7342

ApPLIED A D ENVIRO ME TAL MICROBIOLOGY, Feb. 1995, p. 518-525 Vol. 61, 0.2 0099-2240/95/$04.00+0

Cloning and Characterization of a Gene Encoding a Cell-Bound, Extracellular r3-Glucosidase in the Yeast Candida wickerhamii

CHRISTOPHER D. SKORY* AD SHELBY . FREER Fermentation Biochemistry Research Unit, National Center for Agricultural Utilization Research, USDA Agricultural Research Service, Peoria, Illinois 61604

Received 23 September 1994/Accepted 9 December 1994

The ability of yeasts to ferment cellodextrins is rare. Candida wickerhamii is able to use these sugars for alcohol production because of a cell-bound, extracellular, J3-glucosidase that is unusual by not being inhibited by glucose. A cD A expression library in lambda phage was prepared with mRNA isolated from cellobiose­ grown C. wickerhamii. Immunological screening of the library with polyclonal antibodies against purified C. wickerhamii cell-bound, extracellular J3-glucosidase yielded 12 positive clones. Restriction endonuclease anal­ ysis and sequence data revealed that the clones could be divided into two groups, bgLA and bglB, which were shown to be genetically distinct by Southern hybridization analyses. Efforts were directed at the study of bglB since it appeared to code for the cell-bound J3-glucosidase. Sequence data from both cD A and genomic clones showed the absence of introns in bglB. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis and immu­ noblotting of cell lysates from Escherichia coli bglB clones confirmed the presence of an expressed protein with an apparent molecular mass of 72 kDa, which is consistent with that expected for an unglycosylated form of the enzyme. Amino acid comparisons of BglB with other J3-glucosidase sequences suggest that it is a member of family 1 glycosyl hydrolases but is unusual in that it contains an additional 100 to 130 amino acids at the N terminus. This sequence did not have homologies to other known protein sequences and may impart unique properties to this J3-glucosidase.

The production of fuel ethanol from cellulosic biomass, such to glucose inhibition (7, 19). The cell-bound, extracellular ~­ as wood, newspaper, and agricultural residues, has been exten­ glucosidase accounts for approximately 90% of the total ~-glu­ sively studied from both an economic and a microbiological cosidase activity (7), and this 'enzyme is primarily responsible standpoint. While it is currently possible to produce ethanol for the efficient conversion of cellodextrins to glucose (9). with cellulose as the carbon source, economics dictate that the We have isolated a gene (bgLB) from C. wickerhamii that efficiency of production must be improved if this method is to codes for the cell-bound, extracellular ~-glucosidase. In this be competitive with starch-based processes. In the current fer­ paper we describe the cloning and nucleotide and amino acid mentation scheme, the cellulosic feedstock is typically chemi­ sequence analyses of the bglB gene and compare it with other cally or physically pretreated (16) and then subjected to enzy­ ~-glucosidases. matic degradation involving a mixture of glucanases that hydrolyze cellulose to cellobiose and cellodextrins. These prod­ ucts are subsequently hydrolyzed by ~-glucosidase (1,4-~-D­ MATERIALS AND METHODS glucoside glucohydrolase, EC 3.2.1.21) to glucose, which can Strains and culture conditions. C. wickerhamii (Capriotti) Meyer and Yarrow ultimately be converted to ethanol. Difficulties in regulating NRRL Y-2563 (syn. Torulopsis wickerhamii Capriotti) was acquired from the such a system often arise because of accumulation of glucose, Agricultural Research Service Culture Collection, ational Center for Agricul­ which typically leads to end product inhibition of most ~-glu­ tural Utilization Research, Peoria, Ill. Growth of this yeast was carried out at 28°C. cosidases (33). This results in an accumulation of cellobiose, Chemicals. Yeast extract, malt extract, and peptone were bought from Difco which in turn is a potent inhibitor of endoglucanases and Laboratories (Detroit, Mich.). Chemicals, unless specifically referenced other­ exoglucanases (22). Supplementation with additional amounts wise, were purchased from Sigma Chemical Co. (St. Louis, Mo.). of ~-glucosidase can be used to overcome this rate-limiting Enzyme assay. fj-Glucosidase activity was measured as previously described (7) by incubating appropriately diluted cells or other samples in 2.5 mM p-ni­ step (5, 33). An alternative is to use a cellulase system that has trophenyl-fj-o-gJucopyranoside in either 50 mM sodium acetate buffer (pH 4.75) a ~-glucosidase which is naturally resistant to end product or 100 mM phosphate buffer (pH 7.00). One unit of enzyme activity is defined as (glucose) inhibition. This is a potential strategy that this labo­ 1 f.lmol of product, p-nirrophenol, formed per min. ratory is actively exploring. Isolation and analysis of yeast D A. DA was isolated from C. wickerhamii protoplasts by the method of Philippsen et a1. (29). Zymolyase 60000 was pur­ The yeast Candida wickerhamii ferments soluble cellodex­ chased from Miles Labs (Elkhart, Ind.). Restriction endonuclease (RE) enzymes trins, with degrees of polymerization of 2 to 6, to ethanol in were purchased from Gibco-BRL (Gaithersburg, Md.) and used according to the high yields (8, 10, 11). This organism appears to contain an supplier's instructions. Southern hybridization analyses were performed with the intracellular ~-glucosidase (23), a secreted ~-glucosidase (19, Genius System (Boehringer Mannheim, Indianapolis, Ind.) using the manufac­ turer's recommendations. ~-glucosidase 23), and a cell-bound, extracellular that is re­ Construction of the C. wick.erhamii cD 'A library. C. wick.erhamii RRL leased only upon enzymatic digestion of the cell wall (7, 13). Y-2563 was grown in yeast extract-peptone-malt extract medium (10% cellobi­ The secreted and cell-bound ~-glucosidases are highly resistant ose) with shaking under anaerobic conditions to promote production of fj-glu­ cosidase (12). The cells were harvested by centrifugation at 11,000 x g, washed once with 0.9% aCl, frozen in liquid 2, and then ground with glass beads (OA5-mm diameter) with a prechilled mortar and pestle. Total RA was isolated from the disrupted cells by a hot-phenol extraction protocol (1) and used to * Corresponding author. Mailing address: USDA-ARS- CAUR­ obtain mR A by further purification with oligo(dT) cellulose purification col­ FBR, 1815 N. University St., Peoria, IL 61604. Phone: (309) 681-6276. umns (Gibco-BRL) according to the manufacturer's recommendations. A cD A Fax: (309) 681-6686. library was prepared from the mR A with the ZapII cD A synthesis kit (Strat-

518 VOL. 61, 1995 C. WICKERHAMII GE FOR [j-GLUCOSIDASE 519 agene, La Jolla, CaliL). The cD. A library was amplified with Escherichia coli A. B. SURE (Stratagene), while in vivo excision of the pBluescript phagemid, contain­ 2 3 2 3 ing the cD A inserts, from the Uni-ZapII XR vector was performed according to the manufacturer's protocol using E. coli XLl-Blue and SOLR (Stratagene). Plasmid pBluescript II (KS-) was transformed by electroporation into E. coli SOLR and served as a control strain containing only vector sequences. 23.1- 23.1- Construction of the C. wickerhamii lambda genomic library. ADashII DA (Stratagene) was digested with RE Xhol, and then the 5' overhangs were par­ 9.4­ 9.4- tially filled in with dTTP and dCfP. Insert DA was prepared by partially digesting C. wickerhamii DA with RE Sau3A, and the 5' overhangs were 6.6- 6.6- partially filled in with dGTP and dCfP to create compatible ends with the digested ADashII vector. Ligation, packaging, and amplification were performed 4.4- 4.4- according to the manufacturer's recommendations and were followed by pack­ aging with the Gigapack-II XL Lambda packaging extract (Stratagene). E. coli LE392 was used for all manipulations and screenin3 of the lambda genomic library. 2.3­ 2.3- Screening colonies and plaques for ~-glucosidase activity. Phage were used to 2.0- 2.0- infect E. coli SURE and XLI-Blue, which were then inoculated ;..•0 a top agar medium containing 3 mM isopropyl-thiol-galactose (IPTG) and 0.5 mM meth­ '. ylumbelliferyl-~-D-glucoside 5-bromo-4-chloro-3-indoyl-~-D­ (MUG) or 0.5 mM 0.56 - 0.56- glucopyranoside (X-Glu) as an indicator. The lambda genomic library was screened as described above, except E. coli LE392 was used as the host and IPTG was omitted from the medium. Luria-Bertani (LB) medium was used for both phage and colony growth. The ~-glucosidase activity resulted in a dark-blue formation with X-Glu and a blue fluorescence, visible under long-wave UV (366-nm) light, with MUG. Production of polyclonal antibodies against purified ~·glucosidase protein. FIG. 1. Southern hybridization analysis of total D A, isolated from C. wick­ Polyclonal antiserum was prepared from ew Zealand White rabbits by standard erhamii, hybridized with XhoI-linearized pBG2 (bglB [AJ) or pBGl (bglA [BD as protocols (I) and purified C. wickerhamii cell-bound ~-glucosidase (7) as an a probe. Lanes: I, BamHI digested; 2, EcoRl digested; 3, HindIII digested. antigen. Antiserum was purified by precipitation of immunoglobulin G molecules Molecular sizes of A/HindIII standards are given in kilobases. with ammonium sulfate followed by column chromatography using DEAE­ Affi-Gel Blue resin (Bio-Rad Laboratories, Hercules, CaliL). Western analysis (immunoblotting) of ~-glucosidase protein. The protocol used for detection of phage isolates expressing ~-glucosidase was that of Sam­ brook et al. (30) with the modification of employing the Genius System (Boehr­ cosidase activity could be detected in any PFU from either the inger Mannheim) with ytran membranes (Schleicher & Schuell, Keene, N.H.) ADashII genomic library or the AZapII eDNA library. for detection of bound antibodies made against ~-glucosidase. Immunological Because of the absence of detectable enzymatic activity, im­ detection was also used to detect the presence of the ~-glucosidase protein in E. coli SOLR isolates containing the recombinant phagemids. Isolates were grown munological screening was implemented to ascertain whether to the mid-log phase of growth in LB containing 40 J.l.g of ampicillin per mI, a recombinant [3-glucosidase was present. Conditions for washed, resuspended in sodium dodecyl sulfate (SDS)-polyacrylamide gel load­ Western analyses using the anti-[3-gIucosidase antibody were ing buffer, and lysed with a boiling water bath according to the method of optimized to obtain a detection limit of 1 pg of purified [3-glu­ Sambrook et al. (30). Total protein homogenates were separated by SDS-poly­ acrylamide gel electrophoresis (PAGE; 10% polyacrylamide) and eIectroblotted cosidase protein. Screening the AZapII cD A library using E. to polyvinylidene difluoride membrane (Bio-Rad) by the methods of Matsudaira coli XLI-Blue cells revealed that 1 of 275 PFU resulted in a (27). The membrane was cut to remove the lanes containing molecular weight strong antibody binding signal over the AZapII control. markers so they could be stained with Coomassie R-250. The remaining mem­ Twelve of these plaques were isolated and purified for fur­ brane was then used for ~-glucosidase detection as discussed above. ucleotide sequence analysis. Sequence analysis was performed using the Taq ther analysis. DyeDeoxy Terminator Cycle Sequencing Kit (Applied Biosystems, Foster City, Plasmid analysis of (3-glucosidase clones. Plasmids were Calif.). Reaction mixtures were purified with Quick Spin columns (Boehringer recovered by in vivo excision, with E. coli SOLR as the recip­ Mannheim) and analyzed with an Applied Biosystems 373A DA sequencer. ient, from all 12 AZapII cD A clones that yielded a positive Oligonucleotide primers M13 and P13 were used for sequencing inserts con­ tained in phagemid vector pBluescript II. Primer P13 was chosen as a substitute Western analysis hybridization signal to the anti-[3-gIucosidase for the primers reverse M13 and T3, because of its higher thermal denaturation immunoglobulin. RE digestion of the plasmid isolates showed point and t::..Go (4). that all but two clones, plasmids pBG1 and pBG28, had inserts Computer analyses of nucleotide data were performed using the University of of approximately 2 kb in length and virtually the same diges­ Wisconsin Genetics Computer Group package. Protein sequences were com­ pared and overlapped in a multiple-sequence format with Pileup. Conservative tion patterns with BamHI, SalI, PstI, and EcoRI. This predom­ amino acid substitutions in the multiple-sequence format were defined as those inant group of clones was designated group B, or bglB ([3­ amino acids having a similarity of 2:0.1 by using the symbol comparison table, glucosidase B). Three of the bglB eDNA clones that appeared PileupPep-cmp, provided with the University of Wisconsin Genetics Computer to be full length were sequenced in the region of the 5' end of Group package. Amino acid sequence analysis. Purified cell-bound, extracellular ~-glucosidase the insert, and it was found that they differed in length by 24 from C. wickerhamii (7) was analyzed by the Macromolecular Structure Facility, bp. The largest of the three inserts was from plasmid pBG2. Michigan State University, East Lansing. The amino-terminal end and several Sequence analysis of the 5' and 3' ends of pBG28 showed that internal fragments obtained by digestion with trypsin were sequenced by the it was an incomplete 0.9-kb bglB cD A fragment whose se­ Edman degradation procedure. ucleotide sequence accession number. The nucleotide sequence of bglB has quence matched up identically with analogous regions of been submitted to GenBank under accession number 13672. pBG2. The 1.1-kb cD A insert from plasmid pBGl was se­ quenced and found to be different from any of the bglB clones (data not shown). This clone was designated bgLA. Southern RESULTS analysis confirmed that both bgLA and bglB were distinct genes that hybridized to different loci (Fig. 1). Library screening. Approximately 24,000 PFU from the Isolation of genomic bglB clones. The ADashII C. wicker­ lambda genomic library and over 100,000 PFU from the cD A hamii genomic library was probed with XhoI-linearized pBG2 library were screened for activity (with and without IPTG in­ which contains the nearly full-length bglB cD A fragment. cluded in the growth medium) with MUG as a fluorescent Approximately 1 of 320 plaques yielded a strong hybridization indicator of [3-glucosidase activity. This was again repeated signal. DA was purified from eight of these isolated clones with the chromogenic indicator X-Glu. In all cases, no [3-glu- and used for RE digestion and Southern analysis using the 520 SKORY D FREER APPL. E VIRON. MICROBIOL.

GAATTCAATTGAAGTCATGATGTTTGCATTTTGCCAATTTGTCAGTTTACTTAGAATTTCATTATTTAAATGATACTTTTTGCCTTTGTGGAAGTATTTG 100 101 AAATTTTATCAATTAAAAACTGTTAAGAAAAGATGTTCTCACAAAAGTATCTTTTATCATTAGCTGCAATAATTGCAATCGCTAAAGCAGCTCCAGCTGA 200 1 MFSQKYLLSLAAI IAIAKAAPADtm 23 201 TGACGCTTCTAAACCAGGTATTGGGAAATTTGCACCAGGTCAATTAGGTTTCCGTTATTATATCGACACCACCACCGAGTACGCAACTCCTGCCACTGCT 300 24 0 ASK P GIG K FAPGQLG FRY Y lOT TTEYAT PAT A 56 (2)

301 ACTGCTCCTGCAAGTTCCACTACGTACGCTGCACCATATGCTGAATTGTCATCCTTGGTTGGAAACTTGTCCACGACGACATGGGGTAATTGGTATCCTG 400 57 TAP ASS TTY AA PYA ELSSLVGNLSTTT W GN W Y P 0 90

401 ACGCTACCGAGGCTGCCACGGATACTGATGACCCATATGGACAATACGCATGGTCTCAATTATGGGAAGCTACCACTTTCCCAAATTTTACTCGTGGTAT 500 91 ATE AATDTDDPYGQ YAW SQL W EAT TFPNFTRGI 123

501 TTACAGTACCACGGTGGATCCAACACCGATCCCAACCGAGAGTTTAGTTGTGCCACCAGATGACCCAGTCAAGAGGGCATTCCAAGATTTGGGAATCAAA 600 124 Y S T T V 0 P T PIP TESLVVP POD PV K R A F Q 0 L G I K 156

601 TTCCCTCTGGGTTTCATTCAAGGTGTTGCCGGTTCCGCTGCTCAAATTGAAGGTGCCGTCGCCGATGAAGGTAGATCACCAACTAATTTAGAAGTTAGTT 700 157 FPLGFIQGVA GSA AQ lEG AVA0EGRSPTN LEV SS 189

701 CCGCTAGTAGACATTTACCTGAAGATTTCGTCACTAATGAAAATTATTATCTTTACAAACAAGATATCACCAGATTAGCAGCTATCGGCGTTGAATACTA 800 190 ASR HLP EDFVTNENYY LYle Q0 ITRLAAIGVEYY 223 801 TTCGTTCACTATCCCATGGACTAGAATCTTACCATTCGCCTATCCCGGTTCTCCTGTGAATCAACAAGGTTTAGATCATTATGACGACTTGATCAACACT 900 224 SF TIP W TRILP FAY PGSPVNQQGL0HY DOL INT 256

901 GTCTTAGCATATGGAATGAAACCAATTGTCACATTGATCCATTTCGATTCACCATTACAACTTGTCGACTTCAATGCCACATTGGAATTGGGACTGCCAG 1000 257 V LAY GM Ie PIVTLIHFDSPLQLV0F NAT LELG LPG 290

1001 GTGGATACGAAGGTGAAGATTTCGTCGAGGCATTTGTCAATTACGGTAAAATCGTCATGACCCATTTCGCTGATCGTGTCCCATTATGGATCATCTTTAA 1100 291 GYEGEDFVEAFVNYG K IV M TH FAD RVPL W IIFN 323

1101 TGAACCTGTCCAATTCGCCACTAATGGACTCGGTGTCAAACATGTCGTCCAAGCCACGGCTCAATTATACGATTTCTACCATAACGAGATCAACGGGTCC 1200 324 EPVQ FAT NGLGV K HVV QAT AQLY0FYHNEINGS 356

1201 GGTAAGATTGGTATGAAGTTCAGTCACATCTTCGGGTTCCCTGAGGATCCAACTAACCCAGAACATGTTGCTGCCGCAGACAGATCAAATGAATTGCAAT 1300 357 G Ie IGM K FSHIFGFPEDPTNPEHVAAA0RSNELQL 390 (3) 1301 TAGGTCTCTTTGCTGATCCATTGTTCTTAGGTGAAGACTACCCAGACAGTTTCAAGACCACATTATTGAAAACGCAGCCAGCACTGGCTTGGACACTGGA 1400 391 GL FAD PLFLGE0YP0SF K TTL L K T Q PAL A W T L 0 423 1401 TGAATTAGCCGCTGTTAAGGGTAAATGTGATTTCTTCGGTGTTGATCCATACACTTATAACACTATCAAGCCATTGGATAACGGTACTGCATCATGTGAA 1500 424 ELAAV K G Ie C0FFGV0PYTYNTI K P LON GTASCE 456 1501 GCCAACGTCACCGACACTTACTGGCCAACGTGTGTCAATGTCACCGTTACTGAAGCTGATAACTGGAGTATCGGTTACCGTTCCCAATCCTATGTCTACA 1600 457 ANV TOT Y W PTCVNVTV TEA 0 N W SIGYRS Q SY v YI 490 1601 TCACACCAAGACAATTAAGAGTCTCGTTGAACTACATCTGGCAACACTGGCACGTTCCTATCTTCATCACGGAATTTGGTTTCCCTGAATGGAGAGAAGG 1700 491 TPR Q LRVSLNYI W Q H W HVPI FIT EFGF PEW REG 523

1701 TGAGAAACTCTTAGTTGACCAAGTCCAAGATTTGGACAGATCCATTTACTACAGATCTTTCTTGACTGCAGCATTAGAGGCATCTCAGTACGACGGTGTC 1800 524 E K L L V 0 Q V Q 0 LOR SlY YRSFLTA AlE A S Q Y 0 G V 556 (4) . x 1801 GAGATAATGGGTGCCTTGGCTTGGAGTTTTGCCGATAATTGGGAATTCGGTGATTATAACCAACAATTCGGTTTACAAGTCGTTAATAGAACTACTCAGG 1900 557 ElM GAL A W S FAD N W EF GOY NQQFGL Q VVNRTTQE 590

1901 AGAGATTCTATAAGAAGAGTTTCTTTGATTTTGTCGGTTTTATTAATGATAATAGAGCTTGAGATCCCTAATCCATTTATGATTATATTTTTTAAAAAAC 2000 591 RFY K K S F F 0 F V G FIN 0 N R A * 610 (5) . 2001 TTTCTATGATGACTATTATTTCTTTAATGACATTACTACACTAAATGTCAATTTCTTTACTTACGCTTCTTTTATTTTATAACCCAGAAAAGTTGATATA 2100

2101 CAATTTACTTGTCTTTTACCAATTTAAATAAAAAATTAAATAAATAAAATTCAAGCTT 2158

FIG. 2. Nucleotide sequence of bglB and deduced amino acid sequence (accession number U13672). Amino acid sequences confirmed by Edman degradation analyses are shown with double underlines and numbered (shown in parentheses) for reference. , processing site for removal of signal sequence; &, polyadenylation site; /\, glutamic acid 514 and aspartic acid 529 proposed to be involved as a nucleophile and acid-base catalyst, respectively.

same probe as above. All clones appeared to have overlapping and M13 were utilized to sequence the 5' and 3' ends of the regions and contained fragments that hybridized strongly to cD A insert contained in plasmid pBG2. Synthetic oligonu­ the probe. One of these isolates, LBG2-4, was shown by se­ cleotides specific to the cD A were then used as primers to quence analysis to possess a 6.5-kb HindUI fragment that con­ continue sequencing inward until overlapping sequences pro­ tained the entire structural portion of the bglB gene. vided an unambiguous nucleotide sequence of the entire Amino acid and nucleotide sequence analyses. Primers P13 cD A insert. These same primers were again used to sequence VOL. 61, 1995 C. WICKERHAMII GE. E FOR (j-GLUCOSIDASE 521

TABLE 1. Comparison of amino acid compositions Further investigation revealed that BglB had approximately of (j-glucosidases from C. wickerhamii the same degree of similarity (50 to 55%) and identity (25 to 30%) to all of the proteins in this family (Fig. 3). They included Amt (mol)/mol of indicated enzyme Residue bacterial ~-glucosidases (14, 15, 25, 35, 38), phospho-~-gluco­ BglB deduced Cell boundo Secreted/; sidases (6, 31), ~-galactosidases (24), phospho-~-galactosi­ ~-glucosi­ Ala 51 51 68 dases (3), plant cyanogenic and noncyanogenic Cys 3 4 6 dases (20, 28), and mammalian lactase phlorizin hydrolases Asp or Asn 70 68 72 (26). Glu or Gin 58 57 74 BglB enzyme characterization and analysis. one of the 12 Phe 38 40 45 AZapII cDNA isolates that reacted with anti-~-glucosidase Gly 42 42 65 antibody had any detectable ~-glucosidase activity associated His 10 11 13 with the plaques by using MUG-containing LB medium with or Ile 29 30 35 without IPTG. However, whole cells of E. coli SOLR carrying Lys 18 22 25 bglB cD A clones, grown in LB-ampicillin medium, were Leu 45 46 52 Met 4 0 5 found to have optimal activity at approximately pH 7 and Pro 38 37 30 temperatures between 37 and 40°C. These activities usually Arg 18 22 24 were in the range of 5 to 30 mUfml for the substrate p-nitro­ Ser 31 29 30 phenyl-~-D-glucopyranoside. X-Glu and esculin were also Thr 50 49 64 cleaved by the recombinant bglB isolates, but no activity Val 38 37 45 against cellobiose was detected. Examination by phase-con­ Tyr 34 35 42 trast microscopy of the bgLB clones grown overnight in LB­ ampicillin medium revealed that most cells contained cytoplas­ Total 577 580 695 mic inclusion granules that occupied approximately 10 to 20% ° From the work of Freer (7). of the cell volume. In attempts to obtain a cell-free enzyme /; From the work of Himmel et al. (19). preparation from the E. coli SOLR or XLI-Blue bglB clones, all detectable ~-glucosidase activity was lost when cells were lysed. Methods of breakage included the use of a French press, DA obtained from the genomic clone LBG2-4 (Fig. 2). The sonication, and enzymatic treatment with lysozyme followed by sequences in the bglB coding region were identical between the deoxycholate or Triton X-100. Various methods that used ei­ cDNA and the genomic D A, establishing the absence of any ther 100 mM phosphate (pH 7.00) or 50 mM acetate (pH 4.75) introns. An open reading frame of 1,830 bp was present in the buffer, with or without several different combinations of pro­ genomic sequence. By examining sequence data of the 3' ends tein stabilizers (e.g., glycerol, dithiothreitol, and ~-mercapto­ of several bglB cD A clones, the polyadenylation site was ethanol) and proteinase inhibitors (e.g., phenylmethylsulfonyl concluded to be at nucleotide 2121. On the basis of the fluoride, pepstatin A, and leupeptin), were employed. In all genomic DA sequence, it was found that the cD A insert cases, ~-glucosidase activity was not detected after cell break­ from plasmid pBG2 was not quite full length, since its se­ age. quence started at nucleotide 150 in the 5' region and did not E. coli SOLR clones containing plasmids pBG2 and pBG6 include the proposed ATG start codon. expressed a novel 72-kDa protein that was not only partially The peptide sequence obtained from Edman degradation visible in an SDS-polyacrylamide gel stained with Coomassie analysis of the amino terminus of the processed C. wickerhamii R-250 but also readily evident by Western analysis (Fig. 4). E. extracellular ~-glucosidase was identical to the putative amino coli SOLR containing plasmid pBG28 produced a unique 33­ acids 22 to 40 deduced through translation of the bgLB nucle­ kDa protein that could be easily detected by Coomassie R-250 otide sequence (Fig. 2, fragment 1). One exception was that all staining and Western analysis. These recombinant proteins are of the proposed glycines at amino acid positions 29, 31, 36, and consistent with the anticipated sizes of unglycosylated, nearly 39 appeared to be threonines by Edman sequencing. A tryptic full-length (pBG2 and pBG6) and unglycosylated, truncated digestion was performed on the enzyme, and several fragments (pBG28) proteins, with a small portion of translated polylinker were separated by high-performance liquid chromatography fused to the amino terminus. No hybridization signals could be and analyzed by Edman sequencing. The amino acid sequences detected in E. coli SOLR harboring the control plasmid pBlue­ from fragments 2 to 5, totaling 36 different amino acids, coin­ script II (KS-). cided with the protein sequence obtained from bglB. The only exceptions were that the glycines present in fragments 3 and 5 DISCUSSIO appeared as threonines by Edman analysis. Comparisons of the deduced BglB amino acid composition The goal of this study was to clone the gene from C. wick­ with that obtained experimentally for the secreted (19) and erhamii responsible for expression of the cell-bound, extracel­ cell-bound, extracellular (7) C. wickerhamii ~-glucosidases lular ~-glucosidase. Previous attempts by this laboratory to (Table 1) reveal similar patterns of amino acid usage between clone this gene by using a pBR322 plasmid library and screen­ these proteins. However, the composition of the deduced BglB ing E. coli transformants for ~-glucosidase activity were unsuc­ is nearly identical to that of the cell-bound, extracellular en­ cessful. It is not known whether difficulties were encountered zyme, except for the methionine content, further supporting because of instability of the protein or nonrecognition of the the idea that bglB codes for this enzyme. promoter in a gram-negative host. By using a cD A expression Amino acid similarity to other J3-glucosidases. The University library and immunological screening, two different genes, bgLB of Wisconsin Genetics Computer Group program TFASTA and bgLA, were isolated. Several convincing lines of evidence was used to search the GenEMBL nucleotide sequence data­ confirm that the bglB gene encodes the previously purified base for any similarities to the putative BglB peptide sequence. cell-bound, extracellular enzyme (7). First, the recombinant Several ~-glucosidases belonging to the same family, family 1 BglB protein not only reacts with antibodies against the cell­ (17, 18), of glycosyl hydrolases were detected in this search. bound, extracellular ~-glucosidase but also shows activity 522 SKORY D FREER APPL. E VIRON. MICROBIOL.

Bpolyma 1 . Ctherma 1 Csacch 1 Bpolymb 1 Agrobact 1 Mbispor 1 Ecoli 1 Cwicker 22 ADDASKPGIGKFAPGQLGFRYYIDTTTEYATPATATAPASSTTYAAPYAE

Bpolyma 1 . Ctherma 1 Csacch 1 Bpolymb 1 Agrobact 1 Mbispor 1 •••••••••••••••••••••••••••••••••••••••••••••••• MT Ecoli 1 Cwicker 72 LSSLVGNLSTTTWGNWYPDATEAATDTDDPYGQYAWSQLWEATTFPNFTR

Bpolyma 1 Ctherma 1 Csacch 1 Bpolymb 1 Agrobact 1 Mbispor 3 Ecoli 1 Cwicker 122

Bpolyma 21 Ctherma 21 Csacch 20 Bpolymb 23 Agrobact 26 Mbispor 53 Ecoli 19 Cwicker 172

Bpolyma 65 Ctherma 65 Csacch 64 Bpolymb 67 Agrobact 70 Mbispor 95 Ecoli 69 Cwicker 212

Bpolyma 113 Ctherma 113 Csacch 112 Bpolymb 114 Agrobact 118 Mbispor 143 Ecoli 118 Cwicker 262

Bpolyma 154 Ctherma 154 Csacch 153 Bpolymb 155 Agrobact 159 Mbispor 184 Ecoli 160 Cwicker 312

Bpolyma 204 Ctherma 204 Csacch 203 Bpolymb 205 Agrobact 209 Mbispor 231 Ecoli 209 Cwicker 350

FIG. 3. Amino acid sequence similarities of several ~-glucosidase enzymes found in family 1 of glycosyl hydrolases. Conserved amino acids are shown with a gray background, and identical amino acids have a black background. Bpolyma Bacillus polymyxa BglA; Ctherma, Clostridium thermocellum BgIA; Csacch, Caldocellum saccharolyticum BgIA; Bpolymb, B. polymyxa BglB; Agrobact, Agrobacterium sp. Abg; bispor, Microbispora bispora BglB; Ecoli, E. coli Bgl; Cwicker, C. wickerhamii processed BglB.

against synthetic and natural f3-glucosides. Second, the molec­ bglB gene. The only exception was the inconsistency of differ­ ular masses of the deduced peptide sequence and the recom­ entiating threonines and glycines by Edman analysis. However, binant BglB protein correspond to that calculated for the na­ two different examinations of the same region (fragments 1 tive cell-bound, extracellular enzyme by Western analysis. and 2, Fig. 2) yielded conflicting results concerning the pres­ Third, the amino acid composition of the putative peptide ence of threonines, which strongly suggests faultiness of the sequence is nearly identical to that obtained from the purified Edman analysis procedure. native cell-bound protein. Finally, the amino acid sequence There were obvious differences between the recombinant from Edman degradation analysis of five different regions, BglB and the native cell-bound C. wickerhamii f3-glucosidase. totalling 54 amino acids, of the purified f3-glucosidase matched The native enzyme cleaves cellobiose and has a pH optimum of identically the peptide sequence deduced from the isolated 4.75, while the recombinant enzyme could not cleave cellobi- VOL. 61, 1995 C. WICKERHAMJI GE E FOR ~-GLUCOSIDASE 523

Bpolyma 248 D Ctherma 248 Csaeeh 253 Bpolymb 249 Agrobaet 253 Mbispor 271 Reoli 252 Cwieker 391

Bpolyma 292 Ctherma 292 Csaeeh 301 Bpolymb 294 Agrobaet 295 Mbispor 314 Reoli 296 Cwieker 436

Bpolyma 328 Ctherma 330 Csaeeh 338 Bpolymb 330 Agrobaet 334 Mbispor 353 Reoli 337 Cwieker 486

Bpolyma 36'9 Ctherma 373 Csaeeh 381 Bpolymb 373 Agrobaet 376 Mbispor 395 Reoli 379 Cwieker 534

Bpolyma 417 I Ctherma 421 Csaeeh 429 Bpolymb 421 Agrobaet 424 Mbispor 442 Reoli 428 Cwieker 582

FIG. 3-Conlinued.

ose and had a pH optimum of approximately 7. Additionally, It is interesting to note that most members of family 1 have the recombinant enzyme was stable only when associated with the conserved sequence Ile-Thr-Glu-Asn-Gly surrounding the whole cells, and even then, degradation was often severe. glutamic acid nucleophile. Furthermore, Asn and Gly follow These anomalies may be a consequence of improper folding in immediately after Glu in all of these enzymes described to the E. coli host or a manifestation resulting from the lack of date, except for C. wickerhamii BgIB, which contains a Phe in glycosylation. The instability in a cell-free state may explain place of Asn. In the Agrobacterium sp., it has been demon­ why we were unable to detect any [3-glucosidase activity in the strated by site-directed mutagenesis that replacement of Asn­ recombinant Aphage plaques, whereas activity was observed in 359 with Ser results in only a threefold reduction in [3-gluco­ E. coli clones that contained plasmids with the same cD A sidase activity (34). However, Ser is a very conservative inserts. substitution for Asn, while Phe might be expected to alter the The deduced C. wickerhamii BglB protein has been found to characteristics of this protein region. be a member of family 1 glycosyl hydrolase (17), previous­ ly called family BGA (2), which is composed of enzymes Perhaps the most striking difference between BglB and other from archaebacteria, eubacteria, plants, and mammals. This members of family 1 glycosyl hydrolases is the presence of is the first description of a yeast protein belonging to this additional amino acids at the amino terminus of the protein. family. If the proteins in this family are compared in a multiple­ Closer examination of the amino acid sequence with other sequence alignment, both the amino and carboxyl termini dif­ family 1 glycosyl hydrolases reveals that Glu-514 in C. wicker­ fer in position by only a few amino acids (Fig. 3). However, hamii BglB is positioned in the same relative location as Glu­ the processed BglB contains between 100 and 130 additional 358 in an Agrobacterium sp. This amino acid in the Agrobacte­ amino acids at the amino end. This sequence could be in­ rium sp. has been determined by site-directed mutagenesis (34, volved in linking the protein to the cell wall or perhaps aids 36) and inhibitor studies (37) to be the nucleophilic amino acid in substrate binding. Similarities to this region could not be involved in catalysis. It has also been proposed that Asp-374, in found with implied protein sequences in the GenEMBL data­ the Agrobacterium sp., serves as an acid-base catalyst that pro­ base. tonates the leaving group and subsequently deprotonates the Preliminary results suggest that bgLA codes for the intracel­ water as it hydrolyzes the glycosyl enzyme intermediate (21, 32, lular [3-glucosidase. Efforts are currently being aimed at con­ 37). This highly conserved amino acid residue occurs in the firming this, as well as improving the stability and expression of same location as Asp-529 in C. wickerhamii (Fig. 3). The only bglB in other hosts. High expression of bglB in cellulolytic fungi glycosyl hydrolase from family 1 that differed in this conserva­ may yield a non-glucose-inhibited cellulase system. Such a sys­ tion was from creeping white clover (28), which contained an tem could be used in cellulose degradation and ultimately asparagine at this position. improve methods of fuel ethanol production. 524 SKORY AND FREER A!'PL. E1 VIRO . MICROBIaL.

A. B. 2 3 4 2 3 4

116 ­ 116­ 97- 97-

66- 66-

45-

45-

29- 29-

FIG. 4. SOS-PAGE and Western analysis of bglB clones. Total protein homogenates from recombinant E. coli SOLR were resolved by SOS-PAGE (10% polyacrylamide) and then stained with Coomassie R-250 (A) or transferred to polyvinylidene difluoride membranes for Western analysis using anti-l3-glucosidase antibody (B). E. coli SOLR isolates contained plasmids as indicated: lane 1, pBluescript II (KS-); lane 2, pBG2; lane 3, pBG6; lane 4, pBG28. Approximate molecular masses of MW-SDS-200 standards (Sigma) are given in kilodaltons.

ACKNOWLEDGMENTS 14. Gonzalez-Candelas, L., D. Ramon, and J. Polaina. 1990. Sequence and homology analysis of two genes encoding beta-glucosidases from Bacillus We thank Ellen Ball at University of ebraska, Lincoln, for her polymxa. Gene 95:31-38. efforts with production of the antisera against the purified cell-bound 15. Graebnitz, F., M. Seiss, K. P. Ruecknagel, and W. L. Staudenbauer. 1991. 13-glucosidase. The critical review of the manuscript by Robert Hespell Structure of the beta-glucosidase gene bgLA of Clostridium thermocellum: is gratefully acknowledged. sequence analysis reveals a superfamily of cellulases and beta-glycosidases including human lactase/phlorizin hydrolase. Eur. J. Biochem. 200:301-309. REFERE CES 16. Grethlein, H. K, and T. Dill. 1993. Final report: the cost of ethanol produc­ 1. Ausubel, F. M., R. Brent, R. K Kingston, D. D. Moore, J. A. Smith, J G. tion from lignocellulosic biomass-a comparison of selected alternative pro­ Seidman, and K. Struhl (ed.). 1987. Current protocols in molecular biology. cesses. U.S. Department of Agriculture, Philadelphia. John Wiley & Sons, Inc., New York. 17. Henrissat, B. 1991. A classification of glycosyl hydrolases based on amino 2. Beguin, P. 1990. Molecular biology of cellulose degradation. Annu. Rev. acid sequence similarities. Biochem. J. 280:309-316. Microbiol. 44:219-248. 18. Henrissat, 8., and A. Bairoch. 1993. ew families in the classification of 3. Breidt, F., Jr., and G. C. Stewart. 1987. ucleotide and deduced amino acid glycosyl hydrolases based on amino acid sequence similarities. Biochem. J. sequences of the Staphylococcus aureus phospho-l3-galactosidase gene. Appl. 293:781-788. Environ. Microbiol. 53:969-973. 19. Himmel, M. K, M. P. Tucker, S. M. Lastick, K. K. Oh, J. W. Fox, D. D. 4. DeBellis, G., M. Manoni, R. Pergolizzi, M. K Redolfi, and M. Luzzana. 1992. Spindler, and K. Grohman. 1986. Isolation and characterization of a 1,4-13­ A more stringent choice of primers can improve the performance of fluo­ D-glucan glucohydrolase from the yeast, Torulopsis wickerhamii. J. BioI. rescent automated 0 A sequencers. BioTechniques 13:892 97. Chern. 261:12948-12955. 5. Desrochers, M., L. Jurasek, and M. G. Paige. 1981. High production of 20. Hughes, M. A., K. Brown, A. Pancoro, B. S. Murray, K Oxtoby, and J. l3-glucosidase in Schizophyllum commune: isolation of the enzyme and effects Hughes. 1992. A molecular and biochemical analysis of structure of the of culture filtrates on cellulose hydrolysis. Appl. Environ. Microbiol. 23:222­ cyanogenic l3-glucosidase (linamarase) from cassava (Manihot esculenta 228. Cranz). Arch. Biochem. Biophys. 295:283-279. 6. EI Hassouni, M., B. Henrissat, M. Chippaux, and F. Barras. 1992. ucle­ 21. Kempton, J. 8., and S. G. Withers. 1992. Mechanism of Agrobacterium otide sequences of the arb genes, which control l3-g1ucoside utilization in l3-glucosidase: kinetic studies. Biochemistry 31:9961-9969. Erwinia chrysanlhemi: comparison with the Escherichia coli bgl operon and 22. Ladisch, M. R., J. Hong, M. Voloch, and G. T. Tsao. 1981. Cellulase kinetics. evidence for a new l3-glycohydrolase family 1 including enzymes from eu­ Basic Life Sci. 18:55 3. bacteria, archaebacteria, and humans. J. Bacterial. 174:765-777. 23. Leclerc, M., P. Gonde, A. Arnaud, R. Ratomahenina, P. Galzy, and M. 7. Freer, S. . 1985. Purification and characterization of the extracellular l3-glu­ Nicolas. 1984. The enzyme systems in a strain of Candida wickerhamii Meyer cosidase produced by Candida wickerhamii. Arch. Biochem. Biophys. 243: and Yarrow participating in the hydrolysis of cellodextrins. J. Gen. Appl. 515-522. Microbiol. 30:509-521. 8. Freer, S. '. 1991. Fermentation and aerobic metabolism of cellodextrins by 24. Little, S., P. Cartwright, C. Campbell, A. Prenneta, J. McChesney, A. Moun­ yeasts. Appl. Environ. Microbial. 57:655-659. tain, and M. Robinson. 1989. ucleotide sequence of a thermostable beta­ 9. Freer, S. N. 1993. Kinetic characterization of a l3-glucosidase from a yeast, glucosidase from Sulfolobus solfataricus. ucleic Acids Res. 17:7980-7980. Candida wickerhamii. J. BioI. Chern. 268:9337-9342. 25. Love, D. R., and P. L. Bergquist. 1988. Sequence structure and expression of 10. Freer, S. ., and R. W. Detroy. 1982. Direct fermentation of cellodextrins to a cloned beta-glucosidase gene from an extreme thermophile. Mol. Gen. ethanol by Candida wickerhamii and C. lusitaniae. Biotechnol. Lett. 4:453­ Genet. 213:84-92. 458. 26. Mantei, ., M. Villa, T. Enzler, H. Wacker, W. Boll, P. James, W. Hunziker, 11. Freer, S. N., and R. W. Detroy. 1983. Characterization of cellobiose fermen­ and G. Semenza. 1988. Complete primary structure of human and rabbit tations to ethanol by yeasts. Biotechnol. Bioeng. 25:541-557. lactase-phlorizin hydrolase: implications from biosynthesis, membrane an­ 12. Freer, S. ., and R. W. Detroy. 1985. Regulation of 13-1,4-glucosidase ex­ choring, and evolution of the enzyme. EMBO J. 7:2705-2713. pression by Candida wickerhamii. Appl. Environ. Microbiol. 50:152-159. 27. Matsudaira, P. 1987. Sequence from picomole quantities of protein electro­ 13. Freer, S. ., and R. V. Greene. 1990. Transport of glucose and cellobiose by blotted onto polyvinylidene difluoride membranes. J. BioI. Chern. Candida wickerhamii and Clavispora lusitaniae. J. BioI. Chern. 265:12864­ 262: 10035-1 0038. 12868. 28. Oxtoby, K, M. A. Dunn, A. Pancoro, and M. A. Hughes. 1991. Nucleotide VOL. 61, 1995 C. WICKERHAMII GE E FOR [3-GLUCOSIDASE 525

and derived amino acid sequence of the cyanogenic ~-glucosidase (linama­ mutagenesis of residues surrounding the active site nucleophile in ~-gJuco­ rase) from white clover (Trifolium repens L.). Plant Mol. BioI. 17:209-219. sidase from Agrobacterium faecalis. J. BioI. Chern. 267:10248-10251. 29. Philippsen, P., A. Stotz, and C. Scherf. 1991. DNA of Saccharomyces cerevi­ 35. Wakarchuk, W. W., N. M. Greenberg, D. G. Kilburn, R. C. Miller, and R. A. siae. Methods Enzymol. 194:169-182. Warren. ] 988. Structure and transcription analysis of the gene encoding a 30. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a cellobiase from Agrobacterium sp. strain ATCC 21400. J. Bacteriol. 170:301­ laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold 307. Spring Harbor, ..Y. 36. Withers, S. G., K. Rupitz, D. Trimbur, and R. A. J. Warren. ]992. Mecha­ 31. Schnetz, K., C. Toloczyki, and B. Rak. ] 987. ~-Glucosidase (bgl) operon of nistic consequences of mutation of the active site nucleophile Glu358 in Escherichia coli K-12: nucleotide sequence, genetic organization, and possi­ Agrobacterium ~-glucosidase. Biochemistry 31:9979-9985. ble evolutionary relationship to regulatory components of two Bacillus sub­ 37. Withers, S. G., R. A. J. Warren, I. P. Street, K. Rupitz, J. B. Kempton, and tilis genes. 1. Bacteriol. 169:2579-2590. R. Aebersold. ]990. Unequivocal demonstration of the involvement of a 32. Sinnott, M. L. 1990. Catalytic mechanisms of enzymatic glycosyl transfer. glutamate residue as a nucleophile in the mechanism of a "retaining" gly­ Chern. Rev. 90:1171-]202. cosidase. J. Am. Chern. Soc. 112:5887-5889. 33. Sternberg, D., P. Vijayakumar, and E. T. Reese. 1977. ~-Glucosidase: mi­ 38. Wright, R. M., M. D. Yablonsky, Z. P. Shalita, A. K. Goyal, and D. E. crobial production and effect on enzymatic hydrolysis of cellulose. Can. J. Eveleigh. ] 992. Cloning, characterization, and nucleotide sequence of a gene Microbiol. 23:139-147. encoding Microbispora bispora BglB, a thermostable ~-glucosidaseexpressed 34. Trimbur, D. E., R. A. J. Warren, and S. G. Withers. ] 992. Region-directed in Escherichia coli. Appl. Environ. Microbiol. 58:3455-3465.

upplied by U.S. Dept. of Ag I ationa' Center for Agricultura Utilization Research, Peoria, IHin