein Eng 2003;16:535–541. 71. Vinogradov AA, Kudryashova EV, Levashov AV and van Dongen WM. Solubilization and refolding of inclusion body proteins in reverse micelles. Anal Biochem 2003;320:234–238. 72. Tanaka N, Nakao S, Wadai H, Ikeda S, Chatellier J and Kunugi S. The substrate binding domain of DnaK facilitates slow protein refolding. Proc Natl Acad Sci USA 2002;99: 15398–15403. 73. Buchberger A, Schroder H, Hesterkamp T, Schonfeld HJ and Bukau B. Substrate shuttling between the DnaK and GroEL systems indicates a chaperone network promoting protein folding. J Mol Biol 1996;261:328–333. 74. Diamant S and Goloubinoff P. Temperature-controlled activity of DnaK-DnaJ-GrpE chaperones: protein-folding arrest and recovery during and after heat shock depends on the substrate protein and the GrpE concentration. Biochemistry 1998;37:9688–9694. 75. Zavilgelsky GB, Kotova VY, Mazhul MM and Manukhov IV. Role of Hsp70 (DnaK-DnaJ-GrpE) and Hsp100 (ClpA and ClpB) chaperones in refolding and increased thermal stability of bacterial luciferases in Escherichia coli cells. Biochemistry (Mosc) 2002;67: 986–992. 76. Lilie H, Lang K, Rudolph R and Buchner J. Prolyl isomerases catalyze antibody folding in vitro. Protein Sci 1993;2:1490–1496. 77. Maskos K, Huber-Wunderlich M and Glockshuber R. DsbA and DsbC-catalyzed oxidative folding of proteins with complex disulfide bridge patterns in vitro and in vivo. J Mol Biol 2003; 325:495–513. 78. Zapun A and Creighton TE. Effects of DsbA on the disulfide folding of bovine pancreatic trypsin inhibitor and alpha-lactalbumin. Biochemistry 1994;33:5202–5211. 79. Tang B, Zhang S and Yang K. Assisted refolding of recombinant prochymosin with the aid of protein disulphide isomerase. J Biochem 1994;301(Pt 1):17–20. 80. Hofmann A, Tai M, Wong W and Glabe CG. A sparse matrix screen to establish initial conditions for protein renaturation. Anal Biochem 1995;230:8–15.
50 81. Chen GQ and Gouaux E. Overexpression of a glutamate receptor (GluR2) ligand binding domain in Escherichia coli: application of a novel protein folding screen. Proc Natl Acad Sci USA 1997;94:13431–13436. 82. Armstrong N, de Lencastre A and Gouaux E. A new protein folding screen: application to the ligand binding domains of a glutamate and kainate receptor and to lysozyme and carbonic anhydrase. Protein Sci 1999;8:1475–1483. 83. Tobbell DA, Middleton BJ, Raines S, Needham MR, Taylor IW, Beveridge JY and Abbott WM. Identification of in vitro folding conditions for procathepsin S and cathepsin S using fractional factorial screens. Protein Expr Purif 2002;24:242–254. 84. Meyer DE, Trabbic-Carlson K and Chilkoti A. Protein purification by fusion with an environmentally responsive elastin-like polypeptide: effect of polypeptide length on the purification of thioredoxin. Biotechnol Prog 2001;17:720–728. 85. Bloomfield VA. Static and dynamic light scattering from aggregating particles. Biopolymers 2000;54:168–172. 86. Richter W, Hermsdorf T, Lilie H, Egerland U, Rudolph R, Kronbach T and Dettmer D. Refolding, purification, and characterization of human recombinant PDE4A constructs expressed in Escherichia coli. Protein Expr Purif 2000;19:375–383. 87. Gorman PM, Yip CM, Fraser PE and Chakrabartty A. Alternate aggregation pathways of the Alzheimer beta-amyloid peptide: Abeta association kinetics at endosomal Ph. J Mol Biol 2003; 325:743–757. 88. Ellman GL. Tissue sulfhydryl groups. Arch Biochem Biophys 1959;82:70–77. 89. Bures EJ, Hui JO, Young Y, Chow DT, Katta V, Rohde MF, Zeni L, Rosenfeld RD, Stark KL and Haniu M. Determination of disulfide structure in agouti-related protein (AGRP) by stepwise reduction and alkylation. Biochemistry 1998;37:12172–12177. 90. Dykes CW, Bookless AB, Coomber BA, Noble SA, Humber DC and Hobden AN. Expression of atrial natriuretic factor as a cleavable fusion protein with chloramphenicol acetyltransferase in Escherichia coli. Eur J Biochem 1988;174:411–416. 91. Butt TR, Jonnalagadda S, Monia BP, Sternberg EJ, Marsh JA, Stadel JM, Ecker DJ and Crooke ST. Ubiquitin fusion augments the yield of cloned gene products in Escherichia coli. Proc Natl Acad Sci USA 1989;86:2540–2544. 92. LaVallie ER, DiBlasio EA, Kovacic S, Grant KL, Schendel PF and McCoy JM. A thioredoxin gene fusion expression system that circumvents inclusion body formation in the E. coli cytoplasm. Biotechnology (NY) 1993;11:187–193. 93. Nilsson B, Abrahmsen L and Uhlen M. Immobilization and purification of enzymes with staphylococcal protein A gene fusion vectors. Embo J 1985;4:1075–1080. 94. Matsumoto M, Misawa S, Tsumoto K, Kumagai I, Hayashi H and Kobayashi Y. On-column refolding and characterization of soluble human interleukin-15 receptor alpha-chain produced in Escherichia coli. Protein Expr Purif 2003;31:64–71. p 95. Letourner O. Molecular cloning, overexpression in Escherichia coli, and purification of 6 his-tagged C-terminal domain of Clostridium difficile toxins A and B. Protein Expr Purif, in press. 96. Lee HJ and Im H. Purification of recombinant plasminogen activator inhibitor-1 in the active conformation by refolding from inclusion bodies. Protein Expr Purif 2003;31:99–107.
51
Towards a systems biology understanding of human health: Interplay between genotype, environment and nutrition Frank Desiere* Nestle´ Research Center, P.O. Box 44, 1000 Lausanne 26, Switzerland; Institute for Systems Biology, Seattle, Washington, USA Abstract. Sequencing of the human genome has opened the door to the most exciting new era for the holistic system description of human health. It is now possible to study the underlying mechanisms of human health in relation to diet and other environmental factors such as drugs and toxic pollutants. Technological advances make it feasible to envisage that in the future personalized drug treatment and dietary advice and possibly tailored food products can be used for promoting optimal health on an individual basis, in relation to genotype and lifestyle. Life-Science research has in the past very much focused on diseases and how to reestablish human health after illness. Today, the role of food and nutrition in human health and especially prevention of illness is gaining recognition. Diseases of modern civilization, such as diabetes, heart disease and cancer have been shown to be effected by dietary patterns. The risk of disease is often associated with genetic polymorphisms, but the effect is dependent on dietary intake and nutritional status. To understand the link between diet and health, nutritional-research must cover a broad range of areas, from the molecular level to whole body studies. Therefore it provides an excellent example of integrative biology requiring a systems biology approach. The current state and implications of systems biology in the understanding of human health are reviewed. It becomes clear that a complete mechanistic description of the human organism is not yet possible. However, recent advances in systems biology provide a trajectory for future research in order to improve health of individuals and populations. Disease prevention through personalized nutrition will become more important as the obvious avenue of research in life sciences and more focus will need to be put upon those natural ways of disease prevention. In particular, the new discipline of nutrigenomics, which investigates how nutrients interact with humans, taking predetermined genetic factors into account, will mediate new insights into human health that will finally have significant positive impact on our quality of life. Keywords: systems biology, genomics, transcriptomics, proteomics, metabolomics, nutrigenomics, pharmacogenetics, pharmacogenomics, health, nutrition, diet, disease, prevention, diagnostics, bioinformatics, molecular databases, metabolism, networks, cells, polymorphisms, SNPs, epigenetics.
Introduction Systems biology is gaining importance in today’s life-science research. Interestingly, the first attempts to systems biology, go back to the 1960s. At the time such attempts were called modeling of cellular processes by study means of ‘‘systems theory and biology.’’ Many mathematicians and engineers tried to develop approaches that allow analyzing biological systems in a physical way. It was realized at the time, that when they tried to interact with an organism as a physical system, they found themselves interacting with it in many *Corresponding author: E-mail:
[email protected] BIOTECHNOLOGY ANNUAL REVIEW VOLUME 10 ISSN: 1387-2656 DOI: 10.1016/S1387-2656(04)10003-3
ß 2004 ELSEVIER B.V. ALL RIGHTS RESERVED
52 more ways than they had instrumentations for. It then became clear that for a complex system as cells or organisms there were many more capabilities, and several more modes of interaction than just a limited set of canonical rules [1]. Up until that point, many scientists and biologists alike have focused on reducing life to its constituent parts, first focusing on the cell, then working their way down to the molecular level. Today, two apparently opposing opinions are in discussion. The first claims that a cellular system can readily be described in all its parts and even be simulated, maybe using the tools of systems biology. The other opinion cast serious doubts that this can be achieved due to fundamental reasons and limitations. This opinion is certainly well documented by the work of Robert Rosen (1934–1998), a theoretical biologist who strived to answer the question the Nobel physicist Erwin Schro¨dinger posed in 1943: ‘‘What is Life?’’ To this day, what it is that makes an organism alive has remained unanswered by conventional biology, chemistry and physics. Schro¨dinger’s works on complexity and biological systems claim that these cannot be decomposed or predicted because of their anticipatory nature and that a biological system is not a just a complex machine [2]. But let us now focus on the achievements that have catalyzed the massive advances in understanding of biological systems through the field of biotechnology and later genomics, leading finally to the more holistic (and mysterious) term of ‘‘systems biology.’’ With the development of highthroughput technologies for molecular biology in the 1980s and 1990s, that amongst other achievements have resulted in the completion of the human genome [3,4], quantitative data on the transcriptome [5], proteome [6] and metabolome level [7], an increasing interest in formal mathematical models of cellular activities as gene expression and regulation has been triggered. It was realized that the vast amount of data available would require new concepts for the understanding and new tools for the description of life as a whole. Systems biology and systems theory, which study the organization and behavior of living systems, seemed indeed a natural conceptual framework for such a task. Systems biology attempts to reconstruct living systems as a series of overlapping models. It exploits all the theoretical and experimental advances of the various genome projects, allying them to computational, mathematical and engineering disciplines. This is done in an attempt to create predictive models of cells, organs, biochemical processes, and complete organisms. Consequently, systems biology today has the potential to advance our knowledge and understanding of complex biological systems, from simple cells to complete organisms and potentially to whole ecosystems. The understanding of biological systems is not an altruistic matter for the benefit of advancing philosophy and theoretical sciences. No, there are real problems to be solved. The world’s demography has pushed medical treatment to higher levels over the last decades, mainly due to aging populations and increased life-expectancy [8]. With individuals realizing that they will enjoy longer lives, the issue of disease-prevention has become an important concern.
53 The quest for new treatments and prevention of illness has let pharmaceutical companies become powerful, big corporations which drive many areas of modern life science and a big proportion of the biotechnology industry to this date. The discovery of new pharmaceutical treatments, especially those which will bring large amounts of cash back to the industry, the so called ‘‘block-buster drugs,’’ seem to be the ultimate goal for research and development in academic and public institutes, biotechnology companies and the private medical research centers alike. In parallel, researchers have promoted disease prevention also through adequate nutrition and it was realized that scientific breakthroughs in both areas would require a massive investment into modern nutrition research through ‘‘systems biology.’’ To accelerate the mission, research institutes and scientific groups dealing with systems biology have been created in recent years. Founded in the year 2000, the ‘‘The Institute for Systems Biology’’ (www.systemsbiology.org) is for many people the pioneer in the new field, and has managed to influence the pace and direction of modern biology. That trend has now gained broad acceptance as a new scientific field, prompting the National Institutes of Health (NIH) to identify in 2003 systems biology and multidisciplinary research as key components in a new set of agency initiatives for the NIH Roadmap for Medical Research of the next decade. New projects under the theme of New Pathways to Discovery would include Bioinformatics and Computational Biology, Structural Biology, Building Blocks and Pathways, Molecular Libraries and Molecular Imaging, and Nanomedicine. Starting from the year 2004, the NIH will fund these topics which will at the same time require an improved computational infrastructure for biomedical research, libraries of chemical molecules, new molecular and cellular imaging tools, and nanoscale technology devices for viewing and interacting with basic life processes. This policy describes clearly the challenges ahead of us to investigate biological systems, particularly in the context of human health, treatment of disease and prevention of illness. Technologies of systems biology Systems biology focuses on complex biological systems that are composed of molecular components. Understanding systems biology requires the integration of experimental and computational research data [9]. Systems biology is the attempt to systematically study all the concurrent physiological processes in a cell or tissue by global measurement of differentially perturbed states. The ultimate goal of systems biology is the integration of data from these observations into models that might, eventually, represent and make possible the simulation of the physiology of the cell [10,11]. Although biological systems are made-up of their components, the essence of a system lies in dynamics and it cannot be described merely by enumerating components of the system. At the same time, it is inappropriate to believe that only system structures, such as network topologies, are important without
54 Table 1. Web resources and databases for systems biology.
Institute for Systems Biology (www.systemsbiology.org) MIT Computational and Systems Biology Initiative (CSBI) (csbi.mit.edu/) Bauer Center for Genomics Research (CGR) at Harvard University (www.cgr.harvard.edu/) Bio-X at Stanford University (biox.stanford.edu/) Cell Systems Initiative at the University of Washington (csi.washington.edu/) Genomes to Life program at the US Department of Energy (DOE) (doegenomestolife.org/) Biomolecular Systems website at the Pacific Northwest National Laboratory (PNNL) (biomolecular.org) Institute for Computational Biomedicine at the Weill Medical College of Cornell University (icb.med.cornell.edu/)
paying sufficient attention to diversities and functionalities of the components (Table 1). Both structure of the system and its components play indispensable roles forming a holistic view of the state of the system. The goals of systems biology are: (1) (2)
(3) (4)
Understanding of the components of a biological system, such as genes, proteins and metabolites, as well as their physical structures, Understanding of dynamics of the system, both quantitative and qualitative analysis as well as construction of theories/models with powerful prediction capability, Understanding of control methods of the system, and Understanding of design methods of the system.
The following sections will give a more detailed overview of the sub-disciplines of systems biology, which characterize the cellular components. Finally these components will have to be put into context, which will be the focus towards the end of this review. Genomics The availability of completely sequenced genomes catalyzed the emergence of systems biology and has truly revolutionized biology. For the first time since the advent of molecular biology, biological questions are now addressed by studying the complete set of a system in contrast to the previous investigation of function(s) of individual genes and gene products one or a few at a time. Before, high-throughput analytical instruments like the DNA sequencer or mass spectrometers for protein determination had been invented, this reductionist approach proved to be extremely fruitful, leading to the discovery of an impressive number of biological principles. However, it was quickly realized that in nature, cellular components function together with other components. As Henri Poincare´ already pointed out in 1952 [12] ‘‘the aim of science is not things in themselves, but the relations between them; outside these relations there is no reality knowable.’’ Indeed, biological processes should be considered as a
55
Number of completed genomes
complex network of interconnected components. In other words, for any biological process, one might consider a ‘‘modular approach’’ in which the behavior and function of the corresponding network are studied as a whole. In addition to studying some of its components individually, the first step to reach that goal was the determination of complete genomes of organisms. The significance of the finished human genome sequence [3,4] and other genomes of model organisms for the field of systems biology cannot be overstated. Without these genomes, holistic studies would simply not be possible. Still, our knowledge is steadily increasing, which is underlined by the latest detailed analysis of human chromosome 6 [13]. A great abundance of biological information was revealed that was previously unrecognized within the draft of the human genome. Comparative genomics using the genomes of the mouse, rat, puffer- and zebra-fish allowed refined predictions of which stretches of DNA are actually genes, and a more sophisticated interpretation of the underlying genomic data. The power of comparative genomics is quickly growing as the genome sequences of other nematodes are sequenced [14], as well as chicken, chimpanzee, frog, and cow that are already in the production queue, become available. Currently there are about 203 complete genomes of living organisms in the public domain (www.ebi.ac.uk/genomes/, Fig. 1), with more than 800 on their way of being finished. These numbers underline the growing importance of comparative genomics. However, it must be stated that gene-prediction remains to be a significant challenge and it can be anticipated that our current data about location and number of genes will constantly have to be updated [15]. The genome of an organism represents an ideal coordinate system for systems biology, a precisely definable digital core of information for an organism [16]. Genes are the ‘‘genetic parts list’’ to which all other biological information can be linked. Transcripts are directly related to genes. Proteins are related to transcripts and then to genes. All the information is hierarchical in
160 140 120 100 80 60 40 20 0 1995 1996 1997 1998 1999 2000 2001 2002 2003
Fig. 1. Number of completed genomes (http://www.ebi.ac.uk/genomes).
56
Fig. 2. Regulatory gene network for endomesoderm specification: the view from the genome. The architecture of the network is based on perturbation and expression data, on data from cisregulatory analyses for several genes, and on other experiments (reproduced with permission from Hamid Bolouri and Eric Davidson, http://sugp.caltech.edu/endomes/) [17].
nature: DNA, mRNA, protein, protein interactions, informational pathways, informational networks, cells, tissues or networks of cells, an organism, populations and whole ecologies. It is therefore tempting to construct a geneindex in which every gene of organisms are listed and numbered and to use it as a central core for linking any kind of biological information to it. This concept has partially been applied to publicly accessible genome resources e.g., Ensembl (www.ensembl.org) and RefSeq (www.ncbi.nlm.nih.gov/RefSeq). Genomic sequences also provide access to regulatory sequences in genomes, which are a vital component to solving the regulatory code [17]. Also, genomic sequences open access to polymorphism studies; some of these variations are responsible for differences in physiology and disease predisposition. These components combined make-up the elements in the ‘‘periodic table of life.’’ With these components in hand, the immediate challenge is to place them in the context of their informational pathways and networks.
57 The logical extension to studying the genome is the determination of interindividual differences within the genome of people. Only a small number of common polymorphisms explain the bulk of heterozygosity [18]. Human genetic diversity appears on the level of individual polymorphisms, known as single nucleotide polymorphisms (SNPs), as well as in the specific combinations of alleles (haplotypes) as observed at closely linked sites. The goal of the International HapMap Project for example is to develop a haplotype map of the human genome, to describe the common patterns of human DNA sequence variation. The HapMap is expected to be a key resource for researchers in finding genes affecting health, disease, responses to diet and other environmental factors. SNPs, single-nucleotide polymorphisms, are small genetic variations between people that can significantly alter the function of proteins. Most importantly, the altered function may have significant effects on how the individual reacts to treatment of drugs, allergies to environmental substances and digestion of foods [19]. The latest release of dbSNP (118) at the NCBI contains an impressive amount of 5,798,183 SNPs for human of which 2,359,534 are validated. These polymorphisms now have to be investigated for their significance in altering biological function of proteins and pathways. Knowledge about SNPs is most important for treatment using drugs. Altered protein function might not carry a drug to its target cells or tissues cripple the enzymes that activate a drug or aid its removal from the body, or alter the structure of the receptor to which a drug is supposed to bind. Variation in immune-system genes can also influence how particular drugs are tolerated. Together, these subtle genetic variations mean that the dose at which a drug will work may vary hugely from person to person. The so widely utilized ‘‘one-size-fits-all’’ prescription leads to life-threatening adverse reactions and to drugs completely failing to do their job. Well-documented examples of active SNPs are available from the P450 protein family, enzymes in the liver that oxidize foreign chemicals. Three of these P450 genes that are particularly important for drug metabolism of commonly prescribed drugs, have been shown to be highly polymorphic and some have already been linked to failure in certain patients [20]. Other examples show that the efficiency of the painkiller codeine depends on a particular polymorphism [21] and that the anticoagulation drug warfarin can cause serious adverse drug reactions depending on the genotype of the patient [22]. Another example shows that the base excision repair enzyme MED1 is associated with nonpolyposis colorectal tumors, a very common form of hereditary cancer. The gene’s protein product, MED1, is an enzyme that normally helps cells repair potentially cancer-causing damage to genes. However, a defective MED1 enzyme did not only prevent repairs in normal cells and permitted a cancer to start, but in particular, the enzyme also interfered with the effectiveness of some types of chemotherapy [23]. Genomic polymorphisms will only be able to be investigated with many complete human genomes available, an achievement that can be envisaged by the
58 end of the first decade of the 21st century. It is anticipated that within about 10 years, advances in nanotechnology and other methods will allow the fast and cheap sequencing of individuals’ genomes, which in turn will lead to advances in predictive medicine. As scientists are able to look at 30,000 or more genes for each patient, doctors could use such genome sequences to predict what health problems the individual patient is likely to face. Genome shotgun sequencing and microarrays have given us the tools to identify people with SNPs [24]. Individuals can now be profiled with increasing efficiency, and used to highlight polymorphic genes that influence our response to specific drugs or foods. These developments have resulted in a completely new discipline called ‘‘Pharmacogenetics’’ – the study of the influence of genetic variation on drug responses [25]. Similarly, the science of nutrigenomics seeks to provide a molecular understanding for how common dietary chemicals (i.e., nutrition) affect health by altering the expression and/or structure of an individual’s genetic makeup. Thus, the new field of nutrigenomics opens the way for ‘‘personalized nutrition.’’ In other words, by understanding our nutritional needs, our nutritional status, and our genotype, nutrigenomics should enable individuals to manage better their health and well-being by precisely matching their diets with their unique genetic makeup. The success of these methods will largely depend on the large-scale discovery of SNPs, their validation and the discovery of diet-related genes. To achieve this task more research into nutritional sciences using systems biology will have to be initiated thus identifying nutritionally relevant genes in order to study their response to nutrients systematically. The understanding of the human genome is not completed with the genome sequence established and the polymorphisms determined. New discoveries further complicate the understanding of genomes. Inheritable changes in gene function can occur without a change in the DNA sequence. Epigenetic mechanisms such as DNA methylation, histone acetylation, and RNA interference, and their effects in gene activation or inactivation might be involved in imprinting and parental imprinting in which a gene’s activity depends on whether it is inherited from the mother or the father [26]. There is evidence to suggest that factors such as lifestyle and diet leave a trail of epigenetic footprints across our genome, which is then inherited [27]. In a striking example, Duke University researchers have demonstrated recently in mice how extra vitamin doses during pregnancy in the mother’s diet changes the color of pups [28]. This study is the first one to find a clear mechanism of the effect of maternal nutrition on disease and phenotype. The nutrients used in the study, B12, folic acid, choline and betaine, had silenced the gene that rendered mice fat and yellow, but had not altered its sequence. The gene was in fact methylated, and thus switched off, linking prenatal diet to diseases like diabetes, obesity and cancer. Thus, knowledge about the genomic make-up of individuals will be crucial in health research (Table 2).
59 Table 2. Web resources and databases for genomics.
The National Human Genome Research Institute (www.nhgri.nih.gov) Nature Genome Getaway (www.nature.com/genomics/human/) Ensembl Human Genome browser (www.ensembl.org) European Bioinformatics Institute EBI (www.ebi.ac.uk/) National Center for Biotechnology Information NCBI (www.ncbi.nlm.nih.gov/genome/guide/ human/) International HapMap Project (www.hapmap.org) The Human Epigenome Project HEP (www.epigenome.org) Database of single nucleotide polymorphisms dbSNP (www.ncbi.nlm.nih.gov/SNP)
Transcriptomics The genome describes the ultimate potential of an organism, and the transcriptome, all complementary DNA sequences, describes the utilization/ expression of that potential. Transcripts can readily be identified by expressed sequence tags (ESTs). EST sequencing efforts still represent an economic and fast way to characterize expressed genes. EST sequencing still remains an essential resource for genome exploitation and annotation. This is particularly important with the increasing availability of draft genome sequences from different organisms and the mounting emphasis on gene function and regulation [29]. Simultaneous analysis of gene-expression can be performed using the technology that allows synthesis or immobilization of known complementary DNA sequences on microscopic arrays and later hybridizing RNA obtained from living cells onto the array. Microarrays exploit the preferential binding of complementary single-stranded nucleic-acid sequences and the underlying principle is the same for all microarrays. An unknown sample is hybridized to the array of immobilized DNA molecules whose sequence is known. Each array features thousands of different DNA probe sequences arranged in a defined matrix and thus can identify thousands of genes simultaneously, which means that genetic analysis can be done on a huge scale. Transcriptome profiling, using microarrays [30,31] or serial analysis of gene expression (SAGE) [32], can measure the relative abundance of transcripts simultaneously for thousands of genes under various experimental conditions. This technology has revolutionized the way in which researchers analyze gene expression in cells and tissues. It allows researchers to determine which genes are being expressed in a given cell type at a particular time and under particular conditions. They can be used to compare the status of gene expression in two different cell types or tissue samples, for example, healthy versus diseased tissue, and to examine changes in gene expression-profile at different stages in the cell cycle or during embryonic development. Other uses of microarrays include comparative genomic hybridization studies [33], genotyping individuals for genetic differences that might be associated with disease [34], assignment of probable functions to newly discovered genes by comparison with the expression patterns of known genes,
60 to identify key players in signaling pathways and to uncover new categories of genes [35] (Fig. 3). Other areas of application engulf today the identification of new targets for therapeutic drugs, in disease diagnosis, and in toxicogenomics [36], the study of the genetic basis of an individual’s response to environmental factors such as drugs and pollutants. Transcription profiling is today applied in all major areas of biology. One of the most remarkable studies to date and a great example for a systems biology approach is the description of a geneco-expression network for global discovery of conserved genetic modules [5].
Experimental Design
Experiment
Data analysis
Data storage
RT-PCR Labeling Pooling of samples Hybridization to array Scanning
Image analysis Statistics Normalization Clustering Annotation
MAGE-ML, MIAME Relational database ArrayExpress GEO Stanford microar. DB
Protein extraction Sample fractionation Separation (2D-GE, LC) Digestion ESI/MALDI/FT-MS
Analysis of spectra Statistical evaluation Identification: Database search, Quantification Annotation
mzXML, PEDRO Relational database No repository yet! GenBank, EMBL, BIND, DIP, etc.
Analysis of spectra Database search Statistical evaluation Annotation, Clustering Network Modeling
SBML Relational database No repository yet! KEGG, E-cell, EMP etc.
Transcriptome Determination of genome-wide transcript levels via DNA array: Treated vs. non-treated Normal vs. abnormal tissue
Proteome Determination of all proteins in a cell or body fluid via (quantitative) mass spectrometry: Treated vs. non-treated Normal vs. abnormal tissue
500
Intensity, counts
400
Metabolome Determination of all metabolites in a cell or body fluid: Treated vs. non-treated Normal vs. abnormal tissue
300 200 100 0 600
1000
1400 1800 m/z , au
2200
2600
Metabolite extraction Sample fractionation Separation (LC, GC) Identification: MS, GC, NMR
y
x z
Fig. 3. Comparison of data analysis strategies for transcriptome, proteome and metabolome studies. Abbreviations: RT-PCR, reverse transcriptase polymerase chain reaction; MAGE-ML, MicroArray Gene Expression Markup Language; MIAME, Minimum Information About a Microarray Experiment; GEO, Gene Expression Omnibus; 2D-GE, 2 Dimensional Gel Electrophoresis; ESI, Electro Spray Ionization; MALDI, Matrix Assisted Laser Desorption/ Ionization; FT, Fourier Transform; MS, Mass Spectrometry; mzXML, mass spectrometry eXtensible Markup Language [37]; PEDRO, software, to support the capture, storage and dissemination of proteomics experimental data [37,38]; EMBL, European Molecular Biology Laboratory; BIND, Biomolecular Interaction Network Database; DIP, Database of Interacting Proteins; LC, Liquid Chromatography; GC, Gas Chromatography; NMR, Nuclear Magnetic Resonance spectrometry; SBML, Systems Biology Markup Language; KEGG, Kyoto Encyclopedia of Genes and Genomes; EMP, database of Enzymes and Molecular Pathways.
61 In a truly massive approach co-expressed pairs of genes were identified over 3182 DNA microarrays from humans, flies, worms, and yeast. An estimated number of 22,163 conserved co-expression relationships were identified using statistical clustering algorithms providing new evidence for the involvement of genes in core biological functions. The relative ease of producing such a large number of data is obscured by the difficulty of dealing with the results due to a lack of simple and accepted approaches to analyzing large-scale gene expression data. Visualizing and presenting such large gene expression data is not trivial [30]. Despite these difficulties, the field of gene expression analysis is helping devise strategies that also allow distinguishing between the expressions of alternatively spliced transcripts. It has been estimated that 30%–60% of all human genes encode for more than one transcript. The impact of these alternative gene-products on function and regulation has become a major focus for research and has led to the establishment of various databases harboring information on alternatively spliced transcripts [39,40]. Further investigation is required to determine the cause and effect of alternative splicing in a genome, transcriptome and proteome context. The impact of transcriptome studies in human health research has been shown for many fields in recent times. Their applications include assessing the safety of food, drugs, vaccines, medical devices and other products of consumer interest [41–46]. DNA arrays for the identification of food-borne bacterial pathogens and viruses [47] can be used to reduce the incidence of food poisoning, illness and death associated with bacterial or viral contamination of meat, seafood, dairy products and other foods. Also in clinical settings, the identification of organisms in patients admitted to hospitals with systemic bacterial infections can be envisaged. The capacity to type unambiguously all the common bacteria on a single chip within a few hours of sampling will allow high-speed testing in agricultural, manufacturing and clinical settings. It might be possible that gene-expression patterns will be used to simplify widely used diagnostic descriptions of cancers. When currently as many as 7000 disease-concepts with 42,000 names (and synonyms) are used worldwide to describe different cancers and the number of validated gene-expression profiles for cancers grows, these profiles may offer a useful way to streamline this list and standardize cancer classification on a rational basis [48]. Another possible application will be the test of efficacy and safety of pharmaceuticals, both in clinical trials and treatment. Genotyping by DNA arrays could be used to stratify patients participating in clinical trials into populations of responders and non-responders to enhance the accuracy of drugtesting results, and allowing drugs to be tailored to specific subsets of the population according to clearly identifiable markers in the patient population [30]. DNA arrays could also be used to examine the physiological effects of a specific diet, allowing the analysis of pathways and the identification of reactions in which food and its components are involved in Ref. [49].
62 Table 3. Web resources and databases for transcriptomics. Microarray Informatics at the EBI (www.ebi.ac.uk/microarray) Pat Brown’s lab at Stanford University (cmgm.stanford.edu/pbrown/) Stanford Microarray Database (genome-www5.stanford.edu/) Microarray Gene Expression Data (MGED) Society (www.mged.org/) Database of alternatively spliced proteins ASP at UCLA (www.bioinformatics.ucla.edu/ HASDB/)
This technology will be used as a valuable tool to identify mechanisms by which nutrients interact with the body and how individuals respond to food intake in a specific diet (Table 3). DNA microarrays are currently becoming useful analytical tools for disease profiling. However, there is a pressing need for other profiling technologies that go beyond measuring RNA levels, particularly for disease-related investigations. DNA microarrays have limited utility for the analysis of biological fluids and for the discovery of markers directly in the fluid. To reach that goal, there is a need to assay protein levels and activity. Numerous alterations may occur in proteins that are not reflected in changes at the RNA level, providing a compelling rationale for additional, direct analysis of gene expression at the protein level. The next challenge is to integrate RNA data with protein data [50]. Proteomics Proteomics technologies attempt the large-scale determination of gene and cellular function directly at the protein level. Mass spectrometry (MS) has increasingly become the method of choice for analysis of complex protein samples. MS-based proteomics is a discipline made possible by the availability of gene and genome sequence databases and technical and conceptual advances in instrumentation technology [51]. Proteomics has also established itself as an indispensable technology to interpret the information encoded in genomes. Protein analysis by MS so far, has been most successful when applied to small sets of proteins isolated in specific functional contexts. The systematic analysis of the much larger number of proteins expressed in a cell is now also rapidly advancing, mainly due to the development of new experimental approaches. A single bacterial cell may produce 4000 proteins whose abundances and activities may vary throughout an experiment, while the number of proteins expressed in higher eukaryotes is likely to be at least 10-fold greater. Attempts to catalogue, visualize, and analyze proteomics experiments have therefore become a major challenge (Table 4). Further to the identification of proteins, their quantification can now be addressed. However, no single method or instrument exists that is capable of identifying and quantifying the components of a complex protein sample. Two methods are popular: 2-dimensional electrophoresis (2DE) followed by MS or
63 Table 4. Steps involved in a typical proteomics experiment. Protein isolation from a biological sample (e.g., a cell extract) following some experimental treatment. Fractionation of the resulting proteins (or peptides, the products of proteome digestion) by methods such as two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) or liquid chromatography (LC). Protein or peptide detection by MS. Protein identification through manual interpretation or database correlation of mass spectra.
limited protein purification with automated peptide MS/MS. When accurate quantification is desired, stable-isotope tagging of proteins or peptides is made. While 2DE clearly has its shortcomings [52–54], the use of liquidchromatography combined with tandem MS (LC MS/MS) experiments appears to be an extremely promising technology. However, mass spectrometers are inherently poor quantitative devices. The data collected by this method is comprehensive requiring more sophisticated tools for its analysis than what is presently available. Current challenges for the analysis of MS based data are the development of tools for their high-throughput analysis [55–57]. First steps into that direction do however exist. For example, statistical models to estimate the accuracy of the peptide assignments called ‘‘PeptideProphet’’ [58]. For computing probabilities that proteins are present in a particular sample can be done by ‘‘ProteinProphet’’ [59]. In order to tackle the quantitative analysis of peptide LC-MS/MS experiments, stable-isotope tagging has been developed. Different stable-isotopes can be readily differentiated in a mass spectrometer owing to their mass difference leading to an accurate indication of the abundance ratio for the two samples. This new technique has been applied successfully in several experiments [6,60–62]. An interesting aspect for studying the inner workings of a cellular system is to study their protein machines. Most proteins exert their function by way of protein–protein interactions. Enzymes are often held in tightly controlled regions of the cell by such protein–protein interactions. Thus, protein–protein interactions provide a wealth of information on the fundamental aspects of cellular life. The first of such screens is the yeast-two-hybrid (Y2H) technology [63]. Recently, two studies of biochemical purifications combined with mass spectrometry (MS) were conducted: one uses ‘‘high-throughput MS protein complex identification’’ (HMS-PCI) [64], the other employing ‘‘Tandem affinity purification (TAP) followed by MS identification’’ [65]. Some of the most biologically informative results have come from the analysis of large protein complexes, like the analyses of the spliceosome, followed by the yeast nuclear pore complex [66,67]. The complexity of the biological system on the protein level is further rendered more difficult through protein post-translational modification. These posttranslational modifications modulate the activity of most eukaryote proteins.
64 Their analysis is now pursued using mass spectrometric peptide sequencing and analysis technologies. Furthermore, stable isotope labeling strategies in combination with mass spectrometry have been applied successfully to study the dynamics of modifications [68]. Proteomics is an essential component of systems biology research because proteins are responsible for many crucial processes in the cell. This technology became extremely valuable for the description of biological processes such as protein abundance, linkage maps to other proteins or to other types of biomolecules including DNA and lipids. Proteomics can also address for example protein expression profiling, activities, modification states, and subcellular location. Unfortunately, with the exception of quantitative protein expression profiles and protein–protein interactions none of these properties can currently be measured systematically, quantitatively and with high throughput. But rapid advances in technology suggest that these limitations may be only momentary. The few studies where the same biological system was subjected to different types of systematic measurements already offer insights into the power of the method. For instance, mRNA expression profiles and protein expression profiles seem to be largely complementary and therefore contribute to a more refined description of the system that each observation by itself is unable to provide [10]. Combining different genomic and proteomic results obtained from the same biological system will substantially increase our understanding of complex biological processes. More specifically, the systems biology studies based on diverse and high-quality proteomic data are already defining functional biological modules and reveal previously unrecognized connections between biochemical processes and modules. The new hypotheses that are generated by this approach can be tested either by traditional methods or by the targeted generation of more genomic and proteomic data [10,69–71]. A promising quantitative proteomic profiling method (MS/MS) has recently been reported for glycoproteins using isotope protein tagging as well as automated tandem mass spectrometry [72]. The method is based on the conjugation of glycoproteins to a solid support using hydrazide chemistry, stable isotope labeling of glycopeptides and the specific release of formerly N-linked glycosylated peptides via peptide-N-glycosidase F. The application of this approach to the analysis of plasma membrane proteins and human blood serum proteins promises great potential for the functional analysis of biological systems and for clinical diagnostics or prognostics. The result could be that an individual global-health profile based on protein identifications will become feasible, revolutionizing the field of disease diagnosis and health monitoring. It may be possible in the future that a small sample of blood can reveal an image of the physiological and pathological states of every tissue in the body [73]. In conclusion, the ever-advancing proteomics research represents one of the most promising technologies for the investigation of human health. Only two years ago, scientists reported a simple blood test based on proteomics a technology that successfully detects ovarian cancer even in its early stages [74].
65 Table 5. Web resources and databases for proteomics. ExPASy Proteomics tools (expasy.org/tools/) A Research Pointer to the Applied Proteomics and Proteomics Technologies (http:// proteomicssurf.com) spectroscopyNOW (http://www.spectroscopynow.com) Human Proteome Organization HUPO (http://www.hupo.org) Institute for Systems Biology (http://www.systemsbiology.org/)
Now, clinical laboratories are ready to employ the test. The emerging field of clinical proteomics will provide early diagnostic methods leading the way for potentially curing such diseases (Table 5). Metabolomics Our metabolism is an expression of a transient steady state in the dynamics of cellular biosynthesis. Proteins function either as enzymes, receptors, transporters, channels, hormones and other signaling molecules or provide structural elements for cells, organs or the skeleton. Metabolites, in contrast, serve in an extensive broad range of functions within the cell. Metabolites are usually rapidly ‘‘converted’’ in enzymatic and chemical reactions, serve as building blocks for macromolecules or may serve as transient energy-storage. Therefore, the identification, quantification and the reactions of metabolites are important in the context of systems biology. Metabolomics is considered to be the study of the entire set of metabolites in a cell, tissue or organ sample [75–78]. In many respects, metabolites are the final stage of biological cellular activity along the line from gene to mRNA to protein to function to phenotype (Fig. 4). Analytical approaches that take the chemical complexity and dynamic range of the metabolome into account employ usually an extraction of metabolites from a cell by different techniques followed by parallel analyses of those subfractions. This strategy is required to segregate the metabolome into more manageable subclasses with similar chemical properties that also helps minimizing chemical side-reactions between them. The subclasses are subjected to parallel analytical techniques to record metabolite profile information. Segregation of the subclasses while parallel analyses helps visualize a greater portion of the metabolome. In most cases the methods use classical chromatographic separation techniques that may comprise Fourier-transform infrared spectroscopy (FTIR), electrospray mass spectrometry (ESI-MS) and nuclear magnetic resonance (NMR) spectroscopy. A promising route to the metabolome is the comprehensive metabolic analysis coupled with statistical methods of cluster and phenotype analysis alike. An individual’s health status is rapidly reflected at the metabolic state. Thus, it might be possible for health-care and nutrition practitioners to make recommendations for a specific treatment or food for their condition. To reach this goal, a suitable database based on a large
66
Fig. 4. A network of metabolic pathways illustrating the complexity of metabolism as it is known today (excerpt reproduced with permission from Roche Applied Science’s Biochemical Pathways Michal: Biochemical Pathways, 1998 ß Spektrum Akademischer Verlag, Heidelberg, Berlin).
number of measurements of accurate metabolite concentrations from healthy people is required. Consequently, the development of a public metabolite atlas might be necessary. Specific quantification of metabolites has been used to characterize metabolic processes in a multitude of focused metabolic pathways studies. The developed methods have been optimized to produce high-quality data that describe the compounds of interest. Today, these data constitute of the metabolic states of individuals. However, this type of analysis is poorly suited to simultaneously gathering information on the multitude of metabolites that characterize an organism’s nutritional processes. Another technique, metabolic profiling, has been devised to monitor, in parallel, hundreds or even thousands of metabolites, using high-throughput techniques. This is done to enable screening for relative changes rather than absolute concentrations of compounds. Most analytical techniques for profiling small molecules consists of HPLC or gas chromatograph (GC) coupled to mass spectrometry. Mass spectrometers are generally more sensitive and more selective than any other types of detectors. When coupled with the appropriate sample-introduction and ionization techniques, mass spectrometers can selectively analyze both organic and inorganic compounds. Nevertheless, the metabolites have to be separated prior to detection, by chromatographic techniques that are coupled online to the mass detector. Gas chromatography is used to separate compounds on the basis of their relative vapor pressures and affinities for the material in the chromatography column, but is restricted to compounds that are volatile and heat stable. HPLC separations are better suited for the analysis of labile and high-molecular-weight compounds and for the analysis of non-volatile polar compounds in their natural form. The vast information gathered using high-throughput screening with GC-and HPLC-MS techniques require advanced informatics technologies for analysis. Yet proton Nuclear Magnetic Resonance (1H-NMR) Spectroscopy is dealing with metabolite profiling and allowing information to be gathered on the
67 flow of metabolites through biological processes and the control of the pathways. High-resolution 1H-NMR spectroscopy, with the advantage of detection of any proton-containing metabolite, appears to become more important in the future in metabolite profiling. NMR-techniques have been used in the past mainly to analyze metabolite changes in mammalian body fluids and tissues and this method may be extended by detecting other nuclei, for example 31P or naturalabundance isotopes such as 13C. When metabolomics is applied in studies where substrates enriched in 13C, metabolite analysis can even be taken onto a dynamic level by allowing the fluxes to be determined quantitatively. Such automated biochemical profiling techniques will become an important component of multi-disciplinary integrated approaches in metabolic and functional genomics studies. The previously described technologies of genomics, transcriptomics and metabolomics, have produced a complete ‘‘parts-catalog’’ of the molecular components in many organisms. The next challenge would be to reconstruct and simulate the overall cellular functions. Recently, advances have been made in the area of flux balance analysis and mathematical modeling [79]. Fundamental physicochemical laws and principles are used to systematically describe the living cell. However, serious limitations to this goal are the inability to rationally and exhaustively analyze biochemical networks and to accurately take all parameters into account, e.g., conservation of mass, energy and redox potential as well as mass transfer. An attempt to derive a global model of metabolisms of a cell is presented in the E-Cell software for cell simulation. Given a set of reaction rules and initial values, users can run simulations and observe dynamic changes in quantities and concentrations of intra- and extracellular metabolites and substances through graphical user interfaces. Activities of biochemical reactions can be monitored, as well as amounts of substances can be subject of change (increased/ decreased) by the users at any time during the simulation. e-Cell system makes it possible to conduct in silico metabolic experiments [80]. Furthermore, the availability of many annotated genomes paves the way for a systematic application of flux-balance methods to a large variety of organisms. However, such a high-throughput goal crucially depends on the capacity to build metabolic flux models in an automated fashion [81] (Table 6). Pulling it together The availability of genome sequences, expressed protein repertoires and identified metabolites for several organisms, including humans have allowed the transition from classic analytical biology to ‘‘systems biology.’’ In this new approach, biological processes of interest, mostly systems, are studied as complex networks of functionally interacting macromolecules and reactions. These functional genomics approaches can be helpful to accelerate the identification of the genes and gene products involved in particular modules,
68 Table 6. Web resources and databases for metabolomics.
Metabolomics at University of Wales Aberystwyth (http://dbk.ch.umist.ac.uk/metabol.htm) Biochemical pathways (ExPASy) (http://us.expasy.org/tools/pathways/) Biopathways consortium (http://www.biopathways.org/) BRENDA, the Comprehensive Enzyme Information System (http://www.brenda.uni-koeln.de) EcoCyc and MetaCyc (http://www.ecocyc.org/) GeneCards (http://bioinformatics.weizmann.ac.il/cards/) KEGG – Kyoto encyclopedia of genes and genomes (http://www.genome.ad.jp/kegg/) E-cell project (http://www.e-cell.org/) Main metabolic pathways on Internet (http://home.wxs.nl/ pvsanten/mmp/main.htm) Metabolic Control Analysis (MCA) (http://dbk.ch.umist.ac.uk/mca_home.htm) MPI for Molecular Plant Physiology (http://www.mpimp-golm.mpg.de/fiehn/index-e.html) PathDB Biochemical Pathways (http://www.ncgr.org/pathdb/) Compugen’s Biocarta (http://www.biocarta.com/) Interactive metabolic reconstruction on the web WIT (http://wit.mcs.anl.gov/WIT2/) EMP Database of Enzymes and Metabolic pathways (http://emp.mcs.anl.gov)
Table 7. Useful databases for protein interaction
Database of Interacting Proteins DIP [86] (dip.doe-mbi.ucla.edu/) BIND [87] (http://bind.ca) PathCalling Yeast Interaction Database [63] (portal.curagen.com/) Mammalian protein–protein interaction database (PPI) [88] (fantom21.gsc.riken.go.jp/PPI/) Molecular Interaction database MINT [89] (160.80.34.4/mint/) General Repository Interaction Datasets GRID [90] (biodata.mshri.on.ca/grid/servlet/Index)
and to describe the functional relationships between them. However, the data emerging from individual ‘‘omic’’ approaches should be viewed with caution because of the occurrence of false-negative and false-positive results [82]. One of the problems biologists face is that the data set too large to comprehend in full. Novel and useful databases are being developed in recent times reflecting progress in different aspects of genomics [83], prompting the saying that we live in ‘‘the age of databases.’’ In the new age of computational biology, it is not enough to publish scientific results in the literature, but the data has to be stored in a structured way both for retrieval and to connect to other resources on the web. Computer databases first rose to prominence in life science as central repositories for nucleic acid and protein sequences. Their interrogation via e.g., the BLAST sequence search tool [84] is now performed frequently by biologists. After the establishment of GenBank in 1982 [85], many other databases have been developed that will be important for systems biology (Table 7). Some of these databases for example contain searchable indices of known protein-protein interactions. The current limiting factor in these databases however is the quality of information. High-quality information of validated protein–protein interactions
69 is so far only available for yeast [91] and the fruit-fly [92]. Very few largescale high quality data sets for mammalian systems are available in the public domain. TRANSFAC [93] and SCPD [94] catalog interactions between proteins and DNA (i.e., transcription factor interactions), and databases of metabolic pathways have also recently been established e.g., EcoCyc [95], KEGG [96], and WIT [97]. A growing number of databases are under development for storing gene-expression data sets, as for example ArrayExpress [98], Gene Expression Omnibus [99] and the Stanford Microarray Database [100]. This recent explosion, in both the variety and volume of information of interest poses two challenges to database users and developers alike. First, the information must be maintained systematically in a format that is compatible with both single queries and global searches. Often, the desired information is present in the database but is not annotated consistently for all entries. We therefore need systems that integrate data globally [11]. Apart from computer-generated databases, high-quality databases require very often manual work of curators. This time intensive approach is well exemplified in the Human Protein Reference Database (HPRD) [101] (www.hprd.org/). Information relevant to the function of human proteins in health and disease is collected including protein–protein interactions, post-translational modifications, enzyme/ substrate relationships, disease associations, tissue expression, and sub-cellular localization. The data is collected from more than >300,000 published articles for a non-redundant set of 2750 human proteins. The HPRD database as well as others of its kind put existing information in computer-readable format. They represent bioinformatics platforms that are useful in cataloging and mining the large number of proteomic interactions and alterations that are about to be discovered with systems biology approaches. Storing existing knowledge in structured ways is the key challenge and the cornerstone for the new biology (Fig. 5).
PEX14 SEC35 VMA22 TIP20
YPR105C YLR315W YMR181C YOR164C YOR331C
Fig. 5. Visualization of protein interaction using the PathCalling resource. The TIP20 protein, a transport protein that interacts with Sec20p, required for protein transport from the endoplasmic reticulum to the golgi apparatus, shows interaction with protein YPR105C, which itself interacts with many other proteins. This information allows for a rapid evaluation of the functionality of a protein within the context of whole proteome.
70
YBR093C
YAL038W YCR012W
YOL127W YIL0697
YIL13 YER074W YDR171W
YHR174W YGR254W YOL086C
YPL075W YLR127W
YOL120C YML024W
YDR050C
YNL301C YNL216W YIL0697 YER179W
YNL199C YPR048W
YPR048W
YPR048W
Fig. 6. Visualization of a selection of the 331 genes containing network described in Ref. [11] using Cytoscape version 1.1.1. Proteins were selected from the full yeast genome based on their having significant expression change at least 1 of 20 conditions: The wild type (wt) strain and nine genetically altered yeast strains, perturbed environmentally by growth in the presence (+gal) or absence ( gal) of 2% galactose sugar. Each altered strain has a complete deletion of one of GAL genes, which encode proteins needed for the metabolism of galactose. Cytoscape is used to display all information regarding nodes (proteins) and edges (interactions). Here, nodes are represented by grey circles, and interactions/edges are represented by colored lines.
The most complex adventure we are facing now is to achieve a description of cellular biology. Current theories are able to capture and model only a small portion of the data at a time. General approaches to integrate, visualize and model information about cells that will help broaden biological understanding are necessary. To increase the reliability of gene function annotation, multiple independent datasets need to be integrated. Such integration will be crucial for systems biology to achieve its promise (Fig. 6). In order for databases to interact, and researchers to exchange information about their biological observations of a system, a common representationlanguage for storing biochemical models is required. The Systems Biology Markup Language (SBML) was created for that purpose [102]. It is a machinereadable format for the representation of computational models in systems biology. It is expressed in XML (www.w3.org/XML/), and contains structures for representing compartments, species and reactions, as well as optional unit definitions, parameters and rules. SBML will be crucial for the storage and exchange of data between databases. The rapidly expanding biological datasets of physical, genetic and functional interactions present a daunting task for data visualization and evaluation [103].
71 Completely new concepts are required in order to help scientists understand complex data. The Cytoscape software, for example, attempts to integrate biomolecular interaction networks with high-throughput expression data and other molecular states into a unified conceptual framework. It is applicable to any system of molecular components and interactions, and most powerful when used in conjunction with large databases of protein–protein, protein– DNA, and genetic interactions that are increasingly available for humans and model organisms. The tools provide functionality to layout and query interaction networks; visually integrate the network with expression profiles, phenotypes, and other molecular states and linking to databases for functional annotations. An important facet of the tool is that it is extensible through plug-ins, allowing rapid development of additional computational analyses [104]. Another approach is presented in the Osprey software [90] that represents interactions in a flexible and expandable graphical format and provides options for functional comparisons between datasets. Systems biology involves interaction between experiment and simulation, attempting to create ever more accurate models of processes, such as the functioning of an organ over a period of time. Initially, a rough working model is created and used to design experiments that will verify or refute the predictions of that model. The model is modified to incorporate results and new simulations that in turn require further experiments. In this way, both the model and experiments evolve together until a satisfactory simulation can be achieved (Fig. 7). The above-mentioned databases are only covering the cellular level. However, the final goal to capture information about individuals will require databanks TCTTGTCGCACGCAACTT TTGAGGATTTTTAAAGGG TGTCTATACCAAACGGA GAGGAGTAATGATGAGT GGTTAAGAATCCATACTT CAAGCAGAATTCGGGGC GGTTACCAAGCGAC
Biological question
DNA Cells
RNA
Vmax .[S]
In-silico experiment, simulation
[S]+ K m Models Networks
Proteins Intensity, counts
v=
New hypothesis
500 400
Experiments
300
New data
Metabolites
200 100 0 600 1000 1400 1800 2200 2600 m/z, au
Biological System Databanks
Result: New insight
Fig. 7. Systems biology iterative research. Data about the living cell will reside in structured databases that are used to test-out new hypothesis and for proof of new models describing the system. In an iterative process, going back and forth between in-vitro, in-vivo, and in-silico experiments, new insight is created.
72 with information about people, biopsies or body fluid samples, stored to be analyzed for genetic and biochemical assessment. The UK Biobank project (www.ukbiobank.ac.uk/) will pursue exactly this goal. Up to half a million participants aged between 45 and 69 years will be involved in the study. They will be asked to contribute a blood sample, lifestyle details and their medical histories to create a national database of unprecedented size. With such databases, fears and uncertainties dealing with ethical issues have surfaced and are under continuous debate. The concerns associated with single-gene disorders, such as privacy, confidentiality, potential employment, or insurance discrimination and the rights of family members, are relevant. Additional factors include the nuanced meaning of genetic risk in complex diseases that result from genetic, environmental, and lifestyle interactions. The blurred boundary between medicine and genetic enhancement and the social implications of predicting diseases among a large fraction of the population, not to mention the gulf between identifying susceptibility and providing preventive treatment are subject of discussion. There is clearly a need to foster a public debate about the customization of diets or medical treatment to match the genetic profiles of consumers in the interest of preventing or managing chronic health conditions. This discussion needs to be initiated as fast as possible. To ease those fears, the latest decision of the US senate passed a bill, which would bar employers from using genetic information in making employment decisions, and prohibit health insurers from using genetic information to deny coverage or set rates [105] (Table 8). Health: the focus of systems biology ‘‘Let food be your medicine and medicine be your food.’’ Hippocrates, the father of modern medicine, c. 400 BC Biological research using molecular information on all cellular levels is addressing human health in a completely new ways. Disease prevention through Table 8. Web resources and databases for data-integration.
Physiome Project (http://www.physiome.org/) Systems Biology Software at the Keck Graduate Institute (http://www.cds.caltech.edu/ hsauro/) Virtual Cell Project of The National Resource for Cell Analysis and Modeling (http:// www.nrcam.uchc.edu/) E-Cell Project (http://www.e-cell.org/) Microbial Cell Project (http://microbialcellproject.org/) World Wide Web Instructional Committee Virtual Cell (http://www.ndsu.nodak.edu/instruct/ mcclean/vc/) Cytoscape (www.cytoscape.org/) GoMiner (http://discover.nci.nih.gov/gominer/) Database of functional networks at EMBL (http://www.ebi.ac.uk/research/pfmp/)
73 nutritional intervention and/or intelligent medical treatment is realized to be crucial for increased human quality of life. The combination of individual assessment of health status and the resulting personalized interventions can be envisaged in this decade. It can be estimated that by the year 2010 predictive genetic test will be available for as many as a dozen common disorders [106]. Individuals who choose to learn about their susceptibility to these diseases will be able to use this information to take preventive measures. For example, a woman at increased risk for developing breast cancer may want to have more frequent mammograms. A man susceptible to coronary heart disease may take medication to lower his cholesterol. Other people may reduce their risk for disease by changing their diet, getting more exercise and avoiding environmental agents that trigger disease. Genes are being identified that influence how a person responds to a given drug. Increasingly, doctors will prescribe drugs based on the genetic profiles of their patients. This individualized treatment will allow using the drug most likely to treat disease symptoms and also to minimize adverse drug reactions. Such an approach will usher in an era of personalized medicine. The tools of systems biology, by virtue of measuring all constituents of an organism, will have large implications for disease prevention via diet or other environmental factors such as lifestyle. Understanding human health will depend of a holistic view of our body’s biology and the numerous environmental cues to which we are constantly exposed. These include pollutants, toxins, pathogens, commensals and also radiation. Our gastrointestinal system, for example, is the organ with greatest contact to our environment; it is inhabited with a large number of bacteria, termed the microbiome [107]. Exploring the human microbiome during the different states of health, using molecular techniques have partly been initiated. These studies will lead to crucial insights about the relationship of the micro cosmos in our gut and us. These surveys are important for a number of reasons [108]. As adults, our total microbial population that is residing mainly in the intestine is composed of 500 to 1000 species. Their total number is at least one order of magnitude bigger than our somatic and germ cells altogether, with their total number of genes exceeding our own genes by a factor of 100. The microbiota residing in our body functions as a multifunctional organ with multiple implications for our health. In addition to the numerous but poorly characterized beneficial effects of the endogenous microflora on human health, a proper understanding of abundance and variations therein will be critical for recognizing potential patterns that are predictive of health or disease. We have virtually no information on the levels of microbial diversity and abundance that are optimal for maintenance of human health, or of those that are associated with disease. With only few gut microorganisms sequenced [109–112] we are just starting to learn about microbial partitioning within human micro-environments. We still understand little about inter-individual variability or variability as a function of time. Gut bacteria have also been
74 Duodenum and Jejunum: 102-105 cfu ml−1 Lactobacillus Streptococcus Bifidobacterium Enterobactericeae Staphylococcus Yeast
Ileum and Caecum: 103-109 cfu ml−1 Bifidobacterium Bacteroides Lactobacillus Streptococcus Enterobactericeae Staphylococcus Clostridium Yeast
Stomach: 100-103 cfu ml−1 Lactobacillus Streptococcus Staphylococcus Enterobactericeae Yeast
Colon: 1010-1012 cfu g−1 Bacteroides Eubacterium Clostridium Peptostreptococcus Streptococcus Bifidobacterium Fusobacterium Lactobacillus Enterobactericeae Staphylococcus Yeast
Fig. 8. Composition of the human gastro-intestinal micro-biota. The overall number of microorganisms in our body is estimated to be bigger than the number of all our somatic and germ cells [108,114,115].
implicated in colon cancer development, but their role in tumor invasion, which is modulated by environmental factors, has been unclear. One report states that a metalloprotease from Listeria monocytogenes, in combination with a host protease, produces a peptide that stimulates motility and invasion of colon cancer cells [113]. The pro-invasive factor was identified as peptide derived from bovine b-casein. This peptide could be generated in vitro by the combined actions of the L. monocytogenes metalloprotease Mpl and a trypsin-like serine protease present in the collagen used in the cell invasion assay. That data shows convincingly that the combined action of diet, bacteria and host elements can produce health impairments. Thus, detailed knowledge about the somatic and germ cells which make up our corpus has to be extended with knowledge about the microbiome and its interaction with our body (Fig. 8). Intensive research has focused in the past on protection of individuals from various stresses using food ingredients such as anti-oxidants [116]. However, a recent report underlines the significance of natural products for human health apposed to purified ingredients and especially their importance for prevention of disease. Lycopene, a carotenoid found in tomato products, was long known for its anti-oxidant effects [117]. It is used frequently as a purified additive to foods. In this new study [118] tomato powder was shown to inhibit the development of prostate cancer compared with a control diet, whereas a diet containing a pure synthetic lycopene supplement did not. The authors also found that the equally measured restriction on energy intake due to experimental conditions during the experiment produced a reduction in prostate cancer mortality that was independent of the effect of tomato powder. This new
75 Soul
Food Drugs Pollu tants
Body composition
Genome
Health
Exogenous bacteria
Age Disease
Stress Lifestyle Other factors
Microflora
Physiology
Fig. 9. Interaction of the environment with the human body. Environmental factors will influence healthy state of an organism taking individual genetic factors into account. Genetic predisposition may lead to body-dysfunction later in life that is modulated by nutrition and other environmental factors.
study is important on several levels. Perhaps most important, it weighs heavily in the debate about whether cancer prevention is best achieved via whole foods versus via single compounds. Another striking aspect is that caloric restriction can readily lead to disease prevention [119]. It has to be realized from this study that the ultimate biologic activity of a given food or nutrient depends on a large number of variables, including food processing and preparation method, gastrointestinal tract physiology, interactions between compounds in the food, and interactions between foods eaten together at the same meal. Clearly, we have barely begun to scratch the surface of understanding how the nutrient compounds within natural food interact within our biologic systems. The promise of systems biology is to grasp the potential complexity of the relevant effects in humans, untangling these interactions in the laboratory (Fig. 9, Table 9).
Conclusions and outlook What has systems biology achieved for the complete description of human biology? With all the advances we have to realize that getting from a gene to a human being is not as straightforward as some had hoped. Although a starting point of a genomic approach to health research is identifying the mechanisms how components interact in a healthy state and also which genes
76 Table 9. Web resources for health genomics.
European Nutrigenomics Organisation NuGO (www.nugo.org) Nutrigenomics.UCDavis.edu (nutrigenomics.ucdavis.edu) The Centre for Human NutriGenomics (http://www.nutrigenomics.nl/) The IFR Food and Health Network (http://www.foodandhealthnetwork.com/) Nutrition, Metabolism and Genomics Group http://nutrigene.4t.com Center for Nutrigenomics TU Munich (http://www.nutriogenomics.com/) Human Genome Project Information from the DOE (link) NCBI Science primer Pharmacogenomics (http://www.ncbi.nlm.nih.gov/About/primer/ pharm.html) International Society of Pharmacogenomics (www.pharmacogenomics.org.uk/) PharmGKB (http://www.pharmgkb.org)
are associated with disease, the sheer complexity of our biology is projecting this goal far out into the future. Increasing evidence suggests that the genetic makeup may partially explain why people of different ancestry experience disease or metabolize nutrients differently. Yet, these genetic clues have to become firm enough to guide medical practice. Genomics, proteomics, and bioinformatics are just beginning to influence the practice of medicine, most notably in diagnosis of disease and development of drugs and recommendations for nutritious foods. To accelerate this influence, physicians must be better prepared. They need to understand the nature of the tests and the kinds of information from which they will make clinical inferences and assist patients in making clinical decisions, always taking cultural and ethical considerations into account [120]. Translating genomic information into successful clinical trials will require advances on several fronts. Despite extensive preclinical studies, the vast majority of clinical trials fail because the drugs do not work as anticipated in patients or lead to intolerable side effects, mainly due to the lack of basic information about physiology and the difficulty to predicting which treatments is likely to succeed. Current medical practice treats illnesses after they appear. However, with the extended human lifespan, averting one illness enables a person to live long enough to contract another. Therefore, disease prevention and how to reach a global healthy state of our body must be the new focus of research. Nutrition and life-style of individuals can play a primary role in that battle. At some point in the future, genomic information and individual susceptibility data will be part of our healthcare system in which we will try to intervene or prevent at the earliest possible time, rather than what we are doing now, that is treating after an event occurred. Crucial factors for our new health-care consciousness of the public will be genome, proteomic, metabolic and informatics technologies, moving away from the reactive ‘‘fix-it’’ medical treatment towards a proactive, prospective and preventive medicine. This new health concept will start with a personalized assessment of individual nutritional status, environmental factors and life-style factors like sport and risk to disease and finalize in an individual lifestyle and healthcare plan.
77 Therefore, we need to obtain detailed knowledge about how proteins participate in the physiological processes in our bodies. We must know what the normal healthy state of our body is and how we might intervene to prevent any inconvenient condition of health. As for predictive medicine, we will require extensive information to analyze the genetic contribution to disease and, at least for the common afflictions, we will need to know the environmental components as well. Nutrition can certainly help in dealing with the growing problem of obesity. But how can we change the eating habits of our children? Should we exercise greater control over what people eat? But do we know which diet is adapted for us? How can we judge if there is too few scientific data. Clearly, the completion of our genome sequence is not the end of our quest, neither is it the beginning; maybe, it is only the end of the beginning. References 1. Wolkenhauer O. Systems Biology: The Reincarnation of Systems Theory Applied in Biology?. New York, Columbia University Press, 2001, pp. 258–270. 2. Rosen R. Essays on life itself. New York, Columbia University Press, 1999. 3. Lander ES, Linton LM, Birren B, Nusbaum C, Zody MC, Baldwin J, Devon K, Dewar K, Doyle M, FitzHugh W, Funke R, Gage D, Harris K, Heaford A, Howland J, Kann L, Lehoczky J, LeVine R, McEwan P, McKernan K, Meldrim J, Mesirov JP, Miranda C, Morris W, Naylor J, Raymond C, Rosetti M, Santos R, Sheridan A, Sougnez C, StangeThomann N, Stojanovic N, Subramanian A, Wyman D, Rogers J, Sulston J, Ainscough R, Beck S, Bentley D, Burton J, Clee C, Carter N, Coulson A, Deadman R, Deloukas P, Dunham A, Dunham I, Durbin R, French L, Grafham D, Gregory S, Hubbard T, Humphray S, Hunt A, Jones M, Lloyd C, McMurray A, Matthews L, Mercer S, Milne S, Mullikin JC, Mungall A, Plumb R, Ross M, Shownkeen R, Sims S, Waterston RH, Wilson RK, Hillier LW, McPherson JD, Marra MA, Mardis ER, Fulton LA, Chinwalla AT, Pepin KH, Gish WR, Chissoe SL, Wendl MC, Delehaunty KD, Miner TL, Delehaunty A, Kramer JB, Cook LL, Fulton RS, Johnson DL, Minx PJ, Clifton SW, Hawkins T, Branscomb E, Predki P, Richardson P, Wenning S, Slezak T, Doggett N, Cheng JF, Olsen A, Lucas S, Elkin C, Uberbacher E, Frazier M, Gibbs RA, Muzny DM, Scherer SE, Bouck JB, Sodergren EJ, Worley KC, Rives CM, Gorrell JH, Metzker ML, Naylor SL, Kucherlapati RS, Nelson DL, Weinstock GM, Sakaki Y, Fujiyama A, Hattori M, Yada T, Toyoda A, Itoh T, Kawagoe C, Watanabe H, Totoki Y, Taylor T, Weissenbach J, Heilig R, Saurin W, Artiguenave F, Brottier P, Bruls T, Pelletier E, Robert C, Wincker P, Smith DR, DoucetteStamm L, Rubenfield M, Weinstock K, Lee HM, Dubois J, Rosenthal A, Platzer M, Nyakatura G, Taudien S, Rump A, Yang H, Yu J, Wang J, Huang G, Gu J, Hood L, Rowen L, Madan A, Qin S, Davis RW, Federspiel NA, Abola AP, Proctor MJ, Myers RM, Schmutz J, Dickson M, Grimwood J, Cox DR, Olson MV, Kaul R, Raymond C, Shimizu N, Kawasaki K, Minoshima S, Evans GA, Athanasiou M, Schultz R, Roe BA, Chen F, Pan H, Ramser J, Lehrach H, Reinhardt R, McCombie WR, de la BM, Dedhia N, Blocker H, Hornischer K, Nordsiek G, Agarwala R, Aravind L, Bailey JA, Bateman A, Batzoglou S, Birney E, Bork P, Brown DG, Burge CB, Cerutti L, Chen HC, Church D, Clamp M, Copley RR, Doerks T, Eddy SR, Eichler EE, Furey TS, Galagan J, Gilbert JG, Harmon C, Hayashizaki Y, Haussler D, Hermjakob H, Hokamp K, Jang W, Johnson LS, Jones TA, Kasif S, Kaspryzk A, Kennedy S, Kent WJ, Kitts P, Koonin EV, Korf I, Kulp D, Lancet D, Lowe TM, McLysaght A, Mikkelsen T, Moran JV, Mulder N, Pollara VJ, Ponting CP,
78
4.
5. 6. 7. 8. 9. 10. 11.
Schuler G, Schultz J, Slater G, Smit AF, Stupka E, Szustakowski J, Thierry-Mieg D, ThierryMieg J, Wagner L, Wallis J, Wheeler R, Williams A, Wolf YI, Wolfe KH, Yang SP, Yeh RF, Collins F, Guyer MS, Peterson J, Felsenfeld A, Wetterstrand KA, Patrinos A, Morgan MJ and Szustakowki J. Initial sequencing and analysis of the human genome. Nature 2001;409: 860–921. Venter JC, Adams MD, Myers EW, Li PW, Mural RJ, Sutton GG, Smith HO, Yandell M, Evans CA, Holt RA, Gocayne JD, Amanatides P, Ballew RM, Huson DH, Wortman JR, Zhang Q, Kodira CD, Zheng XH, Chen L, Skupski M, Subramanian G, Thomas PD, Zhang J, Gabor Miklos GL, Nelson C, Broder S, Clark AG, Nadeau J, McKusick VA, Zinder N, Levine AJ, Roberts RJ, Simon M, Slayman C, Hunkapiller M, Bolanos R, Delcher A, Dew I, Fasulo D, Flanigan M, Florea L, Halpern A, Hannenhalli S, Kravitz S, Levy S, Mobarry C, Reinert K, Remington K, Abu-Threideh J, Beasley E, Biddick K, Bonazzi V, Brandon R, Cargill M, Chandramouliswaran I, Charlab R, Chaturvedi K, Deng Z, Di FV, Dunn P, Eilbeck K, Evangelista C, Gabrielian AE, Gan W, Ge W, Gong F, Gu Z, Guan P, Heiman TJ, Higgins ME, Ji RR, Ke Z, Ketchum KA, Lai Z, Lei Y, Li Z, Li J, Liang Y, Lin X, Lu F, Merkulov GV, Milshina N, Moore HM, Naik AK, Narayan VA, Neelam B, Nusskern D, Rusch DB, Salzberg S, Shao W, Shue B, Sun J, Wang Z, Wang A, Wang X, Wang J, Wei M, Wides R, Xiao C, Yan C, Yao A, Ye J, Zhan M, Zhang W, Zhang H, Zhao Q, Zheng L, Zhong F, Zhong W, Zhu S, Zhao S, Gilbert D, Baumhueter S, Spier G, Carter C, Cravchik A, Woodage T, Ali F, An H, Awe A, Baldwin D, Baden H, Barnstead M, Barrow I, Beeson K, Busam D, Carver A, Center A, Cheng ML, Curry L, Danaher S, Davenport L, Desilets R, Dietz S, Dodson K, Doup L, Ferriera S, Garg N, Gluecksmann A, Hart B, Haynes J, Haynes C, Heiner C, Hladun S, Hostin D, Houck J, Howland T, Ibegwam C, Johnson J, Kalush F, Kline L, Koduru S, Love A, Mann F, May D, McCawley S, McIntosh T, McMullen I, Moy M, Moy L, Murphy B, Nelson K, Pfannkoch C, Pratts E, Puri V, Qureshi H, Reardon M, Rodriguez R, Rogers YH, Romblad D, Ruhfel B, Scott R, Sitter C, Smallwood M, Stewart E, Strong R, Suh E, Thomas R, Tint NN, Tse S, Vech C, Wang G, Wetter J, Williams S, Williams M, Windsor S, Winn-Deen E, Wolfe K, Zaveri J, Zaveri K, Abril JF, Guigo R, Campbell MJ, Sjolander KV, Karlak B, Kejariwal A, Mi H, Lazareva B, Hatton T, Narechania A, Diemer K, Muruganujan A, Guo N, Sato S, Bafna V, Istrail S, Lippert R, Schwartz R, Walenz B, Yooseph S, Allen D, Basu A, Baxendale J, Blick L, Caminha M, Carnes-Stine J, Caulk P, Chiang YH, Coyne M, Dahlke C, Mays A, Dombroski M, Donnelly M, Ely D, Esparham S, Fosler C, Gire H, Glanowski S, Glasser K, Glodek A, Gorokhov M, Graham K, Gropman B, Harris M, Heil J, Henderson S, Hoover J, Jennings D, Jordan C, Jordan J, Kasha J, Kagan L, Kraft C, Levitsky A, Lewis M, Liu X, Lopez J, Ma D, Majoros W, McDaniel J, Murphy S, Newman M, Nguyen T, Nguyen N and Nodell M. The sequence of the human genome. Science 2001;291: 1304–1351. Stuart JM, Segal E, Koller D and Kim SK. A Gene-coexpression Network for global discovery of conserved genetic modules. Science 2003;302:249–255. Zhou H, Ranish JA, Watts JD and Aebersold R. Quantitative proteome analysis by solidphase isotope tagging and mass spectrometry. Nat Biotechnol 2002;20:512–515. Price ND, Reed JL, Papin JA, Wiback SJ and Palsson BO. Network-based analysis of metabolic regulation in the human red blood cell. J Theor Biol 2003;225:185–194. Arias E, Anderson RN, Kung HC, Murphy SL and Kochanek KD. Deaths: final data for 2001. Natl Vital Stat Rep 2003;52:1–115. Kitano H. Computational Systems Biology. Nature 2002;420:206–210. Ideker T, Galitski T and Hood L. A new approach to decoding life: systems biology. Annu Rev Genomics Hum Genet 2001;2:343–372. Ideker T, Thorsson V, Ranish JA, Christmas R, Buhler J, Eng JK, Bumgarner R, Goodlett DR, Aebersold R and Hood L. Integrated genomic and proteomic analyses of a systematically perturbed metabolic network. Science 2001;292:929–934.
79 12. Poincare´, H. Science and hypothesis; with a preface by J. Larmor. 13. Mungall AJ, Palmer SA, Sims SK, Edwards CA, Ashurst JL, Wilming L, Jones MC, Horton R, Hunt SE, Scott CE, Gilbert JG, Clamp ME, Bethel G, Milne S, Ainscough R, Almeida JP, Ambrose KD and Andrews. The DNA sequence and analysis of human chromosome 6. Nature 2003;425:805–811. 14. Stein LD, Bao Z, Blasiar D, Blumenthal T, Brent MR, Chen N, Chinwalla A, Clarke L, Clee C, Coghlan A, Coulson A, D’Eustachio P, Fitch DH, Fulton LA, Fulton RE, GriffithsJones S, Harris TW, Hillier LW, Kamath R, Kuwabara PE, Mardis ER, Marra MA, Miner TL, Minx P, Mullikin JC, Plumb RW, Rogers J, Schein JE, Sohrmann M, Spieth J, Stajich JE, Wei C, Willey D, Wilson RK, Durbin R and Waterston RH. The genome sequence of caenorhabditis briggsae: A platform for comparative genomics. PLoS Biol 2003; 1:E45. 15. Pennisi E. Bioinformatics, Gene counters struggle to get the right answer. Science 2003;301: 1040–1041. 16. Hood L and Galas D. The digital code of DNA. Nature 2003;421:444–448. 17. Davidson EH, Rast JP, Oliveri P, Ransick A, Calestani C, Yuh CH, Minokawa T, Amore G, Hinman V, Arenas-Mena C, Otim O, Brown CT, Livi CB, Lee PY, Revilla R, Rust AG, Pan Z, Schilstra MJ, Clarke PJ, Arnone MI, Rowen L, Cameron RA, McClay DR, Hood L and Bolouri H. A genomic regulatory network for development. Science 2002;295:1669–1678. 18. Lander ES. The new genomics: global views of biology. Science 1996;274:536–539. 19. Jasny BR and Roberts L. Are we there yet? Science 2003;302:587. 20. Pirmohamed M and Park BK. Cytochrome P450 enzyme polymorphisms and adverse drug reactions. Toxicology 2003;192:23–32. 21. Staddon S, Arranz MJ, Mancama D, Mata I and Kerwin RW. Clinical applications of pharmacogenetics in psychiatry. Psychopharmacology (Berl) 2002;162:18–23. 22. Higashi MK, Veenstra DL, Kondo LM, Wittkowsky AK, Srinouanprachanh SL, Farin FM and Rettie AE. Association between CYP2C9 genetic variants and anticoagulation-related outcomes during warfarin therapy. JAMA 2002;287:1690–1698. 23. Cortellino S, Turner D, Masciullo V, Schepis F, Albino D, Daniel R, Skalka AM, Meropol NJ, Alberti C, Larue L and Bellacosa A. The base excision repair enzyme MED1 mediates DNA damage response to antitumor drugs and is associated with mismatch repair system integrity. Proc Natl Acad Sci USA 2003. 24. Melton L. Pharmacogenetics and genotyping: on the trail of SNPs. Nature 2003;422:917. 25. Johnson JA. Pharmacogenetics: potential for individualized drug therapy through genetics. Trends Genet 2003;19:660–666. 26. Dennis C. Epigenetics and disease: Altered states. Nature 2003;421:686–688. 27. Wolffe AP and Matzke MA. Epigenetics: regulation through repression. Science 1999;286: 481–486. 28. Waterland RA and Jirtle RL. Transposable elements: targets for early nutritional effects on epigenetic gene regulation. Mol Cell Biol 2003;23:5293–5300. 29. Clark MS, Edwards YJ, Peterson D, Clifton SW, Thompson AJ, Sasaki M, Suzuki Y, Kikuchi K, Watabe S, Kawakami K, Sugano S, Elgar G and Johnson SL. Fugu ESTs: New resources for transcription analysis and genome annotation. Genome Res 2003;13:2747–2753. 30. Stears RL, Martinsky T and Schena M. Trends in microarray analysis. Nat Med 2003;9: 140–145. 31. Lockhart DJ, Dong H, Byrne MC, Follettie MT, Gallo MV, Chee MS, Mittmann M, Wang C, Kobayashi M, Horton H and Brown EL. Expression monitoring by hybridization to highdensity oligonucleotide arrays. Nat Biotechnol 1996;14:1675–1680. 32. Velculescu VE, Zhang L, Vogelstein B and Kinzler KW. Serial analysis of gene expression. Science 1995;270:484–487. 33. Hackett CS, Hodgson JG, Law ME, Fridlyand J, Osoegawa K, de Jong PJ, Nowak NJ, Pinkel D, Albertson DG, Jain A, Jenkins R, Gray JW and Weiss WA. Genome-wide array
80
34. 35.
36. 37.
38.
39. 40.
41. 42.
43. 44.
45. 46. 47. 48. 49.
50.
CGH analysis of murine neuroblastoma reveals distinct genomic aberrations which parallel those in human tumors. Cancer Res 2003;63:5266–5273. Howbrook DN, van der VaA, O’Shaughnessy MC, Sarker DK, Baker SC and Lloyd AW. Developments in microarray technologies. Drug Discov Today 2003;8:642–651. Yamada K, Lim J, Dale JM, Chen H, Shinn P, Palm CJ, Southwick AM, Wu HC, Kim C, Nguyen M, Pham P, Cheuk R, Karlin-Newmann G, Liu SX, Lam B, Sakano H, Wu T, Yu G, Miranda M, Quach HL, Tripp M, Chang CH, Lee JM, Toriumi M, Chan MM, Tang CC, Onodera CS, Deng JM, Akiyama K, Ansari Y, Arakawa T, Banh J, Banno F, Bowser L, Brooks S, Carninci P, Chao Q, Choy N, Enju A, Goldsmith AD, Gurjal M, Hansen NF, Hayashizaki Y, Johnson-Hopson C, Hsuan VW, Iida K, Karnes M, Khan S, Koesema E, Ishida J, Jiang PX, Jones T, Kawai J, Kamiya A, Meyers C, Nakajima M, Narusaka M, Seki M, Sakurai T, Satou M, Tamse R, Vaysberg M, Wallender EK, Wong C, Yamamura Y, Yuan S, Shinozaki K, Davis RW, Theologis A and Ecker JR. Empirical analysis of transcriptional activity in the arabidopsis genome. Science 2003;302:842–846. Neumann NF and Galvez F. DNA microarrays and toxicogenomics: applications for ecotoxicology? Biotechnol Adv 2002;20:391–419. Pedrioli PGA, Eng J, Hubley R, Pratt B, Nilsson E and Aebersold R. A standard open representation of mass spectrometry data and its application in a proteomics research environment, 2004, Ref Type: Unpublished Work. Taylor CF, Paton NW, Garwood KL, Kirby PD, Stead DA, Yin Z, Deutsch EW, Selway L, Walker J, Riba-Garcia I, Mohammed S, Deery MJ, Howard JA, Dunkley T, Aebersold R, Kell DB, Lilley KS and Roepstorff. A systematic approach to modeling, capturing and disseminating proteomics experimental data. Nat Biotechnol 2003;21: 247–254. Lee C, Atanelov L, Modrek B and Xing Y. ASAP: the alternative splicing annotation project. Nucleic Acids Res 2003;31:101–105. Huang HD, Horng JT, Lee CC and Liu BJ. ProSplicer: a database of putative alternative splicing information derived from protein, mRNA and expressed sequence tag sequence data. Genome Biol 2003;4. Kipps TJ. Advances in classification and therapy of indolent B-cell malignancies. Semin Oncol 2002;29:98–104. al Khaldi SF, Martin SA, Rasooly A and Evans JD. DNA microarray technology used for studying foodborne pathogens and microbial habitats: minireview. J AOAC Int 2002;85: 906–910. Soini H and Musser JM. Molecular diagnosis of mycobacteria. Clin Chem 2001;47: 809–814. Paweletz CP, Charboneau L, Bichsel VE, Simone NL, Chen T, Gillespie JW, EmmertBuck MR, Roth MJ, Petricoin EF, III and Liotta LA. Reverse phase protein microarrays which capture disease progression show activation of pro-survival pathways at the cancer invasion front. Oncogene 2001;20:1981–1989. Beaucage SL. Strategies in the preparation of DNA oligonucleotide arrays for diagnostic applications. Curr Med Chem 2001;8:1213–1244. Chizhikov V, Rasooly A, Chumakov K and Levy DD. Microarray analysis of microbial virulence factors. Appl Environ Microbiol 2001;67:3258–3263. Gene chip for viral discovery. PLoS Biol 2003;1:139–140. Covitz PA. Class struggle: expression profiling and categorizing cancer. Pharmacogenomics J 2003;3:257–260. Berger A, Mutch DM, Bruce GJ and Roberts MA. Unraveling lipid metabolism with microarrays: effects of arachidonate and docosahexaenoate acid on murine hepatic and hippocampal gene expression. Lipids Health Dis 2002;1:2. Hanash S and Creighton C. Making sense of microarray data to classify cancer. Pharmacogenomics J 2003.
81 51. Aebersold R and Mann M. Mass spectrometry-based proteomics. Nature 2003;422: 198–207. 52. Rabilloud T. Two-dimensional gel electrophoresis in proteomics: old, old fashioned, but it still climbs up the mountains. Proteomics 2002;2:3–10. 53. Unlu M, Morgan ME and Minden JS. Difference gel electrophoresis: a single gel method for detecting changes in protein extracts. Electrophoresis 1997;18:2071–2077. 54. Gauss C, Kalkum M, Lowe M, Lehrach H and Klose J. Analysis of the mouse proteome. (I) Brain proteins: separation by two-dimensional electrophoresis and identification by mass spectrometry and genetic variation. Electrophoresis 1999;20:575–600. 55. Link AJ, Eng J, Schieltz DM, Carmack E, Mize GJ, Morris DR, Garvik BM and Yates JR. Direct analysis of protein complexes using mass spectrometry. Nat Biotechnol 1999;17: 676–682. 56. Han DK, Eng J, Zhou H and Aebersold R. Quantitative profiling of differentiation-induced microsomal proteins using isotope-coded affinity tags and mass spectrometry. Nat Biotechnol 2001;19:946–951. 57. Washburn MP, Wolters D and Yates JR. Large-scale analysis of the yeast proteome by multidimensional protein identification technology. Nat Biotechnol 2001;19:242–247. 58. Keller A, Nesvizhskii AI, Kolker E and Aebersold R. Empirical statistical model to estimate the accuracy of peptide identifications made by MS/MS and database search. Anal Chem 2002;74:5383–5392. 59. Nesvizhskii AI, Keller A, Kolker E and Aebersold R. A statistical model for identifying proteins by tandem mass spectrometry. Anal Chem 2003;75:4646–4658. 60. Conrads TP, Issaq HJ and Hoang VM. Current strategies for quantitative proteomics. Adv Protein Chem 2003;65:133–159. 61. Mirgorodskaya OA, Kozmin YP, Titov MI, Korner R, Sonksen CP and Roepstorff P. Quantitation of peptides and proteins by matrix-assisted laser desorption/ionization mass spectrometry using (18)O-labeled internal standards. Rapid Commun Mass Spectrom 2000;14: 1226–1232. 62. Yao X, Freas A, Ramirez J, Demirev PA and Fenselau C. Proteolytic 18O labeling for comparative proteomics: model studies with two serotypes of adenovirus. Anal Chem 2001;73: 2836–2842. 63. Uetz P, Giot L, Cagney G, Mansfield TA, Judson RS, Knight JR, Lockshon D, Narayan V, Srinivasan M, Pochart P, Qureshi-Emili A, Li Y, Godwin B, Conover D, Kalbfleisch T, Vijayadamodar G, Yang M, Johnston M and I. A comprehensive analysis of protein-protein interactions in Saccharomyces cerevisiae. Nature 2000;403:623–627. 64. Ho YP and Hsu PH. Investigating the effects of protein patterns on microorganism identification by high-performance liquid chromatography-mass spectrometry and protein database searches. J Chromatogr A 2002;976:103–111. 65. Gavin AC, Bosche M, Krause R, Grandi P, Marzioch M, Bauer A, Schultz J, Rick JM, Michon AM, Cruciat CM, Remor M, Hofert C, Schelder M, Brajenovic M, Ruffner H, Merino A, Klein K, Hudak M, Dickson D and Rudi. Functional organization of the yeast proteome by systematic analysis of protein complexes. Nature 2002;415:141–147. 66. Rout MP and Aitchison JD. The nuclear pore complex as a transport machine. J Biol Chem 2001;276:16593–16596. 67. Neubauer G, Gottschalk A, Fabrizio P, Seraphin B, Luhrmann R and Mann M. Identification of the proteins of the yeast U1 small nuclear ribonucleoprotein complex by mass spectrometry. Proc Natl Acad Sci USA 2001;94:385–390. 68. Mann M and Jensen ON. Proteomic analysis of post-translational modifications. Nat Biotechnol 2003;21:255–261. 69. Griffin TJ, Gygi SP, Ideker T, Rist B, Eng J, Hood L and Aebersold R. Complementary profiling of gene expression at the transcriptome and proteome levels in Saccharomyces cerevisiae. Mol Cell Proteomics 2002;1:323–333.
82 70. Betts JC, Lukey PT, Robb LC, McAdam RA and Duncan K. Evaluation of a nutrient starvation model of Mycobacterium tuberculosis persistence by gene and protein expression profiling. Mol Microbiol 2002;43:717–731. 71. Guina T, Purvine SO, Yi EC, Eng J, Goodlett DR, Aebersold R and Miller SI. Quantitative proteomic analysis indicates increased synthesis of a quinolone by Pseudomonas aeruginosa isolates from cystic fibrosis airways. Proc Natl Acad Sci USA 2003;100:2771–2776. 72. Zhang H, Li XJ, Martin DB and Aebersold R. Identification and Quantification of N-linked Glycoproteins Using Hydrazide Chemistry, Stable Isotope Labeling and Mass Spectrometry. Berlin, New York, Springer-Verlag, 2003, pp. 660–666. 73. Liotta LA, Ferrari M and Petricoin E. Clinical proteomics: written in blood. Nature 2003; 425:905. 74. Petricoin EF, Ardekani AM, Hitt BA, Levine PJ, Fusaro VA, Steinberg SM, Mills GB, Simone C, Fishman DA, Kohn EC and Liotta LA. Use of proteomic patterns in serum to identify ovarian cancer. Lancet 2002;359:572–577. 75. Nicholson JK and Wilson ID. Opinion: understanding ‘global’ systems biology: metabonomics and the continuum of metabolism. Nat Rev Drug Discov 2003;2:668–676. 76. Watkins SM and German JB. Toward the implementation of metabolomic assessments of human health and nutrition. Curr Opin Biotechnol 2002;13:512–516. 77. German JB, Roberts MA, Fay L and Watkins SM. Metabolomics and individual metabolic assessment: the next great challenge for nutrition. J Nutr 2002;132:2486–2487. 78. German JB, Roberts MA and Watkins SM. Personal metabolomics as a next generation nutritional assessment. J Nutr 2003;133:4260–4266. 79. Kauffman KJ, Prakash P and Edwards JS. Advances in flux balance analysis. Curr Opin Biotechnol 2003;14:491–496. 80. Takahashi K, Ishikawa N, Sadamoto Y, Sasamoto H, Ohta S, Shiozawa A, Miyoshi F, Naito Y, Nakayama Y and Tomita M. E-Cell 2: Multi-platform E-Cell simulation system. Bioinformatics 2003;19:1727–1729. 81. Segre D, Zucker J, Katz J, Lin X, D’Haeseleer P, Rindone WP, Kharchenko P, Nguyen DH, Wright MA and Church GM. From annotated genomes to metabolic flux models and kinetic parameter fitting. OMICS 2003;7:301–316. 82. Ge H, Walhout AJ and Vidal M. Integrating ‘omic’ information: a bridge between genomics and systems biology. Trends Genet 2003;19:551–560. 83. Desiere F, German B, Watzke H, Pfeifer A and Saguy S. Bioinformatics and data knowledge: The new frontiers for nutrition and foods. Trends Food Sci Technol 2002;12:215–229. 84. Altschul SF, Gish W, Miller W, Myers EW and Lipman DJ. Basic local alignment search tool. J Mol Biol 1990;215:403–410. 85. Benson DA, Karsch-Mizrachi I, Lipman DJ, Ostell J and Wheeler DL. GenBank. Nucleic Acids Res 2003;31:23–27. 86. Xenarios I, Salwinski L, Duan XJ, Higney P, Kim SM and Eisenberg D. DIP, the Database of Interacting Proteins: a research tool for studying cellular networks of protein interactions. Nucleic Acids Res 2002;30:303–305. 87. Bader GD, Betel D and Hogue CW. BIND: the Biomolecular Interaction Network Database. Nucleic Acids Res 2003;31:248–250. 88. Suzuki H, Saito R, Kanamori M, Kai C, Schonbach C, Nagashima T, Hosaka J and Hayashizaki Y. The mammalian protein-protein interaction database and its viewing system that is linked to the main FANTOM2 viewer. Genome Res 2003;13:1534–1541. 89. Zanzoni A, Montecchi-Palazzi L, Quondam M, Ausiello G, Helmer-Citterich M and Cesareni G. MINT: a Molecular INTeraction database. FEBS Lett 2002;513:135–140. 90. Breitkreutz BJ, Stark C and Tyers M. Osprey: a network visualization system. Genome Biol 2003;4. 91. Ho Y, Gruhler A, Heilbut A, Bader GD, Moore L, Adams SL, Millar A, Taylor P, Bennett K, Boutilier K, Yang L, Wolting C, Donaldson I, Schandorff S, Shewnarane J, Vo M, Taggart J,
83
92.
93.
94. 95. 96. 97.
98.
99. 100.
101.
102.
103.
Goudreault M, Muskat B, Alfarano C, Dewar D, Lin Z, Michalickova K, Willems AR, Sassi H, Nielsen PA, Rasmussen KJ, Andersen JR, Johansen LE, Hansen LH, Jespersen H, Podtelejnikov A, Nielsen E, Crawford J, Poulsen V, Sorensen BD, Matthiesen J, Hendrickson RC, Gleeson F, Pawson T, Moran MF, Durocher D, Mann M, Hogue CW, Figeys D and Tyers M. Systematic identification of protein complexes in Saccharomyces cerevisiae by mass spectrometry. Nature 2002;415:180–183. Giot L, Bader JS, Brouwer C, Chaudhuri A, Kuang B, Li Y, Hao YL, Ooi CE, Godwin B, Vitols E, Vijayadamodar G, Pochart P, Machineni H, Welsh M, Kong Y, Zerhusen B, Malcolm R, Varrone Z, Collis A, Minto M, Burgess S, McDaniel L, Stimpson E, Spriggs F, Williams J, Neurath K, Ioime N, Agee M, Voss E, Furtak K, Renzulli R, Aanensen N, Carrolla S, Bickelhaupt E, Lazovatsky Y, DaSilva A, Zhong J, Stanyon CA, Finley RL, Jr., White KP, Braverman M, Jarvie T, Gold S, Leach M, Knight J, Shimkets RA, McKenna MP, Chant J and Rothberg JM. A protein interaction map of Drosophila melanogaster. Science 2003;302:1727–1736. Matys V, Fricke E, Geffers R, Gossling E, Haubrock M, Hehl R, Hornischer K, Karas D, Kel AE, Kel-Margoulis OV, Kloos DU, Land S, Lewicki-Potapov B, Michael H, Munch R, Reuter I, Rotert S, Saxel H and Scheer M. TRANSFAC: transcriptional regulation, from patterns to profiles. Nucleic Acids Res 2003;31:374–378. Zhu J and Zhang MQ. SCPD: a promoter database of the yeast Saccharomyces cerevisiae. Bioinformatics 1999;15:607–611. Karp PD, Riley M, Saier M, Paulsen IT, Collado-Vides J, Paley SM, Pellegrini-Toole A, Bonavides C and Gama-Castro S. The EcoCyc database. Nucleic Acids Res 2002;30:56–58. Kanehisa M, Goto S, Kawashima S and Nakaya A. The KEGG databases at GenomeNet. Nucleic Acids Res 2002;30:42–46. Overbeek R, Larsen N, Pusch GD, D’Souza M, Selkov E, Kyrpides N, Fonstein M, Maltsev N and Selkov E. WIT: integrated system for high-throughput genome sequence analysis and metabolic reconstruction. Nucleic Acids Res 2000;28:123–125. Brazma A, Parkinson H, Sarkans U, Shojatalab M, Vilo J, Abeygunawardena N, Holloway E, Kapushesky M, Kemmeren P, Lara GG, Oezcimen A, Rocca-Serra P and Sansone SA. ArrayExpress–a public repository for microarray gene expression data at the EBI. Nucleic Acids Res 2003;31:68–71. Edgar R, Domrachev M and Lash AE. Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res 2002;30:207–210. Sherlock G, Hernandez-Boussard T, Kasarskis A, Binkley G, Matese JC, Dwight SS, Kaloper M, Weng S, Jin H, Ball CA, Eisen MB, Spellman PT, Brown PO, Botstein D and Cherry JM. The stanford microarray database. Nucleic Acids Res 2001;29:152–155. Peri S, Navarro JD, Amanchy R, Kristiansen TZ, Jonnalagadda CK, Surendranath V, Niranjan V, Muthusamy B, Gandhi TK, Gronborg M, Ibarrola N, Deshpande N, Shanker K, Shivashankar HN, Rashmi BP, Ramya MA, Zhao Z, Chandrika KN, Padma N, Harsha HC, Yatish AJ, Kavitha MP, Menezes M, Choudhury DR, Suresh S, Ghosh N, Saravana R, Chandran S, Krishna S, Joy M, Anand SK, Madavan V, Joseph A, Wong GW, Schiemann WP, Constantinescu SN, Huang L, Khosravi-Far R, Steen H, Tewari M, Ghaffari S, Blobe GC, Dang CV, Garcia JG, Pevsner J, Jensen ON, Roepstorff P, Deshpande KS, Chinnaiyan AM, Hamosh A, Chakravarti A and Pandey A. Development of human protein reference database as an initial platform for approaching systems biology in humans. Genome Res 2003;13:2363–2371. Hucka M, Finney A, Sauro HM, Bolouri H, Doyle JC, Kitano H, Arkin AP, Bornstein BJ, Bray D, Cornish-Bowden A, Cuellar AA, Dronov S, Gilles ED, Ginkel M, Gor V, Goryanin II, Hedley WJ and Hodgman TC. The systems biology markup language (SBML): a medium for representation and exchange of biochemical network models. Bioinformatics 2003;19:524–531. Vidal M. A biological atlas of functional maps. Cell 2001;104:333–339.
84 104.
105. 106. 107. 108. 109.
110.
111.
112.
113.
114. 115. 116. 117.
118. 119.
120.
Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, Amin N, Schwikowski B and Ideker T. Cytoscape: A software environment for integrated models of biomolecular interaction networks. Genome Res 2003;13:2498–2504. Collins FS and Watson JD. Genetic discrimination: time to act. Science 2003;302:745. Collins FS, Green ED, Guttmacher AE, Guyer MS and US National Human Genome Research Institute. A vision for the future of genomics research. Nature 2003;422:835–847. Lederberg J. Getting in tune with the enemy – Microbes. The Scientist 2003;17:20. Xu J and Gordon JI. Inaugural article: Honor thy symbionts. Proc Natl Acad Sci USA 2003; 100:10452–10459. Xu J, Bjursell MK, Himrod J, Deng S, Carmichael LK, Chiang HC, Hooper LV and Gordon JI. A genomic view of the human-Bacteroides thetaiotaomicron symbiosis. Science 2003;299:2074–2076. Schell MA, Karmirantzou M, Snel B, Vilanova D, Berger B, Pessi G, Zwahlen MC, Desiere F, Bork P, Delley M, Pridmore RD and Arigoni F. The genome sequence of Bifidobacterium longum reflects its adaptation to the human gastrointestinal tract. Proc Natl Acad Sci USA 2002;99:14422–14427. Paulsen IT, Banerjei L, Myers GS, Nelson KE, Seshadri R, Read TD, Fouts DE, Eisen JA, Gill SR, Heidelberg JF, Tettelin H, Dodson RJ, Umayam L, Brinkac L, Beanan M, Daugherty S, DeBoy RT, Durkin S, Kolonay J, Madupu R, Nelson W, Vamathevan J, Tran B, Upton J, Hansen T, Shetty J, Khouri H, Utterback T, Radune D, Ketchum KA, Dougherty BA and Fraser CM. Role of mobile DNA in the evolution of vancomycinresistant Enterococcus faecalis. Science 2003;299:2071–2074. Pridmore RD, Berger B, Desiere F, Vilanova D, Barretto C, Pittet AC, Zwahlen MC, Rouvet M, Altermann E and Barrangou R. Proc Natl Acad Sci U.S.A. 2004;101: 2512–2517. Oliveira MJ, Van Damme J, Lauwaet T, De C, V, De Bruyne G, Verschraegen G, Vaneechoutte M, Goethals M, Ahmadian MR, Muller O, Vandekerckhove J, Mareel M and Leroy A. beta-casein-derived peptides, produced by bacteria, stimulate cancer cell invasion and motility. EMBO J 2003;22:6161–6173. Savage DC. Gastrointestinal microflora in mammalian nutrition. Annu Rev Nutr 1986; 6:155–78.:155–178. Berg RD. The indigenous gastrointestinal microflora. Trends Microbiol 1996;4:430–435. Urso ML and Clarkson PM. Oxidative stress, exercise and antioxidant supplementation. Toxicology 2003;189:41–54. Goodman GE, Schaffer S, Omenn GS, Chen C and King I. The association between lung and prostate cancer risk and serum micronutrients: results and lessons learned from beta-carotene and retinol efficacy trial. Cancer Epidemiol Biomarkers Prev 2003;12:518–526. Gann PH and Khachik F. Tomatoes or lycopene versus prostate cancer: Is evolution antireductionist? JNCI Cancer Spectrum 2003;95:1563–1565. Hursting SD, Lavigne JA, Berrigan D, Perkins SN and Barrett JC. Calorie restriction, aging and cancer prevention: mechanisms of action and applicability to humans. Annu Rev Med 2003;54:131–52:131–152. Omenn GS. Genetic advances will influence the practice of medicine: examples from cancer research and care of cancer patients. Genet Med 2002;4:15S–20S.
85
Public health issues related with the consumption of food obtained from genetically modified organisms Andrea Paparini1 and Vincenzo Romano-Spica1,2,* 1
University of Rome ‘‘Foro Italico’’ (IUSM), Rome, Italy Catholic University Medical School, Rome, Italy
2
Abstract. Genetically Modified Organisms (GMOs) are a fact of modern agriculture and a major field of discussion in biotechnology. As science incessantly achieves innovative and unexpected breakthroughs, new medical, political, ethical and religious debates arise over the production and consumption of transgenic organisms. Despite no described medical condition being directly associated with a diet including approved GM crops in large exposed populations such as 300,000,000 Americans and a billion Chinese, public opinion seems to look at this new technology with either growing concern or even disapproval. It is generally recognized that a high level of vigilance is necessary and highly desirable, but it should also be considered that GMOs are a promising new challenge for the III Millennium societies, with remarkable impact on many disciplines and fields related to biotechnology. To acquire a basic knowledge on GMO production, GM-food consumption, GMO interaction with humans and environment is of primary importance for risk assessment. It requires availability of clear data and results from rigorous experiments. This review will focus on public health risks related with a GMO-containing diet. The objective is to summarize state of the art research, provide fundamental technical information, point out problems and perspectives, and make available essential tools for further research. Are GMO based industries and GMO-derived foods safe to human health? Can we consider both social, ethical and public health issues by means of a constant and effective monitoring of the food chain and by a clear, informative labeling of the products? Which are the so far characterized or alleged hazards of GMOs? And, most importantly, are these hazards actual, potential or merely contrived? Several questions remain open; answers and solutions belong to science, to politics and to the personal opinion of each social subject. Keywords: biotechnology, food safety, genetically modified organisms, genetically engineered organisms, genetically manipulated organisms, transgenic, animals, plants, horizontal transfer, DNA uptake, DNA intake, genetic modification, genetic manipulation, novel food, GMO, GE, crops, recombinant, agriculture, food allergies, diet, regulations, labeling, food intolerance, antinutrients, herbicide tolerance, insect-resistant, EPSPS, BT, BAR, cry, nptII, public health, antibiotic resistance.
Introduction For thousands of years, thoroughly unaware of even the existence of nucleic acids, farmers and plant breeders have been performing a rudimental yet effective form of gene transferring and selection, a process called genetic ‘‘manipulation’’ or ‘‘modification.’’ Breeding animals is an example, but most commonly crossing and saving seeds of the strongest or most fruitful plants enabled to slowly maintain or improve the most desired characteristics for farming. As all crop and domesticated animal species have undergone human selection since the dawn *Corresponding author: E-mail:
[email protected] BIOTECHNOLOGY ANNUAL REVIEW VOLUME 10 ISSN: 1387-2656 DOI: 10.1016/S1387-2656(04)10004-5
ß 2004 ELSEVIER B.V. ALL RIGHTS RESERVED
86 of time, they can all be considered ‘‘genetically modified’’ with respect to the wild types, even if old farmers did not know they were selecting genes. Today, humans have knowledge of the underlying mechanisms and, above all, can manipulate genetic information. The general principle is that the genetic code is universal as also the mechanisms involved in transcription and translation. Among the several advantageous traits, that drove natural or genetically engineered (GE) selection, there was an improved growth and yield, an enhanced parasite or herbicide resistance, a better nutritional value or quality. Employing novel and more powerful instruments and knowledge, modern agronomists and zootechnologists are performing the same old process, somehow, trying to reach the same old goals. The ability to modify a specific organism at a molecular level represents the newest instrument of the contemporary form of genetic selection. New promising perspectives arise, but also new questions and concerns. The recombinant DNA revolution discloses key developments for the societies of the ‘‘biotech’’ Millennium. By means of the recombinant DNA (rec-DNA) technology, it is now possible to remove or introduce a new trait, in a plant or animal, in a much faster and more precise fashion than ever. Furthermore, novel biomolecular strategies allow scientists to selectively control, or prevent the transgene escape from genetically modified plants (Table 1) [1,2]. About three decades ago, the recombinant DNA technology broke into the scientific scene with its impressive impact. In 1973, Stanley Cohen and
Table 1. List of useful websites. Each listed website address is intended to begin with ‘‘http://www.’’. FDA U.S. Food and Drug Administration fda.gov/ cfsan.fda.gov/ lrd/biotechm.html cfsan.fda.gov/ lrd/biotechm.html#inf cfsan.fda.gov/ lrd/biotechm.html#prod cfsan.fda.gov/ lrd/biotechm.html#label cfsan.fda.gov/ lrd/biotechm.html#reg fda.gov/cvm/biotechnology/bioengineered.html
FDA Home Page Biotechnology Information for Consumers Completed Consultations on GM foods Food labelling Regulations and Guidance on Safety Assessments Biotechnology in Animals and Feeds
EPA U.S. Environmental Protection Agency epa.gov/ epa.gov/ebtpages/pesticides.html epa.gov/pesticides/biopesticides epa.gov/ebtpages/pestpesticherbicides.html epa.gov/epahome/lawregs.htm epa.gov/ebtpages/treatechnobiotechnology.html epa.gov/ebtpages/humafoodsafety.html
EPA Home Page Pesticides Biopesticides Herbicides Laws and Regulations Biotechnology Food Safety
(Continued)
87 Table 1. (Continued) USDA U.S. Department of Agriculture usda.gov/ usda.gov/agencies/biotech/index.html
USDA Home Page Agricultural biotechnlogy
APHIS – USDA animal and Plant Health Inspection Service (USDA) aphis.usda.gov/ aphis.usda.gov/programs/programs.html aphis.usda.gov/programs/biotechregsvcs.html
APHIS Home Page APHIS Programs Biotechnology Regulatory Services
GIPSA – USDA Grain Inspection, Packers and Stockyards Administration (USDA) usda.gov/gipsa/biotech/biotech.htm GIPSA Home Page WHO World Health Organization who.int/en/ who.int/health_topics/food_safety/en/ FAO Food and Agriculture Organization of the United Nations fao.org/ fao.org/ag/ fao.org/ag/guides/subject/b.htm fao.org/biodiversity/index.asp?lang ¼ en fao.org/biotech/index.asp?lang ¼ en fao.org/ethics/index_en.htm fao.org/ag/AGA/AGAP/FRG/Feedsafety/ feedsafety.htm http://apps.fao.org/ EU European Union europa.eu.int/index_en.htm europa.eu.int/pol/food/index_en.htm europa.eu.int/comm/biotechnology/ introduction_en.htm europa.eu.int/comm/index_en.htm europa.eu.int/pol/agr/index_en.htm europa.eu.int/pol/food/index_en.htm europa.eu.int/comm/food/food/ biotechnology/index_en.htm europa.eu.int/comm/food/food/foodlaw/ principles/index_en.htm europa.eu.int/comm/environment/index_en.htm europa.eu.int/comm/food/food/biotechnology/ novelfood/index_en.htm europa.eu.int/comm/food/plant/ gmplants/index_en.htm europa.eu.int/comm/food/food/biotechnology/ strategy/index_en.htm europa.eu.int/comm/food/food/biotechnology/ gmfood/legisl_en.htm europa.eu.int/comm/food/food/biotechnology/ authorisation/list_author_gmo_en.pdf europa.eu.int/comm/food/food/ labellingnutrition/foodlabelling/index_en.htm
(Continued)
WHO Home Page Food safety
FAO Home Page Agriculture Biotechnology Biological Diversity in Food and Agriculture Biotechnology in Food and Agriculture Ethics in Food and Agriculture FAO Feed and Food Safety Gateway FAO Statistical Databases
EU Home Page Food Safety Biotechnology European Commission European Commission European Commission European Commission feed safety European Commission General Food Law European Commission European Commission
– Agriculture – Food Safety – Food and – Principles of – Environment – Novel Food
European Commission – GM plants and seeds European Commission – Strategy for Europe on Life Sciences and Biotechnology European Commission – Legislation of GM Food & Feed Genetically modified (GM) foods authorised in the European Union European Commission – Food Labelling
88 Table 1. (Continued) Colorado State Univesity colostate.edu/programs/lifesciences/ TransgenicCrops/index.html colostate.edu/programs/lifesciences/ TransgenicCrops/terminator.html colostate.edu/programs/lifesciences/ TransgenicCrops/hotlabel.html colostate.edu/programs/lifesciences/ TransgenicCrops/hotrice.html colostate.edu/programs/lifesciences/ TransgenicCrops/current.html colostate.edu/programs/lifesciences/ TransgenicCrops/future.html colostate.edu/programs/lifesciences/ TransgenicCrops/risks.html colostate.edu/programs/lifesciences/ TransgenicCrops/defunct.html
Transgenic crops Terminator Technology Food Labelling Golden Rice Transgenic Crops Currently on the Market Future Transgenic Products Risks and Concerns Discontinued Transgenic Products
ISAA International Service for the Acquisition of Agri-biotech Applications isaaa.org/ ISAA Home Page isaaa.org/Publications/pubs.htm ISAAA Publications isaaa.org/Publications/Downloads/ James, C. 2000. Global Status of Briefs%2021.pdf Commercialized Transgenic Crops isaaa.org/Publications/Downloads/ James, C. 2001. Global Status of Briefs%2024.pdf Commercialized Transgenic Crops isaaa.org/Publications/Downloads/ Brookes G and Barfoot P. GM Rice: Briefs%2028.pdf Will This Lead the Way for Global Acceptance of GM Crop Technology? Union of Concerned Scientists ucsusa.org/ Union of Concerned Scientists Home Page ucsusa.org/food_and_environment/ Biotechnology biotechnology ucsusa.org/food_and_environment/ Engineered foods allowed on the market biotechnology/page.cfm?pageID ¼ 337 GMO-Watch Internet site of the Biosafety Assessment, Technology and Sustainability (BATS) Institute gmo-watch.org GMO-Watch Home Page gmo-watch.org/GVO-report140703.pdf Bruderer S and Leitner KE. Modified (GM) Crops: molecular and regulatory details. PBS pbs.org/wgbh/harvest/
University of Sussex http://www.biols.susx.ac.uk/Home/ Neil_Crickmore/Bt/index.html boils.susx.ac.uk/home/Neil_Crickmore/ Bt/toxins2.html (Continued)
‘‘Harvest of fear’’ – Exploring the intensifying debate over genetically-modified (gm) food crops. Bacillus thuringiensis toxin nomenclature Bacillus thuringiensis delta-endotoxin list
89 Table 1. (Continued) FURTHER DOCUMENTS AND LINKS General topics ncbi.nlm.nih.gov ncgr.org cast-science.org ific.org croplifeamerica.org usinfo.state.gov/gi/global_issues/ biotechnology.html biome.ac.uk betterfoods.org fb.org bioigene.it/biotech iss.it Commercial websites agbios.com agbios.com/dbase.php aventis.com monsanto.com mycogen.com Bayer.com bejo.com/ pioneer.com Syngenta.com seminis.com dupont.com
National Center for Biotechnology Information National Center for Genome Resources Council for Agricultural Science and Technology International Food Information Council CropLife America U.S. State International Information Programs BIOME The Alliance for Better Foods American Farm Bureau IUSM Biotechnology Home Page Istituto Superiore di Sanita` AgBios Home Page AgBios – GM Crop Database Aventis Home Page Monsanto Home Page Mycogen Home Page Bayer Bejo Zaden Pioneer Syngenta Seminis Vegetable Seed DuPont
Herbert Boyer, developed the technique of DNA cloning, which allowed genes to be manipulated and transferred between different biological species [3,4]. Their discovery marked the birth of genetic engineering together with the discovery by Temin and Baltimore of the reverse transcriptase [5,6]. Almost 10 years later, in 1982, the first transgenic mice were obtained by microinjection, into fertilized mouse eggs, of a DNA fragment containing the promoter of the mouse metallothionein-I gene fused to the structural gene of rat growth hormone [7]. Successively, in 1988 the first tobacco plant was successfully transfected. Thereafter, it only took about two years for genetically engineered crops to enter food production, in the early 1990s and only another decade for the total area, cultivated with GM crops in the world, to reach 44.2 million hectares in 2000 [8]. Between the 1970s and the early 1980s, molecular biology and genetic engineering have been continuously, rapidly and effectively contributing to biology, medicine and biotechnologies. This evolution represents an example of the rapid reduction in the gap between basic research advancements and know-how applications, almost reciprocally overlapping in the biotech fields, showing potentialities, concerns and promises for the next millennium.
90 Recombinant DNA technology enabled scientists to change an organism’s genetic endowment by direct manipulation of DNA sequences. This procedure, also known as genetic engineering, involves (i) elimination (e.g., knock out animals) or (ii) introduction of specific foreign genes, even belonging to unrelated species. In particular, the latter strategy (ii) produces a ‘‘transgenic’’ organism, in which a foreign DNA (a transgene) is incorporated into the genome during an early stage of development. The transgene is present in both somatic and germ cells, is expressed in one or more tissues and is inherited by the offspring in a Mendelian fashion. A gene, to be transcribed by the cell, requires regulatory sequences: a promoter and a terminator. These genetic elements determine the activity of a gene and the time and modalities of its expression. The product of a specific coding sequence can be modulated or ‘‘switched on’’ or ‘‘of’’ by the presence or absence of such an element. Therefore, a broader definition of ‘‘transgenesis’’ includes the introduction of foreign regulatory sequences in the hosting organisms, and not only the specific coding sequences. Transgenic plants Creation of the first transgenic plants dates back to the early 1980s, when four groups working independently at Washington University in St. Louis, Missouri, the Rijksuniversiteit in Ghent, Belgium, Monsanto Company in St. Louis, Missouri, and the University of Wisconsin successfully inserted foreign genes in plant cells. Their scientific achievements were then published in three different journals, in 1983 [9–12]. Today, transgenic plants are currently produced by introduction of genes conferring several properties such as: resistance to insects, viruses or herbicides, improved nutritional value and flavor, resistance to environmental stresses (such as drought, salinity, pollution, extreme temperatures), capacity to produce heterologous substances with pharmacological properties, prolonged organoleptic stability, extended conservation and improved value of flowers (floriculture). Noteworthy, not all of the above traits are already present in commercially available plants or employed in edible crops. Indeed, some of them are just successful applications of recombinant DNA technology, may have an exclusive scientific value or be yet only at an experimental phase. Furthermore, several transgenic food products that received approval for marketing have been discontinued for a variety of reasons, even after being available on the market for years. A few examples of such abandoned products can be found on the University of Colorado website (Table 1). Examples of discontinued foods include transgenic tomatoes that soften more slowly than conventional, a tomato paste made of another line of transgenic tomatoes with the same trait, some insectresistant potato lines, a herbicide-tolerant flax, some insect-resistant corn lines. Interestingly, beside various clear failures based on either medical or environmental issues, other products that showed satisfactory results during trials,
91 Identification of the desired trait
Identification of the source of the gene (donor organism)
Isolation of the gene from that source
Adjustment of the gene to confer the desired trait
Transfection of the plant
Test for the presence of the desired trait
Field trials, to make sure that :
Initiation of product safety trials
1) There are no detrimental effects of the gene 2) The gene works the way it was conceived
Transmission to regulatory agencies as required Fig. 1. Development of a transgenic plant.
were discontinued only due to the reluctance of buyers and/or the adverse public opinion about GM-food. Transgenic plants, currently available on the market, include corn, tomato, potato, rape soybeans, maize, canola, potatoes and papayas (Table 1) [8,13]. A new transgenic crop can be developed through a complex serial procedure (Fig. 1). As shown in Fig. 2, once a specific gene of interest has been chosen and isolated, the transfection may be carried out by several different strategies: (1)
Agrobacterium tumefaciens. This method involves the use of a plant parasite, A. tumefaciens that is a well-known bacterium, causing large tumors in some dicotyledons. Its infectious capacity is associated with the
92 Isolation of the desired gene a) A. tumefaciens method
b) Gene gun method
c) Other strategies Gene inserted intoTi plasmid Particles coated with DNA and transformation of A. tumefaciens
Cells shot with Gene gun and Bacterium mixed with plant cells. DNA incorporated into Ti plasmid moves into cells and inserts plant chromosome DNA into plant chromosome Selection and screening of transformed cells. Regeneration of the plant from a single transformed cell Fig. 2. Alternative methods of plant transformation.
(2)
(3)
presence of a plasmid called Ti (Tumor inducing), which is eventually transferred to the infected plant cell, and that carries the tumor-associated genes. A coding sequence of the plasmid is then integrated into the host chromosome and hence inherited by all the cells [14]. This phenomenon results in permanent cell transformation and unregulated massive growth (tumor). By substituting the naturally integrating region of the Ti plasmid with the transgene of interest, it is possible to insert foreign genes in plant cells infected by A. tumefaciens and to obtain a transgenic organism. A disadvantage of this system is that the bacterium does not infect all plant species. Gene gun. The gene gun method circumvents the host-restriction limitation, typical of the A. tumefaciens method. Tiny gold or tungsten particles are preventively coated with the DNA fragment to be inserted, and successively shot into the cell [15]. Once DNA enters the cell, it becomes free to integrate into the host chromosome. When the integration occurs, permanent transfection of the host plant cell is achieved. Further approaches, commonly used for plant transformation, include infiltration, electroporation of cells and tissues, direct protoplast transformation, electrophoresis of embryos, microinjection, pollen-tube pathway, silicon carbide- and liposome-mediated transformation [15]. In particular, electroporation was first employed in 1985, to transfect tobacco and maize protoplasts [16], whereas transformation of tobacco protoplasts by direct DNA microinjection was carried out about a year later [17].
93 The nucleic acid fragment that is used to transfect the host plant cell is formed by one or more units, each containing three genetic elements: the promoter, the transgene and the terminator [13]. The whole unit is called ‘‘gene cassette’’ and the promoter and the terminator represent its regulatory sequences. The promoter of the transgene is one of the most critical choices to make, when a genetically modified organism is to be constructed. It locates at the 50 end of the gene cassette and its sequence can affect the expression level of the transgene, the histological fate of the product and the time of the synthesis. Most of the manipulated crops, approved and commercialized today, utilize the 35s constitutive promoter (P-35s) of the Cauliflower Mosaic Virus (CaMV), although its sequences, available from several sources (e.g., patents, gene bank or petitions), show some differences when compared. Table 2 and Fig. 3 include a list of promoters used in transgenic crops and their occurrences.
Table 2. The frequency of occurrence of introduced promoters into approved GM crops. Used promoters
Donor organisms (origin)
An anther specific promoter Bacterial dP-35s E-OCS nda P-35s P-4AS1 P-5126del P-ALS P-Als P-CDPK P-E35s P-E8 P-FMV P-HelSsu P-Kti3 P-mac P-mas P-napin P-nos and 2 P-nos P-OCS,35s P-PCA55 P-PEPC P-Ptac P-ract P-Ssu P-TA29
/ / Cauliflower Mosaic Virus Agrobacterium tumefaciens / Cauliflower Mosaic Virus Cauliflower Mosaic Virus Zea mays Nicotiana tabacum Arabidopsis thaliana Zea mays Cauliflower Mosaic Virus Lycoperiscon esculentum (tomato) Figworth Mosaic Virus Heliantus annus Glycine max (soybean) A. tumefaciens and Cauliflower Mosaic Virus Agrobacterium tumefaciens Brassica rapa Agrobacterium tumefaciens Cauliflower Mosaic Virus and A. tumefaciens Zea mays Zea mays Bacterial Oryza sativa (rice) Arabidopsis thaliana Nicotiana tabacum
(Continued)
Number of occurences of each promoter 2 22 1 1 1 42 1 1 1 1 1 12 1 8 1 1 1 1 1 10 1 1 1 1 2 9 6
94 Table 2. (Continued) Used promoters
Donor organisms (origin)
P-ubiZM1(2) P-b-Conglycinin
Zea mays Glycine max (soybean)
Number of occurences of each promoter 1 1
The donor organisms of promoters are indicated. Some promoters may be present in more than one copy in a single product, since a regulatory sequence may have been used for more than one transgene and since several copies of a transgene may be present in the same product. This frequency of appearance is not taken into account in the table. dP-35s: double 35s promoter, promoter region from Cauliflower Mosaic Virus. The double (d) represents a duplicated region in the promoter. E-OCS: octopine synthase enhancer from A. tumefaciens Ti plasmid, pTiACH5. nda: No Data Available. P-35s: 35s Cauliflower Mosaic Virus promoter. P-4AS1: promoter containing four tandem copies of AS1 (activating sequence 1) and a single portion of 35s Cauliflower Mosaic Virus promoter synthetic polylinker sequence. P-5126del: a modified Z. mays anther specific promoter. P-ALS: tobacco ALS1 promoter. P-PCDK: promoter derived from a corn calcium dependent protein kinase (CDPK) gene. P-E35s: 35s promoter from the Cauliflower Mosaic Virus with the duplicated enhancer region. P-E8: ethylene responsive gene promoter. P-FMV: a promoter derived from Figworth Mosaic Virus (FMV). P-HelSsu: RuBisCo SSU (ribulose-1,5-bisphosphate carboxylase small subunits 1A) promoter, from Helianthus annuus. P-Kti3: Kunitz trypsin inhibitor 3 (Kti 3) promoter. P-mac: P-mas and P-35s hybrid. P-mas: promoter region of mannopine synthase gene of pTiA6. P-napin: the promoter of the nopamin gene from Brassica rapa which functions in developing seeds. P-nos: promoter region of the nopaline synthase gene. P-ocs: promoter of the octopine synthase gene. P-PCA55: the promoter region of the anther specific gene CA55 from Zea mays. P-PEPC: green tissue-specific phosphoenolpyruvate carboxylase (PEPC) promoter from corn. P-Ptac: bacterial Ptac promoter. P-ract: 50 region of the rice actin 1 gene containing the promoter and first intron. P-Ssu: (also called P-SsuAra) the A. thaliana ribulose-1,5-bisphosphate carboxylase small subunits1A promoter. P-TA29: the promoter region of anther-specific gene TA29 from Nicotiana tabacum. P-ubiZM1(2): the ubiquitin promoter plus ubiquitin intron and a 50 untranslated region from Z. mays. P-b-Conglycinin: seed-specific promoter derived from the a0 -subunit of the Glycine max b-Conglycinin gene (modified from Bruderer S and Leitner KE, 2003).
More than 39 genes have already been used in currently approved GM crops. Among the most common are the bacterial neomycin-phosphotransferase II gene (nptII), the phosphinothrycin acetyl transferase from S. hygroscopicus (BAR) the cry gene and the 5-enolpyruvylshikimate-3-phosphate synthase (EPSPS) gene, which will be described later. Table 3 includes a list of transgenes and the donor organisms, whereas their frequency of occurrence is depicted in the Fig. 4. The third element of the gene cassette is the terminator. It functions as a regulatory sequence controlling the halt of the transcription by RNA polymerase and the poly-adenylation signal. It situates at the 30 end of the transgene. Table 4 and Fig. 5 include the terminators most commonly found in GMOs and their frequency of occurrence. Improved resistance to parasites The cry gene encodes a bacterial insecticidal crystal protein (Cry protein or ICP) first described in the gram-positive soil bacterium Bacillus thuringiensis (Bt) [18,19]. To be toxic, the Cry protein must be ingested by the insect larva.
95 Bacterial (16.5%)
nda (0.8%) P-FMV (6.0%)
P-35s (41.4%)
P-nos and 2xP-nos (7.5%) P-Ssu (6.8%) P-TA29 (4.5%) Others (16.5%) Fig. 3. Frequency of occurrence of the most often used promoters in the currently approved genetically engineered crop plants. P-35s: 35s Cauliflower Mosaic Virus promoter. P-35s includes P-35s, P-E35s and dP-35s. nda: No Data Available. P-FMV: Figworth Mosaic Virus Promoter. P-nos and 2 P-nos are, respectively, the promoter region of the nopaline synthase gene, from A. tumefaciens, and the tandem duplicate promoter region of the nopaline synthase gene, from A. tumefaciens. P-Ssu: A. thaliana ribulose-1,5-bisphosphate carboxylase small subunits 1A promoter. P-TA29: promoter region of anther-specific gene TA29 from Nicotiana tabacum (modified from Bruderer S and Leitner KE, 2003).
The mechanism of action involves solubilization of the crystal in the insect midgut, proteolytic processing of the protoxin by midgut proteases, binding of the Cry toxin to midgut receptors, and insertion of the toxin into the apical membrane to create ion channels or pores [20]. Lethality is believed to be due to destruction of the transmembrane potential, with the subsequent osmotic lysis of cells lining the midgut [21]. Each B. thuringiensis strain synthesizes up to five different Cry proteins. Besides them, cytolysins (Cyt toxins), further toxins that act by a different mechanism, are also found within the crystal. Both these two classes of toxins are referred as delta-endotoxins (d-endotoxins). Other than d-endotoxins, Bt produces also various further virulence factors, including secreted insecticidal protein toxins, alfa-exotoxins, beta-exotoxins, hemolysins, enterotoxins, chitinases and phospholipases. The Cry proteins have different specificity and their combination within a given strain, defines the activity spectrum of that strain [22]. The ecology of Bt is not fully understood yet. Discording hypothesis have been postulated about the evolutionary advantage for the bacterium, associated with the production of the toxin, but according to Aronson et al., the bacterium is likely to have a subtle symbiotic interaction, perhaps with plants, to account for the extensive production of the highly specific and efficacious toxins [21]. Several subfamilies of the cry gene have been discovered, named and classified and in 1993, a B. thuringiensis d-endotoxin nomenclature committee was created
96 Table 3. Frequency of occurrence of introduced genes in approved GM crop plants with the corresponding donor organisms. Multiple insertions of a gene into a genome were counted as one event. Introduced genes
Donor organisms (origin)
aad accd AccS ALS bar barnase barstar Bay TE bla Chimeric S4-HrA CMV cp CMV/PRV cp
E. coli Pseudomonas chlororaphis Lycoperiscon esculentum (tomato) Arabidopsis thaliana Streptomyces hygroscopicus Bacillus amyloliquefaciens Bacillus amyloliquefaciens Umbrellaria californica (California bay) E. coli Nicotiana tabacum Cucumber Mosaic Virus strain C Papaya Ringspot Virus and Cucumber Mosaic Virus Watermelon Mosaic Virus 2 strain FL and Cucumber Mosaic Virus Zucchini Yellow Mosaic Virus strain FL and Cucumber Mosaic Virus Agrobacterium tumefaciens sp. strain CP4 B. thuringiensis subsp. Kurstaki B. thuringiensis subsp. Kurstaki HD-73 B. thuringiensis var. aizawai B. thuringiensis subsp. kurstaki B. thuringiensis subsp. Tenebrionis B. thuringiensis subsp. kumamotoensis B. thuringiensis subsp. Tolworthi E. coli Corynebacterium E. coli Glycine max (soybean) Achromobacter sp. Strain LBAA E. coli Zea mays Klebsiella ozaenae Agrobacterium tumefaciens E. coli Streptomyces viridochromogenes Lycoperiscon esculentum (tomato) Potato Potato Leaf Roll Virus (PLRV) Potato Virus Y (PVY) strain O E. coli bacteriophage T3 E. coli
CMV/WMV2 cp CMV/ZYMV cp CP4EPSPS cry1Ab cry1Ac cry1F cry2Ab cry3A cry3Bb1 cry9C dam dapA gentR GmFAD2-1 gox GUS mEPSPS nitrilase nos nptII pat PG pinII PLRVrep PVYcp sam-K tetR
Number of occurrences of each gene 7 1 1 1 14 8 6 1 6 ( þ 7 part.*) 1 1 1 2 2 12 6 5 1 1 6 1 1 1 1 1 1 7 5 1 5 1 28 ( þ 1 part.*) 11 2 1 2 1 1 1
(*) denotes the number of GM crops containing only partial copies of the corresponding genes. It should be noted that plants containing only partial genes were not counted towards the total. aad (from E. coli): 300 (9)-O-aminoglycoside adenylyltransferase. accd: 1-amino-cyclopropane-1-carboxylic acid deaminase, an
97 [23,24]. A full and up-to-date list of the known d-endotoxins, along with the respective NCBI (National Center for Biotechnology Information) accession number, authors, publication year, source strain and further comments is available online on the University of Sussex website (Table 1) [25]. Numerous subspecies of Bt have been described so far. Together, all these subspecies can kill a large variety of host insects and even nematodes, but each strain does so with a high degree of specificity [26]. In a sprayed form, Bt and its purified toxins, have been used around the world for about 40 years as highly selective and inexpensive insecticides, for their recognized safety to humans, animals and environment [20]. However, despite its advantages, when exposed to physical factors (e.g., UV light), the toxins rapidly degrade into nontoxic/ environmental friendly compounds, with no further efficacy on insect larvae. To circumvent this rapid inactivation, frequent applications of suspensions of spores and inclusions are required to maintain a constant and effective level of pesticide in the field [21]. A novel approach to such limitation is represented by the application of biotechnology. Transgenic plants, modified to express Bt toxins, also known as Bt-protected plants, were first created during the early 1990s by cloning the cry genes into different crops varieties [27]. Today, the most frequently found genes are cry1Ab and cry3A, present in about 29% of the insect-protected plants
essential precursor for the biosynthesis of the plant hormone ethylene. AccS: 1-amino-cyclopropane-1carboxylic acid synthase, an essential precursor for the biosynthesis of the plant hormone ethylene. bar (from S. hygroscopicus): phosphinothricin acetyl transferase. barnase: ribonuclease enzyme (RNAse). barstar: the coding region of the barstar gene from B. amyloliquefaciens. The barstar gene encodes for a ribonuclease inhibitor (barstar enzyme Bay TE: the 12:0 acyl carrier protein (ACP) thioesterase gene which codes for an enzyme in the fatty acid biosynthetic pathway found in developing seeds. bla: beta-lactamase. Chimeric S4-HrA encodes an acetolactate synthase (ALS) enzyme from Nicotiana tabacum. CMV cp: Cucumber Mosaic Virus coat protein gene. CMV/PRV cp: coat protein gene of Papaya Ringspot Virus (PRV) HA 5-1. CMV/WMV2 cp: coding region of the WMV2 cp gene fused to the 48 nucleotides from the 50 terminus of the CMV cp gene. CMV/ZYMV cp: ZYMV cp coding region fused to the CMV translation initiation codon. CP4EPSPS: 5-enolpyruvylshikimate-3-phosphate synthase, isolated from Agrobacterium sp. (strain CP4). cry1Ab, cry1Ac, cry3A, cry9C, cry1F, cry3Bb1, cry2Ab cry: insecticidal bacterial crystal protein (Cry protein) found in the bacterium Bacillus thuringensis. Dam: DNA adenine methylase from E. coli. DapA: enzyme dihydrodipicolinic acid synthase. GentR: gentamycin resistance gene. GmF AD2-1: delta-12 desaturase. gox: glyphosate oxidase (GOX) from the bacterium Ochrobactrum anthropi. GUS: beta glucuronidase. mEPSPS: a modified form of wild type 5-enolpyruvyl-3-phosphoshikimate synthase gene from Zea mays which encodes an insensitive enzyme to inactivation by glyphosate nitrilase (from K. pneumoniae subspecies ozaenae): nitrilase. nos: nopaline synthase. nptII: neomycin phosphotransferase aminoglycoside (30 ) phosphotransferase type II gene, from E. coli transposon Tn5 (or Kanamycin resistance gene). pat: phosphinothricin acetyltransferase from Streptomyces viridochromogenes. PG: polygalacturonase. PinII: potato a potato DNA containing 18 bp untranslated leader, pinII protein coding region with intron and about 920 bp of 30 sequence (30 untranslated region of the RNA and putative transcription termination region), which encodes for a protease inhibitor. PLRVrep: full-length ORF1 and ORF2 from Potato Leaf Roll Virus (PLRV), which encode a fusion protein having both helicase and RNA-dependent RNA polymerase activity. PVYcp: coat protein from potato virus Y strain 0. sam-K: modified Sadenosylmethionine hydrolase gene derived from E. coli bacteriophage T3 that encodes an enzyme, Sadenosylmethionine hydrolase (SAMase). Tet-R: tetracycline resistance gene (modified from Bruderer S and Leitner KE, 2003).
98 cry family (13.5%)
aad (4.5%) bar (9.0%)
barnase (5.2%) bla (3.9%) Others (23.2%)
CP4EPSPS (7.7%)
gox (4.5%) nitrilase (3.2%) pat (7.1%) nptII (18.1%) Fig. 4. Frequency of occurrence of the most often used genes in the currently approved genetically engineered crop plants. aad (from E. coli): 300 (9)-O-aminoglycoside adenylyltransferase. bar (from S. hygroscopicus): phosphinothricin acetyl transferase. barnase: ribonuclease enzyme (RNAse). bla: beta-lactamase. CP4EPSPS: 5-enolpyruvylshikimate-3-phosphate synthase, isolated from Agrobacterium sp. (strain CP4). gox: glyphosate oxidase (GOX) from the bacterium Ochrobactrum anthropi. nitrilase (from K. pneumoniae subsp. ozaenae): nitrilase. nptII: neomycin phosphotransferase aminoglycoside (30 ) phosphotransferase type II gene, from E. coli transposon Tn5 (or Kanamycin resistance gene). pat: phosphinothricin acetyltransferase from Streptomyces viridochromogenes. cry: insecticidal bacterial crystal protein (Cry protein) found in the bacterium Bacillus thuringiensis. The cry gene family was grouped as a whole and includes: cry1Ab, cry1Ac, cry3A, cry9C, cry1F, cry3Bb1, cry2Ab (modified from Bruderer S and Leitner KE, 2003).
carrying cry genes (Table 3). These latter transgenes are mostly modified or truncated forms of the native sequence [13]. Many synthetic variants have been obtained to modulate the expression of the endotoxin and to modify its features in general. Physical inclusion of the toxin within the transgenic plant cell: (i) provides a protected environment, that prolongs the activity of the insecticide by delaying its degradation, (ii) reduces costs and needs of higher dosages, and (iii) enables the toxin to reach even insects presents within the stalk, and not only on the plant surface. Moreover, by cloning the cry transgene under control of different promoters it may be possible to selectively express the Bt toxin in certain tissues of the plant or in specific time lags. Thus, Bt-protected plants, by providing highly effective control of major insect pests, such as the European corn borer, southwestern corn borer, tobacco budworm, cotton bollworm, pink bollworm, and Colorado potato beetle, ensure better yields, lower costs and reduced reliance on conventional chemical pesticides with wider insecticidal spectra and lower specificity [28]. Especially this
99 Table 4. A lists of terminators, the organism from which they originated, and how often they are found in current GM crops. Used terminators
Donor organisms (origin)
Bacterial nda T-35s T-7s T-ALS T-Als T-E9 T-g7 T-Kti3 T-mas T-napin T-nos T-ocs T-ORF25 T-phaseolin T-pinII T-SSU T-tahsp 17 T-tml T-Tr7
/ / Cauliflower Mosaic Virus Glycine max (soybean) Nicotiana tabacum Arabidopsis thaliana Pea Agrobacterium tumefaciens Glycine max (soybean) Agrobacterium tumefaciens Brassica rapa Agrobacterium tumefaciens Agrobacterium tumefaciens Agrobacterium tumefaciens Phaseolus vulgaris (green bean) Selanum tuberosum Glycine max (soybean) Triticum aestivum (Wheat) Agrobacterium tumefaciens Agrobacterium tumefaciens
Number of occurrences of each terminator 22 3 17 2 1 1 12 3 1 2 1 35 5 1 1 2 1 1 4 2
Some terminators may be present in more than one copy in a single product, since a regulatory sequence may have been used for more than one transgene and several copies of a transgene may be present in the same product. This frequency of appearance is not taken into account in the table. T-7s: the 30 untranslated region of the soybean alpha subunit of the beta-Conglycinin gene. T-ALS: ALS: tobacco ALS1 terminator. T-Als: Arabidopsis thaliana ALS1 terminator. T-E9: 30 untranslated region of the pea ribulose-1,5-bisphosphate carboxylase small subunit E9 gene. T-g7: 30 untranslated end of the TL-DNA gene 7. T-Kti3: Kunitz trypsin inhibitor 3 (Kti 3) terminator. T-mas: polyadenylation region from mannopine synthase gene of pTiA6. T-napin: napin gene terminator. T-nos: 30 untranslated region of the nopaline synthase gene. T-ocs: terminator of the octopine synthase gene. T-ORF25: terminator from A. tumefaciens. T-phaseolin: 30 fragment of the phaseolin gene of green bean. T-PinII: terminator sequence from Solanum tuberosum proteinase inhibitor II gene. T-SSU: the 30 untranslated region from the G. max ribulose-1,5-bisphosphate carboxylase small subunit gene. T-tahsp 17: 30 untranslated region of the coding sequence for the heat shock protein 17.3. T-tml: polyadenylation region of tml gene from pTiA6. nda: no data available. T-35s: 30 nontranslated region of the Cauliflower Mosaic Virus 35s gene. T-Tr7: the 30 region from A. tumefaciens t-DNA transcript 7 (modified from Bruderer S and Leitner KE, 2003).
latter advantage is considered highly desirable for what concerns public, animal and environmental safety. A few examples of the specificity of some used recombinant Cry proteins are listed in Table 5. Cry protein expression, in transgenic bacteria that naturally colonize tomato plants, has been reported as well. This approach involves the use of GM microorganisms with insecticidal activity, capable of surviving on leaf surfaces for several weeks. Also such a novel strategy should allow for a reduction in pesticide application [29].
100 others (16.2%)
Bacterial (18.8%)
T-tml (3.4%) nda (2.6%) T-ocs (4.3%)
T-35s (14.5%)
T-nos (29.9%)
T-E9 (10.3%)
Fig. 5. The frequency of occurrence of the most often used terminators introduced into the currently approved genetically engineered crop plants. T-nos: 30 untranslated region of the nopaline synthase gene. T-ocs: terminator of the octopine synthase gene. T-tml: polyadenylation region of tml gene from pTiA6. nda: no data available. T-35s: 30 untranslated region of the Cauliflower Mosaic Virus 35s gene. T-E9: 30 untranslated region of the pea ribulose-1,5-bisphosphate carboxylase small subunit E9 gene (modified from Bruderer S and Leitner KE, 2003).
Table 5. Specificity of some recombinant, insecticidal d-endotoxins, currently employed in transgenic plants. Susceptible insect orders (and species in same cases) are indicated. d-endotoxin
Toxic to these insect orders/species
Cry1Aa Cry1Ab Cry1Ac
Lepidoptera Lepidoptera Lepidoptera (cotton bollworm, tobacco budworm and pink bollworm) Lepidoptera Lepidoptera, Diptera Coleoptera (Colorado potato beetle, elm leaf beetle and yellow mealworm) Coleoptera (corn rootworm species) Diptera Lepidoptera, Coleoptera
Cry1B, Cry1C, Cry1D, Cry1F Cry2 Cry3A Cry3Ab, Cry3Bb1 Cry4 Cry5
Examples of Bt insect-resistant crops (approved but not necessarily currently commercialized) include corn (primarily for control of European corn borer, but also corn earworm and Southwestern corn borer), cotton (for control of tobacco budworm and cotton bollworm), potato (for control of Colorado potato beetle), tomato (for control of lepidopteran pests including, but not limited to, cotton bollworm, pink bollworm, tobacco budworm) (Table 1) [8,13]. Despite their approval, extensive use of Bt-protected plants is raising serious concerns among scientists, politicians and the public. There are worries regarding the possible spread of Bt-resistant insects. This eventuality may present
101 a threat to the environment as well as for the durability of this novel insect control technology. As in the presence of any constant selective pressure, the possible spread of individuals carrying mutant genes, conferring advantages against the challenging factor, is favored. Since the mid-1980s many cases of resistance to B. thuringiensis, mostly induced experimentally under laboratory conditions, have been discovered. Insects have also demonstrated their enormous genetic plasticity with hundreds of species found resistant to various insecticides [30]. To prevent the onset of undesired Bt-resistant pests, strict directives were given by regulatory agencies, such as the American Environmental Protection Agency (EPA) (Table 1). One of the adopted strategies is called ‘‘Hi-dose/ Refuge’’ [20]. Such an approach involves the coexistence of both transgenic fields expressing high levels of crystal protein (i.e., 25 times the protein concentration necessary to kill susceptible larvae or more) and conventional crop fields. The latter are called structured refugia and consist of restricted areas, devoid of GM plants that represent the selective pressure. The principle is to express Cry toxins at such a high dose that nearly all heterozygotic carriers of resistance alleles will be killed. Assuming an extensive random mating between rare resistant individuals and the numerous sensitive insects, harbored in the nearby refuge, a population of homozygous resistant insects would be unlikely to emerge. Resistance to B. thuringiensis toxins seems to be inherited in a recessive fashion at least in some species [31,32]. Studies on the feeding behavior of bollworm and tobacco budworm larvae, in mixed stands of traditional and transgenic insect-resistant cotton suggest that larvae of both species frequently moved among plants, feeding indiscriminately on BTK and non-BTK plants [33]. A further ardent debate, associated with the use of insect protected plants was sparked in 1999, by a short paper published by the entomologist John Losey and colleagues [34]. This study pointed out unpredicted toxic properties of the pollen from Bt corn, on larvae of the monarch butterfly, Danaus plexippus. The caterpillars of this species feed exclusively on milkweed leaves and, thus, are not a target pest of Bt corn. Nonetheless, when insects were reared on milkweed leaves, dusted with pollen from Bt corn, higher mortality, with respect to controls, could be observed. Based on these findings, several controversial studies have been conducted [35]. According to the results of, at least, three of them, the commercial large-scale cultivation of current Bt-maize hybrids did not seem to indicate a significant risk for the monarch population [36–38]. Herbicide tolerance Besides insects, another serious problem in agriculture is the weed infestation of the cultivation. ‘‘Weeds’’ is a generic word to describe wild plants growing where they are not wanted, as in gardens or crops fields. These highly aggressive and fast-growing plants represent a plague for the cultivation for their (i) ability
102 to compete for water, light and nutrients, (ii) possible contamination of the crop seeds with undesired seeds and toxins and (iii) interference with the crop harvest. In the worst cases, weeds infestation can kill crops, with a dramatic loss in terms of yield percentage [39]. A specific biochemical pathway involving the 5-enolpyruvylshikimate-3phosphate synthase (EPSPS) gene, to produce aromatic amino acids, has been described in plants, fungi and bacteria. This enzyme is the target of the glyphosate (N-phosphonomethyl glycine), one of the most popular herbicides whose reaction kinetics was described in 1983 [40–42]. In 1985, Stalker et al. showed that the glyphosate-resistant phenotype was associated with a mutation, resulting in a Pro to Ser amino acid substitution at the 101st codon of the protein [43]. Bacteria with the mutant form of the gene were viable and able to make aromatic amino acids in presence of glyphosate. A few years later, Klee et al. cloned an Arabidopsis thaliana EPSPS gene and then fused it with a Cauliflower Mosaic Virus 35s Promoter, before reintroducing the recombinant gene into Arabidopsis. The resultant overproduction of EPSPS led to glyphosate tolerance in transformed callus and plants [44]. Thus, unlike such transgenic plants, weeds infesting herbicide tolerant (HT) crops, are selectively susceptible to glyphosate treatments and can be selectively killed. Since Klee’s experiment, employment of approved HT plants has grown remarkably, world-wide. Between 1996 and 2001, herbicide tolerance has constantly been, by far, the most prevalent trait among GM plants (with insect protection being second) and in 2001, about 77% of the global area grown with transgenic crops (52.6 million hectares) was occupied by herbicide tolerant plants (largely soybean) [45]. Examples such transgenic crops (not exclusively for food or feed purposes) include sugar beet, Argentine canola, Polish canola, chicory, carnation, soybean, cotton, flax, linseed, tobacco, rice and maize (Table 1). Regarding the safety aspects associated with the adoption and consumption of this transgene, several studies have been conducted. In a comparative analysis, in 1996, Padgette et al. demonstrated the nutritional equivalence of seeds and selected processing fractions, from glyphosate-tolerant soybean lines and their parental, conventional cultivar [46]. A similar comparative approach was also conducted to evaluate the composition of a glyphosate-tolerant line of corn, with respect to that of conventional corn, grown in the United States in 1998 and in the European Union in 1999. Also this study suggested that the GM corn was compositionally equivalent to, and as safe and nutritious as, conventional corn hybrids [47]. The safety of EPSPS derived from Agrobacterium sp. strain CP4 (CP4 EPSPS) was assessed [48]. CP4 EPSPS is introduced and expressed in glyphosatetolerant soybeans (Table 3). An in vitro digestion model was employed to infer and demonstrate the digestibility of the transgenic product, whereas acute administration of high dosages of CP4 EPSPS, showed no toxic effects of the heterologous protein, in mice. In that paper, potential allergen concerns were excluded as well.
103 Additional studies suggested that glyphosate-tolerant soybeans are as safe as traditional soybeans, with respect to food and feed safety, even after treatment with commercial levels of glyphosate [49–52]. Recently, Chang et al. demonstrated that the EPSPS protein in GM soybean, cloned, expressed and purified from an E. coli strain, showed no significant allergenicity in the Sprague Dawley rats [53]. Golden rice For underdeveloped and low-income countries including highly populated areas extending throughout Asia, Africa and Latin America, reliance on rice, as a primary food staple, contributes to vitamin A deficiency (VAD), a serious illness causing juvenile blindness, pregnancy-related mortality and death in million people per year. Indeed, upon milling, provitamin A content in rice endosperm, which ultimately represents the edible part of the rice grain, is insufficient to meet the required daily allowance of several hundreds micrograms. Golden rice is the name given to a transgenic, nutritionally-enhanced plant, created in 2000 to obtain a functioning provitamin A (beta carotene) biosynthetic pathway in rice endosperm [54]. The transgenic plant accumulates beta carotene in the endosperm and may provide a supplementary dietary source of provitamin A for people feeding mainly on rice [55]. With regard to this, however, medical criticism seems to be mainly focused on the effective adequacy, of the provided dietary supplement, as an effective solution for VAD [56]. Thus, current experimental efforts on Golden rice, are aimed to increase the provitamin A accumulation in the endosperm, by the identification of the metabolic, rate-limiting bottlenecks that caused a limited beta carotene synthesis in the prototype lines. To enhance or improve nutritional properties is another task achievable by development of genetically modified foods. Golden rice represents an immediate example of the complex debate involving not only scientific or technical aspects, but also ethical, social, economical and political decisions related to its commercialization and diffusion. Modified and widely spread species, such as rice or wheat, that have provided important and precious foods in the history of human population survival may represent a successful and promising advancement, but open up reasonable concerns related mostly to the novelty of the GM procedures [57,58]. World population growth rate and globalization processes are changing the traditional view of food and agriculture, pushing toward new solutions for a sustainable development. Transgenic animals and animal technology One of the scientific aims of animal transgenesis in livestock, is to understand the various regulatory mechanisms of genes and the physiological role of the different proteins in vivo. Downstream applications of such acquired knowledge
104 include maintenance, improvement or attainment of profitable or desirable traits in farm animals. After preliminary studies, conducted on embryos of different organisms such as sea urchin, Candida elegans, Xenopus, Drosophila and mice, successive experiments involved swine, rabbits and sheep that were modified in the attempt to obtain animals, with higher levels of circulating hormones [59]. Among the first employed transgenes, were the human growth hormone (hGH) and the bovine growth hormone (bGH), which are normally secreted by the pituitary gland [60]. A mouse metallothionein-I promoter was used to control the synthesis in the first manipulated animals. In these studies, the monitored growth rate was not enhanced in any of the transgenic animals, however definite biological effects were observed, in comparison to the littermate controls. Although the complexity of the field promptly revealed to the scientific world, these pioneer studies showed the enormous potential of animal transgenesis and thoroughly fulfilled the initial ambitious expectations. Today, though in a more developmental phase with respect to GM crops, research in animal technology is a very important and active sector of biotechnology. Advances in transgenic biology, gene therapy, ‘‘knock-out’’ gene technologies, and cloning may lead to other novel products/strategies that enhance productive efficiency. Nevertheless, passionate and ardent debates constantly arise worldwide. For instance, there are common fears of a reduced genetic diversity caused by the intensification of livestock production associated to the adoption of genetically uniform varieties of animals. To avoid this unlikely event, several strategies (such as cryopreservation of semen and embryos, coupled with artificial insemination and embryo transfer, as well as somatic cloning) have been devised. Today, it seems unlikely that in a near future genetically modified livestock may play a major role, in developing countries, as a major food source. Among food animals, only engineered fish (salmon in particular) are presently under active consideration by US regulators. Current research and development in zootechnology, are mainly focused on (a) the quality of livestock feeds, through nutrient content improvement of forages, and the monitoring of food chain safety, (b) the digestibility of low quality feeds and auxological aspects (c) the effective control of several animal diseases, (d) the possible secretion of heterologous substances (e.g., antibodies, vaccines, pharmaceuticals, dietary supplements) in milk of transgenic, dairy animals. Such research lines are likely to represent nearest applications of genetic engineering in animal technology. The recent observation of the Bovine Spongiform Encephalopathy (BSE), showed the need to respect the natural animal physiology and consider adequate preventive actions. The main public health aspects related to zootechnology developments will particularly need to deal not only with the definition of health safety aspects for human beings, food chain, environment, but also with the development of effective screening tools and monitoring procedures for surveillance and traceability.
105 Social, ethical and legal issues The so-called ‘‘consumer sovereignty’’ requires that information be made available, so that people may make food choices based on their own ethical, social, cultural or religious values. Proper policy provisions and clear regulations to inform people about the degree, to which food has been genetically engineered, is highly desirable and appropriate. Regulation of genetic engineering in the US is overseen by three agencies: the Animal and Plant Health Inspection Service (APHIS) of the United States Department of Agriculture (USDA), the United States Food and Drug Administration (FDA), and the aforementioned EPA. The regulatory paradigm, used by the FDA to approve novel foods, is based on the concept of ‘‘substantial equivalence’’ [61,62]. According to this guideline, the allergen, nutrient and toxin content of the new GM food must fall within the normal range of the equivalent, conventional food. The concept of substantial equivalence provides the framework for a comparative approach to identify the similarities and differences, between GMOs and their traditional counterparts that have a known history of safe use. Using this method for the evaluation of several GM crops already approved world-wide, Cockburn concluded that foods and feeds derived from genetically modified crops are as safe and nutritious as those derived from traditional crops [63]. On the other hand, Kuiper et al. argue that the concept is not a safety assessment in itself because it identifies hazards, but does not assess them. Moreover, application of the concept of substantial equivalence encounter several difficulties in its application [64]. The European Union plans to establish a ‘‘farm to fork’’ tracking system that would regulate the traceability for transgenic foods, and several proposals are in progress. It is noteworthy that public health would favor these directives not only for GMOs per se, but also for those foods derived from GMOs (e.g., oils or processed materials), that may not contain heterologous DNA or protein, but consider rec-DNA technology in their production processes. Food labeling represents a complex issue with several concerns. A hypothetical ‘‘zero tolerance’’ approach is not favorably seen or applicable (Table 1: ‘‘Harvest of fear’’ PBS website). Some argue that warning labels on approved GM foods may imply harmful effects on health and create false alarm, even though those effects have been excluded, by the pre-market safety trials. It seems also plausible that distinctive food labeling, segregation of GM and non-GM foodstuff (during storage, shipping and processing) and transgenic organism traceability would be difficult to achieve and may have a significant impact on the costs of the goods. Thus, for the coming years, food labeling may still represent a major field of discussion, for novel as well as for conventional foods [65]. Indications, laws and regulations are frequently modified due also to the rapid changes in the biotechnology field and the relatively short juridical experience of society in this new area.
106 Labeling enables consumers to get a clear and comprehensive information on the contents and the composition of food and helps buyers to make an informed choice while purchasing their products. A recent survey was set up to determine whether perceiving obvious benefits from eating genetically modified soybeans, would have altered, among consumers, personal risk assessment and desire for labeling in general [66]. The results of this study showed that consumers reading about the GM soybean with obvious consumer benefits were significantly more comfortable eating it, than those reading about the GM soybean with no obvious consumer benefits. This interesting relationship, observed between perceived risks and benefits, had also been shown in two previous studies [67,68]. Thus, activities or technologies that are judged high in risk tend to be judged low in benefit, and vice versa. Such an observation has suggested the hypothesis that creation and commercialization of new transgenic cultivars with more direct benefits for consumers, would increase public acceptance [66]. Indeed, as one can tell from the current most used transgenes, to date most modifications to crop plants have mainly benefit producers. Consumers stand to benefit by development of food crops with increased nutritional value, medicinal properties, enhanced taste and esthetic appeal [69]. As national regulations testify (Table 1) extremely different public and governmental sensibilities concerning biotechnology, may be observed throughout the world [70–73]. This extremely heterogeneous condition should be addressed with effective strategies such as new labeling systems and consensus conferences, to stimulate public information and public debate and to defeat the spread of misinformation, easily possible in a new and fast growing issue [74,75]. It is generally recognized that each adopted approach, aimed at the reconciliation between science, or technology, and the community, should be adapted to the cultural context of each country and society. Health risks: Present knowledge and potential hazards Health risks associated with GMOs can be classified as risks for the person and risks for the environment (Table 6). Present data do not provide evidences of specific hazards for populations exposed to a diet containing approved GMOs [76]. Epidemiological and experimental studies are in progress in different countries and are supported by different institutions including universities, research centers, companies, or supranational organizations. Prudential criteria are required due to the novelty of the procedures and of the possible derivable hazard, and the wide and quickly growing exposure levels that the world population encountered in a relatively short period of time. Risks for health comprehends: food poisoning, food intolerance, auto-intoxication, anti-nutrients. In principle, upon manipulation, including a genetic modification, a food organism may become toxic. Food poisoning is an acute disturbance occurring
107 Table 6. Classification of risks related to production and consumption of GM foods. Risks for the person #
Risks for the environment #
Toxic food Food poisoning
Extinction of existing species or varieties Of animals Of plants Interference with environmental balance Uncontrolled spread of transgenic species Gene transfer Unpredictable risks
Non toxic food Non immune-mediated (a) food intolerance (b) auto-intoxication (c) anti-nutrients Immune-mediated (food allergy) Gene Transfer Unpredictable risks
after consumption of food that is contaminated with toxic agents that are inherently unsuitable for human consumption. Apparently this may appear as a fearful and worrying risk. Instead, this possibility is easily detectable and controllable, as it is for any natural or synthetic novel food humans have met before. Major concerns are related to effects over long periods or overreactions on susceptible population subgroups. Food intolerance and auto-intoxication are food-induced morbid states that do not involve a specific immune system reaction. Individuals with food intolerances may lack an enzyme that is needed to digest a certain food and show a pathological sensitivity to it. Autointoxications are caused by an accumulation of harmful metabolic intermediates or substances, through an endogenous origin. Genetic differences in the population may determine or enhance these effects. The susceptibility to such reactions can be characterized by planning extended and accurate pilot studies. Anti-nutrients, although not necessarily toxic per se, can be plant compounds that decrease the nutritional value of that plant. Anti-nutrients usually make an essential nutrient unavailable or indigestible when consumed. For example, phytate, a common component of most seeds and cereals, forms a complex with many important minerals, making less of the minerals available [77]. Trypsin inhibitor, lectins, isoflavones, stachyose and raffinose are further examples of anti-nutrients that may be studied during nutritional assessments of novel foods [46]. The presence of these compounds in a plant is of essential importance not only for human nutrition but also for animal feeding. Further research will likely provide effective protocols to focus on whether genetic engineering has accidentally changed the nutritional components associated with conventional cultivars of a crop. As antibiotic resistance markers are routinely used for the selection of transformed plant cells, concerns have been raised about whether the enzyme product of the DNA might be produced in transgenic plant cells. Although
108 various processing procedures would inactivate the enzyme in processed foods, ingestion of fresh or raw transgenic organisms may result in intake of active enzymes. Such an occurrence may cause the inactivation of orally ingested antibiotics but preliminary pilot studies showed that antibiotics would remain effective. Moreover, antibiotic resistance is widely diffused in nature and is part of ancient phylogenetic processes involving both intrinsic and transferable resistance. It should be noted that all approved GMOs are currently created by genetic engineering of GRAS (Generally Recognized As Safe) organisms that have already been used or eaten without risks by humans and animals, since the dawn of time. Presently, current food safety regulations for traditionally bred food crops are, in practice, less stringent compared to those applied to GM foods [62]. Food allergies Food allergy is a specific immunomediated reaction to one or more food components; it is characterized by altered bodily reactivity (hypersensitivity) to an antigen, in response to a first exposure. Atopic individuals may overreact to an allergen with different clinical manifestations, both localized or systemic. This event can show different degrees of severity and can also be fatal as in the anaphylactic shock. Agricultural biotechnology implies the introduction of novel proteins into the modified foods, and proteins can be allergens. There are two situations that may occur: first, a known allergen may be transferred from a donor crop into a nonallergenic target crop. The second scenario is the creation of a novel allergen with a possible de novo sensitization of the population [78]. The case of the Brazil nut is a well known example, even if its conclusions underwent opposite interpretations: GMOs represent an ‘‘allergic hazard’’ or the allergic risk ‘‘can be prevented.’’ Methionine in relative low concentration in the protein fraction of soybean (Glycine max) seeds, compromises the nutritional value of such crops. Research was conducted in order to improve the quality of soybean meal as an animal feed and a transgene, coding for a storage protein from a Brazil nut (Betholletia excelsa), was successfully introduced and expressed into soybean. While it was known at that time that Brazil nuts were allergenic to some consumers, no one had ever identified which gene product from Brazil nuts was the responsible allergen, thus, several laboratory tests had to be conducted in turn, to determine whether potential allergens had been transferred to the GM plant. Investigations by Nordlee et al., on the transgenic soybeans, identified the methionine-rich 2S albumin as a major Brazil nut-derived allergen and demonstrated the possibility of transfer of allergens from a food, known to be allergenic, into another food, through genetic engineering. These data were the conclusion of a safety assessment study, begun in 1993, in collaboration with the University of Nebraska [79]. Following the 1993 preliminary findings,
109 all the field trials were discontinued and all plant material and seeds, not held for laboratory study, destroyed. Despite this case, it should be noted that consumers’ exposure to approved and commercialized transgenic crops, is reaching significant levels, especially in some countries such as the US or China. However, no reports exist regarding allergic reactions to the crops that have been approved for human consumption, and available data seem to confirm the safety of the approved GM foods [80–82]. Biotechnology also offers many promising perspectives, as concerns food allergies. In the future, biotechnology may be employed to characterize, eliminate or attenuate the potential of food allergens. Even if at an experimental phase, this represents a novel, yet encouraging approach. For instance, Herman et al. employed transgene-induced gene silencing to prevent the accumulation in soybean seeds of Gly m Bd 30 K protein, a major (i.e., immunodominant) soybean allergen [83]. Further, a reduced content of a known allergen was shown in GM rice. In this study, antisense RNA strategy was applied to repress the allergen gene expression in maturing rice seeds. Immunoblotting and ELISA analyses of the seeds showed that allergen content of seeds, from several transgenic rice plants, was markedly lower than that of the seeds from parental wild type rice [84]. Several approaches are currently being employed to assess the possible allergenic potential of the transgenic proteins introduced in GMOs [85,86,51,78]. The safety evaluation of transgenic foods is relatively easy when the allergenicity of the gene source is known. New and powerful tools may be represented by bioinformatics and modern databases. A typical use of bioinformatics, involves the search for possible homologies between the newly introduced protein and known allergens present in up-to-date databases [87]. Sequence analysis plays an important role in assessing the potential allergenicity of proteins used in transgenic foods, particularly for proteins that have not previously been part of the food supply. Sequence comparisons are used to indicate potential unexpected cross reactivity to existing allergens and to assess the potential for developing new sensitivities [88]. Bioinformatics plays an important role in this process and new laboratory protocols and surveillance procedures are being performed. Immunological and biomolecular assays involving the reactivity assessment of a novel food are available, including analysis of IgE antibodies from the serum of individuals with known allergies to the source of the transferred DNA or to materials that are broadly related to that protein or protein domains. The immunoreactivity can be tested also in appropriate animal models. Biochemical tests, focusing on the evaluation of the chemical stability of the foreign products after food processing, storage, cooking and digestion are usually performed, as well. Since food allergens may share physico-chemical properties that distinguish them from nonallergens, characterizing those properties may serve as an effective
110 preventive tool to predict the inherent allergenicity of transgenic proteins introduced into novel foods. A candidate property is the stability to digestion. In a study involving an in vitro model of gastric digestion, Astwood JD et al. conclude that the stability to digestion is a significant and valid parameter that distinguishes food allergens from non-allergens [89]. Additional factors, such as expression level of the novel protein in the edible portion of the food, may also bear significant information. Further, the stability of potential allergens from GM plants, is also tested by heat stability analysis, to simulate cooking. As transgenic foods commonly eaten may contain transgenes (in variable quantities depending on the food and processing), concerns regarding the exposures to transgenic DNA sequences and the onset of specific allergies, aroused. Transgenic DNA degradation was tested in vitro, by means of a simulated digestion model, that showed how heterologous DNA fragments from GM foods, may survive passage through the small intestine [90]. However, in a comprehensive review Jonas et al. conclude that DNA from GMOs is equivalent to DNA from conventional, non-transgenic organisms and, consequently, any risks associated with the consumption of DNA will remain, irrespective of its origin [91]. Especially for infants, dietary nucleotides have been reportedly beneficial, since they positively influence lipid metabolism, immunity, and tissue growth, development and repair [91–93]. Gene transfer and foreign DNA intake As previously mentioned, fragments of transgenic nucleic acids may be present in GM food or derivates. Transgenic sequences include the promoter, the transgene and the terminator [13]. The intake of such sequences, the ‘‘gene cassette,’’ is causing increasing worries in case a transfer event would occur upon ingestion. Hereafter, with ‘‘horizontal gene transfer,’’ we will refer to the passage, from engineered foods to gut epithelium, or cells from other tissues, or gut microflora. But, is this horizontal transfer from ‘‘diet’’ to human cells possible? Can a DNA sequence of hundreds of nucleotides pass the intestinal barrier? Can it integrate into the host genome or even be inherited by the offspring? Can this event be possible in animal feeding and represent a risk for the food chain? Can gut microflora acquire antibiotic resistance through transgenic DNA? Researchers or public health are currently considering these and other related issues. Preliminary data are available to open up a discussion more than to clearly define the problem. Insertion of a DNA sequence in the genome of a human cell represents a mutation event and is feared to cause neoplastic transformation or other genetic modifications. However, to address this issue, it should be immediately noted that in principle the transgenic DNA is considered biochemically equivalent to the DNA from any other source and that almost every food contains DNA, either of animal or plant origin. Since the dawn of time, human gut, and that of other methazoans, has been
111 Table 7. ILSI Europe Workshop on Safety Considerations of DNA in Food, 26–28 June 2000 (modified from Jonas DA et al., 2001). rec-DNA: recombinant DNA. GMOs: genetically modified organisms. Statements of the ILSI Europe Workshop on Safety Considerations of DNA in Food, 26–28 June 2000 1. All DNA, including rec-DNA is composed of the same four nucleotides. 2. In view of the variability of dietary intake of DNA, consumption of foods derived from GMOs does not measurably change the overall amount of DNA ingested through the diet. 3. Taking into account the natural variations of DNA sequences, the present use of recombinant techniques in the food chain does not introduce changes in the chemical characteristics of the DNA. 4. There is no difference in the susceptibility of rec-DNA and other DNA to degradation by chemical or enzymatic hydrolysis. 5. The metabolic fate of DNA digestion products is not influenced by the origin of the DNA. 6. DNA is not toxic at levels usually ingested. Where there is potential for adverse effects, e.g., in gout, this is due to excessive intake, not the origin of DNA. 7. Ingested DNA showed no indication of allergenic or immunogenic properties, that would be of relevance for consumption of GMOs-derived foods. 8. Uptake, integration and expression of any residual extracellular DNA fragments from foods by microorganisms of the gastro-intestinal tract cannot be excluded. However, each of these circumstances is a rare event and would have to happen sequentially. 9. In vivo uptake of DNA fragments by mammalian cells after oral administration has been observed. However, there are effective mechanisms to avoid genomic insertion of foreign DNA. There is no evidence that DNA from dietary sources has ever been incorporated into the mammalian genome.
exposed to a constant flow of dietary nucleic acids. This also implies that if some risks were associated with such an event, representing a selective pressure, counteracting evolutionary mechanisms would likely have arisen and spread. Nevertheless, even if the probability of gene transfer from GMOs to mammalian cells seems extremely low and statistically insignificant, it is highly important considering hypothetical health risks on the long period scale [91]. Evidence is available to address the issue of the bioequivalence and horizontal transfer, as stated by the ILSI Europe Workshop on Safety Considerations of DNA in Food (Table 7) [94]. Concerns regarding possible adverse effects associated with gene transfer, should take into account the metabolic fate of DNA and RNA, and the content of rec-DNA with respect to the total amount ingested. Apart from differences in the sequence, recombinant DNA is composed of the same four nucleotides as nonrecombinant DNA. Chemical properties of such DNA molecules are not altered by rec-DNA techniques that are currently employed in the production of the approved GMOs. It seems reasonable, therefore, to expect, for transgenic sequences, the same degradation pathways of any other DNA fragment. Degradation is a generic term including several kinds of reactions either catalyzed by enzymes or caused by chemical
112 as well as physical agents [95,96]. Virtually all these events, can occur during food processing, storage, cooking and digestion, thus allowing only smaller nucleic acid fragments to reach the intestinal lumen [90]. During digestion, the acid environment of the stomach, along with pancreatic nucleases, and those secreted by intestinal epithelial cells, cause extensive hydrolysis of dietary DNA. Such breakage releases nucleotides that are subsequently processed through the gastro-intestinal (GI) tract to sugars, purine and pyrimidine bases. Proteolytic enzymes are responsible for nucleoprotein cleavage in the gut. As mentioned, nucleic acids content in raw foods can differ significantly, depending on the type of cells present in the foodstuff, and can be considerably affected by food processing. DNA and RNA in raw plant storage tissues, is lower than in animal muscle tissues and edible offals, however, significant amounts of dietary DNA and RNA are assumed daily [91]. The organisms with the highest content of nucleic acids are fungi, bacteria and yeast. In humans, with a certain variability depending on the diet and on the effects of processing, this intake may range from 0.1 to 1.0 g/person per day. Processed foods have a lower nucleic acids content which can sometimes drop, to almost undetectable quantities, in case of highly refined products. As regards rec-DNA content in GM crops, it should be noted that in some homozygous insect-protected and herbicide-tolerant cultivars, it may range from 0.00018 to 0.0011% of the plant genome, allowing an estimate of total per-capita intake of rec-DNA (mg/d) from GM maize, soya and potatoes as 0.38 mg/d, assuming that only GM crops would be consumed. This is about 0.00006% of a typical daily DNA intake of 0.6 g [91]. Some approved and commercialized plants, however, are heterozygous, and the nucleic acids intake is lower when this crop is consumed. Under physiological dietary exposure to GM crops, several individual events, most of them considered somewhat rare, would have to occur sequentially to cause gene transfer to gut microflora or mammalian cells (Fig. 6). In microorganisms, this event can occur by one of three different mechanisms: conjugation, transduction, transformation. In particular, bacterial transformation involves the uptake of free extracellular DNA with an inheritable incorporation [95]. The host cell may either undergo a (i) transformation without expression of foreign proteins, or (ii) transformation with expression of heterologous products. In order to be able to actively take up DNA from the environment, bacteria have to achieve a state of competence, during which DNA molecules may be admitted to the cell. Stable incorporation relies on rare events of recombination, which ultimately depend on the extent of the genetic homology between the recipient cell and the fragment, or by formation of independent replicons. However, foreign DNA acquisition is not sufficient for the production of a heterologous protein in a transformed host cell. It always requires the presence of a coding sequence, proper regulatory elements, proper frequency of codon usage and, more generally, of a favorable cell environment which allows processing and expression of the heterologous protein.
113 The complete TE would have to be released into the GI tract, from the ingested GM food.
The TE would have to survive degradation by host nucleases or food associated nucleases.
The TE would have to compete for uptake with other DNA present in the GI.
Host mammalian or bacterial cells would have to be competent for transformation.
Uptake of the TE from a host cell would have to occur.
The TE would have to survive further degradation within the cell.
The TE would have to be inserted, into the host DNA, by repair or recombination events.
The TE would have to be expressed. Fig. 6. Individual events required for the transfer, in the gastro-intestinal (GI) tract, of ingested transgenic elements (TE), from a genetically modified plant, to microbial or mammalian cells, under physiological dietary exposure.
A foreign DNA fragment, that after cellular internalization is not integrated into the resident chromosome, might be expressed only transiently. Noteworthy, in the absence of a specific selective pressure that confers an advantage to the transformed cell with respect to the others, the spread of the trait may not occur. The importance of the gene transfer issue in gut microflora, may be better perceived considering how common this event is in nature. Unlike eukaryotes, bacteria have obtained a significant proportion of their genetic diversity through the acquisition of sequences from distantly related organisms. Horizontal gene transfer produces extremely dynamic genomes and has, in turn, a great impact on bacterial population dynamics as well as on bacterial evolution and speciation [97,98]. Microenvironment influences bacterial competence acquisition.
114 Brautigam et al. assessed the effects of food matrices on the natural transformation of B. subtilis and showed how this species may develop natural competence when grown in milk products. Frequency of transformation varied with the content of fat in the foodstuff [99]. This general phenomenon may play a role in milk products obtained from GM dairy animals, but also in gut or soil microflora. On the other hand, over a million years of evolution, bacteria have also adopted an effective mechanism to protect themselves from foreign DNA infection. The modification/restriction system, for instance, is used by E. coli to distinguish between its own genome and foreign DNA introduced by bacteriophage infection, plasmid transfer or transfection [100]. DNA with familiar patterns of methylation are immune to attack, whereas those not recognized by the specialized cellular machinery are earmarked for degradation and promptly degraded by endogenous restriction enzymes [95]. An area of concern focuses on the possibility that antibiotic resistance genes used as markers in transgenic crops may be transferred to pathogenic bacteria. The most frequently used transgene is nptII, originating from the E. coli transposon 5 (Table 3 and Fig. 4). This gene confers resistance to selected aminoglycoside antibiotics. In 1997, nptII was found to be present in 61% of the surveyed GM crops. However, six years later, its employment in GM plants was found to be reduced to about 44% of the surveyed transgenic crops [13]. Antibiotic-resistance markers that are presently employed in GM crops for selection, belong to a class of the limited clinical importance of the antibiotic they inactivate, and their frequent occurrence in nature. It is also generally recognized that the increased number of resistant bacterial strains is more likely due to the widespread use, abuse and misuse of antibiotics, in human and zootechnical applications, rather than to the recent adoption of GM crops [101]. In 1993, Fuchs et al. showed that ingestion of genetically engineered plants, expressing the NPTII protein, did not determine particular safety concerns for human health. Purified NPTII protein produced in E. coli, was shown to be chemically and functionally equivalent to the NPTII protein produced in genetically engineered cotton seed, potato tubers and tomato fruit and to degrade rapidly under simulated mammalian digestive conditions [102,103]. As for other transgenic sequences, the chain of events that would transfer an antibiotic resistance marker, from a GM plant to a pathogenic bacterium is quite unlikely. A negative impact of GMOs on antibiotic efficacy has been assessed and considered improbable by scientists and regulatory agencies [101]. However, in response to concerns about this remote possibility scientists are starting to use, in transgenic plants, other marker genes, such as the GFP gene from Aequorea Victoria, and alternative strategies [104]. Further approaches to the problem include removal of the selection marker after successful gene transfer.
115 Besides bacteria, it is also relevant for public health to assess the likelihood of uptake of free undigested transgenic fragments by mammalian cells. In recent studies, Jennings JC et al. failed to detect fragments of the cp4 epsps transgene in a variety of tissue samples from pigs, fed glyphosate-tolerant soybeans. Immunoreactive fragments of the transgenic protein were not detectable either. In a similar study, broiler chickens were fed a diet containing insectprotected plants. In accordance with the previous experiments, fragments of transgenic and endogenous plant DNA, as well as transgenic protein, were not detected in the chicken breast muscle samples [105,106]. These evidences are in contrast with those previously published by Schubbert et al. (1994), who observed that high doses of orally administered naked M13 phage DNA survived, transiently, in the GI tract and entered the bloodstream of mice [107]. Further experiments suggested a possible transport mechanism of foreign DNA, through the intestinal wall and Peyer’s patches, to peripheral blood leukocytes and into several organs [108]. Food-ingested M13 DNA fed to pregnant mice, could also be detected in several cells of various organs of the foetuses and newborn animals, suggesting a possible transfer through the transplacental route, but not via the germ line [109]. Thus, to challenge these preliminary observations with a more natural scenario, mice were fed soybean leaves, and the fate of the small subunit of the ribulose-1,5bisphosphate carboxylase (Rubisco) gene was followed in the mouse organism. A short fragment of the authentic Rubisco gene, was observed to be transferred to spleen and liver, of the mice [110]. In the same paper, authors failed to detect any GFP expression, in gut, spleen or liver, upon oral administration of GFPencoding constructs. Conversely, green fluorescence was observed in mouse skeletal muscle tissues, after injection of the same constructs. Therefore, it seems plausible that small amounts of ingested DNA are not broken down under physiological digestive processes. This DNA may either enter the bloodstream or be excreted. It is generally accepted, however, that body’s normal defence systems eventually would destroy this DNA thus preventing potential adverse effects. The significance of the observations of Schubbert and Doerfler have been questioned by other recent studies that failed to detect transgenic proteins or transgenes in tissues from animals fed on GMOs [111–113]. It is generally accepted that the available data are not sufficient to demonstrate that plant DNA can be transferred, stably maintained and expressed in mammalian cells. Further, it should be considered that many genes used for the genetic modification of food organisms come from the so called GRAS organisms, with a long and safe history of human coexistence. However the novelty of the problem requires further accurate research, experimental reproduction and prudence in result interpretation. Carcinogenesis, mutagenesis, reactivation of dormant viruses and even generation of new viruses have been recently postulated, for example in the case of horizontal transfers involving the promoter 35s (P-35s) from the
116 Cauliflower Mosaic Virus (CaMV). This regulatory sequence is used in most of the currently approved GM plants to give constitutive overexpression of transgenes (Fig. 3). Critics maintain that the naked recombinant promoter may have harmful effects, for instance due to genetic instability. Rebuttals to these hypothesis are based on several evidences [114]. CaMV P-35s is a nucleic acid sequence and, as previously considered (Fig. 6), a multi-step chain of events would have to occur to escape the normal digestive breakdown process, penetrate a cell, insert itself into a human chromosome, and take on the control of expression of the resident genes. Moreover, a significant percentage of the cauliflowers (10%) and cabbages normally sold in our markets and consumed, are found to be naturally infected by CaMV [115]. Therefore, it has been estimated that, historically, humans have been ingesting CaMV and its 35s promoter at levels that are over 10,000 times greater than those present in uninfected transgenic plants. It is actually this line of argument that led the USDA to endorse the use of CaMV P-35s as a safe promoter in GM crops. An integrated, multidisciplinary approach is required to address these and other issues related to gene transfer events and foreign DNA intake consequences. Accurate analysis of gene cassette sequence properties by molecular biology and bioinformatics tools is needed together with an open and rigorous interpretation of the results. Detection of possible health effects by classical epidemiology studies is an important but limited strategy. Appropriate risk assessment requires a deep comprehension of the basic mechanisms involved in gene transfer and uptake, considering their potential role on extended exposed populations along a long-term scale. Conclusions Years of debate and intensive scientific work seem to exclude the presence of evident health hazards associated with the consumption of the authorized GMOs and of GMO-derived ingredients. When approved and marketed, such food was confirmed to be devoid of additional risks with respect to its conventional, non-transgenic counterpart [76,86,100,116]. Risk assessment procedures and authorization protocols are available and consider the costbenefit ratio of the introduction of GMOs in the food chain. However, it is not appropriate to generalize and the assessment of risks associated with the consumption of novel foods per se, would be meaningless if not conversely focused on a specific product. It is not presently possible to state that every GMO is potentially safe, but neither that a food is unsafe if ‘‘GM.’’ Availability of a rigorous and deep knowledge on the basic mechanisms involved in genetic modification is the major limiting point for an effective risk assessment. Intensive scientific effort is in progress to thoroughly understand and foresee possible consequences on humans, animals and environment. Special care is required due to the particular novelty of these foods and their rapid diffusion in the absence of a co-evolution process. Social aspects and economical implications
117 strongly influence the commercialization of GMO products, in both the directions. In different countries, several regulations are available and frequently updated. GMO foods represent a challenging issue for public health science, involving safe nutrition, health risks, monitoring tools, policy making. In the new Millennium, prevention tasks include also the survaillance on novel foods and the management of the health safety of Genetically Modified products.
References 1. Kuvshinov VV, Koivu K, Kanerva A and Pehu E. Molecular control of transgene escape from genetically modified plants. Plant Sci 2001;160(3):517–522, Feb 5. 2. Daniell H. Molecular strategies for gene containment in transgenic crops. Nat Biotechnol 2002 Jun;20(6):581–586; Erratum in: Nat Biotechnol 2002;20(8):843, Aug. 3. Cohen SN, Chang AC, Boyer HW and Helling RB. Construction of biologically functional bacterial plasmids in vitro. Proc Natl Acad Sci USA 1973;70(11):3240–3244, Nov. 4. Morrow JF, Cohen SN, Chang AC, Boyer HW, Goodman HM and Helling RB. Replication and transcription of eukaryotic DNA in Escherichia coli. Proc Natl Acad Sci USA 1974;71(5): 1743–1747, May. 5. Temin HM and Mizutani S. RNA-dependent DNA polymerase in virions of Rous sarcoma virus. Nature 1970;226:1211–1213. 6. Baltimore D. RNA-dependent DNA polymerase in virions of RNA tumour viruses. Nature 1970;226:1209–1211. 7. Palmiter RD, Brinster RL, Hammer RE, Trumbauer ME, Rosenfeld MG, Birnberg NC and Evans RM. Dramatic growth of mice that develop from eggs microinjected with metallothionein-growth hormone fusion genes. Nature 1982, Dec 16;300(5893):611–615. 8. James C. Global Status of Commercialized Transgenic Crops: 2000. ISAAA Briefs No. 21: Preview ISAAA, Ithaca, NY, 2000. 9. Herrera-Estrella L, Depicker A, van Montagu M and Schell J. Expression of chimaeric genes transfered into plant cells using a Ti-plasmid-derived vector. Nature 1983;303:209–213. 10. Bevan MW, Flavell RB and Chilton MD. A chimaeric antibiotic resistance gene as a selectable marker for plant cell transformation. Nature 1983;304:184–187. 11. Fraley RT, Rogers SG, Horsch RB, Sanders PR, Flick JS, Adams SP, Bittner ML, Brand LA, Fink CL, Fry JS, Galluppi GR, Goldberg SB, Hoffmann NL and Woo SC. Expression of bacterial genes in plant cells. Proc Natl Acad Sci USA 1983;80:4803–4807. 12. Murai N, Sutton DW, Murray MG, Slightom JL, Merlo DJ, Reichert NA, SenguptaGopalan C, Stock CA, Barker RF, Kemp JD and Hall TC. Phaseolin gene from bean is expressed after transfer to sunflower via tumor-inducing plasmid vectors. Science 1983;222: 476–482. 13. Bruderer S and Leitner KE. Modified (GM) Crops: molecular and regulatory details. 1. Version 2-30.03.2003. 14. Zambryski P, Tempe J and Schell J. Transfer and function of T-DNA genes from agrobacterium Ti and Ri plasmids in plants. Cell 1989;56(2):193–201, Jan 27. 15. Rakoczy-Trojanowska M. Alternative methods of plant transformation-a short review. Cell Mol Biol Lett 2002;7(3):849–858. 16. Fromm M, Taylor LP and Walbot V. Expression of genes transferred into monocot and dicot plant cells by electroporation. Proc Natl Acad Sci USA 1985;82(17):5824–5828, Sep. 17. Crossway A, Oakes JW, Irvine JM, Ward B, Knauf VC and Shewmaker CK. Integration of foreign DNA following microinjection of tobacco mesophyll protoplasts. Mol Gen Genet 1986;202:179–185.
118 18. Somerville HJ and Pockett HV. An insect toxin from spores of Bacillus thuringiensis and Bacillus cereus. J Gen Microbiol 1975;87(2):359–369, Apr. 19. Bulla LA Jr, Kramer KJ and Davidson LI. Characterization of the entomocidal parasporal crystal of Bacillus thuringiensis. J Bacteriol 1977;130(1):375–383, Apr. 20. Schnepf E, Crickmore N, Van Rie J, Lereclus D, Baum J, Feitelson J, Zeigler DR and Dean DH. Bacillus thuringiensis and its pesticidal crystal proteins. Microbiol Mol Biol Rev 1998;62(3):775–806, Sep. 21. Aronson AI and Shai Y. Why Bacillus thuringiensis insecticidal toxins are so effective: unique features of their mode of action. FEMS Microbiol Lett 2001;195(1):1–8, Feb 5. 22. de Maagd RA, Bravo A and Crickmore N. How Bacillus thuringiensis has evolved specific toxins to colonize the insect world. Trends Genet 2001;17(4):193–199, Apr. 23. Hofte H and Whiteley HR. Insecticidal crystal proteins of Bacillus thuringiensis. Microbiol Rev 1989;53(2):242–255, Jun. 24. Crickmore N, Zeigler DR, Feitelson J, Schnepf E, Van Rie J, Lereclus D, Baum J and Dean DH. Revision of the nomenclature for the Bacillus thuringiensis pesticidal crystal proteins. Microbiol Mol Biol Rev 1998;62(3):807–813, Sep. 25. Crickmore N, Zeigler DR, Schnepf E, Van Rie J, Lereclus D, Baum J, Bravo A and Dean DH. Bacillus thuringiensis toxin nomenclature, 2002; http://www.biols.susx.ac.uk/Home/Neil_ Crickmore/Bt/index.html. 26. Wei JZ, Hale K, Carta L, Platzer E, Wong C, Fang SC and Aroian RV. Bacillus thuringiensis crystal proteins that target nematodes. Proc Natl Acad Sci USA 2003;100(5):2760–2765, Mar 4. 27. Tian YC, Qin XF, Xu BY, Li TY, Fang RX, Mang KQ, Li WG, Fu WJ, Li YP, Zhang SF, et al. Insect resistance of transgenic tobacco plants expressing delta-endotoxin gene of Bacillus thuringiensis. Chin J Biotechnol 1991;7(1):1–13. 28. Betz FS, Hammond BG and Fuchs RL. Safety and advantages of Bacillus thuringiensisprotected plants to control insect pests. Regul Toxicol Pharmacol 2000;32(2):156–173, Oct. 29. Theoduloz C, Vega A, Salazar M, Gonzalez E and Meza-Basso L. Expression of a Bacillus thuringiensis delta-endotoxin cry1Ab gene in Bacillus subtilis and Bacillus licheniformis strains that naturally colonize the phylloplane of tomato plants (Lycopersicon esculentum, Mills). Appl Microbiol 2003;94(3):375–381. 30. Ferre J and Van Rie J. Biochemistry and genetics of insect resistance to Bacillus thuringiensis. Annual Review of Entomology 2002;47:501–533. 31. Tabashnik BE. Evolution of resistance to Bacillus thuringiensis. Annual Review of Entomology 1994;39:47–79, Jan. 32. Tabashnik BE, Liu YB, Dennehy TJ, Sims MA, Sisterson MS, Biggs RW and Carriere Y. Inheritance of resistance to Bt toxin crylac in a field-derived strain of pink bollworm (Lepidoptera: Gelechiidae). J Econ Entomol 2002;95(5):1018–1026, Oct. 33. Halcomb JL, Benedict JH, Cook B, Ring DR and Correa JC. Feeding behavior of bollworm and tobacco budworm (Lepidoptera: Noctuidae) larvae in mixed stands of nontransgenic and transgenic cotton expressing an insecticidal protein. J Econ Entomol 2000;93(4): 1300–1307, Aug. 34. Losey JE, Rayor LS and Carter ME. Transgenic pollen harms monarch larvae. Nature 1999; 399(6733):214, May 20. 35. Shelton AM and Sears MK. The monarch butterfly controversy: scientific interpretations of a phenomenon. Plant J 2001;27(6):483–8, Sep; Erratum in: Plant J 2002;29(5):679, Mar. 36. Hellmich RL, Siegfried BD, Sears MK, Stanley-Horn DE, Daniels MJ, Mattila HR, Spencer T, Bidne KG and Lewis LC. Monarch larvae sensitivity to Bacillus thuringiensispurified proteins and pollen. Proc Natl Acad Sci USA 2001;98(21):11925–11930, Oct 9. 37. Sears MK, Hellmich RL, Stanley-Horn ED, Oberhauser KS, Pleasants JM, Mattila HR, Siegfried BD and Dively GP. Impact of Bt corn pollen on monarch butterfly populations: a risk assessment. Proc Natl Acad Sci USA 2001;98(21):11937–11942, Oct 9.
119 38. Gatehouse AM, Ferry N and Raemaekers RJ. The case of the monarch butterfly: a verdict is returned. Trends Genet 2002;18(5):249–251, May. 39. Yoder JI. Parasitic plant responses to host plant signals: a model for subterranean plant–plant interactions. Curr Opin Plant Biol 1999;2(1):65–70, Feb. 40. Steinrucken HC and Amrhein N. The herbicide glyphosate is a potent inhibitor of 5-enolpyruvyl-shikimic acid-3-phosphate synthase. Biochem Biophys Res Commun 1980; 94(4):1207–1212, Jun 30. 41. Steinrucken HC, Schulz A, Amrhein N, Porter CA and Fraley RT. Overproduction of 5-enolpyruvylshikimate-3-phosphate synthase in a glyphosate-tolerant Petunia hybrida cell line. Arch Biochem Biophys 1986;244(1):169–178, Jan. 42. Boocock MR and Coggins JR. Kinetics of 5-enolpyruvylshikimate-3-phosphate synthase inhibition by glyphosate. FEBS Lett 1983;154(1):127–133, Apr 5. 43. Stalker DM, Hiatt WR and Comai L. A single amino acid substitution in the enzyme 5-enolpyruvylshikimate-3-phosphate synthase confers resistance to the herbicide glyphosate. J Biol Chem 1985;260(8):4724–4728, Apr 25. 44. Klee HJ, Muskopf YM and Gasser CS. Cloning of an Arabidopsis thaliana gene encoding 5-enolpyruvylshikimate-3-phosphate synthase: sequence analysis and manipulation to obtain glyphosate-tolerant plants. Mol Gen Genet 1987;210(3):437–442, Dec. 45. James C. Global Review of Commercialized Transgenic Crops: 2001. ISAAA Briefs No. 24: Preview. ISAAA: Ithaca, NY, 2000. 46. Padgette SR, Taylor NB, Nida DL, Bailey MR, MacDonald J, Holden LR and Fuchs RL. The composition of glyphosate-tolerant soybean seeds is equivalent to that of conventional soybeans. J Nutr 1996;126(3):702–716, Mar. 47. Ridley WP, Sidhu RS, Pyla PD, Nemeth MA, Breeze ML and Astwood JD. Comparison of the nutritional profile of glyphosate-tolerant corn event NK603 with that of conventional corn (Zea mays L.). J Agric Food Chem 2002;50(25):7235–7243, Dec 4. 48. Harrison LA, Bailey MR, Naylor MW, Ream JE, Hammond BG, Nida DL, Burnette BL, Nickson TE, Mitsky TA, Taylor ML, Fuchs RL and Padgette SR. The expressed protein in glyphosate-tolerant soybean, 5-enolpyruvylshikimate-3-phosphate synthase from Agrobacterium sp. strain CP4, is rapidly digested in vitro and is not toxic to acutely gavaged mice. J Nutr 1996;126(3):728–740, Mar. 49. Hammond BG, Vicini JL, Hartnell GF, Naylor MW, Knight CD, Robinson EH, Fuchs RL and Padgette SR. The feeding value of soybeans fed to rats, chickens, catfish and dairy cattle is not altered by genetic incorporation of glyphosate tolerance. J Nutr 1996;126(3):717–727, Mar. 50. Cromwell GL, Lindemann MD, Randolph JH, Parker GR, Coffey RD, Laurent KM, Armstrong CL, Mikel WB, Stanisiewski EP and Hartnell GF. Soybean meal from roundup ready or conventional soybeans in diets for growing-finishing swine. J Anim Sci 2002;80(3): 708–715, Mar. 51. Nair RS, Fuchs RL and Schuette SA. Current methods for assessing safety of genetically modified crops as exemplified by data on Roundup Ready soybeans. Toxicol Pathol 2002; 30(1):117–125, Jan–Feb. 52. Taylor NB, Fuchs RL, MacDonald J, Shariff AR and Padgette SR. Compositional analysis of glyphosate-tolerant soybeans treated with glyphosate. J Agric Food Chem 1999;47(10): 4469–4473, Oct. 53. Chang HS, Kim NH, Park MJ, Lim SK, Kim SC, Kim JY, Kim JA, Oh HY, Lee CH, Huh K, Jeong TC and Nam DH. The 5-enolpyruvylshikimate-3-phosphate synthase of glyphosatetolerant soybean expressed in Escherichia coli shows no severe allergenicity. Mol Cells 2003; 15(1):20–26, Feb 28. 54. Ye X, Al-Babili S, Klo¨ti A, Zhang J, Lucca P, Beyer P and Potrykus I. Engineering provitamin A (b-carotene) biosynthetic pathway into (carotenoid-free) rice endosperm. Science 2000;287: 303–305.
120 55. Beyer P, Al-Babili S, Ye X, Lucca P, Schaub P, Welsch R and Potrykus I. Golden Rice: introducing the beta-carotene biosynthesis pathway into rice endosperm by genetic engineering to defeat vitamin A deficiency. J Nutr 2002;132(3):506S–510S, Mar. 56. Nestle M. Genetically engineered ‘‘golden’’ rice unlikely to overcome vitamin A deficiency. J Am Diet Assoc 2001;101(3):289–290, Mar. 57. Potrykus I. Nutritionally enhanced rice to combat malnutrition disorders of the poor. Nutr Rev 2003;61(6 Pt 2):S101–S104, Jun. 58. Brookes G and Barfoot P. GM Rice: Will This Lead the Way for Global Acceptance of GM Crop Technology? ISAAA Briefs No. 28 - 2003. ISAAA: Ithaca, NY. 59. Hammer RE, Pursel VG, Rexroad CE Jr, Wall RJ, Bolt DJ, Ebert KM, Palmiter RD and Brinster RL. Production of transgenic rabbits, sheep and pigs by microinjection. Nature 1985; 315(6021):680–683, Jun 20–26. 60. Pursel VG, Rexroad CE Jr, Bolt DJ, Miller KF, Wall RJ, Hammer RE, Pinkert CA, Palmiter RD and Brinster RL. Progress on gene transfer in farm animals. Vet Immunol Immunopathol 1987;17(1–4):303–312, Dec. 61. Martens MA. Safety evaluation of genetically modified foods. Int Arch Occup Environ Health 2000;73:Suppl:S14–S18, Jun. 62. Kuiper HA, Kleter GA, Noteborn HP and Kok EJ. Assessment of the food safety issues related to genetically modified foods. Plant J 2001;27(6):503–528, Sep. 63. Cockburn A. Assuring the safety of genetically modified (GM) foods: the importance of an holistic, integrative approach. J Biotechnol 2002;98(1):79–106, Sep 11. 64. Kuiper HA, Kleter GA, Noteborn HP and Kok EJ. Substantial equivalence – an appropriate paradigm for the safety assessment of genetically modified foods? Toxicology 2002;181–182: 427–431, Dec 27. 65. Joshi P, Mofidi S and Sicherer SH. Interpretation of commercial food ingredient labels by parents of food-allergic children. J Allergy Clin Immunol 2002;109(6):1019–1021, Jun. 66. Brown JL and Ping Y. Consumer perception of risk associated with eating genetically engineered soybeans is less in the presence of a perceived consumer benefit. J Am Diet Assoc 2003;103(2):208–214, Feb. 67. Alhakami AS and Slovic P. A psychological study of the inverse relationship between perceived risk and perceived benefit. Risk Anal 1994;14(6):1085–1096, Dec. 68. Frewer LJ, Howard C and Shepherd R. Understanding public attitudes to technology. J Risk Res 1998;1:221–235. 69. Falk MC, Chassy BM, Harlander SK, Hoban TJ 4th, McGloughlin MN and Akhlaghi AR. Food biotechnology: benefits and concerns. J Nutr 2002;132(6):1384–1390, Jun. 70. Moseley BE. Safety assessment and public concern for genetically modified food products: the European view. Toxicol Pathol 2002;30(1):129–131, Jan–Feb. 71. Harlander SK. Safety assessments and public concern for genetically modified food products: the American view. Toxicol Pathol 2002a;30(1):132–134, Jan–Feb. 72. Hino A. Safety assessment and public concerns for genetically modified food products: the Japanese experience. Toxicol Pathol 2002;30(1):126–128, Jan–Feb. 73. Harlander SK. The evolution of modern agriculture and its future with biotechnology. J Am Coll Nutr 2002b;21(3 Suppl):161S–165S, Jun. 74. Braun R. People’s concerns about biotechnology: some problems and some solutions. J Biotechnol 2002;98(1):3–8, Sep 11. 75. Romano-Spica V, Orsini M, Riccio F and Laurenti P. Rischi connessi alla produzione e al consumo di alimenti geneticamente modificati. L’Igiene Moderna 1999;112:1971–2000. 76. Lachmann P. Health risks of genetically modified foods. Lancet 1999;354(9172):69, Jul 3. 77. Urbano G, Lopez-Jurado M, Aranda P, Vidal-Valverde C, Tenorio E and Porres J. The role of phytic acid in legumes: antinutrient or beneficial function? Physiol Biochem 2000;56(3): 283–294, Sep. 78. Lack G. Clinical risk assessment of GM foods. Toxicol Lett 2002;127(1–3):337–340, Feb 28.
121 79. Nordlee JA, Taylor SL, Townsend JA, Thomas LA and Bush RK. Identification of a Brazilnut allergen in transgenic soybeans. N Engl J Med 1996;334(11):688–692, Mar 14. 80. Taylor SL and Hefle SL. Genetically engineered foods: implications for food allergy. Curr Opin Allergy Clin Immunol 2002a;2(3):249–252, Jun. 81. Herman EM. Genetically modified soybeans and food allergies. J Exp Bot 2003a;54(386): 1317–1319, May. 82. Lehrer SB, Horner WE and Reese G. Why are some proteins allergenic? Implications for biotechnology. Crit Rev Food Sci Nutr 1996;36(6):553–564, Jul. 83. Herman EM, Helm RM, Jung R and Kinney AJ. Genetic modification removes an immunodominant allergen from soybean. Plant Physiol 2003b;132(1):36–43, May. 84. Nakamura R and Matsuda T. Rice allergenic protein and molecular-genetic approach for hypoallergenic rice. Biosci Biotechnol Biochem 1996;60(8):1215–1221, Aug. 85. Lehrer SB and Reese G. Recombinant proteins in newly developed foods: identification of allergenic activity. Int Arch Allergy Immunol 1997;113(1–3):122–124, May–Jul. 86. Taylor SL. Protein allergenicity assessment of foods produced through agricultural biotechnology. Annu Rev Pharmacol Toxicol 2002b;42:99–112. 87. Helm RM. Food biotechnology: is this good or bad? Implications to allergic diseases. Ann Allergy Asthma Immunol 2003;90(6 Suppl 3):90–98, Jun. 88. Gendel SM. Sequence analysis for assessing potential allergenicity. Ann N Y Acad Sci 2002; 964:87–98, May. 89. Astwood JD, Leach JN and Fuchs RL. Stability of food allergens to digestion in vitro. Nat Biotechnol 1996;14(10):1269–1273, Oct. 90. Martin-Orue SM, O’Donnell AG, Arino J, Netherwood T, Gilbert HJ and Mathers JC. Degradation of transgenic DNA from genetically modified soya and maize in human intestinal simulations. Br J Nutr 2002;87(6):533–542, Jun. 91. Jonas DA, Elmadfa I, Engel KH, Heller KJ, Kozianowski G, Konig A, Muller D, Narbonne JF, Wackernagel W and Kleiner J. Safety considerations of DNA in food. Ann Nutr Metab 2001;45(6):235–254. 92. Carver JD. Dietary nucleotides: effects on the immune and gastrointestinal systems. Acta Paediatr Suppl 1999;88(430):83–88, Aug. 93. Gil A. Modulation of the immune response mediated by dietary nucleotides. Eur J Clin Nutr 2002;56(Suppl 3):S1–S4, Aug. 94. ILSI Europe. Safety assessment of viable genetically modified micro-organisms used in food, ILSI Europe Report Series, ILSI Europe, Brussels, 1999. 95. Sambrook J, MacCallum P and Russell D, Molecular Cloning (third edition): A Laboratory Manual, Cold Spring Harbour Laboratory Press, USA, 2001. 96. Lindahl T. Instability and decay of the primary structure of DNA. Nature 1993;362(6422): 709–715, Apr 22. 97. Lorenz MG and Wackernagel W. Bacterial gene transfer by natural genetic transformation in the environment. Microbiol Rev 1994;58(3):563–602, Sep. 98. Ochman H, Lawrence JG and Groisman EA. Lateral gene transfer and the nature of bacterial innovation. Nature 2000;405(6784):299–304, May 18. 99. Brautigam M, Hertel C and Hammes WP. Evidence for natural transformation of Bacillus subtilis in foodstuffs. FEMS Microbiol Lett 1997;155(1):93–98, Oct 1. 100. Aber W and Linn S. DNA modification and restriction. Ann Rev Biochem 1969;38: 467–500. 101. Malcom AD. Health risks of genetically modified foods. Lancet 1999;354(9172):69–70, Jul 3. 102. Fuchs RL, Heeren RA, Gustafson ME, Rogan GJ, Bartnicki DE, Leimgruber RM, Finn RF, Hershman A and Berberich SA. Purification and characterization of microbially expressed neomycin phosphotransferase II (NPTII) protein and its equivalence to the plant expressed protein. Biotechnology (NY) 1993a;11(13):1537–1542, Dec.
122 103.
104. 105.
106.
107.
108.
109.
110.
111.
112.
113.
114. 115.
116.
Fuchs RL, Ream JE, Hammond BG, Naylor MW, Leimgruber RM and Berberich SA. Safety assessment of the neomycin phosphotransferase II (NPTII) protein. Biotechnology (NY) 1993b;11(13):1543–1547, Dec. Scutt CP, Zubko E and Meyer P. Techniques for the removal of marker genes from transgenic plants. Biochimie 2002;84(11):1119–1126, Nov. Jennings JC, Kolwyck DC, Kays SB, Whetsell AJ, Surber JB, Cromwell GL, Lirette RP and Glenn KC. Determining whether transgenic and endogenous plant DNA and transgenic protein are detectable in muscle from swine fed Roundup Ready soybean meal. J Anim Sci 2003;81(6):1447–1455, Jun. Jennings JC, Albee LD, Kolwyck DC, Surber JB, Taylor ML, Hartnell GF, Lirette RP and Glenn KC. Attempts to detect transgenic and endogenous plant DNA and transgenic protein in muscle from broilers fed YieldGard Corn Borer Corn. Poult Sci 2003;82(3): 371–380, Mar. Schubbert R, Lettmann C and Doerfler W. Ingested foreign (phage M13) DNA survives transiently in the gastro-intestinal tract and enters the bloodstream of mice. Mol Gen Genet 1994;242(5):495–504, Mar. Schubbert R, Renz D, Schmitz B and Doerfler W. Foreign (M13) DNA ingested by mice reaches peripheral leukocytes, spleen, and liver via the intestinal wall mucosa and can be covalently linked to mouse DNA. Proc Natl Acad Sci USA 1997;94(3):961–966, Feb 4. Doerfler W and Schubbert R. Uptake of foreign DNA from the environment: the gastrointestinal tract and the placenta as portals of entry. Wien Klin Wochenschr 1998;110(2): 40–44, Jan 30. Hohlweg U and Doerfler W. On the fate of plant or other foreign genes upon the uptake in food or after intramuscular injection in mice. Mol Genet Genomics 2001;265(2):225–233, Apr. Beever DE and Kemp F. Safety issues associated with the DNA in animal feed derived from genetically modified crops. A review of scientific and regulatory procedures. Nutr Abst & Revs 2000;70:197–204. Phipps RH and Beever DE. Detection of transgenic DNA in bovine milk: Preliminary results for cows receiving a TMR containing Yieldguard TM MON810. Proc Int Anim Agr & Food Sci Conf Indianapolis July 2001, Abst. 476. Einspainer R, Klotz A, Kraft J, Aulrich K, Poser R, Scheagele F and Flachowsky G. The fate of foreign plant DNA in farm animals: A collaborative case-study investigating cattle and chicken fed recombinant plant material. Eur Food Res Technol 2001;212:129–134. Hodgson J. Scientists avert new GMO crisis. Nat Biotechnol 2000;18(1):13, Jan. Damgaard PH, Hansen BM, Pedersen JC and Eilenberg J. Natural occurrence of Bacillus thuringiensis on cabbage foliage and in insects associated with cabbage crops. J Appl Microbiol 1997;82(2):253–258, Feb. Bakshi A. Potential adverse health effects of genetically modified crops. J Toxicol Environ Health B Crit Rev 2003;6(3):211–225, May–Jun.
123
p75 Neurotrophin receptor signaling in the nervous system Yuiko Hasegawa1,2, Satoru Yamagishi1, Masashi Fujitani1,2, and Toshihide Yamashita1,* 1
Department of Neurobiology, Graduate School of Medicine, Chiba University, 1-8-1 Inohana, Chuo-ku, Chiba 260-8670, Japan 2 Department of Anatomy and Neuroscience, Graduate School of Medicine, Osaka University, 2-2 Yamadaoka, Suita, Osaka 565-0871, Japan Abstract. The neurotrophin receptor p75NTR has long been known as a receptor for neurotrophins that promote survival and differentiation. Consistent with the role of neurotrophins, p75NTR is expressed during the developmental stages of the nervous system. However, p75NTR is re-expressed in various pathological conditions in the adult. We now know that p75NTR has the ability to elicit bi-directional signals, that result in the inhibition as well as the promotion of the neurite outgrowth. p75NTR is a key receptor for myelin-derived inhibitory cues that contribute to the lack of regeneration of the central nervous system. Keywords: neurotrophin, growth factor, myelin, peptide, neuron, oligodendrocyte, glia, receptor, p75, signal, G protein, rho, axon, neurite, regeneration, central nervous system, apoptosis, cell survival, synapse, migration.
Discovery of neurotrophins Pictures of isolated embryo-derived nerve cells in culture were presented by Rita Levi-Montalcini [1]. She explained that she was planning to investigate their growth under different experimental conditions. This was a momentous achievement at that time and has lead to many advances in current biotechnology. She presented evidence that an ‘‘agent’’ released by sarcoma fragments could stimulate outgrowth of nerve fibers from sensory and autonomic ganglia in culture [1]. The nature of the growth-stimulating agent led to the discovery of nerve growth factor (NGF). Remarkable events that followed were the use of snake venom to inactivate the agent. It resulted in the recognition that the venom itself had growth-stimulating properties. Since the venom had been derived from salivary glands, it was possible that salivary glands from other animals might also contain a similar factor. This lead to the discovery that adult male mouse salivary glands are an abundant source of the factor [2]. Thanks to the success in developing antibodies to the factor, she eventually identified NGF. However, the factor did not immediately command interest, as it stimulated neurite growth to a pathological degree but had no effect on motor neurons or some neurons in the central nervous system (CNS). The careful crafting of her experiments *Corresponding author: E-mail:
[email protected] BIOTECHNOLOGY ANNUAL REVIEW VOLUME 10 ISSN: 1387-2656 DOI: 10.1016/S1387-2656(04)10005-7
ß 2004 ELSEVIER B.V. ALL RIGHTS RESERVED
124 contributed to growing recognition of the significance of her work, and in 1986 lead to her being awarded the Nobel Prize. One of her important strategies was to examine the growth of neurites from the neurons in culture to test the effects produced by extracts made from the target tissue. It became a powerful tool in the search for other trophic factors [3]. It is also important for developing strategies to overcome the inability of CNS neurons to regenerate, which is described in this chapter. These observations set the stage for a molecular analysis of the mechanism of action of NGF by a number of groups, including those of Levi-Montalcini, Shooter, Thoenen and Barde. The second neurotrophin to be identified, brainderived neurotrophic factor (BDNF), was isolated in 1982 from pooled extracts of porcine brain [4]. The isolation of BDNF helped establish the concept that the fate and the shape of most vertebrate neurons can be regulated by diffusible growth factors, as only a small number of neurons are NGF responsive in the CNS. Neurotrophins and their receptors The neurotrophins are a family of structurally related, secreted proteins that have a profound influence on the development and functioning of the nervous system [5]. Four members of this family have been identified in birds and mammals: NGF, BDNF, neurotrophin-3 (NT-3) and neurotrophin-4/5 (NT-4/5). All consist of dimers of a small basic peptide, held together by disulfide linkages between the conserved cystein residues. The dimer appears as a symmetrical protein with variable, basic regions that determine receptor specificity. Neurotrophins affect essentially all biological aspects of vertebrate neurons, including the survival, differentiation, growth and apoptosis of neurons by using a two-receptor system, which consists of the Trk tyrosine kinases and the p75 neurotrophin receptor (p75NTR) [6,7] (Fig. 1). Three kinds of Trk receptors: Trks A, B and C mediate the biological activity of neurotrophins with the following specificity (Fig. 2). TrkA interacts with NGF, TrkB with BDNF and NT4, while TrkC is the preferred receptor of NT-3, although this molecule could interact with TrkA and B. Trk receptor tyrosine kinases undergo rapid transphosphorylation following ligand binding, leading to a cascade of protein phosphorylations in the cell. The distinctive neuronal deficiencies in mice with null mutations in the trkA, trkB and trkC genes are similar to those observed in mice with null mutations in the NGF, BDNF and NT3 genes, respectively. These findings suggest that the Trks mediate the survival-promoting actions of neurotrophins on developing neurons. p75NTR was the first member of a large family of receptors, which includes both TNF receptors, Fas (Apo-1/CD95), RANK, CD40, and approximately 25 other members to be molecularly cloned [8,9]. The defining motifs of this receptor superfamily are cystein repeats in the extracellular domain, which form the ligand-binding domain.
125
Fig. 1. The two receptor classes of the neurotrophins. p75NTR is a transmembrane glycoprotein receptor of approximately 75 kDa. There are four cystein rich repeats in the extracellular domain. Signaling of p75NTR occurs through cytoplasmic interactors which bind with the juxtamembrane linker region or the helical domain. The Trk receptors are transmembrane glycoproteins of approximately 140 kDa. They are tyrosine kinases with an extracellular ligand binding domain containing multiple repeats of leucine-rich motifs, two cystein clusters, two immunoglobulin-like domains.
Fig. 2. Specificity and cross-talk in the interactions of neurotrophins with their receptors.
All neurotrophins bind to p75NTR with an affinity of 10 9 M [10]. This is a lower affinity than that required for neurotrophin binding on neurons (typically 10 11 M). However, high affinity binding of neurotrophins to neurons cannot simply be explained by the presence of Trk receptors, as the majority of the neurotrophin-binding sites formed by Trk receptors are of low affinity. In addition, NT-3 binds p75NTR with high affinity in embryonic chick sympathetic neurons, whereas this binding does not promote their survival [11]. Thus, it appears that the formation of high affinity binding and specific sites for neurotrophins on neurons are most likely, but not all, a result of the association of p75NTR with Trk receptors [12].
126 Although p75NTR was the first neurotrophin receptor to be identified, its characterization had been in the shade for some years. This is, at least partly, due to the nonenzymatic activities of p75NTR, in sharp contrast to the inherent tyrosine kinase and the potent biological activities of Trk receptors. None of the receptor family members exhibit any intrinsic catalytic activity, and they pass their signals by associating with, or dissociating from, cytoplasmic interactors. In addition, studies on p75NTR have been complicated by the fact that it can interact with Trk receptors [13], and by the finding that its signaling capacity is modified by the coincident activation of Trk receptors. Nontheless, recent years have seen the emergence of a consensus regarding the signaling pathways activated by p75NTR and of potential biological function, and have lead to the elucidation of a number of p75NTR-interacting proteins [14]. We review here p75NTR signaling in the nervous system, especially focusing on the recent discovery that it transduces the signal from several myelin-derived inhibitors of neurite outgrowth, which are involved with the inability of CNS neurons to regenerate after injury. Interactions between p75NTR and the Trk receptors The present data suggest that p75NTR has two main physiological functions: modulating Trk receptor signaling and initiating autonomous signaling cascades. The precise molecular mechanisms that allow p75NTR to enhance NGF binding to TrkA and increase TrkA responsiveness to NGF remain uncertain, but two hypotheses have been put forward. First, p75NTR acts as a co-receptor that binds NGF and either concentrates it locally or presents it to TrkA in a favorable binding conformation. A number of studies have shown that disrupting NGF binding to p75NTR inhibits NGF-induced TrkA activation [15–18]. Complementary studies have shown that a mutant form of NGF, which binds TrkA but does not bind p75NTR, is less effective than the wild type NGF in activating TrkA in cells where the two receptors are co-expressed [15,19]. These findings support the notion that NGF binding to p75NTR is necessary to facilitate TrkA activation in response to low levels of NGF. Second, p75NTR has an allosteric effect on TrkA that confers high-affinity NGF binding to the TrkA receptor irrespective of NGF binding to p75NTR. In support of this hypothesis, high affinity NGF binding sites can be generated when TrkA is co-expressed with either a p75NTR mutant deficient in neurotrophin binding, or using a chimaeric receptor consisting of the extracellular domain of EGF receptor and the cytoplasmic domain of p75NTR [20]. Further work will be required to reconcile these two models. Diverse function of p75NTR One of the most prominent biological functions of p75NTR (Fig. 3) may be that it induces cell death, as it contains a death domain sequence distantly related to the
127
Fig. 3. p75NTR is involved in several different biological activities. p75NTR regulates both cell death and cell survival. Pro NGF seems to be the most effective activator of cell death. Several interactors have been found to associate with these actions presumably depending on the cellular context. Some of the proteins that interact with p75NTR block cell cycle progression. p75NTR also regulates axon elongation by regulating the activity of RhoA both during development and after lesion.
intracellular domains of the Fas and TNF receptors [6,7]. This domain consists of a bundle of six short a-helices spanning 90 amino acids that form a novel type of fold [21]. Direct evidence of p75NTR-mediated apoptosis was first described in 1993 by Bredesen and his colleagues who reported that p75NTR overexpression facilitates apoptosis, which is inhibited by NGF [22]. The cell death may be caused by the spontaneous signaling that occurs when the receptor multimerizes. In contrast, many in vitro studies have subsequently demonstrated that neurotrophins binding to p75NTR induce apoptosis. For example, NGF binding to p75NTR elicits apoptosis of differentiated rat oligodendrocytes [23–25], Schwann cells [26], hepatic stellate cells [27], sympathetic neuron precursor (MAH) cells [28], mesodermal cells [29], chick isthmo-optic nucleus neurons [30], trigeminal mesencephalic sensory neurons [31], and retinal ganglion
128 cells [32]. Therefore, the precise ligand dependency of this phenomenon is still in controversy and there are some reports that indicate p75NTR-promoted survival of the neurons. These in vitro studies may represent the extreme supraphysiological limits of p75NTR action, and the physiological role of the receptor should be determined at least in vivo. Basal forebrain cholinergic neurons express p75NTR at high levels, and recent evidence shows that there are indeed significantly more cholinergic neurons in the complete p75NTR mutant mice [33]. Cell death in the avian retina in the developmental stages is reduced following the addition of the antibodies against NGF or the extracellular domain of p75NTR, which is expected to act as an inhibitor of endogenous p75NTR [34,35]. Mice that transgenically overexpress the intracellular domain of p75NTR show reductions in cortical, sympathetic and sensory neurons [36]. These findings show that p75NTR activates the cell death pathway in vivo. Neurotrophins are synthesized as precursors or pro-proteins, and proteolytic cleavage is necessary for the production of the mature neurotrophins. Recent evidence shows that uncleaved pro-NGF binds to p75NTR with high affinity and causes cell death at significantly lower concentrations than does mature NGF [37]. The binding affinity of pro-NGF to TrkA is not as strong as that for mature NGF, suggesting that proteolytic processing is crucial in determining the signaling elicited by these two kinds of receptors. This aspect might be relevant, as increased levels of pro-NGF are found in the brains of patients of Alzheimer’s disease [38]. p75NTR also binds non-neurotrophin ligands, which include the neurotoxic prion protein fragment PrP (26–106) and the Ab-peptide of the amyloid precursor protein (APP) [39–42]. These peptides induce the cell death via p75NTR in culture. p75NTR is also a receptor for a glycoprotein of the rabies virus envelope, allowing the virus to enter the nervous system [43], and for the invertebrate ligand cystein-rich neurotrophic factor [44]. Other p75NTR-mediated activities have been proposed, including enhancing axonal outgrowth [45–47], influencing Schwann cell migration [48], promoting myelin formation [49], modulating synaptic transmission [50] and regulating the function of sensory neurons [51] and calcium currents [52]. Most of these biological activities can be attributed to neurotrophins binding to p75NTR. Recent work also implicates p75NTR in the regulation of axonal elongation that is elicited by several myelin-derived proteins that may contribute to the lack of regeneration of the injured adult CNS. Finally, it is noted that a short isoform of p75NTR has also been found. The transmembrane and intracellular domains of the short isoform are identical to that of the full-length p75NTR, but the short isoform lacks three of the four cystein rich repeats in the extracellular ligand binding domain [53]. Thus, this short isoform does not bind neurotrophins. Mice carrying a mutation in the p75NTR gene still express this short isoform, although the full-length of p75NTR is completely deleted [54]. The mice generated were the initial p75NTR-targeted
129 ones and have proven very useful for determining p75NTR action. However, the phenotype of the mice is different from that of the complete p75NTR knockout mice [53], which shows up to 40% lethality, presumably as a result of blood vessel defects. The involvement of this short isoform in the formation of the blood vessels is not clear. This numerous amount of work uncovers surprisingly diverse function of p75NTR. p75NTR is involved in the pathogenesis of neurological diseases Although the expression of p75NTR is developmentally regulated in the nervous system, marked increases in p75NTR levels are observed under certain pathological conditions. In rats subjected to pilocarpine-induced seizure, expression of p75NTR is induced in entorhinal, piriform and hippocampal cortices, and its expression is associated with the cell death [55]. In the dorsal root ganglia, reduction of the p75NTR levels by antisense oligonucleotides prevents the loss of axotomized neurons [56]. Similarly, motor neuron loss occurring after transection of the neonatal facial nerve is reduced in mutant mice that carry a mutation in the p75NTR gene [57,58]. Conversely, administration of NGF into transected neonatal facial nerve of animals produces increased cell death [59]. p75NTR is re-expressed in neurons, at levels comparable to those seen during the developmental stages. Upregulation of p75NTR is also observed in the cerebral cortex in Alzheimer’s disease [60]. In a mouse model of amyotrophic lateral sclerosis, in which a mutant form of superoxide dismutase is overexpressed, there is re-expression of p75NTR in lumbar motor neurons, which are destined to die 4 months after birth [61]. In fact, expression of p75NTR is observed in motoneurons in the cervical spinal cords of patients with amyotrophic lateral sclerosis. After axotomy, cortical spinal neurons strongly reexpress p75NTR three days after lesion, when the neuronal death occurs. Activation of p75NTR by NT3 causes the death of these neurons, as shown by experiments employing antibodies against p75NTR or to NT3 [62]. These findings establish a link between p75NTR and neuronal death in neurological disorders. Signal transduction through p75NTR The first indication of the signaling function of p75NTR was the observation that p75NTR mediates sphingomyelin hydrolysis and production of ceramide following neurotrophin binding [63]. Ceramide production is known to follow TNF binding to its receptor and to lead to NF-kB activation [64]. Likewise, in Schwann cells expressing p75NTR but not catalytic Trk receptors, NF-kB activation was observed following the addition of NGF [65]. As is the case with TNF receptor 1 signaling, it appears that activation of NF-kB prevents cell death [66,67].
130 In contrast to NF-kB activation that seems to mediate cell survival, there should be signals that elicit programmed cell death. These signals involve caspase activation, as well as Bax/Bad, Bcl-2 and Bcl-xL [68]. Inhibition of Jun kinase (JNK) activity blocks apoptosis through p75NTR, suggesting that JNK plays a significant role in p75NTR-mediated apoptosis [69]. However, activation of JNK does not always explain p75NTR-mediated apoptosis, and there is evidence demonstrating that p53 and the p53-related protein p73 play a role [70]. As p75NTR has no intrinsic catalytic activity, cytoplasmic interactors should be recruited to produce the signals. A number of interactors have been identified mainly by yeast two-hybrid screening, trying to explain multiple intracellular signals (Fig. 4). The interactors, presumably involved in the cell death, include a ubiquitously expressed zinc finger protein designated as NRIF (neurotrophin receptor-interacting factor) [71]. The retinas of the NRIF / mice show reduced cell death, and this reduction is quantitatively similar to that seen in mice carrying a mutation in the p75NTR gene. As it localizes in the nucleus as well as in the cytoplasm, it is suggested that the neurotrophin-binding to p75NTR facilitates release from the intracellular domain of p75NTR, resulting in the translocation to the nucleus. A protein named NADE, the p75NTR-associated cell death executor, is found to be associated with p75NTR when activated by NGF, but not by BDNF,
Fig. 4. p75NTR recruits cytoplasmic interactors to signal. Several interactors of p75NTR intracellular domain have been identified that mediate different biological functions. NADE, NRIF and NRAGE associate with cell death. NRIF, NRAGE and SC-1 are involved in cell cycle arrest. The GTPase RhoA is a regulator of axon elongation. TRAFs associate with p75NTR to activate NF-kB. The physiological relevance of some interactors is not clear.
131 NT3 or NT4/5 [72]. It also seems to contribute to the cell death inducing activity of p75NTR. The NRAGE (neurotrophin receptor-interacting MAGE) homolog was also identified as an interactor of p75NTR, and was shown to mediate NGF-dependent apoptosis in sympathetic neuron precursor cells [28]. When NRAGE is overexpressed in the transfected cells, it causes cell cycle arrest, suggesting that p75NTR may play a role in the control of growth. Nestin-positive neural stem cells proliferate at a higher rate than the wild type cells in the absence of the fulllength p75NTR, and the activation of p75NTR by BDNF promotes differentiation into neurons [73]. Therefore, the differentiation promoting effect of p75NTR might be mediated by NRAGE, which should be the subject of future studies. Although NRIF is a zinc finger protein, another zinc finger protein, the Schwann cell factor-1 (SC-1), is involved in cell cycle arrest [74]. In transfected COS cells, the localization of SC-1 changes from the cytoplasm to the nucleus following NGF stimulation, but not BDNF stimulation. Expression of SC-1 in the nucleus results in a loss of BrdU incorporation. Several kinases have been shown to interact with the intracellular domain of p75NTR. A p75-associated kinase [75], as well as ERK1 and ERK2 [76], are also the interactors of p75NTR, although the functional significance of the interactions is not clear. A variant of the b catalytic subunit of cAMP-dependent protein kinase (PKACb) is shown to be a p75NTR interacting protein, which phosphorylates p75NTR at serine 304 [77]. Intracellular cAMP in cerebellar neurons is transiently accumulated by ligand binding to p75NTR. Activation of cAMP-PKA is required for translocation of p75NTR to lipid rafts, and for biochemical and biological activities of p75NTR, such as inactivation of Rho and neurite outgrowth. Therefore, PKACb may be necessary for the proper recruitment of the activated p75NTR to lipid rafts, structures that represent specialized signaling organelles.
Axon elongation In good correlation with the function of neurotrophins, p75NTR is expressed abundantly in neurons during developmental stages. Motor neurons in the spinal cord, most sympathetic and sensory neurons in the peripheral nervous system, as well as cerebellar Purkinje cells and retinal ganglion cells all express p75NTR at high levels during the outgrowth of axons [78–82]. In dendrite-bearing neurons, p75NTR is also expressed during the time of dendritic arborization. Some neurons markedly up-regulate p75NTR after lesion or seizure [55,83]. Mice carrying a mutation in the p75NTR gene show deficits in the outgrowth of thoracic intercostal nerves and forelimb motor axons [45] as well as retarded axonal arborization of the opthalamic branch [84]. As adults, these mice have deficits in sensory and sympathetic target innervation [85]. Notably, these mice show a marked reduction of visual cortex innervation by thalamic axons, which are thought to use the pathway pioneered by the subplate axons as a scaffold,
132 and the growth cones are smaller and have a markedly reduced number of filopodia [86]. Although these in vivo biological effects are likely to be due to reduced Trk activation, there are indications, at least in vitro, that the ligands binding to p75NTR promotes axon outgrowth. NGF stimulates neurite outgrowth from embryonic rat hippocampal neurons and chick ciliary neurons [45,87], which express p75NTR but not TrkA. These findings suggest that p75NTR plays some primary roles in the developmental stages as well as during pathological states. The key molecule that regulates these effects may be small GTPase Rho. Rho GTPases are a family of highly related proteins that are best known for their effects on the actin cytoskeleton. The representatives of the Rho family are Rho, Rac, and Cdc42. Several isoforms of Rho have been reported, and in neurons, RhoA is expressed at higher levels than RhoB and RhoC [88]. RhoA was shown to interact with the intracellular domain of p75NTR [45]. Interestingly, overexpression of p75NTR in 293 cells results in the activation of RhoA, whereas ligand binding to p75NTR abolishes the activation (Fig. 5). Inactivation of RhoA is suggested to be implicated in the neurite outgrowth of chick ciliary neurons, as incorporation of the active mutant of RhoA into the cells attenuates the effect of NGF. This suggestion is substantiated by the fact that blocking Rho activity
Fig. 5. p75NTR is a bi-directional regulator of RhoA. RhoA was shown to interact with the intracellular domain of p75NTR [45]. Overexpression of p75NTR in 293 cells results in the activation of RhoA, presumably by the clustering of the receptor. Ligand binding to p75NTR, however, abolishes the activation of RhoA.
133 with the botulinus toxin C3 mimicks the effects of NGF. It should be noted that neurite outgrowth by the inactivation of Rho is not specific to ciliary neurons [89,90]. The missing link between p75NTR and inactivation of RhoA was shown recently, where ligand binding to p75NTR was demonstrated to increase intracellular cAMP (Fig. 6) [91]. NGF induces inactivation of RhoA in cerebellar neurons and 293T cells, and this effect is PKA dependent. PKA phosphorylates many target proteins, and one such target identified is RhoA. When serine 188 is phosphorylated, RhoA becomes inactive [92]. Taken together, it is possible
Fig. 6. Mechanisms of the axon elongation by p75NTR and neurotrophins. The ligand binding to p75NTR increases intracellular cAMP [77]. Inactivation of RhoA by neurotrophins binding to p75NTR in cerebellar neurons and 293T cells is PKA dependent. PKA phosphorylates serine 188 of RhoA, leading to inactivation of RhoA [92].
134 that inactivation of RhoA is the downstream component of cAMP-PKA. However, another interpretation of the data is that the inhibition of the PKA signal blocks the translocation of the receptor to lipid rafts and might result in the failure of transduction of the downstream signal. Inhibition of axon elongation by p75NTR Recent reports indicate that p75NTR is involved in the inhibition of axonal elongation by myelin, in sharp contrast with the notion that p75NTR contributes to promotion of axon elongation. Transgenic mice were generated in which NGF was expressed by astrocytes in the CNS under the control of the GFAP promoter. Sympathetic axonal sprouting into the brains was observed in these mice, however much more axonal sprouting occured if a mutation was inserted into the p75NTR gene. Interestingly, abberant axonal elongation is observed in myelin-rich areas where these axons would normally not grow [93]. The hippocampus of mice carrying a mutation in the p75NTR gene is hyperinnervated by cholinergic afferents [94]. These seemingly contradictory findings suggest that p75NTR transduces bi-directional signals that elicit inhibition of neurite growth as well as axonal outgrowth. Recent surprising reports uncover the molecular mechanism of these biological effects, providing molecular targets for the development of the therapies against injuries to the CNS. Inability of the adult CNS to regenerate Injury to the adult CNS is devastating because of the inability of central neurons to regenerate correct axonal and dendritic connections. It is now well established that axons of the adult CNS are capable of only a limited amount of regrowth after injury, and that an unfavorable growth environment plays a major role in the lack of regeneration. In 1911, F. Tello described the first successful transplantation of a peripheral nerve into the adult mammalian CNS [95]. Previously denervated sciatic nerve pieces were implanted into the cortex of rabbits, and he observed fascicles and individual nerve fibers that invaded into these peripheral nerves 2 to 4 weeks after surgery. He and Ramon y Cajal concluded that peripheral nerve Schwann cells reacted to the loss of their axons by the synthesis of attractive and neuritepromoting cues [96]. They suggested further that CNS glia would be devoid of such a reaction. The morphological features of axonal injury and degeneration in vivo were elegantly described by Ramon y Cajal. Aguayo’s group in the early 1980s showed that many neurons can regenerate over long distances if offered a peripheral nerve as a substrate [97–99]. That CNS myelin is involved in the prevention of axonal regeneration in adult mammals was first suggested by Berry [100]. He pointed out that nonmyelinated axons in the CNS would regenerate after chemical axotomy if damage did not occur to nearby myelinated fibers, but not after mechanical axotomy, which
135 damages myelinated fibers. As damage to the myelinated fibers leads to the release of degeneration products of CNS myelin, it was proposed that this damage would be inhibitory to axonal growth. Subsequently, Schwab’s group tested this hypothesis by exposing perinatal DRG or sympathetic neurons to optic and sciatic nerve explants of adult rats in the presence of NGF. However, they observed few or no axons in the optic nerves during 2 weeks in culture, whereas abundant nerve fibers invaded into the sciatic nerves [101]. As repeated freezing and thawing of the explants prior to culture gave the same results, the absence of neurite outgrowth in the adult optic nerve explants results from an intrinsic property of the adult CNS tissue rather than to reactions to the lesion or the culture condition. They postulated that myelin from the adult CNS is an inhibitory substrate for neurite outgrowth.
Three distinct myelin proteins inhibit axon growth Nogo Initially, biochemical analysis of rat brain myelin showed two protein constituents of MW 35 kDa and 250 kDa which were potent inhibitors of neurite outgrowth [102]. One monoclonal antibody called inhibitor-neutralizing antibody (mAB IN-1) was obtained and used extensively for subsequent in vitro and in vivo experiments. The inhibitory activity of a crude myelin extract was decreased to approximately 50% by this antibody, and that of purified bovine NI-220 (the homolog of rat NI-250) was decreased to 0–20% of initial levels by the antibody [103]. Starting with large amounts of bovine spinal cord, Schwab’s group succeeded in purifying the bovine homolog bNI-220. The corresponding cloned cDNA has the characteristics of a type 2 membrane protein and is derived from a gene which gives rise to three mRNAs [104–106]. This gene is designated Nogo. The three splice variants of Nogo are called NogoA, NogoB and NogoC, the latter two of which are widely expressed outside the CNS. NogoA possesses a unique amino-terminal region not shared by NogoB and NogoC. The two most strongly predicted transmembrane domains are separated by the 66-residue extracellular or lumenal loop, called Nogo-66. Nogo-66 causes growth cone collapse [107]. The Nogo-A specific aminoterminal region is also inhibitory for neurite outgrowth, and prevents the spreading of fibroblasts. Immnunohistochemical studies have shown that Nogo proteins are present in neuronal cell bodies and axons as well as oligodendrocytes. Specifically, Nogo-A is most strongly expressed in oligodendrocytes in the white matter, although it was also detected in neuronal perikarya including those in the cerebral cortex, spinal motor neurons and DRG neurons, as well as in axons [108]. Whether neuronal Nogo-A plays a role in axonal growth or guidance in the developing nervous system should be determined in the future.
136 MAG Myelin-associated glycoprotein (MAG) is a transmembrane protein of the immunoglobulin superfamily, found in both peripheral and CNS myelin, where it plays a role in the formation and maintenance of myelin sheath. MAG was identified as the first myelin-derived growth inhibitory protein by two groups. McKerracher et al. detected MAG inhibitory activity in myelin after extraction with octylglucoside, fractionation by ion exchange chromatography, and screening for inhibitory activity [109]. Filbin’s group demonstrated that MAG that was ectopically expressed in CHO cells inhibits neurite outgrowth [110]. Interestingly, MAG is a bifunctional regulator of axon growth. MAG can stimulate neurite outgrowth of young neurons, where endogenous levels of cAMP may be critical for this effect of MAG [111]. A soluble form of MAG, capable of inhibiting neurite outgrowth from P6 DRG neurons, is released from damaged CNS myelin [112]. Soluble MAG constitutes the great majority of the neurite outgrowth inhibiting factors released from damaged myelin. OMgp Oligodendrocyte-myelin glycoprotein (OMgp) is the most recently identified protein that is an inhibitory component of myelin. In the course of the isolation of MAG as an inhibitory protein, Braun’s group observed two peaks of inhibitory activity, with MAG present in the first peak. The group led by He separated the inhibitory protein in the second peak, and identified OMgp [113]. They identified OMgp as an inhibitor with the hypothesis that any GPI-anchored myelin proteins act as regeneration inhibitors [114]. OMgp, which is abundant in myelin, has potent growth cone collapsing and neurite outgrowth inhibitory activities. The available evidence suggests that OMgp is principally a neuronal protein with a limited amount of OMgp being expressed by oligodendrocytes [115], whereas the functions of neuronal OMgp have not been explored. The three inhibitors, Nogo, MAG and OMgp, have similar inhibitory activity and distribution in the myelin sheath, suggesting that all of them probably contribute to growth inhibition in the adult CNS. The Nogo receptor A protein that binds Nogo-66 was identified with high affinity by an alkalinphosphatase-fusion protein expression screening strategy [116]. Transfection of the cDNA encoding this putative receptor into retinal ganglion cells at a developmental stage when they otherwise are unresponsive to Nogo-66 promotes growth cone collapse by GST-Nogo-66. Mutated forms of the receptor eliminates growth inhibition by Nogo-66. Therefore, this protein is a receptor for Nogo-66 (NgR). NgR is a glycosylphosphatidylinositol (GPI) anchor protein
137 that attaches to the outer leaflet of the plasma membrane, and is expressed in the CNS neurons as well as their axons [117,118]. As release of GPI-anchored proteins by phosphatidylinositol-specific phospholipase C from embryonic DRG results in the abolishment of growth cone collapse in response to Nogo-66, NgR mediates the signal from Nogo-66 in at least these neurons. Surprisingly, two other inhibitory components, MAG and OMgp, also bind to NgR. In an expression screening for NgR-interecting proteins, Strittmatter’s group isolated MAG as a binding partner for NgR [119]. Filbin’s team identified it by direct binding studies based on the similarity in molecular weight to candidates revealed in a previous characterization of MAG binding proteins [120]. NgR was also obtained by screening for proteins that bind to OMgp [114]. Therefore, NgR is necessary for inhibition of axon growth by MAG, Nogo-66 and OMgp in vitro, and ectopic expression of NgR leads insensitive neurons to become sensitive to these myelin-derived proteins. These findings bring these various molecules to an intersection at the level of NgR. Surprisingly, NgR expression is not very altered by axotomy [121], which suggests that it has a physiological role in the intact CNS, unrelated to injury and regeneration. Moreover, Niederost et al. claim that phospholipase treatment, to remove NgR and other GPI-linked cell surface molecules, does not block all of the inhibitory effects of MAG on neurite outgrowth from cerebellar granule cells grown on polylysine [122]. Their results contrast markedly with the findings of others [119,120]. From the perspective of trying to develop a therapeutic approach, it is important to note that a fragment of Nogo-66 binds to NgR as a high affinity antagonist [123]. The antagonist peptide, NEP1-40, reduces endogenous inhibitory activity, to promote sprouting of corticospinal tract axons, long distance growth and functional recovery. p75NTR transduces the signal from MAG, Nogo and OMgp Although NgR is a binding partner for MAG, Nogo-66 and OMgp, the GPIlinked nature of NgR suggests that there may be a second receptor subunit that spans the plasma membrane and mediates signal transduction. Identification of the signal transducer of these proteins came from the experiments by Filbin’s team showing that nerve cells pretreated with neurotrophins overcome MAG’s power to squelch growth [124]. The finding hinted at a connection between p75NTR and MAG. Perhaps MAG cannot signal when neurotrophins occupy p75NTR, we reasoned. To learn whether the receptor might be playing both sides – as a growth stimulant and suppressor – we tested whether MAG requires p75NTR to relay its message. MAG’s effect on nerve elongation in normal mice and in animals lacking the receptor was examined. Without p75NTR, MAG’s clout in blocking nerve extension withered [125]. Colocalization of p75NTR and MAG binding is seen in neurons. These results show that p75NTR may be a signal
138
Fig. 7. Signal transduction by the complex of the Nogo-66 receptor (NgR) and p75NTR. MAG, Nogo-66 and OMgp are ligands for the Nogo-66 receptor (NgR), and are all expressed by oligodendrocytes. These ligands do not share any recognized protein domains. p75NTR interacts with NgR as well as ganglioside GT1b, and mediates the inhibitory signaling of these myelinderived proteins by activating RhoA.
transducer of MAG (Fig. 7). Next, we looked for conspirators in the molecules’ ability to suppress nerve cell extension. p75NTR’s talent for eliciting nerve growth relies on Rho. As Rho shuts down and nerves branch out when p75NTR binds neurotrophins, we reasoned that p75NTR’s nerve-constraining alter ego also relies on Rho. To test the idea, Rho’s function in cells was crippled and then, MAG susceptibility vanished. Then we exposed normal and p75NTR-deficient cells to MAG and measured the amount of active Rho by affinity precipitation. MAG activates Rho only in the presence of the receptor, verifying p75NTR’s part in the effects. Extrapolating from the observations that MAG is a ligand for NgR, the possibility that p75NTR associates with NgR to form a receptor complex for MAG, Nogo and OMgp was tested [126,127]. He’s group as well as Poo’s group demonstrated that at least a fraction of p75NTR binds with NgR using co-immunoprecipitation experiments. Postnatal cerebellar neurons from mice carrying a mutation in the p75NTR gene are insensitive to GST-Nogo-66 and OMgp-AP [126]. The inhibitory activity of these proteins in cerebellar granule neurons is decreased by the ectopic expression of a dominant negative form of p75NTR that lacks a cytoplasmic domain. Soluble p75NTR-Fc fusion protein also attenuates the effects. These observations not only suggest that p75NTR is required for the inhibitory activity of these myelin-derived proteins, but also
139 provide a potent molecular target for developing therapeutic agent against injuries to the CNS. The p75NTR knockout mouse has become a prime target for regeneration experiments. However, it should be noted that p75NTR has diverse functions, including bi-directional signals regulating axon growth, thus alarming that silencing all the functions of p75NTR would be a less attractive treatment strategy. Axon growth inhibition signals from p75NTR Downstream from the receptor complex of p75NTR and NgR, Rho appears to be a key intracellular effector for growth inhibitory signaling by myelin. In neurons, myelin and MAG inhibit growth, that is abolished by the botulinus toxin C3 which inactivates Rho [88]. More specifically, it is directly shown that Rho is activated by MAG-Fc in the cerebellar granule neurons shown by the affinity precipitation of GTP-bound form of RhoA [125]. The precise mechanism of action of p75NTR is suggested by the finding that p75NTR releases Rho from Rho guanine nucleotide dissociation inhibitor (RhoGDI) (Fig. 8) [128], thus eliciting activation of Rho. Rho-GDI is an essential part of the signaling mechanism that suppresses the activity of Rho. Rho proteins are regulated either by enzymes that enhance GTP binding and activity (guanine nucleotide exchange factors) or by proteins that increase the hydrolysis of GTP (GTPase activating proteins) and thus decrease activity. Rho is kept in an inactive state in cells by Rho-GDI [129]. Rho-GDI inhibits the activity of Rho by binding to and sequestering Rho in the cytoplasm, by inhibiting the formation of active RhoGTP, and by blocking the binding of Rho to its effectors. As mentioned above, RhoA was previously identified as an interactor of p75NTR by the yeast two-hybrid screening method [45]. As only the wild type of RhoA, which is predominantly in a GDP-bound form, but not the constitutive active form of RhoA, interacts with p75NTR, as shown by the co-immunoprecipitation assay, it is suggested that activation of RhoA is dependent on a direct interaction of RhoA and p75NTR. Interestingly, overexpression of the intracellualr domain of p75NTR as well as the full-length p75NTR activates RhoA ligand independently, suggesting that p75NTR may be a constitutive activator of Rho. However, the intracellular domain of p75NTR shows no similarity with the Dbl homology domain, which is shared by conventional guanine nucleotide exchange factors, demonstrating that p75NTR does not mediate the process of exchange reaction, in which GDP is replaced by GTP. Instead, direct interaction of the Rho-GDI with p75NTR initiates the activation of RhoA, by promoting the release of prenylated RhoA from RhoGDI, enabling RhoA to be activated by guanine nucleotide exchange factors. These findings establish Rho as a key player in inhibiting the regeneration of the CNS, and launched a new wave of studies that aimed to promote regeneration of injured axons by modulating this inhibitory pathway. For example, an inhibitor of Rho kinase, a downstream effector of Rho, called Y-27632 has
140
Fig. 8. Mechanisms of axon growth inhibition by p75NTR. In the absence of MAG or Nogo, growth and regeneration occur as a result of Rho-GDI-induced suppression of Rho activity. Rho-GDI maintains Rho in an inactive state by binding, and prevents Rho from interacting with its effectors. Activation of p75NTR promotes dissociation of Rho-GDI from RhoA, allowing RhoA to become activated through the exchange of GDP for GTP. The activated RhoA then interacts with its signaling molecules to elicit axon growth inhibition in some neurons.
been used to probe the role of Rho in growth inhibiting signaling [130,131]. Treating neurons with C3 transferase, a bacterial endotoxin that inactivates Rho, or with Y-27632, promotes growth on inhibitory substrates. Intriguingly, a peptide that blocks the pathway elicited by MAG, Nogo and OMgp was found [128]. The binding region of Rho-GDI on p75NTR was identified as the fifth alpha helix in the p75NTR intracellular domain. The short sequence of the fifth helix is similar to mastoparan, a 14-residue peptide of wasp venom that is capable of activating Rho [132]. A peptide ligand bonded to this region was previously reported by Ilag’s group [133] by screening a combinatorial library using a variation of the selectively-infective phage method. This peptide, designated Pep5, inhibits the interaction of p75NTR with Rho-GDI in vitro and in vivo. The inhibitory peptide completely abolishes the effects
141
Fig. 9. Hypothesis of p75NTR function. p75NTR positively and negatively regulates axon elongation. Balancing mechanisms of the opposite cues through p75NTR might be necessary for plasticity and axon-glial interaction as well as neural development and regeneration.
mediated by MAG or Nogo-66 in adult DRG neurons and postnatal cerebellar granule neurons [128], establishing the Rho-GDI-p75NTR association as an important mecahnism of p75NTR-induced suppression of axon growth by myelin proteins. An especially notable aspect is that the peptide does not inhibit other functions of p75NTR, such as axon elongation or cell death by neurotrophins.
142 Conclusion One of the most striking actions of neurotrophins – the feature for which NGF was originally named – is the promotion of neurite outgrowth, whereas that of the myelin-derived proteins is the inhibition of it. Two rivers originated from Levi-Montalcini and the collaborative work by Tello and Ramon y Cajal met during the investigations of p75NTR function, both, in part, attempting to elucidate the chain of events initiated by p75NTR that culminates in changes in the cytoskeleton and axonal growth. At first glance, p75NTR seems to reveal Dr. Jekyll and Mr. Hyde sides, as the latter work is predominantly related to pathological conditions. However, new findings suggest that myelin-derived inhibitors are not only important for regeneration, but have an important role in regulating plasticity and axon-glial interactions (Fig. 9). Thus, balancing mechanisms of these opposite cues would be necessary for the proper regulation of the development and maintenance as well as regeneration of the nervous system. If we can further understand the precise mechanisms of p75NTR at the molecular level, then it would open the door for the development of new biotechnological strategies to promote CNS regeneration. References 1. Levi-Montalcini R. Neuronal regeneration in vitro. In: Regeneration in the Central Nervous System, Windle WF (ed), Springfield, Thomas, CC, 1955, pp. 54–65. 2. Levi-Montalcini R. Nerve Growth Factor 35 years later. Science 1987;237:1154–1162. 3. Barde YA, Edgar D and Thoenen H. New trophic factors. Ann Rev Physiol 1983;45:601–612. 4. Barde YA, Edgar D and Thoenen H. Purification of a new neurotrophic factor from mammalian brain. EMBO J 1982;1:549–553. 5. Lewin GR and Barde YA. The physiology of neurotrophins. Annu Rev Neurosci 1996;19: 289–317. 6. Kaplan DR and Miller FD. Neurotrophin signal transduction in the nervous system. Curr Opin Neurobiol 2000;10:381–391. 7. Lee FS, Kim AH, Khursigara G and Chao MV. The uniqueness of being a neurotrophin receptor. Curr Opin Neurobiol 2001;11:281–286. 8. Johnson D, Lanahan A, Buck CR, Sehgal A, Morgan C, Mercer E, Bothwell M and Chao M. Expression and structure of the human NGF receptor. Cell 1986;47:545–554. 9. Radeke MJ, Misko TP, Hsu C, Herzenberg LA and Shooter EM. Gene transfer and molecular cloning of the rat nerve growth factor receptor. Nature 1987;325:593–597. 10. Rodriguez-Tebar A, Dechant G and Barde YA. Binding of brain-derived neurotrophic factor to the nerve growth factor receptor. Neuron 1990;4:487–492. 11. Dechant G, Tsoulfas P, Parada LF and Barde YA. The neurotrophin receptor p75 binds neurotrophin-3 on sympathetic neurons with high affinity and specificity. Neuroscience 1997; 17:5281–5287. 12. Mahadeo D, Kaplan L, Chao MV and Hempstead BL. High affinity nerve growth factor binding displays a faster rate of association than p140trk binding. Implications for multisubunit polypeptide receptors. J Biol Chem 1994;269:6884–6891. 13. Bibel M, Hoppe E and Barde YA. Biochemical and functional interactions between the neurotrophin receptors trk and p75NTR. EMBO J 1999;18:616–622.
143 14. Dechant G and Barde YA. The neurotrophin receptor p75(NTR): novel functions and implications for diseases of the nervous system. Nat Neurosci 2002;5:1131–1136. 15. Barker PA and Shooter EM. Disruption of NGF binding to the low affinity neurotrophin receptor p75LNTR reduces NGF binding to TrkA on PC12 cells. Neuron 1994;13:203–215. 16. Verdi JM, Birren SJ, Ibanez CF, Persson H, Kaplan DR, Benedetti M, Chao MV and Anderson DJ. p75LNGFR regulates Trk signal transduction and NGF-induced neuronal differentiation in MAH cells. Neuron 1994;12:733–745. 17. Clary DO and Reichardt LF. An alternatively spliced form of the nerve growth factor receptor TrkA confers an enhanced response to neurotrophin 3. Proc Natl Acad Sci USA 1994;91:11133–11137. 18. Lachance C, Belliveau DJ and Barker PA. Blocking nerve growth factor binding to the p75 neurotrophin receptor on sympathetic neurons transiently reduces trkA activation but does not affect neuronal survival. Neuroscience 1997;81:861–871. 19. Ryden M, Hempstead B and Ibanez CF. Differential modulation of neuron survival during development by nerve growth factor binding to the p75 neurotrophin receptor. J Biol Chem 1997;272:16322–16328. 20. Esposito D, Patel P, Stephens RM, Perez P, Chao MV, Kaplan DR and Hempstead BL. The cytoplasmic and transmembrane domains of the p75 and Trk A receptors regulate high affinity binding to nerve growth factor. J Biol Chem 2001;276:32687–32695. 21. Liepinsh E, Ilag LL, Otting G and Ibanez CF. NMR structure of the death domain of the p75 neurotrophin receptor. EMBO J 1997;16:4999–5005. 22. Rabizadeh S, Oh J, Zhong LT, Yang J, Bitler CM, Butcher LL and Bredesen DE. Induction of apoptosis by the low-affinity NGF receptor. Science 1993;261:345–348. 23. Casaccia-Bonnefil P, Carter BD, Dobrowsky RT and Chao MV. Death of oligodendrocytes mediated by the interaction of nerve growth factor with its receptor p75. Nature 1996;383: 716–719. 24. Yoon SO, Casaccia-Bonnefil P, Carter B and Chao MV. Competitive signaling between TrkA and p75 nerve growth factor receptors determines cell survival. J Neurosci 1998;18: 3273–3281. 25. Gu C, Casaccia-Bonnefil P, Srinivasan A and Chao MV. Oligodendrocyte apoptosis mediated by caspase activation. J Neurosci 1999;19:3043–3049. 26. Soilu-Hanninen M, Ekert P, Bucci T, Syroid D, Bartlett PF and Kilpatrick TJ. Nerve growth factor signaling through p75 induces apoptosis in Schwann cells via a Bcl-2-independent pathway. J Neurosci 1999;19:4828–4838. 27. Trim N, Morgan S, Evans M, Issa R, Fine D, Afford S, Wilkins B and Iredale J. Hepatic stellate cells express the low affinity nerve growth factor receptor p75 and undergo apoptosis in response to nerve growth factor stimulation. Am Pathol 2000;156:1235–1243. 28. Salehi AH, Roux PP, Kubu CJ, Zeindler C, Bhakar A, Tannis LL, Verdi JM and Barker PA. NRAGE, a novel MAGE protein, interacts with the p75 neurotrophin receptor and facilitates nerve growth factor-dependent apoptosis. Neuron 2000;27:279–288. 29. Cotrina ML, Gonzalez-Hoyuela M, Barbas JA and Rodriguez-Tebar A. Programmed cell death in the developing somites is promoted by nerve growth factor via its p75(NTR) receptor. Dev Biol 2000;228:326–336. 30. von Bartheld CS, Heuer JG and Bothwell M. Expression of nerve growth factor (NGF) receptors in the brain and retina of chick embryos: comparison with cholinergic development. J Comp Neurol 1991;310:103–129. 31. Davey F and Davies AM. TrkB signalling inhibits p75-mediated apoptosis induced by nerve growth factor in embryonic proprioceptive neurons. Curr Biol 1998;8:915–918. 32. Frade JM. Unscheduled re-entry into the cell cycle induced by NGF precedes cell death in nascent retinal neurones. J Cell Sci 2000;113:1139–1148. 33. Naumann T, Casademunt E, Hollerbach E, Hofmann J, Dechant G, Frotscher M and Barde YA. Complete deletion of the neurotrophin receptor p75NTR leads to long-lasting
144
34. 35. 36.
37. 38.
39.
40. 41.
42.
43.
44.
45. 46. 47.
48.
49. 50. 51.
52.
increases in the number of basal forebrain cholinergic neurons. J Neurosci 2002;22: 2409–2418. Frade JM, Rodriguez-Tebar A and Barde YA. Induction of cell death by endogenous nerve growth factor through its p75 receptor. Nature 1996;383:166–168. Frade JM and Barde YA. Microglia-derived nerve growth factor causes cell death in the developing retina. Neuron 1998;20:35–41. Majdan M, Lachance C, Gloster A, Aloyz R, Zeindler C, Bamji S, Bhakar A, Belliveau D, Fawcett J, Miller FD and Barker PA. Transgenic mice expressing the intracellular domain of the p75 neurotrophin receptor undergo neuronal apoptosis. J Neurosci 1997;17:6988–6998. Lee R, Kermani P, Teng KK and Hempstead BL. Regulation of cell survival by secreted proneurotrophins. Science 2001;294:1945–1948. Fahnestock M, Michalski B, Xu B and Coughlin MD. The precursor pro-nerve growth factor is the predominant form of nerve growth factor in brain and is increased in Alzheimer’s disease. Mol Cell Neurosci 2001;18:210–220. Yaar M, Zhai S, Pilch PF, Doyle SM, Eisenhauer PB, Fine RE and Gilchrest BA. Binding of beta-amyloid to the p75 neurotrophin receptor induces apoptosis. A possible mechanism for Alzheimer’s disease. J Clin Invest 1997;100:2333–2340. Kuner P, Schubenel R and Hertel C. Beta-amyloid binds to p57NTR and activates NFkappaB in human neuroblastoma cells. J Neurosci Res 1998;54:798–804. Della-Bianca V, Rossi F, Armato U, Dal-Pra I, Costantini C, Perini G, Politi V and Della G. Valle, Neurotrophin p75 receptor is involved in neuronal damage by prion peptide-(106–126). J Biol Chem 2001;276:38929–38933. Perini G, Della-Bianca V, Politi V, Della Valle G, Dal-Pra I, Rossi F and Armato U. Role of p75 neurotrophin receptor in the neurotoxicity by beta-amyloid peptides and synergistic effect of inflammatory cytokines. J Exp Med 2002;195:907–918. Tuffereau C, Benejean J, Blondel D, Kieffer B and Flamand A. Low-affinity nerve-growth factor receptor (P75NTR) can serve as a receptor for rabies virus. EMBO J 1998;17: 7250–7259. Fainzilber M, Smit AB, Syed NI, Wildering WC, Herman, van der Schors RC, Jimenez C, Li KW, van Minnen J, Bulloch AG, Ibanez CF and Geraerts WP. CRNF, a molluscan neurotrophic factor that interacts with the p75 neurotrophin receptor. Science 1996;274: 1540–1543. Yamashita T, Tucker KL and Barde YA. Neurotrophin binding to the p75 receptor modulates Rho activity and axonal outgrowth. Neuron 1999;24:585–593. Bentley CA and Lee K-F. p75 is important for axon growth and Schwann cell migration during development. J Neurosci 2000;20:7706–7715. Brann AB, Scott R, Neuberger Y, Abulafia D, Boldin S, Frainzilber M and Futerman AH. Ceramide signaling downstream of the p75 neurotrophin receptor mediates the effects of nerve growth factor on outgrowth of cultured hippocampal neurons. J Neurosci 1999;19:8199–8206. Anton ES, Weskamp G, Reichardt LF and Matthew WD. Nerve growth factor and its low-affinity receptor promote Schwann cell migration. Proc Natl Acad Sci USA 1994;91: 2795–2799. Cosgaya JM, Chan JR and Shooter EM. The neurotrophin receptor p75NTR as a positive modulator of myelination. Science 2002;298:1245–1248. Blo¨chl A and Sirrenberg C. Neurotrophin stimulate the release of dopamine from rat mesencephalic neurons via Trk and p75Lntr receptors. J Biol Chem 1996;271:21100–21107. Stucky CL and Koltzenburg M. The low-affinity neurotrophin p75 regulates the function but not the selective survival of specific subpopulations of sensory neurons. J Neurosci 1997;17: 4398–4405. Jiang H, Takeda K, Lazarovici P, Katagiri Y, Yu Z-X, Dickens G, Chabuk A, Liu X-W, Ferrans V and Guroff G. Nerve Growth Factor (NGF)-induced calcium influx and
145
53.
54.
55. 56.
57. 58. 59.
60. 61.
62.
63.
64.
65.
66.
67.
68.
69.
intracellular calcium mobilization in 3T3 cells expressing NGF receptors. J Biol Chem 1999; 274:26209–26216. von Schack D, Casademunt E, Schweigreiter R, Meyer M, Bibel M and Dechant G. Complete ablation of the neurotrophin receptor p75NTR causes defects both in the nervous and the vascular system. Nat Neurosci 2001;4:977–978. Lee KF, Li E, Huber LJ, Landis SC, Sharpe AH, Chao MV and Jaenisch R. Targeted mutation of the gene encoding the low affinity NGF receptor p75 leads to deficits in the peripheral sensory nervous system. Cell 1992;69:737–749. Roux PP, Colicos MA, Barker PA and Kennedy TE. p75 neurotrophin receptor expression is induced in apoptotic neurons after seizure. J Neurosci 1999;19:6887–6896. Cheema SS, Barrett GL and Bartlett PF. Reducing p75 nerve growth factor receptor levels using antisense oligonucleotides prevents the loss of axotomized sensory neurons in the dorsal root ganglia of newborn rats. J Neurosci Res 1996;46:239–245. Ferri CC, Moore FA and Bisby MA. Effects of facial nerve injury on mouse motoneurons lacking the p75 low-affinity neurotrophin receptor. J Neurobiol 1998;34:1–9. Ferri CC and Bisby MA. Improved survival of injured sciatic nerve Schwann cells in mice lacking the p75 receptor. Neurosci Lett 1999;272:191–194. Sendtner M, Holtmann B, Kolbeck R, Thoenen H and Barde YA. Brain-derived neurotrophic factor prevents the death of motoneurons in newborn rats after nerve section. Nature 1992; 360:757–759. Mufson EJ and Kordower JH. Cortical neurons express nerve growth factor receptors in advanced age and Alzheimer disease. Proc Natl Acad Sci USA 1992;89:569–573. Lowry KS, Murray SS, McLean CA, Talman P, Mathers S, Lopes EC and Cheema SS. A potential role for the p75 low-affinity neurotrophin receptor in spinal motor neuron degeneration in murine and human amyotrophic lateral sclerosis. Amyotroph Lateral Scler Other Motor Neuron Disord 2001;2:127–134. Giehl KM, Rohrig S, Bonatz H, Gutjahr M, Leiner B, Bartke I, Yan Q, Reichardt LF, Backus C, Welcher AA, Dethleffsen K, Mestres P and Meyer M. Endogenous brain-derived neurotrophic factor and neurotrophin-3 antagonistically regulate survival of axotomized corticospinal neurons in vivo. J Neurosci 2001;21:3492–3502. Dobrowsky RT, Werner MH, Castellino AM, Chao MV and Hannun YA. Activation of the sphingomyelin cycle through the low-affinity neurotrophin receptor. Science 1994;265: 1596–1599. Wiegmann K, Schutze S, Machleidt T, Witte D and Kronke M. Functional dichotomy of neutral and acidic sphingomyelinases in tumor necrosis factor signaling. Cell 1994;78: 1005–1015. Carter BD, Kaltschmidt C, Kaltschmidt B, Offenhauser N, Bohm-Matthaei R, Baeuerle PA and Barde YA. Selective activation of NF-kappa B by nerve growth factor through the neurotrophin receptor p75. Science 1996;272:542–545. Liu ZG, Hsu H, Goeddel DV and Karin M. Dissection of TNF receptor 1 effector functions: JNK activation is not linked to apoptosis while NF-kappaB activation prevents cell death. Cell 1996;87:565–576. Hamanoue M, Middleton G, Wyatt S, Jaffray E, Hay RT and Davies AM. p75-Mediated NF-kappaB activation enhances the survival response of developing sensory neurons to nerve growth factor. Mol Cell Neurosci 1999;14:28–40. Coulson EJ, Reid K, Barrett GL and Bartlett PF. p75 neurotrophin receptor-mediated neuronal death is promoted by Bcl-2 and prevented by Bcl-xL. J Biol Chem 1999;274: 16387–16391. Yoon SO, Casaccia-Bonnefil P, Carter B and Chao MV. Competitive signaling between TrkA and p75 nerve growth factor receptors determines cell survival. J Neurosci 1998;18: 3273–3281.
146 70. Pozniak CD, Radinovic S, Yang A, McKeon F, Kaplan DR and Miller FD. An anti-apoptotic role for the p53 family member, p73, during developmental neuron death. Science 2000;289: 304–306. 71. Casademunt E, Carter BD, Benzel I, Frade JM, Dechant G and Barde YA. The zinc finger protein NRIF interacts with the neurotrophin receptor p75(NTR) and participates in programmed cell death. EMBO J 1999;18:6050–6061. 72. Mukai J, Hachiya T, Shoji-Hoshino S, Kimura MT, Nadano D, Suvanto P, Hanaoka T, Li Y, Irie S, Greene LA and Sato TA. NADE, a p75NTR-associated cell death executor, is involved in signal transduction mediated by the common neurotrophin receptor p75NTR. J Biol Chem 2000;275:17566–17570. 73. Hosomi S, Yamashita T, Aoki M and Tohyama M. The p75 receptor is required for BDNF-induced differentiation of neural precursor cells. Biochem Biophys Res Commun 2003; 301:1011–1015. 74. Chittka A and Chao MV. Identification of a zinc finger protein whose subcellular distribution is regulated by serum and nerve growth factor. Proc Natl Acad Sci USA 1999;96:10705–10710. 75. Canossa M, Twiss JL, Verity AN and Shooter EM. p75(NGFR) and TrkA receptors collaborate to rapidly activate a p75(NGFR)-associated protein kinase. EMBO J 1996;15: 3369–3376. 76. Volonte C, Ross AH and Greene LA. Association of a purine-analogue-sensitive protein kinase activity with p75 nerve growth factor receptors. Mol Biol Cell 1993;4:71–78. 77. Higuchi H, Yamashita T, Yoshikaw H and Tohyama M. PKA phosphorylates the p75 receptor and regulates its localization to lipid rafts. EMBO J 2003;22:1790–1800. 78. Buck CR, Martinez HJ, Black IB and Chao MV. Developmentally regulated expression of the nerve growth factor receptor gene in the periphery and brain. Proc Natl Acad Sci USA 1987;84:3060–3063. 79. Ernfors P, Hallbook F, Ebendal T, Shooter EM, Radeke MJ, Misko TP and Persson H. Developmental and regional expression of beta-nerve growth factor receptor mRNA in the chick and rat. Neuron 1988;1:983–996. 80. Large TH, Weskamp G, Helder JC, Radeke MJ, Misko TP, Shooter EM and Reichardt LF. Structure and developmental expression of the nerve growth factor receptor in the chicken central nervous system. Neuron 1989;2:1123–1134. 81. von Bartheld CS, Kinoshita Y, Prevette D, Yin QW, Oppenheim RW and Bothwell M. Positive and negative effects of neurotrophins on the isthmo-optic nucleus in chick embryos. Neuron 1994;12:639–654. 82. Yan Q and Johnson EM, Jr. An immunohistochemical study of the nerve growth factor receptor in developing rats. J Neurosci 1988;8:3481–3498. 83. Ernfors P, Henschen A, Olson L and Persson H. Expression of nerve growth factor receptor mRNA is developmentally regulated and increased after axotomy in rat spinal cord motoneurons. Neuron 1989;2:1605–1613. 84. Bentley CA and Lee KF. p75 is important for axon growth and schwann cell migration during development. J Neurosci 2000;20:7706–7715. 85. Lee KF, Bachman K, Landis S and Jaenisch R. Dependence on p75 for innervation of some sympathetic targets. Science 1994;263:1447–1449. 86. McQuillen PS, DeFreitas MF, Zada G and Shatz CJ. A novel role for p75NTR in subplate growth cone complexity and visual thalamocortical innervation. J Neurosci 2002;22: 3580–3593. 87. Brann AB, Scott R, Neuberger Y, Abulafia D, Boldin S, Fainzilber M and Futerman AH. Ceramide signaling downstream of the p75 neurotrophin receptor mediates the effects of nerve growth factor on outgrowth of cultured hippocampal neurons. J Neurosci 1999;19:8199–8206. 88. Lehmann M, Fournier A, Selles-Navarro I, Dergham P, Sebok A, Leclerc N, Tigyi G and McKerracher L. Inactivation of Rho signaling pathway promotes CNS axon regeneration. J Neurosci 1999;19:7537–7547.
147 89. Luo L, Jan LY and Jan YN. Rho family GTP-binding proteins in growth cone signalling. Curr Opin Neurobiol 1997;7:81–86. 90. Luo L. Rho GTPases in neuronal morphogenesis. Nat Rev Neurosci 2000;1:260–264. 91. Higuchi H, Yamashita T, Yoshikawa H and Tohyama M. PKA phosphorylates the p75 receptor and regulates its localization to lipid rafts. EMBO J 2003;22:1790–1800. 92. Lang P, Gesbert-Carmagnat F, Delespine M, Stancou R, Pouchelet M and Bertoglio J. Protein kinase A phosphorylation of RhoA mediates the morphological and functional effects of cyclic AMP in cytotoxic lymphocytes. EMBO J 1996;15:510–519. 93. Walsh GS, Krol KM and Kawaja MD. Absence of the p75 neurotrophin receptor alters the pattern of sympathosensory sprouting in the trigeminal ganglia of mice overexpressing nerve growth factor. J Neurosci 1999;19:258–273. 94. Yeo TT, Chua-Couzens J, Butcher LL, Bredesen DE, Cooper JD, Valletta JS, Mobley WC and Longo FM. Absence of p75NTR causes increased basal forebrain cholinergic neuron size, choline acetyltransferase activity, and target innervation. J Neurosci 1997;17:7594–7605. 95. Tello F. La influencia del neurotropismo en la regeneracion de los centros nerviosos. Trab Lab Invest Biol 1911;9:123–159. 96. Ramon y Cajal S. Degeneration and Regeneration of the Nervous System, London, Oxford Unversity Press, 1928. 97. David S and Aguayo AJ. Axonal elongation into peripheral nervous system ‘‘bridges’’ after central nervous system injury in adult rats. Science 1981;214:931–933. 98. Richardson PM, Issa VM and Aguayo AJ. Regeneration of long spinal axons in the rat. J Neurocytol 1984;13:165–182. 99. Keirstead SA, Rasminsky M, Fukuda Y, Carter DA, Aguayo AJ and Vidal-Sanz M. Electrophysiologic responses in hamster superior colliculus evoked by regenerating retinal axons. Science 1989;246:255–257. 100. Berry M. Post-injury myelin-breakdown products inhibit axonal growth: An hypothesis to explain the failure of axonal regeneration in the mammalian central nervous system. Bibliotheca Anatomica 1982;23:1–11. 101. Schwab ME and Thoenen H. Dissociated neurons regenerate into sciatic but not optic nerve explants in culture irrespective of neurotrophic factors. J Neurosci 1985;5:2415–2423. 102. Caroni P and Schwab ME. Two membrane protein fractions from rat central myelin with inhibitory properties for neurite growth and fibroblast spreading. J Cell Biol 1988;106: 1281–1288. 103. Spillmann AA, Bandtlow CE, Lottspeich F, Keller F and Schwab ME. Identification and characterization of a bovine neurite growth inhibitor (bNI-220). J Biol Chem 1998;273: 19283–19293. 104. Prinjha R, Moore SE, Vinson M, Blake S, Morrow R, Christie G, Michalovich D, Simmons DL and Walsh FS. Inhibitor of neurite outgrowth in humans. Nature 2000;403: 383–384. 105. GrandPre T, Nakamura F, Vartanian T and Strittmatter SM. Identification of the Nogo inhibitor of axon regeneration as a Reticulon protein. Nature 2000;403:439–444. 106. Chen MS, Huber AB, van der Haar ME, Frank M, Schnell L, Spillmann AA, Christ F and Schwab ME. Nogo-A is a myelin-associated neurite outgrowth inhibitor and an antigen for monoclonal antibody IN-1. Nature 2000;403:434–439. 107. Fournier AE, GrandPre T and Strittmatter SM. Identification of a receptor mediating Nogo-66 inhibition of axonal regeneration. Nature 2001;409:341–346. 108. Huber AB, Weinmann O, Brosamle C, Oertle T and Schwab ME. Patterns of Nogo mRNA and protein expression in the developing and adult rat and after CNS lesions. J Neurosci 2002;22:3553–3567. 109. McKerracher L, David S, Jackson DL, Kottis V, Dunn RJ and Braun PE. Identification of myelin-associated glycoprotein as a major myelin-derived inhibitor of neurite growth. Neuron 1994;13:805–811.
148 110.
111. 112.
113.
114.
115.
116. 117.
118.
119. 120.
121. 122.
123. 124.
125. 126. 127.
128. 129.
Mukhopadhyay G, Doherty P, Walsh FS, Crocker PR and Filbin MT. A novel role for myelin-associated glycoprotein as an inhibitor of axonal regeneration. Neuron 1994;13: 757–767. Cai D, Qiu J, Cao Z, McAtee M, Bregman BS and Filbin MT. Neuronal cyclic AMP controls the developmental loss in ability of axons to regenerate. J Neurosci 2001;21:4731–4739. Tang S, Qiu J, Nikulina E and Filbin MT. Soluble myelin-associated glycoprotein released from damaged white matter inhibits axonal regeneration. Mol Cell Neurosci 2001;18: 259–269. Kottis V, Thibault P, Mikol D, Xiao ZC, Zhang R, Dergham P and Braun PE. Oligodendrocyte-myelin glycoprotein (OMgp) is an inhibitor of neurite outgrowth. J Neurochem 2002;82:1566–1569. Wang KC, Koprivica V, Kim JA, Sivasankaran R, Guo Y, Neve RL and He Z. Oligodendrocyte-myelin glycoprotein is a Nogo receptor ligand that inhibits neurite outgrowth. Nature 2002;417:941–944. Habib AA, Marton LS, Allwardt B, Gulcher JR, Mikol DD, Hognason T, Chattopadhyay N and Stefansson K. Expression of the oligodendrocyte-myelin glycoprotein by neurons in the mouse central nervous system. J Neurochem 1998;70:1704–1711. Fournier AE, GrandPre T and Strittmatter SM. Identification of a receptor mediating Nogo-66 inhibition of axonal regeneration. Nature 2001;409:341–346. Josephson A, Trifunovski A, Widmer HR, Widenfalk J, Olson L and Spenger C. Nogo-receptor gene activity: cellular localization and developmental regulation of mRNA in mice and humans. J Comp Neurol 2002;453:292–304. Wang X, Chun SJ, Treloar H, Vartanian T, Greer CA and Strittmatter SM. Localization of Nogo-A and Nogo-66 receptor proteins at sites of axon-myelin and synaptic contact. J Neurosci 2002;22:5505–5515. Liu BP, Fournier A, GrandPre T and Strittmatter SM. Myelin-associated glycoprotein as a functional ligand for the Nogo-66 receptor. Science 2002;297:1190–1193. Domeniconi M, Cao Z, Spencer T, Sivasankaran R, Wang K, Nikulina E, Kimura N, Cai H, Deng K, Gao Y, He Z and Filbin M. Myelin-associated glycoprotein interacts with the Nogo66 receptor to inhibit neurite outgrowth. Neuron 2002;35:283–290. Hunt D, Mason MR, Campbell G, Coffin R and Anderson PN. Nogo receptor mRNA expression in intact and regenerating CNS neurons. Mol Cell Neurosci 2002;20:537–552. Niederost B, Oertle T, Fritsche J, McKinney RA and Bandtlow CE. Nogo-A and myelinassociated glycoprotein mediate neurite growth inhibition by antagonistic regulation of RhoA and Rac1. J Neurosci 2002;22:10368–10376. GrandPre T, Li S and Strittmatter SM. Nogo-66 receptor antagonist peptide promotes axonal regeneration. Nature 2002;417:547–551. Cai D, Shen Y, De Bellard M, Tang S and Filbin MT. Prior exposure to neurotrophins blocks inhibition of axonal regeneration by MAG and myelin via a cAMP-dependent mechanism. Neuron 1999;22:89–101. Yamashita T, Higuchi H and Tohyama M. The p75 receptor transduces the signal from myelin-associated glycoprotein to Rho. J Cell Biol 2002;157:565–570. Wang KC, Kim JA, Sivasankaran R, Segal R and He Z. p75 interacts with the Nogo receptor as a co-receptor for Nogo, MAG and OMgp. Nature 2002;420:74–78. Wong ST, Henley JR, Kanning KC, Huang KH, Bothwell M and Poo MM. A p75(NTR) and Nogo receptor complex mediates repulsive signaling by myelin-associated glycoprotein. Nat Neurosci 2002;5:1302–1308. Yamashita T and Tohyama M. The p75 receptor acts as a displacement factor that releases Rho from Rho-GDI. Nat Neurosci 2003;6:461–467. Sasaki T and Takai Y. The Rho small G protein family-Rho GDI system as a temporal and spatial determinant for cytoskeletal control. Biochem Biophys Res Commun 1998;245: 641–645.
149 130.
131. 132. 133.
Dergham P, Ellezam B, Essagian C, Avedissian H, Lubell WD and McKerracher L. Rho signaling pathway targeted to promote spinal cord repair. J Neurosci 2002;22: 6570–6577. Fournier AE, Takizawa BT and Strittmatter SM. Rho kinase inhibition enhances axonal regeneration in the injured CNS. J Neurosci 2003;23:1416–1423. Koch G, Haberman B, Mohr C, Just I and Aktories K. Interaction of mastoparan with the low molecular mass GTP-binding proteins rho/rac. FEBS Lett 1991;291:336–340. Ilag LL, Rottenberger C, Liepinsh E, Wellnhofer G, Rudert F, Otting G and Ilag LL. Selection of a peptide ligand to the p75 neurotrophin receptor death domain and determination of its binding sites by NMR. Biochem. Biophys Res Commun 1999;255:104–109.
151
Phage display for epitope determination: A paradigm for identifying receptor–ligand interactions Merrill J. Rowley*, Karen O’Connor, and Lakshmi Wijeyewickrema Department of Biochemistry and Molecular Biology, Monash University, Clayton, Victoria 3800, Australia Abstract. Antibodies that react with many different molecular species of protein and non-protein nature are widely studied in biology and have particular utilities, but the precise epitopes recognized are seldom well defined. The definition of epitopes by X-ray crystallography of the antigen–antibody complex, the gold standard procedure, has shown that most antibody epitopes are conformational and specified by interactions with topographic determinants on the surface of the antigenic molecule. Techniques available for the definition of such epitopes are limited. Phage display using either gene-specific libraries, or random peptide libraries, provides a powerful technique for an approach to epitope identification. The technique can identify amino acids on protein antigens that are critical for antibody binding and, further, the isolation of peptide motifs that are both structural and functional mimotopes of both protein and non-protein antigens. This review discusses techniques used to isolate such mimotopes, to confirm their specificity, and to characterize peptide epitopes. Moreover there are direct practical applications to deriving epitopes or mimotopes by sequence, notably the development of new diagnostic reagents, or therapeutic agonist or antagonist molecules. The techniques developed for mapping of antibody epitopes are applicable to probing the origins of autoimmune diseases and certain cancers by identifying ‘‘immunofootprints’’ of unknown initiating agents, as we discuss herein, and are directly applicable to examination of a wider range of receptor–ligand interactions. Keywords: antibody, epitope, conformational epitope, discontinuous epitope, critical contact residues, phage displayed libraries, random peptide libraries, gene specific library, mimotopes, carbohydrate antigens, DNA antigens, filamentous bacteriophage, pIII, pVIII, pVI, receptor–ligand interactions, cysteine-constrained libraries, unconstrained libraries, immunofootprinting, autoimmunity, tumor-specific antigens, X-ray crystallography, homology modeling, vaccines.
Introduction Antibodies can bind with high affinity and specificity to molecules of virtually any shape, and to antigens ranging from small organic compounds to large proteins. These characteristics have led to the widespread use of antibodies as laboratory reagents, in diagnostic tests, and for therapeutic purposes. The immune system, upon stimulation, produces antibodies of increasing affinity by a process of natural selection. This antigen-driven selection that governs the production of antibodies has three key features: the generation of millions of different antibody genes through the rearrangement of a limited number of germ-line gene segments; the expression of this repertoire of rearranged genes on the surface of B lymphocytes where the antibody functions as an *Corresponding author: E-mail:
[email protected] BIOTECHNOLOGY ANNUAL REVIEW VOLUME 10 ISSN: 1387-2656 DOI: 10.1016/S1387-2656(04)10006-9
ß 2004 ELSEVIER B.V. ALL RIGHTS RESERVED
152 antigen receptor; and the further antigen-driven selection of clones of B lymphocytes for proliferation and differentiation into antibody-producing plasma cells [1]. Antibodies are raised by the immune system against regions on the surface of a protein known as epitopes. Epitopes or antigenic determinants are often classified as either continuous or discontinuous: a continuous or linear epitope typically comprises three or four adjacent amino acids over a short segment in the primary sequence, whereas discontinuous or conformational epitopes comprise residues that are distant in primary sequence but brought together in the folded native conformation. Whilst an epitope is often specified as a linear sequence of an antigenic molecule, antibodies in fact react optimally with structures formed by protein folding by which disparate residues for the antibody paratope come into contiguity on the surface of the antigenic molecule; the actual affinity of antigen–antibody binding, and the conferring of specificity, can depend on the interaction of just a few contiguous albeit discontinuous amino acid residues [2]. Accordingly the distinction between the two types of epitopes to a degree is artificial. Thus, a linear peptide itself can adopt particular conformations and, within the tertiary structure of a protein, an epitope defined as linear according to reactivity with a short peptide sequence may well represent just one part of a larger conformational epitope [2,3].
Techniques of epitope mapping The epitope–paratope interface involves a surface area of about 700 A˚2. The use of X-ray crystallography to examine structures of antibody–protein complexes has indicated that most epitopes contain 15–22 protein residues in contact with the combining site of the antibody [4–6]. Molecular modeling and site-directed mutagenesis studies on antibodies and protein antigens revealed that within a structural or conformational epitope, there is a subset of residues that contributed most of the free binding energy [7,8]. These contact residues constitute the functional epitope and can be scattered over two or three discontinuous polypeptide segments. On average at least three noncontiguous amino acids dominate a functional epitope, and about eight amino acid residues contribute to binding [8]. Complete definition of the structure of epitopes requires X-ray crystallography, but this requires a purified and homogenous source of antibody and of antigen, and is not readily applicable to most antigens. Initially, epitopes were most readily defined by comparing the cross-reactivity of antibodies against defined proteins with naturally occurring variants [9]. This technique was used extensively in early studies of antigen–antibody interactions, for antigens that included hen egg-white lysozyme in which lysozymes could be compared from different species of birds [9,10]. However, such studies were limited by the availability of naturally occurring and sequenced variants.
153 With the development of molecular biology, epitope mapping with both monoclonal and polyclonal antibodies has been performed by dissecting the antigen into overlapping polypeptides in the form of recombinantly expressed fusion proteins or truncation mutants [11–13]. However the assembly of appropriate panels of truncation mutants is time-consuming and the procedure usually resolves the epitope region only to a sequence of some 100–200 amino acids. With other approaches, attempts are made to maintain the conformation of the original antigen. Hybrid molecules have been prepared, exemplified by hybrids of the two isoforms of glutamic acid decarboxylase, GAD65 and GAD67 [14,15], or antigens are altered by site-directed mutagenesis to identify amino acids that contribute particularly to antibody binding [15,16]; these procedures likewise provide limited insight into the conformational structure of the true epitope. Alternatively, linear synthetic peptides can be used to identify epitopes. Geysen introduced the technique of synthesizing overlapping oligopeptides and probing for reactivity with the antibody under investigation [17,18], and showed the applicability of this technique to the mapping of ‘‘discontinuous’’ epitopes [18]. Since then many peptide-based technologies have been developed for mapping both B and T cell epitopes [19]. It was these studies that lead to the designation of ‘‘mimotopes,’’ as mimics of an epitope, with the term applied to peptides that bound to antibodies reactive with conformational epitopes, and hence acted as mimics of the epitopes without necessarily any linear sequence homology with the original antigen. Although peptide epitope mapping has been widely applied [20–22] and the peptides identified may react directly by ELISA with antibodies, or absorb reactivity, this reactivity is usually weak relative to that of the parent molecule; hence these sequences are likely fragments of a more complex antigenic structure [3]. For this reason, scanning for antibodies with overlapping sets of synthetic or recombinant peptides covering the antigen sequence has had only limited success. Screening large libraries of random peptides would allow the selection of any peptides that could fulfil the 3-dimensional requirements for recognition, whether or not the peptide occurred in the primary antigenic sequence, but studies that utilize such random peptides have been restricted by the lack of techniques to synthesize and screen appropriate libraries. Accordingly, the development of phage display has provided a powerful new methodology for epitope mapping that allows the ready identification of noncontiguous ‘‘critical contact residues’’ that contribute to binding.
Phage display Phage display technology is the product of two elementary concepts. First, any insertion mutation at an appropriate location within a structural gene of a virus if it does not disrupt an essential function conferred by the gene product,
154 will lead to the display of the mutation-encoded peptide on the surface of the viral particle. Second, if multiple inserts are all random oligonucleotides, the resulting phage particles will comprise a vast library of peptides, each one displayed on the virus and linked to the DNA that encodes it in the mutated coat protein surrounding the enclosed mutant DNA [23]. The physical association between phenotype (the displayed peptide) and genotype (the encoding DNA) in the same phage particle is the unique and highly advantageous feature of phage displayed peptide libraries. Phage display technology was first developed by Smith [24] who cloned a restriction enzyme digest of plasmid DNA into the gene III insertion site of the filamentous phage f1, thus creating a library of fusion proteins with the foreign sequence in the middle, which was displayed on the virion surface. Smith showed that, after transfection in Escherichia coli, each of the ‘‘fusion phage’’ clones that were produced contained the insert, which encoded part of Eco R1 endonuclease, and could be affinity purified from a library of random inserts using antibody to Eco R1 endonuclease. Cwirla et al. [25] further developed the technology by creating a large and diverse oligonucleotide library using inserts with randomly synthesized residues representing many variations. In this case, the phage were affinity selected using a monoclonal antibody specific for the N-terminus of b-endorphin (YGGF). Almost all clones selected displayed YG on the N-terminus of the variable peptide, in agreement with the known specificity of the monoclonal antibody. Of importance for the acceptance of the technique, the screening method required no previous knowledge of the structure of the peptide, or its antibody specificity. The above studies pointed to the utility of the technology for identifying antibody epitopes on a wide range of antigens. Although for Smith’s initial study, the DNA insert was a restriction digest of plasmid DNA, and so in fact represented a gene-specific library, the possibility of selecting ligands from large libraries of random peptides with very diverse amino acid sequences created immediate interest. There are now various phage-displayed random peptide libraries commercially available, and phage display technology has become an economical and practical approach to ascertain the epitopes for a wide range of antibodies, both monoclonal and polyclonal [26–33]. Antibody screening readily discloses sequences that mimic linear, discontinuous, and even nonpeptide epitopes of antigens, such as DNA or carbohydrates [34–36]. Moreover, phage display technology has been extended to the study of diverse types of receptor–ligand interaction, using the techniques generally applicable to antibody–epitope interactions.
Characteristics of phage particles Foreign polypeptides have been displayed on viruses [37], eukaryotic cells [38], bacteria [39], as well as on bacteriophage l [40], T4 [41–43], T7 [44–46] and
155 P4 [47]. However most published work cites filamentous phage, a class of single stranded bacteriophage that infect only male bacteria. Filamentous phage have several properties that make them attractive for use as peptide display vectors, being well characterized, easy to work with, and having a surface of low complexity. Also, the phage grow to high titres, and are resistant to pH 2 or 12, which simplifies the breakdown of phage–antibody complexes. The filamentous phage M13, closely related to filamentous phage fd and f1, has been used most extensively; it is a nonlytic phage, so that the problem of contamination from E. coli host proteins is much reduced, with less time spent isolating the phage. By electron microscopy, the wild-type M13 virus particle appears as a flexible rod, about 1 mm in length, and 6 nm in diameter, depending on the strain used. The single-stranded circular genome is stretched along the entire particle length and is coated by the helically arranged molecules of the 50-residue major coat protein pVIII (about 0.42 copies of pVIII per nucleotide or for the wild-type 6408 nucleotides, about 2670 copies of pVIII) [48]. At one tip of the virus particle there are five copies each of the pIII and pVI proteins (genes III and VI, respectively) that are involved in host cell binding and termination of the assembly process, and the minor coat proteins pVII and pIX (genes VII and IX) are at the other tip (Fig. 1). The phage receptor on the bacterial surface is the tip of a thread-like structure, the sex pilus, encoded by the F episome in male strains, and thus referred to as the F pilus; phage infect strains of E. coli that display the F pilus. Infection is initiated by attachment of the N-terminal domain of pIII (about 200 amino acids) to the tip of the pilus, which is the end of the particle that first enters the cell [49]. After the virion is brought to the cell surface and the single stranded genome (or ( þ ) strand) is delivered to the cytoplasm, host polymerases employ the ( þ ) strand as a template for synthesis of a complementary () strand, yielding a double-stranded phage genome [50]. Unlike lytic phage, which are released by cell lysis after assembly in the host cell cytoplasm, M13 phage are continuously extruded through the host cell envelope in a process that couples assembly with export. A feature of
Fig. 1. Schematic diagram of the structure of filamentous phage.
156 filamentous phage of importance for their use in phage display technology is their ability, merely by further addition of pVIII subunits, to package longer genomes than the wild type [51]. pIII is required not only for F-pilus absorption but also for terminating virion assembly and stabilising the viral particle; deficiency of pIII leads to the production of multilength viral particles (polyphage) containing two or more unit-length phage genomes [52–54]. Display systems in filamentous phage Phage libraries in filamentous phage have been generated by fusion of foreign peptides into three coat proteins, pIII, pVIII and, less often, pVI, with each having particular advantages and disadvantages [55]. The most widely used phage display vectors, in which peptides are displayed close to the signal peptide in the pIII coat protein, (type 3, type 3 þ 3, type 33), or the pVIII coat protein (type 8, type 8 þ 8 or type 88) are shown in Table 1 [49,56]. The first phage system used was a type 3 vector in which the phage genome coded also for a fusion protein [24], but every copy of a coat protein containing a peptide has disadvantages: the size of peptides that can be inserted in every copy of a coat protein without disrupting function is limited; and a lower density of recombinant peptide on the surface of a bacteriophage should allow for better discrimination of peptides that bind with high affinity [57]. The need to limit the number of peptide-displaying coat proteins prompted the development of phagemid vectors (type 3 þ 3 or type 8 þ 8 vectors) [58,59]; phagemids are plasmids that possess the usual plasmid origin of replication and selectable antibiotic-resistance markers together with all the bacteriophage elements required for single-strand synthesis and encapsidation. If a cell that is harboring a phagemid is co-infected with a helper phage, there is generated a mixture of particles that contain either the phagemid or helper-phage genome, with a combination of wild-type and hybrid molecules on the surface. Through selection for their antibiotic-resistance markers, phagemid virions can be selected from the helper virus. Alternative systems have been developed based on Table 1. Most common phage display vectors developed for use in filamentous phage. System
Coat protein
Type
No. of peptides
References
3 33 3þ3 8 88 8þ8
pIII pIII pIII pVIII pVIII pVIII
Single recombinant gene III with DNA insert Two genes III, one wild type, one with insert Two genes III, recombinant gene on plasmid Single recombinant gene VIII with DNA insert Two genes VIII, one wild type, one with insert Two genes VIII, recombinant gene on plasmid
5 40%, the dose of epoetin alfa should be held until the hematocrit decreases to 35%. The dose then should be reduced by 25% upon restarting therapy Contraindicated in patients with uncontrolled hypertension or with known hypersensitivity to any of the constituents or benzoic acid
Darbepoetin alfa
Chemotherapy-induced anemia in patients with nonmyeloid cancers; anemia associated with chronic renal failure
Contraindicated in patients with Chemotherapy-induced anemia: recommended uncontrolled hypertension. Use with starting dose is 2.25 mg/kg/week. The dose caution in patients with history of should be adjusted to maintain target Hgb hypertension and/or cardiovascular concentration. Dose should be increased to disease. Contraindicated in patients with 4.5 mg/kg/week if Hgb concentration increase is known hypersensitivity to mammalian