Enzymes for Carbohydrate Engineering

ENZYMES FOR CARBOHYDRATE ENGINEERING Progress in Biotechnology Volume 1 New Approaches to Research on Cereal Carbohyd...

Author: Kwan-Hwa Park | John F. Robyt | Yang-Do Choi

127 downloads 1459 Views 11MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

ENZYMES FOR CARBOHYDRATE ENGINEERING

Progress in Biotechnology Volume 1 New Approaches to Research on Cereal Carbohydrates (Hill and Munck, Editors) Volume 2 Biology of Anaerobic Bacteria (Dubourguier et al., Editors) Volume 3 Modifications and Applications of Industrial Polysaccharides (Yalpani, Editor) Volume 4 Interbiotech '87. Enzyme Technologies (Bla~ej and Zemek, Editors) Volume 5 In Vitro Immunization in Hybridoma Technology (Borrebaeck, Editor) Volume 6 Interbiotech '89. Mathematical Modelling in Biotechnol0gy (Bla~ej and Ottova, Editors) Volume 7 Xylans and Xylanases (Visser et al., Editors) Volume 8 Biocatalysis in Non-Conventional Media (Tramper et al., Editors) Volume 9 ECB6: Proceedings of the 6th European Congress on Biotechnology (Alberghina et al., Editors) Volume 10 Carbohydrate Bioengineering (Petersen et al., Editors) Volume 11 Immobilized Cells: Basics and Applications (Wijffels et al., Editors) Volume 12 Enzymes for Carbohydrate Engineering (Kwan-Hwa Park et al., Editors)

Progress in B i o t e c h n o l o g y 12

ENZYMES FOR CARBOHYDRATE ENGINEERING Edited by Kwan-Hwa Park Department of Food Science and Technology, and Research Center for New Bio-Materials in Agriculture, Seoul National University Suwon, Korea J o h n F. R o b y t Department of Biochemistry and Biophysics, Laboratory of Carbohydrate Chemistry and Enzymology, Iowa State University Ames, Iowa, U.S.A. Yang-Do Choi Department of Agricultural Chemistry, and Research Center for New Bio-Materials in Agriculture, Seoul National University Suwon, Korea

ELSEVIER Amsterdam

- Lausanne - New York - Oxford - Shannon

- Tokyo

1996

Published by: Elsevier Science B.V. P.O. Box 211 1000 AE Amsterdam The Netherlands

ISBN 0-444-82408-1 ©1996 Elsevier Science B.V. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher, Elsevier Science B.V., Permissions Department, P.O. Box 521, 1000 AM Amsterdam, The Netherlands. No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, the Publisher recommends that independent verification of diagnoses and drug dosages should be made. Special regulations for readers in the USA - This publication has been registered with the Copyright Clearance Center Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01293, USA. Information can be obtained from the CCC about conditions under which photocopies of parts of this publication may be made in the USA. All other copyright questions, including photocopying outside the USA should be referred to the copyright owner, Elsevier Science B.V., unless otherwise specified. This book is printed on acid-free paper. Printed in the Netherlands

PREFACE The recent advances in carbohydrate engineering largely can be attributed to the advancement of biotechnology. In particular, successful applications of the state of the art biotechnology for the development and improvement of carbohydrate enzymes has been achieved by pioneers in carbohydrate enzymology. Two consecutive agricultural biotechnology symposia on Enzymes for Carbohydrate Engineering were organized by the Research Center for New Bio-Materials in Agriculture, and were held in Suwon, Korea in August 1994 and September 1995. Most up-to-date information and achievements in enzymology, as applied to carbohydrate engineering, have been compiled through these two symposia and this resulting book. Currently, carbohydrate engineering is an important part of agricultural biotechnology. New types of amylases are being isolated from various microorganisms and the functional properties of carbohydrates produced by treatments of above enzymes are of great interest. This book describes the fundamental research and application of carbohydrate enzymes to agricultural biotechnology. It is concentrated on the structures and biochemical properties of various enzymes, their production, and application in plant and foods. Since the application of enzymology to carbohydrate engineering has developed rapidly, based primarily on discoveries in biochemistry and molecular biology, the gap between basic research and applications in the bio-industries has become greater. Through the symposia and this book, we hope that a bridge connecting basic sciences with their application for the development of carbohydrate engineering can be found. The scientists involved in the symposia were from the fields of basic chemistry, molecular biology, and enzymology. We trust the current information in this book will contribute to studies and applications of basic and applied scientists in the field of carbohydrate enzymology, chemistry, biotechnology, agricultural engineering, and various interrelated fields. K. H. Park John F. Robyt Y. D. Choi

vi ACKNOWLEDGEMENTS We gratefully wish to t h a n k the Korea Science and Engineering Foundation for their supports that have made possible the realization of the Symposia and the publication of the Proceedings. T h a n k s are extended to members of the Research Bio-Materials in Agriculture for their contribution.

Center

for

New

Special t h a n k s are due to Professors Yong Hwan Lee and Yeon Woo Ryu for their critical reading of articles and editorial efforts. We also acknowledge manuscripts.

Mr. Myo Jeong

Kims assistance

in p r e p a r i n g

the

°°

VII

CONTENTS

Preface Mechanism and action of glucansucrases John F. Robyt Cyclodextrin producing enzyme (CGTase) Shoichi Kobayashi

23

Modulation of Bacillus amylolytic enzymes and production of branched oligosaccharides Tae-Kyu Cheong, Tae-Jip Kim, Myo-Jeong Kim, Yang-Do Choi, In-Cheol Kim, Jung-Wan Kim and Kwan-Hwa Park

43

A novel maltotetraose-forming alkaline s-amylase from an alkalophlic Bacillus strain, GM8901 Yong Chul Shin and Si Myung Byun

61

Structural studies on cellulases, pectinases and xylanases Peter W. Goodenough

83

Structure and activity of some start-metabolising enzymes E. Ann MacGregor

109

Properties and uses of dextransucrases elaborated by a new class of Leuconostoc mesenteriodes mutants Doman Kim and John F. Robyt

125

Molecular determinants of thermozyme activity and stability: analysis of xylose isomerase and amylopullulanase J. Gregory Zeikus

145

Crystal structure of Bacillus licheniformis o~-Amylase at 1.7/k resolution Hyun Kyu Song, Kwang Yeon Hwan, Chansoo Chang and Se Won Seo

163

Characteristics of carbohydrase reactions in heterogeneous enzyme reaction system utilizing swollen extrusion starch as the substrate Yong-Hyun Lee and Dong-Chan Park

171

Manipulation of storage compounds in transgenic plants David W. Stalker, Kevin E. McBride and Christine K. Shewmaker

189

Overproduction of bacterial amylases in recombinant Escherichia coli systems Jin-Ho Seo, Woo-Jong Lee, Myoung-Dong Kim, Chang-Sup Kim and Yong-Chul Park

201

INDEX OF AUTHORS

215

This Page Intentionally Left Blank

Enzymesfor Carbohydrate Engineering K.H. Park, J.F. Robyt and Y-D. Choi (Editors) 9 1996 Elsevier Science B.V. All rights reserved.

Mechanism and Action of Glucansucrases John F. Robyt Laboratory of Carbohydrate Chemistry and Enzymology, D e p a r t m e n t of Biochemistry and Biophysics, Iowa State University, Ames, Iowa 50011 USA

Abstract

There are several different kinds of glucansucrases that synthesize glucans with different structures from sucrose. Glucans that have contiguous a-1---6 linked glucose residues are known as dextrans. The dextrans differ from each other by the type of branch linkages (a-1-,2, a-1-~3 or a-1-~4), the percentage of branching, the length of the branch chams, and their spatial arrangement. Some of these dextrans are very highly branched with 3 5 % - 5 0 % branch linkages. Other glucans that have contiguous a-1-,3 linked glucose residues in the main chains are known as mutans, and glucans with alternating a-1--6 and c~-1--~3 linked glucose residues in the main chains are known as alternans. The enzymes that catalyze glucan synthesis from sucrose form covalent glucosyl- and glucanyl-enzyme complexes. They transfer the glucosyl umt to the reducing-end of the growing glucan chain by a two-site insertion mechanism. Branching of the glucans occurs when a glucan chain acts as an acceptor and attacks the covalent glucosyl- or glucanyl-enzyme complex. Glucose or glucan chain are transferred to the acceptor chain where they are attached by a branch linkage. When other carbohydrates, in addition to sucrose, are present in the enzyme digest, the enzyme transfers glucose to the carbohydrate acceptors in a secondary reaction that diverts some of the glucose from incorporation into glucan. Many carbohydrate acceptors have been recognized and the products that result are dependent on the particular enzyme and the structure of the particular acceptor.

S t r u c t u r e s of the g l u c a n s and s o u r c e s of the e n z y m e s

Glucansucrases comprise a family of enzymes that synthesize glucans fiom sucrose. The enzymes are secreted into the culture media by a number of Leuconostoc and Streptococcus species and strains. These two genera are gram-positive, facultative anaerobic cocci that are closely related to each other. One notable difference between them is that, until recently, Leuconostoc species required sucrose in the culture medium to induce the formation of the enzyme(s), whereas the Streptococcus species did not require sucrose in the culture medium for the formation of the enzymes. Thus, the Leucolmstoc species were inducible for the formation of the glucansucrases, and the Streptococcus species were constitutive for their formation. Jeanes, et al., 1954 reported the formation of glucan by 96 strains of bacteria that were primarily Leuconostoc strains. There is a question here as to whether they are strains or species, especially with regard to the formation of different

kinds of polysaccharides. The classification of the time was to place them into one species, mesenteroides, that had several different strains. This classification stands today. The polysaccharides were characterized by various properties such as optical rotation, viscosity, periodate oxidation, and physical appearance after alcohol precipitation. The latter were observed to have different appearances that were described by Jeanes, et al., 1954 in qualitative terms such as pasty, fluid, stringy, tough, long, short, flocculent, crumbly, etc. It was also found that certain strains (species) of organisms elaborated more than one kind of polysaccharide. Wilham, et al., 1955 reported the separation of these polysaccharides by differential alcohol precipitation. Table 1 lists several Leuc. mesenteroides strains and some Streptococcus species along with the percentages of the linkages found in the various glucans that are synthesized by the enzymes elaborated by the organisms. Table 1 illustrates a wide variation in the types of linkages and their percentages found in the main chains and in the branch linkages. The structures of the polysaccharides have been determined by methylation (Van Cleve, et al., 1956; Hare, et al., 1978" Jeanes and Seymour, 1979; Seymour et al. 1977, 1979a; and Shimamura, et al., 1 9 8 2 ) a n d I~C-NMR (Seymour, et al., 1976, 1979b, 1979c, 1979d, 1979e, 1980a, 1980b, 1980c). Initially all of the glucans synthesized from sucrose were considered to be dextrans. A dextran is defined as a glucan with main chains being composed of contiguous ~-1-~6 linked glucopyranose residues. Differences in the different dextrans involve the types, amount, length, and arrangements of the branch chains. The principal type of branch linkage is a-1~3, but a-1~2 and a-1--~4 branch linkages have also been observed (see Table 1). As the structures of the different glucans have been studied, it became obvious that there were glucans that did not fit the above definition of a dextran. In particular, there were glucans that had contiguous a-l--*3 linkages in the main chains and glucans that had alternating a-1---6 and a-l--*3 linked glucose residues in the main chains. Leuc. mesenteroides B-512F(M) produces only one glucan, a dextran, that has 95% ~-1~6 linkages in the main chains and 5% a-1~3 branch linkages(Van Cleve, et al., 1956; Jeans and Seymour, 1979). The branches consist of two types, single glucose units and relatively long a-1~6 linked chains attached to an a-1--6 linked chain by an a-1--3 branch linkage (see Fig.l). Other dextrans contain a much higher percentage of a-1-~3 branch linkages. Leuc. mesenteroides B-742 dextran-S has 50% ~-1-~6 linkages and 50% a-1-~3 branch lingkages of which the majority are single glucose residues (Seymour, et al. , 1979a). This is the highest degree of branching that could be obtained. The branches predominantly consist of single glucose residues. The structure that results is a bifurcated comb in which each of the single c~-1-~3 linked glucose residues are like teeth of a comb on a backbone of c~-1-~6 linked chains (see Fig. 1). A dextran of this type would be highly resistant to endo-dextranase hydrolysis. Leuc. mesenteroides B-742 also produces another dextran that has 7% a-1-~4 branch linkages rather than a-1~3 branch linkages. Strep. m u t a n s 6715 also elaborates two glucansucrases. One (glucosyltransferase-soluble or GTF-S) is a dextransucrase that synthesizes a water-soluble dextran reported to have 35% c~-1-~3 branch linkages consisting primarily of single glucose residues (Seymour, et al. 1980c; Shimamura, et al.,

Table 1 Glucans Synthesized from Sucrose by Glucansucrease from Selected Leuconostoc and Streptococci Species and Solubility Strain No. a Class b

1--6

L.m. B-512F

L

95

L.m. B-742

L

87

Percent of Linkages Description of the 1--3 1-*3Br c 1-*2Br c 1-*4Br c ethanol precipitate translucent gel 13

L.m. B-742

S

50

50

L.m B-1299

L

66

1

L.m B-1299

S

65

L.m B-1355

L

95

L.m B-1355 L.m B-1191

S L

54 94

L.m. B- 1308

L

95

S.s.

I

83

B-1526

fine ~ 27

flocculent ~

35

fine ~

5 35

11 2

heavy, opaque

translucent gel 4

heavy, opaque cohesive, stringy d

5

pasty, crumbly

17

fluid, stringy ~

L.m. B-523

I

100

water-insoluble

L.m. B-1149

I

100

water-insoluble

S.m. 6715

S

64

S.m. 6715

I

4

S.v. B-1351

S

89

36 96

- Leuconostoc mesenteroides" Strep. s p e c i e s S . v . - Strep. viridans in the N o r t h e r n Regional R e s e a r c h U S D A Laboratory, Peoria, IL. bL p r e c i p i t a t e d by 34-37% ethanol; S recipitated by 40-44% ethanol; I escription t a k e n from J e a n e s et al. aL.m.

2 11

heavy, opaque water-insoluble short ~

S.m. - Streptococcus mutants; B - n u m b e r s refer to the s t r a i n L a b o r a t o r y Collection (NRRL) - less soluble referring to - more soluble referring to water-insoluble. Br c - b r a n c h (5).

S.s. = number of the glucans glucans linkage.

1982). This also is a relatively high degree of branching, in which one out of every two glucose residues along the a - 1 ~ 6 linked main chain has an a-1-~3 linked single glucose branch. If the single branch glucose residues are uniformly distributed along the a-1---6 chain, the result is an alternating, bifurcated comb structure in which the single branch glucose residues are attached by a-1--3 linkages to every other glucose residue in the main chains (see Fig. 1). This structure also would be resistant to endo-dextranase hydrolysis. The second enzyme elaborated by Strep. mutans synthesizes a water-insoluble glucan t h a t has contiguous a-l--*3 linked glucose residues in the main chain not a instead of c~-1~6 linkages (Seymour, et al., 1979e) and obviously is dextran. It is totally resistant to endo-dextranase and is called mutan and its enzyme is called mutansucrase or GTF-I.

/~C~/'~~uc.

ruescn~ro~t~s

mum~

Atlermttb,l~ Comb 13~xtrun

Fig. 1. Structural representation of segments of different glucans synthesized by glucansucrases from sucrose. represents a glucose residue linked a-1--6 to another glucose residue represents a glucose residue linked a-l-*3 to another glucose residue

%

represents a glucose residue linked c~-1--~2 to another glucose residue

Leuc. mesenteroides B-1355 also elaborates two glucansucrases. The first enzyme synthesizes a dextran very similar in structure to B-512F dextran. The second enzyme synthesizes a glucan that has alternating a-1---6 and a-1--3 linked glucose residues in the main chains with 11% c~-1---3 branch linkages. This also is not a dextran; it has been caUled alternan and its enzyme, alternansucrase. Alternan also is totally resistant to endo-dextranase hydrolysis. The dextran produced by Leuc. mesenteroides B-1299 is highly unusual. It has single glucose branches that are hnked by c~-1-~2 glucosidic bonds (Seymour, et al., 1979a, 1979e). The c~-1---2 bond is quite rare in biological systems. Recently (Kim and Robyt, 1994, 1995) have obtained Leuc. mesenteroides mutants from strains B-512FM, B-742, B-1299, and B-1355 that are constitutive for the glucansucrases instead of being inducible. This permits the organism to elaborate active enzymes when the mutants are grown in media containnmg glucose or fructose instead of sucrose. In addition, some of the mutants elaborated only one of the two glucansucrases that were elaborated by the wild type organism, such as B-742, B-1355, and B-1299. These mutants greatly facihtate the purification of the enzymes and give enzyme in the culture supernatant that are devoid of the polysaccharides that they synthesize.

Mechanism of polysaccharide synthesis The reaction of glucansucrases with sucrose can be simply formulated by the following: n Sucrose -* (Glucose)n_m.w + n-m Frucose + m Leucrose + w Glucose !

The reaction is essentially irreversible. The main products are high molecular weight glucan (1 X 107-1 X 10s Da) and fructose; the minor products are glucose and leucrose (5-O-c~-D-glucopyranosyl-D-fructopyranose) (where n > > m or w). Small amounts of glucose are formed from an acceptor reaction with water and leucrose is formed from an acceptor reaction with the primary product, fructose. A discussion of the acceptor reactions is given later. When dextransucrase was first described (Hehre, 1941), Cori and Cori, 1939 and Swanson and Cori, 1948 were studying the action of muscle phosphorylase and Hanes, 1940 was studying potato phosphorylase. These investigators observed that phosphorylase could elongate glycogen and starch chains by the transfer of glucose from a-glucose-l-phosphate (G-l-P) to the nonreducing-end glucose residues of glycogen and starch. The reaction did not take place unless a glycogen primer chain or a starch primer chain was present. It, thus, resulted that a primer was an absolutely required constituent in the enzyme digest to obtain chain elongation. The phosphorylase reaction was later shown not to be the mechanism for either glycogen or starch synthesis, but rather to be a degradative process in which phosphorylase catalyzes the reaction of inorganic phosphate with the nonreducing glucose residue of the glycogen or starch chain to give a-G-I-P (Stetten and Stetten, 1960). The so-called synthetic reaction, requiring a primer, was the reverse of the degradative reaction that indeed requires the "primer" chain for degradation. This, however, was not appreciated and the primer mechanism for polysaccharide synthesis became firmly

established. Thus, in the 1940's and 50's. the primer mechanism was assumed for the synthesis of dextran (Hehre, 1951; Koepsell, et al., 1953; Tsychiya, et al., 1955; Tsuchiya, 1960). The primer mechanism for dextran synthesis was strengthened in the 1970's when Germaine, et al., 1976, 1977 found that the addition of dextran to Strep. mutans dextransucrase digests increased the rate of dextran synthesis. Kobayashi and Matsuda, 1980, 1986 also reported that the purified dextransucrases of both Leuc. mensenteroides B-512F and Strep. mutans were stimulated by dextran, although both enzymes could synthesize dextran without the addition of dextran primer. The reaction was accompanied by a lag that could be abolished by the addition of exogenous dextran. Both groups interpreted their results as evidence for a primer-based mechanism for dextran synthesis. Kobayashi and Matsuda, 1986 proposed a model for a primer dependent synthesis of dextran (see Fig. 2). They postulated that the active-site has two types of substrate binding-sites, one for sucrose and the other for the primer dextran. It was proposed that the primer is elongated when an enzyme nucleophile (a carboxylate anion) makes an attack on C-1 of the glucose moiety of sucrose to give a glucosyl-enzyme covalent intermediate. This glucose is then transferred to the nonreducing end of the primer chain when the C-6 hydroxyl group of the nonreducing-end glucose of the primer chain makes an attack on C-1 of the glucose unit of the glucosyl-enzyme complex, givmg elongation of the chain by one glucose residue. For synthesis to continue, the elongated chain

PrimerBinding-site ,,,

~

~

e

~_

LX_ 0

SucroseBinding-site

etc.

.....

~ m e r C~2

t~ ~~di~ociadon of I ~ dexu-anchain

5 Fig. 2. Primer-dependent mechanism for the synthesis of Leuc. mesel~teroides B-512F dextransucrase. Q - ~ is sucrose, 9 is glucose, ~ is fructose, Xrepresents an enzyme nucleophile, O - O represents two glucose residues linked a- 1~6.

must dissociate and a new primer molecule must brad in the primer binding-site with its nonreducing glucose residue in position to accept the next glucose residue. The primer mechanism, thus, is a discontinuous process in which glucose residues are added one at a time to many individual primer molecules. No firm evidence, however, was ever presented for this mechanism. There is a significant difference in the synthesis of dextran by dextransucrase and the chain elongation of glycogen and starch by phosphorylase. Even though the rate of dextran synthesis is stimulated by the addition of exogenous dextran, the dextransucrase can still synthesize dextran in the absence of any added dextran. A number of other problems developed for the primer mechanism. Robyt and Corrigan, 1977 found that dextrans modified by a blocking group (triisopropyl benzene sulfonyl or tripsyl) on the noneducing-end C-6 hydroxyl group increased the rate of dextran synthesis equally as well as did unmodified dextran. The modified dextran could not participate in a priming reaction as the requisite site for the addition of glucose, the C-6 hydroxyl of the nonreducing-end glucose residue, was blocked by a tripsyl group. This showed that the added dextran was not stimulating the reaction by acting as a primer but by some other mechanism. Robyt, et al. (1995) reported that added dextran was activating dextransucrases from Leuc. mesenteroides B-512FMC (a mutant constitutive for dextransucrase) and from Strep. m u t a n s 6715 (a species constitutive for dextransucrase) by an allosteric mechanism and not by a primer mechanism. They showed that neither of these enzymes required a primer for dextran synthesis. Using pulse and chase techniques with 14C-sucrose and Bio-Gel P2 immobilized dextransucrase, Robyt, et al., 1974 showed that glucose and dextran were covalently attached to the enzyme during synthesis and that the glucose is added to the reducing-end of the growing chain by a two-site insertion mechanism. This mechanism postulated that there are two sucrose binding-sites and two nucleophiles (presumably carboxylate anions) that attack the two sucrose molecules to give two covalent glucosyl-enzyme intermediates (see Fig. 3). The C-6 hydroxyl of one of the glucosyl intermediates makes a nucleophilic

L-x'~-:~b---@

2

x ~

(0)A

Fig. 3. Two-site insertion mechanism for the synthesis of Leuc. mese~teroides B-512F dextran by dextransucrase. The symbols are the same as in Fig. 2. X orients the glucosyl units so that their C-6 hydroxyl groups can make an attack onto C-1 of the apposed glucosyl unit.

attack onto C-1 of the other glucosyl intermediate to form an a - 1 ~ 6 glycosidic linkage and an isomaltosyl mtermediate. The newly released nucleophilic then attacks another sucrose molecule to give a new glucosyl-enzyme intermediate. The C-6 hydroxyl of this new glucose-enzyme intermediate then attacks the C-1 of the isomaltosyl-intermediate to give an a - l o 6 linkage and the formation of an isomaltotriosyl-enzyme intermediate. The process contmues in a similar manner between the two sites, giving the synthesis of an c~-1~6 linked glucan chain by the addition of glucose to the reducmg-end of the growing chain and the apparent insertion of glucose between the enzyme and the dextran chain. The dextran chain remains attached to the enzyme and is extruded from the active-site as the glucose residues are added to the reducing-end of the chain. The dextran chain, thus, is synthesized de n o v o in a contmuous manner

OH

xV Reaction

)

2_._OH

Ctt2--OH

HO

H 2--.-OH OH

OH

f--

Xo

H

Reaction 2

/

,

c~

OH ~

~......~T_

-"N / 4 ~ - ~

X

H

.__J

\

H

x---~

Fig. 4. Mechanism for the cleavage of sucrose and the formation of a a-1--6 glycosidic bond by dextransucrase. Reaction 1 : nucleophilic displacement and protonation of the fructose moiety to form a glucosyl-enzyme intermediate. Reaction 2 : formation of an c~-1-~6 glycosidic bond by attack of a C-6 hydroxyl group onto C-1 of a glycosyl-enzyme complex ; the attack is faciliated by abstraction of a proton from the hydroxyl group by the imidazole group.

\~t

without the need for any preformed primer dextran and without the need for the dissociation of the dextran cham from the active-site before the next glucose residue is added. An additional requirement for the reaction to take place is the transfer of a hydrogen ion to the displaced fructosyl moiety of sucrose (Robyt and Eklund, 1982). Chemical modification of dextransucrease by diethyl pyrocarbonate and Rose Bengal dye photo-oxidation (Fu and Robyt, 1988) showed that two imidazolium groups of histidine were essential for dextran synthesis. It was postulated that these two imidazolium groups donate their hydrogen ions to the leaving fructose units (see Fig. 4) and that the resulting imidazole group, in a second step, becomes reprotonated by abstracting a proton from the attacking C-6 hydroxyl group of the glucosyl-enzyme intermediate, facilitating the nucleophilic attack and the formation of the a-l-*6 hnkage. The imidazole group, thereby becomes reprotonated for the next reaction with sucrose. In a similar pulse and chase study, Robyt and Martin, 1983 showed that both Strep. mutans GTF-S and GTF-I synthesized alternating comb dextran and mutan from the reducing-end by an insertion mechanism. The mechanism for mutan synthesis is shown in Fig. 5. This mechanism differs from that of dextran synthesis only by having the glucosyl-enzyme intermediates oriented so their C-3 hydroxyl groups are stereochemically placed to make the nucleophilic attack onto C-1 to give the formation of ~-1---3 glycosidic linkages. The synthesis of alternan can likewise be postulated to take place by a two-site insertion mechanism. In this synthesis, the two glucosyl-enzyme intermediates are stereochemically positioned differently. On one site (the X-site), the glucosyl-intermediate is positioned so that its C-6 hydroxyl makes the attack onto C-1 of the opposite glucosyl- enzyme intermediate to give an a-1 ~ 6 linkage, while on the other site (the Y-site), the glucosyl-enzyme intermediate is stereochemically positioned so that its C-3 hydroxyl makes the attack onto C-1 of the opposite glucosyl metrmediate to give an a - l ~ 3 linkage (see Fig.6). In this manner the synthesis goes back and forth between the two

C

._.y.') n-times

Fig. 5. Two-site insertion mechanism for the synthesis of Strep. mutans mutan by mutansucrase. The symbols are the same as in Fig. 3 with the substitution of Y for X as the nucleophiles and O - Q represents a glucose residue hnked c~ -1---3 to another glucose residue. Y orients the glucosyl units so that their C-3 hydroxyl groups can make an attack onto C-1 of the apposed glucosyl unit.

lO

•

Fig. 6. Two-site insertion mechanism for the synthesis of Leuc. meseltte- roides B-1355 alternan by alternansucrease. The symbols are the same as in Fig. 5 with the addition that the two nucleophiles are X and Y; X orients its glucosyl unit so that its C-6 hydroxyl group can make an attack onto C-1 of the apposed glucosyl unit and Y orients its glucosyl unit so that its C-3 hydroxyl group can make an attack onto C-1 of the apposed glucosyl units.

sites giving the alternating synthesis of a-l-*6 and a-1-~3 glycosidic linkages. The two-site insertion mechanism for Leuc. mesenteroides B-521FM dextransucrase was confirmed (Su and Robyt, 1994) using equilibrium dialysis with 6-deoxy-sucrose, a strong competitive inhibitor for enzyme. They showed that there are two sucrose binding sites at the active-site. They further showed that the two sites are required for dextran synthesis by studying the relative rates of dextran synthesis and acceptor product synthesis as a function of diethylpyrocarbonate modification of histidine. The argument was based on the hypothesis that if two sites were required for glucan synthesis (Fig. 3) and one of the sites is modified so that it cannot function, synthesis of glucan would stop, but if only one of the two sites is required for the acceptor-reaction (Fig. 7), the acceptor reaction can still occur when only one of the two sites is modified. The experimental results verified the hypothesis as the enzyme lost the ability to synthesize dextran more rapidly than it did the ability to synthesize acceptor-product.

M e c h a n i s m for the branching of dextran Bovey(1959) postulated that dextran was branched by a specific enzyme similar to the branching enzyme found for the biosynthesis This dextran branching enzyme, however, has never been isolated or The energy for a dextran branching enzyme that transfers an a-1---6

branching of starch. observred. linkage to

11

Fig. 7. TLC autoradiogram of the acceptor products formed in the reaction of B-512 FM dextransucrase with ~4C- sucrose and D-glucose. The first acceptor product of D-glucose is isomaltose followed in decreasing amounts by isomaltotriose through isomaltooctaose. A small amount of leucrose results from the acceptor reaction of D-fructopyranose. Dextran remains at the origin. an a-1---3 branch linkage is unfavorable as the a-1--3 branch linkage is ot higher energy than the a-1---6 linkage. Robyt and Taniguchi 1976 studied the branching of dextran using Bio-Gel P2 immobilized dextransucrase. The immobilized enzyme was labeled by incubating it with a relatively low concentration of [14C]-sucrose. In a second procedure, the immoblilized enzyme was first incubated with nonlabeled sucrose, washed, and then labeled with a low concentration of [14C]-sucrose. In both experiments, the labeled material was shown to be glucose and dextran. When either of the labeled, immobilized enzymes were incubated with a low molecular weight, nonlabeled dextran, all ot the enzyme bound label was released as [ltC]-dextran. No [14C]-labeled dextran was released when the labeled enzyme was incubated in buffer alone. The

12 released [14C]-dextran was shown to be slightly branched by hydrolysis with an exo-dextranase. Acetolysis (a process that is relatively specific for cleaving c~-1 -*6 linkages) of the labeled dextran gave 7.3% of the 14C in nigerose. Reduction of the labeled nigerose, followed by acid hydrolysis, gave all of the label in glucose, demonstrating that the nigerose was exclusively labeled in the nonreducing glucose residue. The results of the experiments showed that the [14C]-label was being released by the action of the added low moleccular weight dextran (acceptor dextran) and that this action gave the formation of a new a-1 -*3 branch linkage. Robyt and Taniguchi, 1976 proposed a mechanism for the synthesis of branch linkages by Leuc. mesenteroides B-512FM dextransucrase in which a C-3 hydroxyl of an interior glucose residue on an acceptor dextran makes a nucleophilic attack onto C-1 of either the glucosyl-enzyme complex or onto C-1 of the dextranyl-enzyme complex, thereby forming an a-1---3 branch linkage by displacing glucose or dextran from the enzyme (Fig. 8). Thus, branching can take place without a seperate branching enzyme by the action of an acceptor dextran on the glucosyl- and dextranyl-dextrasucrase complexes.

A

~

.

.

.

____ A

. . . . . ~~

~

A ~

~

~----

8 A W

A W

A W

A fA~

A W

,,, A ~

A W

-

Fig. 8. Mechanism for the synthesis of a-1--+3 branch linkages by Leuc. mesenteroides B-512F dextransucrase. The C-3 hydroxyl of an acceptor dextran chain makes an attack onto (A) the glucosyl unit to give a single branched glucose linked a-1-.3 or 03) the C-3 hydroxyl group of the acceptor dextran makes attack onto C-1 of the glucosyl unit of the dextranyl chain to give long c~-1---3 linked branched dextran chain.

13 A c c e p t o r r e a c t i o n s of

glucansucrases

In addition to catalyzing the synthesis of dextran from sucrose, dextrasucrase also catalyzes the transfer of glucose from sucrose to other carbohydrates that are present or are added to the digest (Koepsell, et al., 1953; Tsuchiya, et al., 1955). The added carbohydrates are called acceptors and the reaction is called When the acceptor is a monosaccharide or disaccharide an acceptor-reaction. there usually is produced a series of ohgosaccharide acceptor-products (Robyt and Eklund, 1983). Fig.8 shows a chromatographic analysis of acceptor products that result when D-glucose is the added acceptor. Actually there are two classes of acceptors, those that give a homologous series of oligosaccharides, each differing one from the other by one glucose residue, and those acceptors that only form a single acceptor-product containing one glucose residue more than the acceptor. When D-glucose, methyl-c~-D-glucopyranoside, maltose, and isomaltose are the acceptors, the glucose from sucrose is transferred to the C-6 hydroxyl of the monosaccharide or to the C-6 hydroxyl of the nonreducing-end glucose residue of the disaccharides to give a series of isomaltodextrins of degree of polymerization (d.p.) of 2 to 7 attatched to the acceptor (Robyt and Walseth, 1978; Robyt and Eklund, 1983). The first product in the series with isomaltose is isomaltotriose and the first product in the series with maltose is panose [62-a-D-glucopyranosyl maltose] (Killey, et al., 1955). The next product in the maltose series is a tetrasaccharide, 6~-c~-isomaltosyl maltose, and the other members of the series have isomaltodextrin chams of increasing degrees of polymerization linked to the C-6 hydroxyl group of the nonreducing-end glucose residue of maltose (Robyt and Walseth, 1978). Similar homologous series are obtained from mgerose, 1,5-anhydro-D-glucitol, and turanose (Robyt and Eklund, 1983). The amount of each saccharide product in the series decreases as the d.p. increases, usually terminating at d.p. 6 or 7. Cellobiose gives an unusual series in which the first product is 21-c~-Dglucopyranosyl cellobiose with glucose attatched to the C-2 hydroxyl group of the reducing-end glucose residue (Barley, et al., 1958; Yamauchi and Ohwada, 1969). The succeeding products of the cellobiose series had the glucose unit of sucrose transferred to the C-6 hydroxyl of the glucose attatched to C-2 of the reducing residue of cellobiose. When the cellobiose analog, lactose, was the acceptor only one acceptor-product was formed, 21-a-D-glucopyranosyl lactose (Barley, et al., 1955; Bourne, et al., 1959; Yamauchi and Ohwada, 1969). There seems to be a pattern that when D-galactose composed part of the acceptor structure, only one acceptor product was formed, for example, raffinose [6~1c-(1 -D-galactopyranosyl sucrose] also gave only a single acceptor product, 2Glc-a -D-glucopyranosyl raffinose (Neely, 1959). When fructose is the acceptor, there are two products formed, depending on the ring form of the fructose acceptor. The major product, leucrose [5-O-c~-D-glucopyranosyl-D-fructopyranose], is formed from D-fructopyranose, and the minor product, isomaltulose [4-O-c~-Dglucopyranosyl-D-fructofuranose], is formed when D-fructofuranose is the acceptor (Stodola, et al., 1952, 1956; Sharpe, et al., 1960). Because D-fructose is a major product in the dextransucrase synthesis of dextran from sucrose, it acts as an acceptor to give leucrose in all dextransucrase-sucrose digests. A small amount of D-glucose also is formed when water acts as an acceptor

14 (Robyt and Eklund, 1983). This reaction represents the hydrolysis of sucrose. One study has suggested that dextransucrase has distinct active sites for sucrose hydrolysis and dextran sythesis (Yamashita, et al., 1989). Luzio et al. (Luzio and Mayer, 1983; Luzio, et al., 1983), however, showed that the three reactions catalyzed by dextransucrase, (a) sucrose hydroysis, (b) polymerization of the glucose moiety of sucrose, and (c) glucosyl transfer to acceptors, were competitive and therefore taking place at the same active-site. Other unusual acceptor-products result from the reaction of D-mannopyranose and D-galactofuranose. D-mannopyranose gave a nonreducing, a,B-trehalose isomer, ~-D-glucopyranosyl-B-D-mannopyranoside and D-galactofuranose gave aD-glucopyranosyl-B-D-galactofuranoside (Iriki and Hehre, 1969). c~,~-Trehalose gave two products, B-isomaltosyl-a-D-gluco pyranoside and c~isomaltosyl-6-D-glucopyranoside (Yamauchi and Ohwada, 1969) and ~,~-trehalose gave one product, B-isomaltosyl-~-D-glucopyranoside. a,~-Trehalose, however, was not an acceptor. Fu and Robyt, 1990 studied the structures of the maltodextrin, maltotriose to maltooctaose (G3-G8), acceptor products produced by Leuc. mesenteroides B-512FM dextransucrase and Strep. m u t a n s dextransucrase (GTF-S) and mutansucrase (GTF-I) (Fu and Robyt, 1991). They found that B-512FM dextransucrase transfers D-glucose from sucrose to C-6 hydroxyl of both the nonreducing-end and the reducing-end residues of G3-G8. G3, thus, gave two tetrasaccharides, 63-a-D-glucopyranosyl maltotriose and 61-a-D-glucopyranosyl maltotriose. The former acceptor-product was also an acceptor giving a homologous series of isomaltodextrins attatched to the C-6 hydroxyl of the nonreducing-end glucose residue. The acceptor-product with glucose attached to the reducing-end residue, however, was not acceptor. This same pattern was observed for the other maltodextrins studied. None of the glucose residues between the reducing-end glucose and the nonreducmg-end glucose served as acceptor sites. The maltose acceptor-products of GTF-S and GTF-I gave panose like the B-512FM dextransucrase. Panose, however, served as an acceptor to give two products, 6~-a-isomaltosyl maltose and 62-a-mgerosyl maltose. Whereas, panose reacting with B-512FM dextransucrase as an acceptor only gave 6-a-lsomaltosyl maltose, which served as an acceptor for all three enzymes to give a homologous series of 62-a-isomaltodextrinyl maltoses. Like B-512FM dextransucrase, GTF-S and GTF-I transferred glucose to the C-6 hydroxyl of both the nonreducing-end and the reducmg-end residues of the G3-G8 maltodextrins, and when glucose was transferred to the nonreducing-end residues, a series of homologous oligosacchrides resulted. When GTF-I reacted with G3 as an acceptor, four tetrasaccharide products resulted with glucose transferred to the C-6 hydroxyl and to the C-3 hydroxyl of both the nonreducing-end glucose and the reducing-end glucose. The acceptor product with glucose substituted on C-6 of the nonreducing-end residue of G3 served as an acceptor to give the homologous series. Only two acceptor-products, however, were imtially formed wih G4-G7 in which glucose was substituted at C-6 on the reducing residue and on the nonreducing residue. As with G3, the latter acceptor-products also served as acceptors to give a homologous series. C6t6 and Robyt, 1982 studied the acceptor products catalyzed by alternansucrase. They found that alternansucrase was capable of forming both 2

9

15 a - l ~ 6 and a-1~3 glycosidic bonds with acceptors. Isomaltose gave both 2isomaltotriose and 3~-a-D-glucopyranosyl isomaltose. These initial acceptor products also acted as acceptors, and the structures of the products of higher d.p. show that an 4-1--3 glycosidic bond is formed only when the nonreducing-end glucose residue is linked by an a-1~6 bond to another glucose residue. Nigerose, thus, gave 62-a-glucopyranosyl nigerose. Maltose gave 6~-aglucopyranosyl maltose but this saccharide gave an unusal tetrasaccharide, 6'Lanigerosyl maltose in which there are three types of glycosidic linkages in Table 2 Acceptor Products Formed by Glucansucrases a Acceptor

First Acceptor Product b

isomaltose maltose nigerose methyl-c~-D-glucopyranoside 1,5-anhydro-D- glucitol D-glucose lactose cellobios D-fructose raffinose melibiose D-glucitol D-mannose D-galactose theanderose c~,B-trehalosec

isomaltotriose* panose* 6z-ct-D-glucopyranosyl. nigerose* methyl-a-isomaltoside 1,5-anhydro-isomaltitol* isomaltose 2alc-a-D-glucopyranosyl lactose 21-a-D-glucopyranosyl cellobiose* leucrose (5-c~-D-glucopyranosyl D-fructose) -a-D-glucopyranosyl raffinose structure not determined structure not determined c~-D-glucopyranosyl ~-D-mannopyranoside a-D-glucopyranosyl [~-D-galactofuranoside -~-lsomaltosyl sucrose ~-isomaltosyl-a-D-glucopyranoside ct-isomaltosyl-[~-D-glucopyranoside B-isomaltosyl-B-D-glucopyranoside not an acceptor 63-a-D-glucopyranosyl maltotrmse* 6~-ct-D-glucopyranosyl maltotrxose 63-c~-D-glucopyranosyl maltotriose* 6~-a-D-glucopyranosyl maltotnose 33-a-D-glucopyranosyl maltotriose 3 l-a-D-glucopyranosyl maltotriose 64-ct-D-glucopyranosyl maltotetraose* 61-a-D-glucopyranosyl maltotetraose 6z-a-nigerosyl maltose*

[I,B-trehalose ~,a-trehalose maltotriose c maltotriose d

maltotetraose c panose e

alc

.-,

,,~Glc

9

.

-

*

aThe products are those produced by Leuc. mesenteroides B-512F dextransucrase unless otherwise indicated, bThe starred products are also acceptors that give a homologous series with ~-isomaltodextrins attached to the acce~tor. CTwo products are formed by Leuc. mesenteroides B-512F dextransucrase. Four products are formed by Strep. mutans GTF-I. eproduct formed by alternansucrase.

16 sequence from nonreducing-end: a-1-.3, a-1--6, and a-1--4. Thus, alternansucrase can synthesize both a-1---6 and a-1--3 acceptor product linkages. When the nonreducing residue acceptor is linked by an a-1--6 linkage, alternansucrase can transfer glucose to either C-6-OH or C-3-OH to give a-1-.6 or a-1--~3 linked glucose unit, but when the nonreducing glucose umt of the acceptor is linked by a ~-1-.3 or ~-1--4 bond, alternansucrase will only transfer glucose to C-6-OH of the nonreducing glucose residue. Another unusual feature was that nigerose was a better acceptor than isomaltose. Table 2 summarizes the major acceptors and their products. Robyt and walseth(1978) studied the mechanism of the acceptor reactions of Leuc. mesenteroides B-512FM dextransucrase. A purified dextransucrase was incubated with sucrose, and the resulting fructose, glucose, leucrose, and unreacted sucrose were removed from the enzyme by chromatography on a Bio-Gel P-6 column. The charged enzyme was incubated 9 9 [14,~ with C]-D-glucose, [14C]-D-fructose, and [14C]-reducing-end labeled maltose acceptors. Each of the three acceptors gave two types of labeled products, a high molecular weight product, identified as dextran, and a low molecular weight product that was an oligosaccharide. It was found that all three of the acceptors were incoporated

A _X---~

f ~

~

-X--O

B

9

Hs9 )n"

Fig. 9. Mechanism for the acceptor reaction of Leuc. mesenteroides B-512F dextransucrase. A disaccaride acceptor binds in the acceptor binding-site so that (A) its nonreducing C-6 hydroxyl group can make an attack on C-1 of the glucosyl unit releasing it from the active-site to give a trisaccharide or (B) its nonreducing C-6 hydroxyl can make an attack onto C-1 of the glucanyl chain releasing it from the active-site.

17 into the products at the reducing-end. Similar results were obtained when the enzyme and labeled acceptors were reacted in the presence of sucrose. The only difference being higher yields of the labeled products and a series of homologous oligosaccharides from the glucose and the maltose acceptor-reactions. Because both a labeled oligosaccharide and a labeled dextran was produced when labeled acceptor and enzyme were incubated together with and without sucrose, it was concluded that the acceptor reactions were taking place by the acceptor making a nucleophilic displacement of the glucosyl and dextranyl groups from the covalent enzyme intermediates. It was further concluded that the acceptor reactions serve to terminate polymerization of dextran by displacing the growing chain from the active-site in contrast to previous ideas that acceptors were serving as primers for dextran synthesis. Robyt and walseth(1978) proposed the mechamsm shown in Fig. 9 for the acceptor reaction. In this mechamsm, the acceptor is bound at an acceptor-binding site (Tanriseven and Robyt, 1992) and when maltose is the aceptor, its C-6 hydroxyl group at the nonreducmg-end attacks C-1 of the glucosyl or dextranyl groups in the enzyme complex to give an oligosaccharide or a dextran acceptor product, respectively. When glucose is the acceptor, its C-6-hydroxyl group makes the attack, and when fructose is the acceptor, its C-5-hydroxyl group makes the attack. For acceptors that form a homologous series, Robyt and Walseth(1978) also showed that when the concentration of the first acceptor-product becomes sufficiently high, it too can act as an acceptor to give the next higher homolog, which in turn can act as an acceptor so that a series of homologous oligosaccharides are formed. This was later confirmed by Mayer et a/.(1981) for Strep. sanguis dextransucrase. Thus, it was determined that the mechamsm of action of the acceptors is one of terminatmg dextran synthesis by the release of the glucosyl and dextranyl umts from the covalent enzyme-intermediate rather than one of priming the synthesis of dextran. The number of acceptor binding-sites was determined to be one for Leuc. mesenteroides B-512FM dextransucrase by Su and Robyt(1994), using maltose in an equilibrium dialysis experiment. Thus, the active-site of Leuc. mentereroides B-512FM dextransucrase has two sucrose binding-sites and one acceptor binding-site.

18

Literature Cited Bailey, R. W., Barker, S. A., Bourne, E. J. and M. Stacey. 1955. synthesis of a "branched" trisaccharide, Nature. 1 7 6 : 1164-1165.

Enzymic

Bourne, E. J., Hartigan, J. and H. Weigel. 1959. Mechanism of the enzymic

synthesis of a branched trisaccharide containing the c[-1,2 glucosidic linkage,

J.

Chem. Soc. 2332-2335. Bovey, F. A. 1959. Enzymatic Polymerization. I. Molecular Weight Branching during the Formation of Dextran, J. Polym. Sci. 35 : 167-182.

and

Chludzinski, A. M., Germaine, G. R. and C. F. Schachtele. 1976. Streptococcus mutans Dextransucrase : Purification, Properties, and Requirement for Primer Dextran, J. Dent. Res. special Issue C. 55 : C75-C86. Cori, G. T. and J. F. Cori. 1939. The activating effect of glycogen on the enzymatic synthesis of glycogen from glucose-l-phosphate, J. Biol. Chem. 131 : 397-398. C6t~, G.L. and C. F. Robyt. 1982. Acceptor reactions of alterllansucrase from Leuconostoc mesenteroides NRRL B- 1355, Carbohydr. Res. 111 : 127-142. Ditson, S. L. and R. M. Mayer. 1984. Dextransucrase : The direction of chain growth during autopolymerization, Carbohydr. Res. 1 2 6 : 1 7 0 - 1 7 5 . Fu, D. and J. F. Robyt. 1988. Essential histidine residues in dextransucrase:

chemical

modification

by

diethyl

pyrocarbonate

and

dye photo-oxidation,

Carbohydr. Res. 1 8 3 : 9 7 - 1 0 9 . Fu, D. and J. F. Robyt. 1990. Acceptor reactions of maltodextrins with Leuconostoc mesenteroides B-512FM dextransucrase, Arch. Biochem. Biophys. 283 :379-387. Fu, D. and J. F. Robyt. 1991. Maltodextrin acceptor reactions Streptococcus 6715 glucosyl transferases, Carbohydr. Res. 217 : 201-211.

mutans

Germaine, G. R., Chludzinski, A. M., and C. F. Schachtele. 1976. Streptococcus mutans Dextransucrase : Requirement for Primer Dextran, J. Bacteriol. 120C : 287-294. Germaine, G. R., Harlander, S. IC, Leung, W-L. S. and C. F. Schachtele. 1977.

Streptococcus mutans Dextransucrase : Functioning of Primer Dextran and Endogel~ous Dextranese in water-soluble and Water-insoluble Glucan Synthesis, Infect. Immun. 16 : 637-648. Hanes, C. S. 1940. The reversible formation of starch from glucose-l-phosphate catalysed by potato phosphorylase, Proc. Royal Soc. London. series B. 1 2 9 : 174-208.

19 Hare, M. D., Svensson, S. and G. J. Walker. 1978. Characterization of the extracellular, water insoluble ct-D-glucans of oral streptococci by methylation analysis, and by enzymic synthesis and degradation, Carbohydr. Res. 66 : 245-255. Hehre, E. J. 1941. Production from sucrose of a serologically polysaccharide by a sterile bacterial extract, Science 93 : 237-238.

reactive

Iriki, Y. and E. J. Hehre. 1969. The scope of interanomeric glycosyl transfer

reactions. Heterodialdoside synthesis by enzymic glucosylation of D-galactose and D-mannose, Arch. Biochem. Biophys. 134 : 130-137. Jeanes, A., Haynes, W. C., Wilham, C. A., Rankin, J. C., Melvin, E. H, Austin, M. J., Cluskey, J. E., Fisher, B. E., Tsuchiya, H. M. and C. E. Rist. 1954.

Characterization and classification of dextrans from ninety-six strains of bacteria, J. Am. Chem. Soc. 7 6 : 5041-5052. A. and F. R. Seymour. 1979. The a-D-glucopyranosidic linkages o! dextrans : comparison of percentages form structural analysis by periodate oxydation and by methylation, Carbohydr. Res. 74 " 31-40.

Jeanes,

Killey, M., Dimler, R. J. and J. E. Cluskey. 1955. Preparation of panose by the action of NRRL B-512 dextransucrase on a sucrose-maltose mixture, J. Am. Chem. Soc. 77 : 3315-3318. Kim, D. and J. F. Robyt. 1994. Selection of Leucolmstoc mesenteroides mutants constitutive for glucansucrases, Enzyme Microbiol. Technol. 1 6 : 1010-1015. Kim, D. and J. F. Robyt. 1995. Properties of Leuconostoc mesenteroides B-512FMC constitutive dextransucrase, Enzyme Microbiol. Technol. 17 : 689-695. Kobayashi, M. and IC Matsuda. 1980. Characterization of the multiple forms and main component of dextransucrase from Leuconostoc mesenteroides NRRL B-512F, Biochim. Biophys. Acta 614 : 46-62. Kobayashi, M. and IC Matsuda. 1986. Electrophoretic analysis J. Biochem. 1 0 0 : 615-621.

of the multiple

forms of dextransucrase,

Kobayashi, M., Yokoyama, I. and K. Matsuda. 1986. Substrate binding sites o! Leuconostoc dextransucrase evaluated by inhibition kinetics, Agric. Biol. Chem. 5 0 : 2585-2590. Koepsell, H. J., Tsuchiya, H. M., Hellman, N. N., Kazenko, A., Hoffman, C. A., Sharp, E. S. and R. W. Jackson. 1953. Enzymatic synthesis of dextran : acceptor specificity and chain initiation, J. Biol. Chem. 200 : 793-801. Luzio,

G.

A.

and

R.

M.,

Mayer.

1983.

dextransucrase, Carbohydr. Res. 1 1 1 : 3 1 1 - 3 1 8 .

The

hydrolysis

of sucrose

by

20 Luzio, G. A. Parnaik, V. K. and R. M. Mayer. 1983. A D-glucosylated form o!

dextransucrase, Carbohydr. Res. 1 2 1 : 2 6 9 - 2 7 8 . Neely, W. B. 1959. Studies on the enzyme dextransucrase: II. The role of raffinose as an acceptor, Arch. Biochem. Biophys. 79 : 154-161. Robyt, J. F., Kimble, B. K. and T. F. Walseth. 1974. The mechanism o! dextransucrase: direction of dextran biosynthesis, Arch. Biochem. Biophys. 165 : 634-640. F. and H. Taniguchi. 1976. The mechanism of dextransucrase : biosynthesis of branch linkages by acceptor reactions with dextran, Arch.

Robyt, J.

Biochem. Biophys. 174 : 129-135. Robyt, J. F. and A. J. Corrigan. 1977. The mechanism of dextransucrase : activation of dextransucrase from Streptococcus mutans OMZ 176 by dextran and modified dextran and the lmnexistence of the primer requirement for the synthesis of dextran, Arch. Biochem. Biophys. 183 : 726-731. Robyt, J. F. and T. F. Walseth. 1978. The mechanism of acceptor reactions o! Leuconiostoc mesenteroides B-512F dexrtransucrase, Carbohydr. Res. 61 : 433-445 Robyt, J. F. and S. H. Eklund. 1982. Streochemistry involved in the mechanism of action of dextransucrase in the synthesis of dextran and the formation o] acceptor products, Bioorg. Chem. 11 : 115-132 Robyt, J. F. and P. J. Martin. 1983. Mechanism of synthesis of D-glucans by D-glucanosyltransferases from Streptococcus mutans 6715, Carbohydr. Res. 113 : 301-315 Robyt, J. F. and S. H. Eklund. 1983. Relative, quantitative effects of acceptors in the reaction of Leuconostoc mesenteroides B-512F dextransucrase, Carbohydr. Res. 1 2 1 : 2 7 9 - 2 8 6 . Robyt, J. F., Kim, D. and L. Yu, 1995. Mechanism of dextran activation ol dextransucrase, Carbohydr. Res. 2 2 6 : 293-299. Seymour, F. R., Knapp, R. D. and S. H. Bishop. 1976. Determination of the structure of dextran by 13C-nuclear magnetic resonance spectroscopy, Carbohydr. Res. 51 : 179-194. Seymour, F. R., Slodlh, M. E., Plattner, R. D. and A. Jeanes. 1977. Six unusual

dextrans : methylation structural analysis by combined g.l.c, per-O-acetyl-aldononitriles, Carbohydr. Res. 53 : 153-166.

m.s.

o[

Seymour, F. R., Chen, E. C. M. and S. H. Bishop. 1979a. Methylation structural analysis of unusual dextrans by combined gas-liquid chromatography-mass spectrometry, Carbohydr. Res. 68 : 113-121.

21 Seymour, F. R., Knapp, R. D.~ Bishop, S. H. and A. Jeanes. 1979b. High-temperature enhancement o! C-n.m.r. chemical shifts of unusual dextrans, and correlation with methylation structural analysis, Carbohydr. Res. 68 : 123-140. Seymour, F. R., Knapp, R. D. and S. H. Bishop. 1979c. Carbon-13 spin-lattice relaxation studies for resonance assignment to specific carbon positions o! dextrans, Carbohydr. Res. 72:229-234. Seymour, F. R., Knapp, R. 1979d. Strutural analysis glucopyranosyl residues at resonance spectroscopy and Res. 7 1 : 2 3 1 - 2 5 0 .

D., Chen, E. C. M., Jeanes, A. and S. H. Bishop. of dextrans containing 2-O-a-D-glucosylated a-Dthe branch points, by use o! 13 C-nuclear magnetic gas-chromatography-mass spectrometry, Carbohydr.

Seymour, F. R., Knapp, R. D., Chen, E. C. M., Jeanes, A. and S. H. Bishop. 1979e. Structural analysis of dextrans containing 4-O-a-D- glucosylated a -D-glucopyranosyl residues at the branch points, by use o! 13C-nuclear magnetic resonance spectroscopy and gas-chromatography-mass spectrometry, Carbohydr. Res. 7 5 : 2 7 5 - 2 9 4 . Seymour, F. R. and R. D. Knapp. 1980a. Structural analysis of a-D-glucans by 13C-nuclear magnetic resonance, spin-lattice relaxation studies, Carbohydr. Res. 81 : 67-103. Seymour, F. R. and R. D. Knapp. 1980b. Structural analysis of dextrans, from strains of Leuconostoc and related genera, that contain 3-O-a-D- glucosylated a -D-glucopyranosyl residues at the branch points, or in consecutive, linear positions, Carbohydr. Res. 81 : 105-129. Seymour, F. R., Knap, R. D. and B. L. Lamberts. 1980c. Structural analysis o! soluble D-glucans from strains of Streptococcus mutans by 13C-nuclear magnetic resonance spectrometry, Carbohydr. Res. 84 : 187-195. Sharpe, E. S., Stodola, F. H. and H. J. Koepsell. 1960. Formation isomaltulose in enzymatic dextran synthesis, J. Org. Chem. 25 : 1062-1064.

o!

Shimamura, A., Tsumori, H. and H. Mukasa. 1982. Purification and properties of stereptococcus mutans extracellular glucosyltransferase, Biochim. Biophys. Acta 7 0 2 : 72-80. Stetten, D. Jr. and M. R. Stetten. 1960. Glycogen Metabolism, Physiol. Rev. 40 : 505-537. Stodola, F. H., Koepsell, H. J. and E. S. Sharpe. 1952. A new disaccharides produced by Leuconostoc mesenteroides, J. Am. Chem. Soc. 74 : 3202.

22 Stodola, F. H., Sharpe, E. S. and H. J. Koepsell. 1956. The preparation, properties, and structure of the disaccharide leucrose, J. Am. Chem. Soc. 78 : 2514-2518. Su, D. and J. F. Robyt. 1994. Determination of the number of sucrose and acceptor binding sites for Leuconostoc mensenteroides B-512FM dextral~sucrase, and the confirmation of the two-site mechanism for dextran synthesis, Arch. Biochem. Biophys. 3 0 8 : 471-476. Swanson, M. A. and C. F. Cori. 1948. Relation of structure to activation o! phosphorylases, J. Biol. Chem. 1 7 2 : 815-824. Tsuchiya, H. M., Hellman, N. N., Koepsell, H. J., Corman, J., Stringer, C. S., Rogovin, S. P., Bogard, M. O., Bryant, G., Feger, V. H., Hoffman, C. A., Senti, F. R. and R. W. Jackson. 1955. Factors affecting molecccular Weight o! enzymaticaUy synthesized dextrans, J. Am. Chem. Soc. 77 : 2412-2419. Tsuchiya, H. M. 1960. Dextransucrase, Bull. Soc. Chim. Biol. 42 : 1777-1787. Van Cleve, J. W., Schaefer, W. C. and C. E. Rist. 1956. The structure of NRRL B-512 dextran, methylation studies, J. Am. Chem. Soc. 78 : 4435-4438. Wilham, C. A., Alexander, B. H. and A. Jeanes. 1955. Heterogeneity in dextran

preparations, Arch. Biochem. Biophys. 5 9 : 61-75. Yamashita, Y., Hanada, N., Itoh-Andoh, M. and T. Kakehara. 1989. Evidence for the presence of two distinct sites of sucrose hydrolysis and glucosyl transfer activities on 1, 3-c[-D-glucan synthase of Streptococcus mutans, FEBS Letters 243:343-346. Yamauchi, F. and Y. Ohwada. 1969. Synthesis of oligosaccharides by growing

culture of Leuconostoc mesenteroides. Part IV, oligosaccharide formatiol~ in the presence of various types of glucobioses as acceptors, Agr. Biol. Chem. 33 : 1295-1300.

Enzymes for Carbohydrate Engineering K.H. Park, J.F. Robyt and Y-D. Choi (Editors) 9 1996 Elsevier Science B.V. All rights reserved.

23

Cyclodextrin Producing Enzyme (CGTase) Shoichi Kobayashi Carbohydrate Laboratory, National Food Research Institute Ministry of Agriculture, Forestry and Fisheries 2-1-2, Kannondai, Tsukuba, Ibaraki, 305 J a p a n

Introduction Formal name of cyclodextrin producing enzyme is cyclomaltodextrin glucanotransferase [1,4-a-D-glucan 4-c~-D-(1,4-c~-glucano)-transferase(cyclizing), EC 2.4.1.19, CGTase], which catalyzes the formation of cyclodextrin(CD) from starch by cyclization reaction, the formation of various open chain oligosaccharides by coupling reaction in the presence of acceptors such as glucose and sucrose, and the formation of various DPs of cF1,4-glucan from an ~-l,4-glucan by disproportionation reaction. Furthermore, the enzyme has starch hydrolyzing activity. All of these reactions are transglycosylation, in which cyclization is intramolecular, coupling and disproportionation are intermolecular, and hydrolysis is transferation of sugar to H20. As in Table 1A, 1B, B. macerans was discovered by Schardinger as CD producing bacterium, CDs were named after this achievement as Schardinger dextrins, and cell free enzyme was first discovered by Tilden and Hudson[I]. Since then, cyclization, coupling and disproportionation were found by Freudenberg et al.[2], Cori, Myrback[3], French et al.[4], Norberg and French, Pazur[5] one after another. But it has been always noted that if these three actions proceed on the same active site or not, and that hydrolyzing activity might be on contaminated enzyme, and that CGTase has not been crystalized because of its poor purity. And finally, B. macerans CGTase could be crystalized[6], and by use of the crystal enzyme and the same reaction condition, cyclization proceeds effectively with surface active agents such as SDS, on the other hand, hydrolysis depressed[7]. At present, it is generally accepted that CGTase has 4 activities of cyclization, coupling, disproportionation and hydrolysis. As for the main reaction(cyclization), the enzymes are classified 3 types such as cFCD, ~-CD and ~-CD producing. Table 2 shows properties of various CGTases. c~-CD producing enzymes are from B. macerans[8], B. stearothermophilus[9] and Klebsiella oxytoca[lO], ~-CD producing are from B. circulans[ll], B. licheniformis[12], B. megaterium[13], B. ohbensis[14], alkalophilic Bacillus sps.[15] and Thermoanaerobacter sp.[16], and ~'-CD producing enzymes are from B. firmus 290-3[17], B. subtilis No.313118], Bacillus sp. AL6119]. CGTase, which produces mainly ~-CD, seems to exist widely. The number of ~'-CD producing enzyme is increasing, partly because of ~'-CD's high solubility, its large cavity and its potential usefulness. In all of these CGTase producing bacteria, B. macerans was selected for our research, and cultivated to obtain the enzyme containing broth.

24

T a b l e 1A H i s t o r y of C D a n d C G T a s e Year

CD

research Enzyme

Name

1891

Isolation of ~-CD

Villiers

1908

Isolation of a and ~-CD

Schardinger

1904

Isolation of B.macerans

Schardinger

1912

a,~-Series

Pringsheim Karrer Miekeley

1934

Fractionation scheme (a,7-CD)

Freudenberg

1939

Cyclic formula (a- 1,4 linkage)

1942

a=6,~=7 D-glucose units

CD formation mechanism

Freudenberg

Discovery of B.macerans amylase Tilden French Transglucosylation

1945 1948-50 7=8 D-glucose units

Coupling reaction

1948 1950

(new CD and new strain of B.macerans) Disproportionation Model for artificial enzyme

1957 1968

Production of ~-CD

1983

Production of CD syrup

R.E.Gordon et al. Okada and Kitahata Horikoshi Hayashibara Co.,Ltd. Teijin Co.,Ltd.

Production of Alkalophilic B.sp. enzyme

Nihon Shokuhin Kako Co.,Ltd. Ensuiko Sugar Refining Co.,Ltd. Amano Pharmaceutical Co.,Ltd.

Production of macerans enzyme 1991

Production of 7-CD

1992 Production of CDs

Cramer

B.circulans B.megaterium enzyme Alkalophilic B.sp. enzyme B.stearothermophilus enzyme ~-CD sample supply

1976

French et al. Akiya and Watanabe Norberg and French Pazur CPC Co.,Ltd.

Test production of ~-CD

1973

Cori, Myrback Freudenberg

Wacker-Chemie Co.,Ltd. Production of 7-CD producing enzyme by use of B.firmus gene introduced B.subtilis Production of CGTase by use of Novo Nordisk Denmark Co.,Ltd. Thermoanaerobacter gene introduced Bacillus Novo Nordisk U.S.A.

25 Table 1B History for CD research in Carbohydrate Lab. 1968

Preparation of Cyclodextrin Producing Enzyme(CGTase) and Cyclodextrins(CDs) in Tokyo Noko Univ. 1970 Employed in Ministry of Agriculture and Forestry. 1973 Analysis of CDs Glucoamylase method, Paperchromatography Action pattern of CGTase Pathway of CDs formation, Effect of helical structure on cyclization 1975 Preparation of CDs in a large scale Action mechanism of CGTase and its application to CD production 1977 Preparation of c~-CD, branched- and HE-CDs 1978 Purification and some properties of CGTase Crystallization of CGTase Possible uses of branched-CDs 1979 Preparation of branched-CDs Properties of branched-CDs 1978 Production of CDs 1981 Production of CD syrup by starch adsorbed CGTase Production of CDs by a continuous process 1980 Summary of the enzyme action Enzyme model proposition 1981 Production of branched-CDs 1982 Preparation of large ringed-CDs Structure of large ringed-CDs 1986 Preparation of branched-glucan branched-CDs 1987 Preparation of neotrehalose and centose by the action of CGTase Preparation of doubly branched-CDs 1988 Preparation of doubly glucosyl branched-CDs 1989 Conversion of CGTase action by immobilization 1991 Separation of (G1)2-c~-CD~, (G1)2-a-CDAc and (G1)2-ct-CDAD

26

Table 2 Properties

of various

Origin

B. circulal~s B. firmus 290-3 B. licheniformis (?) B. macerans

CGTases Optimum pH

pH Stability

Optimum Temperature

-

5.2-5.7

7.0-9.0

55

75,000

6.0-8.0

-

Molecular weight

72,000>~ 2

5.5

75,000

5.2-5.7

H e a t Stability -Ca +Ca(~ 50

65-70 8.0-10.0

6O

7.0-10.0

55

5O 55

B. megaterium

75,000

5.2-6.2

B. ohbensis

35,000

5.5

6.5-9.5

6O

55

B. stearothermophilus

68,000

6.0

8.0-10.0

70-75

65

70

B. subtilis No.313

64,000

8.0

6.0-8.0

65

5O

50

88,000

7.0

6.0-8.0

50

60

70

45

65

65

90-95

75

BaciUus sp.(alkalophilic) Neutral-

88,000

4.5-4.7

6.0-10.0

Bacillus sp.AL6

Acid-

45,000

7.0

6.0-10.7

Klebsiella oxytoca

69,000

6.0-7.2

5.0-7.5

Thermoanaerobacter sp.

75,000

6.0

-

Origin

B. circulans

Ratio of CD formation a-CD : B-CD :~'-CD

45

Total Yield(%)

1.0 : 6.4 : 1.4

B. firmus 290-3

~-CD ) B-CD

B. licheniformis(?)

0 : 5.0 : 1.0

50

B. macerans

5.7 : 1.0 : 0.4

50-60

B. megaterium

1.o:6.3:1.3

B. ohbensis B. stearothermophilus

0:

5.0:1.0

Discovered by

Year

Gordon et al.

1973

O k a d a and K i t a h a t a

1973

Wacker-Chemie Co.

1991

Allelix Co., Ltd.

1987

Tilden

1942

O k a d a et al.

1972

Mercian Co.

1986

H a y a s h i b a r a K.K.

1973

Kato, Horikoshi

1986

1976

1.7 : 1.0 : 0.3

62

7-CD only

5

Neutral-

~-CD ) ~-CD, ~'-CD

75-80

N a k a m u r a , Horikoshi

Acid-

~-CD ) cFCD, y-CD

73

N a k a m u r a , Horikoshi

1976

Ozaki

1986

Bender

1977

Novo Nordisk Co.

1992

B. subtilis No.313 Bacillus sp.(alkalophilic)

BaciUus sp.AL6 Klebsiella oxytoca Thermoanaerobacter sp.

0:1.0

: 2.7

~-CD > ~-CD, ~,-CD I~-CD ) ~ - C D > ~-CD

(Data collected by Kitahara and Kobayashi)

35

27

P u r i f i c a t i o n of

macerans

CGTase

1. Starch adsorption and desorption[6] The heat-moisture treated starch was added to the crude enzyme p r e p a r a t i o n ( p H 7.5-8.5) and the suspension was stirred for 15hr at 4~ The starch(5g) adsorbed 90% of the activity from 5,000U in the crude enzyme solution. After desorption, the starch was reused to sorb r e m a i n i n g enzyme. Residual activity(4%) was recovered by repeated starch adsorption, with 5g of the starch. The starch t h a t adsorbed enzyme on the first t r e a t m e n t was used for further purification. The starch, containing adsorbed enzyme, was washed 4 times with 250ml portions of cold 33% ethanol solution, to remove substances having no enzyme activity. Finally, the ethanol solution was filtered t h r o u g h a sintered-glass filter. Only 0.1-0.2% of the activity was lost during the washings, providing an extremely efficient purification of the enzyme. In order to desorb the enzyme, the washed starch was suspended in distilled water, stirring at 50~ During the initial 15min of stirring, 65% of the initially applied activity was desorbed, and the curve then leveled off. To recover more enzyme from the adsorbed starch, the desorption process was repeated by using fresh distilled water. After the second t r e a t m e n t , - 8 0 % of the initial activity had been recovered. The extract, in which the specific activity of the enzyme had increased 23-fold, was almost colorless and clear. 2. Column chromatography on DEAE-cellulose. Partially purified enzyme solution(200 ml, 172 mg of protein) obtained by starch adsorption, was applied to a DEAE-cellulose column, and the enzyme was eluted from the column with a gradient system of sodium chloride(0-0.5M) in 50mM acetate buffer(pH 6.0) containing mM calcium chloride. The active fractions were combined and dialyzed against 50raM acetate buffer containing mM calcium chloride(pH 6.0) and then rechromatographed. 3. Crystallization The active fractions were collected and concentrated to - 3 - 4 % of protein in collodion bags by dialysis against mM calcium chloride solution under diminished pressure. Solid a m m o n i u m sulfate was slowly added to the concentrated enzyme solution to 10% saturation. The solution was kept at 3-5~ Within a week, the enzyme was observed to have crystallized. The crystals of the enzyme are rod-like shape. By use of the crystal CGTase preparation, we have been studying on the action of the enzyme. A c t i o n of C G T a s e

1. ~-CD accumulation[20] Up to the present, it is well known t h a t starch solution reacted under intense reaction condition to form mainly ~-CD. To elucidate this

28 phenomenon, various maltooligosaccharides were reacted with CGTase. A qualitative result is shown in Table 3, and a-CD formation was predominant, the longer substrates chain became, the more the formation degree of CDs were, though the formation degree of Gg-G,, was slightly lower t h a n those of G8 and G,2. As for B-CD, the clear spot was detected at GT-G,,, and the formation degree was fairly high at G9 and G10. Table 3 CD formation from various DP of substrates G, a-CD

-

~-CD

-

G~

G3

G4

G5

G6

G7

G8

G9

G,0

G,,

G,2

•

+

+

++

+++

++++

+++

+++

+++

+++

+

+

+

++

++

+

+

- " n o formation • " v e r y slight formation + - + + + + 9 degree of formation

Rate of Coupling action is shown in Table 4, a-CD was effectively coupled with G2 to form straight chained maltooligosaccharides, but the reaction rate of B-CD was considerably low. Disproportionation was considered to be at random reaction, but, a certain p a t t e r n was found in the reaction course, and its main pattern was 2Gn ~ G2n-2 + G2. Reaction rate of hydrolysis was 15/10,000 of that of cyclization. By the combination of all of these results, we proposed schematic explanation of ~-CD accumulation as in Figure l. Table 4 Rate of coupling (Coupled CD 9 ~M/min/mg prot.) Acceptor

G,

None

G~

G3

G4

2/3

4/3

6/3

2/3

2/3

2/3 (• 102M)

2.4

1.9

c~-CD

0.2

2.9

5.3

8.3

4.3

g-CD

0

0

0

-

O.5

2. Cyclization (intramolecular transglycosylation)[7] It was not known from which side cyclization proceeded on the substrate. To elucidate the action, various maltooligosaccharides having radioactive glucose unit on its reducing end was used. Reducing end labelled radioactive glucose, G2, G3, G4,"- were formed from reducing end labelled radioactive GT, G8, G9, G,o,..., and a-CD was detected by staining with iodine acetone solution, but radioactive CDs were not detected. Namely, cyclization is exo-wise attack at the nonreducing end of the glucans as in Figure2A. It is noted that the reaction preferably proceeds on G8 and G9,

29 and t h a t slower on GI~ and longer glucans t h a n these. By use of non radioactive maltosaccharides, the same results were obtained as in Table 3.

7

ot

ss . % B Figure 1. Schematic explanation for ~-CD accumulation. SS: Soluble starch, a: a-CD, ~: B-CD and MD: Maltodextrin.

SDS was considerably effective to proceed cyclization, the rate of a-CD formation from soluble starch with SDS(containing SDS by 10% of substrate) was 1.7 to 1 without SDS as in Table 5, also various kinds of branched CDs were formed as in Figure 2B. These results suggest t h a t cyclization proceeds on helices and helices attached with branches to from CDs and branched CDs, because SDS is effective reagent for helices formation.

Table 5 Rate of cyclization (Formed CD : ~M/min/mg prot.) Substrate

G5

G7

G8

G9

G~0

SS

-SDS

-

-

22

22

22

21

+SDS

7

12

26

33

-

36

-SDS : without SDS

+SDS : with SDS

SS : Soluble starch

2-1. The effect of surfactants on cyclization Surfactants, whose skeletal structures of hydrophobic moiety are straight 12 and 18 carbon chains, were effective for c~-CD forming. According to space filling molecular model, 65-helices fit straight carbon chain and 6 glucose units make 1 pitch. Therefore, these surfactants could be effective to form a-CD. On the other hand, surfactants having more bulky hydroptfobic moieties t h a n straight carbon chain were effective to form ~-CD.

30

*--'-')" 0 - 0 - 0 - ~ * + a-CD

O-O--Q* + a-CD 0"0"0-0"~~ ~

*"

~ * - - - ) "

> 0 " ~ * + a-CD (~* + a-CD

Figure 2A. Cyclization 1 (D* : C~4-1abelled glucose and reducing end ~) " Action point of cyclization

.O-Q

O-O.O-

Figure 2B.

Cyclization 2

2-2. Specific formation of ~- and fl-CDs It is thought t h a t specific formation of a- and B-CDs is possible by use of a- and $-CD forming surfactants respectively. If a substrate, which is r e s i s t a n t to the action of CGTase and from which any CD is not formed in the absence of surfactants, is obtainable, the substrate will be used to elucidate the m e c h a n i s m of the enzyme, and from this standpoint, aamylase and B-amylase treated waxy corn starch was prepared. By use of the substrate, without surfactant, trace of a-CD was formed, on the contrary, by the addition of surfactant such as SDS, a considerable a m o u n t of ct-CD was formed. $-CD was formed by the addition of surfactant which have bigger hydrophobic moiety t h a n t h a t of a-CD forming surfactant. T h a t is, hydrophobic moiety may act as main factor to form a- and $-CD through forming different helices.

31

2-3. Effect of S D S on hydrolyzing activity It is well known that SDS inhibits endo attack amylases. This has been thought because the enzymes become difficult to attack the inner part of substrates which take helical structure with SDS. And SDS was very effective in repressing the hydrolyzing activity of CGTase. This result suggests that CGTase attacks loose parts of the molecules to transfer the part to water, and when the molecules take helical structure, cyclization proceeds predominantly from non-reducing end. 2-4. Effect of S D S on the enzyme protein There is another possibility, the enzyme may change its conformation with SDS, and may show different action patterns. UV and circular dichroism spectra were used to solve this problem. The enzyme solutions(1.12mg/3ml of water) with and without SDS(final concentration 0.2%) were prepared, kept at room temperature for 15hr, and took UV spectra by use of SM-401(Union Giken, Kyoto, Japan). For circular dichroism spectra, 1.12mg protein/2ml of water with and without SDS(final concentration 0.2%) were prepared, and treated the solutions as same as UVs. The apparatus was J-20(Jasco, Tokyo, Japan). No difference of the UV spectra of the enzyme was observed between with and without SDS. The circular dichroism spectra were also the same in the range of 250 to 300nm, but different in the range of 190 to 250nm. The ORD(optical rotary dispersion) spectrum of the substrate with SDS was different from the one without SDS, probably due to the formation of helical complex. The authors infer that cyclization proceeds on helices from non-reducing end of substrates; ~-CD is effectively formed by the addition of 65 helix forming surfactants, and B-CD is forming by the addition of 76 helix forming surfactants. The results suggest that the action pattern of CGTase depends not only on the specificity of the enzyme itself, but also on the conformation of the substrates modified by helicogenic complexing agents. Even though, the authors could not entirely neglect the effect of conformational change of the enzyme on the change of action pattern. 2-5. Production of T-CD By use of K.pneumonia CGTase, production of ~'-CD is increased as much as 19% in the presence of 200 mM sodium acetate(which may modify conformation of the substrate to be favourable for formation of ~'-CD) and by addition of bromobenzene(complexing agent) after preincubation for 7hr[21]. The addition of stevioside and glycyrrhizin as clathrate-forming compounds is greatly effective for ~'-CD formation in the case of B. ohbensis CGTase[22]. And, cyclododecanone and cyclotridecanone were most effective insoluble ~'-CD complex forming agents(purity of ~-CD in the precipitate is more than 98%). Therefore, by taking out the precipitate from the CGTase-substrate reaction system, highly pure ~-CD preparation was produced[17].

32

3. Coupling (intermolecular transglycosylation) Rate of coupling is different according to acceptors and their concentration. Among glucose and maltooligosaccharides, G2 was the most effective acceptor as in Table 4. As for the structural features of the acceptor, the effective monosaccharides and their derivatives are classified into three groups depending on the efficiency as in Figure 3[23]. D-Glucose, 6-deoxy-D-glucose, D-xylose, L-sorbose, methyl-m, and ~-D-glucosides and phenyl-c~- and B-D-glucosides are a group of most effective acceptors(Group A). 2-Deoxy-D-glucose and 3-O-methyl-D-glucose are also effective acceptors (Group B), but slightly less than those of Group A. Other sugars and sugar alcohols are little or no effective acceptors(Group C). These results show that the structure of an effective acceptor is pyranosyl type having the same configurations as glucopyranose, namely with free C2-, C3- and C4-hydroxyl groups. D-Glucuronic acid is almost ineffective acceptor.

Group A (CH3)

D-Glucose

D-Xylose

(6-Deoxy-D-Glucose)

L-Sorbose

Group B

.,L CHzOH

~L C H z O H

HO~o~OH

CN 3-O-Methyl-D-Glucose

2-Deoxy-D-Glucose Group C HO

~OH

~OH

O H

HO

D-Galactose

D-Mannose

D-Ribose

.,L CHzOH HO~"I

,A~t'10 ~L CHzOH OH

'

Transfer 9 Position of Substrates

Figure 3. Constitutional structure of various mono saccharides in aqueous solution (Reported by Dr. Kitahata).

33 CGTase transfers glycosyl residues only to the hydroxyl group at the C4 position of D-glucose, D-xylose, 6-deoxy-D-glucose, 2-deoxy-D-glucose and 3-O-methyl-D-glucose with the exception of C3-hydroxyl group of L-sorbose. D-Galactose is a very poor acceptor and the yield of transfer products to D-galactose is only 2~-3% of those to D-glucose, L-sorbose and D-xylose. However, the transglycosylation to D-galactose occurs at several hydroxyl groups of the sugar; the proportion accepted at C1-, C3- and C2-(C4-) hydroxyl groups are 26:10:1, respectively. The transglycosylation to gluco-disaccharides occurs mainly or exclusively at the C4-hydroxyl group of non-reducing end glucose residue as in Figure 4. Each CGTase has fairly different efficiency of transglycosylation in the presence of various kinds of acceptors as in Table 6, and stearothermophilus CGTase has high efficiency of transglycosylation. Recently, we found that CGTase can also transfer glucose moiety to C1, C2, C3, C6 position of glucose to form neotrehalose, nigerose, kojibiose and isomaltose as in Figure 5, but trehalose was not detected[25].

Figure 4. Transferation of glucose to various glucodisaccharides by the action of CGTase (Reported by Dr. Kitahata).

34 Table 6 The Rate of Starch Degradation by CGTase from Bacillus stearothermophilus, B. circulans and B. macerans in the Presence of Acceptors*[24]. Acceptor

B. stearothermophilus

B. circulans

B. macerans

None

100

100

100

D-Glucose

680

590

363

D-Galactose

126

114

105

D -Ribose

111

125

105

D-Mannose

126

114

105

D-Fructose

125

110

105

D-Glucosamine

113

105

100

D -Arabino se

180

117

118

L-Arabinose

155

116

106

D-Fucose

121

110

105

L-Fucose

130

115

110

L-Rhamnose

125

110

109

*Relative rate of starch degradation in the presence of acceptor to t h a t in its absence.

Figure 5.

Products from maltose by the action of CGTase.

35

Action

of C G T a s e g l u c o s e [26]

on

the

mixture

of

GI-ct-CD

and

C14-1abelled

Branched oligosaccharides formed at the initial stage of the action of CGTase on the mixture of GI-C~-CD and C~4-1abelled glucose. Fraction Bs was the m a i n radioactive oligosaccharide, and after long reaction radioactive fraction Bs, fraction B6, fraction BT, fraction Bs and higher radioactive oligosaccharides were also formed. After exhaustive reaction conditions, a non-radioactive B4 fraction which was the smallest branched oligosaccharide, was formed as in Figure 6. Figure 6 illustrates the structures of the branched oligosaccharides found in each fraction B s - B s . The structures were identified using well known action specificities of individual enzymes on each fraction B ~ - B s . Fraction B5 was found to consist of a single component. It was characterized as 64-O-a-glucosylmaltotetraose by the action of isopullulanase(IPul, [27]), which gave isomaltose and linear radioactive G3. This was also confirmed by use of PPA which degrades the linkage between reducing and the next glucose unit of 64-O-c~gl uco sylm alto t e t r a ose.

B5

B6a

B7a

B8a

B4

B6b

B7b

B7c

B8b

B8c

Figure 6. Structure of branched oligosaccharides obtained from GI-a-CD and glucose by the action of CGTase. (2)*: C14-1abelled glucose and reducing end, O: glucose residue, - : c~-1,4 linkage and ;: c~-1,6 linkage.

Fraction BG was found to consist of two components, which were mainly 64-O-a-glucosylmaltopentaose mixed with a small a m o u n t of 65-O-a glucosylmaltopentaose. Radioactive glucose was main product from fraction BG by the action of PPA. This suggests t h a t the structure of m a i n component in fraction B6 is 64-O-a-glucosylmaltopentaose. PPA can also degrade G~ to G1 though the action is very slow and thus 65-O-a glucosylmaltopentaose was considered to be a minor component of this

36 fraction. If IPul can degrade 65-O-a-glucosylmaltopentaose to isomaltose and G4 completely, the a m o u n t of this sugar in fraction B6 will be checked. Unfortunately, the action of the enzyme on B6§ .... )is considerably slow. Therefore, by using IPul, the formation of G4 merely shows the existence of 65-O-a-glucosylmaltopentaose. Branched oligosaccharide formed from faction B6 by the action of glucoamylase was only 64-O-a glucosylmaltotetraose, with after its formation, G~, G~, and G3 were formed. This result also confirms t h a t the structure of main component in fraction B6 is 64-O-a-glucosylmaltopentaose. Faction B7 was found to consist of three components, which was mainly 65-O-a-glucosylmaltohexaose mixed with a small a m o u n t of 64. and 66-O-a glucosylmaltohexaose. By the action of ~-amylase, 64-O-a-glucosylmaltotetraose was formed from faction B7. This shows t h a t this fraction contains 64-O-a-glucosyl maltohexaose. And also, by the action of IPul, a trace of G5 was formed, consequently, the fraction contains 66-O-a-glucosylmaltohexaose. To confirm the structure of the main product, the sample free from 64-O-a glucosylmaltohexaose was prepared as follows: 250~1 of fraction B7 solution and 50~1 of B-amylase solution were mixed and reacted at 40~ for 24hr. The mixture was streaked on a 5cm wide paper and an a u t o r a d i o g r a m was made and the band of non-reacted fraction B7 was removed. 300~1 of the sample solution was obtained. Branched oligosaccharide formed from the non-reacted fraction B7 by the action of glucoamylase at the initial stage of the reaction was only B6, after its formation, G3, G2, G~ were gradually formed. This result shows t h a t the m a i n structure of ~-amylase-non-reactive fraction B7 is 65-O-a-glucosyl maltohexaose. In addition the percentage of 64-O-a-glucosylmaltotetraose and 65-O-a-glucosylmaltohexaose were 11 and 84, compared with the total radioactive count as 100. The structures of oligosaccharides in fraction B8 was also d e t e r m i n e d in a similar m a n n e r on fraction B7. From fraction B8, considerable a m o u n t of 64-O-~-glucosylmaltopentaose was formed by the action of ~-amylase. The percentage of 6-O-a-glucosylmaltopentaose 4 and 66-O-a-glucosylmalto heptaose were 29 and 68, compared with the total count as 100. Thus, fraction B8 contains 66-O-a-glucosylmaltoheptaose as the major component, and 65. and 64-O-a-glucosylmaltoheptaose as the minor component. F u r t h e r m o r e 64-O-a-glucosylmaltotetraose was degraded by exhaustive action of CGTase to 63-O-a-glucosylmaltotriose and glucose, and 6~-O-a glucosylmaltotriose was completely non-reactive to the enzyme. Consequently, 63-O-a-glucosylmaltotriose is the limit dextrin for CGTase. Therefore, the smallest radioactive branched oligosaccharide was 64-O-c~glucosylmaltotetraose and the non-radioactive one was 63-O-~-glucosylmalto triose. Fraction B8 formed at initial stage of the reaction was collected and the 4 structure was studied by use of glucoamylase. 6-O-a-glucosylmaltotetraose, 65-O-c~-glucosylmaltopentaose, and 6-O-ct-glucosylmaltohexaose 6 were formed. This implies t h a t fraction B8 contains Baa,b,c(Figure 6). Accordingly, Baa m a y be easily hydrolyzed by the action of CGTase from the non-reducing end to form B7a, B6a and Bs. And in the same way, Bsh may be

37 converted to BTb and BGb, and Bsc to BTc. Thus, by disproportionation, various kinds of oligosaccharides may be formed. Radioactive G2 and G3 were formed at the initial stage of the reaction. This implies that CGTase can also hydrolyze radioactive fraction Bs from the reducing end. By the combination of the results described above, we proposed an enzyme model of CGTase active site[28] as in Figure 7. From the model, diglucosyl-CD will be not reacted with CGTase depending on the structure.

Figure 7.

Active site model of CGTase.

A c t i o n o f CGTase on the m i x t u r e of (G1)2-~-CD and glucose[29]

The mixture of glucose and (G1)2-c~-CD was reacted with CGTase, and the reaction mixture injected to obtain HPLC profiles as in Figure 8. The main product formed by the coupling action of CGTase was branched Gg(BBg). To study the structure of the branched oligosaccharides, peak 2 in Figure 8 was separated and reacted with glucoamylase, and the reaction mixture was periodically sampled and injected to obtain HPLC profiles. From the profiles, the possible structure of BB9 was 6~,64-O-c~-diglucosyl maltoheptaose and 65,64-O-a-diglucosyl maltoheptaose, because glucoamylase can more rapidly degrade a-1,4 linkages from the non-reducing end o~ glucan molecules than a-1,6 linkages. Formed branched Gs was separated, reacted again with glucoamylase, and injected to obtain HPLC profile having peaks of glucose, branched Gs, and unreacted branched Gs. Branched G7 gave an HPLC profile having glucose, branched Gs, and unreacted branched GT. From the results, branched Gs and G7 were assumed to be 6G,64-O-a-diglucosyl maltohexaose and 65,64-O-c~-diglucosyl maltopentaose, respectively.

38

Figure 8. Formation of BB9 from the mixture of glucose and (G,)~-c~-CD by the coupling action of CGTase. 1: Glucose, 2: BBg, 3:(G,)2-c~-CD and 4: (G,)3-c~-CD (?). HPLC column: Inertsil ODS-2, solvent: aqueous 1% ethanol, flow rate: 0.7mL/min and temperature: 50~

Unreacted (G,)2-c~-CD which was separated from peak 3 of Figure 8 was again reacted with CGTase on addition of glucose to obtain BB9 again, but the amount of BB9 formed in this second reaction was less than half oi that obtained in the first reaction. After repeating this treatment several times, about 37% (from (G,)~-c~-CD preparation) of completely nonreactable (G1)~-a-CD was obtained. We deduce that the structure of nonreactable (G1)2-c~-CD is an AD type of doubly branched c~-CD. According to our enzyme model, any part of the molecule of an AD type of (G,)2-a-CD can't fit on the catalytic site. A fragment of peak 2 in Figure 8 was collected and reacted with CGTase, and two new peaks(1 and 3) appeared as shown in Figure 9. Fragment of peak 1 was determined to be glucose from its HPLC retention time, peak 2 was unreacted BBg, and the molecular weight and HPLC retention time of peak 3 were the same as that of (G,)~-a-CD, the latter fragment being completely resistant to the action of glucoamylase. The molar ratio of fragments 1 and 3 was 1:1. Thus, fragment 3 was determined to be (G,)~-a-CD. Moreover, the type of the CD should be AC and AB, as described above. This reaction was reversible, and more than 50% of (G,)~-c~-CD w a s formed from BB9 by no addition of glucose. Addition of yeast itself to the reaction mixture was considerably effective in

39 increasing the yield of (G,)2-c~-CD, whereas addition of glucose decreased the yield.

A c t i o n of the m i x e d e n z y m e s y s t e m of CGTase a n d g l u c o a m y l a s e (GI)2-a-CD

on

To a solution of (G1)2-a-CD preparation was added the mixed enzyme solution of CGTase and glucoamylase, and the reaction mixture was analyzed. Only glucose was formed, and 37% of the (G,)~-c~-CD preparation remained as completely nonreactable (G1)2-c~-CD. This completely nonreactable (G,)~-c~-CD should be type AD. Fragments formed by the hydrolyzing and coupling action of CGTase were simultaneously degraded to glucose by the action of glucoamylase. Therefore, the pure AD type oi (G1)2-c~-CD was prepared by use of the combined enzyme.

Figure 9. Reversible formation of (G,)2-c~-CD from BB9 by cyclization of CGTase. 1: Glucose, 2 : B B 9 and 3: (G,)2-c~-CD. HPLC column: Inertsil ODS-2, solvent: aqueous 1% ethanol, flow rate: 0.7mL/min and temperature: 50~ Conclusion

By use of various Substrates, action of CGTase was partly elucidated, and it is known that "Action pattern could be varied by changing the substrates' conformation". Needless to say, enzyme protein takes the most important role to produce various kinds of useful food stuffs, but it should be also noted that species of products will be widely varied by changing reaction conditions.

40

Acknowledgments The author thanks Mr.Kohichi Nakashima for skilled technical assistance, Dr.Yoshiyuki Sakano for donating isopullulanase, Dr.Kyoko Koizumi for donating authentic (G1)2-c~-CDs.

References 1

Tilden, E. B. and Hudson, C. S., J. Am. Chem. Soc., 1939, 63, 29002902.

2

Freudenberg, K., Schaaf, E., Dumpert, Naturwissenschaften, 27, 850-853.

3

Cori, C. F., Federation Proc., 1945, 4, 226. Myrback, K. and Gjorling, L-G., Arkiv for Kemi, Mineralogi och Geologi, 1945, 20A(5), 1-13.

4

French, D., Pazur, J., Levin, M. L. and Norberg, E., 1948, 70, 3145.

5

Norberg, E. and French, D., J. Am. Chem. Soc., 1950, 72, 1202-1205. Pazur, J. H., 1950, Ph.D. Thesis, Iowa State College.

6

Kobayashi, S., Kainuma, K. and Suzuki, S., Carbohydr. Res., 1978, 61, 229-238.

7

Kobayashi, S. and Kainuma, K. and French, D., J. Jap. Soc. Sci., 1983, 30, 62-68.

8

Tilden, E. B., Adams, M. and Hudson, C. S., J. Am. Chem. Soc., 1942, 64, 1432-1433.

9

Shiosaka, M. and Bunya, H., Proceedings Amylase(in Japanese), 1973, 8, 43-50.

G.

and

Ploetz,

of the

T.,

1939,

Symposium

on

10 Bender, H., Arch. Mikrobiol., 1977, 111, 271-282. 11 Gordon, R. E., Haynes, W. C. and Pang, C. H-N., 1973, Table 20 Bacillus circulans Jordan pp208-211 in "The Genus Bacillus", Agricultural Handbook No.427, Agricultural Research Service, United States Department of Agriculture, Washington,D.C. Okada, S. and Kitahata, S., Proceedings of the Symposium on Amylase(in Japanese), 1973, 8, 21-27. 12 Allelix Co., Ltd.,

C&EN U.S.A. 1987, May 18, pp24-26.

13 Okada, S., Tsuyama, N. and Kitahata, S., Proceedings of the Symposium on Amylase(in Japanese), 1972, 7, 61-68.

41 14 Yagi, Y., .M. and Ishikura. T., J. Jpn. Soc. Starch Sci.(in Japanese), 1986, 33, 144-151. 15 Nakamura, N. and Horikoshi, K., Agric. Biol. Chem., 1976, 40, 735-757. 16 Norman, B. E. and Jfrgensen, S. T., Denpun Kagaku, 1992, 39, 101-108. 17 Wacker Chemie GmbH(Schmid,G.), 1991, Preparation and application of ~-cyclodextrin, in "New trends in cyclodextrins and derivatives" Editions de Sante, Paris, France, pp27-55. 18 Kato, T. and Horikoshi, K., J. Jpn. Soc. Starch Sci.(in japanese), 1986, 33, 137-143. 19 Ozaki, A., 1986, Japan Kokai Tokkyo Koho 274680. 20 Kobayashi, S. and Kainuma, K., J. Jap. Soc. Sci., 1981, 28, 132-141. 21 Bender, H., Carbohydr. Res., 1983, 124, 225-233. 22 Sato, M., Nagano, H., Yagi, J. and Ishikura, T., 1985, Japanese Patent No. JP 60,227,693, 12 November. 23 Kitahata, S., Kagaku to Kogyo (in Japanese), 1982, 56, 127-130. 24 Kitahata, S., Hara, K., Fujita, K., Nakano, H., Kuwahara, Koizumi, K., Biosci. Biotech. Biochem., 1992, 56, 1386.

N. and

25 Shiota, M. and Kobayashi, S., Cabbohydr. Res., 1991, 215, 203-209. 26 Kobayashi, S. Ashraf, H. L., Braun, P. and French, D., Carbohydr. Res., 1988, 173, 324-331. 27 Sakano, Y., Masuda, N. and Kobayashi, T., Agric. Biol. Chem., 1971, 35, 971-973. 28 French, D. and Kobayashi, S., Fed. Proc., 1980, 39, 1856. 29 Kobayashi, S. and Nakashima, K., Carbohydr. Chem., 1991, 10, 701-709.


Enzymesfor CarbohydrateEngineering K.H. Park, J.F. Robyt and Y-D. Choi (Editors) 9 1996 Elsevier Science B.V. All rights reserved.

43

Modulation of Bacillus amylolytic enzymes and production of branched oligosaccharides Tae-Kyu Cheong, Tae-Jip Kim, Myo-Jeong Kim, Yang-Do Choi ~ In-Cheol Kim 2, Jung-Wan Kim3,and Kwan-Hwa Park. Department of Food Science and Technology, ~Department of Agricultural Chemistry and Research Center for New Bio-Materials in Agriculture, Seoul National University, Suwon, 441-744, Korea. 2Samyang Genex Research Institute, Taejeon, 305-348, Korea. 3Department of Biology, University of Inchon, Inchon, 402-749, Korea. Introduction

Many interesting and useful findings of new type amylases have been reported. They catalyze hydrolysis of a-l,4-glucosidic linkages not only in starch but also in other carbohydrates such as pullulan and cyclodextrin (Kim et al., 1992 ; Bender and Wallenfels, 1961; Sakano et al., 1971; Nakamura and Horikoshi, 1976). Debranching enzymes which hydrolyze a-l,6-glucosidic linkage were found in higher plants or microorganisms (David et al., 1987; Pazur and Ando, 1960). Some amylolytic enzymes exhibit glucose transferring activity as well as hydrolyzing activity (Kim et al., 1992 ; David et al., 1987; Matsumoto and Matsuda, 1983; Kaneko et al., 1987; Imanaka and Kuriki, 1989). Several microbial amylases which produce oligosaccharides with specific lengths or structures have been reported by numerous researchers (Robyt and Ackerman, 1971; Takasaki, 1985; Kainuma, 1988). Design of specific oligosaccharide synthesis became possible due to the discovery and modification of new amylases. Maltooligosaccharides mixture, maltotetraose syrup, and branched oligosaccharide mixture are used as substitutes for sucrose and other saccharides in the food industry due to their lower viscosity, less sweet taste, and smaller freezing point depression (Osaki et al., 1988). They also can be used to prevent crystallization of sucrose in foods and are useful in controlling microbial contamination as well as retrogradation of starchy foods because they have low water activity and high moisture-retaining capacity (Kweon et al., 1994; Komoto et al., 1993). Branched oligosaccharides are of benefit in in preventing dental caries (Glor et al., 1988) and effective for the growth of Bifidobacteria (Park et al., 1992) in human intestinal tract. They are hardly digested by human body. Therefore, the low calorie content of branched oligosaccharides has a great appeal to low calorie dieters. Due to these useful properties, the production of BOS mixtures has been increased by 50-100% per year during last few years. Generally, branched oligosaccharides are manufactured by a 2-step procedure that uses a-amylase, [~-amylase, and transglycosidase on starch solution (Takaku, 1988). The starch slurry is liquefied to a degree of

P P

Table 1. Comparison of the physicochemical properties of various amylolytic enzymes Substrate Specificity Substrate Major product

Transferring activity

Reference

Maltose Panose Maltose Maltopentaose

Formation of a-1,6-linkage

Kim et al., 1992

Enzyme

Origin

M.W. (daltons)

BLMA

B. licheriiforrnis

64,000

Starch Pullulan Cyclodextrin

BLTA CDase

B. licheriiforrnis Bacillus sp.1-5

55,000 63,000

Starch Starch Pullulan Cyclodextrin

CGTase

Bacillus sp.1-5

76,000

Starch

Maltose Panose Maltose Cyclodextrin

Pullulanase

K. prieurnoriiae

66,000

Pullulan

Panose

Isopullulanase

A. ritger

62,000

Pullulan

Isopanose

Neopullulanase

B. stearotherrnophilis

62,000

Starch Pullulan

Maltose Panose

a-amylase

T. uulgaris

71,000

Starch Pullulan

BMA

B. rnegateriurn

55,000

Starch Pullulan Cyclodextrin

Maltose Panose Oligomer Panose Maltose

~

Kim, 1991

+ +

Oh, 1993

Kim, 1994 Ohba & Ueda, 1975 Sakano et al., 1971

Formation of a-1,4-, 1,6linkage

Formation of a-1,4-linkage

Kuriki et al.. 1988 Sakano et al., 1982 David et al., 1987

45 hydrolysis (DE) 6-10 by the action of a thermostable c~-amylase, then saccharified by t r e a t m e n t with soybean ~-amylase and transglycosidase of AspergiUus niger at 60~ for 72 hours. It is a time-consuming procedure. A series of amylolytic enzymes including a thermostable a-amylase (BLTA), a maltogenic amylase (BLMA), a CDase, and a CGTase were isolated from two Bacillus species,B, licheniformis and alkalophilic Bacills I-5 strain, in our laboratory. The genes responsible for them were cloned in E. coli and the physicochemical properties and the action patterns of these amylases on various carbohydrates were characterized. Recently, we developed a one-step procedure which simplified and shortened the reaction for the production of branched oligosaccharides. It applies BLTA and BLMA to starch solution simultaneously (Kim et al., 1994). The two amylases were prepared from an E. coli transformant harboring a recombinant plasmid which carried both of genes for the enzymes. The reaction was carried out for 24 hours and 45% of the product was branched oligosaccharides of various lengths. The amount of branched oligosaccharides in the mixture was comparable to that of the Alo (anomalously linked oligosaccharide) mixture in market (Takaku, 1988). However, it contained much less glucose than the Alo mixture (27% vs. 40%). Development of a continuous process of branched oligosaccharides production using immobilized BLMA and yeast cells would also improve the efficiency of the whole process and the quality of the product. Modification of the enzymes by recombinant DNA techniques would make application of the amylases to food industry more diverse and effective. Improving thermostability and inducing secretion of BLMA would be of significant benefit in applying the enzyme to the production of branched oligosaccharides. The predicted amino acid sequences indicated that they share homology with various amylolytic enzymes at four conserved active domains (Ihara et al., 1985; Ryoichi et al., 1986). Based on amino acid sequence comparison to that of Taka amylase, whose tertiary structure is known (Matsuura et al., 1984), BLTA and BLMA are likely to have the structure of 8 (~/a) barrels with intervening loops (MacGreger, 1993; Jespersen et al., 1993). Mutagenesis of the genes responsible for the enzymes is under investigation as an effort to improve the process for the production of oligosaccharides. Results and Discussion

Catalytic properties of starch degrading enzymes Physicochemical properties of the starch hydrolyzing enzymes isolated in the laboratory and those reported by others are summarized in Table 1. Substrate specificity of BLMA was compared with that of BLTA. Soluble starch, pullulan, and cyclodextrin were hydrolyzed by BLMA, while BLTA hydrolyzed soluble starch only. Optimum temperature of BLTA was 70~ and it was increased to 90~ in the presence of 5raM Ca +* (Kim, 1991). Optimum temperature of BLMA was 50~ (Kim et al., 1992). The activity of BLMA was inhibited by most of divalent cations such as Ca ++, Zn §247and Mn §247 CGTase produced a-, B-, ~-cyclodextrins at the ratio of 0.7 : 3.3 : 1.0

46 from starch in the mother strain. The ratio among various CD products was changed when the enzyme prepared from recombinant DNA in E. coli (see below) was used. CDase hydrolyzed cyclodextrin very rapidly into maltose. BLMA and CDase exhibited transferring activity in addition to hydrolyzing activity in the presence of excess glucose. A BOS mixture is prepared from liquefied starch using the hydrolyzing and the transferring activities of BLMA (Kim et al., 1994). The BOS mixture was analyzed by various methods including high performance liquid chromatography (HPLC) and found to contain various small branched sugars such as isomaltose, panose, and isopanose etc. Based on the results, a model of BLMA action mode was proposed as shown in Fig. 1. BLMA is likely to synthesize branched oligosaccharides in a complicated m a n n e r : it hydrolyzes liquefied starch further to glucose and maltose, and at the same time transfers the resulting molecules onto cleavage sites of other sugar moieties by forming c~-l,6-1inkages. The whole reaction is completed by repeating the coupled hydrolysis and transfer reactions, thereby creating a new population of branched oligosaccharide molecules.

Pullulan

4, Starch

s ~

/ t ~' "" " ~ l

%

i

Figure 1. A proposed model of BLMA action mode involved in the production of branched oligosaccharides from pulllulan or maltooligosaccharide. BLMA hydrolyze pullulan to panose and the resulting molecule is transferred to acceptor (in this case glucose) by forming a-l,6-glycosidic linkage. Maltooligosaccharide is hydrolyzed mainly to maltose, but also to glucose or maltotriose etc. They are then transferred to acceptor molecules such as glucose, maltose, or maltotriose. Resulting branched molecules are likely to be hydrolyzed further by BLMA.

47

Continuous production of the BOS mixture The procedure for continuous production of the BOS mixture and a highly concentrated BOS (high BOS) mixture is illustrated in Fig. 2. BLMA has an optimal temperature of 50~ and its thermostability could be enhanced by immobilization on CPC Silica. Decimal reduction time of immobilized BLMA at 55~ was 96.8 min, while that of free BLMA at the same temperature was only 20.5 min. Thirty percents (w/v) corn starch suspension in 50mM maleate-NaOH buffer (pH 6.8) was liquefied by thermostable c~-amylase (Termamyl ; Novo Nordisk, Denmark). The sample was heated in a boiling water bath for 10 min and the reaction was stopped by autoclaving the hydrolysate for 10 min when the DE value was about 20-25. Five millimoles of EDTA was added to the liquefied solution and its final pH was adjusted to 6.8 using 1N NaOH. For the synthesis of the BOS mixture, the liquefied starch was run through a column of immobilized BLMA at 45-50~ Residence time of the

Starch slurry

l

30% Soluble starch

Liquefaction

~-

alpha-Amylase 100~ 10min

Saccharification

--

45~

}_

~-

BLMA

15hrs

Filtration

BOS

i

Yeast fermentation ~-

I_.

Immobilized Yeast

27~ 2days Freeze drying

High B O S

Figure 2. mixtures.

Procedure

for

the

production

of the

BOS

and

high

BOS

48 liquefied starch solution in the column was 2.5 hours. It took 24 hours at 40~ to prepare a BOS mixture with similar composition when free BLMA was used. The reaction was stopped by boiling, filtered through a W h a t m a n paper No. 5, and then dried using a freeze-drier. The BOS mixture was analyzed by HPLC (Fig. 3) and its composition is shown in Table 2. It contained BOS over 60% including panose, branched DP4, and branched DP5 etc. The shoulder of the maltose peak indicated that isomaltose might be present in the mixture. Further analysis of the mixture using ion-chromatography showed that isomaltose and isopanose were also present in the mixture (data will be discussed elsewhere). The BOS mixture contained about 30% glucose and maltose. In order to get rid of sweet taste of the mixture, it was fermented by yeast. As the result of glucose and maltose fermentation by yeast, BOS would be concentrated to make a high BOS mixture. For the preparation of a high BOS mixture, the BOS mixture (30%; w/v) prepared as described above was fermented by Saccharomyces cerevisiae var. ellipsoideus that was immobilized on sodium-alginate matrix. Fermentation was carried out at 27-28~ for 2 days to remove glucose and maltose contained in the BOS mixture. Upon the completion of fermentation, the mixture was filtered and dried as described above. As the result of fermentation, almost all of glucose and maltose was removed and the BOS content increased to over 90% (Table 2).

o-o +

0

aD --"

N

Retention Time (min)

A

Retention Time (min)

B

Figure 3. High performance liquid chromatography analysis of the BOS (A) and high BOS (B) mixtures. Saccharide which is likely to constitute each peak is shown above the peaks. HPLC analysis was carried out as described previously (Kim et al., 1992).

49 Table 2 The compositions of the BOS and the high BOS mixtures Saccharide

BOS mixture

High BOS mixture

Glucose

10.2 %

1.3 %

Maltose

18.9 %a

_

Isomaltose

-

2.7 %

Maltotriose

6.8 %

7.2 %

Panose b

15.4 %

20.4 %

Branched tetraose

30.4 %

43.8 %

Branched pentaose

18.3 %

24.6 %

> Branched pentaose

trace

trace

Total a m o u n t of BOS

64 %

91.5 %

a : m a y contain isomaltose b : m a y contain isopanose

Application of the BOS mixture to bread as h u m e c t a n t for starchy foods was a t t e m p t e d . It retarded starch retrogradation and lowered w a t e r activity (manuscript submitted). The high BOS mixture would be of great benefit to apply it as h u m e c t a n t to food.

Comparison of amino acid sequences of various amylases The

genes

responsible

for

BLTA

or

BLMA

were

isolated

from

B.

licheniformis by shotgun cloning EcoRI or BamHI/EcoRI genomic DNA digests into an E. coli vector, pBR322. A CDase gene was isolated from an alkalophilic Bacillus strain I-5 using the same method. The positive clones were screened for their starch hydrolyzing phenotype in E. coli HB101 using iodine test. The starch hydrolyzing activity of BLMA or CDase was observed only when cell m e m b r a n e of the t r a n s f o r m a n t carrying one of the genes was disrupted by D-cycloserine. This indicated the two enzymes were cytoplasmic proteins. In order to isolate a CGTase gene from Bacillus I-5 strain, two oligonucleotides were synthesized based on homology among CGTase genes. An 1.1kb DNA fragment was amplified from I-5 genomic DNA and used as a probe to screen a genomic DNA library constructed using EMBL3 k phage DNA (Amersham, U.S.A.). The insert of a plaque with positive signal was subcloned into pUC18. The CGTase positive phenotype of the clone was confirmed by a phenolphthalein tested (Park et al., 1989) and HPLC analysis of the product produced by the enzyme p r e p a r e d from the clone. The restriction m a p s of pIJ322, pTA322, pCGTJ322, and pTJ3, which

50 contain the BLMA, BLTA, CGTase, and CDase gene, respectively, are shown in Fig. 4. Each gene was expressed from its own promoter and stably m a i n t a i n e d in E. coli. The gene products were characterized in detail and three of them, BLMA, BLTA, and CGTase, had properties t h a t were in good correlation with those found in the mother strains. The activity of CDase has not been detected in the mother strain, alkalophilic Bacillus I-5. To analyze the structure of the BLTA and BLMA genes, nucleotide sequencing was carried out. Nucleotide sequences of the two B. licheniformis amylases were determined by sequencing the plasmids

EcoRI

EcoRI

Hindlll

Pvull

SlII

/ X

c.,

cloning

m"'

Pstl

Pvull

~

J J

EcoRI EcoRI

Hlndlll

hmHI

BIImHI

EcoRI

Pvull

Sail

oa,

BamHI

pBR322

BImHI

Smal

Hindlll

EcoRI

Figure 4. Restriction maps of pTA322, pIJ322, pTJ3, and pCGTJ322. pTA322 carries the BLTA gene on a 3.1kb insert; pIJ322, the BLMA gene on a 3.5kb insert; pTJ322, a CDase gene on a 3.2kb insert; pCGTJ322, a CGTase gene on a 4.8 kb insert. All of them are cloned on an E. coli vector, pBR322, and selected by resistance to ampicillin in E. coli.

51 carrying the genes according to Sanger's chain t e r m i n a t i o n method (Sanger et al., 1977) and using Sequenase kit purchased from U.S.B. Corp. The DNA f r a g m e n t s were sequenced in both strands. The BLTA gene coded for a 55 kD protein of 483 amino acids. The m a t u r e BLTA protein was proceeded by 29 amino acids which are likely to function as a signal sequence for secretion (Kim, 1991; Yuuki et al., 1986). The protein was localized in both periplasmic and cytoplasmic spaces. The BLMA gene was capable of encoding 584 amino acids from a promoter located 5' flanking region and the molecular weight of the gene product was predicted to be 66.5 kD. There was no signal sequence like sequence found at the amino t e r m i n a l of the gene. The protein was found only in cytoplasm of E. coli and the mother strain. The deduced amino acid sequences were aligned with those of various a m y l a s e s reported previously (Jespersen et al., 1993) and four highly conserved regions were found in Table 3. Conserved regions III and IV are considered to include active sites of amylases. The specificity of the anti-BLTA antibody was tested using Ouchterlony double immunodiffusion analysis (Jang et al., 1994). The antibody formed a precipitate with Taka-amylase, which is a thermostable amylase. However, neither pullulanase nor glucoamylase cross-reacted with the antibody. Termamyl (Novo Inc., Denmark) formed a precipitate with the antibody (data not shown). In a recent study, the antibody raised in the laboratory against another amylase of B. licheniformis, BLMA, did not cross-reacted with BLTA (Shim, 1994). Pullulanase, Termamyl, and glucoamylase are also Table 3 Comparison of amino acid sequences in the conserved domains Domains

I

II

III

IV

Reference

Enzymes

DAVINH GFRLDAAKH

EVIH

FVDNHD

BLTA

DVVINH GFRLDAVKH

EYWQ

FVDNHD Kim, 1991

BLMA

DAVFNH GWRLDVANE

EIWH

LLDSHD Kim et al., 1991

Neopullulanase

DAVFNH GWRLDVANE

EIWH

LLGSHD Kuriki & Imanaka, 1989

Pullulanase

DVVYNH GFRFDLMGI

EGWD

YVESHD Nakajima et al., 1985

CGTase

DFADNH GIRVDAVKH

EYHQ

FIDNHE Kaneko et al., 1988

ct-amylase a

DAVFNH GWRLDVANE

EIWH

LLDSHD Tonozuka et al., 1993

a: neopullulanse-type a-amyalse of T. vulgaris

52 known to have four active domains conserved among various amylases (Kim et al., 1992). The results obtained in this study suggest that the antigenic epitope determinant is located on a different portion of the protein than the conserved domains. Also, it is likely that the two amylases (BLTA and BLMA) of B. licheniformis should be easily separated from each other by immunoaffinity chromatography due to the specificity of the antibodies. AP BLTA BP DG GD IA KP BSMA BLMA CD NP OG PP TA

62 10 35 17 28 21 28 31 36 36 36 17 5 28

27 37 41 27 37 46 103 18 27 27 27 26 34 34

5 160 g 74 177 423 130 134 134 134 8 9 9

1~7

9

:3

12

21 ' lO 53 I 27 101 151 151 15 , 19 ' 25

5

o

8

~4

20 5 3 6 12 9 9

' 72 1 261

156 48 50 60 61 60 "/6 75 6,3

g

20

t

i.4

~

27 2O

271

9

[ BLTA BP DG GD IA KP BLMA BSMA CD NP OG PP TA

15 8

I105

34 37 45

27 27 14 53 27

~'JP~ R ~

Sttarcls

9i 4 4

5 2 4

4 4 4

7 4

4

3 2 1 4

1 1 1 1 1 2 1

6

6 34 6 5

5 6 6 6 6 3 6

4

7

4

6

7 2

5 5

4 4

1

~0

4

4

To

G-t~minai domtms

4

3 3 3 3 6 15 3

9 g

1 4 4

9

4

5

5

6

4

8

2

Figure 5. Comparison of numbers of amino acid residues in loops between the 8 (~/c0 barrels in BLTA, BLMA, and other amylolytic enzymes. The presence of 8 (B/a) barrels in BLTA and BLMA was predicted by aligning the amino acid sequences to that of Taka-amylase. Numbers in the boxes represent the number of amino acid residues in loops between B-strand (El-E8) and c~-helix (HI-H8 and H) or vice versa. Abbreviations for the enzymes are : AP, c~-amylase-pullulanase from C. thermohydrosulfuricum; BA, BLTA ; BP, pullulanase from B. stearothermophilus; DG, dextran glucosidase; GD, gycogen debranching enzyme; IA, isoamylase; KP, pullulanase from K. pseudomonas; MA, BLMA; MB, maize branching enzyme; NP, neopullulanase; OG, oligo-1,6-glucosidase from B. cereus; PP, porcine pancreatic mamylase; TA, Taka amyalse A. The format of the figure was cited from Jespersen et al., 1993.

53

Mutagenesis of the BLTA and BLMA genes The c~-amyalse family are known to have catalytic domain consisted of a barrel of eight parallel ~-strands surrounded by eight helices (MacGreger, 1993; Jespersen et. al., 1993). Differences in specificities of starch metabolizing enzymes are in the numbers of subsites at active and catalytic sites. Structure of such (~/c0s-barrel was likely to be present in BLTA and BLMA based on circular dichroism a n a l y s i s , sequence comparison analysis, and prediction according to Chou and Fasman's method (Fig. 5). Enzymes specific for forming a-l,6-1inkage might resemble a-amylases at subsites of active site but differ from them in loops 4 and 5. The lysine-histidine residues of loop 4 (in conserved domain II) should be important for specificity of amylases, since they are absent from enzymes that hydrolyze or synthesize a-l,6-glucosidic bonds. Therefore, mutation in one or more residues might alter the ratio of a-l,4-bond hydrolysis to a-l,6-bond hydrolysis. Among the amylolytic enzymes, only a-amylase from Aspergillus oryzae (Taka-amylase A ; Matsuura et al., 1984) and from porcine pancreas (Buisson et al., 1987) have been investigated by X-ray crystallographic analysis. Three-dimensional structures of enzymes acting on a-l,6-glucosidic linkages are yet to be elucidated. Deletion mutation of the BLTA gene Spacing among the four conserved regions of various amylolytic enzymes is compared in Table 4. A large variation among the enzymes in the spacing between conserved regions I and II was found. Spacing between the conserved region I and II of BLTA was changed using Bal 31 exonuclease after linearizing the plasmid carrying the BLTA gene at the unique KpnI site occurring between the first and the second conserved domains (Fig. 6). The DNA fragments with various lengths of deletion Table 4 Comparison of the spacing between the conserved domains Domain N-terminal 1st-2nd-3rd-- 4th/C-terminal Enzyme'~'-----_~ BLTA 101 127 33 63 159 BSA a

101

129

34

62

189

BLMA

244

78

33

62

168

Neopullulanase

242

82

33

62

169

Pullulanase GG producing amylase G4 producing amyalse

281

67

33

83

194

107

133

92

112

80

95

a: c~-amylase of Bacillus stearothermophilus (Ihara et al., 1985)

54 were ligated and transformed into E. coli. Mutants were tested for their starch hydrolyzing activity on a starch agar plate by staining with iodine solution. Four of them showed less activity than wild type BLTA. These m u t a n t s were less stable at 75~ and had narrower pH range for stability than wild type BLTA. Also, the sizes of starch hydrolysis products were changed. The mutants produced glucose and maltose from starch, while wild type BLTA produced oligosaccharides of G l-G6. This draws a hypothesis : the size of the reaction product might be closely related to the spacing between the first and second conserved regions. The spacings between the conserved domains could reflect the chemical nature of the reaction and product specificity of an enzyme.

pUCNF18

RI

Kpnl

RI

Kpnl

r--''

i

l

"

Hd RIm i Hd

Hindlll deletion

pUCNF18 A Hd

Hd

Ba131 deletion Hd

! pUCNF~I

n

i ~ pUCNF~.2 "

pUCNF~3

Figure 6. Construction of BLTA deletion mutants using Bal31 exonuclease. The 3.1kb EcoRi fragment of pTA322 harboring the BLTA gene was subcloned into pUC18 (pUNCF18). The KpnI site on the polylinker was removed by simply digesting pUCNF18 with HindIII and religating it. The resulting plasmid, pUCNF18AHd, was linearized at the unique KpnI site and treated with Ba131 to create deletion between the 1st and 2nd conserved domains, the open box represents the BLTA coding region; the line, the vector; the dark boxes, four conserved domains. The restriction sites shown are : RI, EcoRi; Hd, HindIII.

55

Figure. 7. TLC analysis of starch hydrolyzed by wild type BLTA or by the m u t a n t s (A1-A4).

Site directed mutagenesis of the BLMA gene The BLMA gene was mutated by synthetic oligonucleotide mediated site directed mutagenesis (Kunkel, 1985; Vieira and Messing, 1987). The resulting mutations were confirmed by DNA sequence analysis and activities of the mutants were tested by DNS method (Kim et. al., 1992). Three of the mutants are listed in Table 5. Glu-356 in the conserved domain III of BLMA was assumed to be an active site and was replaced with Asp by site directed mutagenesis. The m u t a n t did not exhibit the activity of wild type BLMA at all. The result indicated t h a t Glu-356 may constitute the active center of the Table 5 BLMA m u t a n t s created by site-directed mutagenesis Conserved Domain I II BLMA Wild type

DAVFNH

H250Q

.....

GWRLDVANE

III

EIWH

IX"

LLDSHD

Q

E356D D420G

- -G-

- -

55 enzyme as known for other enzymes (Kuriki et al., 1991). The mutant, H250Q, in which the histidine residue at the position 250 was substituted with glutamine retained only 10% cyclodextrin hydrolyzing activity of the wild type BLMA. The mutant, D420G, in which the aspartic acid residue at the position 420 was substituted with glycine showed only 5% cyclodextrin hydrolyzing activity of the wild type BLMA. The conserved histidine residue at 250 is known to be one of the substrate binding sites. Substitution of the residue with another basic amino acid caused significant loss of enzyme activity. Neopullulanase of B. stearothermophilus (Imanaka and Kuriki, 1989) carries out c~-l,6-glucosidic linkage hydrolysis and c~-l,4-transglycosylation reactions as well as the a-l,4-glucosidic linkage hydrolysis and a-l,6-transglycosylation reactions. BLMA shares significant homology (over 80%) with neoplullanase at the amino acid sequence level, especially around the four conserved regions. All conserved residues except one amino acid, aspartic acid at 420 (glycine in neoplullanase), are identical in these enzymes (Kuriki and Imanaka, 1989). The aspartic acid was substituted with glycine to test how the BLMA activity would be changed. In addition to decrease in activity, change of substrate specificity was observed in these mutants. The wild type BLMA hydrolyzed cyclodextrin best, then pullulan, and then soluble starch. The relative activity of the wild type BLMA on these substrate was 8 : 1 : 0.4. The ratio was changed to 10 : 1 : 1.4 and 27 : 1 : 1.1 by H250Q and D420G, respectively. The two m u t a n t s did not lose the transglycosylation activity of BLMA. More BOS was produced by the mutants (Table 6) when equal amount of pullulan hydrolyzing activity of the wild type BLMA or the m u t a n t s was added to 15% (w/v) liquefied starch solution. Especially, increase of branched DP4 molecules in the product was significant. This could be due to less efficient hydrolysis of oligosaccharides by the m u t a n t BLMA enzymes. More mutants have been created and they are under investigation in the laboratory.

Table 6 Production of BOS by the BLMA mutants Reaction 12 hours a time Enzyme BOS (%)

BDP4 (%)

36 hours a BOS (%)

BDP4 (%)

Wild Type

55.6

19.2

40.4

8.7

H250Q

54.1

19.2

56.2

17.3

D420G

56.9

25.3

58.8

25.9

a : the reaction was carried out using 15% (w/v) liquefied starch solution

57

Conclusions Four genes encoding various amylolytic enzymes were cloned from two strains. Two of them, the BLTA and BLMA genes were characterized in detail and their gene products were applied to the production of BOS. Immobilization of BLMA enabled continuous production of BOS at higher temperature than the process using free BLMA. Fermentation of the BOS mixture using immobilized yeast was useful to remove glucose and maltose, and the resulting high BOS mixture contained 91.5% BOS. DNA sequencing analysis of the BLTA and the BLMA genes suggested that the enzymes are likely to have the (~/~)s structure common to many amylolytic enzymes. Deletion mutation of the BLTA and BLMA enzymes indicated that the spacing between conserved domain I and II is important in determining the enzyme activity and the reaction product. Histidine at 250 and aspartic acid at 420 of BLMA were important for the hydrolysis activity of the enzyme. However, the transferase activity of the enzyme was not changed by the mutation introduced to the residues. The m u t a t e d enzymes showed preference to starch over pullulan as substrate. This would be of benefit for controlling the production of oligosaccharides.

Bacillus

References Bender, H., and K. Wallenfels. 1961. Untersuchungen an pullulan. II. Spezifisher abbau durch ein basterielles enzyme. Biochem. Z. 334 : 1913-1920 Buisson, G., E. Du~e, R. Haser, and F. Payan. 1987. Three dimensional structure of porcine pancreatic a-amylase at 2.9/k resolution, role of calcium in structure and activity. EMBO J. 6:3909-3916. David, M. H., H. Gunter, and H. Roper. 1987. 39:436-440

Catalytic properties of

Bacillus megaterium amylase. Starch

Glor, E. B., C. H. Miller, and D. F. Spandan. 1988. Degradation of starch and hydrolytic products by oral bacteria. J. Dent. Res. 67 : 75-81. Ihara, H., T. Sasaki, A. Tsuboi, H. Yamakata, N. Tsukagoshi, and S. Udaka. 1985. Complete nucleotide sequence of a thermophilic a-amyalse gene : homology between procaryotic and eukaryotic a-amyalses at the active site. J. Biochem. 9 8 : 9 5 - 1 0 3 . Imanaka,

T.

and

T.

Kuriki.

1989.

Pattern of action of Bacillus J. Bacteriol. 171 : 369-374.

stearothermophilus neopullulanase on pullulan.

Jang, S. Y., T. K. Cheong, W. Shim. J. W. Kim, and K. H. Park. 1994. Purification of Bacillus licheniformis thermostable c~-amylase by immunoaffinity chromatography. Korean Biochem. J. 27 : 38-41.

58 Jespersen, H. M., E. A. MacGregor, B. Henrissat, M. R. Sierks, and B. Svensson. 1993. Starch- and glycogen-debranching and branching enzymes : prediction of structural features of the catalytic (B/c0s barrel domain and evolutionary relationship to other amylolytic enzymes. J. Prot. Chem. 12 : 791-805 Kainuma, K. 1988. a-Amylases which produce specific oligosaccharides. In "Handbook of Amylases and Related Enzymes: Their Sources, Isolation Methods, Properties and Applications," The Amylase Research Society of Japan, Osaka, Japan (Ed.). pp 50-62, Pergamon Press, New York Kaneko, T., T. Hamamoto, and K. Horikoshi. 1988. Molecular cloning and nucleotide sequence of the cyclomaltodextrin glucanotransferase gene from the alkalophilic Bacillus sp. strain No. 38-2. J. Gen. Micro. 134 : 97-103. Kaneko, T., T. Kato, N. Nakamura, and K. Horikoshi. 1987. Spectrophotometric determination of cyclization activity of beta-cyclodextrin-forming cyclomaltodextrin glucanotransferase. J. Japan. Soci. Starch. Sci. 3 4 : 45-48. Kim, I. C. 1991. Molecular cloning of thermostable a-amylase and maltogenic amylase from Bacillus licheniformis and characterization of their enzymatic properties. PhD thesis, Seoul National University Kim, I. C., J. H. Cha, J. R. Kim, S. Y. Jang, B. C. Seo, T. K. Cheong, D. S. Lee, Y. D. Choi, and K. H. Park. 1992. Catalytic properties of the J. Biol. Chem. 267: cloned amylase from Bacillus licheniformis. 22108-22114. Kim, I. C., S. H. Yoo, S. J. Lee, B. H. Oh, J. W. Kim, and K. W. Park. 1994. Synthesis of branched oligosaccharides from starch by two amylases cloned from Bacillus licheniformis. Biosci. Biotech. Biochem. 58 : 416-418 Kunkel, T. A. 1985. Rapid and efficient site-specific mutagenesis without phenotypic selection Proc. Nat. Acad. Sci. 82:488-492 Kuriki, T., H. Takata, S. Okata, and T. Imanaka. 1991. Analysis of the active center of Bacillus stearothermophilus neopullulanase. J. Bacteriol. 173:6147-6152. Kuriki, T., T. Imanaka. 1989. Nucleotide sequence of the neopullulanase gene from Bacillus stearothermophilus. J. Gen. Microbiol. 135 : 1521-1528. Kweon, M. R., C. S. Park, J. H. Auh, B. M. Cho, N. S. Yang, and K. H. Park. 1994. Phospholipid hydrolysate and antistaling amylase effects on retrogradation of starch in bread. J. Food. Sci. 59 : 3953-3958 MacGreger, E. A. 1993. Relationships between structure and activity in the a-amylase family of starch-metabolizing enzymes, Starch 45: 232-237.

59 Matssura, Y., M. Kusunoki, W. Harada, and M. Kakudo. 1984. Structure and possible catalytic residues of Taka-amylase. J. Biochem., 95 : 697-702. Matsumoto, A., and K. Matsuda. 1983. Role of branching enzyme in glycogen biosynthesis in Neurospora crassa. J. J a p a n Soc. Starch Sci., 30 : 212-222. Nakajima, R., T. Imanaka, and S. Aiba. 1985. Nucleotide sequence of the Bacillus stearothermophilus ~-amylase gene. J. Bacteriol., 163 : 401-406. Nakamura, N. and K. Horikoshi. 1976. Purification and properties of neutral cyclodextrin glucanotransferase of an alkalophilic Bacillus sp. Agri. Biol. Chem. 40 : 1785-1790. Osaki, S., Z. Yoshino, Y. Tsujisaka, H. Takaku. 1988. Manufacture of Oligosaccharides. In "Handbook of Amylases and Related Enzymes: Their Sources, Isolation Methods, Properties and Applications," The Amylase Research Society of Japan, Osaka, Japan (Ed.). pp 210-217, Pergamon Press, New York Park, C. S., K. H. Park and S. H. Kim. 1989. A rapid screening method for alkalophilic ~-cyclodextrin glucanotransferase using phenolphthalein methylorange containing solid medium. Agri. Biol. Chem. 53: 1167-1169. Park, J. H., J. Y. Yoo, O. H. Shin, H. K. Shin, S. J. Lee, and K. W. Park. 1992. Growth effect of branched oligosaccharides on principal intestinal bacteria. Kor. J. Appl. Microbiol. Biotechnol. 20 : 237-242. Pazur, J. H., and T. Ando. 1960. The hydrolysis oligosaccharide with a-D-(1,4) and a-D-(1,6) bonds amyloglucosidase J. Biol. Chem. 235 : 297-302.

of by

glycosyl fungal

Robyt, J. R. and R. J. Ackerman. 1971. Isolation, purification, and characterization of a maltotetraose-producing amylase from Pseudomonas stutzeri. Arch. Biochem. Biophys. Acta. 145 : 105-114. Ryoichi, N., T. Imanaka, and S. Aiba. 1986. Comparison of amino acid sequence of eleven different a-amylase. Appl. Microbiol. Biotechnol. 23 : 355-360 Sakano, Y., N. Masuda, and T. Kobayashi. 1971. Hydrolysis of pullulan by a novel enzyme from Aspergillus niger. Agric. Biol. Chem. 35:971-973 Sanger, F., S. Nicklen and A. R. Coulson. 1977. DNA sequencing with chain terminating inhibitors. Proc. Natl. Acad. Sci. USA. 74:5463-5467 Shim, W. 1994. Purification of Bacillus licheniformis maltogenic amylase by Immunoaffinity chromatography, MS thesis, Seoul National University.

60 Takaku, H. 1988. in "Handbook of Amylases and Related Enzymes: Their Sources, Isolation Methods, Properties and Applications," The Amylase Research Society of Japan, Osaka, Japan (Ed.), Pergamon Press, New York, 1988, pp.215-217. Takasaki, Y. 1985. An amylase producing maltotetraose and maltopentaose from Bacillus circulans. Agric. Biol. Chem. 47 : 2193-2199. Tonozuka, T., M Ohtsuka, S, Mogi, H, Sakai, T. Ohta, and Y. Sakano. 1993. A neopullulanase-type ~-amylase gene from Thermoactinomyces vulgaris R-47. Biosci. Biotech. Biochem. 57 : 395-401 Vieira, J. and J. Messing. 1987. Production of single-stranded plasmid DNA. Methods in Enzymology. 153 : 3-11. Yuki, T., Nomura, T., Tezuka, H., Tsuboi, A., Yamagata, H., Tsukagoshi, N., and Udaka, S. 1985. Complete nucleotide sequence of a gene coding for heat- and pH-stable a-amylase of Bacillus licheniformis : comparison of the amino acid sequences of three bacterial liquefying a-amylase deduced from the DNA sequences. J. Biochem. 98:1147-1156.

Enzymesfor CarbohydrateEngineering K.H. Park, J.F. Robyt and Y-D. Choi (Editors) 1996 Elsevier Science B.V.

61

A N o v e l M a l t o t e t r a o s e - F o r m i n g Alkaline a - A m y l a s e from an A l k a l o p h i l i c B a c i l l u s Strain, GM8901 Yong Chul Shin 1 and Si Myung Byun 2 ~Department of Microbiology, Gyeongsang National University, Chinju 660-701, Republic of Korea 2Department of Life Science, Korea Advanced Institute of Science and Technology, Daejeon 305-701, Republic of Korea

Abstract An alkalophilic bacterium, Bacillus sp. strain GM8901, grown at pH 10.5 and 50~ produced five alkaline amylases in culture broth. At an early stage of the bacterial growth, amylase I (Amyl I) was produced initially and then, as cultivation progressed, four alkaline amylases, Amyl II, Amyl llI, Amyl IV, and Amyl V, were produced from proteolytic degradation of Amyl I. A serine protease present in the culture medium was believed to be involved in Amyl I degradation. We purified Amyl I from the culture supernatant by ammonium sulfate precipitation, heparin-Sepharose CL-6B column chromatography, phenylToyopearl column chromatography, and Mono Q HR5/5 high-performance liquid chromatography. The molecular weight of Amyl I was estimated to be about 97,000 by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Amyl I had an extremely high optimal pH of 11.0 to 12.0 and was stable in a broad pH range of 6.0 to 13.0. Amyl I had an optimal temperature of 60~ and was stable up to 50~ Thermostability was increased in the presence of Ca 2§ and soluble starch. The enzyme required metal ions such as Ca 2§ Mg 2§ Cu 2§ Co2§ Ag+, Zn2§ and Fe 2§ for its enzyme activity and was inhibited by 1 mM EDTA and 1 mM phenylmethylsulfonyl fluoride. According to the mode of action of Amyl I on starch, Amyl I was classified as an exo- and a-amylase. Amyl I produced maltotetraose predominantly from starch via intermediates such as maltohexaose and maltopentaose. The gene encoding the maltotetraose-forming alkaline amylase (Amyl I) was cloned from Bacillus sp. strain GM8901 into Escherichia coli JM83. The nucleotide sequence of the cloned 2.8 kb DNA revealed that it contains one open reading frame of 2,554 nucleotides without a translational stop codon. The deduced amino acid sequence for these 2,554 nucleotides is 848 amino acids including a signal peptide of 29 residues at its NH2-terminal end. The deduced amino acid sequence of Amyl I contains four conserved regions that constitute the active center of a-amylases. However, The deduced amino acid sequence of the mature Amyl I showed relatively low

62 homology, 20-40%, with those of the saccharifying type a-amylases and the liquefying type a-amylases. The conserved regions and homologous region of Amyl I with other known a-amylases were restricted in the NH2-terminal part (about 450 amino acids) of Amyl I and the remaining COOH-terminal part (about 400 amino acids) showed no homology with other known amylases. When the amino acid sequence of Amyl I was compared with those of exo-a-amylases and alkaline amylases, homologous regions which seemed to be involved in the alkalophilicity or the exo-cleavage mode of Amyl I were not found.

Introduction Alkaline amylases that have optimum pH values higher than 8.0 have potential applications for hydrolyzing starch under high pH conditions in the starch and textile industries and as an ingredient in detergents for automatic dish washers and laundries (1, 2, 3). They can also be used as model proteins for understanding the molecular basis of the alkalophilicity of the enzymes, which may be of value in protein engineering. Since Horikoshi (4) first reported an alkaline amylase of an alkalophilic Bacillus strain, A-40-2, alkaline amylases have been identified in Bacillus sp. strain NRRL B-3881 (5), Bacillus sp. strain H-167 (6), Bacillus licheniformis TCRDC-B13 (7), Bacillus alcalothermophilus A3-8 (3), and Streptomyces sp. strain KSM-9 (2). However, there is only a limited amount of information concerning the molecular basis of reaction mechanism of these alkaline amylases. Previously, we isolated an alkalophilic Bacillus strain, GM8901, optimally grown at pH 10.5 and 50~ (8). Alkaline a-amylase produced from the bacterium has unique properties in that it produces predominantly maltotetraose from starch at an extremely high pH of 11.0 to 12.0 by exo-cleavage mechanism. We thought that the alkaline a-amylase of Bacillus sp. strain GM8901 might be a good model enzyme for elucidating a molecular basis of alkalophilicity of the alkahne amylases. In addition to this, our maltotetraose-forming alkaline aamylase might also be a good model enzyme for elucidating a molecular basis of exo-cleavage mechanism of the exo-type a-amylases. The exo-type a-amylases are of value for the production of maltoohgosaccharides that have potential uses in the food, phamaceutical, and fine-chemical industries because of its low sweetness, superior moisture retention, high viscosity and freezing point, and other special properties, compared with conventional sugar syrups (9). To understand the molecular basis of alkalophilicity and exo-cleavage mechanism of the alkahne a-amylase of Bacillus sp. strain GM8901, enzymatic and molecular biological studies have been conducted (10-12). In this article, we

53 describe the purification and characterization of the enzyme, and cloning and nucleotide sequence analysis of the gene coding for the alkaline a-amylase.

Results

Isolation of Bacillus sp. GM8901 Strain GM8901 was isolated as an alkaline amylase producer from soil on the shore of a river in Jangsung, Korea. The taxonomic study of the strain was done in our previous report (8). This strata belongs to the genus Bacillus and has a close resemblance to B. licheniformis. We designated this bacterium tentatively Bacillus sp. strain GM8901. Amylase production and proteolytic degradation of the enzyme Alkaline amylase from Bacillus sp. strain GM8901 is an extracellular enzyme induced by soluble starch and produced maximally at pH 10.5 and 50~ During the cultivation of the bacterium in 3 liters of medium containing 1% soluble starch as a sole carbon source, the changes in cell growth, alkaline amylase activity, and pH were examined (Fig. la). The amylase activity increased with the increase of cell growth and reached a maximum level at 24 h (0.75 U/ml ); thereafter, cell concentration and amylase activity decreased gradually. To gain information on the change of extracellular alkaline amylase pattern during cultivation, culture supernatants taken at various time intervals were concentrated fivefold with 80% saturation of ammonium sulfate and separated on a native 12% polyacrylamide gel; alkaline amylases on the gel were then visualized by an activity staining method (Fig. lb). From the sample taken at 8 h of cultivation, one major alkaline amylase was observed. However, thereafter, the number of alkaline amylases increased to five with the increase of culture time. We designated these alkaline amylases Amyl I, Amyl II, Amyl Ill, Amyl IV, and Amyl V. As shown in Fig. lb, Amyl I and Amyl II of culture supernatant taken at 8 or 12 h appeared at much higher position on the native gel than those of the other culture supernatant. The delayed migration of Amyl I and Amyl II was due to their high affinity to soluble starch present in the concentrate of culture supernatant taken at 8 or 12 h. The activity of Amyl I with the highest molecular weight showed a peak at 24 h of cultivation by native polyacrylamide gel electrophoresis. However, after that time, the activity of Amyl I decreased gradually, and while the activities of the other alkaline amylases, Amyl II, Amyl HI, Amyl IV, and Amyl V, increased with time. When Amyl I, separated by electroelution from the native gel without the contamination of the other alkaline amylases, was incubated with a small amount of 24 h culture supernatant, it was converted into Amyl II, Amyl HI, Amyl IV, and Amyl V. This result suggested that Amyl II, Amyl HI, Amyl IV, and Amyl V are produced from Amyl I, presumably by proteolytic degradation. We found an alkaline protease activity

54

Figure 1. Time courses of cell growth, amylase production, and pH change in culture medium. Bacillus sp. strain GM8901 was cultivated in a jar fermentor with a working volume of 3 liters under the initial conditions of pH 10.5, 50 ~ 200 rpm, and 1 vvm. (a) Changes in cell growth Co), amylase production (~), and pH ([3). (b) change of the isozyme pattern of alkaline amylase by culture time. Fifty milliliters of culture supernatant was concentrated fivefold by ammonium sulfate precipitation (80% saturation) and dialysis, and the concentrate was analyzed by native 12% PAGE and activity staining. OD, optical density. in the culture supernatant. The alkaline protease was partially purified from the culture supernatant by ammonium sulfate precipitation (80% saturation), phenyl-Toyopearl column chromatography, and Mono Q HR 5/5 HPLC (Fig. 2a). The partially purified protease was completely ~nhibited by PMSF, suggesting a subtilisin-like serine protease. As shown in Fig. 2b, this enzyme degraded Amyl I into Amyl II, Amyl Ill, Amyl IV, and Amyl V. This result clearly showed that Amyl II, Amyl Ill, Amyl IV, and Amyl V are produced from proteolytic degradation of Amyl I by an alkaline serine-protease present in the culture supernatant.

Purification of alkaline amylase Am yl I After 8 h of cultivation of Bacillus sp. strain GM8901, crude enzyme was obtained from the culture supernatant by ammonium sulfate precipitation (80% saturation) and dialysis. To minimize the proteolytic degradation of Amyl I by the serine protease, all of the purification steps were carried out at 4~ PMSF, a serine protease inhibitor, could not be used for purification steps because it

65

Figure 2. Proteolytic degradation of an alkaline amylase (Amyl I) by an alkaline protease present in culture medium of Bacillus sp. strain GM 8901. (a) Alkaline protease was partially purified from culture supernatant by ammonium sulfate precipitation (80% saturation), phenyl-Toyopearl column chromatography, and Mono Q HR 5/5 HPLC. The Mono Q HR5/5 HPLC is shown. The alkaline protease peak is marked with shadow. (b) Degradation of Amyl I into Amyl ]], Amyl HI, Amyl IV, and Amyl V by the partially purified alkaline protease. The purified Amyl I was incubated with the partially purified a lkahne protease, and the reaction samples taken at 0 h (lane 0), lh (lane 1), 2 h (lane 2), and 3 h (lane 3) were analyzed by native 12% PAGE and activity staining. Lane C is the enzyme sample obtained from a 24 h culture supernatant described in the legend to Figure lb. OD, optical density. irreversibly inactivated Amyl I. Amyl I was purified serially by heparinSepharose CL-6B column chromatography, phenyl-Toyopearl column chromatography and Mono Q HR 5/5 HPLC. The purification steps are summarized in Table 1. The final enzyme preparation had a specific activity of 157.5 U/mg of protein and gave a single protein band by SDS-PAGE with a molecular mass of about 97,000 (Fig. 3). The native polyacrylamide gel stained with the KI-I2 solution showed a single alkaline a_m_ylase band which coincided with that of Amyl I. Through the purification steps described above, Amyl I was purified to homogeneity without detectable contamination of the other alkaline amylases produced from proteolytic degradation of Amyl I.

Effect of p H on the enzyme activity The activity of Amyl I was assayed in buffers of various pH values, and the relative activities are shown m Fig. 4a. The optimum pH of Amyl I was 11.0 to 12.0. Amyl I shows high enzyme activity (above 90%) at an alkaline pH range of 10.5 to 12.0. However, at pH values below 9.0, the enzyme activity dropped

66 Table 1 Summary of the purification steps of alkaline Amyl I from the culture supernatant of Bacillus sp. strain GM 8901 Purification step

Total activity (U)

Culture supernatant 870 Ammonium sulfate precipitation 473 (80% saturation) and dialysis Heparin-Sepharose CL-6B 180 chromatography Phenyl-Toyopearl chromatography 139 Mono Q HR 5/5 HPLC 57

Total protein (mg)

Specific activity (U/mg)

675 359

1.29 1.32

8.4

21.4

1.9 0.06

74.5 157.5

Figure 3. Native PAGE (a) and SDS-PAGE (b) of the purified alkaline amylase (Amyl I). Amyl I was visualized by Coomassie brilliant blue staining (lanes C) and activity staining (lane A). Lane M, molecular mass markers (in kilodaltons). Arrows indicate the alkaline amylase. sharply to below 50% of maximal activity. To examine the pH stability of Amyl I, the enzyme was incubated in buffers with various pH vaues for 1 h at 50 ~ and the residual enzyme activity was assayed. As shown in Fig. 4b, Amyl I retained more than 85% of its initial enzyme activity at the pH range of 6.0 to 13.0. However, the pH stability of Amyl I was very low at pH values below 5.0 (80% and nearly 100% inactivations at pH 5.0 and 3.0, respectively). These results

57

indicated that Amyl I is an extremely alkalophilic enzyme that has a high optimal pH of 12.0 and that shows high stability in the pH range of 6.0 to 13.0 To the best of our knowledge, such an alkaline amylase has not been reported to date. Amyl I can be used as a good model protein for investigating the molecular basis of alkalophilicity of alkaline enzymes.

120

120

100

100 tip .,-I :>

.,4

.tJ

,rU

/

80

/

60

J J

r

>

.,4

40 ,-I

20

2

4

pH

pH

a

b

Fiqure 4. Effect of pH on purified Amyl I activity (a) and stability (b). The buffers used were 50raM citrate-sodium citrate buffer (pH 3.0 to 7.0) ( 9 50mM TrisHC1 buffer (pH 7.0 to 9.0) ( ~ ) , 50mM glycine-NaOH buffer (pH 9.0 to 11.0) ( E] ), 50raM sodium phosphate-NaOH buffer (pH 11.0 to 12.0) ( v ) , and 50mM hydroxide-chloride buffer (pH 12.0 to 13.0) ( x ).

Effect of temperature on the enzyme activity The enzyme activity of Amyl I, buffered at pH 10.5, was measured at various temperatures. As shown in Fig. 5a, Amyl I has an optimum temperature of 60 ~ To examine the thermostability of Amyl I, the enzyme buffered at pH 10.5 was incubated at various temperatures and samples were taken at appropriate time intervals to assay the residual activity (Fig. 5b). After 8 h of incubation at 40 ~ and 50 ~ Amyl I retained nearly 100 and 85% of its initial activity, respectively. However, only 37 and 12% remained after 2 h incubation at 60 ~ and 70 ~ respectively. By the addition of lmM CaC12, the thermostability of Amyl I at 60 ~ was distinctively enhanced: after 2 h incubation of the enzyme with and without lmM CaC12, 78 and 37% of the initial activity remained, respectively. However, the addition of lmM CaC12 could not stabilize the enzyme activity above 70 ~ The thermostability of Amyl I also increased in the presence of 1%

68

soluble starch: after 30-min incubations of the enzyme with and without 1% soluble starch at 60 ~ 54 and 22% of the initial activity remained, respectively. 120

120 100

,,,,,,,,

10Q O ~

dO

.~

0

.

.

.

0

.

[]

80

~ .~,

80

.,4

[] ~

E]

I1) ~

.,4 ~ r-t

40

,-4

[] 40

20

20

oo Temperature

a

(~

~

~

4

6

8

,o

T i m e (hr)

b

Figure 5. Effect of temperature on purified Amyl I activity (a) and stability (b). For determining the thermostability, Amyl I was incubated at each temperature (i.e., 40oC [ 9 50oC [ ~ ], 60oC [[~], and 70~ [ v]) in the presence (dotted lines) or absence (solid lines) of lmM Ca2§ after that, the remaining activity was measured.

Effect of metal ions and chemical reagents The activity of Amyl I buffered with 10raM glycine-NaOH (pH 10.5) was assayed in the presence of various metal ions and chemical reagents. Amyl I is activated in the presence of lmM Ca 2§ Mg 2§ Cu 2§ Co2§ or Ag + by 120 to 154%. The addition of lmM Zn2§ or Fe 2§ does not significantly affect the enzyme activity. However, lmM Hg 2§ completely inhibited the enzyme activity. When Amyl I was preincubated with lmM EDTA, the enzyme activity dropped to 18% of its initial activity (82% inhibition). The enzyme recovered its original activity by the addition of 2raM Ca 2§ Mg 2+, Cu 2+, Co2+, Zn2§ Ag +, or Fe 2§ after the preincubation with lmM EDTA. When the activity of Amyl I buffered at pH 10.5 was assayed in the presence of lmM PMSF, significant inhibition was not observed. However, with the increase of incubation time of Amyl I with l m M PMSF at 4 ~ the enzyme inhibition was gradually increased: after 10 h of incubation, the residual enzyme activity of PMSF-treated sample was 24% of the nontreated enzyme activity. The addition of 0.1% detergent, including SDS,

59 CHAPS {3-[(3-cholamidopropyl)-dimethyl-ammonio]-l-propanesulfonate}, Triton X-100, did not significantly affect the enzyme activity.

and

Mode of action of Amyl I on starch hydrolysis Table 2 shows the relative activity of Amyl I on various substrates. Amyl I most easily hydrolyzed long-chain polysaccharides such as amylose and amylopectin and also easily hydrolyzed maltooligosaccharides, including maltopentaose (GS), maltohexaose (G6), and maltoheptaose(G7). Short-chain maltooligosaccharides such as maltose (G2), maltotriose (G3), and maltotetraose (G4) were resistant to Amyl I. Amyl I could not hydrolyze pullulan, an a-l,6polysaccharide of G3, or a-, ~-, and 7-cyclodextrins at all, even after overnight reaction. When Amyl I was incubated with 5% G5 or 5% G6 and the reaction products of samples taken at appropriate time intervals were analyzed by TLC, we could not detect any products formed via transglycosylation. These results indicated that Amyl I hydrolyzes the a-l,4-glycosidic linkage of starch or preferably polysaccharides longer than G4, without glycosyl transferase activity or hydrolytic activity of the a-1,6-glycosidic linkage of starch. Table 2 Relative activities of purified Amyl I on various substrates Substrate

Relative activity (%)

Soluble starch ...................................................... 100 Amylose ................................................................. 140 Amylopectin ........................................................... ND PuUulan .................................................................. ND Cyclodextrins ( a, ~, 7) ........................................ ND Maltose ................................................................... ND Maltotriose ............................................................ ND Maltotetraose ....................................................... ND Maltopentaose ...................................................... 35 Maltohexaose ........................................................ 56 Maltoheptaose ...................................................... 75

To examine the mode of action of Amyl I on starch hydrolysis, amylose and maltooligosaccharides, including G2 to G7, were treated with Amyl I, and the reaction products of samples taken at different time intervals were analyzed by TLC. As shown in Fig. 6, at an early stage of amylose hydrolysis, G6 was produced predominantly with the concomitant productions of a small amount of G4 and a trace amount of G5. After 2 h of reaction, a large amount of G4 and

70 some G2 accumulated with the disappearance of G6 and G5. When 1% G2 or 1% G3 was treated with amyl I at 50 ~ for 1 h, it did not produce any hydrolytic products on TLC. G4 was also resistant to Amyl I but produced a trace amount of degradation products after 1 h of reaction. Maltooligosaccharides such as G5, G6, and G7 were rapidly hydrolyzed into G4 by Amyl I via the intermediate of G5 (minor intermediate) or G6 (major intermediate) (Fig. 7). When we quantitatively analyzed reaction products of Amyl I from 5% soluble starch with HPLC, nearly the same results as those of TLC analysis were obtained. As shown in Fig. 8, at an early phase of hydrolysis (1 h of reaction), 1% (wt/wt) G 1, 12% G2, 1% G3, 26% G4, 6% G5, and 54% G6 maltooligosaccharides were produced. As starch hydrolysis progressed, G4 and G2 increased gradually and while G6 and G5, inversely, decreased. After 20 h of hydrolysis, maltooligosaccharides compositions were 9% G1, 27.6% G2, 8.3% G3, 53.2% G4, 1.3% G5, and 0.6% G6. These results indicated that Amyl I produced G4 as a major end product from starch through the intermediates of G6 (major intermediate) and G5 (minor intermediate).

Figure 6. Thin layer chromatogram of the reaction products of the purified Amyl I on amylos. Samples taken at 2 min (lane 1), 10 min (lane 2), 20 rain (lane 3), 30 min (lane 4), 60 min (lane 5), 2 h (lane 6), and 15 h (lane 7) were analyzed by TLC. Lane S represents the standard sugars: glucose (G 1), maltose (G 2), maltotriose (G 3), maltotetraose (G 4), maltopentaose (G 5), maltohexaose (G 6), and maltoheptaose (G 7).

71

Figure 7. TLC of the reaction products of purified Amyl I on maltotetraose (G4), maltopentaose (G5), maltohexaose (G6), and maltoheptaose (G7). Samples taken at 5 min (lane 1), 10 min (lane 2), 20 rain (lane 3), 30 rain (lane 4), 40 min (lane 5), and 60 mm (lane 6) were analyzed by TLC. Standard sugars (lane S) are as described in the legend to Figure 6. TLC of reaction products of Amyl I on the various maltooligosaccharides described above (Fig. 6 and Fig. 7) suggested that Amyl I hydrolyzes starch by an exo-cleavage mode rather than an endo-cleavage mode. We examined various methods to determine whether Amyl I is an exo- or endo-type enzyme. When insoluble blue starch (starch azure; Sigma chemical Co.), a substrate for colorimetric determination of endo-type a-amylases, was used as a substrate, Amyl I could not hydrolyze the blue starch at all. In addition to this, when Amyl I was reacted with 0.1% soluble starch and the blue values of the reaction mixture taken at appropriate time intervals were measures at 620 nm after blue color development with the final solution 0.5% KI-0.05% I2-0.1 N HC1, the blue color of the reaction mixture diminished slowly as hydrolysis proceeded: only 8% of the blue vale was reduced after 17% hydrolysis of the soluble starch. These results indicated that Amyl I follows an exo-cleavage mechanism. The exo-cleavage mechanism of Amyl I is also supported by the fact that Amyl I could not hydrolyze a-, ~-, and T-cyclodextrins. It has been reported that exo-amylases such as glucoamylase, ~-amylase, and maltohexaose-forming amylase from Klebsiella pneuminiae could not hydrolyze the cyclodextrins but endo-amylases such as Taka-amylase and bacterial saccharifying amylase could (13).

72

A

8

v

// ~,

o o

J /

s G2

,

2 G1

G3

(I

;

|

,o

,5

20

G5 G6

25

T i m e (hr)

Figure 8. Time course of maltooligosaccharides production from 5% soluble starch by purified Amyl I. Samples taken at appropriate time intervals were analyzed by HPLC. Symbols (G1 to G6) are as described in the legend to Figure 6. The anomeric form of the products of Amyl I was examined with a polarimeter. Optical rotation decreased sharply after the reaction stopped. This showed that the product has an a-configuration. Hence, our Amyl I was classified as an exo-aamylase. Cloning of Amyl I gene A genomic library of Bacillus sp. strain GM8901 DNA was prepared in E. coli JM83 by using pUC19. From about 30,000 transform ants, one positive clone showing ~lk_aline amylase activity was obtained by a top-agar overlay method (Fig. 9). The recombinant plasmid from the positive clone was designated pALA10 and the restriction map of pAI~10 is shown in Fig.10. Southern hybridization revealed that the labeled 1.2 kb EcoRI fragment from pALA10 hybridized only with a 1.2 kb fragment of chromosomal DNA of Bacillus sp. GM8901 digested with EcoRL This result indicated that only one copy of the insert DNA of pAI~10 is present in the chromosome of Bacillus sp. GM8901 and that no apparent rearrangement of the DNA occurred during gene cloning. The recombinant E.coli JM83 carrying pAIA10 accumulated alkaline amylase in intracellular fraction of cells without secretion of the enzyme into culture broth. Maximum 0.38U/ml of alkaline amylase was produced from the recombinant E. coli strain in LB broth, which is comparable to the amount of alkaline amylase

73

Figure 9. Starch-hydrolysis of E. coli JM83 carrying pALA10 on agar plates. E. coli JM83 carrying pAI~10 was tooth-picked on LB agar plate containing 50 ~g/ml ampicillin. After the incubation of the agar plates for 24 hr at 37 ~ top agar solution containing 1% soluble starch dissolved in 50mM N a-phosphate buffer (pH 7.5) or 1% soluble starch dissolved in 50raM glycine-NaOH buffer (pH 10.5) was overlayed onto the agar plates. After htrther 20 hr incubation of the agar plates at 37 ~ the starch-hydrolysis was visualized by spraying 2% KI0.2% I9 solution. produced from Bacillus sp. strain GM8901. The alkaline amylase was partially purified from a cell extract of the recombinant E. coli strain by a phenylToyopearl column chromatography. Some properties of the partially purified enzyme from the recombinant E. coli strain were compared with those of Amyl I purified from the culture supernatant of Bacillus sp. strain GM8901. As shown in Fig. 11, the alkaline amylase from the recombinant E. coli strain produced initially G6 from soluble starch and then, accumulated G4 as a major end product with the increase of reaction time. This result was coincided with the starch-hydrolytic pattern of Amyl I (Fig. 6). The alkaline amylase from the recombinant E. coli strain showed the same pH-activity profile and temperatureactiviW profile to those of Amyl I. These results indicated that the alkaline amylase encoded in p ~ 1 0 is Amyl I of Bacillus sp. strain GM8901. Nucleotide sequence analysis of Amyl I gene According to the strategy indicated in Fig. 10, nucleotide sequence of the insert DNA (2,776 bp) of pALA10 was determined by the dideoxy chain termination method described by Sanger et al. (14) (Fig. 12).

74

Figure 11. Thin layer chromatogram of reaction products of the alkaline amylases produced from Bacillus sp. GM 8901 (lane B), E. coli JM83 (pUC 19) (lane E), and E. coli JM 83 (pALA10) (lane ET). 1% soluble starch in 50raM glycine-NaOH buffer (pH 10.5) was hydrolyzed with alkaline amylases for a given time interval at 50 ~

75 The nucleotide sequence contained translation start codon ATG at position 232 to 234. This ATG codon presumably is the correct initiation codon for translation because it is the only ATG which is preceded by a putative ribosome-binding site, GGAGG on position 223 to 227. In addition to this, there is a putative promoter site at the upstream of the ATG codon: -35 region, GTGAAG at position 164 to 169 and -10 region, TACAAAT at position 185 to 191. Because the direction of transcription of Amyl I gene in pALA10 is opposite to that of/ac promoter of pUC19 vector, this putative promoter site was believed to be used for transcription of the Amyl I gene by RNA polymerase of E. coli. Starting from the translation start codon ATG, one open reading frame of 2,554 nucleotides was found. However, we found no translational stop codon in the sequence of the insert DNA of pAIA10. The open reading frame in pALA10 codes a protein of 848 amino acid residues with a predicted molecular mass of 95,354 Da, which is sligthly lower than 97,000 Da of Amyl I purified from Bacillus sp. strain GM8901. The Amyl I gene appeared to be much longer than 2,554 bp. The deduced amino acid sequence indicated that the first 29 amino acids, from the initiator Met to Ser, might constitute a signal peptide invloved in the secretion of the exported proteins. The deduced amino acid sequence of the putative signal peptide had typical structural elements as a bacterial signal peptide having NH2-terminal cationic residues followed by hydrophobic core residues, and COOH-terminal polar residues (15). When the amino acid sequence of Amyl I was compared with those of other known a-amylases, four regions (I to IV) which are conserved in a-amylases and considered to constitute active centers of a-amylases, were found at position 174 to 438 amino acid residue (Fig. 13). This result indicated that Amyl I has the active center, common to other known a-amylases. However, the amino acid sequence of Amyl I showed relatively low homology with those of other known a amylases of Bacillus origin. The sequence of Amyl I was 27, 31, and 38% homologous to those of the liquefying type a-amylases of B. stearothermophilus (16), B. amyloliquefaciens (17), and B. licheniformis (18), respectively. These liquefying type a-amylases have been known to have more than 65% homology in their amino acid sequences. The sequence of Amyl I showed 39% homology with that of the saccharifying type a-amylase of B. subtilis (19). The sequence of the saccharifying type a-amylase has been known to have less than 30% homology with those of the liquefying type a-amylases. These results indicated that Amyl I belongs to a new group of a-amylase having different amino acid sequence from those of the saccharifying type a-amylase of B. subtilis and the liquefying type aamylases of Bacillus species. It seemed that Amyl I might be derived through an different evolutionary route from those of the saccharifying type and liquefying type a-amylases, starting from a common ancestor of a-amylases. The conserved

76 138

I C ~ % T L ~ C T 2 T A A A A T A A A G A A A A A A ~ G G A ~ ~ T A A A G ~ ~ T C A ~ ~ T r A C G A 2 - 1 T A T T ~ A A C ~ ~

{

G L ~ I ~ / ~ T C ' I ' s

-35

276

-10

S D

.

N A s Y

K

TCA'I"TrGTC~T~~ACATTCTCTATCAAT~~(r._AA~TGGGATA~W3CC,

v

~

. . . .

G

L

414

A

T

V

F

A

I

R

F

H

D

V

~ K

T ~ L N E

A G

~ A ~ Y F T E

L

G

I

A A ~ T ~ T G G A T C A ~ A ~ ~ ~ T r A C G C T T A C C A T G G T T A C T A 2 X ] C A T T A G ~ C r A C A ~ T C ~ T C A A A A T A ~ C G N A I W I T A P Y E Q v H G W V G G G S D G D F A H Y A Y H G Y Y A L D Y T M I D Q N

M

G

T

Gr~GATATC~TI~TACA(~2k3~GCAAGG(I~ V E D M R E F V H T A H

G

V

S

C

G

N

C C ~ T A ~ C T G G A ~ A ~ D K V E Y E D W I V P

A

V

I

Q

Q

S

G G ~ T C C A A C A A C ~ G T C A T A ~ C ~ ~ G ~ C G ~ T G G ~ C A T T G G C , ~ C ~ G N P S N N Q S Y G R P Q E D A W G Q N I G T F H G

G

D

GAACAT~ E

H

W

T

T

L

P

P

G

L

Q

G

F

S

P

T

F

StI

E

Q

N

G

S

I

V

R

D

V

A

V

S

L

O

D

G

V

G

E

N

W

H

D

V

H

Q

Q

I

N

Y

E

D

T

C

L

S

W

D

Q

~ G

L

TGAATCA~ATAACACC, M N H P G Y N

V

C A A A A T ~ T ~ T C A A C A G A q ' r A A C T A ~ C ~ Q

F

A

G

A

W

R

H

T

W

~

W

G

T

P

A G T ~ 2 ~ r A ~ T ~ A G ~ A L - ~ A ( ~ C T C A C A C A T ~ A G ~ A T T A G A A ~ ~ S E I T M C L A G L P D F R T E L T H S V G L P P L L E R 5 G GC~TGAC~TTTAGG~GCTGATTA~TCGTTAAA A N D L R R D L G V A P A

D

Y

I

V

K

W

L

C

A

W

V

Q

E

F

L

K

D

A

A

N

E

A

L

W

K

W

~TC-ATGGATTTGATT~~TrI~TT~GCA D H G F D S V I N F T F

R

Q

E

G

E

E

CTAAGCTAC2~rATL~CATGATAC~ATATC~CTEA L S Y L S Q H D T Q L Y P

N

H

G

R

{y

P

E

E

A

P

G

A

Y

W

T

D

~

A

W

I

T

I

K

N

R

I

X

A

G

(

~

F

R

D

~

~

T

Y

E

E

T

G

C

T

Q

S

~

C

~

G

2

N

A

A

V

A

Q

L

~

C

T

G

E

T

3

~

V

A

V

L

0

A

V

(

T

R

"

L

L

T

Y

T

]

G

E

Y

A

C

~

~

N

V

_

~

E

T

G

Q

A

A

N

G

H

~

V

T

A

H

E

A

T C

A

T

A

F C

E

G

A

P _

W

E

C

A

A A A G T ~ C A C ~ T O T ~ ~ I ~ I V , K

V

T

L

H

T

R

N

V

E

G

G

T

Y

A

S

G

~

K

I

H ~

G

A

2

R

L

R

N

Y

~

E

F

N

V

A

P

E

P

F

A

A

G

T

C

T

S

K

.

G

V

~

D

A

R

P

A

P

A

C

A

K

A

G

L

Y

D

~

E

G

P

O

A

E

Q

'

G

E

Q

A

F

T

T

G

L

T

I

Y

F

D

A

A

A

T ~ ( r _ . ~ ' r rAff..~A~GGGGAT'IVb2'ATAC(F._AT A f f . . , A T T u k ~ ( ~ Ac~L.~C

~

W

S

E

A

P

E

M

V

T

H

T

I

E

G

v

E

K

D

T

Y

K

G

E

W

S

D

I

D

K

S

R

P

P ~

E

966

T

S

C

V

W

G

H

G

V

G

{{{

G E

A G

V

T

L

P

~

T

T

F

T

A

A

L

V

~

D

F

'

S

q

P

'

A

C

P

T

A

A

r

N

W

~

S

G

G

E

~

D

I

L

S

M

I

G

K

-

~

T

~

W

G

T

P

c r ATTA'I'r'rA~;C~' R

L

L

F

K

D

G

A

H

~

C P

r A

D

V

Y

~

F

~ F

A

1656

G S

~

S

L

A

T

~ D

C

V

P

F

G

P

T

1794 GGGGAACATCEACAAA~ G E H R Q I 1932

~ G Q

6 T A G G E F

I

1242

~

1518

; ~ _ ~ A G C _ ~ A O ~ R S N H L A V G

S

1104

13~o TATq~rT

qx3qx3C~AGTATTTr~G{~%GAT~OCTACA G V Q v F Y G D E T A R

G

L

T

G

T

~

D

E

G

E

Y

T

G

I

F

Y

T

~

I

~ L

E

H

Y

G

~ V I

A G

N

E

T

F

~ T ~ T C ~ V V V V G A S

E

C~,(~

E

I

T

G V

G

G Q

T

L

E

A~TGAGV_A~A~T~GGGA~TTATGATT~TAC~~GACT;T_AC2L-V_AGCrT

D

AA~G~d~.,AA~CAATAGAAGAAACATACACATACA~TCAA~TTTA2TFr K

~

~

W

R

L

G

A G T A G T T I D V S S V

~

~

Y

ACC~AGAT~TGGAAA~AL~~TGCAATCAATA~TT~TTCAATGTG L D T M E S T F S R Y A D A I N T D S

M

690

II V

~ T

G R

A

F

~ / ~ A T T T A A I D G G T Y L

~ D

G

Y

L

T ~ C A T A T T Y V E G D V D

~

E

R

M

GL~I~CAAG'r~ T A A S P Y

G R

A

S

P

W

F

~ S

L

R

G

F

Q

L F

I

552

82/} N

N

N

D

D

A

~ C C C T C - A ~ ~ ~ T G A A ~ T r C r G T C A A ~ A ~ ~ T G O G G C A A T T r G S D P H Q G T R S S M N W D S V N E E V L H H W Q K M G A T

3

G A A T T C G G r A T r C ~ T G G A ~ T A L ~ ~ I D G F R V D T A K H V E L S R

G

C A A T T A A A G G A T ~ T G A A G C T ~ T C ~ G G A G A A C ~ G L ' ~ ~ T C ~ 2 ~ Q

~

T

T

ATTAAAGATATGCATf~GTAC~~ATCA I K D M H E Y G F G

T ~

~

Y

C

".

T

F

~

L

R

S

AACGGTATA~Tr

Y

~

Q

~ D

G

A

v

~

R

~

R

2070

T

E

~

D

L

~

2208 N

I

~46

~ [ ' ~ G D T L

2484 I

~

L

E

A

A

~

Y

Y

l

~

'

~

R

G

T

N

G

T

S

~

E

~

T

~

T

T

L

Q

L

V

D

2622

Y

D

T

N

P

K

Q

w

P

G

P

G

27 60 C$,GCCT E

P

~rTI'I~CCC~GA~ G F F R D

Figure 12. Nucleotide sequence of the alkaline amylase gene of Bacillus sp. strain GM8901. The cleavage site between the putative signal peptide and extracellular mature Amyl I is indicated by the upward arrow. The lines marked with -35 and 10, and the line marked SD indicate putative promoter site and ribosomebinding site, respectively. The four conseved regions of a-amylases (I to IV) are underlined.

77

regions (I to IV) and homologous region of Amyl I with other known a-amylases were highly restricted in the NH2-terminal part of 450 amino acid residues of Amyl I. The remaining COOH-terminal part of 400 amino acid residues of Amyl I showed no homology with those of other known a-amylases, glucoamylases, and ~-amylases. We thought that this unique structure of the COOH-terminal part of Amyl I may be closely related with the unique properties of Amyl I such as extremely high optimum pH or exo-cleavage mechanism of the enzyme.

BamA~IY BpoAMY BstAMY BsuAMY I 16AMY B7OAMY -~- B89AMY Pst~NY

I

98 821 101 D 97 160 135 174 133

~H

~ i

AorAMY HsaAMY

177 99

BmaCGT

135

BstCGT Pam!AM BstNPL KaePUL

1311i 291 242 600 _

NH NH

A

NH NH NH NH NH NH NH I NH NH NH NH NH

N___~H

LD

II F-R

L

I A

~FR F

L

II

I

I

I V

I~~

~FR FR FR

TA

LR

T

IR IR

I

GFR I D

vI ~FRWR V I

V

I GFR

F

S

V V LAS V V ANE L MG Y

I II 261 ~Y 931 E l 264 EY 208 E l EV EF 371 EV EL 221 EV 230 EV EV

WQ FH WS L Q FG WK WG WK I T L D I D

323 992 326 264 415 361 432 308 288 292 298

F V EE_E~N D T F Fv ~T Wv ~ T Y ~ T Fv Y F FA I T T]H Oll F VEIN H DIN F VID N HDIN

258 253 454 357 704

F L F L SV WH WD

324 319 5O2 419 827

F liD N H DIM F lID N H DI M F I []~_F~]HDIG L L G SIH DiT Y V S KL~_~] N

299

236

EW EW EW El EG

o

IV

IDN~~ V

i

S

Ao

Figure 13. Conseved regions I, II, HI and IV of a-amylase type starch-degrading enzymes. Amino acid residues which correspond to the consensus sequence are boxed. Enzymes are abbreviated as: AMY, a-amylase; CGT, cyclodextrm glucanotransferase; IAM, isoamylase; NPL, neopullulanase; PUL, pullulanase. Enzyme sources are abbreviated as: Barn, Bacillus amyloliquefaciens; Bpo, B. polymyxa; Bst, B. stearothermophilus; Bsu, B. subtilis; I16, isolate 163-26; B70, Bacillus sp. #707; B89, Bacillus sp. GM8901 (indicated with arrow); Pst, Pseudomonas stutzeri; Ahy, Aeromonas hydrophilia; Aor, Aspergillus oryzae; Hsa, human saliva; Bma, B. macerans; Pare, P. amyloderamosa; Kae, Klebsiella aerogenes. $ , Catalytic residues and A, substrate binding residues. When the amino acid sequence of Amyl I was compared with those of typical exo-cleavage type amylases such as glucoamylase (20) and ~-amylases (21, 22), no significant homology was found. This result indicated that the evolutionary origin and catalysis mechanism of Amyl I is more closer to those of a-amylases than to those of gluco- or ~-amylases.

78 As mentioned previously, our Amyl I is exo-cleavage type a-amylase. We expected that the primary structure of Amyl I will have homologous regions with those of other known exo-type a-amylases such as G6-forming alkaline aamylase of an alkalophilic Bacillus sp. #707 (23), G4-forming a-amylase of Pseudomonas stutzeri(24), and G5-forming a-amylase of an alkalophilic bacterial strain 163-26 (25). However, no homologous regions were found in the primary structures of these exo-cleavage type a-amylases. When the alkaline amylases of G4-forming Amyl I of Bacillus sp. GM8901, G5-forming a-amylase of strain 16326, and G6-forming a-amylase of Bacillus sp. #707 were compared in their primary structures, no homologous regions were found. These results suggested that the regions involved in the exo-cleavage mode of exo-type a-amylases or in the alkalophilicity of the alkaline amylases are given by the tertiary structures of the enzymes rather than by their primary structures.

Discussion

Most of the Bacillus amylases have optimum pH values of from 5.0 to 8.0 (26, 27, 28, 29, 30), and an acidic amylase from Bacillus acidocaldarius has an optimum pH 3.5 (31). Alkaline amylases reported previously had optimum pH values of 9.0 to 10.5 (2-5, 7). However, our alkaline amylase Amyl I has an extremely high optimal pH value of 11.0 to 12.0, which to our knowledge is higher than those of any other amylases reported to date. Amyl I is stable in an alkaline pH range of 7.0 to 13.0: more than 90% enzyme activity remained after 1 h of incubation in this pH range. These results indicated that Amyl I is well adjusted to catalyze starch hydrolysis under extremely alkaline conditions. It has been known that all enzymes of the a-amylase family are considered to catalyze the same basic reaction, a nucleophilic double-displacement mechanism with a transient covalent intermediate: two acidic amino acid residues (Asp or Glu) of protein are essentially involved in catalysis, i.e., one has an ionized carboxylic acid group and the other has an un-ionized carboxylic acid group (32). However, it is doubtful whether Amyl I can have an un-ionized carboxylic acid group in an active center even at pHs above 11.0 because the deionization of the carboxylic acid group of Asp and Glu residues of proteins is expected under this pH condition. With respect to this view, compared with other neutral and acidic amylases, Amyl I is believed to have a unique structure and reaction mechanism to catalyze starch hydrolysis under extremely high pH conditions. However, there are no reports describing the catalytic mechanism of alkaline amylases and the structure-function relationship of the enzymes. Further study of Amyl I will reveal the molecular basis of the alkalophilicity of the enzyme. The activity of Amyl I is greatly influenced by metal ions. Amyl I is activated by lmM Ca 2+, Mg ~-+,Cu e+, Co~-+. or Ag+ by 120 to 154%. The addition of lmM Zn 2+

79 or Fe 2§ does not significantly affect the enzyme activity. Amyl I is inhibited by a chelating agent, EDTA, and its full activity is recovered by adding Ca 2§ Mg 2§ Cu 2§ Co2§ Zn2§ Ag+, and Fe 2§ This result suggested that metal ions are essential for the enzyme activity of Amyl I and that Amyl I does not show a strict specificity for metal ions. Ca 2§ not only activates the activity of Amyl I but also enhances its thermostability. It has been reported that alkaline amylases of Bacillus sp. strain A-40-2 (4), Bacillus sp. strain NRRL B-3881 (5), and B. alcalothermophilus A3-8 (3) are stable in response to EDTA treatment, while the neutral ~mylases of B. amyloliquefaciens, Taka-~mylase A, and the liquefying amylase of B. subtilis are sensitive to EDTA treatment (30). Amyl I shows unique properties in the catalytic mode of action. Amyl I produces predominantly G4 from soluble starch via the intermediates of G6 (major intermediate) and G5 (minor intermediate). Previously, G4-forming amylases were discovered in P. stutzeri (33, 27). and Bacillus circulans MG-4 (34). However, these enzymes showed their optimal pH to be in neutral range. G4forming alkaline amylase from olkalophilic bacteria has not been reported to date. Amyl I is the first alkaline amylase producing G4 from starch as a major end product. On the basis of its mode of action, Amyl I can be classified as an exoand a-amylase. It has been known that most a-amylases are endo-type enzymes that randomly hydrolyze starch. However, Amyl I is a rare a-amylase that hydrolyzes starch by an exo-cleavage mechanism as shown in Fig. 14.

I;

Amylose

...... 9

o0

Amylopectin

Figure 14. A proposed reaction mode of the alkaline amylase I on starch. Thickness of arrows indicates relative activity on cleavage sites, e, reducing end. Deduced amino acid sequence of Amyl I from the nucleotide sequence showed that Amyl I has four conserved regions (I to IV) that constitute the active centers

80 of a-amylases. However, the sequence of Amyl I showed relatively low homology, about 20-40%, with those of other known saccharifying and liquefying type aamylases. This result suggested that Amyl I is derived through a different evolutionary route from those of other saccharifying and liquefying type aamylases, starting from a common ancestor of a-amylases. We suggest that Amyl I belongs to a new group of a-amylases which is different from the saccharifying and liquefying type a-amylases in their primary structures. The amino acid sequence of Amyl I showed the highest homology, about 66%, with that of G5-forming amylase of an alkalophflic bacterial strain 163-26 (25). This GS-forming amylase can be sorted into the same a-amylase group as Amyl I. These two amylases showed similar enzymatic properties: they have optimum pHs at alkaline condition and followed exo-cleavage mechanism. The amino acid sequence of Amyl I showed no significant homology with those of gluco- and ~amylases. This results suggested that the evolutionary origin and catalytic mechanism of Amyl I is more closer to those of a-amylase than to those of glucoand ~-amylases. The conserved regions and homologous part of Amyl I with other known a amylases were restricted only in the NH2- terminal part (about 450 amino acids) of Amyl I. The COOH- terminal part (about 400 amino acids) of Amyl I showed no homology with other known a-amylases and gluco- or ~-amylases. Though no evidences are present, the unique structure of COOH-terminal part of Amyl I may be related wth the unique properties of Amyl I, including an extremely high optimum pH or exo-cleavage mode of the enzyme on starch. Investigation on the role of COOH-terminal part of Amyl I will give an insight on the molecular basis of these enzymatic properties. Futher molecular biological study of Amyl I will contribute to an understanding of the molecular basis of alkalophilicity and exo-cleavage mechanism of the enzyme.

Acknowledgement This study was supported by a research grant from K O S E F for SRC(Research Center for N e w Bio-Materials in Agriculture, Seoul National University).

References Grant WD, Horikoshi K. 346-366. In Dacosta MS, Duarte JC, Wilfiams AD. eels. Microbiology of extreme environments and potential for biotechnology. Elsevier Science Publishers Ltd., Essex, 1989; 346-366.

81 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Nakai R, Sato T, Okamoto K. Japanese Kokai Koho patent 86,209,588. Ozaki A, Tanaka A. Japanese Kokai Koho patent 9,049,584. Horikoshi K. Agric Biol Chem. 1971; 35" 1783-1791. Boyer EW, Ingle MB. J Bacteriol 1972; 110:992-1000. Hayashi T, Akiba T, Horikoshi K. Agric Biol Chem 1988; 52" 443-448. Bajpai P, Bajpao P. Biotechnol Bioeng 1989; 33:72-78. Shin YC, Kim TU, Lee SY, Byun SM. Korean J Food Sci Technol 1991; 23" 349-357. Fogarty W M , Kelly CT. In: Fogarty W M , Kelly T, eds. Microbial enzymes and biothechnology, 2nd ed. Elsevier Science Publishers Ltd., Essex, 1976; 71132. Kim TU, Gu BG, Jeong JY, Byun SM, Shin YC. Appl Envron Microbiol 1995; 61: 3105-3112. Kim TU. Ph. D. thesis.Gyeongsang National University, Chinju: 1995. Jeong JY. M. S. thesis. Gyeongsang National University, Chinju: 1996. Saito N. Arch Biochem Biophys 1973; 155: 290-298. Sanger F, Nicklen S, Coulson AR. Proc Natl Acad Sci U S A 1977; 74: 54635467. Perman D, Halvorson HO. J Mol Biol 1983; 167:391-409. Sohma A, Fujita T, Yamane KJ. Gen Microbiol 1987; 133: 3271-3277. Takkinen K, Petterson RF, Kalkkinen N, Palva I, Soderlund H, Kaariainen L. J Biol C h e m 1983; 258: 1007-1013. Yauki T, Nomura T, Tezuka H, Tsuboi A, et al. J Biochem 1985; 98: 11471156. Yamazaki H, Ohmura K, Nakayama A, Takeichi Y, et al. J Bacteriol 1983; 156: 327-337. Itoh T, Ohtsuki I, Yamashita I, Fukui S. J Bacteriol 1987; 169: 4171-4176. Nanmori T, Nagai M, Shimizu Y, Shinke R, et al. Appl Environ Microbiol 1993; 59: 623-627. Kawazu T, Nakanishi Y, Uozumi N, Sakaki T, et al. J Bacteriol 1987; 169: 1564-1570. Tsukamoto A, Kimura K, Ishii Y, Takano T et al. Biochem Biophys Res C o m m u n 1988; 151: 25-31. FujitaM, Torigoe K, Nakada T, Tsusaki K, et al. J Bacteriol 1989; 171" 1333-1339. Canussio A, Schmid G, Bock A. Eur J Biochem 1990; 191: 177-185. Pfueller SL, ElliottWH. J Biol Chem 1969; 244: 48-54. Sakano Y, Kashiyama E, Kobayashi T. Agric Biol C h e m 1973; 47: 1761-1768. Shetty JK, Allen WG. Cereal Foods World 1988; 33: 929-934. Taniguchi H, Jae CM, Yoshigi N, Maruyama Y. Agric Biol C h e m 1983; 7: 511-519. Toda H, Narita N. J Biochem 1986; 63: 302-307.

82 31 Buonocore V, Caporale C, Rosa MD, Gambacorta A. J Bacteriol 1976; 128: 515-521. 32 Tao BY, Reilly PJ, Robyt JF. Biochem Biophys Acta 1989; 995:214-220. 33 Robyt JF, Ackerman RJ. Arch Biochem Biophys 1971; 145:105-114. 34 Hayashi S, Imada K. Agric Biol Chem 1991; 55:1715-1720.

Enzymesfor CarbohydrateEngineering K.H. Park, J.F. Robyt and Y-D. Choi (Editors) 1996 Elsevier Science B.V.

83

Structural studies on cellulases, pectinases and xylanases Peter W. Goodenough Protein Engineering Department, Institute of Food Research, Reading Laboratory, Earley Gate, Whiteknights Road, Reading, RG6 6BZ, U.K.

Abstract The importance of plant cell walls in agriculture, food technology and h u m a n health is so large that considerable efforts are being made to understand the synthesis and breakdown of the complex carbohydrates which form the cell wall structure. In particular the last three years has seen the discovery of many of the tertiary structure of proteins involved in cell wall degradation. Initially there was an extensive classification based on amino acid homologies but the tertiary structures of the proteins has revealed surprising data about the structural homologies that are found between cellulases and xylanases. Although the structural data has possibly reduced the usefulness of the earlier classification it has given much information on the evolutionary trends amongst these enzymes. Perhaps even more exciting has been the discovery that enzymes that modi~ pectin have a unique and unexpected structural motif.

84

Introduction

The Carbohydrates Hexose sugars in the pyranose configuration are commonly found in biological systems as polymers. In particular a sequence of homopolymers is synthesised during development of plant cell-walls. Initially a plant cell-wall consists of middle lamellae which are mainly constructed of polygalacturonate (Figure 1).

Figure 1. Schematic representation of a plant cell wall showing that the middle lamella (ml) is formed first followed by the primary cell wall (1) and a layered secondary cell wall (2). The plasmallema surrounds the cell wall (pm). No single plant has all of these walls exactly as shown.

Depending on the genera the polygalacturonate can be highly methylated at C 6 to give pectin and various other sugars can form branches from the main chain. Arabinose (a pentose sugar) occurs in the furanose form and galactose is also found as a branch on the chain. Over the last fifteen years the structure of pectins have been extensively investigated and as well as homogalacturonate there are well defined heteropolymers with rhamnose (Rhamnogalacturan I and II) xylose (xylogalacturan) galactan, apiose and fucose. A unique sugar found in rhamnogalacturan II is aceric acid. These heteropolymers occur in highly ramified regions of pectin which occur along a largely linear chain of homogalacturonate. There are a range of enzymes which are specific for these ramified areas and the build up or break down of these areas will be a major area of research in the future. This middle lamellae is rarely found alone and a primary wall often forms at the same time. In its purest form the primary wall is composed of polyglucose. The linkages are ~-1,4 and the resulting polymer is cellulose. Cellulose consists of ~glucopyranoside units linked by ~-1,4 linkages. The chains are unbranched, but various straight chains are cross-linked by hydrogen bonds, to form crystalline

85 fibres, in patterns that vary to form different structures (1). The great majority of cellulose in plants is in this compact crystalline form, which hinders enzymatic attack (Figure 2). There is rarely anything other than a homopolymer of glucose in cellulose. I

? 5~ 14

x

OH

O

="

-

Cellot)iose repealing unit

_

n

~. ~nm

Conformation of ce{lulo~e molecule 10nl

50

Figure 2. The internal structure of a cellulose microfibril.

As secondary walls are formed the cellulose becomes embedded in a matrix of both carbohydrates and polyphenolic polymers. The carbohydrate polymers are known as hemicellulose. Hemicellulose, a term first coined in the 19th century to signify the easily hydrolysable part of the cell wall is an amorphous polymeric carbohydrate matrix of variable composition. In hardwoods (which generally contain more hemicellulose than softwoods) it consists mainly of xylan, while in softwoods the polymers are mainly of glucomannan although xylan is still a major component. In both types of woods small amounts of hemicellulose consisting mainly of galactose are also present. The xylan most common in hardwood (4-0-

86 methyl-D-glucuronoxylan) consists of a backbone of 150 to 200 xylopyranose units (some of which O-acetylated at C2 or C 3) linked in a straight chain through a ~-1,4 linkage, with side-chains of 1,2-~-linked 4-O-methyl-D-glucuronic acid (Figure 3a). Softwood xylan is similar to hardwood xylan, but it has no O-acetylation, and in addition to 4-O-methyl-D-glucuronic acid it contains side-chains of Larabinofuranose attached at the C3 position (2; 3) (Figure 3b). Many variations on these general themes are found in different plant species.

(a) "0

0

Ar

9

0

0

.

0 "..

0

COOH i CH~~ OH

(b)

/--o OH

J--o ~

J--o

,oo 9 :i 0

CH~O~

oA,:

J---o /"

0

oM

HOH2

OH

OH

Figure 3. (a) The structure of hardwood xylan, mainly 4-0-methyl-D-glucuronxylan. (b) The structure of softwood xylan has additional side-chains of L-arabinofuranose.

The polyphenolic component is known as lignin and is a major component of woody perennials. As it is hydrolysed by completely different classes of enzymes it is not discussed here.

87

The Enzymes The activity of the enzyme which hydrolyses pectic acid or the methylated form of pectin has been classified as endo-~,l-4, polygalacturonase, EC 3.2.1.15. This enzyme randomly cleaves ~-l,4-hnkages in a chain of polygalacturonate. A pectin methyl esterase EC 3.1.1.11. enzyme removes the methyl group from CGto produce chains of pectic acid, Finally pectate lyase (EC 4.2.2.2), randomly cleaves the chains of polygalacturonate by ~-elimination to produce a double bond. There are also recognised exopolygalacturonate lyase (EC 4.2.2.9) and pectin lyase (EC 4.2.2.10). The latter enzyme also cleaves randomly (Figure 4).

a)

(b)

Pectmmethylesterase I~d,- I'.~.

--

~"

co.oc~

IE~.RM~

o/~ E,,,IJ.P.G

~-

I'.G.

P.M.E.

-o-~ ~ _ ~ o

CO-OH

/ ~

+c.,-o.

NO CHAIN BREAKAGE

(c)

Pectic lyases H

CO'OCH3

OH

I

O

CO'OCH3

Cs

Oi./~

!%,: H

o

OH

I Tmns elimination

H

-o

CO.OCH~

OH

~

CO-OCH 3 I

~ "1

H

!~

OH

Figure 4. (a) shows the activity of pectin methyl esterase. (b) shows the activity of polygalacturonase. (c) shows the activity of pectate lyase.

88 All pectolytic and pectin esterase enzymes are found in a soluble form unlike cellulolytic systems. Cellulolytic systems are produced as soluble, freely diffusing proteins or, by a different group of organisms, as soluble aggregates which tend to bind to the substrate (4). In non-aggregating systems there are essentially three types of activities: ~-l,4-endoglucanase or endocellulase (EC 3.2.1.4) which cleaves internal ~-1,4 glycosidic bonds in cellulose, cellobiohydrolase or exocellulase (EC 3.2.1.91) which removes cellobiose from the non-reducing end of cellulose; ~glucosidase (EC 3.2.1.21) which hydrolyses cellobiose to glucose. Although the three activities are present on several separate enzymes, they act in synergy to degrade crystalline cellulose. The commonly accepted model is that the endoglucanase cleaves internal bonds in those 'amorphous' regions where the cellulose chains are not so closely packed together, producing many non-reducing ends t h a t can be attacked by the cellobiohydrolase. The main function of the glucosidase (not strictly a cellulase) is to remove the product inhibition by cellobiose on cellobiohydrolase. This model is based more on common sense than hard scientific evidence, and is by necessity oversimplified firstly because not in all nonaggregating systems it is possible to identify a cellobiohydrolase activity, and secondly because the enzyme specificities are often not clearly defined. For example some exo-enzymes also have some endo-activity and vice versa. Also, some cellulases show xylanase activity and vice versa. Aggregating systems, which are found in several cellulolytic anaerobic bacteria, are those in which the secreted cellulases aggregate to form highly organized multicellular enzyme complexes associated with the cell surface where they mediate the attachment of the cell to its insoluble substrate. In Clostridium thermoceUum the complex has been called the cellulosome and comprises at least fourteen distinct polypeptides, among which there are various endoglucanases and xylanases, at least one ~-glucosidase and one cellobiohydrolase, and S1, a nonenzymic structural subunit. Ruminococcus albus, Ruminococcus flavefaciens, Fibrobacter succinogenes and various other Clostridia have been shown to have a similar organization of cellulases in a multienzyme complex. The cellulase complex of Clostridium ceUulovorans also contains a nonenzymatic cellulose binding protein (5). This q u a t e r n a r y organization is necessary for an efficient hydrolysis of the natural substrate, as individual components have been shown to have little activity towards ordered substrates, such as crystalline cellulose or filter paper, but the combination of exoglucanase, cellobiohydrolase and ~-glucosidase activities often restores almost full activity (6). Although models for the operation of the cellulosome have been proposed, the advantage of the association of the various polypeptides in a quaternary structure and the synergistic relationships between them are not at all obvious. Mayer et al (1987) (7) showed by electron microscopy that the cellulosome can exist in two different physical states at different stages of culturing. In the early stages cellulosomes are closely packed together (tight cellulosomes), while they subsequently decompose to loosely packed complexes. In these loose cellulosomes, rows of equally spaced, equally sized polypeptides are visible. Mayer and coworkers propose that this regular structure is responsible for cutting cellulose in cello-oligosaccharides eight glycosyl units in length, which are eventually hydrolysed by other components of the cellulosome to cellobiose. The

89 architecture of C. thermoceUum seems to be at least partly held together by interactions between the enzymatic components with the structural subunit $1 as well as between the individual components themselves. Interaction with the $1 subunit is mediated by a highly conserved duplicated segment 22 residues long (8). The hydrolysis ofxylan makes use of xylanases (1,4 ~-xylanases, EC 3.2.1.8) as well as enzymes able to hydrolyse the side-chains such as arabinfuranosidases and mannases. Aggregating xylanolytic systems seem to be much rarer than their cellulose-degrading counterparts. The only "xylanosome" reported so far is the multienzyme complex found in Butyrivibrio fibrisolvens (9). Some cooperativity between endo- and exo-xylanases is documented eg for a xylanases and a ~xylosidase from Neurospora crassa (10). ~-xylosidase may increase xylanase activity by removal of product inhibition. Cooperativity between endo-xylanases and debranching enzymes such as acetylesterases (11) and arabinofuranosidases (12) is very important and has been shown to almost double the rate and/or extent of xylan hydrolysis as compared with the activity of xylanase alone.

Classification of the enzymes Recently Henrissat and Bairock (13) used sequence homology and hydrophobic cluster analysis to classify 482 sequences of glycosyl hydrolases and produce 45 families. The xylanase and cellulase families belong to families 1, 2, 5-12 and possibly families 26, 44 and 45. Polygalacturonase is in family 28 but pectate lyase is not classified in this system as it is a lyase rather than a hydrolase. However, the structure of pectate lyase has been solved and evidence from circular dichroism may be cited as showing that pectate lyase and polygalacturonase have similar folds (14). There is evidence that structural features are more highly conserved in cellulases and xylanases t h a n sequence evidence alone would indicate. This had led to proposals from Jenkins & Henrissat (15 & 16) that a superfamily could be formed from Henrissat families 1,2,5, 10 and 17.

S t r u c t u r a l s t u d i e s on c e l l u l a s e a n d x y l a n a s e f a m i l i e s 1,2,5, 10 a n d 7

A very important discovery has been that the number of tertiary protein structural folds making up all known globular proteins may not exceed 1-2000 (17). The discovery in 1990 that the cellobiohydrolase II from Trichoderma reesei is an 8-fold ~/(~ barrel (similar to the well documented triose phosphate isomerase type fold but differing in having seven strands not eight) has been endorsed and extended by the further discovery that enzymes from families 1, 2, 5, 10 and 17 also have this common fold. The structure of the active site in these families is highly conserved and moreover the alignment of the residues involved in substrate binding to certain loops of protein between the sheets a n d the helices is distinctive. The loops can be identified by reference to the family 10 Pseudomonas fluorescens xylanase solved in 1994 by Harris et al (18). Although the enzyme is a multidomain

90 system it has been possible to genetically manipulate the gene so that the catalytic core can be expressed separately. The other components of the complex consist of the cellulose binding domain and a flexible linker. The catalytic core of xylanase A consists of 347 amino acid residues and has the architecture of the 8-fold ~/(~ barrel, the TIM barrel fold. This is the most common protein fold as it has been found in 1 in 10 enzyme structures solved (Fig. 5 ).

Figure 5. Three-dimensional structure of Pseudomonas fluorescens xylanase.

The ~-barrel core is elliptical in cross section as are the majority of [~/(~barrels. The substrate binding has been identified by incorporation in the crystal of a xylan substrate. The identification of the binding sites was made possible by a m u t a t i o n in the molecule which converted an active site glutamate to a cysteine E246C. The molecule was rendered inactive by this mutation and so a substrate was bound but not hydrolysed.It was found that the substrate binding cleft was formed by long loops at the carboxy-terminal end of ~-strands 4 and 7 and short loops at the

91 c a r b o x y - t e r m i n a l end of ~-strands 5 and 6. The active site acid/base and nucleophile of xyl A are glutamate-127, located close to the carboxy-terminal end of ~4 and glutamate-246 at the carboxy-terminal end of ~7. It was found by Jenkins et al (15) t h a t the structures of ~-1,4 and ~-(1,3)(1,4) glucanases, solved by Varghese et al (19), can be superimposed on the xyl A structure with only a small s t a t i s t i c a l deviation. The ~-strands of the 8-fold ~/~ barrel superimposed much b e t t e r t h a n the (~-helices. The glucanases are classified by H e n r i s s a t & Bairoch (13) as family 17. The nucleophiles identified by both groups of workers in the respective enzymes are close together in the superimposition but there is some d i s a g r e e m e n t about the acid base glutamate. Glu 288 in the glucanases does not have an equivalent in the superimposed xylanase but Glu 93/94 does have an e q u i v a l e n t Glu at position 127 in the xylanase structure. The residue at this position seems the most likely candidate for the acid base if the chemical reaction is a retention mechanism, as the distance between the Oe atoms of Glu 93/94 and the nucleophile is optimal. On the other hand the distance between Glu 288 and the nucleophile would be consistent with a chemical reaction known as the inversion mechanism. The final a s s i g n m e n t of the catalytic g l u t a m a t e s in the glucanases therefore still has to be made. However the residues which are known from the crystallographic xylanase data to form the binding site for the xylopentaose can be shown to be present in the same position in the glucanases. These are an Asn or Gln, and two Trp or Phe residues depending on the individual enzyme. Aromatic residues, such as Trp, are often involved when carbohydrates are bound and so it proves in these cases. The structure of E.coli ~-galactosidase, classified as family 2 (EC 3.2.1.23), is also an 8-fold ~/~ barrel (20). Although the barrel is very distorted (c~-helix 5 is missing and ~-strand 6 is distorted) the mechanism seems to involve g l u t a m a t e s at the carboxy terminus of ~-strands 4 and 7. It has proved possible to examine the amino acid sequences of quite a n u m b e r of carbohydrases and J e n k i n s et al (15) had already identified, from sequence alignment with the P s e u d o m o n a s fluorescens x y l a n a s e structure, some other carbohydrase enzymes which seemed to have conserved g l u t a m a t e s in positions which would probably position t h e m at the carboxy terminus of ~-strands 4 and 7. These enzymes included the ~-glucosidase family (part of family 1 (EC 3.2.1.21)) and the family 5 cellulases (such as the cellulase from Ruminococcus flavofaciens). The discovery of structures of family 5 cellulases (21) & (22) (EC 3.2.1.4)) and the family 1 cyanogenic ~-glucosidase from white clover (23) confirmed these predictions because the enzymes have a cylindrical 8-fold ~/a barrel with active site g l u t a m a t e s positioned as desribed above. It is proposed t h a t the enzymes are known as the 4/7 superfamily by J e n k i n s et al 1995. It is interesting to note that when the structures of c~ and ~ amylase (EC 3.2.1.1. & 3.2.1.2) were determined (24) and (25) these were also 8-fold ~/c~ barrels with their major axis r u n n i n g between strands 1 and 5, in the case of ~ amylase, and between strands 2 and 6 in the case of ~ amylase. The former is the closest of these

92 two structures to the superfamily exemplified by the Pseudomonas fluorescens xylanase structure. These were classified by Henrissat as families 13 & 14. There are other enzymes, such as xylose isomerase, which appear to fall into this superfamily. Henrissat et al (16) were also able to use hydrophobic cluster analysis to detect similarities between families 1,2,5, and 10. They went on to identify members of family 30 (Glucocerebrosidase EC 3.2.1.45), 35(~-galactosidase EC 3.2.1.23), 39(~-xylosidase EC 3.2.1.37 and (~-L-iduronase EC 3.2.1.37 as also being part of this evolutionary grouping. It is clear that family 17 (EC 3.2.1.73) should have been included in Henrissat et al's classification. The review by Davies & Henrissat (26) shows ribbon diagrams of the main folds of the catalytic domains of the glycosyl hydrolases with well resolved tertiary structures. Even this simple representation makes it possible to identify close similarity at a gross level of structure between families 1,2,5, 10,13,14, 17 and 18. Structures of families 30, 35 and 39 are not yet available. Although the amylases (families 13 & 14) are outside of the scope of this article it is very interesting how the basic tertiary structure of 8-fold ~/(~ barrels has been able to evolve to generate different substrate specificities whilst retaining the same disposition of catalytic residues. The detailed interactions which enable binding of different substrates and give different cleavage patterns (exo or endo) have been identified and discussed by Davies and Henrissat 1995 (26). The active sites of the 4/7 superfamily form a distinct cleft which allows a random binding of several sugar units in the polymeric substrates. The detailed chemical interactions between different carbohydrate polymers and amino acid residues relies upon subtle positioning of hydrophobic residues, such as tryptophanes and phenylalanines, so that different hexose and pentose ring structures are bound. The end result is different specificities as shown in Figure 1 of Jenkins et al (15)

S t r u c t u r a l s t u d i e s on c e l l u l a s e s from family 6 ( c o r r e s p o n d i n g c e l l u l a s e F a m i l y B) Two structures have been reported for proteins belonging to this family. One is the cellobiohydrolase II (CBHII) from Trichoderma reesei (27), and the endocellulase 2 (E2) from Thermomonospora fusca (28). CBHII, the first of the two to be solved, has a catalytic core with a region of about 35 amino acids, which is conserved in all T.reesei cellobiohydrolases and endoglucanases, located at the Nterminus in CBHII but at the C-terminus in CBHI and EGI. This conserved region, the A box, is always connected to the main body of the enzyme by a region containing a high percentage of threonine and serine residues (29). In CBHII the residues are highly glycosylated (30). The molecule of CBHII was cleaved into two fragments with the endoprotease papain. The core fragment (the largest) retained some catalytic activity but had reduced activity on solid cellulose. The core region (residues 83-447, molecular weight 46,000) was crystallized in the monoclinic space group P21 from a PEG solution in hanging drop (31) (C-terminal catalytic domainresidues 83-447). The smaller portion of the molecule (residues 3-38) was obtained by papain catalysed hydrolysis of the whole protein and the fragment structure

93 solved by n.m.r. The structure of the catalytic core revealed a large, single domain ~/a-protein, similar but different to the (~/(~)8 barrel of triose phosphate isomerase fold (TIM barrel) (Figure 5). The structure, solved by multiple isomorphous replacement (MIR) used inhibitors containing S and I atoms as additional heavy metal derivatives and consisted of a central ~/~-barrel formed by seven parallel ~strands, the first six of which are connected by a-helices, while the 6th and 7th strand are irregularly connected. The shear in the barrel is 8. Despite previous suggestions, based on theoretical considerations, that packing inside a sevenstranded barrel would be poor, hydrogen-bonding interactions inside the barrel are as extensive as in the TIM barrel, except for the closure of the barrel between strand 1 and 7. The active site, identified as a 20A long tunnel at the carboxyterminal end of the ~/(~-barrel, is almost completely enclosed by two long loops, each containing a disulphide bridge.

Figure 6. Structures of the catalytic cores of (a) Trichoderma reesei CBHII (Rouvinen et al, 1990) and (b) Thermomonospora fusca Endoglucanase 2 (Spezio et al, 1993).

94 The structure of the catalytic domain of endoglucanase 2 from T. fusca, which is in this case the N-terminal end of the molecule, is very similar in topology to CBHII, as to be expected since these two proteins belong to the same family. There is an additional ~-strand (strand VIII) which interacts in antiparallel fashion with s t r a n d VII forming a ~-bridge and is also involved in electrostatic interactions which probably aid barrel closure. As in the CBHII structure, barrel closure involves only hydrogen bonding, unlike in the TIM fold. Crystals used for this analysis were grown using a m m o n i u m sulphate as precipitant and belonged to the monoclinic space group P21. Microseeding was used to obtain crystals of a suitable quality. After a t t e m p t s to solve the structure by molecular replacement failed, probably because the identity between the two structures is only 26%, the structure was solved by MIR. One main difference between the two structures is the fact t h a t the active site of E2 is a much more open cleft than the CBHII tunnel, because one of the two loops enclosing the active site of the latter is deleted in the E2 structure and the other loop adopts a different conformation. This difference is of great biological consequence, as it offers a rationale for the difference in the specificity of the two enzymes, the first an exo- and the second an endo-cellulase (E2 is 300 times more active on CMC and 20 times more active on swollen cellulose t h a n CBHII). The enclosed, tunnel-like active site of CBHII only allows efficient binding of cellulose at its terminal end. This partly confirms hypotheses which had been made previously on the basis of sequence comparison only, predicting deletions of loops in endo-cellulases resulting in a more open active site. This is not always the case, as in the E2 structure, as the loops may also be removed from the active site by changes in the conformation. As both structures have been solved also in the presence of inhibitors, this allows further u n d e r s t a n d i n g of the catalytic mechanism. Four subsites for glucosyl binding were clearly identified in the structure of CBHII, with cleavage of the glycosidic bond occurring most likely between subsites B and C. Asp175 and Asp221 appeared to be the most likely candidates as the proton donors in a single displacement reaction (family 6 members are inverting glycosidases), as they are conserved in all four and three members of family A respectively, and in a suitable position for this role. Site directed mutagenesis has confirmed Asp221 as the catalytic residues (the m u t a n t D221A was shown to be completely inactive), while m u t a t i o n of Asp 175 resulted in a residual activity of 4-20 %. Asp263 is the most likely candidate as the nucleophile involved in the activation of the water molecule at the opposite side of the ring. In E2, Asp117 and Asp265 are closest to the proposed cleavage site. Of these Asp 117 is most likely to function as proton donor because it is unlikely to be ionized in its environment, while Asp265 forms a strong salt bridge with Arg221. Sitedirected m u t a g e n e s i s of these residues has not yet confirmed this hypothesis.

95

S t r u c t u r a l s t u d i e s on family 7 c e l l o b i o h y d r o l a s e I and e n d o g l u c a n a s e Although the cellobiohydrolase enzymes cleave from the chain ends it is probably cellobiohydrolase I that is the key enzyme for efficient hydrolysis of native crystalline cellulose. In the specific case of the enzyme from Trichoderma reesei, where the tertiary structure was solved by Divne et al (32), it was found that the enzyme was the most a b u n d a n t of the proteins involved in hydrolyzing cellulose. Removal of the gene from the fungus reduces overall activity against cellulose by 70%. Based on amino acid homology this enzyme was classified as a family 7 enzyme. As in many other instances this protein was multidomain with a cellulose binding domain and a heavily glycosylated linker to a catalytic domain. When crystallized and the structure solved by x-ray crystallography the catalytic domain had one-third of the 434 residues arranged in two antiparallel ~-sheets that are face to face forming a ~ sandwich. Thus i~hey are quite unlike the 4/7 superfamily described earlier and, apart from four short ~-helices, the protein consists almost entirely of loops connecting the sheets, stabilisation seems to be provided by nine of the ten disulphide bonds. This is in contrast to the 8 fold ~/a-barrels which have only a low number of disulphide bridges(one in the case of the Psezldomonas xylanase). The interface between the sheets in cellobiohydrolase I has a considerable volume which is packed with hydrophobic side chains but contains one hydrophillic patch formed by residues from both sheets. However despite the clearly different secondary and tertiary stucture between the TIM barrels and cellobiohydrolase I the catalytic and binding residues are found to be similar. The catalytic residues in CBHI comprise Glu 212, Asp 214 and Glu 217. There are seven binding sites identified in a tunnel made of the loops and the binding sites often involve the hydrophobic face of the glycosyl unit stacking onto an indole ring of Trp. This is also the case in this enzyme. It is believed t h a t Glu 212 is the probable nucleophile. It has been suggested that the tunnel has become lost in the evolution of related endoglucanases, creating structures with a cleft type of motif. (27). This has been confirmed in the cellobiohydrolase family when an homologous enzyme-endoglucanase I-from Trichoderma reesei was modelled on the CBHI structure and several deletions map to active site loops. This is supported by the structure of Humicola insolens endoglucanse (27) The CBHI core shows local similarity to the ~ 1,3-1,4 glucanases. This was originally used to indicate structural homology (32) but once the structure was solved it was obvious that there is a considerable difference in the two structures.

S t r u c t u r a l studies on family 9 cellulase e n z y m e ( c o r r e s p o n d i n g c e l l u l a s e F a m i l y E) The structure of a recombinantly expressed endoglucanase CelD from Clostridium thermoceUum, has been solved by Multiple Isomorphous Replacement ((33 & 34) (Figure 7). The enzyme crystallized in the trigonal space group P3121 using ammonium sulphate or CaCI._~as precipitant. CelD has a globular, somewhat elongated shape, consisting of two structural domains: a small N-terminal ~-barrel

96

Figure 7. Structure of Clostridium thermoceUum CelD (Juy et al, 1992).

with an immunoglobulin-type fold, of unknown function, closely packed against a large domain consisting of 12 major (z-helices, with a novel fold which the authors t e r m a 'twisted a-barrel'. The barrel is formed by six inner helices, all with the same orientation, and six outer helices, also all with the same orientation but opposite to the one of the inner helices. The helices are organized in a nearest neighbour up and down pattern. The active site is probably a large cleft r u n n i n g across the top of the helical domain, which is partially occupied by glucose and by the inhibitor o-iodobenzyl-l-thio-[~-cellobioside in the enzyme-ligand structures t h a t have also been solved. The active site is likely to consist of at least six subsites. In the crystal structure the groove is partially occupied by a signal peptide from a neighbouring chain, which does not belong to the native CelD sequence. It is likely t h a t no hindrance to the binding of cellulose at either sides of the groove exists in the native enzyme, and a n u m b e r of a r o m a t i c and polar groups are positioned favourably to i n t e r a c t with the s u g a r chain. Sequence c o m p a r i s o n w i t h m e m b e r s of the same family shows t h a t they can be divided in two groups according to the presence or absence of the small [~-sheet domain. As m e n t i o n e d previously, site-directed mutagenesis identified five carboxylates the m u t a t i o n of which causes severe reductions in catalytic activity. Of these, Glu555, Asp201 and Asp198 are positioned towards the centre of the active site cleft. As

97 Asp 198 is 8A away from the labile glycosidic bond, being pulled away from the active site by the neighbouring His197 which is involved in Zn binding, and can be seen to form hydrogen bonds with other protein atoms in the structure, it is an unlikely candidate for direct involvement in catalysis. The carboxylate of Glu555 is in a equivalent position as the one of Glu35 of lysozyme, if the two active sites are superimposed and is therefore likely to be the proton donor. However, unlike Asp52 of lysozyme, Asp201 of CelD is too far from the C1 of the charged reaction intermediate to interact directly with it, but it is in an appropriate position to be the nucleophile base activating a water molecule in an inversion-type mechanism. This is indeed the stereochemistry of reaction observed for CelD and other members of family E (35).

Structural studies on family 11 xylanases (corresponding xylanase family

G) This family consists of small (around 20 kDa in molecular weight), soluble xylanases. They are therefore relatively easy to crystallize with respect to, for example, cellulase components of multienzyme complexes. This is reflected in the relative abundance of crystallization reports for members of this family. Production of crystals suitable for X-ray analysis have been reported for xylanase I and II from Trichoderma reesei (36), an endo-xylanase from Aspergillus oryzae (37), a thermophilic xylanase from a Bacillus sp. (38), a xylanase from Trichoderma harzianum (39), and the 3-D structure of one of them, a xylanase from Bacillus p u m i l u s has been solved (40) (details of crystallizations are shown in the table). The solution of this structure revealed a novel fold consisting of three antiparallel ~-sheets forming a sandwich-like structure (Figure 8). This is reminiscent of the 3-D structure reported for a 1,3-1,4-~-glucanase from a Bacillus sp., also forming a ~- sandwich, but the connectivity appears to be rather different (41). On the assumption that the active site is located in a cleft region about 3 nm long and 1.5 nm in diameter (which is large enough to accommodate a xylan fibre about 1.1 nm in diameter) site directed mutagenesis was performed on three out of thirteen carboxylic residues of the enzyme which were both situated in the cleft and conserved in other Family G enzymes. Mutation of two of these residues, Glu93 and Glu182 abolished activity, indicating that these residues are directly involved in catalysis, while residual activities of up to 18% were found for mutations in Asp21, which may cause local conformational changes unfavourable to substrate binding (42). Although other family 11 protein xylanases are expected to have a similar topology, because all these xylanases have slightly different catalytic characteristics, a resolution of the local differences in the 3-D structure would be very valuable in the understanding the differences in pH optimum, differences in product distribution, in terms of structure-function relationships, as well as clarifying the structural features that confer thermostability. As far as the latter is concerned, the approach of random mutagenesis has been taken to improve the thermostability of B. Pumilus xylanases (40).

98

Figure 8. Structure of Bacillus p u m i l u s xylanase (Arase et al., 1993).

TABLE I (References are indicated in the text) PROTEIN

SPACE GROUP PRECIPITANT METHOD

T. reesei XYNI C2 T. reesei X Y N I I P21

ammonium sulphate ammonium sulphate

B. p u m i l u s T. h a r z i a n u m Bacillus sp.

P21 P212121 P1 ethylene glycol

A. oryzae

P21

PEG

hanging drop batch and seeding in hanging drop

RES.LIMIT 2.2A 1.5s 2.53, 2.8s

equilibrium dialysis hanging drop

2.5s 2.2A

99

S t r u c t u r a l s t u d i e s on family 28 p e c t a t e m o d i f y i n g e n z y m e s Although there are not any structures available for these enzymes the closely related enzymes pectin and pectate lyase have been crystallized (43) (44) and (45). An example of the crystallization conditions for the gram-positive bacterium Bacillus subtilis is as given in Jenkins et al (42). The structure of the B. subtilis and Erwinia enzyme has given the first structures of a new superfamily (45) (46). Mature B. subtilis pectate lyase (BPel) consists of 399 amino-acid residues (47 & 48), molecular weight 43,505 daltons, in comparison to PelC (Erwinia pectase lyase) which has 353 residues.

The parallel It-helix domain BPel has seven complete turns in the parallel ~-helix form strands one through seven of PB1, three through nine of PB2 and two through eight of PB3 (Figure 9). In total PB1 and PB3 have eight strands and PB2 has ten. PB1 and PB2 form an anti-parallel ~-sandwich with PB3 approximately perpendicular to PB2. Before the first turn of parallel ~-helix the polypeptide chain forms the first strand of PB2, an a-helix, the second strand of PB2, the first strand of PB3 and the first irregular long loop. In addition there is a 30 residue N-terminal extension that partially shields PB2 from solvent. After turn seven of the parallel ~-helix the polypeptide forms the final strands of PB1 and PB2, a short (~-helix and then the 30 residue Cterminal extension. The shortest turn of parallel ~-helix in BPel has only 19 residues. Residues in (~L-conformation occur frequently at both the amino- and carboxyends of the strands that form PB2 and less frequently at the carboxy-ends of the strands that form PB1 and PB3 of BPel. (~L-residues occur five times at the amino and six times at the carboxy end of PB2, three times at the carboxy end of and PB3 (and twice at the carboxy-end of PB1). These residues in mE-conformation rotate the strand through 180 ~ and change the polypeptide chain direction by approximately 90 ~ Their effect is to bend the ~-strands so that the ~-conformation residues at the end of turn T1, the strand of PB2 and the strand of PB3 can form a contiguous ~-strand with two sharp changes of direction. This is a striking feature of several turns of the BPel parallel ~-helix and a search of the protein databank (49) using graphics programs with fragment libraries (50) shows it to be unique. The effect is to maximize hydrogen bonding between adjacent turns of the parallel ~-helix. The hydrogen bonding of residues at the carboxy-end of T1 and through into PB2 is maintained elegantly for all carbonyls and amides in turns one through four of the parallel ~-helix, but the carbonyls of the residues in (xLconformation in T2 tend to point away from the axis of the parallel ~-helix to make hydrogen bonds with the C-terminal extension and associated water molecules. The carbonyls in the T2 turns point out apparently as a consequence of the following amides pointing into the structure and hydrogen bonding the side-chain of the asparagine that tends to follow the residue in (~L-conformation. This does not occur in the T1 turns as the residues after the residue in mE-conformation are

100

Figure 9. Schematic drawings of BsPel made using MOLSCRIPT. Arrows represent ~-strands and coils represents a-helices. Parallel ~-sheet 1-3 are PB1, PB2 and PB3. The active site calcium is the sphere which binds an aspartate on the surface of the fourth strand of PB1 and two from the long loops.

hydrophobic. These residues in (~L-conformation are a new type of ~-bend of which the aL-Asn turn is a variant that does not maintain the ~-sheet hydrogen bonding so elegantly. The sequence asparagine, histidine, tyrosine occurs on the outside of the parallel ~-helix at the (~L-positions ofT1 turns and T2 turns which is also seen for the T2 turns of PelC, asparagine, histidine and tyrosine occurs almost exclusively at this position. In BPel, like PelC and PelE, the T2 turn is the most regular but it is not invariant since there is a four residue insertion and no residue in aL-conformation in the first T2 turn of the parallel ~-helix of BPel. BPel has an impressive stack of five aromatics, phenylalanines 159 and 201 and tyrosines 242, 273 and 295, within the parallel ~-helix domain (Figure 10) compared to three in PelC and four in PelE. In comparison the asparagine ladder is not as pronounced in BPel with four asparagines. The greatest variability in the turns occurs between the T3 loops.

101

Figure 10. The regularity of the parallel ~-helix domain in all atom representation. The six residue aromatic stack and the asparagine ladder can be seen clearly.

The long loops (T3 turns) The most striking difference between the Erwinia PelC and PelE and Bacillus Pel is the additional structural region formed by three long loops. The first long loop precedes turn one of the seven complete turns of parallel ~-helix but it is effectively a T3 loop as it connects strand one of PB3 with strand one of PB1. The other two long loops are T3 turns two and three of the parallel ~-helix. These loops form a local globular structure of 104 residues consisting of two aLhelices and two extremely distorted ~-hairpins. This local globular structure has its own hydrophobic core and only limited interactions with the rest of the molecule. However, it is unlikely that this structural region could fold independently of the parallel ~-helix of pectate lyase. The loops are aromatic-rich and especially tyrosine-rich. There are no main-chain hydrogen bonds between the three long loops. The first long loop, 58 residues, is formed from residues 64 to major helical elements 85 to 91 and 105 to 122. The rest of the except for a single turn of helix between residues 75 and 78. aromatic and aspartate rich distorted ~-hairpin 24 residues

121 and has two loop is irregular Loop two is an long comprising

102 residues 161 to 184. The loop has five aspartates, no basic residues and six aromatics. The regularity of this hairpin is limited by prolines 16'7 and 171 between which the amides and carbonyls of residues 168 and 170 with the carbonyls and amides of 180 and 178 form a short a'ntiparallel ~-hairpin. Proline 171 causes the polypeptide chain to bulge for a couple of residues before the t u r n at the tip of the hairpin. The amide of residue 172 in this bulge is stabilised by a hydrogen bond to the side-chain of a s p a r t a t e 170. The five residue t u r n at the tip of the hairpin has a hydrogen bond across its base. Towards the base of the hairpin proline 168 kinks the polypeptide chain and causes the hairpin to widen. Aspartate 164 is buried between the ends of the hairpin and makes hydrogen bonds to amides of residues 166 and 182 and to a buried w a t e r molecule. A s p a r t a t e 184 is a ligand to the active site calcium-ion. Loop two makes a number ofhydrophobic interactions with loop three but predominantly hydrophilic interactions with loop one. Proline 167 and the two t r y p t o p h a n s 169 and 179 of loop two interact favourably with tyrosines 215 and 219 of loop three. Loop three is a 22 residue loop, residues 203 to 224, with less regular ~-structure t h a n loop two. There are two intra-loop hydrogen bonds between the carbonyl-oxygen of residue 212 and the amide of 219 and between the amide of 214 and the carbonyl-oxygen of 217. The loop opens up t o w a r d s its base with histidine 221 and tyrosine 163 inserted between the ends. A s p a r t a t e s 203 and 208 are buried at the base of the loop and interact with histidines 245 and 222 respectively. The irregular structure of this loop region is in striking regularity of the parallel ~-helix domain.

contrast to the

Calcium binding Calcium is bound to three aspartates in a pronounced cleft between the parallel ~-helix domain and the loop region. We therefore propose t h a t this is the active site cleft in BPel. The calcium-binding aspartates 184 and 223 occur at the carboxy-end of the second and third long T3 loops respectively and a s p a r t a t e 227 occurs on the ~-strand of PB1 t h a t follows the third long loop. Comparison of the sequences of PelC and BPel shows t h a t this calcium binding site is the site proposed on the basis of heavy atom data (44). However, in PelC the calcium binding ligands are two a s p a r t a t e s and a g l u t a m a t e and the active site cleft is not as pronounced. The calcium in BPel has seven ligands: one carboxy-oxygen from each of a s p a r t a t e s 223 and 227, both carboxy-oxygens of a s p a r t a t e 184 and the oxygens of three clearly defined w a t e r molecules. The two carboxy-oxygens of 184, one of 223 and two waters are approximately co-planar with the carboxy-oxygen of 227 on one side and the t h i r d w a t e r on the other. The calcium-oxygen distances vary between 2.34A and 2.55A with an average of 2.42A; both the number of ligands and the distances are compatible with the known geometry of calcium binding sites. The crystals were soaked in 5.0mM calcium chloride before the high resolution data collection and the calcium-ion has full occupancy in the crystal which suggests t h a t the K M for calcium is lower t h a n 0.5raM.

103 The only obviously conserved residue across all the pectate and pectin lyases close to the active-site calcium is arginine 279 which we believe may be one of the active site residues. This region of the active site is more basic in the pectate lyases than the pectin lyases due to the proximity of arginine 284 and lysine 247 and in BPel arginine 282. Arginine 279 may make short, strong hydrogen bonds to the carboxylate of the galacturonic acid residue thereby increasing the rate of abstraction of the proton from the C5 carbon (Gerlt and Gassman, 1993). If this is so the calcium may be involved in binding the carboxylate of the preceding galacturonic-acid residue. In the pectin lyases, which lack the calcium binding site, arginine 279 may make somewhat weaker hydrogen bonds to the esterified substrate but for this substrate proton abstraction from C5 is easier. An alternative is that arginine 284, which is only present in the pectate lyases, may be involved in making short, strong hydrogen bonds and arginine 279 stabilises the negative charge that develops on abstraction of the C5 proton. This issue will only be resolved when the structure of a pectate lyase-inhibitor complex is solved. Arginine 279 occurs in the sequence arginine, alanine, cis-proline in fungi. The arginine is RMP in plants (Me-aromatic changes) in a~-conformation and the carbonyl of the alanine hydrogen bonds to one of two buried water molecules that hydrogen bond the tyrosines of the aromatic stack of BPel.

Discussion

The structures of the enzymes hydrolysing cellulose and xylanase have recently been shown to form variants of the triose phosphate isomerase (TIM-barrel) superfamily. Although it could be argued that the celD, with twelve helices, six inner and six outer, is the first representative of a new superfamily. The structure of the xylanase TIM barrel from Pseudomonas fluorescens also shows a distinct similarity to glucanases and amylases. Finally those enzymes modifying pectate form yet another structural superfamily. Thus we can speculate that the binding and hydrolysis of the three main plant cell wall carbohydrates has been achieved by both convergent and divergent evolution of different protein structural motifs. Cellulose seems to be bound by a common structure found also in glucoamylase enzymes and so the difference from other carbohydrases is the evolution of a binding domain. However the remarkable finding is that pectin is hydrolysed by structures which are not found in other protein superfamilies. Recent evidence relating to new structures of xylanases and cellulases has been discussed to show how the Henrissat classification can be "collapsed" reducing the number of classes into which these enzymes are placed.

104 References 1. Hotchkiss, A.T. 1989. Cellulose biosynthesis. In Plant Cell Wall Polymers, ACS symposium series, 399, 232-247. 2. Puls, J. and K. Poutaten. 1989. Mechanisms of enzymic hydrolysis of hemicelluloses (xylans) and procedures for determination of the enzyme activities involved. Enzyme Systems Ligno., 151-165. 3. Haun, J.L. 1970. Hemicellulose. HDB Pulp Paper Techn., 25-31. 4. Gilbert, H.J. and G.P. Hazlewood. 1993. Bacterial cellulases and xylanases J. Gen. Mic., 139,187-194. 5. Goldstein, M.A., Takagi, M., Hashida, S., Shoseyov, O., Doi, R.H. and I.H. Segel. 1993. Characterization of the cellulose-binding domain of the Clostridium cellulovorans cellulose-binding protein A. J. Bact., 175, 5762-5768. Creuzet, N., Berenger, J.-F. and C. Frixon. 1983. Characterization of exoglucanase and synergistic hydrolysis of cellulose in Clostridium stercorarium. FEMS Mic. Lett., 20, 347-350. 7. Mayer, F., Coughlan, M.P., Mori, Y. and L.G. Ljungdahl. 1987. Macromolecular organization of the cellulolytic enzyme complex of Clostridium thermocellum as revealed by electron microscopy. App. Env. Mic., 53, 2785-2792. 8. Tokatlidis, K., Salamitou, S., B6guin, P., Dhurjati, P. and J.-P. Aubert. 1991. Interaction of the duplicated segment carried by Clostridium thermocellum cellulases with cellulosome components. FEBS Lett., 291, 185-188. Lin., L.-L. and J.A. Thomson. 1991. An analysis of the extracellular xylanases and cellulases ofButyrivibrio fibrisolvens H17c. FEMS Mic. Lett., 84, 197-204. 10. Deshpande, V., Lachke, A., Mishra, C. Keskar, S. and M. Rao. 1986. Mode of action and properties of xylanase and beta-xylosidase from Neurospora crassa. Biotec. Bioeng., 28, 1832-1837. 11. Biely, P., MacKenzie, C.R., Puls, J. and H. Schneider. 1986. Cooperativity of esterases and xylanases in the enzymatic degradation of acetyl xylan. Bio/Tec., 4, 731-733. 12. Bachmann,S.L. and A.J. McCarthy. 1991. Purification and cooperative activity of enzymes constituting the xylan-degrading system of Thermomonospora fusca. App. Env. Mic., 57, 2121-2130. 13. Henrissat, B. and A. Bairoch 1993. New families in the classification of glycosyl hydrolases based on amino acid sequence similarities. Biochem. J., 293, 781-788. 14. Goodenough, P.W., Clark, D.C., Durrant, A.J., Gilbert, H.J., Hazlewood, G.P. and G. Waksman 1991. Structural analysis by circular dichroism of some enzymes involved in plant cell wall degradation. FEBS Letters, 282(2), 355-358. 15. Jenkins, J., Leggio, L.L., Harris, G.& Pickersgill, R. 1995 B-Glucosidase, B Galactosidase, family A cellulases, family F xylanases and two barley glucanases form a superfamily of enzymes with 8-fold B/a architecture and with two conserved glutamates near the carboxy-terminus ends of B-strands four and seven. FEBS Letters 362, 281-285. .

.

105 16. Henrissat,B., Callebayt, I., Fabrega, S., Lehn, P.,Mornon,J-P. & Davies,G. 1995 Conserved catalytic machinery and the prediction of a common fold for several families of glycosyl hydrolases. Proceedings of the National Academy of Science. USA. 92,7090-7094. 17. Chothia,C. 1992 One thousand families for the molecular biologist. Nature 357, 543-544. 18. Harris,G.W., Jenkins,J.A., Connerton,I., Cummings,N., Leggio,L.L., Scott, M., Hazlewood,G.P., Laurie,J.I., Gilbert,H.J. and Pickersgill, R.W. 1994 Structure of the catalytic core of the family F xylanase from Pseudomonas fluorescens and identification of the xylopentaose-binding sites. Structure 2, 1107-1116. 19. Varghese,J.N., Garrett, T.P.J., Colman,P.M., Chen,L.,Hoj,P.B., and Fincher,G.B. 1994 Three-dimensional structure of two plant b-glucan endohydrolases with distinct substrate specific proceedings of the National Academy of Science. USA. 91, 2785-2789. 20. Jacobson,R.H., Zhang,X-J., Dubose, R.F.and Matthews, B.W. 1994 Nature 369, 761 766. 21. Dominguez,R., Souchon, H., Spinelli,S., Dauter, Z., Wilson, K.S., Chauvax,S.,Beguin, P. and Alzari, P.M. A common protein fold and similar active site in two distinct families of b-glycanases. 1995 Nature Structural Biology 2,569-577. 22. Ducrois,V., Czjzek, M., Belaich,A., Gaudin,C., Fierobe,H-P., Belaich,J-P., Davies, G.J. And Haser,R.1995 Crystal structure of the catalytic domain of a bacterial cellulase belonging to family 5. Sructure 3 939-949. 23. Barrett,T., Suresh, C.G., Tolley,S.P., Dodson,E.J. and Hughes, M.A. 1995 The crystal structure of a cyanogenic b-glucosidase from white clover, a family 1 glycosyl hydrolase. Structure 3 951-960. 24. Boel,E., Brady,L., Brzozowski,M., Derewenda, Z, Dodson,G,G., Jensen,V.J., Petersen,S.B., Swift, H., Thim, L., Woldike,H.F., Calcium-binding in alpha-amylases - an x-ray-diffraction study at 2.1-a resolution of 2 enzymes from Aspergillus.Biochemistry 29, 6244-6249. 25. Mikami,B., Hehre,E.J., Sato,M, Katsube,Y, Hirose,M., Morita,Y, Sacchettini,J.C. (1993) The 2.0-angstrom resolution structure of soybean beta-amylasecomplexed with alpha-cyclodextrin Biochemistry, 32, 6836-6845 26. Davies,G and Henrissat, B (1995) Structures and mechanisms of glycosyl hydrolases. Structure 3, 853-859. 27. Rouvinen, J., Bergfors, T., Teeri, T., Knowles, J.K.C., and T.A. Jones. 1990. Three-dimensional structure of cellobiohydrolase II from Trichoderma reesei. Sci., 249, 380-386. 28. Spezio, M., Wilson, D.B. and P.A. Karplus. 1993. Crystal structure of the catalytic domain of a thermophilic endocellulase. Biochem., 32, 9906-9916. 29. Van Tilbeurgh, H., Tomme, P., Claeyssens, M., Bhikhabhai, R. and G. Pettersson. 1986. Limited proteolysis of the cellobiohydrolase I from Trichoderma reesei. Separation of functional domains. FEBS Letts. 204(2), 223.

106 30. Tomme, P., Van Tilbeurgh, H., Pettersson, G., Van Damme, J., Vandekerckhove, J., Knowles, J., Teeri, T. and M. Claeyssens. 1988. Studies of the cellulolytic system of Trichoderma reesei QM 9414. Analysis of domain function in two cellobiohydrolases by limited proteolysis. Eur. J. Biochem., 170, 575-581. 31. Bergfors, T., Rouvinen, J., Lehtovaara, P., Caldentey, X., Tomme, P., Claeyssens, M., Petterson, G., Teeri, T., Knowles, J. and T.A. Jones. 1989. Crystallization of the core protein of cellobiohydrolase II from Trichoderma reesei. J. Mol. Biol., 209, 167-169. Arase, A., Yomo, T., Urabe, I., Hata, Y., Katsube, Y. and H. Okada. 1993. Stabilization of xylanase of random mutagenesis. FEBS Lett., 316, 123-127. 32. Divne,C., Stahlberg,J., Reinikainen,T., Ruohonen,L., Pettersson, G., Knowles,J.K.C., Teeri,T.T., Jones,T.A. (1994) The three dimensional crystal structure of the catalytic core of cellobiohydrolase I from Trichoderma reesei.Science 265,524-528. 33. Juy, M., Amit, A.G., Alzari, P.M., Poljak, R.J., Claeyssens, M., B~guin, P. and J.-P. Aubert. 1992a. Three-dimensional structure of a thermostable bacterial cellulase. Nature, 357, 89-91. 34. Juy, M., Amit, A.G., Alzari, P.M. and R.J. Poljak. 1992b. Three-dimensional structure of a thermostable bacterial cellulase. In "Cellulose hydrolysis and fermentation". J. Coombs and G. Grassi (eds). pp 18-29 35. Gebler, J., Gilkes, N.R., Claeyssens, M., Wilson, D.B., B~guin, P., Wakarchuk, W.W., Kilburn, D.G., Miller, R.C.Jr., Warren, R.A.J. and S.G.Withers. J.B.C., 267, 12559-125 61. 36. Torronen,A., Rouvinen, J., Ahlgren, M., Harkki, A. and K. Visuri. 1993. Crystallization and preliminary X-ray analysis of two major xylanases from Trichoderma reesei. J. Mol. Biol., 233, 313-316. 37. Golubev, A.M., Kilimnik, A.Yu., Neustroev, K.N. and R.W. Pickersgill. 1993. Crystals ofbeta-xylanase from Aspergillus oryzae. J. Mol. Biol., 230, 661-663. 38. Pickersgill, R.W., Debeire, P., Debeire-Gosselin, M. and J.A. Jenkins. 1993. Crystallization and preliminary X-ray analysis of a thermophilic Bacillus xylanase. J. Mol. Biol., 230, 664-666. 39. Rose, D.R., Birnbaum, G.I., Tan, L.U.L. and J.N. Saddler. 1987. Crystallization and preliminary diffraction studies of a xylanase from Trichoderma harzianu. J. Mol.Biol., 194, 755-756 40. Arase,A., Yomo,T,. Urabe,I., Hata,Y., Katsube,Y. and Okada,H. 1993. Stabilization of xylanase by random mutagenesis. FEBS Letters. 316,123-127. 41. Cambilllau, C. and H. van Tilbeurgh. 1993. Structure ofhydrolases: lipases and cellulases. Curr. Opin. Struc. Biol., 3, 885-895. 42. Ko, E.P., Akatsuka, H., Moriyama, H., Shinmyo, A., Hata, H., Katsube, Y., Urabe, I. and H. Okada. 1992. Site-directed mutagenesis at aspartate and glutamate residues of xylanase from Bacillus pumilus. Biochem. J., 288, 117-121.

107 43. Jenkins, J.A., Nasser, W., Scott, M., Pickersgill, R.W., Vignon, J.C. and J. Robert-Baudouy 1992. Crystallization and preliminary X-ray studies of the pectate lyase from Bacillus subtilis. J. Mol. Biol., 228, 1255-1258. 44. Yoder, M.D., DeChain, D.A. and F. Jurnak. 1990. Preliminary crystallographic analysis of the plant pathogenic factor pectate lyase from Erwinia chrysanthemi. J. Biol. Chem., 265, 11429-11431. 45. Yoder, M.D., Keen, N.T. and F. Jurnak. 1993. New domain motif: the structure of pectate lyase C, a secreted plant virulence factor. Science, 260, 1503-1507. 46. Pickersgill, R.W., Jenkins, J. A., Harris, G.W., Nasser, W. and J. Robert-Baudouy. 1994. The structure of Bacillus subtilis pectate lyase in complex with calcium. Nature Structural Biology, 1,717-723. 47. Nasser, W., Chalet, F. and J. Robert-Baudouy. 1990. Purification and characterisation of extracellular pectate lyase from BaciUus subtilis. Biochemie, 72, 689-695. 48. Nasser, W., Awad~, A.C., Reverchan, S. and J. Robert-Baudouy. 1993. Pectate lyase from Bacillus subtilis: molecular characterisation of the gene and properties of the cloned enzyme. FEBS Letters, 335, 319-326. 49. Bernstein, F.C. et al 1977. The protein databank: a computer based archival file for macromolecular structures. J. Mol. Biol., 112, 535-542. 50. Jones, T.A. and T. Sirup 1986. Using known substructures in protein model-building and crystallography. EMBO J., 5, 819-822.



109

Structure and Activity of Some Starch-metabolising Enzymes E. Ann MacGregor D e p a r t m e n t of Chemistry, University of Manitoba Winnipeg, Manitoba, Canada, R3T 2N2

Abstract Several enzymes that participate in the metabolism of starch are believed to share a common structural feature - their catalytic domain folds as a (~/~)s - barrel i.e. a cylinder of eight ~-strands surrounded by eight (~-helices. In each enzyme, the active site is made up of amino-acid residues situated on the ~-strands or loops protruding from the C-terminal ends of the ~-strands. (~-Amylases are the most widely-studied enzymes in this "family", but other enzymes belonging to the group can catalyse hydrolysis or synthesis of (z-l,4- or a-l,6-glucosidic linkages or both bond types. In this review, a description is given of current ideas of the relationship between key structural variations and specificity differences amongst the enzymes of the a-amylase family.

Introduction Several starch-hydrolysing enzymes are of commercial importance. Notable examples are a-amylase in brewing and baking, isoamylase and pullulanase as well as ~-amylase in the production of glucose syrups, and cyclodextrin glucanotransferases (CGTases) for the formation of cyclodextrins. Specificity, thermostability and pH response of the enzymes are critical properties for industrial use. Where naturally-occurring enzymes have undesirable characteristics, it is hoped that altered enzymes with improved properties may be made available in future by protein engineering. A good u n d e r s t a n d i n g of the relationship between protein structure and enzyme activity under different conditions is, however, a prerequisite for rational design of modified enzymes. Similar knowledge of starch-synthesizing enzymes will be required in order to use plant genetic engineering to allow agricultural production of altered starches with novel properties. In the last ten years, the three-dimensional structures of several starchdegrading enzymes have been determined by x-ray crystallography [1-13]. Thus it is now possible to examine different structures in the light of known enzyme activity and to try to establish the link between them. The studies have shown that enzymes acting on a- 1,4- or a- 1,6-glucosidic linkages with retention of configuration (e.g. a-amylases, CGTases and an oligo-l-6-glucosidase) are structurally distinct from enzymes giving inversion of configuration (e.g. ~-amylase or glucoamylase). In fact, the "retaining" enzymes studied to date have been found to belong to one

llO s t r u c t u r a l family, where the catalytic domain of each enzyme is folded to give a (~/a)8-barrel (see next section). F u r t h e r work has been carried out on starch-metabolising enzymes, where analysis of known amino-acid sequences is used to predict major features of threedimensional structure. Such investigations indicate that several enzymes, with a variety of hydrolysing or transglycosylating activities on a-l,4- or a-l,6-glucosidic bonds, belong to this "a-amylase family", and contain a (~/(z)8-barrel catalytic domain [14-20]. Crystallography of complexes of enzyme with substrate analogues, and molecular modelling, have given insight into the nature and a r r a n g e m e n t of amino acid residues at the active site of some enzymes of the (z-amylase family [1,21-24], and the importance of certain residues has been confirmed by work on m u t a t e d enzymes. Amino acid sequence similarities and differences allow conclusions to be drawn about structure-function relationships amongst members of this protein family, and some current ideas on the subject are discussed in this review.

S t r u c t u r a l f e a t u r e s c o m m o n to e n z y m e s of t h e a - a m y l a s e f a m i l y

All of the enzymes so far examined in this group are multidomain proteins i.e. the molecules contain more than one domain or folding unit. While the functions of some domains are known [25], others are unknown [19], but the catalytic domain can be identified, and in every case folds in the form of a (~/a)8-barrel. In this structure, ~-strands and (~-helices alternate along the polypeptide chain which is folded so that eight parallel ~-strands form an inner cylinder surrounded by eight helices. Usually an "extra" helix follows the sixth ~-strand. The alternating strands and helices are linked by irregular loops, and the active site of each enzyme is made up of amino acid residues on the ~-strands or the loops that join the C-terminal ends of the ~-strands to adjacent helices (Figure 1). Such a structure has been confirmed by x-ray crystallography in a-amylases and an oligo-l,6-glucosidase that catalyse hydrolysis of a-1,4- and (z-l,6-glucosidic bonds, respectively, and in CGTases that bring about hydrolysis and transglycosylation of a-l,4-1inkages [1-3, 6-13]. Structure prediction from amino acid sequences has led to the conclusion that the (~/(~)8-barrel should exist in other enzymes such as isoamylase and pullulanase, with hydrolytic activity on a-l,6-glucosidic bonds only, and also in enzymes like neopullulanase, the a-amylase-pullulanase of Clostridium thermohydrosulfuricum, starch branching enzymes and possibly cyclodextrinases that are active on both (z-1,4- and a-1,6- linkages between glucose residues [20]. Since the length of, and nature of amino acid residues on, loops joining structural elements of the ~-barrel can vary with little effect on the barrel itself, this is an ideal a r r a n g e m e n t for producing a variety of enzymes that are related, yet have different specificities and properties. In general, in this family of enzymes, the third loop linking ~-strand 3 to the third helix (Figure 1) is long, but lengths of other loops are very variable, and loops may even contain "extra" ~-strands and helices. In a-amylases and CGTases, a

111 1

2

3

4

5

6

7

8

P ~ N-term inal end of protein

= l~-strand

~

= co-helix

C-terminal end of protein

Figure 1. A r r a n g e m e n t , on a polypeptide chain, of ~-strands and helices of a (l~/a)8barrel. The enzyme active site is formed from amino acid residues on the l~-strands or on loops 1 to 8 t h a t join C - t e r m i n a l ends of ~-strands to adjacent helices.

calcium ion is needed for s t r u c t u r a l integrity [1-3, 7-11, 13], but it is not k n o w n w h e t h e r calcium is required in all e n z y m e s of this group. The active site of any enzyme acting on a polysaccharide can be described in t e r m s of contiguous subsites, where a subsite is a section of the active site t h a t i n t e r a c t s with one monosaccharide residue of the s u b s t r a t e (Figure 2). In the (~a m y l a s e family of enzymes, the subsites are composed of amino acid residues on the loops t h a t join ~-strands to helices; t h u s variations in lengths and n a t u r e of loops bring about differences in the n u m b e r and kinds of subsite at the active sites of the enzymes.

substrate enzyme surface subsite number o _ __glucose residue

catalytic site

Figure 2. Subsite s t r u c t u r e of the active site of an enzyme belonging to the cza m y l a s e family. Each subsite can interact with a glucose residue of the s u b s t r a t e . A g l u c a n chain is shown bound at the active site; bond splitting would occur b e t w e e n the glucose residues at subsites -1 and +1.

112 Modelling of enzyme-substrate interactions suggests that for enzymes acting on (z-l,4-glucosidic bonds, subsites binding the glycone part of the substrate (subsites -1 to -7 in Figure 2) can be formed from amino acid residues on loops 1, 2, 3, 7 and 8 of the [~-barrel, while subsites interacting with the aglycone segment of the substrate (subsites 1 to 3 in Figure 2) may consist of residues from loops 3 to 7. Loop 3, and in at least one case loop 7, are long enough to contribute residues to subsites on both sides of the catalytic site [1, 21, 22]. Examination of substrate structure (Figure 3) shows that the segment of substrate on the non-reducing side of the CI-0 bond to be attacked by an enzyme remains essentially the same whether the glucosidic bond has (~-1,4- or r configuration. Thus we may expect the branch of a branched substrate (rings A and B of Figure 3b) to be bound by subsites involving amino acid residues on loops 1 3, 7 and 8 of an enzyme active on (z-1,6- linkages. Obviously the aglycone part of a substrate differs according to whether the enzyme attacks an a-1,4- or a-1,6glucosidic bond [rings C and D of Figure 3a in the first case, and C1 to F of Figure 3b in the second). While rings C1 and D of the branched substrate (Figure 3b) may well be accommodated by subsites composed of residues from loops 3 to 7 of an enzyme, insufficient information is available to allow assignment of the binding of rings E and F (Figure 3b) to particular loops.

10 [73]. This loop does not appear to be present in cereal or thermophilic bacterial (z-amylases or Taka amylase A, so it is not surprising that such enzymes exhibit less or no multiple attack. Conserved residues of the sequences of Figure 5 are seen to be located in the active sites of (~-amylases of AspergiUus oryzae, barley, pig and h u m a n pancreas, B. licheniformis and a maltotetraose-forming amylase [1, 8, 11-13, 22]. The carboxylic acid groups equivalent to those of Asp 206, Glu 230 and Asp 297 (Figure 5), already implicated in catalysis, are seen close to the bond of the substrate likely to be hydrolysed, while side chains equivalent to His 122 and His 296 are found at subsite -1, in a position to interact with glucose residue B of the substrate (Figure 3) and stabilize a transient intermediate. The side chains equivalent to His 210 and Lys 209 are found close to subsites 1 and 2 in Taka, mammalian and B. licheniformis amylases, and are believed to have important implications for specificity. Cereal enzymes contain glycine as the residue equivalent to His 210 of Taka amylase that forms part of subsite 1, and are found to produce more glucose on glucan hydrolysis than other (~-amylases. This glucose is probably released from the reducing end of a substrate [62], suggesting that the histidine-.glycine substitution facilitates binding of a reducing chain end at subsite 1 in the cereal (zamylases. The importance of this histidine for specificity has been shown by chemical modification and mutation. In both pig pancreas and Bacillus amyloliquefaciens (z-amylases, the modified enzyme had increased activity for hydrolysing the bond between an (z-glucose residue and p-nitrophenol in pnitrophenyl oligosaccharides, and for the pig enzyme the pH for optimum activity

117 on amylose was lowered [56, 74-76]. That no similar change was found in barley a-amylase is consistent with the lack of the equivalent histidine residue. Although it has been suggested that this histidine side chain forms part of subsite 1, mutation of the equivalent residue to asparagine in pig and human ~-amylases appeared to have an effect at subsite 2 [77]. Such an effect may come about by distortion of the active site at subsites 1 and 2 in response to the mutation. The residue equivalent to lysine 209 of Taka amylase is also believed to be important in determining specificity. In enzymes acting on ct-l,4-glucosidic linkages only, this residue is always lysine, or less commonly arginine, while other residues are invariably found at this position in enzymes that hydrolyse ~-1,6bonds or have dual specificity for both bond types [20]. Matsuura et al. [1] have proposed that this residue makes up part of subsites 1 and 2, and participation in subsite 2 has been confirmed by Qian et al., [22]. Modification of the equivalent lysine in pig pancreas (~-amylase and Taka amylase A increased maltosidase activity, decreased (x-amylase activity and changed the pH-activity characteristics of the pig enzyme [57, 77]. The substitution of arginine for lysine at this position in isozyme 1 of barley (~-amylase has been implicated in the production of more glucose from substrates by the arginine-containing compared to the lysinecontaining isozyme 2 [62]. An equivalent mutation in Taka amylase A increased maltosidase activity [29], and in Saccharomycopsis fibuligera (x-amylase gave an enzyme releasing more glucose from maltotetraose [78, 79]. Although the relationship between transglycosylation activity and structure is not yet understood, mutation of a single amino acid residue at subsite -2 in S.fibuligera (~-amylase increased the transglycosylation catalysed by the enzyme. It was suggested that the change caused tighter binding of the glycone fragment of the substrate to the enzyme after hydrolysis, providing time for a second oligosaccharide to bind at subsites 1-3 and allow for formation of a new a-1,4linkage between the glycone fragment and the incoming saccharide [80]. A change in interaction with the 'catalytic' water molecule has also been proposed [81]. Greater changes in (x-amylase structure can also affect specificity. Truncation of a bacterial (~-amylase by removal of a segment from one end of the polypeptide chain appeared to affect the action pattern of the enzyme [48], while production of a chimaeric protein consisting of the polypeptide chain of a fungal a-amylase fused to a segment of glucoamylase chain yielded an enzyme with higher activity on starch granules [82].

Cyclodextrin glucanotransferases These enzymes bring about starch breakdown by removing segments from the non-reducing ends of (x-l,4-1inked glucans and cyclizing these segments to give cyclodextrins. They are also able, however, to hydrolyse polyglucans and catalyse transglycosylation reactions of smaller oligosaccharides. It is therefore not surprising that their active sites have some similarities to a-amylase active sites. The active site is believed to consist of subsites, and variations in position of the catalytic site within an array of nine subsites have been suggested to explain differences in the ratio of a-, ~- and y-cyclodextrins produced by CGTases from different sources [83]. That subsites 1 and 2 (Figure 2) exist in CGTases has been

118 confirmed by studies of acceptors for transglycosylation [84], and structural features of substrates required for binding at specific subsites have been investigated [8587]. Three carboxylic acid groups and two histidine side chains are invariably found at the active sites of CGTases, equivalent to Asp 206, Glu 230, Asp 297 and His 122 and 296 of Figure 5. They probably fulfill the same role in CGTases and (~-amylases [15, 21, 24, 30, 35]. In addition, the lysine 209-histidine 210 sequence of Figure 5 occurs in CGTases where the lysine is probably critical for specificity on a-l,4-glucosidic linkages, while the histidine forms part of subsite 1. The similarities between a-amylases and CGTases in the sequences of Figure 5 are so great, that enzymes have been mis-classified as (~-amylases when they may indeed be CGTases [20]. Important differences between the two types of enzyme appear to be an alanine-proline sequence in CGTases at positions equivalent to 119 and 120 of Figure 5, isoleucine in place of leucine 203, and tryptophan-phenylalanine(glycine or alanine) instead of valine 231 - leucine 232 - aspartic acid 233 [15, 88]. Studies of m u t a t e d CGTases have shown that this latter sequence is extremely important for cyclisation, with the phenylalanine residue in particular having an effect on activity [42]. This residue is situated on loop 5 of the ~-barrel and may be expected to be close to subsite 1 or 2 (Figure 2). Other residues, on loop 3, and histidines at the active site are also important for cyclization [35, 42, 46, 89-92]. When a CGTase removes a segment of six to eight a-l,4-1inked glucose residues from a longer chain, the segment will remain bound for a short time to the enzyme (at subsites -6 to -1 of Figure 2 for a hexasaccharide, for example). For cyclisation to occur the glucose residue at subsite -6 must be repositioned at subsite I while the potential reducing-end glucose residue of the fragment remains bound at subsite -1. Flexibility in the enzyme would help to bring this about, and indeed flexible sections of loops 1, 2 and 3 have been reported in a CGTase [3]. Molecular modelling indicates these segments could be involved in the mechanism of cyclisation [21]. CGTase molecules are longer than a-amylases and consist usually of five domains [3]. One domain helps to bind starch granules [25], but the importance for activity of the domains not found in a-amylases is not yet clear. Removal of ninety amino acid residues from the C-terminal end of a KlebsieUa pneumoniae CGTase produced little effect on activity, but removal of ten residues from the same end of a Bacillus species enzyme changed the ratio of a- to ~-cyclodextrin produced [52, 83]. In a study of chimaeric CGTases it was shown that both the beginning, i.e. loops 1 to 3 of the ~-barrel, and the C-terminal non-barrel domain are important for determining the relative amounts of different cyclodextrins produced by the enzymes [93].

Enzymes specific for (~-1,6-bond hydrolysis The structure of an oligo-l,6-glucosidase that removes non-reducing glucose residues linked by an a-l,6-bond has been determined by x-ray crystallography, and the presence of three carboxylic acid groups at the active site has been confirmed [6]. Other debranching enzymes are also believed to contain these residues [20]. Insufficient information is available, however, to identify other active site residues

119 in the oligo-l,6-glucosidase, particularly those making up subsite 1, which m u s t be substantially different from those forming subsite 1 of an (~-l,4-1inkage-specific enzyme of the (~-amylase family. Subsite -1 is, however, expected to be very similar in enzymes specific for (~-1,4- or (~-l,6-glucosidic bonds, as the histidine residues constituting this subsite are conserved. The lysine 209 (Figure 5) of (~-amylases and CGTases is conspicuously absent in debranching enzymes and is replaced by glycine in pullulanases, serine in isoamylase or asparagine in oligo-l,6-gluco sidase [20]. The oligo-l,6-glucosidase would be expected to have only subsite -1, of the subsites interacting with the glycone segment of a substrate, and indeed it has been suggested that loops 1 and 8 are very short, making it unlikely that they could form any of subsites -2 to-7 of Figure 2 [20, 44]. Isoamylase and pullulanase are both able to hydrolyse branch points in amylopectin, although only pullulanase has action on pullulan. Thus the specificities of the two enzymes differ, but the structural basis for this is not understood. Marked differences in lengths of loops 4, 5 and 7 may contribute to these specificity differences [20]. That the active sites of these enzymes do indeed bear some resemblance to the active sites of enzymes such as (~-amylases and CGTases can be concluded from the fact that both isoamylase and pullulanase are competitively inhibited by cyclodextrins [94, 95]. Enzymes active on both ~-1, 4- and c~-1,6-glucosidic bonds These enzymes can be divided into two broad categories- branching enzymes with a narrow specificity and enzymes such as neopullulanase that can hydrolyse and synthesise both (~-1,4- and (~-1,6- linkages. Starch-branching enzymes remove an ohgosaccharide from the non-reducing end of an (~-l,4-1inked glucan chain and reattach the oligosaccharide to a polyglucan by an (~-l,6-bond. Thus a saccharide containing glucose rings A and B of Figure 3a is detached by the branching enzyme and reattached as in Figure 3b. As yet little is known of the structure of these enzymes beyond the fact that they are likely to contain a (~/(~)s-barrel domain and operate by a mechanism involving three carboxylic acid groups at the active site, and two histidine residues for stabilization of the transient intermediate [20]. Starch-branching enzymes in general do not contain an asparagine residue equivalent to Asn 121 of Figure 5. Since the residue is recognised in (~-amylases as a Ca%binding ligand, this may mean the branching enzymes do not contain calcium, and may have a more flexible active site t h a n (~amylases, as is probably required for their complex action. The characteristic Lys 209 (Figure 5) found in (~-l,4-bond-specific enzymes is not present in branching enzymes, but the catalytic residue equivalent to Glu 230 is always followed by a second acid residue (aspartic or glutamic acid) [20]. The significance of this is not yet known. Several forms of branching enzymes are found in maize, rice and peas, with different specificities for forming long or short branches [96-98]. Lack of one form in rice gives the amylose-extender characteristic to the starch, which apparently contains a higher proportion of amylose and unusual amylopectin. The two major isozymes of branching enzyme have substantially different N- and Ct e r m i n a l ends, and these, with variations in active site loops 7 and 8, may confer a specificity difference on the isozymes [96, 98].

120 Several enzymes appear to be able to hydrolyse and/or synthesize (z-1,4- and (z1,6-glucosidic bonds. There is apparently some confusion about the naming and classification of these enzymes. Neopullulanases, for example, can hydrolyse pullulan to panose by attacking (z-l,4-1inkages, but can also hydrolyse or synthesize (z-l,6-bonds as well as make new (z-l,4-bonds [99, 100]. (z-Amylase-pullulanases or amylo-pullulanases bring about hydrolysis of (z-l,4-1inkages in maltodextrins and (z-l,6-1inkages in pullulan [101]. It is believed that some amylase-pullulanases contain two active sites [102-105], while others have one only [20, 32, 101, 106]. Yet another group of enzymes can act on cyclodextrins, maltodextrins, amylose and pullulan to give mainly maltose, but may differ in the ease with which they attack different substrates. Such enzymes can be called cyclomaltodextrinase, maltogenic amylase, pullulan hydrolase or even (z-amylase [33, 107-110]. The name (zamylase, at least, should be reserved for enzymes hydrolysing (z-l,4-glucosidic bonds only. These enzymes are all likely to contain ([~/(z)s-barrel catalytic domains [20] and for a neopullulanase, an amylo-pullulananse and a cyclomaltodextrinase, three carboxylic acid groups have been shown to be necessary for catalysis [28, 32, 111]. The two histidine residues involved in transition state stabilisation appear to be present in these enzymes [20], and their importance in neopullulanase has been shown by mutagenesis [28]. Many of the enzymes have an asparagine-glutamic acid sequence instead of the lysine 209 - histidine 210 of Figure 5 [20, 32, 33, 106, 112, 113], and mutation of these residues in neopullulanase has indicated that they exert some influence on the ratio of activities on a-l,4- versus (z-l,6-glucosidic bonds [28]. Although a good understanding of structure-activity relationships is lacking in these enzymes, they can be useful for starch degradation to give interesting oligosaccharide mixtures [114, 115].

Conclusion Many enzymes with a variety of activities on a-l,4- or a-l,6-glucosidic bonds are recognised as belonging to the a-amylase family of proteins. They share a common (~/(z)s-barrel fold for the catalytic domain, and probably also a common mechanism. Because of the close structural similarities amongst the enzymes, confusion can arise about classification of an individual enzyme. Nevertheless many of the enzymes have important commercial applications, and it is likely that new enzymes in this family will be discovered, having novel and useful specificities.

References .

.

M a t s u u r a Y, Kusunoki M, Harada W, Kakudo M. J Biochem (Tokyo) 1984; 95: 697-702. Boel E, Brady L, Brzozowski AM, Derewenda Z, Dodson GG, Jensen VJ, Petersen SB, Swift H, Thim L, Woldike HF. Biochemistry 1990; 29: 62446249.

121

.

4. .

.

.

8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30.

Klein C, Schulz GE. J Mol Biol 1991; 217: 737-750. Aleshin A, Golubev A, Firsov LM, Honzatko RB. J Biol Chem 1992; 267: 19291-19298. Mikami B, Sato M, Shibata T, Hirose M, Aibara S, Katsube Y, Morita Y. J Biochem (Tokyo) 1992; 112: 541-546. Kizaki H, Hata Y, Watanabe K, Katsube Y, Suzuki Y. J Biochem (Tokyo) 1993; 113: 646-649. Qian M, Haser R, Payan F. J Mol Biol 1993; 231: 785-799. Kadziola A, Abe J-I, Svensson B, Haser R. J Mol Biol 1994; 239: 104-121. Larson SB, Greenwood A, Cascio D, Day J, McPherson A. J Mol Biol 1994; 235: 1560-1584. Lawson CL, van Montfort R, Strokopytov B, Rozeboom HJ, Kalk, KH, de Vries GE, Penninga D, Dijkhuizen L, Dijkstra BW. J Mol Biol 1994; 236: 590-600. Brayer GD, Luo Y, Withers SG. Protein Sci 1995; 4: 1730-1742. Machius M, Wiegand G, Huber R. J Mol Biol 1995; 246: 545-559. Matsuura Y. In: Yamamoto T, ed. Enzyme Chemistry and Molecular Biology of Amylases and Related Enzymes. London: CRC Press, 1995; 137-145. MacGregor EA. J Protein Chem 1988; 7: 399-415. MacGregor EA, Svensson B. Biochem J 1989; 259: 145-152. Raimbaud E, Bul~on A, P~rez S, Henrissat B. Int J Biol Macromol 1989" 11: 217-225. Watanabe K, Kitamura K, Iha H, Suzuki Y. Eur J Biochem 1990; 192: 609620. Watanabe K, Chishiro K, Kitamura K, Suzuki Y. J Biol Chem 1991; 266: 24287-24294. Jespersen HM, MacGregor EA, Sierks MR, Svensson B. Biochem J 1991; 280: 51-55. Jespersen HM, MacGregor EA, Henrissat B, Sierks MR, Svensson B. J Protein Chem 1993; 12: 791-805. Klein C, Hollender J, Bender H, Schulz GE. Biochemistry 1992; 31: 87408746. Qian M, Haser R, Buisson G, Du~e E, Payan F. Biochemistry 1994" 33" 62846294. Casset F, Imberty A, Haser R, Payan F, P~rez S. Eur J Biochem 1995; 232: 284-293. Strokopytov B, Penninga D, Rozeboom HJ, Kalk KH, Dijkhuizen L, Dijkstra BW. Biochemistry 1995; 34: 2234-2240. Svensson B, Jespersen HM, Sierks MR, MacGregor EA. Biochem J 1989; 264:309-311. Tao BY, Reilly PJ, Robyt JF. Biochim Biophys Acta 1989; 995: 214-220. Holm L, Koivula AK, Lehtovaara PM, Hemminki A, Knowles JKC. Protein Eng 1990; 3: 181-191. Kuriki T, Takata H, Okada S, Imanaka T. J Bacteriol 1991; 173: 6147-6152. Nagashima T, Toda S, Kitamoto K, Gomi K, Kumagai C, Toda H. Biosci Biotech Biochem 1992; 56: 207-210. Nakamura A, Haga K, Ogawa S, Kuwano K, Kimura K, Yamane K. FEBS Lett 1992; 296: 37-40.

122 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56.

Takase K, Matsumoto T, Mizuno H, Yamane Y. Biochim Biophys Acta 1992; 1120: 281-288. Mathupala S, Lowe SE, Podkovyrov SM, Zeikus JG. J Biol Chem 1993; 268: 16332-16344. Podkovyrov SM, Zeikus JG. J Bacteriol 1992; 174: 5400-5405. Sogaard M, Kadziola A, Haser R, Svensson B. J Biol Chem 1993; 268: 2248022484. Nakamura A, Haga K, Yamane K. Biochemistry 1993; 32: 6624-6631. Feller G, Lonhienne T, Deroanne C, Libioulle C, Van Beeumen J, Gerday C. J Biol Chem 1992; 267: 5217-5221. Tomazic SJ, Klibanov AM. J Biol Chem 1988; 263: 3092-3096. Suzuki Y, Ito N, Yuuki T, Yamagata H, Udaka S. J Biol Chem 1989; 264: 18933-18938. Declerck N, Joyet P, Gaillardin C, Masson J-M. J Biol Chem 1990; 265: 15481-15488. Vihinen M, Ollikka P, Niskanen J, Meyer P, Suominen I, Karp M, Holm L, Knowles J, Mantsala P. J Biochem (Toyko) 1990; 107: 267-272. Brosnan MP, Kelly CT, Fogarty WM. Eur J Biochem 1992; 203: 225-231. Fujiwara S, Kakihara H, Sakaguchi K, Imanaka T. J Bacteriol 1992; 174: 7478-7481. Feller G, Payan F, Theys F, Qian M, Haser R, Gerday C. Eur J Biochem 1994; 222: 441-447. Watanabe K, Masuda T, Ohashi H, Mihara H, Suzuki Y. Eur J Biochem 1994; 226: 277-283. Conrad B, Hoang V, Polley A, Hofemeister J. Eur J Biochem 1995; 250: 481490. Mattsson P, Battchikova N, Sippola K, Korpela T. Biochim Biophys Acta 1995; 1247: 97-103. Yuuki T, Nomura T, Tezuka H, Tsuboi A, Yamagata H, Tsukagoshi N, Udaka S. J Biochem (Tokyo) 1985; 98: 1147-1156. Kim C-H, Kwon S-T, Taniguchi H, Lee D-S. Biochim Biophys Acta 1992; 1122: 243-250. Nakada T, Kubota M, Sakai S, Tsujisaka Y. Agric Biol Chem 1990: 54: 737743. Candussio A, Schmid G, Bock A. Eur J Biochem 1991; 199: 637-641. Vihinen M, Peltonen T, Iitia A, Suominen I, Mantsala, P. Protein Eng 1994; 7: 1255-1259. Kimura K, Kataoka S, Nakamura A, Takano T, Kobayashi S, Yamane K. Biochem Biophys Res Commun 1989; 161: 1273-1279. Schwermann B, Pfau K, Liliensiek B, Schleyer M, Fischer T, Bakker EP. Eur J Biochem 1994; 226: 981-991. Rodenburg KW, Juge N, Guo X-J, Sogaard M, Chaix J-C, Svensson B. Eur J Biochem 1994; 221: 277-284. Juge N, Rodenburg KW, Guo X-J, Chaix J-C, Svensson B. FEBS Lett 1995; 363: 299-303. Ishikawa K, Matsui I, Honda K, Nakatani H. Biochem Biophys Res Commun. 1992; 183: 286-291.

123 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89.

Kobayashi M, Miura M, Ichishima E. Biochem Biophys Res Commun 1992; 183: 321-326. Robyt JF, French D. J Biol Chem 1970; 245: 3917-3927. Suganuma T, Matsuno R, Ohnishi M, Hiromi K. J Biochem (Tokyo) 1978; 84: 293-316. Omichi K, Ikenaka T. J Biochem (Tokyo) 1988; 104: 881-883. Omichi K, Hase S, Ikenaka T. J Biochem (Tokyo) 1991" 109: 410-415. MacGregor AW, Morgan JE, MacGregor EA. Carbohydr Res 1992; 227" 301313. Ajandouz EH, Abe J-I, Svensson B, Marchis-Mouren G. Biochim Biophys Acta 1992; 1159: 193-202. Robyt JF, French D. Arch Biochem Biophys 1967" 122: 8-16. MacGregor EA, MacGregor AW. Carbohydr Res 1985" 142: 223-236. MacGregor EA, MacGregor AW, Macri LJ, Morgan JE. Carbohydr Res 1994; 257: 249-268. Sakano Y, Kashiwagi Y, Kobayashi T. Agric Biol Chem 1982" 46" 639-646. Outtrup H, Norman BE. Starke 1984; 36: 405-411. Candussio A, Schmid G, Bock A. Eur J Biochem 1990; 191: 177-185. Spreinat A, Antranikian G. Starke 1992; 44: 305-312. Ajandouz EH, Marchis-Mouren GJ. Carbohydr Res 1995; 268" 267-277. Robyt JF. Denpun Kagaku 1989; 36: 287-301. Robyt JF, French D. Arch Biochem Biophys 1970; 138" 662-670. Ishikawa K, Hirahata H. Arch Biochem Biophys 1989" 272: 356-363. Yamashita H, Nakatani H, Tonomura B. J Biochem (Tokyo) 1991" 110" 605607. Nakatani H, Hamaguchi K, Ishikawa K. Carbohydr Res 1994; 257" 155-161. Yamashita H, Nakatani H, Tonomura B. Biochim Biophys Acta 1993; 1202" 129-134. Matsui I, Ishikawa K, Miyairi S, Fukui S. Honda K. FEBS Lett 1992; 310: 216-218. Matsui I, Ishikawa K, Miyairi S, Fukui S, Honda K. Biochemistry 1992; 31" 5232-5236. Matsui I, Ishikawa K, Miyairi S, Fukui S, Honda K. Biochim Biophys Acta 1991; 1077: 416-419. Matsui I, Yoneda S, Ishikawa K, Miyairi S, Fukui S, Umeyama H, Honda K. Biochemistry 1994; 33: 451-458. Shibuya I, Tamura G, Shima H, Ishikawa T, Hara S. Biosci Biotech Biochem 1992; 56: 884-889. Bender H. Carbohydr Res 1990; 206: 257-267. Nakamura A, Haga K. Yamane K. FEBS Lett 1994; 337: 66-70. Abe S, Nagamine Y, Omichi K, Ikenaka T. J Biochem (Tokyo) 1991; 110: 756-761. Kitahata S, Hara K, Fujita K, Nakano H, Kuwahara N, Koizumi K. Biosci Biotech Biochem 1992; 56: 1386-1391. Vetter D, Thorn W, Brunner H, Konig WA. Carbohydr Res 1992" 223" 61-69. Janecek S, MacGregor EA, Svensson B. Biochem J 1995; 305: 685-686. Bender H. Carbohydr Res 1991; 209: 145-153.

124 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115.

Mattsson P, Pohjalainen T, Korpela T. Biochim Biophys Acta 1992; 1122: 3340. Nakamura A, Haga K, Yamane K. Biochemistry 1994; 33: 9929-9936. Penninga D, Strokopytov B, Rozeboom HJ, Lawson CL, Dijkstra BW, Bergsma J, Dijkhuizen L. Biochemistry 1995; 34: 3368-3376. Kaneko T, Kudo T, Horikoshi K. Agric Biol Chem 1990; 54: 197-201. Iwamoto H, Ohmori M, Ohno M, Hirose J, Hiromi K, Fukada H, Takahashi K, Hashimoto H. Sakai S. J Biochem (Tokyo) 1993; 113: 93-96. Iwamoto H, Ohno M, Ohmori M, Hirose J, Tanaka A, Sakai S, Hiromi K. J Biochem (Tokyo) 1994; 116: 1264-1268. Mizuno K, Kawasaki T, Shimada H, Satoh H, Kobayashi E, Okumura S, Arai Y, Baba T. J Biol Chem 1993; 268: 19084-19091. Takeda Y, Guan H-P, Preiss J. Carbohydr Res 1993; 240: 253-263. Burton RA, Bewley JD, Smith AM, Bhattacharyya MK, Tatge H, Ring S, Bull V, Hamilton WDO, Martin C. Plant J 1995; 7: 3-15. Imanaka T, Kuriki T. J Bacteriol 1989; 171: 369-374. Takahata H, Kuriki T, Okada S, Takesada Y, Iizuka M, Minamiura N, Imanaka T. J Biol Chem 1992; 267: 18447-18452. Mathupala S, Saha BC, Zeikus JG. Biochem Biophys Res Commun 1990; 166: 126-132. Sata H, Umeda M, Kim C-H, Taniguchi H, Maruyama Y. Biochim Biophys Acta 1989; 991: 388-394. Kim C-H, Taniguchi H. J Microbiol Methods 1994; 19: 39-49. Ara K, Igarashi K, Saeki K, Ito S. Biosci Biotech Biochem 1995; 59: 662-666. Kim C-H, Kim YS. Eur J Biochem 1995; 227: 687-693. Ramesh MV, Podkovyrov SM, Lowe SE, Zeikus JG. Applied Environ Microbiol 1994; 60: 94-101. Sakano Y, Sano M, Kobayashi T. Agric Biol Chem 1985; 49: 3041-3043. Suzuki Y, Imai T. Appl Microbiol Biotechnol 1985; 21: 20-26. Kim I-C, Cha J-H, Kim J-R, Jang S-Y, Seo B-C, Cheong T-K, Lee DS, Choi YD, Park K-H. J Biol Chem 1992; 267: 22108-22114. Krohn BM, Lindsay JA. Enzyme Microb Technol 1992; 14: 194-196. Podkovyrov SM, Burdette D, Zeikus JG. FEBS Lett 1993; 317: 259-262. Matuschek M, Burchhardt G, Sahm K, Bahl H. J Bacteriol 1994; 176: 32953302. Tonozuka T, Mogi S, Shimura Y, Ibuka A, Sakai H, Matsuzawa H, Sakano Y, Ohta T. Biochim Biophys Acta 1995; 1252: 35-42. Kim I-C, Yoo S-H, Lee S-J, Oh B-H, Kim J-W, Park K-H. Biosci Biotech Biochem 1994; 58: 416-418. Lee S-J, Yoo S-H, Kim M-J, Kim J-W, Seok H-M, Park K-H. Starke 1995; 47: 127-134.


125

Properties and uses of dextransucrases elaborated by a new class ofLeuconostoc mesenteroides mutants Doman ~1

and J o h n F. R o b y t 2

1Department of Biochemical Engineering, Chonnam National University, Kwangju, Korea ~Department of Biochemistry and Biophysics, Iowa State University, Ames, IA, USA

Abstract After chemical mutagenesis using ethyl methane sulfonate, we isolated Leuconostoc mutants constitutive for glucansucrases from several mesenteroides. The mutants produced glucansucrases when grown on D-glucose as well as sucrose. In addition to being constitutive for dextransucrase, each m u t a n t of B-742 produced a dextransucrase t h a t synthesized a dextran of different structure and solubility, and possibly produced only a single dextransucrase. B-512FMC dextransucrases produced in glucose medium was adsorbed on Sephadex G-100 and G-200, but much less enzyme was adsorbed when it was produced in sucrose medium. Sephadex adsorption decreased when the glucose- produced enzyme was preincubated with dextrans of molecular size greater t h a n 10 kDa. The addition of dextran to dextransucrase digests affects the initial velocity of dextran synthesis in a sigmoidal m a n n e r t h a t is the characteristic of an allosteric binding of dextran to the enzyme. The purified B512FMC dextransucrase had a molecular size of 184 kDa on SDS-PAGE. On standing at 4~ for 30 days, the native enzyme was dissociated into three inactive proteins. The native enzyme is believed to be a trimer of two 63 kDa and one 59 kDa monomers. A mixed culture fermentation system of Lipomyces starkeyi, a dextranase constitutive yeast, and Leuconostoc mesenteroides, a dextransucrase producing bacteria, was designed for the production of controlled size - limited dextrans. Introduction Dextran is a generic term given to a group of bacterial polysaccharides t h a t are synthesized from sucrose and composed of chains of D-glucose units connected by a - 1 ~ 2 , a - 1 ~ 3 , or a - 1-~4 linkages to varying degrees, depending on the strain of the organism. In the reaction of glucansucrases, sucrose serves as a high energy donor of glucose units. Sucrose has a glycosidic bond energy equivalent to t h a t of UDP- glucose ( 6 - 7 kcaYmole). Several different Leuconostoc mesenteroides strains and Streptococcus species elaborated glucansucrases t h a t synthesize glucans from sucrose [1]. The L. mesenteroides strains are inducible for glucansucrases and require sucrose in the growth media to elaborate the enzyme(s), whereas the Streptococcus species are constitutive

126 for glucansucrases and do not require sucrose in the growth media to elaborate the enzyme(s). During dextran elongation, the glucosyl unit is transferred to the reducing end of the growing dextran chain by an insertion mechanism (Figure 1). If other carbohydrates in addition to sucrose are present in the reaction digest, they m a y act as acceptors for the glucosyl and dextranosyl intermediates at the active site of the glucansucrases [2, 3] (Figure 2). In the useful conditions of action onto sucrose, dextransucrase yields a high molecular weight (107 - 10 s daltons) dextran polymers. However, the industrial applications of this polysaccharide in the analytical field (chromatography supports) or in the medical field (blood plasma substitute) require lower molecular weight polymers (40,000 or 70,000 daltons and even 5,000 for iron- or sulfate-dextran derivatives) [4]. Such molecules are obtained by acidic hydrolysis of high molecular weight dextran polymers, followed by alcohol-precipitation fractionation steps. On the other hand, the addition of efficient acceptors to the reaction medium results in the synthesis of low molecular weight oligosaccharides instead of high molecular weight dextran polysaccharides [2, 5]. Recently, a mixed culture fermentation system was designed for the production of size-limited dextran [6 - 9].

,x9 --X |

t

_ 2

[-- X-O

(~H2

~--X ~

- - X-O etc.

--------IP-

OH CH2

--X-O

- - •

Figure 1. Reducing end polymerization mechanism of dextransucrase. nucleophiles at the active site 0 - - 4 = sucrose, O = glucose, 4 = fructose

X=

127

~H2 ~H

X

(A)

G

r 0

I C~

IFe

~ ~ fA1A

L,WJb'~

(B) __X | r-1 i . r - l - .

,~, cw.ca--' LI" 4:>--

Figure 2. (A)" Mechanism of acceptor reactions, H = isomaltose, ~ = reducing glucose unit; (B)" mechanism for forming branch linkages in dextran. An exogenous dextran acts as an acceptor and displaces the glucosyl or dextranosyl unit forming an (~- 1 -~ 3 linkage. In this paper, we describe on the isolation of dextransucrase constitutive m u t a n t s of Leuconostoc mesenteroides to overcome some of difficulties relating with dextransucrase inducibility, on the characterizations of constitutive dextransucrases, and on the productions of oligosaccharides and controlled sizelimited dextrans.

128

P r o d u c t i o n and s e l e c t i o n of m u t a n t s c o n s t i t u t i v e for g l u c a n s u c r a s e s

of L e u c o n o s t o c m e s e n t e r o i d e s

L. mesenteroides NRRL B-512 elaborates a single dextransucrase, which has m a n y important industrial and medical uses [10-14]. It is also important because of its theoretical and practical aspects in understanding the mechanism of glucan synthesis [15-17] and in its ability to synthesize a wide variety of oligosaccharides by glucosyl transfer reactions to acceptors [5, 18]. L. mesenteroides B- 1355 produces two sucrose-inducible extracellular glucansucrases. One produces a dextran (or fraction L dextran) and the other produces a glucan referred to as an alternan [19]. AIternan has glucose residues linked alternately by a-1 - 6 and a-1 - 3 linkages. B-742 also produces two extracellular fraction L; contained 14% a-1 - 4 branch linkages and about 1% a-1 - 3 branch linkages and fraction S; contained 50% a - 1 - 3 branch linkages and no - 1 - 4 branch linkages [20]. Fraction L dextran was hydrolyzed by endodextranase, while fraction S was relatively resistant to endo-dextranase hydrolysi~ L. mesenteroides B-1299 is also known to produce two kinds of dextrans containing a - 1 - 2 branch linkages. Because of structural differences, alternan holds the possibility for commercial development as a highly soluble, non-caloric or low-caloric bulking agent t h a t could be used as a filler, binder, and extender in food preparations [21]. AIso, highly branched dextran might be chemically modified to give new gums, gels, or films that could have applications in food preparations, pharmaceuticals, and drug or pesticide delivery systems. Before dextran, alternan, a n d / o r other branched dextrans, however, can be marketed for food and pharmaceutical applications, technologies for m a s s producing t h e m m u s t developed; such as the isolation of hyper glucansucrase producing strains, the development of relatively simple purification methods, and the development of continuous processes for polysaccharide production using enzyme reactors. The previously reported glucansucrases of L. mesenteroides strains are all inducible and require sucrose in the culture medium for elaboration. The enzymes are secreted into the culture broth during growth, the presence of sucrose results in production of glucan in the culture broth and creates a problem in t h a t the glucansucrases are contaminated with relatively large amounts of glucan t h a t imparts a high viscosity to the culture supernatant. This limits the concentration of sucrose t h a t can be used in the medium for the production of glucansucrases to 2-3% (w/v). To purify the glucansucrases, a hydrolase t r e a t m e n t is needed to lower the viscosity and glucan content. Dextrans with a low degree of branching (5-10%), for example B-512F(MC) dextran, can be hydrolyzed by PeniciUium endo-dextranase, but the more highly branched dextrans, such as B- 1142, B-742, and B-1299 are resistant to endodextranase hydrolysis. The non-dextran glucans, such as B-1355 alternan and the a -1 - 3 linked glucans, are completely resistant to endo-dextranase hydrolysis. Hydrolases for these latter glucans have yet to be isolated and described. Heretofore, there have been no reports of L. mesenteroides strains that are constitutive for glucansucrases. In addition, B-742 and B-1299 strains produce mixture of two dextransucrases t h a t can not be separated easily. To overcome difficulties, using ethyl methane sulfonate as a mutagen, we have

129 isolated mutants from L. mesenteroides (B-512FM, B- 1142, B-1355, B-742 and B-1299) that produce glucans having different structures from each other [1, 10] and that have potential commercial uses [11]. We have designated these glucansucrase constitutive mutants as B-512FMC, B-1142C, B-1355C, B-742C, B-742CA, B-742CB, and B-1299C, B-1299CA, and B-1299CB, respectively (Table 1). Table 1. Comparison of glucansucrase activities of parents and mutants grown on glucose and sucrose. Glucansucrase activity (U/ml) Strains Sucrose (2 %) Glucose (2 %) 0.05 B-512F 0.00 3.20 B-512FMC 1.10 0.10 B-1355 0.00 B-1355C 0.34 a 2.52 ~ B-1142 0.00 0.04 B-1142C 0.12 1.28 B-742 0.00 0.04 B-742C 0.14 0.29 0.65 B-742CA 0.43 0.15 0.10 B-742CB aContains both glucansucrase activities, alternansucrase and dextransucrase, as judged by the nature of the glucans produced The mutants produced glucansucrase(s) on glucose with higher activities (from 3 to 22 times) than those produced on sucrose by the wild types. The mutants, however, produced less enzyme when grown on a glucose medium than they did when grown on a sucrose medium. This may indicate that some induction for glucansucrase may still be present in the mutants. There is also the possibility that a complex of glucan with the glucansucrase stabilizes the enzymes in a favorable conformation for the synthesis of glucan and thereby stimulates glucan synthesis, or that the glucan prevents subunit dissociation into less active or inactive subunits. Since the constitutive glucansucrase grown on glucose do not have glucan present, they may have less glucansucrase activity due to these possibilities. The glucans prepared from different mutants were hydrolyzed with Penicillium dextranase, and the hydrolyzate profiles were compared with those of the wild-type strains (Table 2 and Figure 3). The glucans produced by B1142C and B-1355C showed higher dextranase resistance than did the wild types. The glucose- and sucrose-grown B-1142C glucansucrases synthesized glucan that gave similar products from endo-dextranase hydrolysis, 25.9% and 28.3% branched dextrins 50.1% and 49.4% resistant glucan, respectively. This is in contrast to the B-1142 wild type that gave 26.8% branched dextrins and only 18.2% resistant glucan.

130 Table 2. Dextranase hydrolysis products of glucans prepared by different Leuconostoc mesenteroides under various conditions. CHO a % of total c a r b o h y d r a t e 1142 Sttcb

1142C Glc

1142C Suc

1135 Suc

1135C Glc

1135C Suc

512FM Suc

512FMC' Glc

512FM(: Suc

Glc

21.0 6.2 8.7 :1'3.2 6.2 19.7 3 7 1 8 16.0 30.7 34.0' 17.8 13.6 3-2.8 7.6 57.8 57'.5 80.0 61.7 6.6 8.6 9.3 8.0 0.5 1.4 3.0 1.1 1.4 6.8' 7.6 9.4 8.0 0.2 1.8 1.4 1.3 2.7 4.0 6.8 6.2 6.4 0.1 1.i 0.0 0.9 1.9 B. 4.3 2.9 3.4 5.8 0.0 0.8 0.0 0.4' 0.9 3.1 0.0 0.0 4.1 0.0 0.2 0.0 0.3 0.8 B7 B. 2.'0 0.0 0.0 2.1 0.0 0.0 0.0 0.0' 0.0 ORI c 18.2 50.1 49.4 19.6 85.4 17.2 0.0 0.0 0.0 a Carbohydrates separated on TLC and quantitatively determined by densitometry b Carbon source for growth c Dextranase-resistant polysaccharide remaining at the origin of the TLC Glc, Glucose; IM2, Isomaltose; B3-B4, Branched saccharides The glucans produced by the glucansucrases of glucose- and sucrose-grown B - 1 3 5 5 C gave much different distribution of products from dextranase hydrolysis. The glucose-grown B- 1355C glucan gave 0.8% branched dextrins and 85.4% resistant glucan, and the sucrose-grown B- 1355C glucans gave 5.3% branched dextrins and only 17.2% resistant glucan. The B-1355 wild-type glucans gave 34.4% branched dextrins and 19.6% resistant glucan. The polysaccharides at the origin is alternan. The dextrans prepared from the different B-742 m u t a n t s were hydrolyzed with Penicillium endo-dextranase, and the hydrolyzates were compared (Figure 3). Mutant B-742C, grown on glucose, produced a dextransucrase that synthesized a soluble dextran. This dextran was hydrolyzed by endo-dextranase to give glucose, isomaltose, a trace of branched tetra-saccharide, and branched penta-, hexa-, and hepta-saccharides. There was an appreciable amount of higher molecular weight glucan that was not converted into chromatographic products. Mutant B-742C, grown on sucrose, produced dextransucrase(s) that synthesized both a soluble and an insoluble dextran. The soluble dextran was hydrolyzed by endo-dextranase to give a slightly different series of products that included glucose, isomaltose, isomaltotriose, isomaltotetraose and branched tetra-, penta-, hexa-, and hepta-saccharides and appreciable amounts of higher branched dextrins and tmhydrolyzed dextran. The insoluble glucan gave only small amounts of glucose and isomaltose with appreciable amounts of glucan that was not hydrolyzed. Mutant B-742CA, grown on glucose or sucrose, produced a dextransucrase that only synthesized a soluble dextran. The two dextrans appeared to be identical as they were hydrolyzed by endo-dextranase to give identical products:

131

Figure 3. The dextranase hydrolysis products of dextrans (soluble and/or insoluble) prepared by different Leuconostoc mesenteroides B-742 m u t a n t s grown on glucose and on sucrose. Lanes 1- 3 were the dextranase hydrolysis products of dextrans prepared by B -742C dextransucrases; lanes 4 - 5 were the dextranase hydrolysis products of dextrans prepared by B-742CA dextransucrases; lanes 6 - 9 were the dextranase hydrolysis products of dextrans prepared by B-742CB dextransucrases. Glc and Suc indicate enzymes obtained from cultures grown on glucose and sucrose, respectively. S and I indicates soluble and insoluble dextrans, respectively. G: Glucose, 1M2 - 1M4: I s o m a l t o s e - Isomaltotetraose, B 4 - B7: Branched saccharides. glucose, isomaltose, and branched tetra-, penta-, and hexa-saccharides and appreciable amounts of unhydrolyzed glucan. These dextrans appeared similar to the dextran synthesized by the dextransucrase elaborated by m u t a n t B-742C grown on glucose. Because of its relative susceptibility to endo-dextranase hydrolysis, this dextran most probably corresponds to fraction L dextran produced by one of the dextransucrases elaborated by the parental strain. Mutant B-742CB, grown on glucose or sucrose, produced a dextransucrase t h a t synthesized both a soluble and an insoluble dextran. The two glucans (soluble and insoluble) synthesized by dextransucrase elaborated by growth on either glucose or sucrose appeared to have only slight susceptibility to endo-dextranase hydrolysis and appeared to be identical in structure. Further, in both cases, the insoluble dextran was rendered soluble by endo-dextranase treatment. This was

132 similar to endo-dextranase t r e a t m e n t of the insoluble dextran synthesized by the dextransucrase t h a t was elaborated by B-742C, grown on sucrose. Because of its relatively high resistance to endo-dextranase hydrolysis, this dextran m o s t probably corresponds to the highly branched, fraction S dextran produced by one of the two dextransucrases elaborated by the parental strain. M u t a n t B- 1299C dextransucrases produced slightly different dextrans when t h e y were elaborated on a glucose medium and on a sucrose medium. M u t a n t B - 1299CA dextransucrase elaborated on a glucose medium and on a sucrose medium synthesized the s a m e dextran, though the dextran was different from those of other m u t a n t s and the parental strain. M u t a n t B- 1299CB dextransucrase, elaborated on a glucose medium and on a sucrose medium, formed different dextrans. Differences in water solubility, susceptibility to endo-dextranase hydrolysis, and the physical a p p e a r a n c e of the ethanol precipitated dextrans elaborated by different m u t a n t s grown on glucose media and on sucrose media were found (Figure 4).

Figure 4. Differences in water solubility of ethanol precipitates from glucansucrase digests of Leuconostoc mesenteroides B-1299 m u t a n t s . G and S indicate dextrans precipitated from reaction digests of glucansucrases obtained from cultures grown on glucose and sucrose, respectively.

133

P r o p e r t i e s of Leuconostoc mesenteroides B-512FMC c o n s t i t u t i v e dextransucrase After SDS-PAGE of B-512FMC glucose-grown culture s u p e r n a t a n t (Figure 5), the gels were incubated with sucrose, followed by the periodic acid- Schiff stain. It showed one major activity band at 184 kDa and a minor band at 128 kDa. B-512FMC culture s u p e r n a t a n t grown on sucrose showed one major band at 184 kDa and no minor bands. B- 1142C culture supernatant grown on glucose (Figure 5) showed one major band at 173 kDa and a minor band at 240 kDa. The sucrose grown B-1142C culture supernatant contained an additional band at the origin. B-1355C grown on glucose medium (Figure 5) produced one band at 173 kDa. When grown on sucrose, there were additional activity bands at 184 kDa, 125 kDa, and the origin. This major 184 kDa activity band also exist for each of the dextransucrases from B-742 and B-1299 m u t a n t s as well as parental strains, eventhough each of the strains elaborated dextransucrase(s) t h a t synthesized different dextrans. Thus, it is possible t h a t all or m a n y of the dextransucrases elaborated by the various strains of L. mesenteroides dextransucrases have originated from the s a m e gene. During evolution of L. mesenteroides, various portions of the enzyme or its subunits have changed. This change, thus, affects the structure of the dextran(s) synthesized. The purified enzyme had a molecular size of 184 kDa on SDS-PAGE. On standing at 4~ for 30 days, the native enzyme was dissociated into three inactive proteins of 65, 62, and 57 kDa. however, two protein bands of 63 and 59 kDa were obtained on SDS-PAGE aider heat denaturation of the 184 kDa active enzyme at 100~ The amount of 63 kDa protein was about twice t h a t of 59 kDa protein. The native enzyme is believed to be a trimer of two 63 kDa and one 59 kDa monomers. L. mesenteroides B-512FMC was grown on glucose, fructose, or sucrose. The amount of cell-associated dextransucrase was about the s a m e for the same three sugars at different concentrations (0.6% and 3%; Table 3). Enzyme produced in glucose medium was adsorbed on Sephadex G- 100 and G-200, but much less enzyme was adsorbed when it was produced in sucrose medium (Table 4). Sephadex adsorption decreased when the glucose-produced enzyme was preincubated with dextrans of molecular size greater than 10 kDa (Table 5). The release of dextransucrase activity from Sephadex by buffer (20 mM acetate, pH 5.2) was the highest at 28~ - 30~ (Figure 6). The addition of dextran to the enzyme stimulated dextran synthesis but had very little effect on the temperature or pH stability.

M e c h a n i s m of d e x t r a n a c t i v a t i o n of d e x t r a n s u c r a s e In earlier work on the initiation of dextran synthesis, Hehre suggested t h a t the dextran associated with the enzyme could prime the polymerizing reaction [22]. Germaine et al. showed that the addition of dextran to Streptococcus mutans dextransucrase digests increased the rate of dextran [23 - 24]. They found t h a t the rate was dependent on the size of the dextran chain and reached a m a ~ m u m when the average size of the added dextran was 30 glucose residues. Kobayashi and Matsuda also reported t h a t purified dextransucrases elaborated

134 by both L. mesenteroides B-512F and Streptococcus sp. were stimulated by the addition of dextran, although both enzymes could synthesize dextran without the addition of dextran to the digests [25 - 26]. In contrast, Robyt et al., using pulse and chase techniques with [14C] sucrose and Bio-Gel P2-immobilized dextransucrase, showed that glucose and dextran were covalently attached to the enzyme during synthesis and that the glucose is added to the reducing end of the growing dextran chain by a two-site insertion mechanism (Figure 1).

Figure 5. SDS-PAGE of crude glucansucrases of Leuconostoc mesenteroides mutants. Enzyme activity was obtained by incubating the gels in 5% sucrose for approximately 18h. followed by periodic acid-Schiff stain. [I] (A) Enzyme from B-512FMC; (B) enzyme from B-1142C; (C) enzyme from B1355C. Ga, Gb, and Gc were these enzymes, respectively, obtained from cultures grown on glucose, and Sa, Sb, and S c were these enzymes, respectively, obtained from cultures grown on sucrose. [II] (A) Enzyme from B-742 parent strain; (B) enzyme from B-742C; (C) enzyme from B-742CA; (D) enzyme from B-742CB. G and S were enzymes obtained from cultures grown on glucose and sucrose, respectively.

135 In this mechanism, a primer is not necessary for the synthesis to occur. One of the reasons for the primer controversy was the uncertainty of complete dextran removal in purified Leuconostoc dextransucrase preparations. There was no carbohydrate found with the B-512FMC dextransucrase released from Sephadex gel using buffer, and this preparation could synthesize dextran [25 - 28. This purified enzyme preparation does not have any factors that can compromise the experimental results for the study of the enzyme reaction mechanism. Thus, it is clear t h a t the synthesis of dextran by L. mesenteroides B-512FMC dextransucrase does riot require a dextran primer. We also investigated the mechanism of dextran activation of dextransucrase by studying the initial velocity of the synthesis of dextran as a function of the concentration of added dextran, using B-512FMC dextransucrases [29]. The results of the experiments are shown in Figures 7 and 8 in which the initial velocities (~ mol of glucose incorporated into dextran/min) are plotted against the log of the concentration ( g/ml) of added dextran T-40. Table 3. Effect of carbon sources on the location of B-512FMC dextransucrase. Dextransucrase activity (U/ml) Extracellular Cell association Carbon source ratio of E/A (E) (A)a 0.61 Glucose-0.6% 0.14 0.23 Fructose-0.6% 0.79 0.15 0.19 Sucrose-0.6% 1.19 0.26 4.58 Glucose-3.0% 1.02 0.36 2.83 Fructose-3.0% 1.08 0.30 3.60 Sucrose-3.0% 3.02 0.61 4.95 a Cell-associated dextransucrase was determined by measuring the amount of ltC-dextran produced by a suspension of washed cells at an OD of 1.5 at 660 nm. Table 4. Adsorption of Leuconostoc mesenteroides B-512FMC dextransucrase produced on glucose or sucrose medium on Sephadex gels. Relative % of dextransucrase adsorbed a Sephadex b Glucose-grown Sucrose-grown r G-200 18 100 G-100 65 G-75 48 6 G-50 35 4 G-25 17 2 G15 12 1 a { (Dextransucrase activity bound per gram Sephadex) / (Glucose grown dextransucrase activity bound per gram Sephadex G-200)} x 100 b A column (15 x 100 mm) was used c Supernatant from 2% (wlv) sucrose culture medium contained 0.1% (w/v) dextran

136 Table 5. Influence of various saccharides on the adsorption of B-512FMC constitutive dextransucrase on Sephadex G-200. . Saccharides Amount (%) of enz~rme not adsorbed None 0 T1 3 T10 52 T40 76 T70 88 T500 89 T2000 97 Industrial-6rade dextran 67 Alternan 42 Maltose 1 Glucose 2 Fructose 1 ,,,

,,

One milliliter of enzyme (10 IU ml a) solution obtained from a glucose medium was preincubated for l h with 10 mg of each of the carbohydrates, then placed onto Sephadex G-200 column (15 x 100 nm).

100 ~9

80 =

60

x

40

~-

20

~ w,,,4

~

0

"

I

'

I

'

I

10 20 30 Incubation T e m p e r a t u r e (~

'

40

Figure 6. Effect of temperature on the release of dextransucrase from Sephadex G-200. Dextransucrase-Sephadex G-200 complex was incubated with 20 mM acetate buffer (pH 5.2) for 2h at different temperatures. The Sephadex was removed by filtration through glass fiber filters and the enzyme was assayed.

137 The L. mesenteroides B-512FMC dextransucrase reaction gave sigmoidal curves as a function of the concentration of added dextran. Figure 7 gives the results of a digest containing 20mIU of enzyme. It had an initial velocity lag between 1.25 and 3.12 ~ g/ml of exogenous dextran. Thereai~r, the initial velocity increased up

to an exogenous dextran concentration of 40 ~ / m l , where a constant, maximum velocity was attained. In this digest, the initial velocities decreased slightly with an additional increase in the concentration of exogenous dextran. Figure 8 shows a second L. mesenteroides B-512FMC dextransucrase reaction in which there was 3.9 times more enzyme t h a n in the digest of Figure 7. This reaction was similar to the other experiment, giving a sigmoidal curve. It had an initial velocity lag between 1.25 and 1.52 ~ / m l of exogenous dextran, diminished from the size of the lag observed in the experiment of Figure 7. The digests with 2.5 and 10.0 ~g/ml exogenous dextran had significantly higher initial velocities than the equivalent digests of Figure 7. The initial velocity of Figure 8 maximized at 40 ~,/ml of exogenous dextran as it also did for the experiment of Figure 7. The initial activity of the L. mesenteroides B-512FMC dextransucrase was proportional to the amount of enzyme. The amount of enzyme used in the experiment of Figure 8 was 3.75-times the amount of enzyme used for the experiment of Figure 7. The initial activity, without exogenous dextran, of Figure 8 was exactly 3.75-times the initial activity of Figure 7. The plots of Figures 7 and 8 are Michaelis- Menten type-plots t h a t usually hyperbolic curves. The sigmoidal curves obtained are not t h a t of usual Michaelis-Menten kinetics. Sigmoidal curves are particularly characteristic of a cooperative binding of ligands (in this case dextran) to enzymes t h a t have multiple binding sites. The sigmoidal increase in the reaction rate with increasing ligand concentration can be interpreted as due to the binding of the ligand at a noncatalytic site or allosteric site to give a more active enzyme. This positive allosteric effect olden involves the joining together of subunits. This is the very likely the mechanism of exogenous dextran activation of dextransucrases. It was found that S. m u t a n s dextransucrase has a dextran-binding domain t h a t is distinct and some distance from the active-site d o m a i n [ 3 0 - 31. Further, both the S. m u t a n s dextransucrases and the L. mesenteroides B-512FMC dextransucrases bind to lightly crossed-linked dextran (Sephadex G-200) [26]. The binding was inhibited when exogenous dextran is present. This noncatalytic, dextran-binding domain, is very likely an allosteric site to which the dextran binds and induces a favorable conformation for the synthesis of dextran from sucrose. Further, there is evidence t h a t the favorable conformation involves the formation of multiple subunit complexes. As indicated previously the molecular weights of the dextransucrases have ranged from 184 kDa to 64 kDa. Very high-molecularweight aggregates also have been observed, especially when the enzymes are elaborated in a sucrose medium where dextran is synthesized [25, 28. 'This indicates t h a t the presence of dextran in relatively large amounts, with a high molecular weight, acts as a crosslinking matrix t h a t can join m a n y enzyme subunits together. The decrease in the a m o u n t of lag from the experiment of Figure 7 to the experiment of Figure 8, when the amount of enzyme was increased 3.75-fold, also indicates t h a t the activation of dextransucrase is by the association of subunits induced by the binding of dextran. As the concentration of enzyme is

138 220 I-'I

210

.-4

200 o 190

r,.)

180

r~

o

'

170

~,~

' 160

~;-

150

......

L

9

0-=

-

!

0.0

,

!

0.4

]

,

0.8

,

1.2

I

,

1.6

Log [ d e x t r a n

!

,

I

2.0

2.4

T-40 /zg/mL]

Figure 7. Activation of dextransucrase by native dextran. The enzyme solution was incubated with sucrose and samples were taken at different times, giving various concentrations of dextran. Dextransucrase activity was determined on these enzyme solutions by the 14C sucrose radioactivity assay and by the determination of the amount of fructose released by TLC analysis. 800 .el

//

78O 760

740 0

720

7OO r._9 6130 r/l

(D 0

660

640

:t 620 600 580

1

0.0

~

1

0.4

,

~

1

,

0.8

Log [ d e x t r a n

I

=

1.2

[

1.6

_,

1

2.0

|

_

]

2.4

T-40 #g/mL]

Figure 8. Effect of exogenous dextran (T10) addition on pH. Relative dextransucrase activities with exogenous dextran addition. Dextransucrase activity was assayed using 14C-sucrose radioactive assay procedure.

139 increased per unit volume, the concentration of inactive or less active subunits is increased, giving a higher probability of association. Thus, a lower concentration of dextran is required to induce the association into active units and give activation. The maximum degree of activation, however, remains the same as the relative proportion of subunits that can associate remains the same.

Novel p r o c e s s for the production of oligosaccharides and size-limited dextrans Oligoaccharides are the subject of an increasing number of research programs due to both their technological and physiological properties. The technological properties of oligosaccharides have been stressed in the food application field by the introduction of intense sweeteners and the need for stabilizers and/or bulking agents. Their viscosity and water-retention properties are keys for the development of new low-calorie diets. The physiological properties of oligosaccharides result from their physico-chemical and their specific biological characteristics [ 3 2 - 33]. The involvement of these oligosaccharides in specific recognition mechanisms (immunology, cell adhesion etc.) has been underestimated for a while, but is of great interest for both human and animal applications [34]. Low molecular weight dextrans have their greatest application in the pharmaceutical industry. Dextran 70 (average MW 70,000) and dextran 40 (average MW 40,000) have been used as blood plasma extenders and blood flow improvers, respectively. This size range of dextran (MW 75,000 + 25,000) is classed as clinical dextran. Anticoagulant activity with low toxicity has been obtained using the dextran derivative, dextran sulfate (MW 7,300) which also interacts specifically with the major envelope glycoprotein of H1V- 1, inhibiting its ability to bind to CD4 § cells [13, 35 - 37]. In addition, crosslinked dextran, mercaptodextran, and iron dextran have been used for various applications. An important factor in these applications is the size of the dextran derivative. In biological systems, synthesis reactions generally involve cofactors (NAD(P)H, ATP etc.) which are consumed on a stoichiometric basis during the reaction. This raises the problem of cofactor regeneration, which up to now limits the industrial application of this type of enzymatic synthesis reactions. Recent process in enzymology have demonstrated that the alternative potential use of hydrolases in the catalysis of reverse reactions for the synthesis of glucooligosaccharides and galactooligosaccharides [ 3 8 - 39]. However, the reaction yields and the limited specificity of most hydrolases in reverse reactions limit the application of this technology. Another enzymatic approach for oligosaccharide production consists of the use of transfer reactions by transferases. Industrial enzymatic processes are used to synthesize cyclodextrins [40- 41] and fructooligosaccharides [42]. Glucooligosaccharides can be synthesized using glucosyltransferases (E.C. 2.4.1.5). These enzymes catalyze the transfer of glucose units from sucrose onto acceptor molecules, mainly sugars, resulting in the synthesis of glucooligosaccharides (Figure 2). In the absence of acceptors the main product of the reaction is dextran. The chemical structure of dextran was shown to be highly dependent on the glucosyl-

140 transferase producing strain. Different carbohydrate structures will bind at the acceptor binding-site with different affinities and therefore have different acceptor efficiencies[5]. There are reports about the production of glucooligosaccharides via acceptor reactions using the glucosyltransferases from L. mesenteroides strains [2, 3, 22]. The synthesis of oligosaccharides (or acceptor products) and controlled molecular weight dextran by the acceptor reaction using dextransucrase from L. mesenteroides NRRL B-512F(M) have been described [43- 44] and the enzymatically synthesized dextrans were shown to have less branched linkages than the dextran fractions produced by acid hydrolysis [ 4 4 - 45]. These glucosyltransferases were shown to display a specificity in the acceptor reaction close to that observed in high-molecular weight glucan synthesis. The efficiency of the acceptor reactions are dependent on the ratio of the concentrations of the acceptor to sucrose, the absolute concentration of acceptor and sucrose (Figure 9). In general, when the concentration of the acceptor is higher than that of sucrose, the synthesis of the acceptor-product is favored over the synthesis of glucan. Further, with acceptors that give multiple acceptor products, the number of acceptor products is decreased as the ratio of the concentration of acceptor to sucrose is increased. The formation of acceptor-products are also favored when the concentrations of both acceptor and sucrose are relatively high [46]. Commercially, low-molecular weight dextran is produced by whole-culture fermentation of L. mesenteroides NRRL B-512(F), followed by controlled acid hydrolysis and organic solvent fractionation. Yields are relatively low due to losses during hydrolysis and fractionation. There have been several studies on ways to improve clinical dextran production using purified dextransucrase [44 45]. A concept for producing clinical dextran has been to use a maltose acceptor in conjunction with dextransucrase [44]. An enzymatic method replace acid hydrolysis for clinical dextran production has also been patented, but neither of these methods is used commercially [47 - 48]. We developed a new, simpler, and industrially practical method for producing size-limited dextrans. It is the mixed culture fermentation system that required the coproduction of two microorganisms, an ascosporogenous yeast Lipomyces starkeyi (dextranase constitutive mutant assigned as ATCC 74054) and L. mesenteroides NRRL B-512 [6, 7, 9, 14]. Acceptable conditions for cell growth, enzyme production, and enzyme reactions for both species established. Both enzyme have similar pH optima for activity and stability, between 5.2 and 5.5. A pH of 5.2 (• 0.1) was found to be optimum for product formation. The optimum temperature for maximum dextran yield was 28 (+ 0.5)~ The highest yield of clinical dextran, 70% of theoretical yield, was obtained in mixed-culture fermentation with 0.9% yeast extract. A single addition of sucrose (15%) produced more dextran than discontinuous addition, but clinical dextran yields were higher with discontinuous addition (Table 6). Clinical dextran formation in the mixed culture system appears to be the result of both dextranase action and acceptor (dextranase hydrolyzates) reactions. It is logical that the balance of sucrose to acceptor would be important. A ratio of less than 4 to 1 sucrose to acceptor in mixed culture resulted in high levels of clinical dextran. Practically, this means that the sucrose concentration at any given period should be less than 4 times the available acceptor concentration. Operationally there was less than 10%

141

Figure 9. Thin-layer chromatographic diagram of C-sucrose acceptor reactions using different ratios of maltose to sucrose at 100mM constant total carbohydrate and 120 m I U of B-742C prepared from glucose c u l t u r e / 100~LL. The chromatography was conducted on Whatman K5F plates using 3 ascents of 2:5:3 nitromethane - 1-propanol - water. Pn is a saccharide acceptor-product of dp=n, having an isomaltodextrin chain linked a - l ~ 6 to the non-reducing end of maltose; Frc = D-fructose, Mal = maltose, P = panose, IMn = isomaltooligosaccharide standard. (w/v) sucrose in the fermentor at any given time. Several factors are known to affect dextran yields in L. mesenteroides fermentation [49]. These are requirements for s u b s t r a t e dispersion, dextransucrase activity maintenance, viscosity control, and by-product (i.e. fructose) levels. Fructose is a poor acceptor for dextran formation [5]. When fructose concentrations are high, glucose is transferred from sucrose to fructose molecules with the formation of leucrose. Leucrose is not an acceptor for dextransucrase and decreases dextran yield. In mixed culture fermentation, L. starkeyi used fructose as a carbon source, producing dextranase and decreasing leucrose formation. In addition, concentrations of isomaltose and other oligosaccharides (dextran hydrolysis products by dextranase) were higher,

142 Table 6. Effect of discontinuous and single addition of dextran l~roductivit~,. Sucrose concentration 30% 15% Discontinuous Single Sin61e Discontinuous 0.41(_+0.01) 0.41(_+0.005) 0.40(_+0.005) Total Dextran a 0.35(+0.01) b Clinical 94 73 94 79 Dextran (%)c 22.3(+1.3) 100.0(-+2.2) 50.8(+1.6) 306(_+15.6) Lm/Ls Ratio d Total dextran yield (Yp/s: kg d e x t r a n / kg sucrose); average value of total dextran production (100% of the theoretical yield of dextran is Y p/s = 0.48), standard deviation (+0.01) and (< _+0.01) for pure and mixed culture, respectively; values are averages from results of gel permeation chromatography, phenolsulfuric acid, and saccharimeter assays. b Standard deviation (n=4). CClinical dextran (MW 75,000 + 25,000) % portion of the total dextran produced (_+1.5). d Cell number ratio of L. mesenteroides to L. starkeyi.

a

displacing fructose as an acceptor. These oligosaccharides were incorporated into dextran and showed up as increased amounts of chnical dextran. Mixed culture fermentation maintained lower viscosities in the fermentation broth (less than 50 times as viscous as pure culture broths) [8]. Our process for the production of controlled-size dextrans differs in several ways from the traditional process. This process requires co-inoculation and establishment of a co-culture. Because of higher sucrose concentrations (over 30%), discontinuous sucrose addition is required. One of the most expensive steps in the traditional process, resulting in the highest dextran loss, is acid hydrolysis. The new process does not require acid hydrolysis, and because of this, deionization is also not required. The size-limited dextran produced in mixed culture broth has a small polydispersity index. The dextran size can be readily controlled simply by stopping the fermentation at the desired size. Since this process produces small-size dextrans which have lower viscosity, cell removal is easier. In addition, using different L. mesenteroides strains producing a dextransucrase, such as B-742CA or B-742CB, the new process can produce new structure low-molecular weight dextran t h a t has different physico-chemical and biological characteristics.

Acknowledgments This work was partially supported by the research grant of K O S E F for Research Center for New Bio-Materials in Agriculture, Seoul National University.

143

References

4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38

Jeanes, AWC, Haynes C, Wilham A, Rankin JC, et al. J Am Chem Soc 1954; 76: 5041- 5052. Koepsell HJ, Tsuchiya HM, Hellman NN, Kazenko A, et al. J Biol Chem 1953; 200: 793- 801. Tsuchiya HM, Hellcican NN, Koepsell HJ, Corman J, et al. J Am Chem Soc 1955; 77: 2412- 2419. Gronwall, AJT. New York: Academic Press 1957; 159 Robyt JF, Eklund SH. Carbohydr Res 1983; 121: 279- 286. Kim D, Day DF. Enzyme Microbial Technol 1994; 16:844 - 848. Day DF, Kim D. US Pat 1993; 5,229,227. Kim D. Ph.D. Thesis, Louisiana State Univ 1992 Kim D, Day DF. Lett Appl Microbiol 1995; 20:268 - 270. Gronwell, AJT, Ingleman BGA. Acta Physiol Scand 1944; 7:97 - 107. Gronwell, AJT, Ingelcican BGA. Acta Physiol Scand 1945; 9:1 - 27. Ricketts CR, Lorenz L, Maycock DA. Nature 1950; 165:770 - 771. Mbemda E, Cham V, Gluckman JC, Kiatzmann D, et al. Biochem Biophy Acta 1992; 1138: 6 2 - 67. Day DF, Kim D. Ann NY Acad 1992; 672:573 - 576. Ebert KH, Schenk G. Z Naturforsch 1962; 17:732 - 741. Robyt JF, Taniguchi H. Arch Biochem Biophysics 1976; 174:129 - 135. Robyt JF, Kimble BK, Walseth TF. Arch Biochem Biophysics 1974; 165: 634 - 640. Robyt JF, Walseth TE. Carbohydr Res 1978; 61: 433- 445. Cote GL, Robyt JF. Carbohydr Res 1982; 101:57 - 74. Seymour FR, Chen ECM, Bishop SH. Carbohydr Res 1979; 68:113 - 121. Hardin B. Agr Res 1991; 39: 27. Hehre EJ. Adv Enzymol 1951; 11:297 - 337. Germaine GR, Chludzinski AM, Schachtele CF. J Bacteriol 1974; 120: 287- 294. Germaine, GR, Schachtele CF. Infect Immun 1976; 13:365 - 372. Kobayashi M, Matsuda K. Biochem Biophysics Acta 1980; 614:46 - 62. Kobayashi M, Matsuda K. J Biochem 1986; 100:615 - 621. Kim D, Robyt JF. Enzyme Microbial Technol 1994; 16:659 - 664. Kim D, Robyt JF. Enzyme Microbial Technol 1994; 16:1010 -1015. Kim D, Robyt JF. Enzyme Microbial Technol 1995; 17:689 - 695. Kim D, Robyt JF. Enzyme Microbial Technol (in press). Robyt JF, Kim D, Yu L. Carbohydr Res 1995; 266:293 - 299. Mooser G, Wong C. Inf Immun 1988; 56:880 - 884. Wong C, Hei~a SA, Paxton RJ, Shively JE, et al. Infect Immun 1990; 58: 2165- 2170. Monsan P, Paul F, Remaud M, Lopez-Munguia A. Food Biotechnol 1989; 3 : 1 1 1 - 130. Hidaka H, Hirayama M, Tokunaga T, Eida T. In: Furda I, Brine CJ, eds. New developments in dietary fiber. New York: Plem!m, 1990; 105 - 117 Glaser V. Gen Eng News 1994; July: 6 - 7, 11. Ricketts CR. Biochem J 1959; 51:129-133. Takemoto K ~ Liebharber H. Virology 1962; 17: 499- 501.

144 39 40 41 42 43 44 45 46 47 48 49 50 51

Mitsuya H, Looney D, Kuno S, Ueno R, et al. Science 1988; 240:646 649. Ajisaka K, Nishida H, Fujimoto H. Biotechnol Lett 1987; 9:243 - 248. Ajisaka K, Nishida H, Fujimoto H. Biotechnol Lett 1987; 9:387 - 392. Kitahata S, Tsuyama N, Okada S. Agri Biol Chem 1974; 38:387 - 393. Kitahata S, Okada S. J Jap Soc Starch Sci 1982; 29:7 -12. Tanaka T, Yacicamoto S, Oi S. J Biochem (Tokyo) 1981; 90:521 - 526. Lopez-Munguia A, Paul F, Monsan P, Biton J, et al. Ann NY Acad Sci 1990; 613: 717- 722. Remaud M, Paul F, Monsan P, Heyraud A, et al. J Carbohydr Chem 1991; 10:861 - 876. Paul F, Oriol E, Auriol E, Monsan P. Carbohydr Res 1986; 149:433 - 441. Su D, Robyt JF. Carbohydr Res 1993; 248:339 - 348. Novak LJ, Stoycos GS. US Pat 1958; July: 2,841,578. Novak LJ, Wiff EC. US Pat 1961; February: 2,972,567. Alsop RM. Prog Ind Microbiol 1983; 18:1 - 44.


145

Molecular determinants of thermozyme activity and stability" Analysis of xylose isomerase and amylopullulanase J. Gregory Zeikus Department of Biochemistry, Michigan State University East Lansing, Michigan 48824

Introduction

Generalizations on Thermophiles Thermophilic organisms that grow at temperatures as high as 60~ were known since the late 1800's; and, were limited to spore-forming aerobes such as Bacillus stearothermophilus and anaerobes such as Clostridium These organisms were not thought to possess thermosaccharolyticum. inherently stable enzymes and were assumed to have evolved from mesophiles. More recently, microbes that grow above 70~ such as Thermus aquaticus and Methanobacterium thermoautotrophicum, a member of the archae, have been isolated which possess inherently thermal stable enzymes. Furthermore, hyperthermophiles have now been isolated such as Thermotoga neopolitana and Pyrococcus furiosus which grow from 80~ to above 100~ and, these microbes are thought to be among the first forms of life to have evolved on earth.

Generalizations on Thermozymes Enzymes from thermophiles and hypothermophiles (i.e., thermozymes) are inherently thermostable (70~ to 110~ and display a high temperature for irreversible protein denaturation. Thermozymes are also inherently thermophilic and display a high temperature for maximal activity (70~ to 110~ Thermophilicity corresponds to optimal enzyme flexibility that results from molecular movement of the protein at high temperature.

Thermozymes in Biotechnology The major advantages and disadvantages of thermozymes in biotechnology have been reviewed elsewhere (1). Thermozymes can display the following advantages and disadvantages in bioprocessing; Advantages A. Higher reaction rate due to the Q10 relation. B. Increased product formation. C. Lower processing costs. D. Higher physiochemical stability and resistance to denaturants. E. Limited applications (they will not work at low temperatures so are not of medical importance).

146 Disadvantages A. Unwanted chemical by-product formation can occur at high temperatures. B. Some co-factors, substrates and products of enzymes are unstable at high temperatures. Many industrial (e.g., starch processing) and specialty (e.g., PCR) processes require high temperature enzymes for utility and represent application opportunities for thermozymes. The developmental approach employed by my lab for thermozymes involves the following: A. Screen for target enzyme in thermophilic or hyperthermophilic source. B. Clone the thermogene in a mesophilic microbe. C. Perform site-directed mutagenesis to enhance activity in relation to pH and substrate specificity-activity. D. Over-express the engineered gene in a mesophilic industrial host. E. Purify the thermozyme by heat treatment in the presence of high solvent or other chemical denaturants. We have used this approach (See Table 1) for developing saccharidase thermozymes that will expand the utility of these enzymes in starch processing, sweetener production, baking, animal nutrition, waste treatment and fermentation (2-6). Table 1 Thermozyme saccharidase biotechnology Applications: Starch processing Baking and food processing Brewing and fermentation Pulp and Paper Detergents Specialty chemicals Waste treatment Animal Feeds Major Advantages:

Approaches:

Higher reaction rates (Q10) Higher stability Increased product formation Lower processing costs Screen for new thermophiles and hyperthermophiles Clone and overexpress thermogenes in industrial hosts Site-directed mutagenesis to enhance activity Purify product by heat-chemical treatment

147

Scope My lab has worked on thermozymes for the last twenty-five years. This research was aided by some outstanding Korean doctoral and post-doctoral students. These include: Dr. H. H. Hyun, who discovered the first Bamylase and amylopullulanase thermozymes (7-8); Dr. C. Lee, who discovered the first xylose isomerase thermozymes (9-10); Dr. Y.-O. Lee, who discovered the first endoxylanase and ~-xylosidase thermozymes (11-14); and, Dr. J. H. Park, who discovered the first thermostability and thermophilicity domains in amylopullulanase thermozymes (15-16). This paper reviews only our recent work on thermozymes used for processing starch and fructose sweetener production. Figure 1 shows the biochemical process for conversion of starch into fructose. Starch is first hydrolyzed into maltodextrins by a-amylase at 95~ at pH 6.5 in the presence of calcium. This D.E.10 syrup is then converted to glucose by glucoamylase at 55~ and pH 4.5. Glucose is then isomerized by xylose (glucose) isomerase at 58~ and pH 7.0. Industry targeted the need to improve the process by identifying three new enzymes: acid thermostable a-amylase; acid thermostable pullulanase; and, extremely thermostable glucose isomerase. We developed our thermozyme research program around these targets. Furthermore, we have also developed xylose isomerase and amylopullulanase as model thermozymes to understand the fundamental molecular determinants that account for thermozyme" activity and stability at high temperatures.

Amylopullulanase Thermozymes Biotechnological Applications The starch processing industry needed a more acid stable, non-calcium requiring thermostable ~-amylase to improve the starch hydrolysis reaction. Calcium must be removed in the current process, and a lower pH active enzyme would also enhance the stability of the process. An acid stable thermostable pullulanase would allow the starch branch points to be hydrolyzed and would improve the overall glucose yield. We discovered amylopullulanase that was thermal stable and cleaved both ~-1-4 and ~-1-6 bonds in starch (8,15,17-18). Amylopullulanase has now been found in a wide variety of thermophiles and hyperthermophiles (1). Amylopullulanase, however, saccharifies starch directly into maltotetraose, maltotriose and maltobiose. Thus, it is unsuitable for starch solubilization into D.E.10 maltodextrin syrups. However, amylopullulanase can serve as a model for further genetic engineering studies by site-directed mutagenesis to produce a factitious thermozyme that will just hydrolyze ~-1-6 bonds. The enzyme as it is found in nature is suitable for direct production of fermentation-conversion syrups in a single step process from starch (2).

148

General Properties Table different

2 compares the general properties of amylopullulanase from thermophilic sources. The enzyme from Thermoanaerobacter ethanolicus 39E (formerly, Clostridium thermohydrosulfuricum) has become a model enzyme for detailed investigations of structure-function. This amylopullulanase has a molecular weight of about 140,000. Amylopullulanase is optimally active and stable at pH 5.0. The enzyme converts pullulan into maltotriose; whereas, it cleaves starch into maltotetraose, maltotriose and maltobiose.

Process

Stage

Enzyme

pH

Temperature

Metal ions

Liquefaction

Bacterial a-Amylase

6.0-7.0

80-120~

Ca++

Saccharification

Glucoamylase

4.0-5.0

55-60~

Isomerization

Glucose

7.0

58-60~

l Starch |

Malto Dextrin (D.E.10 - 15) (pH Adjustment with Acid)

1

I Glucose 1 Filter, pH Adjustment, Addition of Metal Cofactor Mg ++ Mn ++ C o ++

Glucose and Fructose Mixture (58:42)

Figure 1. Schematic process biochemistry of enzymatic fructose production from starch. Reprinted with permission from Ref. 1. Copyright 1990 Elsevier.

149 Table 2 Comparison of some different thermophiles Property

biochemical

properties

of

amylopullulanase

Microbial Sources 1

2

136.5

3

5.0 75 70 4.5

4

5

450

105

Mol. weight ( x 104) Optimum pH

5.0-5.5

370+85 330+85 5.6

pH stability Optimum temp.(~ Thermal stability Isoelectric point

3.5-5.0 90 90 5.9

4.5-5.0 85-90 65 4.25

maltotriose DP2-DP4 -

same

same DP2-DP4

same

Ca

-

Ca

+

+

Product Pullulan Starch Activator Inhibitor Cyclodextrin (B-and 7-) 1, 2, 3, 4, 5,

from

120

5.5

same DP2-DP4

C. thermohydrosulfuricum strain 39E; C. thermohydrosulfuricum strain E101; Thermoanaerobacter strain B6A; Thermoanaerobium brockii; Thermoanaerobium strain Tok6-B1.

Genetics We have cloned the amylopullulanase gene (APU) from Thermo anaerobacter ethanolicus 39E and have characterized its sequence (19). The structural gene contains a single open reading frame with 4443 base pairs with an estimated molecular weight of 162, 780. Analysis of the deduced amino acid sequence of APU with sequences of a-amylase and a-1-6 debranching enzymes enabled the identification of four conserved regions putatively involved in substrate binding and catalysis (See Figure 2). Alignment of the T. ethanolicus amylopullulanase sequence with those of pullulanase, a-amylase, and glucoamylase showed 82% similarity to T. thermohydrosulfuricum amylopullulanase, 48% to A. oryzae cFamylase, and The overall similarity to AspergiUus 44% to Klebsiella pullulanase. glucoamylase, however, was only 40%. Catalytic Mechanism Amylopullulanase contains a single active site which cleaves both cF1-6 and a-1-4 bonds in starch. Figure 3 compares hydrophobic cluster analysis (HCA) plots of the amylopullulanase active site with Taka cFamylase and neopullulanase which converts pullulan to pannose. These enzymes and all other enzymes of the a-amylase family contain a catalytic triad comprised of

150 two aspartic acids and one glutamic acid. When any single amino acid of this catalytic triad was changed to their amine forms the enzyme loses both a-amylase activity and pullulanase activity. The residues constituting the catalytic triad of T. ethanolicus amylopullulanase, Asp 597, Glu 6~6 and Asp 7~ are located in close proximity.

0

9

t

..

200 i

I

400 I

1

600 1

1

800 I

!

1000 1

!

1200 I

t

1400 I

e.e.

39E amy 1opullulanase E101 c-am 91ase-

-pullulanase

TAA B.am 9

~-amvlase

m-amylase K.ae pullulanase K.pn

pullulanase

Figure 2. Overall alignment of the deduced sequence of amylopullulanase of T. ethanolicus 39E with amylases from microbial and fungal origin. 39E, T. ethanolicus 39E; E101, C. thermohydrosulfuricum E 101; TAA, A. oryzae, B. amy, Bacillus amyloliquefaciens; K. ae, Klebsiella aerogenese; K. pn, K. pneumoniae. The open boxes represent regions putatively identified on all sequences based on the four conserved regions of a-amylase of A. oryzae.

Thermal Properties The thermal features of T. ethanolicus 39E amylopullulanase are shown in figure 4. This enzyme is optimally active and stable at 90~ Calcium is not required for activity which makes it unlike most ~-amylases characterized. The amylopullulanase from T. saccharolyticum is optimally active and stable at 70~ Amylopullulanases characterized from certain hyperthermophiles are active and stable at 110~ Protein Engineering We have initiated protein engineering studies on amylopullulanase thermozymes in order to identify molecular determinants for thermal stability and thermophilicity (16). By generating a series of N- and C-terminus deletions, we have been able to show: an N-terminus region separate from the catalytic domain that confers thermophilicity; and, a C-terminus region that confers thermostability. The native gene expresses a protein in E. coli with a half-life of 50 min at 85~ an optimum temperature for activity of 80~ and a specific activity of 440 ~/mg. Deletion m u t a n t enzyme APU N324 retained stability and activity but,

151 unlike the wild type recombinant enzyme, it was optimally active at a very broad range of 55 to 80~ Deletion mutant APU N106/C379 retained the same specific activity and thermophilicity of the wild-type enzyme, but the enzyme half-life decreased to 5 min at 850C. Analysis of HCA plots of the N-terminus and C-terminus regions o~ amylopullulanase from Thermoanerobacter 39E enzyme (active and stable at 90~ versus Thermoanaerobacterium B6A-RI enzyme (active and stable at 70~ indicate the presence of site-specific proline residues in loop regions o~ the enzyme with higher thermostability. Prolines in enzyme loop regions have been shown to thermostabilize certain enzymes by enhancing amino acid interactions associated with the constraint caused by proline (1). I

I

I

I

I

[

I

I

I',

I

I:

I:

[

I;

I

I '

I

1

I

II

I

I

I

I

I

39E

,

| | -

[

|

1

#

1

I

I

I

I

I

,

I

1

1

x

'

!'

'I

~

r. s

['

[

.J

I

..

,,

I

I

I

',I

1

I

I

I

I

1

I

I

I

1

Taa v

9

A

y

U

c

L

9

",g"~ C ~. D

-~ ;... N ~

D

v

D

c-.

,

9

I

I

:

1

l

1

I

I

J=~,v*~/'~c'r~..,a*.o.=,'~_.E~-~.,~

I

I

1

!

,'

[].L".~...,~

:

i

"

',

I'

~.: . . . .

[

~

'I

L

I '

-,

I

E ,E,,

t

,

I

1

1

t,,~ =.I,,(L~,~,,:L,,,\.~B

~

Neo ,* , - ' , j "

9~ , ~

,.~'~,'f-'c'C=~'(',l) E ;,...Vv.~,.'~..4",.Xt

: "' ~ ' ; ' > ' " " ~ ' ~ ' ~ " ~'

1

'

' "r U

2

' K . "~ ' ~ "

3

' "~',~

"~

Cl

a, o.."c-~,,

4

5

~ v~-F

a' D c

~

6

Figure 3. HCA plots of amylopullulanase (39E), Taka-c~-amylase (Taa), and neopullulanase (Neo). Numbering starts from the first amino acid of the mature protein for Taa and Neo. The first amino acid of the plot for 39E, R, is the 481st amino acid of the mature protein. Proline is symbolized by , , glycine by 0, serine by ~, and threonine by D. Proposed catalytic residues are ringed with circles. For the parts of the sequences containing catalytic residues, the correspondence between hydrophobic clusters (segments 1-6) are shown by vertical lines.

'

152

Xylose Isomerase Thermozymes Biotechnological Applications Immobilized xylose isomerase is used for isomerization of glucose into fructose in the industrial manufacture of high fructose corn syrup. Equilibrium for the isomerization reaction is shifted to fructose at high temperatures. The current process is operated at 58 "C with non-thermozyme xylose isomerases and gives rise to a 40-42% fructose syrup. This necessitates an additional chromatrographic step to obtain a 55% syrup concentration. Performing the isomerization at 95"C would achieve 55% syrup without the chromatographic concentration step. Native xylose isomerases have evolved to convert xylose to xylulose, not glucose to fructose. Consequently, these current industrial enzymes are also not optimized for glucose isomerization; and, they are consequently immobilized to maximize activity and stability.

General Properties Xylose isomerases are categorized into two classes. Class I is represented by enzymes from Arthrobacter and Thermus; and, they have a shorter N-terminus region than type II enzymes. Representative type II xylose isomerases are isolated from Escherichia, Thermatoga and Thermoanaerobacter. We have cloned, sequenced and characterized the xylose isomerase thermozymes from Thermoanaerobacterium, Thermoanaerobacter and Thermotoga species (9, 20-21). Table 3 compares the genetic constants for different thermostable xylose isomerases from both mesophiles, thermophiles, and hyperthermophiles. Notably, Vmax and K~at and Kcat/Km values for xylose isomerase are not dependent on whether the enzymes is mesophilic, thermophilic or hyperthermophilic. Table 4 shows that for xylose isomerases from Thermotoga the fructose to glucose reaction, Vmax increases with increasing temperature, as does Km. The net effect of this situation is that the catalytic efficiency of the xylose isomerase decreases by a factor of two at 98~ compared with its value over the temperature range of 60 to 90~

Genetics Figure 5 compares xylose isomerase type II thermozyme gene sequences to family I Arthrobacter xylose isomerase sequences in hydrophobic cluster analysis plots. Although type I sequences have an N-terminus about 50 amino acids shorter, the two enzyme classes still have a common catalytic triad. The gene for xylose isomerase (xyla) in Thermotoga neopolitana encoded a polypeptide of 444 residues -- with a molecular weight of 40,892. The native enzyme was a homotetramer with a molecular weight of 200,000. The catalytic triad (His-101, Asp-104, Asp-339), as well as almost all the other residues involved either in substrate or metal binding is conserved among the two protein families.

153

100 A

o~ 9~4.A U ==

-

C

~

1 oo

-'3

80

60

~

6o

40

:3 "O (/1 0

er

f

20 O0

2O

1=0 2~0 3'0 4'0 5()

6'0 7'0 (~

Temperature

dO 910 1()0

0

------~\

0

I

50

I

60

l

70

I

80 Temperature (~

l

90

I

1 O0

Effect of temperature on stablility and activity of Clostridium (a) Thermal stability. The enzyme was placed in acetate buffer (50 mM, pH 6.0) with 5 mM CaC12 and preincubated at various temperatures for 30 min, and then residual amylopullulanase activities were assayed. (b) Effect of heat on activity. The enzyme activity was assayed at various temperatures by the standard assay method (30 min incubation). (Reproduced with permission from Ref. 17. Copyright 1988, The Biochemical Society and Portland Press, London). Figure 4.

thermohydrosulfuricum strain 39E amylopullulanase.

The amino acid compositions of xylose isomerase from family II have been compared, and with the exception of a significant decrease in the Asp and Gln content, no obvious amino acid substitution could be detected between enzymes originating from mesophilic, thermophilic or hyperthermophilic organisms. Figure 6 shows the Asn plus Gln content of family II xylose isomerase as a function of organism growth temperature. There is a correlation in family II enzymes between growth temperature of the organism and the Asn plus Gln content of its xylose isomerase. The finding that decreasing numbers of Asn and Gln residues may contribute to the thermostability of Thermotoga xylose isomerase is understandable because if these residues were involved in stabilization of the enzyme structure, they would be subject to deamination at high temperatures.

155 50.0 #r ,,,

45.0

o Z

o 40.0 nu/

0

~ 35.0

13

\

\ \

\

Ul D a

\

~ 300'

\,

W

\

ff

a

0

+

z 25.0 20.0

.... z~_,_. . . . . . . . . . . . . . . . . 20

30

\

\

~,=~_L . . . . .

40 50 60 70 80 eRO~',rfmTEM~ER.AT~RE(:C)

90

Figure 6. Number of Asn plus Gln residues per monomer of xylose isomerase, as function of growth temperature. Growth temperatures were obtained from the DSM catalog (1993), and Asn plus Gln contents were obtained from xylose isomerase sequences in GenBank by using the Sequence Analysis Software Package of the Genetics Computer Group, version 5 (University of Wisconsin).

Catalytic Mechanism Based on site-directed mutation studies (22-23) on active site residues in the T. thermosulfurigenes xylose isomerase, we have proposed the catalytic mechanism shown in Figure 7. In this model for the reaction catalyzed by xylose isomerase, His-101 is locked into one tautomeric form by interaction with Asp-104, and it acts as a hydrogen bond acceptor to stabilize the substrate and the transition state. Hsp-339, acting as a base, attracts the proton from C2-OH of the substrate. The attraction facilitates the subsequent hydride shift from C-1 to C-2 and simultaneously induces the ring opening. Metal [1] stabilizes the substrate and the transition state by coordination and, perhaps, provides the electrostatic force to stabilize the developing negative charge at the C5-0H.

156 Table 3 Comparison of kinetic constants for thermostable xylose isomerase a Temp Vmax kcat K~ kcat/Km Organism (~ ) (~t/mg) (min 1) (mM) (mM-lmin 1) Family I

Arthrobacter sp. A. missouriensis Streptomyces olivochromogenes Streptomyces griseofuleus Thermus acquaticus

60 60

27.4 33.9

1,190 1,494

210 290

5.7 5.2

60

17.6

760

220

3.4

60

5.3

230

250

0.9

70

9.1

294

93

3.2

60

6.0

330

220

1.5

65 65 90 90

6.3 5.3 16.2 22.4

315 265 810 1,139

Family II

Bacillus stearothermophilus T. saccharolyticum T. thermosulfurigenes T. matitima T. neapolitana aActivities for reference 21.

conversion

of

glucose

to

fructose

120 142 118 98.5 are

reported.

2.6 1.9 6.9 12.9 From

Thermal Features We have recently completed a comparative physiochemical study on thermal features of mesophilic, thermophilic and hyperthermophilic xylose isomerase (24). The optimum temperature for activity and thermal stability for xylose isomerase corresponded to the optimum temperature at which the organism grows. This suggests that these enzymes evolved the thermal features required for function in their environment. The more thermophilic the xylose isomerases also correlated with the higher putative Tm's for thermal unfolding of the enzyme. Interestingly, the more thermophilic the xylose isomerase, the more rigid the enzyme, and, the more resistant it was to precipitation by high temperatures.

Protein Engineering Xylose isomerase functions in nature to convert xylose to xylulose; and, it does not function physiologically to convert glucose to fructose. Consequently, the enzyme was targeted for protein engineering to increase glucose isomerization activity (22). Figure 8 shows a diagram illustrating the amino acid changes we made by site-directed mutagenesis to engineer the Thermoanaerobacterium enzyme to be a "true" glucose isomerase. This was achieved by replacing an active site tryptophane with phenylalanine to remove steric hindrance for glucose binding, and, by replacing a valine with serine to provide more hydrogen bonding to the substrate. Interestingly,

157 the p h e n y l a l a n i n e substitution also doubled the half-life of the enzyme (25). The putative explanation for this result was t h a t addition of a more hydrophobic amino acid to the active site stabilized the enzyme by reducing the w a t e r accessible surface area.

Table 4 Kinetic constants for T. neopolitana xylose isomerase b

Substrate

Temp(~

Xylose

Vma•

a

kcat(minl) b

Zm(mM) a

k~at/Zm 1) (mM-lmin -

90

52.2+1.5

2,654

15.9+2.8

166.8

98

75.0+3.7

3,820

52.1+6.0

73.4

Glucose

90 98

22.4+ 1.3 21.2+0.9

1,139 1,079

88.5+ 16.6 159.5+ 19.3

12.9 6.8

Fructose

60 70 85 90 98

3.6+0.1 9.3+0.2 13.3+0.4 21.0+1.2 28.5+1.5

181 471 680 1,070 1,449

15.1+1.5 48.6+3.1 79.5+6.9 106.5+15.5 260.7+30.3

12.0 9.7 8.5 10.0 5.6

aData are m e a n s +_ s t a n d a r d deviations. bThe kcat is the n u m b e r of substrate molecules reacted per active site per min.

Mill

[Asp-339]

Mill .."i

[Asp-339]

MIll

IAsp.339]

o

~OH ....... " OH :

OH NAN --

--H .... O, O

[llis-101]

LAsp'I04]

N#'~N --H .... O. O

[lli,-101]

[Asp-104]

NIXN--H

[His-101]

.... O, O

J

[Asp-'O4l

Figure 7. Proposed catalytic m e c h a n i s m for D-xylose isomerase involving the cyclic substrate, amino acid catalytic triad, and divalent metal in position [I].

158 2

X

x~

__o

0

m -1

~z

Q~:~

ll

k-

o9

-

I-

o9

>

>

>

~-

~

~

>

>

13.. 133

tl G'3 CO

s O~

s O0

~0 CO

,--,~y

O--'b 09

(:~ C'9

(0 CO

CO

Enzymes

Figure 8. Diagram illustrating amino acid changes of substrate preference from xylose (Xyl) to glucose (Glc) associated with the amino acid substitutions in the substrate-binding pocket of xylose isomerase. The ratios of catalytic efficiency (kcat/Km) of enzymes with xylose versus that with glucose, shown in Table 1, are expressed in a logarithmic scale. The negative values shown by factitious enzymes indicate more favored enzyme specificity toward glucose than xylose, which is required of "true" glucose isomerase. Amino acids are indicated by the single-letter code. Conclusion

Thermozymes appear to have evolved inherently thermostable and thermophilic traits by subtle site-specific changes in amino acid residues and not by evolving totally different global protein structures. Site-specific proline placements in loop regions are hypothesized to aid in protein thermal stabilization; whereas, glutamine and asparagines placements which can deaminate at high temperatures effect protein thermolability. Thermozyme saccharidases appear to offer unique advantages for both starch processing and HCFS manufacture. Finally, thermozymes should become model enzymes for protein chemists and biotechnologists. Once

159 cloned, they are easy to purify and crystallize and, they store readily at room temperature. Since high enzyme stability is a desired trait, one could start with a thermozyme and then alter it by site directed mutagenesis to function on a specified substrate at a specific pH and temperature range.

References 1

Vieille, C., D. Burdette and J.G. Zeikus. 1996. Thermozymes. Advances in Biotechnology. (Manuscript submitted).

2

Saha, B. C., S. P. Mathupala, and J. G. Zeikus. 1991. Comparison of amylopullulanase to c~-amylase and pullulanase, pp. 362-371. IN: Leatham, G. C., and M. E. Himmel (eds.), ACS Symposium Series 460, Enzymes in Biomass Conversion, American Chemical Society, Washington, D.C.

3

Zeikus, J. G., C. Lee, Y.-E. Lee, and B. C. Saha. 1991. Thermostable saccharidases: New sources, uses and biodesigns, pp. 36-51. IN: Leatham, G. F., and M. E. Himmel (eds.), ACS Symposium Series 460, Enzymes in Biomass Conversion, American Chemical Society, Washington, D.C.

4

Lee, Y.-E., S. E. Lowe, and J. G. Zeikus. 1992. Molecular biology and physiological biochemistry of xylan degradation by thermoanaerobes, pp. 275-288. I n J. Visser et al. (eds.), Xylans and Xylanases, Elsevier Science Publishers.

5

Lowe, Susan E., Mahendra K. Jain, and J. Gregory Zeikus. 1993. Biology, Ecology, and Biotechnological Applications of Anaerobic Bacteria Adapted to Environmental Stresses in Temperature, pH, Salinity, or Substrates. Microbiol. Rev. 57:451-509.

6

Zeikus, J. G., S. Mathupala, Y.-E. Lee, S. Podkovyrov, B. C. Saha, M. Meng, and M. Bagdasarian. 1992. Thermophilic enzymes: New sources, uses and biodesigns, pp. 110-113. In: M. R. Ladisch and A. Bose (eds.), Proc. ACS Conference, Harnessing Biotechnology for the 21st Century, ACS Publishers. Hyun, H.H. and J.G. Zeikus. 1985. General biochemical characterization of thermo-stable extracellular ~-amylase from Clostridium thermosulfurogenes. Appl. Environ. Microbiol. 49:1162-1167.

7

8

Hyun, H.H. and J.G. Zeikus. 1985. General biochemical characterization of thermo-stable pullulanase and glucoamylase from Clostridium thermohydrosulfuricum. Appl. Environ. Microbiol. 49:11681173.

160 9

Lee, C., L. Bhatnagar, B. C. Saha, Y.-E. Lee, M. Takagi, T. Imanaka, M. Bagdasarian, and J. G. Zeikus. 1990. Cloning and expression of the Clostridium thermosulfurogenes glucose isomerase gene in Escherichia coli and Bacillus subtilis. Appl. Environ. Microbiol. 56:2638-2643.

10 Lee, C., M. Bagdasarian, M. Meng, and J. G. Zeikus. 1990. Catalytic mechanism of xylose (glucose) isomerase from Clostridium thermosulfurogenes. J. Biol. Chem. 265:19082-19090. 11 Lee, Y-E., S. E. Lowe and J. G. Zeikus. 1993. Regulation and characterization of xylanolytic enzymes of Thermoanaerobacterium saccharolyticum B6A-RI. Appl. Environ. Microbiol. 59:763-771. 12 Lee, Yong-Eok and J. Gregory Zeikus. 1993 (B). Genetic organization, sequence and biochemical characterization of recombinant ~-xylosidase from Thermoanaerobacterium saccharolyticum strain B6A-RI. J. Gen. Microbiol. 139:1235-1243. 13 Lee, Yong-Eok, Susan E. Lowe and J. Gregory Zeikus. 1993. Gene Cloning, Sequencing, and Biochemical Characterization of Endoxylanase Appl. Environ. from Thermoanaerobacterium saccharolyticum B6A-RI. Microbiol. 59:3134-3137. 14 Lee, Yong-Eok, S. E. Lowe, B. Henrissat and J. G. Zeikus. 1993. Characterization of the active site and thermostability regions of endoxylanase from Thermoanaerobacterium saccharolyticum B6A-RI. J. of Bacteriol. 175:5890-5898. 15 Mathupala, Saroj P., Jong-Hyun Park and J. Gregory Zeikus. 1994. Evidence for a-1,6 and a-l,4-Glucosidic Bond Cleavage in Highly Branched Glycogen by Amylopullulanase from Thermoanaerobacter Ethanolicus. Biotechnol. Lett. 16:1311-1316. 16 Park, J.H., S. Mathupala, C. Peterson and J.G. Zeikus. Findings.

Unpublished

17 Saha, B.C., S. Mathupala, and J.G. Zeikus. 1988. Purification and characterization of a highly thermostable pullulanase from Clostridium thermohydrosulfuricum. Biochem. J. 252:343-348. 18 Mathupala, S., B. C. Saha, and J. G. Zeikus. 1990. Substrate competition and specificity at the active site of amylopullulanase from Clostridium thermohydrosulfuricum. Biochem. Biophys. Res. Comm. 166:126-132.

161 19 Mathupala, Saroj P., Susan E. Lowe, Sergey M. Podkovyrov and J. Gregory Zeikus. 1993. Sequencing of the Amylopullulanase (aup) Gene of Thermoanaerobacter ethanolicus 39E, and Identification of the Active Site by Site-directed Mutagenesis*. J. Biol. Chem. 268: 16332-16344. 20 Lee, Yong-Eok, Matur. V. Ramesh and J. Gregory Zeikus. 1993. Cloning, sequencing and biochemical characterization of xylose isomerase J. Gen. from Thermoanaerobacterium saccharolyticum strain. B6A-RI. Microbiol. 139:1227-1234. 21 Vieille, Claire, J. Mike Hess, Robert M. Kelly and J. Gregory Zeikus. 1995. xylA Cloning and Sequencing and Biochemical Characterization of Appl. Environ. Xylose Isomerase from Thermotoga neapolitana. Microbiol. 61:1867-1875. 22 Meng, M., C. Lee, M. Bagdasarian, and J. G. Zeikus. 1991. Switching substrate preference of thermophilic xylose isomerase from D-xylose to D-glucose by redesigning the substrate binding pocket. Proc. Natl. Acad. Sci. 88:4015-4019. 23 Meng, Menghsiao, Michael Bagdasarian, and J. Gregory Zeikus. 1993. The role of active-site aromatic and polar residues in catalysis and substrate discrimination by xylose isomerase. Proc. Natl. Acad. Sci. USA 90:8459-8463. 24 Tchernenko, V., C. Vieille, D. Burdette and J.G. Zeikus. 1996. Relationships between enzyme molecular rigidity, stability and activity: analysis of xylose isomerase thermal features. Appl. Biochem. and Biotech. (Manuscript Submitted). 25 Meng, Menghsiao, Michael Bagdasarian and J. Gregory Zeikus. 1993. Thermal Stabilization of Xylose Isomerase from Thermoanaerobacterium thermosulfurigenes. Bio/Technology, 11:1157-1161.


Enzymesfor CarbohydrateEngineering K.H. Park, J.F. Robyt and Y-D. Choi (Editors) 9 1996 Elsevier Science B.V. All rights reserved. Crystal S t r u c t u r e Resolution

of

Bacillus

licheniformis

163 a-Amylase

at

1.7

Hyun Kyu Song, Kwang Yeon Hwang, Changsoo Chang, and Se Won Suh

Department of Chemistry and Center for Molecular Catalysis, College of Natural Sciences, Seoul National University, Seoul 151-742. Korea

Abstract

(x-Amylases ((z-l,4-glucan 4-glucanohydrolase, E.C.3.2.1.1)catalyze the cleavage of a-l,4-glucosidic linkages of starch components, glycogen, and various oligosaccharides. Thermostable ~-amylases from Bacillus species are of great industrial importance in the production of corn syrup or dextrose. Thermostable (x-amylase from Bacillus licheniformis, a monomeric enzyme with molecular mass of 55,200 Da (483 amino acid residues), shows a remarkable heat stability: its optimum temperature is 90 ~ and only 10 % of activity is lost after heat t r e a t m e n t at 90 ~ for 30 min. Thus this enzyme provides an attractive model for investigating the structural basis for thermostability of proteins. The threedimensional structure of thermostable (x-amylase from Bacillus licheniformis has been determined by the multiple isomorphous replacement method of X-ray crystallography. The structure has been refined to a crystallographic R-factor of 0.199 for 58,601 independent reflections with Fo > 2a(Fo) between 8.0 and 1.7 .~ resolution, with excellent stereochemistry. The final model consists of 468 amino acid residues and 294 water molecules. Missing from the model are the segment between Trp182 and Asn192 and both the N- and C-termini. The polypeptide chain fold into three distinct domains. The first domain (domain A), consisting of 291 residues (from residue 3 to 103 and 207 to 396), forms a (~/(x)s-barrel structure. The second domain (domain B), consisting of residues 104 to 206, is inserted between the third ~-strand and the third (x-helix of domain A. The third C-terminal domain (domain C), consisting of residues 397 to 482, folds into an eight-stranded antiparallel ~-barrel.

164

Introduction

(x-Amylases ((x-l,4-glucan-4-glucanohydrolase, E.C.3.2.1.1) catalyze the cleavage of (x-l,4-glucosidic linkages of starch components, glycogen, and various oligosaccharides. They are widely distributed in bacteria, fungi, plants, and m a m m a l i a n tissues [1-3]. Only four highly conserved regions are found in the amino acid sequences of all (x-amylases [4]. Despite low sequence identity, however, crystal structures of fungal, plant, and animal (x-amylases showed a similar tertiary folding [5-15]. Thermostable (x-amylases from Bacillus species are of great industrial importance in the production of corn syrup or dextrose [16]. Thermostable ~amylase from Bacillus licheniformis (abbreviated here as BLA), a monomeric enzyme with molecular mass of 55,200 Da (483 amino acid residues) [17], shows a remarkable heat stability: its optimum temperature is 90 ~ and only 10 % of activity is lost after heat treatment at 90 ~ for 30 min [18]. Its pH optimum is 6 [18]. It is one of the most thermostable natural enzymes used in biotechnological processes. Since it is one of the most widely used biocatalysts in industry, there is a great commercial interest in the protein engineering of this enzyme. Therefore, thermostable (x-amylase from B. licheniformis provides an attractive model for investigating the structural basis for thermostability of proteins. Crystals of thermostable (x-amylase from B. licheniformis suitable for Xray analysis were reported previously [19,20]. Here we present the independent structure determination of thermostable (x-amylase from B. licheniformis by the multiple isomorphous replacement (MIR) method and subsequent refinement at 1.7 A resolution. More detailed description of this work will be published elsewhere.

Results and d i s c u s s i o n

Quality of the refined model The structure has been refined to a crystallographic R-factor of 0.199 for 58,601 independent reflections with Fo > 2a(Fo) between 8.0 and 1.7 /~ resolution, with root mean square (r.m.s.) deviations of 0.013 A, and 1.72 o from ideal bond lengths and bond angles, respectively. The final model consists of 468 amino acid residues and 294 water molecules. Missing from the model are the segment between Trp182 and Asn192 and both the N- and C-termini (Table 1). A Ramachandran plot of the O/qJ angles shows that 86.7 % of the residues are in the most favored regions and only Tyrl50, which has a well-defined electron density, is in the disallowed region as determined by PROCHECK [21]. The upper estimate of the error in the atomic positions given by the Luzzati plot is

165 0.25 ,s [22]. Missing from the model are the segment between Trp182 and Asn192, the N-terminal Alal and Asn2, and the C-terminal Arg483. Calcium ion is not included in the final model and chloride ion has been treated as water molecules.

Table 1 Statistics and quality of refined BLA model R-factor R-factor Resolution range (A) Ramachandran outlier

0.199 0.226 8.0 - 1.7 1 (Tyr150)

R.m.s. deviation from ideal geometry Bond distance (,~) Angles (~ Dihedral angles (o) Improper (o)

0.013 1.72 25.2 1.56

Averaged temperature factor (A 2) All atoms Main-chain atoms Side-chain atoms Water molecules

30.7 26.7 32.6 43.6

No. of residues Missing residues Non-hydrogen atom No. of water molecules

469 1-2, 182-192, 483 3,800 294

Description of the structure The structure of BLA is made of three distinct domains (Fig. 1). The largest domain (domain A), formed by 291 residues (3 to 103 and 207 to 396), is a parallel ~-barrel of eight strands interconnected serially by helices and extensive loops between them. The topology is typical of a commonly observed (~/a)8- or TIM-barrel (Fig. 2). This domain provides a compact, stable and probably rigid core of the structure. The second domain (domain B), inserted between the third ~-strand and the third a-helix of domain A, begins at residue 104 and ends with residue 206. The third C-terminal domain (domain C), containing 86 residues from 397 to 482, folds into an eight-stranded antiparallel ~-barrel of the Greek key topology. This domain lies tightly against helices 7 and 9 of (~/(z)s-barrel of

166 domain A to give the entire a-amylase molecule an ellipsoidal appearance, with dimensions roughly 55 .~ x 55 A x 70 A. As seen in Figure 1, the amino terminal side of the barrel of domain A is rather blunt, because the helix-to-strand connections are fairly brief with most loops having only a few residues. In contrast, the strand-to-helix connections on the carboxy terminal side are much longer and more complex. A deep cleft, presumably containing the sites for substrate binding and catalysis, is formed on this side between the two domains A and B. The eleven residue segment of domain B from residue 182 to 192 do not show the elctron density. This is also observed by Machius, Wiegand, and Huber [23] who showed that a cleavage had taken place between Gln189 and Asp190 due to the contaminating Glu-C endopeptidase.

Figure 1. Schematic representation of a-amylase from Bacillus licheniformis. The structure of BLA is made of three distinct domains. The helices and strands are shown in cyan and red, respectively. Secondary structures were assigned with the program PROCHECK.

167

Figure 2. Schematic representation of (~-amylase from Bacillus licheniformis. The helices and strands of major domain (~/a)s-barrel are shown in cyan and red, respectively. The possible catalytic residues are identified (Asp231, Asp328, and Glu261).

Calcium and chloride binding sites A strongly bound Ca 2§ ion has been identified in ~-amylases, essential for the tertiary structure and catalytic activity of the enzymes. We observe a similar Ca 2§ binding site between the two domains A and B in the structure of BLA as in TAA [5] and PPA [8]. This site is formed by the side-chains of residues Asnl04 and Asp 200 and a main-chain atom of His235. The Ca ~§ binding site is occupied by a water molecule (with a B value of 55.6 A 2) in this structure, presumably because the ADA buffer at 100 mM concentration in the hanging drop acted as the chelator, as pointed out by Machius, Wiegand, and Huber. This is supported by the fact that the electron density at the expected Ca 2§ position is not strong

168 enough and the ligand atoms would be too close if the water is replaced by a Ca 2§ ion. A chloride ion binds to and allosterically activates PPA [2]. The chloride binding site in PPA is located near the Ca 2§ binding site in the C-terminal side of the central barrel of domain A. Machius, Wiegand, and Huber observed a bound chloride ion with a B-value of 9.6 ,s which is coordinated by four ligands (NH2 of Arg229, ND2 of Asn326, and two water molecules). In our structure the electron density at the chloride position is not strong enough and when it is modelled as a chloride ion, the B value becomes very high (43.6 ,~ 2). Instead, when a water molecules is placed, its B value is reasonable (25.8 .~ 2) and the water molecules makes reasonable contacts with the surrounding atoms.

Conserved regions and active site Only four short segments of the polypeptide chain appear to be conserved in all a-amylases from bacteria, fungi, plants, and animals [4]. They are BLA residues 99-106 (region I), 226-235 (region II), 261-264 (region III), and 322-331 (region IV). One of the most striking features in this structure is a deep cleft which runs for about 30 A on one side of the molecule and separates the two globular units of different size. The active site is believed to be situated along this characteristic groove between the carboxy-terminal side of domain A ~barrel and domain B. On the basis of comparison of amino acid sequences and tertiary structures of BLA, TAA, and PPA, we propose that Asp 231, Asp328, and Glu261 are essential catalytic residues in BLA. Corresponding residues in the homologous enzymes are Asp206, Asp297, Glu230 for TAA, Asp197, Asp300, Glu233 for PPA, and Asp229, Asp328, Glu257 for Bacillus circulans cyclodextrin glucanotransferase (CGT) [24]. Aromatic residues such as Phe41 and Tyr56 are present in the cleft and they appear to interact with sugar rings by stacking at the entrance of the cleft. A network of water molecules is formed in the active site cleft, making hydrogen bonds with side chain atoms of Arg229, Asp 231, Asp328, and Glu261.

Conclusions

This study reveals the architecture of the thermostable a-amylase from Bacillus licheniformis. By homology with other (~-amylases, important active site residues can be identified as Asp231, Glu261, and Asp328, which are all located at the C-terminal end of the central (~/(~)s-barrel. Since many of the stabilizing and destabilizing mutations obtained so far fall in domain B or at its border, this region of the enzyme appears to be important for the thermostability.

169

Acknowledgements This work was supported by a grant from Center for Molecular Catalysis. A travel fund from Pohang Light Source is acknowledged. Beamline X12C at Brookhaven National Laboratory's National Synchrotron Light Source is supported by the Office of Health and Environmental Research of the United States Department of Energy. We thank Professor N. Sakabe and Dr. A. Nakagawa for their assistance with data collection at beamline BL-6A2, Photon Factory, Japan and the Inter-University Center for Natural Science Research Facilities, Seoul National University for providing X-ray equipments.

References 1. Takagi T, Toda H, Isemura T. In: Boyer PD, 3rd ed. The Enzymes. Academic Press, New York, 1971; 5: 235-271. 2. Thoma JA., Sparadlin JE, Dygert S. In: Boyer PD, 3rd ed. The Enzymes. Academic Press, New York, 1971; 5: 115-189. 3. Fogarty WN, Kelly CT. In: Rose AH, ed. Microbial Enzymes and Bioconversions. Academic Press, New York, 1980; 115-170. 4. Nakajima R, Imanaka T, Aiba S. Appl Microbiol Biotechnol 1986; 23: 355-360. 5. Matsuura Y, Kusunoki M, Harada W, Kakudo M. J Biochem 1994; 95: 697702. 6. Swift HJ, et al. Acta Crystallogr 1991; B47: 535-544. 7. Buisson G, Duee E, Haser R, Payan F. EMBO J 1987; 6: 3909-3916. 8. Qian M, Haser R, Payan F. J Mol Biol 1993; 231: 785-799. 9. Larson SB, Greenwood A, Cascio D, Day J, McPherson A. J Mol Biol 1994; 235:1560-1584. 10.Qian M, Haser R, Payan F. Protein Sci 1995; 4: 747-755. ll.Wiegand G, Epp O, Huber R. J Mol Biol 1995; 247: 99-110. 12.Boel E, et al. Biochemistry 1990; 29: 6244-6249. 13.Brady RL, Brzozowski AM, Derewenda ZS, Dodson EJ, Dodson GG. Acta crystallogr 1991; B47: 527-535. 14.Kadziola A, Abe J, Svensson B, Haser R. J Mol Biol 1994; 239: 104-121. 15.Farber GK, Petsko GA. Trends Biochem Sci 1990; 15: 228-234. 16.Peppier HJ, Reed G. In: Rehm H-J, Reed G, eds. Biotechnology. VCH Verlags, Weinheim, 1987; 7a: 547-603. 17.Yuuki T, et al. J Biochem 1985; 98: 1147-1156. 18.Endo S. In: The Amylase Society of Japan, ed. Handbook of Amylases and Related Enzymes. Pergamon Press, Oxford, 1988; 47. 19.Suzuki A, Yamane T, Ito Y, Nishio T, Fujiwara H, Ashida, T. J Biochem 1990; 108: 379-381.

170 20.Lee SY, Kim S, Sweet RM, Suh SW. Arch Biochem Biophys 1991; 291: 255257. 21.Laskowski RA, MacArthur MW, Moss DS, Thornton JM. J Appl Crystallogr 1993; 26: 283-291. 22.Luzzatti PV. Acta Crystallogr 1952; 5: 802-810. 23.Machius M, Wiegand G, Huber R. J Mol Biol 1995; 246: 545-559. 24.Klein C, Schulz GE. J Mol Biol 1991; 217: 737-750.


171

Characteristics of Carbohydrase Reactions in Heterogeneous Enzyme Reaction System Utilizing Swollen Extrusion S t ~ c h as a Substrate Yong-Hyun Lee and Dong-Chan Park Department of Genetic Engineering, College of Natural Sciences, Kyungpook National University, Taegu 702-701, Korea Introduction

Starch, a glucose homopolymer, is one of the most widely available plant polysaccharides. The various decomposed products of starch produced by hydrolytic or transglycosylation reaction by various carbohydrase, such as, glucose, maltose, maltooligosaccharides, cyclodextrins, or transglycosylated glucosides, were widely used in foods, beverages, pharmaceuticals, cosmetics, and other industrial purposes[I, 2]. Productions of glucose, maltose, cyclodextrins, and other products derived from starch have been carried out by two stages; liquefaction of starch and then saccharification or transglycosylation[3, 4]. However, above traditional enzyme reaction system utilizing liquefied starch as the substrate for follow-up enzyme reaction by various carbohydrase has the following shortcomings, such as, high energy consumption, low product yield, and complexities of separation and purification of products because the residual oligosaccharides formed at liquefaction step and remained after main reaction by various carbohydrases[5-8]. If the above products can be produced from insoluble raw starch directly without liquefaction of starch, it can be expected that the high purity of products can be obtained without accumulation of undesirable oligosaccharides, for enzyme reaction may be carried out directly from the glycosyl residues in the surface of raw starch. Also, the easy separation of residual insoluble starch by simple separation process will facilitate the purification of various products. However, because the raw starch exists as the compact crystalline structure, the enzyme reaction rate and yield of products from raw starch would be too low for industrial implication[9]. Therefore the structural modification of raw starch is required to increase the susceptibility to enzymes. In order to increase the susceptibility to various carbohydrase, the raw starch was extruded instead of liquefaction by cooing and liquefying enzyme. The extrusion starch exists as nearly water insoluble suspension state, the intermediate states between water soluble liquefied and insoluble raw starch, therefore the practical heterogeneous enzyme reaction system composed of soluble enzyme and nearly insoluble substrate could be maintained. In this way the advantages of direct production of cyclodextrin, maltose, and transglycosylated stevioside from surface of swollen extrusion starch may be achieved.

172 In this work, characteristics of carbohydrase reaction in heterogeneous enzyme reaction system utilizing swollen extrusion starch as substrate were investigated. This work includes, 1) the application for production of cyclodextrin directly from extrusion starch, 2) determination of the optimal conditions for production of cyclodextrin, 3) mechanistic and kinetic analysises of cyclodextrin production reaction, 4) the utilization for production of the high purity concentrated maltose and its reaction characteristics, and then 5) the transglycosylation of stevioside using extrusion swollen starch as a glycosyl donor.

Materials and methods

Enzymes Cyclodextrin glucanotransferase(CGTase) from Bacillus macerans(Amano Co.), the specific activity of 280 units/mg protein, was used for cyclodextrin production and transglycosylation of stevioside. Optimum pH and temperature were 6.0 and 55~ respectively. One unit of enzyme was def'med as the amount of enzyme which produces 1 mg of cyclodextrin from 5 ml of 5%(w/v) soluble starch with 0.1 ml of CGTase at pH 6.0, 55"C for 1 hour. Fungal a-amylase from Aspergillus oryzae(Novo Co.), the specific activity of 1,100 unitlmg protein, was used for maltose formation. One unit of enzyme was defined as the amounts of enzyme which produces 1 tJ mole of maltose from 2 ml of 1% soluble starch with 0.2 ml of fungal a-amylase at pH 5.0, 40 ~ for lmin. Carbohydrates used as substrate Corn starch(moisture content; 12%) was used as raw starch, and soluble starch(Sigma Co.) was also utilized for comparison. Mono- and di-saccharides, such as glucose, xylose, sorbose, inositol, maltose, sucrose; glucosides, such as stevioside, hesperidin, salicin, were used as acceptors for transglycosylation reaction. Extrusion of raw corn starch Corn starch, preconditionated at the equilibrated moisture content of 12, 15, 18, 20, and 25%(w/w), was extruded in a single screw extruder at feed rate of 300 g/rain and screw speed of 150, 200, 250, and 300 rpm, thereafter, dried, and powdered. Determination of structural features of raw and extrusion starches Degree of gelatinization of swollen extrusion starch was determined by Wooton's method[10]. Granular structure of raw and extrusion starch was observed by scanning electron microscope(SEM, ISI-SS 130, Asahi Co.) after coating with ion coater(IB-5, Eiko Co.).

173

Enzyme reaction for productions of cyclodextrin, maltose, and transglycosylated stevioside using extrusion starch 100 g of swollen extrusion starch was suspended in 1 1 of 10 mM Tris-malateNaOH buffer(pH 6.0), and 900 units of CGTase/1 were added, and then the cyclodextrin production reaction was carried out at 200 rpm and 55~ 300 g of swollen extrusion starch was suspended in 1 1 of 10 mM acetate buffer(pH 5.5), and 40 units of fungal a-amylase/g starch was added, and then the maltose production reaction was carried out at 250 rpm and 50 ~ 50 g of swollen extrusion starch was suspended in 1 1 of 10 mM Tris-malateNaOH buffer(pH 6.0), and 50 g/l of acceptors, 900 units of CGTase/1 was added, and then the transglycosylation reaction was carried out at 200 rpm and 55~

Enzyme reactions for production of cyclodextrin, maltose, and transglycosylated stevioside using liquefied starch The reaction conditions utilizing liquefied starch were the same as the above conditions, but after liquefaction at 90~ for 20 min(DE 10), with 2,200 units of a-amylase/1 from B. licheniformis(Sigma Co.).

Analytical methods The profile and content of produced CDs, maltose, stevioside, and transglycosylated stevioside were determined by HPLC(Gilson Co.); Cosmosil 5NH2 colunm(Nacalai Co.), acetonitrile/water(65/35), I ml/min, and RI detector. Concentration of CDs was also determined by spectrophotometric methods[l 1] and reducing sugar was determined by DNS method[12].

Results and discussion

Profile of the enzyme reaction for production of cyclodextrin using extrusion starch as the substrate[13-15] The progresses of enzyme reaction of cyclodextrin(CD) production using raw starch, liquefied starch, and swollen extrusion starch, were compared in Figure 1. The highest concentration of 54 g/1 CDs(total of a - , 13-, and ,f-CD) was obtained after 24 hours for extrusion starch, compared to 45 g/l of that of liquefied starch. Meanwhile only 6 g/l of CD was produced after 24 hours from raw starch indicating that CD production reaction from raw starch is very limited. Figure 1 also showed that the rate of CD synthesis from swollen extrusion starch was slightly lower at the initial stage of reaction compared with that of liquefied starch, however increased steadily, and then exceeded from after 4 hours of reaction. Figure 2 compares the profiles of produced CDs and the other maltooligosaccharides in the reaction mixtures produced from reactions carried out extrusion and liquefied starch as the substrate. In the case of liquefied starch, a

174 60

[(B)

(A) ,

4

40

0 0

6

12

18

24 0

Reaction time, hr Figure 1. Comparison of cyclodextrin produced from raw(z~), swollen extrusion(o), and liquefied starch(J2). 100 g of starch/l, 900 units of CGTase/l, 200 rpm, pH 6.0, and 50"C.

I 5

1

100

5

10

Retention time, min Figure 2. Comparison of HPLC chromatogram of cyclodextrin and maltooligosaccharides in reaction mixtures produced from extrusion starch(A) and liquefied starch(B) after 12 hr.

significant amount of glucose, maltose, and other oligosaccharides were accumulated, which may be in part produced during liquefaction stage of raw starch by means of heating and liquefying a-amylase, or may be in part the residual oligosaccharides remained after CD formation reaction as illustrated in Figure 3. On the other hand, in the case of swollen extrusion starch, CD was mainly produced without accumulation of any significant amount of maltooligosaccharides as can be seen in HPLC chromatogram. Above characteristics may be explained by the action of CGTase[16] that synthesize CD from the nonreducing ends of extrusion starch, and the extrusion starch existed in swollen granular structure without significant fragmentation, so the formation of other maltooligosaccharides was depressed. Also the unreacted residual starch could be easily separated by simple unit operation, such as centrifugation which will facillitate recovery and purification of CD produced after enzyme reaction. It was also suggested that the spent CGTase remained in the reaction mixture could be recovered for reutilization by adsorbing on fresh swollen extrusion starch. Above facts suggest that the CD production in heterogeneous enzyme reaction system utilizing the swollen extrusion starch seems to have potential advantages for CD production in industrial scale, and further study needs to be conducted.

175

Raw Starch

Liquefaction

xtrusion ct-amylase

,;(/,/

!, t

/

~//

Liquefied

~]1'( j ? ~ / ) / ~ / ! t\ (Soluble)

~

' ~ / 1 Extrusion ' / ' J v'-'/ ['--"'~J i starch ~ \ , ,,~\\(Insolub le)

I CGTase Cleavage site

I CGTase (_~)

Subsites CGTas~ ....... "

CGTase Glucose residues

Starch molecule

Figure 3. Schematic representations of micellar organization of starch granule, extrusion starch, liquefied starch, and the proposed hypothetical reaction mechanism of cyclodextrin production from the chain of micelle of swollen extrusion starch and soluble liquefied starch by cyclodextrin glucanotransferase.

Determination of optimal reaction conditions for cyclodextrin production in heterogeneous enzyme reaction system[13] Table 1 shows the effects of the amounts of extrusion starch and CGTase on the production of CD after 24 hours. The highest CD concentration of 58 g/l was obtained at extrusion starch concentration of 150 g/l, and the CD production reaction was hindered at concentration over than 150 g/l due to the lack of the suspension water caused by the penetration of added water molecule into the extrusion starch, which prohibited maintaining the suspension state for enzyme reaction. The ratio of CGTase to substrate(E0/S0, units of CGTase/g of substrate) was found to be an important factor for CD synthesis as can be seen in Table 1. The CD synthesis was increased as the mixing ratio reached up to 15.0; however, it was decreased at over than 15.0. The observed decrease at high ratio may be caused by the various side reaction of CGTase, such as hydrolysis or coupling reaction carried out by excess amount of CGTase[ 17].

176 Table 1 Effect of the amount of CGTase on cyclodextrin production at different extrusion starch concentration.

Amount of CGTase

Concentration of extrusion starch (So, g/l)

(Eo, units~)

50

100

150

200

300 600 900 1,200 1,500

29.8* 32.6 34.2 34.5 33.4

42.3 49.6 54.2 54.9 55.4

46.8 57.6 58.3 58.4 58.1

39.4 43.2 56.5 55.8 54.9

The reaction was carried out at the condition of pH 6.0, 50 ~ , and 200 rpm. * The amount of cyclodextrins( a - , B-, Y -CD)(g/1) after 24 hr

Reaction mechanism of direct production of cyclodextrin from extrusion starch[13-15] Figure 4 compares the microscopic granular structures of raw starch, extrusion starch, and residual starch after reaction with CGTase 2, 4, 8, and 12 hours, respectively, observed by SEM. The starch granules were swollen significantly by extrusion as can be seen in Figure 4(B), the diameter increased around 3-5 times corresponding surface area of around 9-25 times compared to those of raw starch. As the CD production by CGTase proceeded, the swollen starch granules were started to be splitted at the initial stage of reaction, and fragmented into many small particles during initial period of 12 hours. The swollen granules were almost disintegrated and disappeared after 12 hours(Figure 4(E)), and only a few unextrusion raw starch during extrusion process was remained. The extrusion starch that swollen in suspension state as a colloidal form seems to generate significant amount of accessible surface area where the enzyme reaction of CGTase can be carried out, which stimulate enzyme reaction that required for high yield and rate for CD production.

Kinetic analysis of cyclodextrin formation in heterogeneous enzyme reaction system[18, 19] Kinetic equations that can describe CD production reaction in heterogeneous enzyme reaction system utilizing swollen extrusion starch was developed based on the following assumptions and reaction mechanism(Figure 5). 1) Swollen extrusion starch exists in two phase, accessible and inaccessible, and the amount of accessible phases is closely correlated with the degree of gelatinization of swollen starch. 2) CD synthesis reaction initiated by the adsorbed enzyme on the surface of swollen starch, and CGTase exists in two forms, adsorbed and unadsorbed. 3) CD production was subjected to the competitive inhibition by CDs produced, most severely by r And 4) the swollen extrusion starch

177

Figure 4. Scanning electron microscopic photogram of granular structure of raw corn starch(A), extrusion corn strach(B) and residual extrusion starch subjected to the CGTase action after 2, 8, 12, and 24 hours(C-F).

ha E

-

k~d

"~ E *

E* +

kl,a Sa ~" k2,a-

E* +

Si

E S

+ P

-" kl, k2,ii

(1)

E*

9

Sa

E*'Si

kl,ip% E - P~

cL -"k2,ipcx t Si

ks,a

-*

E

+

P

(2)

(3)

(4) (5)

Figure 5. Proposed reaction mechanism of cyclodextrin synthesis in heterogeneous enzyme reaction system utilizing swollen extrusion starch. Where E; CGTase, E*; adsorbed CGTase on extrusion starch, Sa and Si; accessible and inaccessible form of swollen starch, P; produced CDs, P~; (x-CD, respectively.

178 was transformed into the less accessible forms as CD production reaction that can be correlated with the degree of conversion of starch to CDs. The derived kinetic equations can be summarized as shown in Figure 6, and various kinetic constants were evaluated at different reaction conditions as shown in Table 2. Figure 7-10 shows the comparison of theoretical values and experimental values of produced CD at various reaction conditions, such as different concentration of substrate, amount of enzyme, and the degree of gelatinization. The theoretical values were obtained from simulation of Eq.(6), (7) by Runge-Kutta's method. The results of simulation showed that the derived kinetic equations could predict the CD production from swollen starch reasonably well, and that can be utilized for optimization and process development.

RtCD =

V'm~ [S]

(6)

I~'m (1 + [p3IQS~) + [S]

where~ I~m =

I~(l+(rc/(~2-a)) 1 + (KA/KIs)( rc/(D.- a)) '

1 ~2 = x " - - + a F

F-

RE~I}

V'm =

[Si'~ [Sa,o]

~ Vmax 1 + (KA/KIs)(rc/(~-a))

(re =2.532x10 "2, a=15.514x10 2)

(~

I:~2'CD R1,CD

1-([CI~]/[So]) n 1 (1

[S](So])n

V'm~. IS]

(7)

K'm~ (1 +[PjqQs~)+ [S]

where, KA~ (1+(rc/(s I~' m =

1 + (KAdKIs)(rd(s

9 Vmax'

V" m =

1+ (KAJKIs)(x/(s

Figure 6. Summary of derived kinetic equations for cyclodextrin production in heterogeneous enzyme reaction system utilizing swollen extrusion starch. Where RtCD, R~CD; production rate of total CDs and a-CD, ~ ; degree of gelatinization, r'; ratio of inaccessible to accessible portion in substrate, (P; function of structural transformation, RI,CD, R2,CD; initial CD production rate of fresh and residual starch

179 Table 2 Kinetic constants for the reaction of cyclodextrin production according to the degree of gelatinization.

Degree of gelatinization(%)

K "m (g/l)

V "max (g/l" rain)

K "m~ (g/l)

V "max~ (g/l" rain)

41.56 44.51 51.67 58.09 63.52 76.01

8.06 7.42 6.44 5.62 5.03 4.85

0.2532 0.2781 0.3225 0.3472 0.3510 0.3281

5.42 5.22 5.30 5.22 4.60 5.50

0.1722 0.1804 0.2039 0.2112 0.2645 0.2608

,00

r (Si/Sa) 0.0972 0.0892 0.0700 0.0610 0.0505 0.0419

80 ~9 60

% 8o

......

) .......... o ..........

~

6

20

m U

0 0

2

4

6

8

Reaction time, hr

Figure 7. Comparison of theoretical and experimental values of produced cyclodextrin and hydrolyzed starch concentration in heterogeneous enzyme reaction system utilizing swollen extrusion starch. 100 g of starch~, 0.1 unit of CGTase/g starch, pH 6.0, 50 ~ and 200 rpm. Lines, theoretical curve; symbols, experimental values O, --; produced CD, El, ""; hydrolyzed starch

0

2

. I 4

I ..... 6

Reaction time, hr

Figure 8. Comparison of theoretical and experimental values of produced cyclodextrin at different starch concentration. 0.1 unit of CGTase/g starch, pH 6.0, 50 ~ and 200 rpm. Lines, theoretical curve; symbols, experimental values O, --; 50, /~, ---; 75, E], ""; 100 g of extrusion starch/l

180 100 %

8O

80 .........~

o

60

6o

E]

.....

..........

~

o~

f

/

40

.

0

JA

~ 9 40

/

2

4

6

8

Reaction time, hr Figure 9. Comparison of theoretical and experimental values of produced cyclodextrin at different CGTase concentration. 100 g of starch/l, pH 6.0, 50~, and 200 rpm. Lines, theoretical curve; symbols, experimental values O , - - ; 0.01, A , - - - ; 0.1, [-], ""; 1.0 unit of CGTase/g starch

0

i

I

I

2

4

6

8

Reaction time, hr Figure 10. Comparison of theoretical and experimental values of produced cyclodextrin at different degree of gelatinizaiton. 100 g of starch/l, 0.1 unit of CGTase/g starch, pH 6.0, 50~ and 200 rpm. Lines, theoretical curve; symbols, experimental values Initial degree of gelatinizaiton(%): O , - - ; 41.56, A , - - - ; 51.67, [-1,-..; 63.52, V, ---; 76.01

Production of high purity concentrated maltose in heterogeneous enzyme reaction system utilizing extrusion starch[20] The production of concentrated purity maltose using insoluble extrusion corn starch as a substrate was carried out, and compared with the maltose forming reaction using liquefied starch as substrate(Figure 11). Maltose content was much higher in reaction system using extrusion starch than using liquefied starch, whereas, the amount of maltotriose and maltooligosaccharides were reduced. The sugar profiles produced from extrusion starch of different degrees of gelatinization are depicted in Figure 12 and initial maltose forming rate was evaluated as in Figure 13, the optimal degree of gelatinization of extrusion starch suitable for maltose formation was found to be around 70%. The optimal amount of enzyme was 400 units of fungal a-amylase per g of starch, and the reaction time was 12 hours as can be seen in Figure 14. At extrusion

181 250 (B) Liquefied starch

(A) Extrusion starch 200 ~ 1150 0

100 50 0~~=~,~___~ 0 4

8

0

4

8

12

Reaction time, hr Figure 11. Comparison of the progresses of maltose production from extrusion starch(A) and liquefied starch(B). 300 g of starch/l, 400 units of fungal cz-amylase/g starch, pH 5.5, 50~ and 250 rpm. Maltose(O), total maltooligosaccharides(ll), glucose( /~ ), maltotriose(V), maltotetraose(O), other maltooligosaccharides(O)

starch concentration of 300 g/l, maltose concentration was reached up to 220 g/1 and its content among produced reducing sugar was 77%(w/w). The maltose forming reaction was also successfully proceeded at high starch concentration of 700 g/l, however, the conversion yield and content were decreased. By the addition of extrusion starch by fed-batch wise, the maltose concentration, purity, and conversion yield could be improved up to 465 g/l, 70%(w/w), and 0.63, respectively, as shown in Figure 15. The investigated maltose production process seems to have many potential advantages over the conventional process utilizing liquefied starch as compared in Table 3. The production of maltose utilizing swollen extrusion starch seems to have many technical advantages, such as, high reaction rate and high yield, production of high purity concentrated maltose, and low energy consumption, over the conventional method utilizing liquefied starch. The feasibility for industrial application needs to be evaluated.

Characteristics of maltose formation reaction in hetereogeneous enzyme reaction system[21] Characteristics of maltose formation in heterogeneous enzyme reaction system containing swollen extrusion starch were investigated using fungal a-amylase.

182

12 300 I

d

-%

.~

8-

o

4

-

~

"" 0

25

50

75

100

~

I

I

25

50

75

100

Figure 13. Initial rate of maltose formation from extrusion starch of different degrees of gelatinization. Reaction conditions were the same as described in Figure 11.

Figure 12. Sugar profiles produced from extrusion starch of different degrees of gelatinization. [:]; total sugar, O; maltose, V; maltooligosaccharides, A; glucose

100

] ..... "..........................

I

0

Degree of gelatinization, %

Degree of gelatinization, %

300

0( ~-

75

200

50 100

8

25

0

0 0

6

12

18

24

Reaction time, hr Figure 14. Progress of maltose formation reaction from extrusion starch during extended reactiontime. Reaction conditionswere the same as describedin Figure 11. [3; total sugar, O; maltose, V; maltooligosaccharides,A; glucose, ~; maltose content

183 750

800~ (A) Batch wise

(B) Fed-batch wise -[ r

500

7

or r

o 400 250

u

0 0

6

12

18

.0

I

I

I

6

12

18

0 24

Reaction time, hr Figure 15. Comparison of progress of maltose formation reaction at high extrusion starch concentration by batch wise(A) and fed-batch wise feedings(B). Batch conditions; 700 g of extrusion starch/1. Fed-batch conditions; the initial, second, third feedings of 200 g/l, and the forth feeding of 100 g/1 after 0,1, 2, and 3 hours, respectively. Arrow indicates the fedbatch feeding time. Other conditions were same as described in Figure 11. [-]; total sugar, O; maltose, V; maltooligosaccharides, A; glucose, m; residual starch

Table 3 Overall comparison of reaction parameters for maltose production reaction heterogeneous enzyme reaction using insoluble extrusion starch and conventional enzyme reaction using liquefied starch. Extrusion starch Maltose content(%) Maltooligosaccharides conent(%) Glucose content(%) Maltose concentration(g/l) Half reaction time(hr)* Overall productivity(g/1 9hr)

77 17 6 220 0.8 18.3

Liquefied starch 65 31 4 185 1.4 15.4

300 g of extrusion starch/l, 400 units of fungal a-amylase/g starch, 0.05 M acetate buffer(pH 5.5), 50 ~ 250 rpm, and after 12 hours. * Reaction time required for the half of maxium yield

184 Figure 16 shows the effect of extrusion conditions on the structural features of extrusion starch, such as, degree of gelatinization. The relationship between the structural features and maltose forming reaction was investigated, and the result was analyzed in terms of surface reaction of insoluble extrusion starch. The characteristics of maltose formation from swollen extrusion starch was compared using endo-type fungal a-amylase and exo-type B-amylase(Table 4).

Figure 16. Effect of equilibrium moisture content of raw corn starch and screw speed of extruder on the degree of gelatinization of extrusion starch.

Table 4 Effect of enzyme sources on maltose formation from insoluble extrusion starch and liquefied starch. Enzyme

Fungal a-amylase -Amylase

Extrusion starch

Liquefied Starch

G1

G2

~G3

G1

G2

~G3

6*

77

17

4

65

31

1

94

5

1

87

12

300 g of extrusion starch/l, 400 units of each enzyme/g starch, pH 5.5, 50"C, 250 rpm, and 12 hr. * Sugar content (%, w/w)

185

Transglycosylation of stevioside in heterogeneous enzyme reaction system utilizing extrusion starch as a glycosyl donor[22] Transglycosylation reaction of cyclodextrin glucanotransferase was carried out in the heterogeneous enzyme reaction system using extrusion starch as a donor; and mono-, di-saccharide, or glucoside as acceptor. Transglycosylation yield of various acceptors from extrusion starch was much higher than from liquefied starch, as shown in Table 5. Monosaccarides which has the same configuration of C2-, C3-, and C4-OH with D-glucopyranoside, such as glucose, xylose, and sorbose, were identified as a good acceptor for transglycosylation reaction of CGTase[23]. Meanwhile, transglycosylation yield of disaccharides was higher than that of monosaccharides, as observed by Nakamura et al. who reported that acceptor specificity of CGTase to maltose was much higher than that of glucose[24]. Various glucosides that have glucose molecules can be used as acceptor site for transglycosylation reaction of CGTase in their molecular structure, such as stevioside, hesperidin, and salicin, were also used as good acceptors for transglycosylation reaction of CGTase using extrusion starch as a donor.

Table 5 Comparison of transglycosylation yield of various acceptors in heterogeneous enzyme reaction system using extrusion starch and homogeneous enzyme reaction system using liquefied starch as glycosyl donor. Acceptors Monosaccharides Glucose Xylose Sorbose Inositol Disaccharides Maltose Sucrose Cellobiose Glucosides Stevioside Hesperidin Salicin

Extrusion starch

Liquefied starch

52.4* 29.8 50.0 37.0

32.1 25.9 41.8 25.0

82.0 65.0 76.0

69.8 61.0 66.9

81.0 63.4 68.9

65.8 52.2 54.3

50 g of starch/l, 50 g of each acceptor/1, 90 units of CGTase/g starch, pH 6.0, 50Y:, and 200 rpm. * Transglycosylation yield(%) after 24 hr

186 For transglycosylation reaction of stevioside, the transglycosylation rate was similar with and the transglycosylation yield was increased when compared with that of process using liquefied starch as the donor(Figure 17). Also the accumulation of maltooligosaccharides in reaction mixture was minimized. The residual insoluble starch can be easily separated out from reaction mixture, that will facilitate the purification of transglycosylated stevioside.

100 ~===~

75 o

50 c)

25

0

6

12

18

24

Reaction time, hr Figure 17. Comparison of transglycosylation of stevioside in the enzyme reaction system using raw(A), extrusion(O) and liquefied([:]) starch as glycosyl donors. 50 g of starch/l, 50 g of stevioside/1, 90 uints of CGTase/g starch, pH 6.0, 50~ and 200 rpm.

Optimal reaction conditions were determined to be 50~75 g of extrusion starch/l, amount of CGTase to extrusion starch of 90~ 120 units/g of starch, mixing ratio of stevioside to extrusion starch of 2:5(g of stevioside:g of starch). The transglycosylation of stevioside proceeded via two steps; initially C D synthesis from extrusion starch, and then followed by transyglycosylation of produced C D to stevioside. Table 6 compares the reaction parameters for transglycosylation reaction of stevioside in the heterogeneous reaction system using extrusion starch and conventional system using liquefied starch. The transglycosylation using extrusion starch as a glycosyl donor seems to have many potential advantages over conventional methods, hence, is expected to be utilized for industrial production of functional carbohydrates, such as coupling sugar and transglycosylated stevioside.

187

Table 6 Comparison of the heterogeneous reaction system using extrusion starch and conventional system using liquefied starch Transglycosylation yield(%) Concentration of transglycosylated stevioside(g/l) Half reaction time(hr)* Residual cyclodextrinconcentration(g/l) Residual maltooligosaccharides concentration(g/l) Separation of residual maltooligosaccharides Separation of residual starch

Extrusion starch

Liquefied starch

81.0 65.2

71.0 57.1

1.6 21.0 0.2

0.8 22.0 8.0

Not required

Required

Easy

Difficult

Easy Difficult Separation and purification of transglycosylated stevioside * Reaction time required for the half of m a x i m u m transglycosylation yield

Acknowledgement These works were supported by the 1991-1995 research grant to Research Center for New Bio-Materials in Agriculture from the Korea Science and Engineering Foundation.

References Kainuma K. In: Whistler RL, Bemiller JN, Paschall EF, eds. Starch: Chemistry and Technology. N e w York: Academic Press, 1984; 125-152. Szejtli J. In: Cyclodextrin Technology. Dordrecht: Kluwer Academic Publishers, 1988; 34-36. Reilly PJ. In: van Beynum GMA, Roels JA, eds. Starch Conversion Technology. N e w York: Marcel Dekker, 1985; 101-142. Horikoshi K, Nakamura M. U S Patent 1979; 4,135,977. Han IK, Lee YH. Kor J Appl Microbiol Biotechnol 1991; 19: 163-170. Hashimoto H, Hara K, Kuwahara N, Arakawa K. J Jpn Soc Starch Sci 1985; 32: 299-306. Lee YH. In: Proceedings of the Second Korea-US Joint Seminar on Bioprocess Technology. Seoul: KOSEF/NSF, 1991; 12-17.

188

10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Lee YH. In: Proceedings of the Second Korea-China Biotechnology Symposium. Seoul: RCNBMA, 1994; 42-56. Lee YH. In: Proceedings of the First International Symposium on the Development of Natural Resources and Environmental Preservation. Seoul: Korea University, 1992; 104-114. Wooton KW, Weeden C, Munk N. Food Technol 1975; 23: 612-613. Makelti MJ, Korpela TK, Puisto J, Laakso SV. J Agric Food Chem 1988; 36: 83-88. Miller GL. Anal Chem 1959; 31: 426-428. Lee YH, Park DC. Kor J Appl Microbiol Biotechnol 1991; 19: 514-520. Lee YH, Park DC. In: Furusaki S, Endo I, Matsuno R, eds. Biochemical Engineering for 2001. Tokyo: Springer-Verlag, 1992; 127-129. Lee YH, Park DC. Korean Patent 1993; 064852(B1-3187). Schmid G. Tibtech 1989; 7: 244-248. Kitahata S, Okada S. J Jpn Soc Starch Sci 1979; 26: 68-75. Lee YH, Cho MJ, Park DC. Kor J Appl Microbiol Bioeng 1995; 23: 416-424. Cho MJ, Park DC, Lee YH. Kor J Appl Microbiol Bioeng 1995; 23: 425-431. Lee YH, Kim DS, Shin HD, Park JS. Kor J Appl Microbiol Biotechnol 1994; 22: 106-113. Kim DS, Park DC, Cho MJ, Lee YH. Kor J Appl Microbiol Biotechnol 1994; 22: 283- 289. Lee YH, Shin HD, Park DC, Baek SG. Korean Patent 1995; 082729(B13722). Lee YH, Baek SG, Shin HD, Park DC. Kor J Appl Microbiol Biotechnol 1993; 21: 461-467. Nakamura A, Haga K, Yamane K. FEBS Lett 1994; 337: 66-70.


189

M a n i p u l a t i o n of S t o r a g e C o m p o u n d s in T r a n s g e n i c P l a n t s David M. Stalker, Kevin E. McBride and Christine K. Shewmaker Calgene, Inc., 1920 Fifth Street, Davis, CA 95616, USA

Introduction All plants capture CO2 as photosynthate. They use this captured carbon for metabolic and structural purposes and also store it for future use. The three major storage forms of carbon that the plant utilizes are carbohydrates, lipids and proteins. These compounds have wide utility for man both as food and for industrial uses. Genetic engineering of plants offers the ability to be able to modify these compounds for greater nutrition or altered industrial properties or to induce the plant to make more and thus increase yield. The possibilities for doing this with protein, particularly with a goal of improving n u t r i t i o n a l quality by a l t e r i n g amino acid balance, have been recently reviewed [1]. Below we will describe work that has been performed by us and our m a n y of our colleagues t h a t p e r t a i n s to e n g i n e e r i n g s t a r c h e s (carbohydrates) and lipids in plants. We also describe how recent innovations in plastid t r a n s f o r m a t i o n provide even greater opportunities for modifying these plant storage components.

Modification of Storage Carbohydrates The main targets for altered starch production in transgenic plants are shown below in Figure 1.

Starches

Increased yield

Altered amylose/amylopectin ratios

Altered branching and/or molecular weight

Novel carbohydrates

Figure 1. Targets for modification of reserve starch m a t e r i a l in transgenic plants. Starch is often stored in sink tissues such as tubers, roots or grains and our initial work was performed in potatoes as this offered an easily transformed species for which tuber-specific promoters were available.

190 Starch, which is a polymer of a-1-4 glucopyranose units with a-1-6 branch points is usually found in plants as either amylose or amylopectin. For a comprehensive review of starch biosynthesis see Shewmaker and Stalker [2]. The main difference between amylose and amylopectin is the degree of branching; amylopectin being highly branched while amylose contains mostly mainly long chains c a r b o h y d r a t e molecules. A d i a g r a m m a t i c representation of this structure is seen in Figure 2. H2OH

(~H2OH

CH2OH

CH2

CH2OH

1,4-1in~ge

Amylose

Amylopectin J ! ! |

i

i i Figure 2. Structure of a-1-4 and a-1-6 glucan linkages in starch, and a d i a g r a m m a t i c r e p r e s e n t a t i o n of possible s t r u c t u r e s of amylose and amylopectin.

191

The ratio of amylose to amylopectin and the degree of branching of the amylopectin can affect greatly the properties and thus utility of a starch. In an a t t e m p t to alter the b r a n c h i n g p a t t e r n of s t a r c h in potato t u b e r s we expressed a bacterial glycogen s y n t h a s e gene ( g l g A - EC2.41.21) in a tuberspecific m a n n e r and t a r g e t e d it to the a m y l o p l a s t s , the s t a r c h forming organelle of the potato [3]. Glycogen s y n t h a s e is the enzyme responsible for adding glucose units to a growing starch chain. When the t u b e r s produced from the construct (pCGN1457) described above were a n a l y z e d , the first i n d i c a t i o n s of an a l t e r a t i o n in s t a r c h biosynthesis came from a lowered specific gravity and a concomitant decrease in starch in the tuber. F u r t h e r examination (Table 1) showed a decrease in the a m o u n t of amylose and the production of a highly b r a n c h e d s t a r c h realtive to control samples. Comparison of the ratios of chain lengths in some of the transgenic potatoes to other known starches d e m o n s t r a t e t h a t a starch with a novel b r a n c h i n g p a t t e r n has been produced (Table 2). This novel starch was shown to have lower and more stable paste viscosities as well lower initial gelatinization t e m p e r a t u r e s t h a n normal potato starch [3]. The m e c h a n i s m by which expression of the bacterial glycogen s y n t h a s e gene leads to the observed effect is unclear but the findings of increased sugars and decreased starch provide some clues. Nevertheless, these results clearly d e m o n s t r a t e t h a t it is possible to modify the s t r u c t u r e of the starch t h a t is produced in a storage organ of a plant and open the possibilities of doing this in m a n y common starch producing crops such as maize, wheat, etc. Table 1 Analysis of starch branching in transgenic potatoes with g l g A .

Potato Line RB 43 control 57-4 57-17 57-18 57B-15

Percent Amylose 23 12 24 8 9

Chains HPLC B2 + B3% Longer Chains 33 20 28 15 15

Chains HPLC A + Bl% Shorter Chains 66 80 72 85 85

Ratio A + Bl% B2 + B3% 2.0 4.0 2.5 5.7 5.7

The designation A, B1, B2 and B3 refer to individual c a r b o h y d r a t e chains within the amylopectin molecules as revealed by enzymic digestion of starch and HPLC fractionation. These designations conform to the "cluster model" for amylopectin [11]. Fractions A and B1 to B3 are the A and B chains t h a t bind at the carbon-6 of the other chains through their reducing residues (Achains carry no chains and B-chains carry A- or other B-chains). The chains in fractions A and B1 m a k e a single cluster, and the chains in fractions B2 and B3 extend into two and t h r e e clusters, respectively. R e p r e s e n t a t i v e average chain lengths for fractions B1, B2 and B3 are in the ranges 20 to 24, 42 to 48 and 69 to 75, respectively. Average lengths for A-chains are 12 to 16 residues.

192 Table 2 Ratio of chain length comparisons of amylopectins in potatoes and other common starch crops.

glgA t r a n s g e n i c

Low MW/High MW c h a i n s 2 4 2.5 5.7 5.7

Potato-RB43 1457-4 1457-17 1457-18 1457-B-15

Potato* 1.9 Corn* 2.9 Waxy corn* 3.9 Amylomaize* 0.8 Rice* 2.6 Waxy rice* 3.6 Wheat* 3.6 Sweet potato* 2.5 * Values are obtained from Hizukuri [12]. In addition to altering the starch that is present, it should also be possible to produce a novel carbohydrate by modifying starch biosynthesis or using the starch t h a t is present as the substrate for a further reaction to produce a novel carbohydrate. An example of this is the work to produce cyclodextrins in potato tubers. Cyclodextrins are 6(a), 7(~), or 8(T) membered glucose rings t h a t are formed from starch via the enzyme cyclodextringlucosyltransferase (CGT EC2.4.1.19). The basic reaction that this enzyme catalyzes is illustrated in Figure 3.

o_e.--o-e--o-o-o---o,, o o

o o-o O

9

o~o 9 O go

~o o

CGT 9 ..

~o~

o\

. 0 O .O gO

o-o_.o..o_o..o.o-.o .'~

o-o-o-o-o-e-o-o-o 9 4,,_

O~)

,,o-e.

~"~'~

.~g

-.o-t

..-O,-.

T O~

+

,~-o,, .,

o--.o, --..-O'o--o--O"O"OO --

.

"o_e.o

Figure 3. Diagrammatic representation of the action of CGTs on starch. [2]

193 Cyclodextrins have m a n y uses, m a n y of which are based upon there ability to sequester compounds in their hydrophobic center. However, as the outer surface of the cyclodextrin molecule is quite hydrophilic, these sequestered molecules are still able to be solubilized in an aqueous solution. Cyclodextrins are thus finding m a n y uses as odor and flavor enhancers and also as pharmaceutical delivery systems. To engineer the production of these molecules in the potato tuber, we expressed a CGT from Klebsiella in a tuberspecific m a n n e r and targeted it to the amyloplast [4]. When tubers containing this construct were analyzed, small a m o u n t s of cyclodextrins (CDs) were found and the ratio of a/~ CDs was similar to that produced by the KlebsieUa enzyme in vitro [4]. Only small a m o u n t s of cyclodextrins were seen as analyzed by mass spectroscoy (not shown) and it would be desirable to be able to get more cyclodextrins in these tubers. A problem t h a t is frequently observed when expressing bacterial genes in plants is t h a t the codon usage is n o n - o p t i m a l for p l a n t n u c l e a r genes and the gene m a y need to be resynthesized with a codon usage t h a t is similar to t h a t of other nuclear genes in the plant. A second option, is to transform the chloroplast directly and thus be able to circumvent the need for gene resynthesis. This possibility of chloroplast transformation and its utility in altering storage compounds is discussed below.

Modification of Storage Lipids The oils, fats and waxes that many plants store in their seeds constitute a vast and renewable resource. The utility and importance of these plant lipids have been known for centuries; traditional uses such as the m a n u f a c t u r e of soaps and lubricants have increased in volume and diversity. Currently, plant lipids play major roles in a wide variety of applications [5]. For example, lauric acid, derived chiefly from coconut and palm kernel oils, is consumed in soap and detergent m a n u f a c t u r e (5 million tons/year). Castor oil (400,000 tons) h a s specialized use in h i g h - p e r f o r m a n c e l u b r i c a n t s and canola (rapeseed) oil is used directly as a cooking oil and in the m a n u f a c t u r e of margarine (9 million tons). Many decades of breeding work have resulted in impressive yield e n h a n c e m e n t s and other i m p r o v e m e n t s for m a n y oilseed crops. Within the past few years genetic engineering technology has emerged that complements the traditional plant breeding efforts. Many of the lipid compounds produced by wild species would have v a l u a b l e a p p l i c a t i o n s were it not for the difficulties associated w i t h domestication of those plants. By transferring the ability to produce these u n u s u a l lipids to the existing domesticated crop base one can avoid the considerable expense and effort needed to introduce undomesticated species into commercial agriculture. In the longer term there is also the potential to customize plant lipids to produce types and compositions seen in other living organisms, but not yet found in plants. To engineer a defined change in the composition of rapeseed oil, the identities of the gene (or genes) t h a t effect the change m u s t be known. U n d e r s t a n d i n g the metabolism of seed oil formation, at least in sufficient

194 detail to identify and purify the enzymes t h a t catalyze the reactions of most interest, is therefore essential. A considerable knowledge of the metabolic pathways by which seed lipids are synthesized has been gained over the past few decades. The f u n d a m e n t a l metabolic reactions comprising fatty acid biosynthesis require growing acyl groups to be covalently bound to an acyl-carrier protein (ACP). Subsequently, incorporation of these fatty acids into the triglyceride molecules comprising the oil requires t h a t they similarly be esterified to a nonprotein carrier molecule known as coenzyme A (CoA). Thus the gene isolated for manipulation of these pathways encode enzymes t h a t act either acyl-ACPs or acyl-CoAs. The first successful use of genetic engineering to alter the composition of rapeseed oil was d e m o n s t r a t e d [6]. The overall goal has been to alter the balance of s a t u r a t e d and u n s a t u r a t e d fatty acids in the stored triglycerides with a variety of applications by a t t e m p t i n g to reduce the n u m b e r of double bonds, thus producing a higher content of stearic acid (18:0) as opposed to lowering oleic acid (18:1) content by inhibiting the enzyme t h a t introduces the double bond into the s t e a r a t e molecule, stearoyl-ACP d e s a t u r a s e . Because subsequent desaturations depend on oleoyl-ACP formation, lowering the activity of this enzyme should theoretically bring about a reduction in the levels of all u n s a t u r a t e d 18-carbon fatty acids in the oil. The stearoyl-ACP desaturase enzyme was purified from safflower seeds and its sequence was used to obtain the safflower and rapeseed stearoyl-ACP d e s a t u r a s e genes. The gene from rapeseed was then used to transform rapeseed plants in the antisense m a n n e r under the control of a seed-specific promoter, and the seeds of the r e g e n e r a t e d , m a t u r e p l a n t s were e x a m i n e d for changes in oil composition. Dramatic increases were observed in the stearic acid content at the expense of the u n s a t u r a t e d fatty acids, as expected. The usual stearate level of 2% was increased to as much as 35% in this way (Figure 4), without any a p p a r e n t d e t r i m e n t a l effects to the appearance or g e r m i n a t i o n of the seeds. When these plants were field-grown the yield of oil was also normal. The development of commercial, h i g h - s a t u r a t e rapeseed varieties from these plants is now under way. As the first d e m o n s t r a t i o n of the alteration of rapeseed oil by genetic engineering this work is an i m p o r t a n t landmark. However, the fatty acids present in the oil of the genetically engineered rapeseed were still the same as those found in normal rapeseed oil. Only their proportions had been changed. A remaining question pertained to the flexibility of storage triglycerides, and w h e t h e r more d r a s t i c changes could be m a d e w i t h o u t affecting seed formation, oil yield and subsequent seed germination. Oils have 8:0, 10:0, 12:0 or 14:0 fatty acids are commonly referred to as medium-chain triglycerides (MCTs). Lauric acid (12:0) has immense value in the m a n u f a c t u r e of detergents while myristic acid (14:0) is also used in detergents, as well as in cosmetics and personal care applications. Lauric and myristic acids have traditionally been obtained from the tropical coconut and palm kernel oils, but there is considerable interest in producing them in temperate crops such as rapeseed to meet an ever increasing demand. The 8:0 and 10:0 acids are i m p o r t a n t constituents of biodegradable and specialty lubricants, certain foods and p h a r m a c e u t i c a l products. They also are

195 obtained from tropical oils, but because they are only minor components of the oils, they have been relatively scarce and expensive. A renewable, stable and economical source of these compounds could revolutionize the application of these fatty acids.

Figure 4. Comparison of control rapeseed seed oil profile with the seed oil profile from a rapeseed line t r a n s f o r m e d with a seed-specific a n t i s e n s e construct inhibiting the A9-desaturase. Fatty acid formation via the elongation of acyl-ACPs is t e r m i n a t e d by the action of the enzyme acyl-ACP thioesterase. By detaching the growing acyl chain from the ACP molecule, this enzyme effectively removes it from the biosynthetic pathway. In rapeseed the thioesterase is specific for long-chain acyl-ACPs; therefore the products of fatty acid biosynthesis are primarily 18carbon fatty acids. It was found that in some plant species whose seeds store MCTs there is an analogous enzyme that prefers to act on medium-chain acyl ACPs. Its action results in the exclusive formation of MCFAs. Theoretically the introduction of such an enzyme into rapeseed by gene transfer should result in medium-chain formation there as well. Such a thioesterase enzyme was isolated, and the corresponding gene obtained, from the seeds of a local tree known as California Bay. These seeds accumulate MCTs in which the fatty acids are 10:0 and 12:0; the t h i o e s t e r a s e obtained from t h e m is responsible for 12:0 (lauric acid) production. Introduction of this gene into rapeseed plants resulted in the formation of lauric acid in the seed oil [7]. Up to 45% of the rapeseed long-chain fatty acids were replaced by laurate in this way without any apparent ill effects on oil accumulation, seed germination or overall plant performance. The striking difference in fatty acid composition

196 versus controls can be seen in Figure 5. (The small amount of 14:0 also formed results from slight action of the bay thioesterase enzyme on 14:0ACP.)

Figure 5. Comparison of control rapeseed seed oil profile with the seed oil profile from a rapeeed line transformed with a seed-specific construct expressing the California Bay C12-specific thioesterase. A yet more ambitious undertaking would be to replace the triglycerides altogether with a different lipid class. The replacement of rapeseed triglycerides with "long-chain liquid wax" (LCLW), an ester of 20:1 and 22:1 fatty acids and the equivalent alcohols is now underway. This compound is currently obtained from seeds of jojoba, a shrub native to North American deserts and is an important ingredient of cosmetics, shampoo formulations and specialty lubricating oils, but its use has been limited because of the unsuitability of jojoba as a crop plant. LCLW is also similar to sperm whale oil and the reproduction of LCLW in rapeseed could substitute for whale oil and thereby rejuvenate and greatly expand these markets. It could also lead to new applications t h a t would take a d v a n t a g e of LCLW's unique characteristics. Considerable progress in this direction by cloning the gene responsible for long-chain alcohol formation in jojoba seeds. Expression of this gene in rapeseed plants resulted in production of the expected fatty alcohol. These examples illustrate the potential for dramatic modification of rapeseed oil by using the techniques of genetic engineering. Many other seed

197 oil modifications are envisaged with this technology, which also can be applied to soybean, sunflower, cotton and other crops. Plastid Transformation as a P o t e n t i a l Storage Compound Modification.

New

Technology

for Plant

The plastids of higher plants are an extremely attractive target for genetic engineering. Plant plastids (chloroplasts, amyloplasts, chromoplasts, etc.) are the major biosynthetic and photosynthetic centers of the cell, responsible for production of i m p o r t a n t compounds such as amino acids, complex carbohydrates, fatty acids and pigments. Various plastid types are derived from proplastids located in meristematic cells and thus have the same genetic content. Stable transformation of plastid genomes has been achieved in higher plants [8] through homologous recombination of a selectable marker delivered homologous recombination of a selectable marker delivered to the plastid by particle-gun bombardment. Plant cells have been found to contain up to 50,000 copies of the chloroplast genome. This makes it possible by plastid transformation to engineer plant cells to maintain any introduced gene at an extremely high copy number, potentially resulting in a very high level of foreign gene expression. DNA sequence and biochemical data reveal a striking similarity of the plastid organelle's transcriptional and translational machinery and cognate initiation signals to those found in prokaryotic systems. In addition, plastid genes are often organized into polycistronic operons as are genes in prokaryotes. A major drawback in the engineering of plastid gene expression is the lack of tissue-specific developmentally regulated control mechanisms. It is possible t h a t unregulated modification of plastid metabolism and/or the introduction of new biochemical pathways could result in the inability to obtain viable plants. One way to address this problem would be to specifically trans-activate a silent plastid-borne transgene by tissue-specific expression of a nuclear-encoded and plastid-directed RNA polymerase (RNAP). The polymerase chosen for this purpose should have a high degree of specificity for the promoter element associated with the plastid transgene. To establish such a system, a ~-glucuronidase (GUS) reporter gene under control of the phage T7 gene 10 promoter was introduced into the plastid genome of tobacco. GUS expression was now dependent on nuclear-encoded plastid-targeted T7 RNAP activity [9]. To this end a binary vector, pCGN4026, harboring a T7 RNAP chimeric gene with signal sequences for plastid targeting was constructed. The chimeric T7 RNAP gene, lacking its ATG start codon, is expressed from an enhanced CaMV 35S promoter as a translational fusion to the tobacco SSU transit peptide and the first 12 amino acids of mature SSU. Due to the approximately constitutive nature of the 35S promoter, T7 RNAP activity was expected to be present to some degree in plastids of most plant tissues. Twenty-one kanamycin-resistant pCGN4026 tobacco lines were generated by Agrobacterium-mediated transformation and T7 RNAP activity was detected in the leaf tissue from 10 of the primary transformants.

198 Introduction of a m a r k e r gene into the plastid genome was carried out to test for trans-activation by the plastid-localized T7 RNAP activity. The GUS gene (uidA locus of E. coli) was chosen as it codes for an enzyme t h a t is a simple biochemical m a r k e r for p l a n t s and t h a t h a s been shown to be expressed in the plastid under the control of a psbA promoter and 3' region [10]. The GUS gene coding sequence is located downstream from the E. coli phage T7 gene 10 promoter and 5' u n t r a n s l a t e d regions. The chimeric GUS m a r k e r was introduced into the tobacco plastic homology vector pOVZ44B and the r e s u l t i n g plasmid was used to t r a n s f o r m plastids of tobacco by p a r t i c l e - g u n d e l i v e r y of DNA-coated t u n g s t e n microprojectiles. Two i n d e p e n d e n t l y isolated, fertile, t r a n s p l a s t o m i c lines in the T7 RNAPproducing background (4026-3) were g e n e r a t e d in addition to t h r e e fertile, t r a n s p l a s t o m i c lines in the tobacco 'Xanthi' control background (containing no T7 RNAP activity). These tobacco lines were judged to be homoplasmic by Southern blot analysis. To d e m o n s t r a t e t h a t T7/GUS t r a n s c r i p t s c o n t a i n i n g this specialized prokaryotic u n t r a n s l a t e d leader sequence could be t r a n s l a t e d in plastids, GUS specific activity was m e a s u r e d in various tissues (Table 3). Table 3 GUS activity and mRNA levels in transplastomic tobacco [9].

Tissue

Specific Activity* (nmol/min per m6)

4276/4026-3

"4276~anthi

Relative mRNA level 4276/4026t

M a t u r e leaf 2.0 x 105 0 240 Young leaf 1167 0 7 Stem 13 0 7 Root 3 0 1 Petal 497 0 24 Seed 8 0 (1) *4-methylumbelliferone (nmol) produced/minute/mg of total soluble protein. Values are the average of three replicate assays. t Values are the average of two dot blot assays and one N o r t h e r n blot assay nomalized to the value (designated as 1) obtained for seed RNA. These data show t h a t GUS activity is present in all tissues tested, although its level of accumulation varied by >10,000-fold. To determine w h e t h e r the differences in GUS activity correlated with variations of GUS mRNA levels, quantitative N o r t h e r n blot and RNA dot blot assays were performed on total RNA. The q u a n t i t a t i v e results indicated t h a t there was a very high level of GUS mRNA in m a t u r e leaf, corresponding to the high level of GUS mRNA in m a t u r e leaf, corresponding to the high GUS activity value. The next highest level of GUS mRNA was observed in petal tissue, which exhibited the highest level of GUS activity for a non-leaf tissue. E n z y m e activity was lowest in roots and seeds, which accumulated the least a m o u n t of GUS mRNA. The

199 extremely high level of GUS activity observed in mature leaves compared to young leaves and other tissues can be explained not only by a high level of GUS mRNA but also by the fact that mature leaf chloroplasts exhibit an enhanced translational capacity and ability to accumulate proteins relative to other plastid types. These experiments demonstrated the manipulation of plastid transgene expression via the action of nuclear-encoded and plastidtargeted T7 RNAP. The plastid-borne GUS reporter gene, under control of the phage T7 gene 10 promoter and 5' untranslated region, was expressed in all plastid types examined when incorporated into a tobacco line containing an active T7 RNAP gene. We have established that a plastid-encoded monocistronic mRNA can be synthesized by T7 RNAP and that this message is fully capably of being t r a n s l a t e d in plastids of different tissues within the developing plants. Plants can now be designed to express T7 RNAP in a tissue- or inductionspecific m a n n e r limited expression of the plastid transgene appropriately. This may allow modification of plastid metabolism in selected plastid/tissue types without altering normal plant growth and development. The next goal will be to introduce into plastids polycistronic operons encoding multiple traits and/or novel complex biochemical pathways. Operons with the potential for modifying plant storage components already exist in bacteria and could possible by inserted into the plastid genome under the control of the T7 promoter without further modification. References

1 2 3 4 5 6 7 8 9 10 11 12

Habben, J.E. and B.A. Larkins. 1995. Current Opinions in Biotechnology 6" 171-174. Shewmaker, C.K. and D.M. Stalker. 1992. Plant Physiol. 100: 1083-1086. Shewmaker, C.K., Boyer, C.D., Wiesenborn, D.P., Thompson, D.B., Boersig, M.R., Oakes, J.V., and D.M. Stalker. 1994. Plant Physiol. 104: 1159-1166. Oakes, J.V., Shewmaker, C.K. and D.M. Stalker. 1991. Bio/Technology 9: 982-986. Salunke, D.K., Chavan, J.K., Adsule, R.N. and S.S. Kadam. 1991. Van Nostrand Reinhold" New York, 1991. Knutzon, D.S., Thompson, G.A., Radke, S.E., Johnson, W.B., Knauf, V.C. and J.C. Kridl. 1992. Proc Natl Acad Sci USA 89: 2624-2628. Voelker, T.A., Worrell, A.C., Anderson, L., Bleibaum, J., Fan, C., Hawkins, D.J., Radke, S.E. and H.M. Davies. 1992. Science 257: 72-74. Svab, Z., Hajdukiewicz, P. and P. Maliga. 1990. Proc Natl Acad Sci USA 87: 8526-8530. McBride, K.E., Schaaf, D.J., Daley, M. and D.M. Stalker. 1994. Proc Natl Acad Sci USA, 91: 7301-7305. Staub, J.M. and P. Maliga. 1993. EMBO Jour 12: 601-606. Hizukuri, S. 1986. Carbohydrate Res. 147: 342-350. Hizukuri, S. 1985. Carbohydrate Res. 141:295-306



201

Overproduction of Bacterial Amylases in Recombinant Escherichia coli Systems Jin-Ho Seo, Woo-Jong Lee, Myoung-Dong Kim, Chang-Sup Kim and Yong-Chul Park D e p a r t m e n t of Food Science and Technology Research Center for New Bio-Materials in Agriculture, Seoul National Univisity, Suwon, 441-744, KOREA.

Abstract The research was undertaken to produce industrial enzymes using Bacillus licheniformis maltogenic recombinant Escherichia coli cells: amylase(BLMA) used for production of branched oligosaccharides, B. macerans cyclodextrin glucanotransferase(CGTase) for c~-cyclodextrin production and B. subtilis amylase. Fed-batch cultures were employed to improve the expression yield of gene products, as they allow separation of a cloned-gene expression state from a cell growth phase. The optimum operation of fed-batch cultures for the BLMA expression system yielded the maximum BLMA activity of 85.4 U/mL and the final cell mass of 56.2 g/L by keeping acetate concentration below 2.7 g/L throughout the fermentation. These numbers correspond to a 15.5-fold increase in BLMA activity and a 12-fold enhancement in cell mass compared with the simple batch fermentation. BL21(DE3)pLysE was chosen as a host for the T7 promoter-mediated expression of CGTase. The CGTase expression system in the optimized fed-batch fermentation resulted in 62.9 U/mL of CGTase activity and 53.5 g/L of cell mass. CGTase was accumulated in the cell as inclusion body.

Introduction Starch has been widely used in the food and beverage industries. Starch consists of a mixture of linear and branched homopolymers of D-glucose. The first step in utilizing starch as a raw material is to convert it to low molecular weight derivatives. Conversion of starch is catalyzed by many different enzymes [1]. Enzymatic conversion of starch has advantages over traditional acid hydrolysis in terms of higher yields and specificity of products and lower energy consumption. Genetic engineering will make possible quantities of such enzymes to be produced at lower prices and improve process stability and specificity by protein engineering. In order to overproduce industrial enzymes by recombinant DNA technology, a systematic strategy has to be employed as shown in Fig. 1. The selection of a proper cloning system is a key element for the high productivity of a recombinant DNA system. In the design of expression vectors at a gene level, a number of factors have to be considered including a promoter for gene expression, a signal sequence for protein secretion and genetic characteristics of a host cell [2].

202

[ i:,~si~,, PR01:,~,cTI GENES TRANSPORT SIGNAL

PROMOTER

PLASMID

c.~o~oso~

MICROBIAL CELLS

STRUCTURAL (PRODUCT) GENE

~c~o,.~o~

HOSTS [ MAMMALLAN CELLS

J

PLANT CELLS

1 BIOREACTORS BATCH

FED-BATCH

CONTINUOUS

BIOREACTOR OPERATION OPTIMIZATION SELECTIVE PRESSURES

REGULATION OF GROWTH, PROMOTOR GENE REPLICATION

MEDIA COMPOSITION

PRODUCT EXPRESSION & TRANSPORT SECRETED EXTRACELLULAR

CYTOPLASMIC INTRACELLULAR

EXPORTED PERIPLASMIC /

REFOLDING ,,,

1

q/

EF PRooucT I Figure 1.

Process flow chart for a recombinant DNA product.

203 After a cloning system is selected, the next step is to find optimal operating conditions for the chosen system. Operating parameters considered at a process scale include reactor configurations, operation modes (batch, fed-batch or continuous) and medium composition [3]. Escherichia coli has been a popular vehicle as a host for expression of foreign proteins in microbial cells. Escherichia coli has many advantages relative to other hosts such as Bacillus subtilis and Saccharomyces cerevisiae. Remarkable progress has been made toward the understanding of protein folding and secretion in recombinant E. coli cells [4]. Furthermore, E. coli can readily be grown to high cell density and this has led to the development of inexpensive, high-yielding fermentation processes for the production of many enzymes in an industrial scale [5]. This paper summarizes research efforts on production of industrial enzymes in recombinant E. coli systems. In particular, expression of Bacillus subtilis amylase, B. licheniformis maltogenic amylase (BLMA), and B. macerans cyclodextrin glucanotransferase (CGTase) will be discussed with respect to the expression patterns, medium optimization, and high cell density cultures in fed-batch fermentors.

Expression gene

and

secretion

of Bacillus

subtilis

~-amylase

by phoA

The structural gene of the Bacillus subtilis c~-amylase (amyE) was inserted into a plasmid pAMYA behind the E. coli phoA gene in such a way that its expression and secretion to the periplasmic space was controlled by inorganic phosphate concentration in the medium [6]. Recombinant E. coli BW3414 containing the plasmid pAMYA was grown in the MOPS minimal medium containing 100mg/L of ampicillin as selection pressure. About 95% of total c~-amylase activity was found in the periplasmic space of the cells, indicating successful secretion into the periplasmic space of the E. coli cell. The expression pattern of the phoA-amyE fusion gene on the multicopy plasmid was compared with that of the phoA gene coding for alkaline phosphatase (AP) in the chromosome. The expression and secretion of both AP and a-amylase was directed by the same promoter and signal sequence of the E. coli phoA gene, but it was located at different cloning sites. The dependence of inorganic phosphate (Pi), glucose, optical density (OD), specific activity of AP, a-amylase, and ~-lactamase on fermentation time for an initial phosphate concentration of 0.1 mM at 37~ is displayed in Fig. 2. After inorganic phosphate in the medium was exhausted in 3.0 hours, the expression of both AP and a-amylase was initiated accompanying with a slow increase in cell mass. The same type of dependency of cell growth and enzyme expression on inorganic phosphate concentration was also observed in other phoA-directed expression systems [7]. The growth rate of this a-amylase expression system after phosphate starvation, however, was far lower t h a n that of the recombinant strain (BW13704) which produces ~-galactosidase from the phoA-lacZ fusion gene on the multicopy plasmid

204 2400

~,

~~-o-o_.______~ o_______ -

o= 1800

~:

a. ~ .=

1200

~

600

~5-. ~

~

~cr ~

0.6

.E g

~ a

"~ ~ ' 0 . 4

o

~0.2

o r

~.

o

0~

10

"

0.15

~

F2.4

_.

.m

O

10 -~ o

0.0

i

i

6

12

18

Fermentation

Tlme

24

0.0

30

(hourB)

Figure 2. Time trajectories of optical density (G), glucose concentration (v), inorganic phosphate concentration (A), specific alkaline phosphatase activity, specific a-amylase activity, and specific ~-lactamase activity for the E. coli strain BW3414/pAMYA grown in a MOPS medium with 0.1mM initial inorganic phosphate at 37~ and pH 7.4. The specific enzyme activity was defined as a total activity divided by the culture OD. (plasmid pDK110). The specific growth rates of the a-amylase producing (BW3414/pAMYA) and B-galactosidase producing strains (BW13704) after phosphate depletion were estimated as 0.075 h r " a n d 0.28 hr 1, respectively. Even though the copy number of the plasmid pAMYA is approximately 10 times less than that of pDK110, the expression and secretion of c~-amylase showed more deleterious effects on the cell growth compared with the expression of ~-galactosidase which is a cytoplasmic enzyme. The BW3414/pAMYA strain stopped growing in one or two hours after phosphate depletion. Although the reasons are not clear, it seems that the

205 overexpression of the secreted protein ~-amylase may greatly disturb the secretion of the other membrane proteins which are necessary for cell divisions, causing the cessation of cell growth. The expression modes of alkaline phosphatase (AP) and c~-amylase under derepressed conditions were investigated in more details in order to explore growth inhibition phenomena. The expression of both enzyme was controlled by the same promoter; however, their expression patterns are quite different from each other. As soon as inorganic phosphate concentrations in the medium fell to zero, both enzymes were expressed at the maximum rate for 3.5 hours. After this period, the AP expression rate decreased gradually and the maximum specific AP activity was obtained in 4.5 hours. On the other hand, the a-amylase expression continued for 12.5 hours with rather a reduced rate compared to the initial rate. A possible reason for this difference may be the competition for secretion sites between the two enzymes. Both enzymes were expressed in the cytoplasmic space and then secreted through the cytoplasmic membrane. As the number of the secretion sites at the cytoplasmic membrane is constant [8], the overexpression of a-amylase seems to interfere with the secretion of other secreted proteins, including AP. The effect of the competition for the secretion sites on the enzyme activity can also be observed for ~-lactamase, another secreted protein. The specific activity profile of ~-lactamase is similar in shape to that of AP as if ~-lactamase were also under the control of inorganic phosphate concentration in the medium. However, when the total ~-lactamase activity in the culture is plotted against fermentation time, it increases at a constant rate for an initial period of six hours regardless of inorganic phosphate concentrations. This is consistent with the fact that ~-lactamase is expressed constitutively. The specific activity of ~-lactamase reached a maximum value at around 12 hours of fermentation. The expression rate for ~-lactamase changed almost at the same time when the expression rate of both enzymes (AP and a-amylase) changed, suggesting that the competition for the secretion sites between the secreted proteins became severe after 6.5 hours of fermentation. Saturation of the secretion pathway by the overexpression of a secreted protein has also been reported by many research groups. The experimental data displayed in Fig. 2 were analyzed by formulating mathematical equations for cell mass, glucose, inorganic phosphate, c~-amylase (gene product) and alkaline phosphatase. The model equations successfully described cell growth and a-amylase expression in response to inorganic phosphate in recombinant E. c o l i with the phoA-directed expression systems [9].

Production of Bacillus licheniformis maltogenic amylase (BLMA) Bacillus licheniformis maltogenic amylase (BLMA) has distinctive biochemical properties capable of producing branched oligosaccharides from starch [10]. Branched oligosaccharides or iosmaltooligosaccharides are widely used in the food industry as they present a number of advantages including low viscosity, less sweet taste than sugar, Bifidus factor and high

206 m o i s t u r e - r e t a i n i n g capacity [11]. The gene encoding BLMA was cloned and inserted into a plasmid pIJ322 in the Laboratory of Prof. K.H. Park, Seoul National Univisity. Recombinant E. coli TG1 containing plasmid pIJ322 was first grown batchwise to determine optimum growth conditions for m a x i m u m BLMA production. Optimization of m e d i u m composition requires a large n u m b e r of experiments. In a traditional method, a single factor is varied, while others are kept constant. Then another factor is selected for the next set of experiment. This one-factor-at-a-time method was shown to be inefficient and it lacks the abilitiy to detect interactions among factors [12, 13]. The Box and Wilson method was employed to perform m e d i u m optimization in a more efficient and faster manner. The new method, which involves simultaneously varying several factors in a planned fashion, allows precise estimation of the effects and interactions of several factors in a few n u m b e r

, 24

j~ !~ j

6L

I

I o

v

9

~

-"9

~

4

Enzymes for Carbohydrate Engineering

Enzymes for Carbohydrate Engineering

Handbook of Carbohydrate Engineering

Handbook of Carbohydrate Engineering

Handbook of Carbohydrate Engineering

Carbohydrate-active enzymes: structure, function and applications

Enzymes

Carbohydrate Sulphates

Neurotransmitter Enzymes

Enzymes Everywhere

Artificial Enzymes

Carbohydrate Bioengineering

Carbohydrate Chemistry

Understanding enzymes

The Enzymes

Advances in Biochemical Engineering, Volume 010 : Immobilized Enzymes I

Enzymes in Fruit and Vegetable Processing: Chemistry and Engineering Applications

Carbohydrate Chemistry v.23

Carbohydrate Bioengineering Interdisciplinary Apporaches

Drug Metabolism Enzymes

Food Related Enzymes

Enzymes in Food Technology

Quinones and Quinone Enzymes

Enzymes in Food Technology

Enzymes in detergency

Enzymes in Detergency

Quinones and Quinone Enzymes

Textile Processing with Enzymes

Enzymes in Industry

Efficient Carbohydrate Synthesis

Enzymes for Carbohydrate Engineering