Eur. J. Biochem. 219, 261-266 (1994) 0 FEBS 1994

Alteration of the cleavage mode and of the transglycosylation reactions of the xylanase A of Streptomyces Zividans 1326 by site-directed mutagenesis of the Asnl73 residue Alain MOREAU, FranGois SHARECK, Dieter KLUEPFEL and Rolf MOROSOLI Centre de Recherche en Microbiologie AppliquCe, Institut Armand-Frappier, UniversitC du QuCbec, Laval, Canada

(Received August 18lOctober 1, 1993) - EJB 93 1254/4

The amino acid replacement of Asn173 by Asp in the xylanase A (Xln A) of Streptomyces lividans significantly altered its enzymic properties. A time-course hydrolysis of xylan showed that the altered xylanase ([N173D] Xln A) initially produced larger amounts of xylose (X,), xylobiose (X,) and xylotriose (X,) than Xln A, but less xylotetraose (X). The bond-cleavage frequencies were determined for both enzymes using xylopentaose (Xs), xylotetraose (&) and xylotriose (X,) labelled at the reducing end of the molecule. Xln A hydrolysed X5, yielding 56% X, and 44% X3, while [N173D]Xln A liberated 90% X, and only 10% X,. Both enzymes hydrolysed X, into 100% X, and X, into 100% XI. Transglycosylation reactions were detected in HPLC hydrolysis patterns using high substrate concentrations, where larger products than the starting substrates were formed. Their subsequent degradation also affected the yield of hydrolysis products. Using X, as substrate, prod- ucts from xylohexaose (X,) up to xylooligosides larger than xylooctaose (X,) were synthesized by Xln A, while [N173D]Xln A produced only a small amount of xyloheptaose (X,) and X,. Xln A hydrolysed X, into an equivalent amount of X, and X, and 1.5-fold more X,. However, [N173D]Xln A yielded the same amount of X, and X, but half as much X,. With X, as substrate, Xln A synthesized twofold more X, and X, than [N173D]Xln A. Xln A liberated 1.4-fold more X, than X,, while [N173D]Xln A yielded twofold more X, than X,. Xln A liberated almost fourfold more X, than X, from X3, while “173DIXln A produced only twofold more X, than X,. These results indicated that the negative charge introduced by the mutation greatly affected the transglycosylation reactions catalysed by this xylanase.

Streptomyces lividans 1326 produces three different xyla- nases which have been cloned and characterized over the last few years [l-31. The xylanase A belongs to the group of high MJacidic xylanases, while the two others, xylanases B and C, are representative of the low MJbasic xylanases [4]. According to hydrophobic-cluster analysis, the xylanase A belongs to the F family, based upon sequence homologies, while xylanases B and C belong to the G family [5 - 81. Crys- tallographic data from the xylanase of Bacillus pumilus IPO have been used in modelling the three-dimensional structure of the catalytic domain [9]. However, the data only pertain to the G family. In the absence of structural information for the F family xylanases, a site-directed mutagenesis pro- gramme was undertaken to modify highly conserved amino acid residues in the xylanase A of S. lividans 1326 [5, 81. It has previously been shown that both families of xylanases possess some similarities, including two Glu residues in their

Correspondence to R. Morosoli, Centre de Recherche en Micro- biologie Appliquee, Institut Armand-Frappier, Universitk du Quk- bec, 531 boul. des Prairies, Laval, Qubbec, Canada, H7N 423

Abbreviations. xlnA, gene encoding xylanase A ; Xln A, wild- type xylanase A from Streptomyces lividans; [N173D]Xln A, Xln A mutated at position 173 where the asparagine residue was replaced by an aspartic acid; XI, xylose; X,, xylobiose; X,, xylotriose; X,, xylotetraose ; Xs, xylopentaose; X,, xylohexaose; X,, xyloheptaose ; X,, xylooctaose.

Enzyme. Xln A (EC 3.2.1.8).

active sites (unpublished data). Very little is known about the catalytic mode of degradation of xylan by xylanases. Based on previous work on the xylanase of Cryptococcus albidus (family F ) [lo, 111, and recently on that of Schizophyllum commune (family G ) [12], it appeared that the enzymes pref- erentially attack the second and the third xylosidic linkage from the reducing end of xylopentaose. It was deduced from these experiments that the xylanases contained several bind- ing subsites for the xylooligosides. Xylanases (family G), at high substrate concentration, also possess transglycosylation activities which lead to products larger than the starting sub- strates [lo, 111. These reactions are far from negligible, and make it almost impossible to determine kinetic parameters using xylooligosides with a low degree of polymerization.

In this paper, we described the modification of the bond- cleavage frequencies of the xylanase A generated by the mu- tation N173D and its influence on transglycosylation reac- tions.

MATERIALS AND METHODS Bacterial strains and plasmids

Escherichiu coli strain CJ236 dut-1, ung-1, thi-1, relA-1, pCJ105 (chloramphenicol resistant) harbouring plasmid pIAF217 was used to prepare uracil-containing single- stranded DNA for mutagenesis experiments. The vector

262

pIAF217, a derivative of pIAF19.1 (unpublished results), was also used for the expression of the xylanase A gene (xlnA) in E. cali strain MC1061 uruD139, d(ara-leu) 7679, Alac X74, gaEU, gall(, hsr, hsm, strA. S. lividans 10-164, a xylanase-negative and cellulase-negative mutant used as host for the production of the wild-type and N173D xylanases, has been described elsewhere [13]. Plasmid pIAFll was gen- erated by a deletion of a 3.2-kb Sphl fragment from pIAF18.

DNA manipulations and transformations Plasmid isolation, protoplast preparation and transforma-

tion of S. lividuns 10-164 were as described previously [14]. Transformation of E. coli and isolation of plasmid DNA were as described by Sambrook et al. [15]. Restriction and DNA- modifying enzymes were purchased from Pharmacia and used according to the specifications of the manufacturer.

Site-directed mutagenesis

Mutagenesis was carried out by the method of Kunkel [ 161 using a commercial kit (Bio-Rad). The oligonucleotide for mutagenesis was synthesized on a Gene Assembler (Phar- macia). The primer used to mutate the Asn173 codon was the following : S'-GTTCTCGACGTCGTAGTCGTT-3' (the substituted nucleotide is underlined). Plasmid DNA was iso- lated from presumptive mutants and hybridized with the 32P- labelled mutagenic primer in 6X NaClfCit (52.6 g NaCl and 26.5 g sodium citrateAiter, pH 7.0) overnight at 42"C, then the dot blots were washed twice for 20min in a 3-M tetramethylammonium chloride solution at 61 "C [17]. Subse- quently, the mutated gene was sequenced using the dideoxy- chain-termination method with [35S]dATP[S] and T7 DNA polymerase according to the protocol of the manufacturer (Pharmacia).

Enzyme production and purification The production of [N173D]Xln A was with the mutant

strain S. lividuns 10-164 as host. The wild'type xylanase gene was replaced, in pIAFl8, by its mutated counterpart as follows : the Notl -EcoRl restriction fragment containing the mutation in pIAF217/N173D was purified and subcloned into pIAF19.1 deleted from its Notl -EcoRl wild-type DNA fragment. After digestion, the resulting 3.2-kb Sphl fragment containing the mutation was cloned into pIAFll to reconsti- tute pIAF18N173D [13]. The purification of the mutant N173D xylanase was performed as described previously for the wild-type Xln A [l].

Xylanase assays Xylanase activity and Michaelis-Menten parameters were

determined by the reducing-sugar method [l]. Xylan concen- trations of 0.5-16 mg ml-' were employed at a temperature of 60°C. One unit (U) was defined as the amount of enzyme releasing 1 pmol min-' reducing sugar. Hydrolysis of birch- wood xylan was performed in a volume of 1 ml containing 10 m g h - ' birchwood xylan (Sigma) in 50 mh4 sodium cit- rate, pH 6.0, at 60"C, with 20 pg enzyme. Aliquots (50 pI) were withdrawn at various time intervals and the reaction was stopped by heating for 10 min in a boiling-water bath, after which the samples were filtered through an AcroTM fil- ter (0.45 pm pore size) and analyzed by HPLC [2]. Hydroly- sis patterns of X,, X, and X, (X,, xylotriose; X,, xylotet-

u -

0 10 20 30 40 Time (min)

Fig. 1. Time course of birchwood-xylan hydrolysis obtained with Xln A (A) and [N173D]Xln A (B). Concentration of substrate, 1% enzyme concentration, 20 pg I&'. X, (A); X, (0); X, (0); X, (W).

Table 1. Kinetic parameters for xylan hydrolysis by Xln A and [N173D]Xln A.

min-' mg ml-' minP (mg rnI-')-' Xln A 33 927 1.41 24 231 [N173D]XlnA 49 106 3.54 13 872

raose; X,, xylopentaose) were determined by HPLC. The concentration of the xylooligosides was 10 mM and 400 ng enzyme ml-' was added for the hydrolysis of X, and X,. Ten-times more enzyme was used to degrade X,. The reac- tion was allowed to proceed for 10 min at 60°C. The xylooli- gosides used for column-size calibration were obtained from Megazyme.

Determination of bond-cleavage frequencies Bond-cleavage frequencies were determined for Xln A

and [N173D]Xln A according to the method of Biely et al. [lo, 111. The reaction mixtures (10 pl) contained 5 ng xyla- nase, 0.1 mh4 cold substrate supplemented with 2x10, cpm 'H-labelled X3, X, and X, (labelled at the reducing end) in 20 mM sodium citrate, pH 6.0. Tritiated compounds were a gift from P. Biely (Bratislava, Slovakia). The mixtures were incubated at 50°C. At various time intervals, aliquots (1.5 pl) were withdrawn, spotted on cellulose thin layers (Merck) and chromatographed in ethyl acetatefacetic acidfwater (3 : 2 :2, by vol.). The chromatograms were excised and the radio- activity was counted in a liquid-scintillation counter.

RESULTS AND DISCUSSION The mutagenesis experiment was carried out on the X l n A

in E. coli as described earlier (unpublished data). After selec- tion of the mutant and piasmid isolation, the whole gene was sequenced. The sequence of the gene was identical to the wild-type gene except for the introduced mutation. In order to improve xylanase production by the mutant and to com-

263

0 10 20 30 40 50 60 0 10 20 30 40 50 60 70 80 Time (min)

Fig.2. Time course of digestion of [l-3H]xylooligosides (0.1 mM) from XlnA and [N173D]Xln A. (A) [l-3H]Xylopentaose (0); (B) (l-3H]xylotetraose (W); (C) [l-3H]xylotriose (0). Products: X, (A); X, (0); X, (0).

pare its properties to those of the native enzyme, the mutated gene was inserted into plasmid pIAFll [13] by replacement of the wild-type structural gene, and was then used to trans- form S. lividuns 10-164 (xylanase-negative and cellulase- negative mutant). After transformation, the presence of the mutation on the plasmid was confirmed by hybridization with the labelled oligonucleotide primer used to generate the mutation. S. Zividuns IAF18/N173D containing the N173D mutation secreted about 300 U ml-' xylanase, instead of 1.5 U ml-' for the E. coEi IAF217N173D clone. The [N173D]Xln A was purified to homogeneity as described be- fore [ l ] enabling us to determine its enzymic and biochemi- cal properties. The behaviour of the mutated enzyme was identical to that of the wild-type xylanase during the purifica- tion procedure.

The substitution of Asn173 to Asp generated an enzyme which had a specific activity 25% higher than the native xy- lanase A, using birchwood xylan as the substrate, and had the same temperature (60°C) and pH (6.0) optima. Interestingly, different xylan-hydrolysis patterns were observed (Fig. 1). [N173D]Xln A produced initially larger amounts of X,, X, and X, (X,, xylose; X2, xylobiose; X,, xylotriose) but less X, than the wild-type enzyme. Furthermore, no intermediary product larger than X, was observed in the HPLC pattern obtained with [N173D]Xln A, while Xln A produced xylooli- gosides up to X, (xylooctaose; data not shown). There was also a modulation of kinetic parameters of the altered enzyme as compared to those of Xln A, as shown in Table 1. [N173D]Xln A had less affinity for xylan and, notwithstand- ing a higher k,,,, the specificity constant was inferior to that of Xln A. These results are difficult to interpret as it is known that the color yield in the reducing sugar assay is chain- length dependent [18]. As shown by xylan hydrolysis, Xln A produced larger xylooligosides than [N173D]Xln A and this could easily interfere with the method. The effect of this mutation was so obvious in the production of X, that a major change in the substrate-cleavage mechanism was suspected. This hypothesis was further verified by the hydrolysis of [3H]xylooligosides labelled at the reducing end (Fig. 2).

Xln A Xln N173D

x.4 1 .o

0.10 0.90

Fig. 3. Bond-cleavage frequency distribution. The values indicate the bond-cleavage frequencies between D-XylOSe units linked by /3-1,4 linkages. The black circle represents the reducing end of xy- looligosides labelled with 3H. Xln N173D, [N173D]Xln A.

These radioactive substrates were used to accurately deter- mine the bond-cleavage frequencies of the xylanases. Using this method, two modes of cleavage of X, were observed for the xylanase of C. ulbidus [lo, 111. In the first mode, this xylanase hydrolyses X, by preferential cleavage of the se- cond glycosidic linkage from the reducing end, yielding la- belled x,. In the second mode, the third glycosidic linkage from the reducing end is cleaved, producing labelled X,. Our results with labelled X, revealed that the bond-cleavage fre- quencies for the two cleavage modes for the Xln A of S. lividuns were almost equal (Fig. 3). In the case of [N173D]Xln A, the same analysis showed that mode 1 repre- sented 90% and mode 2 only 10%. An alternative could be that the labelled X, generated by mode 2 was an excellent substrate for [N173D]Xln A, which then rapidly cleaved X, into X,, accounting for the large accumulation of X,. This hypothesis was tested by the determination of the cleavage mode of X, by [N173D]Xln A as compared to that of Xln A. Both enzymes hydrolysed the X, by cleavage of the first xylosidic bond from the reducing end, exclusively liberating labelled X, (Fig. 2). Finally, because there was no accumula- tion of labelled XI, during X, hydrolysis, the results clearly confirmed that [N173D]Xln A exhibited a different degrada-

264

Table 2. Hydrolysis of different xylooligosides by Xln A and [N173D]Xln A and analysis of the reaction products. The reaction was carried out for 10 min at 60°C and the substrate concentration was 10 mM. During this period 20-30% of the substrate was hydrolysed. The rate of disappearance of substrate and the rate of appearance of hydrolysis products were calculated from HPLC patterns, and specific activity was expressed as ymole min-' mg-' enzyme. The values represented the mean of three independent determinations.

Reaction X, hydrolysis of X, hydrolysis of X, hydrolysis of products

Xln A "1 7 3DlXlnA Xln A "173DIXlnA Xln A [N173D]Xln

-

3 ? 0.2 5.5 +- 0.6

75 ? 4.5" 60 ? 3.3 44 ? 2.3

-

-

-

1.5 ? 0.2 3 ? 0.4

78 ? 6.2" 45 ? 2.1 99 ? 5.1

-

-

8.5 2 0.6 11 ? 0.9 15 ? 1.2

160 5 9 " 53 ? 4 91 t 6 59 ? 4 -

6 2 0.4 3 5 0.4

-

120 5 7" 26 2 3

103 ? 8 101 5 8 -

a Amount of substrate hydrolysed.

tion mode for X,. Both enzymes hydrolysed X, to X,. These experiments were conducted at low substrate concentration (0.1 mM), and it was shown that these results can be greatly affected at higher substrate concentrations because several transglycosylation reactions take place, which lead to prod- ucts larger than the starting substrates [lo]. It was not pos- sible to determine kinetic parameters for these substrates be- cause degradation and transglycosylation reactions are con- comitant. Nevertheless, the relative activities for the different substrates were determined by the rate of disappearance of the substrate and the appearance of reaction products moni- tored by HPLC (Table 2). The degradation rates were mea- sured at an initial concentration of 10 mM substrate. The er- ror inherent to this method was in the order of 10% and the amount of product resulting from the transglycosylation reactions reached a steady-state level which was constant at least for the first 10min of the reaction (data not shown). The total recovery of reaction products counterbalanced ap- proximately the disappearance of the substrate. The amount of enzyme and incubation time were adjusted to obtain a substrate degradation of 20-30%. The degradation rate was X, < X, < X,. The specific activity for X, hydrolysis was 10-20-fold less than for X, and Xs, respectively. This rela- tive activity was also observed for other xylanases [lo, 121. Both enzymes had more or less the same specific activity for X, and X,, respectively, but Xln A degraded X, more effi- ciently than [N173D]Xln A. The xylanase has both hy- drolytic and transglycosylation activities, as shown by the HPLC hydrolysis patterns. Transglycosylation reactions have been described for many glucanases [19], and the mechanism by which enzymes hydrolyse the P-glucosidic bond with re- tention of the anomeric configuration has been established. The substrate is first bound to the active site of the enzyme, and, after cleavage of the xylosidic linkage, the aglycone (reducing end of the substrate) is released from the aglycone- binding site. The activated glycone (the non-reducing end) forms an intermediate with the nucleophile residue in the active site. Then water or an acceptor attacks this bond to give the hydrolytic product or the transfer product. When transglycosylation occurs, the acceptor interacts with the aglycone-binding site. Glu236 is probably the nucleophile in Xln A (unpublished data), by analogy to the exoglucanase from Cellulomonas fimi which also exhibited a strong xyla- nase activity [20] and belongs, like Xln A, to the F family

A Hydrolysis

Carbonium intermediate X3 aglycone

B Transglycosylation

Glu236 / x5 I

Glu128

Fig. 4. Schematic representation of hydrolysis and transglyco- sylation. (A) Schematic representation of the hydrolysis of xylopen- taose by xylanase A. (B) Proposed mechanism for the transglyco- sylation reaction involving Asnl73, in which Glu236 represents the nucleophile residue.

of xylanases [8]. The acceptor molecule in the present case could be xylooligosides from X, to X,. When X, was used as a substrate, x6, X,, X, (& xylohexaose; X,, xylohep- taose ; X,, xylooctaose) and larger xylooligosides were syn- thesized by Xln A, indicating that strong transglycosylation reactions occurred (Table 2) By comparison [N173D]Xln A yielded a small amount of X, and X,. In the absence of trans- glycosylation, X, hydrolysis should yield a stoichiometric amount of X, and X, with no X,. This distribution was, how- ever, considerably affected by subsequent hydrolysis of the larger xylooligosides synthesized during transglycosylation. Xln A produced an equal amount of X, and X, but 1.5-fold more X,. "173DIXln A yielded an equal amount of X, and X, but half as much &. The X, present was due to hydrolysis of X, and X,. Otherwise, X, degradation was almost stoichio- metric for X, and X2, indicating that the transglycosylation

265

reactions were considerably reduced for [N173D]Xln A. The rate at which X, disappeared faster with Xln A than with [N173D]Xln A was essentially due to transglycosylation re- actions. However, as far as degradation is concerned, [N173D]XIn A was more effective than Xln A considering the larger amounts of X, and X, produced. When X was incubated with Xln A, some X, and X, were detected (Table 2). During incubation with [N173D]Xln A, small amounts of X, and X, appeared in the HPLC elution pattern. The X, was essentially formed from the hydrolysis of X, and X,, while X, resulted mainly from the degradation of X,. These com- plex reactions were indirectly described by the yield of hy- drolysis products such as X, and X,. Xln A liberated 1.4-fold more X, than Xz, while [N173D]Xln A yielded twofold more X, than X,. This clearly indicated that [N173D]Xln A did not perform transglycosylation reactions efficiently and that the degradation of X, into X, was predominant. With X, as substrate, transglycosylation reactions also occurred, but the larger synthesized products were quickly hydrolysed due to the presence of 10-times more enzyme than the amount used for the X, experiments. For this reason, no products larger than X, were detected in the HPLC patterns (Table 2). Under these conditions, Xln A liberated fourfold more X, than XI, while [N173D]Xln A yielded twofold more X, than XI. Transglycosylation reactions involving X, were less efficient with [N173D]Xln A and Xln A was more active on X, than [N173D]Xln A.

The catalytic cleft of the xylanase, where degradation and transglycosylation reactions take place, seems to be consti- tuted of two domains delimited by the catalytic residues. One accommodates the reducing end (aglycone) and the other the non-reducing end (glycone) of the xylooligoside chain. When hydrolysing Xs, Xln A can form two productive enzyme- substrate complexes. This implies that either the second or the third xylosidic linkage from the reducing end of the X, molecule is exposed to the catalytic residues of the enzyme. In the case of [N173D]Xln A, the third xylosidic linkage is in an unfavourable position to be easily hydrolysed. There is a correlation between degradation and transglycosylation reactions involving xylooligoside acceptors of three or more xylose residues. In all cases, the yield of transglycosylation products was lower for [N173D]Xln A than for Xln A. The single amino acid replacement, As11173 for its isosteric form Asp, is conservative in terms of size and it should not disrupt the precise organization of the catalytic cleft [21]. According to biochemical data, xylanases appear to have several sub- strate-binding subsites [lo- 121. Asn residues in the vicinity of the active site of some glycanases are reported to be in- volved in binding interactions with sugar units [22]. We pro- pose that Am173 is implicated in the aglycone-binding sub- site which interacts with the third xylosyl residue from the cleavage site of the substrate (Fig. 4). Accordingly, Asn173 interacts with the third xylosyl residue from the non-reducing end of xylooligoside acceptors in the transglycosylation reac- tion.

Among other xylanases of the F family, the xylanase of Butyrivibrio Jibrisolvens H17c contained a Glu residue in- stead of an Asn at this position [23]. It would have been interesting to compare its cleavage mode and its transgly- cosylation activity to those of our altered xylanase, but un- fortunately the mode of action of this enzyme has not been reported.

This research was supported by grants from the Natural Sciences and Engineering Research Council of Canada under the co-operative

university-industry programme and by a contribution from ICI Canada Ltd. We thank Prof Peter Biely for helpful discussions. The technical assistance of Lise Trempe, Nicole Daigneault and Liette Biron is gratefully acknowledged.

REFERENCES 1. Morosoli, R., Bertrand, J. L., Mondou, F., Shareck, F. & Kluep-

fel, D. (1986) Purification and properties of a xylanase of Streptomyces lividans, Biochem. J. 239, 587 - 592.

2. Kluepfel, D., Vats-Metha, S., Aumont, F., Shareck, F. & Moro- soli, R. (1990) Purification and characterization of a new xyla- nase (xylanase B) produced by Streptomyces’ lividans, Biochem. J. 267, 45-50.

3. Kluepfel, D., Daigneault, N., Morosoli, R. & Shareck, F. (1992) Purification and characterization of a new xylanase (xylanase C) produced by Streptomyces lividans 66, Appl. Microbiol. Biotechnol. 36, 626-631.

4. Wong, K. K. Y., Tan, L. U. L. & Saddler, J. N. (1988) Multiplic- ity of p-1,4 xylanase in microorganisms: functions and applications, Microbiol. Rev. 52, 305-317.

5. Morosoli, R., Shareck, F., Moreau, A. & Kluepfel, D. (1990) Proceedings of the 6th international symposium on genetics o j industrial microorganisms, vol. 2, pp. 935 - 946, Strasbourg, France.

6. Gilkes, N. R., Claeyssens, M., Aebersold, R., Henrissat, B., Meinke, A,, Morrison, H. D., Kilburn, D. G., Warren, R. A. J. & Miller, R. C. (1991) Structural and functional relation- ships in two families of p-1,4-glycanases, Eue J. Biochem.

7. Gilkes, N. R., Henrissat, B., Kilburn, D. G., Miller, R. C. & Warren, R. A. J. (1991) Domains in microbial p-1,4-glyca- nases : sequence conservation, function, and enzyme families, Microbiol. Rev. 55, 303-315.

8. Shareck, F., Roy, C., Yaguchi, M., Morosoli, R. & Kluepfel, D. (1991) Sequences of three genes specifying xylanases in Streptomyces lividans, Gene (Amst.) 107, 75 - 82.

9. Katsube, Y., Hata, Y., Yamaguchi, H., Moriyama, H., Shinmyo, A. & Okada, H. (1989) Proceedings of the second interna- tional conference on protein engineering, pp. 91 -96, Kobe, Japan.

10. Biely, P., Kratky, Z. & Vrsanska, M. (1981) Substrate-binding site of endo-l,4-P-xylanase of the yeast Cryptococcus albidus, Eul: J. Biochem. 119,559-564.

11. Biely, P., Vrsanska, M. & Kratky, Z. (1981) Mechanisms of substrate digestion by endo 1,4-b-xylanase of Cryptococcus albidus, Eul: J. Biochem. 119, 565-571.

12. Bray, M. R. & Clarke, A. J. (1992) Action pattern of xylo- oligosaccharide hydrolysis by Schizophyllum commune xyla- nase A, Eue J. Biochem. 204, 191-196.

13. Mondou, F., Shareck, F., Morosoli, R. & Kluepfel, D. (1986) Cloning of the xylanase gene of Streptomyces Eividaits, Gene (Amst.) 49, 323-329.

14. Hopwood, D. A., Bibb, M. J., Chater, K. C., Kieser, T., Bruton, C. J., Kieser, H. M., Lydiate, D. J., Smith, C. P., Ward, J. M. & Schrempf, H. (1985) Genetic manipulation of Streptomyces: a laboratory manual, The John Innes Foundation, Nonvich.

15. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular cloning: a laboratory manual, 2nd edn, Cold Spring Harbor Laboratoly, Cold Spring Harbor, NY.

16. Kunkel, T. A. (1985) Rapid and efficient site-specific mutagene- sis without phenotypic selection, Proc. Natl Acad. Sci. USA 74, 488-492.

17. Wood, W. I., Gitschier, J., Lasky, L. A. & Lawn, R. M. (1985) Base-composition-independent hybridization in tetramethyl- ammonium chloride: a method for oligonucleotide screening of highly complex gene libraries, Proc. Nut1 Acad. Sci. USA 82, 1585-1588.

18. Robyt, J. F. & Whelan. W. J. (1972) Reducing value methods for maltodextrins : chain-length dependence of alkaline 3,5 dinitrosalicylate and chain-length independence of alkaline copper, Anal. Biochem. 45, 510-516.

202, 367-377.

266

19. Sinnott, M. L. (1990) Catalytic mechanisms of enzymic glyco- syl transfer, Chem. Rev. 90, 1171-1202.

20. Tull, D., Whithers, S . G., Gilkes, N. R., Kilbum, D. G., Warren, R. A. J. & Aebersold, R. (1991) Glutamic acid 274 is the nucleophile in the active site of retaining exoglucanase from Cellulomonas fimi, J. Biol. Chem. 266, 15621 -15625.

21. Bordo, D. & Argos, P. (1991) Suggestion for safe residue substi- tution in site-directed mutagenesis, J. Mol. Bid . 217, 721 - 729.

22. Johnson, L. N., Cheetham, J., McLaughin, P. J., Acharya, K. R., Barford, D. & Phillips, D. C. (1988) in Carbohydrate-Protein interaction, Cur,: Top. Microbiol. Immunol. 139 (Clarke, A. E. & Wilson, I. A., eds) pp. 81-134, Springer-Verlag.

23. Lin L-L. & Thomson, J. A. (1991) Cloning, sequencing and expression of a gene encoding a 73kDa xylanase enzyme from the mmen anaerobe Butyrivibrio fibrisolvens H17c., Mol. Gen. Genet. 228, 55-61.