Mol Gen Genet (1990) 222: 265-269

© Springer-Verlag 1990

Nucleotide sequence of the celA gene encoding a ceUodextrinase of Ruminococcus flavefaciens FD-1

Wenyen Wang and Jennifer A. Thomson Department of Microbiology, University of Cape Town, Rondebosch, 7700, South Africa

Summary. The nucleotide sequence of a 3.6 kb DNA frag- ment containing a cellodextrinase gene (eelA) from Rumino- coccusflavefaciens FD-1 was determined. The gene was ex- pressed from its own regulatory region in Escherichia coli and a putative consensus promoter sequence was identified upstream of a ribosome binding site and a TTG start codon. The complete amino acid sequence of the CelA enzyme (352 residues) was deduced and showed no significant ho- mology to cellulases from other oganisms. Two lysozyme- type active sites were found in the amino-terminal third of the enzyme. In E. coli the cloned CelA protein was trans- located into the periplasm. The lack of a typical signal se- quence, and the results of transposon phoA mutagenesis experiments indicated that CelA is secreted by a mechanism other than a leader peptide.

Key words: Ruminococcus flavefaciens- Cellodextrinase - Sequence - Secretion

Introduction

Cellulolytic organisms degrade crystalline cellulose by the concerted action of a number of different cellulase enzymes with different specificities and modes of action. Ruminococ- cus flavefaciens, one of the most important cellulolytic ru- men bacteria, has been shown to produce enzymes with endo-fl-l,4-glucanase, exo-B-1,4-glucanase and cellodextrin- ase activities (Pettipher and Latham 1979; Gardner et al. 1987; Rasmussen et al. 1988). /?-l,4-glucosidase activity (cellobiase) has not been detected in R. flavefaciens (Petti- pher and Latham 1979) and cellobiose is taken up by the cell, where it is cleaved by cellobiose phosphorylase to yield glucose-l-phosphate and glucose (Ayers 1958).

We have reported the cloning and expression in Eseherichia coli of a cellulase gene from R. flavefaciens FD- 1 (Barros and Thomson 1987) and, because it had activity on carboxymethylcellulose (CMC), we classified it as an endoglucanase. However, further studies on the enzyme en- coded by plasmid pMEB200, revealed that it had cellodex-

Abbreviations: CMCase, carboxymethylcellulase; celA, gene cod- ing for CelA; CelA, cellodextrinase; ORF, open reading frame; phoA, gene encoding alkaline phosphatase; pNPC, p-nitrophenyl- fl-D-cellobioside

Offprint requests to: J.A. Thomson

trinase activity (P. du Preez et al. submitted). It has high activity against p-nitrophenyl-fl-D-cellobioside (pNPC) and releases predominantly cellobiose from cellotetraose, cello- pentaose and cellohexaose. We report here the DNA se- quence of the ceUodextrinase gene, the deduced amino acid sequence of the enzyme, and an analysis of the secretion of the cloned gene product.

Materials and methods

Bacterial strains, plasmicls and growth conditions. Plasmid pMEB200 carries the celA gene of R. flavefaeiens cloned onto plasmid pEcoR251 (Barros and Thomson 1987). The E. coli strains LKI 11 (thrl leu6 thil supE44 tonA2 r~ m~ Iac[ q lacZAM15 lacY1; Zabeau and Stanley 1982), CSH23 [A (lac pro) supE spc thi (F'lac+proA +, B +); Miller 1972], and JM103 [A (lac pro) thil strA supE endA sbcB hsdR4 (F' traD36 proA,B IacI q lacZAM15); Messing et al. 1981] were used as recipient strains for recombinant plasmids. The P h o A - E. coli strain CCl18 (araD139 A(ara, leu)7697 A lacX74 phoA A 20 gale galK thi rpsE rpoB argEam recAl ; Manoil and Beckwith 1985] was used for the TnphoA fusion experiments. Phage 2b221rex: :TnphoA ci857 Pam3 was a gift from C. Manoil (Gutierrez et al. 1987). The Ml3- derived Bluescript plasmid (Stratagene, San Diego, Calif.) was used for subcloning and nucleotide sequencing. Strains were grown in LB medium (Maniatis et al. 1982) and 100 ~tg/ml ampicillin was used for the selection of transfor- mants.

Molecular techniques. The isolation of plasmid DNA and transformation experiments were carried out according to Maniatis et al. (1982). Restriction enzymes were obtained from Boehringer Mannheim and used according to the manufacturer's specifications.

To obtain templates for nucleotide sequencing, plasmid pMEB200 carrying eelA was digested with PvuII and PstI and fragments 1, 2 and 3 cloned into Bluescript SK digested with EcoRV and PstI (Fig. 1). Exonuclease III was used to generate two sets of overlapping deletions of opposite polarity (Henikoff 1984). Sequencing was done by the chain termination method of Sanger et al. (1977) using a Sequen- ase Kit (United States Biochemical Corporation, Cleveland, Ohio).

Preparation of periplasmic and intracelIuIar extracts. Strains were grown overnight at 37°C in 200 ml LB. NaC1 and

266

p M E B 2 0 0 1

z

I l I 4

, , ,

ql

i k b ~ ) 1

4 c e l A b ~- ~ i 4 !¢l-,IF "'

1 2

Ik

4 4

I J I D I

Fig. 1. Nucleotide sequencing strategy for the Rurninocoeeus flavefaciens celA gene. Single, bold and open lines denote vector DNA, eelA gene open reading frame (ORF) and R. flavefaciens insert DNA respectively. Fragments 1, 2 and 3 were independently subcloned into Bluescript for exonuclease III shortening. The arrows represent the extent and direction of sequencing of the templates generated

TRIS-HC1 (pH 7.3) were added to final concentrations of 33 mM. Incubation was continued for a further 10 rain and cells were collected by centrifugation (6000 g for 5 min). The supernatants were stored at - 2 0 ° C. The pellets were resuspended in TRIS-HC1 (33 mM, pH 7.3) at 10 ml/g wet weight of cells. An equal volume of TSE (33 mM TRIS- HC1, pH 7.3, 40% w/v sucrose, 2 mM EDTA) was added. After 5 rain at room temperature the cells were collected as before. They were then resuspended in ice-cold deionised water at 10 ml/g wet weight of cells. After 1 rain, MgC12 was added to a final concentration of 1 mM. The cells were collected as before and the supernatants, which represented the periplasmic fractions, were stored at - 2 0 ° C.

To obtain the intracellular fraction, the cells were resus- pended in 5 ml PC buffer (50 mM K2HPO4, 12.5 mM citric acid pH 6, shown to be the optimal pH for the activity of CelA) (P. du Preez et al., submitted). The cell suspension was cooled on ice and disrupted by sonication on ice (10 s bursts for a total of 200 s) using an MSE Soniprep 150 sonicator. The debris was removed by centrifugation for 15 rain at 27 000 g and the extract stored at - 20 ° C.

Enzyme assays. Cellulase activity was determined according to Deshpande et al. (1984). Enzyme samples were suitably diluted in PC buffer and 250 gl mixed with 250 gl pNPC (Sigma, St. Louis, Mo.; 3.4 mM in PC buffer). This was incubated at 39°C for 30 rain, then the reaction was stopped by the addition of 500 gl sodium carbonate (14% w/v). The absorbances were measured at 405 nm. Protein concentration was measured according to Bradford (1976). Enzyme activity was expressed as micromoles of pNP liber- ated per minute per milligram protein. As controls for E. coli intracellular and periplasmic enzymes, the activities of fl-galactosidase (Miller 1972) and fl-lactamase (Sykes and Nordstrom 1972) respectively were measured.

Analysis o f proteins synthesized in vitro and in vivo. In vitro protein synthesis and analysis was carried out using the prokaryotic DNA-directed translation kit (Amersham) ac- cording to the manufacturer's instructions except that half the recommended quantities of all the components were used. Periplasmic and intracellular proteins, prepared as above, were separated by SDS-polyacrylamide gel electro- phoresis (SDS-PAGE; Laemmli 1970).

TnPhoA mutagenesis. Transposon insertions into pMEB200 were obtained using phage 2:: TnphoA (Manoil and Beck- with 1985) according to the method of De Bruijn and Lupski (1984) with the modifications reported by Scholle et al. (1989), except that early stationary phase cells were used for infection with the phage.

Detection of CMCase activity. Cells were grown in LB agar containing 0.5% (w/v) medium-viscosity CMC (Sigma, St. Louis, Mo.). After 24 to 48 h of growth, the colonies were removed with a filter paper disc and the plates stained with 0.2% w/v Congo red and destained with 1 M NaC1 (Teather and Wood 1982). As the CMCase activity of CelA is low, it was necessary to use 0.5% CMC, rather than the more commonly used 0.1%, and it was preferable to incubate the plates for 48 h.

Results

Nucleotide sequence of the insert on pMEB200

The nucleotide sequence of the 3.6 kb insert of pMEB200 carried an open reading frame (ORF) which, from the pre- sumptive TTG start codon to the stop codon (TAA), con- tained 1059 nucleotides encoding 352 amino acid residues (Fig. 2). The molecular weight of the predicted polypeptide coded for by this region was 39.4 kDa. The ORF was pre- ceded by a ribosome-binding site (AGGA) 7 bp upstream of the start eodon. Two lysozyme-type active sites were found between the amino acids at positions 68 to 85 and 113 to 130 (Paice et al. 1984).

A putative promoter region showing strong homology to the consensus E. coli promoter (0-70) region (Cowing et al. 1985; Hawley and McClure 1983) and the Bacillus subtilis promoter region (Rosenberg and Court 1979) was located upstream of the ORF with a - 3 5 sequence of TTGTCA followed by an 18 bp space and a - 1 0 region of TATAAT (Fig. 2). There is an imperfect inverted repeat situated at the 3' end of the gene followed by the sequence 5'-TTATTTT-3' . The transcript would have the potential to form a stem-loop structure with a A G = - 1 0 kcal/mol (Salser 1977).

Expression ofcelA in vitro

Analysis by SDS-PAGE of in vitro protein synthesis di- rected by pMEB200, revealed the presence of a major pro-

AGATCC Ba~I/BglII

-420 TTTTC~GTACATCGTTGCTATGGCC~GGAAATGGGACTGG~TGCGTGGCGAG~CGT

-360 GGAGACTATGG~CAGCTCCAGCTTCCTCCGAG~C~TCGCGAAATGGCACAGGGTTTC

-300 TTCTTCGATAAACCCATG~CGTCAAAGACTTCGA~ATAAACTCCGTGACAGAGTCTAT

-240 ACTATCTGATAAAAAC~GGG~GCACAT~CTGCTGTACTCATATAG~GCTTTATACG

-180 CATTTGATCGGGGAAAACCTTTCTGAGGAAAGGTTCTTCCCCCGTACCCCCTTTCCAAAG

-120 ACTTTT~TAG~TTTTTGCCCATTATA~GCGCACAC~GTAAAGT~G~TTTGTCAT -35

-60 TTCTTTC~GTGCGCCCTAT~TGGGCAAAAATT~TACTGAAAGTTTT~-~AGGGAGT -10 SD

1 TTGAGGGAG~CCCTTTTTCAAAAGGGTTTCCCTC~TAGACTCGTGTAAAAACTTCTGT 1 f M R E N P F S K G F P S I D S C K N F C 61 ~GAGTACAGCAGTTATGTGCTTTCCTTGTTTTTATCACAGATACTTCAGTATCTCTCAC 21 K S T A V M C F P C F Y H R Y F S I S H 121 GGTTGCTGTCGTACTTGCTGTCGGAAATGCCG~GTCCATTTcCTTGT~TTCCACACTC 41 G C C R T C C R K C R S P F P C N S T L 181 ACGGGAAATGCCGTGCTTGCTG~CACCTTGTTGATGACCTTG~CCACTTCAGCG~TC 61 T G N A V L A E H L V D D L E P L Q R I 241 CTCGGCGGACACGACATCGATCACACCGTATTCACCGCAGTAAAGAGTAGTGCCGTACTT 81 L G G H D I D H T V F T A V K S S A V L 301 CTCGGCCGTGCTTATGGCTGTGGAGAAAAGCTCCTCGAAATACGCCTCGG~GTCTCGCT 1 0 1 L G R A Y G C G E K L L E I R L G S L A 361 GTCCTCG~TTTCATCCTCTCCCC~GGATGATATCGGGAGTCCATGTTGCACCCTGATG 1 2 1 V L E F H P L P G D D I G S P C C T L M 421 AGTG~TTTCAGCGGTTCGTAGCAGT~AAATTGTACACTACCCTGTCA~CGTGCGGCGC 1 4 1 S E F Q R F V A V E I V H Y P V I V R R 481 ATTGAGG~CTGCACTGTGTCCGCTGCATTATTGTGGTAGCTTCCGAG~G~T~TCAC 1 6 1 1 E E L H C V R C I I V V A S E K N N H 541 CTCGGGAGCTATTTTTCT~TCCTTCCGATACAGATATCCGAAATCTCGTTCCACTTGCC 1 8 1 L G S Y F S N P S D T D I R N L V P L A 601 GATG~CGCCTCATCGGTCACCTCATTGAG~GCTCAAAAACGATATTGTCGGTGTCGTG 2 0 I D E R L I G H L I E K L K N D I V G V V 661 GCCGT~TCTTCTTGCTATCTCTTCCCACAGGATGTAG~GCGCTCCTGATACTCTTTAT 2 2 1 A V I F L L S L P T G C R S A P D T L Y 721 TGTCGAAAAAGGCGCTTTCAGGCTCGCCGAAATCAAATGAGAAACCGGCTGTCTTGT~A 2 4 1 C R K R R F Q A R R N Q M R N R L S C G 781 GTCGATGAC~GTTTCAGCCCATATTTGCGGCAAAGTTC~CGACCTTGTCTATTCTGTT 2 6 1 V D D Q F Q P I F A A K F N D L V Y S V 841 ATAGCCGTCCTCGAT~CGCTTCCGTCGTTATTCTG~CGATATTATAGTCGATGGG~G 2 8 1 1 A V L D N A S V V I L N D I I V D G K 901 TCTTACGTGGTCG~TCCCCAGTCTGCTATCTGCTTTATGTCGTTTTCCTTTACAAAGTT 3 0 1 S Y V V E S P V C Y L L Y V V F L Y K V 961 ATCCAGTCTTTCACGGCTGTAGTCACACTGCGACATCCATCCGCCCAGATTGATTCCCTT 3 2 1 I Q S F T A V V T L R H P S A Q I D S L 1021 GAT~TTCCTCTGCTTTT~GTAT~TATTCACCTCTAAAACTCCG~CT~CTqTTCTG 3 4 1 D N S S A F K Y N I H L s t o p 1081 ATA~CAGCTGGCGATTTATG~GGTTTATTTTGGGATGCCTTTACGCAAAATATTGTCT

1141 CTTG~CA

Fig. 2. Nucleotide sequence of the ee~ structural gene and flanking regions of Ruminococcus flavefaciens DNA. The deduced amino acid sequence is given in the single letter code from nucleotide position 1 to 1059 (352 amino acid residues). The putative --35 and - 1 0 regions are un~rlined and the ribosomal binding se- quence (SD) is boxed. The conve~ing arrows indicating imper~ct inverted repeats starting at nucleotide postion 1052 and including the TAA stop codon, represent a possible transcriptional termina- tor. The two lysozyme-type active sites are shown in bold~ce

267

tein band with an apparent molecular weight of ca. 39 kDa (Fig. 3). A protein band with a similar molecular weight was also detected upon expression of pSK1 (fragment 1 of pMEB200 cloned into Bluescript), pSK5 (Bluescript car- rying the large BamHI-EeoRI fragment of pMEB200 which contains the entire R. flavefaeiens insert as well as some pEcoR251 DNA) and pMEB210 (which is pMEB200 with PvuII fragment 2 deleted). This protein band was absent after expression of pSK2 (Bluescript carrying fragment 2 of pMEB200).

Secretion of CelA

We have previously reported that most of the CMCase ac- tivity (75%) of E. coli HB101 (pMEB200) is found in the periplasmic fraction (Barros and Thomson 1987). However, the amino acid sequence of the celA gene product did not reveal a leader peptide such as is characteristic of many secreted proteins. We therefore repeated the assays on a variety of strains, but used the pNPC assay as CelA has higher activity on this substrate than on CMC (results not shown). As controls for E. coli intracellular and periplasmic enzymes, fl-galactosidase and fl-lactamase respectively were assayed. The data (Table 1) confirmed our previous results, although there was considerable strain variation in the level of CelA expression.

It was of interest to note that pNPCase activity was barely detectable in JM103 (pMEB210) and was absent in L K l l l (pMEB210). This was in contrast to the in vitro data where pMEB210 directed readily detectable levels of a 39 kDa protein (Fig. 3). The poor in vivo expression of the celA gene of pMEB210 was confirmed by SDS-PAGE analysis of the periplasmic and cytoplasmic proteins pre- pared from cultures of L K l l l and JM103 carrying pMEB200 and pMEB210 (Fig. 4). The 39 kDa CelA pro- tein band was not detected in LKI 11 (pMEB210) and was present in low amounts in JMI03 (pMEB210).

To determine whether the CelA protein contained an amino acid sequence that could act as a leader peptide, we utilized ,~:: TnphoA mutagenesis (Manoil and Beckwith 1985). The leader sequence of the alkaline phosphatase gene on TnphoA has been deleted and as export is required for

Table 1. Localization of enzyme activity in different cell fractions

Strain Plasmid Enzyme Enzyme activity"

Extra- Peri- Intra- cellular plasmic cellular

JM103 pMEB200pNPCase 2 (3) 56 (89) 5 (8) fl-lactamase 661 (31) 1236 (59) 206(10)

JM103 pMEB210pNPCase 0.4 0.6 4 fl-lactamase 260 (18) 1140 (78) 69 (4)

LKl l l pMEB200pNPCase 42 (22) 99 (51) 53(27) fl-lactamase 333 (56) 239 (40) 27 (4)

LKI l l pMEB210pNPCase 0 0 0 fl-lactamase 96 (10) 802 (86) 40 (4)

CSH23 pMEB200pNPCase 2 (1) 283 (87) 40(12) fl-lactamase 201 (32) 379 (61) 44 (7) fl-galactosidase 1 (2) 0.3( < 1) 62(98)

a Units p-nitrophenyl-fl-D-cellobiosidase (pNPCase) are millimoles p-nitrophenol released per minute. Units fl-lactamase are micromoles ampicillin, hydrolysed per minute. Units fl-galactosi- dase are millimoles o-nitrophenol released per minute. Percentages in parentheses

268

1 2 3 4 5 6 7 kDa

4 39

4 24

20

Fig. 3. SDS-polyacrylamide gel electrophoresis (SDS-PAGE) anal- ysis of aSS-labelled proteins synthesised in vitro. Lane 1, pE- coR251 ; 2, pMEB200; 3, pSK1; 4, pSK5; 5, pMEB210; 6, pSK2; 7, Bluescript

1 2 3 4 5 6 7 8 9 10 kOa kDa

97 -~t~

66

4 5 - - ~

31 ,~

22

,~---39

Fig. 4. SDS-PAGE analysis of periplasmic (even numbered lanes) and intracellular (odd numbered lanes) proteins of stationary phase cultures. Lane 1, molecular weight markers; 2 and 3, LKl l t (pMEB200); 4 and 5, LKl l l (pMEB210); 6 and 7, JM103 (pMEB200); 8 and 9, JMI03 (pMEB210); 10, K5142(pEcoR251)

activity of the enzyme, an in-frame sequence upstream of the transposition site has to function as a signal sequence for activity to be detectable. TnphoA mutagenesis of pMEB200 resulted in the isolation of several recombinant plasmids conferring alkaline phosphatase activity on the Pho- recipient E. coli CCll8. Restriction enzyme analysis of a number of these clones revealed that all the TnphoA inserts were in the vector (results not shown). To ensure that it was possible for TnphoA to transpose into the cloned R. flavefaciens DNA, Apr Pho- transformants were screened on CMC plates. CMCase- mutants were isolated and their plasmids subjected to restriction analysis. It was shown that the TnphoA inserts were in the celA gene (results not shown).

Discussion

Nucleotide sequencing of the celA gene of R. flavefaciens has revealed that although the gene has - 3 5 , - 1 0 se- quences similar to the E. coli consensus sequences, as well as a ribosomal binding site, its initiation codon is TTG. A number of bacterial genes have been shown to initiate with this codon, including the endoglucanase gene from Clostridium acetobutylicum (Zappe et al. 1988), the Bacillus pumilus cat-86 gene (Harwood et al. 1983) and the Staphylo- coccus aureus fl-lactamase gene (McLaughlin et al. 1981). In addition, the lysis gene of the RNA bacteriophage fr has a U U G initiation codon (Adhin and van Duin 1989). Although there is an in-frame ATG codon 75 bp down- stream from the TTG, this is not preceded by a ribosomal binding site. In addition, the protein which would be trans- lated from such an initiation codon would have a molecular weight of 36.7 kDa. The molecular weight of the protein derived from the amino acid sequence starting at the TTG codon, is 39.4 kDa, in agreement with the 39 kDa protein bands found on expression of the celA gene in vivo and in vitro.

When codon usage in the celA gene was compared with that ofE. coIi (Konigsberg and Godson 1983) and B. subtilis (Ogasawara 1985), it was found to resemble the latter as it has relatively little bias. The bias in celA towards the codons UAU and GAU for Tyr and Asp is also reflected in B. subtilis and that towards codons AAA and GAA for Lys and Glu occurs in both B. subtilis and E. coli.

The predicted amino acid sequence of CelA showed no significant homology ( < 8 % ) with a number of different cellulases. This is of interest in the light of the classification of cellulases by hydrophobic cluster analysis (Henrissat etal. 1989). Bacterial and fungal cellulases have been grouped into six broad families according to this method of amino acid sequence comparison.

There are two lysozyme-type active site sequences in the amino-terminal third of the CelA protein. The active site of cellulases is often located towards the amino-terminal end of the enzyme (Knowles et al. 1987). We have shown that CelA is subject to non-competitive inhibition by cello- biose (du Preez et al. submitted). Whether one of these sites is a binding site for cellobiose and the other the active site for cellodextrinase activity will be investigated by site- directed mutagenesis.

It was of interest to note that pNPCase activity was very low in JM103 (pMEB210) and absent in LKl11 (pMEB210). SDS-PAGE analysis of cell extracts showed that this was a result of poor expression of the gene and not to any post-translational inactivation of the enzyme. Plasmid pMEB210 was derived from pMEB200 by the dele- tion of the PvuII fragment 2. The deletion removes part of the hairpin structure present at the 5' end of the celA gene as well as the TTATTTT sequence that follows it. This appears to have a detrimental effect on the in vivo expression of the celA gene.

Most of the pNPCase activity encoded by pMEB200 and its derivatives was located in the periplasmic fraction. The amino acid sequence at the amino-terminal of CelA does not, however, contain a signal sequence. Moreover, TnphoA mutagenesis of pMEB200 resulted in alkaline phosphatase activity only when transposition occurred into vector DNA. That TnphoA was able to transpose into the celA gene was shown when 64 CMCase- mutants were

269

analysed. They were all P h o A - , indicating the absence of export of alkaline phosphatase as this enzyme is only active in the periplasm (Manoil and Beckwith 1985).

Scholle et al. (1989) have also shown that the Vibrio alginolyticus sucrase enzyme is exported across the cytoplas- mic membrane of E. coli by a mechanism other than use of a typical signal sequence. However, in that case the en- zyme was localized entirely in the cytoplasm of the parental V. alginolyticus, whereas in R. flavefaeiens cellulase activity is located extracellularly, either cell associated or in the supernatant fluid (Pettipher and Latham 1979). Although the location of the CelA enzyme has not been studied in R. flavefaciens, it is unlikely that an enzyme which degrades cellodextrins would be located intracellularly. In contrast to our findings, Gong et al. (1989) have found that the Bacteroides suceinogenes periplasmic cellodextrinase was lo- cated intracellulary when the gene was cloned and expressed in E. coli. The analysis of the sequences of more cellulases from rumen anaerobes may lead to an understanding of how these proteins are secreted.

Acknowledgements. The authors wish to thank Bryan A. White for useful discussions during J.A.T.'s visit to his laboratory and for the initiation of the cellopentaose and cellohexaose experiments. This work was funded by African Explosives and Chemical Indus- tries. We are indebted to Linda Foulkes for typing the manuscript and preparing the diagrams and to Jon Lane for photography.

References

Adhin MR, van Duin J (1989) Translational regulation of the lysis gene in RNA bacteriophage fr requires a UUG initiation co- don. Mol Gen Genet 218:137-142

Ayers WA (1958) Phosphorylation of cellobiose and glucose by Ruminococcusflavefaciens. J Bacteriol 76:515 517

Barros MEC, Thomson JA (1987) Cloning and expression in Escherichia coli of a cellulase gene from Ruminococcus flavefa- ciens. J Bacteriol 169:1760-1762

Bradford MM (1976) A rapid and sensitive method for the quanti- tation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72:248-254

Cowing DW, Bardwell JCA, Craig EA, Woolford C, Hendrix RW, Gross CA (1985) Consensus sequence for Escherichia coli heat shock gene promoters. Proc Natl Acad Sci USA 82:2679-2683

De Bruijn FJ, Lupski JR (1984) The use of transposon Tn5 muta- genesis in the rapid generation of correlated physical and genet- ic maps of DNA segments cloned into multicopy plasmids - a review. Gene 27:131-149

Deshpande MV, Eriksson K-E, Pettersson LG (1984) An assay for selective determination of exo-l,4-fl-glucanases in a mixture of cellulolytic enzymes. Anal Biochem 138:481M87

Gardner RM, Doerner KC, White BA (1987) Purification and characterization of an exo-p-l,4-glucanase from Ruminococcus flavefaciens FD-1. J Bacteriol 169:4581 4588

Gong J, Lo RYC, Forsberg CW (1989) Molecular cloning and expression in Escherichia coli of a cellodextrinase gene from Bacteroides succinogenes $85. Appl Environ Microbiol 55:132-136

Gutierrez C, Barondess J, Manoil C, Beckwith J (1987) The use of transposon TnphoA to detect genes for cell envelope proteins subject to a common regulatory stimulus. J Mol Biol 195:289-297

Harwood CR, Williams DM, Lovett PS (1983) Nucleotide se- quence of a Bacillus pumilus gene specifying chloramphenicol acetyltransferase. Gene 24:163-169

Hawley DK, McClure WR (1983) Compilation and analysis of Escherichia coli promoter DNA sequences. Nucleic Acids Res 11:2237--2255

Henikoff S (1984) Unidirectional digestion with exonuclease III creates targeted breakpoints for DNA sequencing. Gene 28 : 351-359

Henrissat B, Claeyssens M, Tomme P, Lemesle L, Mornon J-P (1989) Cellulase families revealed by hydrophobic cluster analy- sis. Gene 81:83-95

Knowles J, Lehtovaara P, Teeri T (1987) Cellulase families and their genes. Trends Biotechnol 5:255-261

Konigsberg W, Godson GN (1983) Evidence for use of rare codons on the dnaG gene and other regulatory genes of Escherichia coli. Proc Natl Acad Sci USA 80:687-691

Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227 : 680-685

Maniatis T, Fritsch EF, Sambrook J (1982) Molecular cloning: A laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY

Manoil C, Beckwith J (1985) TnphoA: A transposon probe for protein export signals. Proc Natl Acad Sci USA 82: 8129-8133

McLaughlin JR, Murray CL, Rabinowitz JC (1981) Unique fea- tures in the ribosome binding site sequence of the Gram-posi- tive Staphylococcus aureus fl-lactamase gene. J Biol Chem 256:11283-11291

Messing J, Crea R, Seeburg PH (1981) A system for shotgun DNA sequencing. Nucleic Acids Res 9 : 309-321

Miller JH (1972) Experiments in molecular genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY

Ogasawara N (1985) Markedly unbiased codon usage in Bacillus subtilis. Gene 40:145 150

Paice MG, Desrochers M, Rho D, Jurasek L, Roy C, Rollin CF, De Miguel E, Yaguchi M (1984) Two forms of endoglucanase from the basidiomycete Schizophyllum commune and their rela- tionship to other fl-l,4-glycoside hydrolases. Biotechnology 2:535-539

Pettipher GL, Latham MJ (1979) Production of enzymes degrading plant cell walls and fermentation of cellobiose by Ruminococcus flavefaciens in batch and continuous culture. J Gen Microbiol 110:29 38

Rasmussen MA, Hespell RB, White BA, Bothast RJ (1988) Inhibi- tory effects of methylcellulose on cellulose degradation of Ru- minococcusflavefaciens. Appl Environ Microbiol 54:890-897

Rosenberg M, Court D (1979) Regulatory sequences involved in the promotion and termination of RNA transcription. Annu Rev Genet 13:319-353

Salser W (1977) Globin mRNA sequences: analysis of base-pairing and evolutionary implications. Cold Spring Harbor Symp Quant Biol 42:985-1002

Sanger F, Nicklen S, Coulson AR (1977) DNA sequencing with chain-terminating inhibitors. Proc Natl Acad Sci USA 74: 5463-5467

Scholle RR, Robb SM, Robb FT, Woods DR (1989) Nucleotide sequence and analysis of the Vibrio alginolyticus sucrase gene (scrB). Gene 80:49-56

Sykes RB, Nordstrom K (1972) Microiodometric determination offl-lactamase activity. Antimicrob Agents Chemother 1 : 94~99

Teather RM, Wood PJ (1982) Use of Congo red-polysaccharide interactions in enumeration and characterization of cellulolytic bacteria from the bovine rumen. Appl Environ Microbiol 43 : 777-780

Zabeau M, Stanley KK (1982) Enhanced expression of cro-fl-galac- tosidase fusion proteins under the control of the PR promoter of bacteriophage. EMBO J 1 : 1217-1224

Zappe H, Jones WA, Jones DT, Woods DR (1988) Structure of an endo-/?-l,4-glucanase gene from Clostridium acetobutylicum P262 showing homology with endoglucanase from Bacillus spe- cies. Appl Environ Microbiol 54:128%1292

Communicated by P. Tiollais

Received December 6, 1989