APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Feb. 1991, p. 359-365
Vol. 57, No. 2
0099-2240/91/020359-07$02.00/0 Copyright C 1991, American Society for Microbiology
Molecular Cloning, Expression, and Characterization of Endoglucanase Genes from Fibrobacter succinogenes ARI RICARDO CAVICCHIOLI*
AND
KENNETH WATSON
Department of Biochemistry, Microbiology, and Nutrition, University of New England, Armidale 2351, New South Wales, Australia Received 10 September 1990/Accepted 9 November 1990
A cosmid gene library was constructed in Escherichia coli from genomic DNA isolated from the ruminal anaerobe Fibrobacter succinogenes ARL. Clones were screened on carboxymethyl cellulose, and 8 colonies that produced large clearing zones and 25 colonies that produced small clearing zones were identified. Southern blot hybridization revealed the existence of at least three separate genes encoding cellulase activity. pRC093, which is representative of cosmid clones that produce large clearing zones, was subcloned in pGem-1, and the resulting hybrid pRCEH directed synthesis of endoglucanase activity localized on a 2.1-kb EcoRI-HindIII insert. Activity was expressed from this fragment when it was cloned in both orientations in pGem-1 and pGem-2, indicating that F. succinogenes promoters functioned successfully in E. coli. A high level of endoglucanase activity was detected on acid-swollen cellulose, ball-milled cellulose, and carboxymethyl cellulose; and a moderate level was detected on filter paper, Avicel, lichenan, and xylan. Most activity (80%) was localized in the periplasm of E. coli, with low but significant levels (16%) being detected in the extracellular medium. The periplasmic endoglucanase had an estimated molecular weight of 46,500, had an optimum temperature of 39°C, and exhibited activity over a broad pH range, with a maximum at pH 5.0.
Fibrobacter succinogenes is one of the most prolific cellulolytic microorganisms in the rumen, possessing a highly active cellulase system (16, 46). However, difficulty in isolating F. succinogenes (43) has resulted in a large proportion of studies concentrating on strain S85, which was originally isolated by Bryant and Doetsch in 1954 (7). Extensive research has focused on the characterization of various enzymes involved in the saccharification of lignocellulosic compounds. Enzyme activities associated with the breakdown of cellulose and hemicellulose from S85 include endoglucanase (30, 31), cellodextrinase (22), esterase (29), cellobiosidase (23), ,B-glucosidase (16), xylanase (41), and lichenase (10). Recently, Bacteroides succinogenes was reclassified to Fibrobacter succinogenes with two subspecies, F. succinogenes subsp. succinogenes and F. succinogenes subsp. elongata (33). Strain AR1 has not been extensively characterized; however, it appears to belong to F. succinogenes subsp. elongata (type strain HM2). Other strains belonging to F. succinogenes subsp. elongata include MM4 and MB4 (1). Evidence that AR1 may belong to F. succinogenes subsp. elongata is suggested by the fact that genomic DNA isolated from strains MM4, MB4, and HM2 (42a), like AR1 (7a), is not successfully digested with the restriction enzyme Sau3Al, whose recognition sequence is GATC. BamHI (whose core recognition sequence is GATC) also fails to digest AR1 DNA (7a). By comparison, genomic DNAs from strains BL2 and S85 are able to be restricted by BamHI (12). This property may be helpful in distinguishing the two subspecies. The cloning of separate genes or operons would facilitate the analysis of individual cellulase enzymes, allowing a better understanding of their role in lignocellulose breakdown in the rumen. Elucidation of the mechanisms involved in gene expression would allow an informed attempt at *
genetic manipulation of these bacteria. Furthermore, it would provide important knowledge that is required before the release of genetically engineered microorganisms can be considered responsible. In this study, a gene library from strain AR1 was constructed to allow analysis of the genetic and enzymatic components responsible for the organism's ability to actively degrade lignocellulose. The isolation from F. succinogenes AR1 of three distinct endoglucanase genes, and in particular, the characterization of one clone with a high level of activity, is reported. The nucleotide sequence and novel aspects of gene control in the endoglucanase gene from AR1 with a high level of activity are presented in a separate report (7b). MATERIALS AND METHODS
Bacterial strains and cloning vectors. F. succinogenes AR1 (formerly B. succinogenes [33]) was provided by Tom Bauchop. Escherichia coli HB101 and DH1 (17) were grown in 2 x YT medium containing (per liter) 16 g of tryptone, 10 g of yeast extract, and 5 g of sodium chloride. Incubations were conducted at 37°C in flasks placed in an orbital shaker operating at 200 opm (orbits per minute). Ampicillin was added at 100 ,ug/ml. For plate media, agar was added at 2%. Cosmid vector pHC79 (20) was used for cosmid cloning, and plasmids pGem-1, pGem-2, pGem7Zf+, and pGem7ZF (Promega, Sydney, Australia) were used for subcloning and exonuclease deletions. Gene library and screening for endoglucanase activity. Genomic DNA was isolated essentially by the procedure of Anderson and McKay (2). Genomic DNA was partially digested with HindlIl, and following separation on a 0.7% agarose gel, appropriately sized fractions were pooled and ligated into the HindIII site of pHC79. Recombinant cosmids were packaged and infected into E. coli DH1 by using the Promega Packagene kit. Cells were plated onto ampicillin (100 ,ug/ml)-containing plates, and 2,000 colonies were screened onto plates containing (per liter) 2 g of carboxy-
Corresponding author. 359
360
CAVICCHIOLI AND WATSON
methyl cellulose (CMC; low viscosity; Sigma Chemical Co., St. Louis, Mo.), 5 g of yeast extract, the minimal salt of Gilardi (13), 100 mg of ampicillin, and 50 mg of cycloheximide. Plates were incubated for 3 days at 37°C and then flooded with 1 mg of Congo red per ml for 15 min. Plates were drained and then washed with 1 M NaCl. Endoglucanase-positive clones were identified by the clearing zones surrounding the colonies (45). Thirty-three clones were identified as endoglucanase positive, with 8 large clearing zones (LCZs; 3 cm in diameter after 2 days of growth) and 25 small clearing zones (SCZs; 1 cm in diameter after 2 days of growth). Recombinants pRC093 and pDA84 were chosen as representatives of cosmid clones, producing large and small clearing zones, respectively (see Table 1). Subcloning and exonuclease deletions. Plasmid DNA was isolated by the alkaline lysis procedure (5). Plasmid pRCE was constructed by cloning an approximately 3.3-kb EcoRI fragment from pRC093 into pGem-1 (Table 1). A 2,075-bp EcoRI-HindIII fragment encoding FSendA (7b) was subcloned from pRCE into pGem-1, pGem-2, pGem7Zf+, and pGem7Zf-, generating pRCEH, pRCEH2, pRCZ+, and pRCZ-, respectively. The 1.55-kb HindIII-ApaI, 1.35-kb EcoRI-SphI, 1.38-kb EcoRI-Sau3AI, 1.62-kb EcoRI-XbaI, and 1.86-kb EcoRI-BglIll restriction enzyme fragments were cloned from pRCEH into pGem7Zf+. Exonuclease deletions of pRCZ+ and pRCZ- were generated by using the Erasea-base kit (Promega). Deletions from the HindIII end of pRCZ+ were obtained by first digesting it with Sacl, to produce a 3' overhang (to inhibit exonuclease III digestion), followed by digestion with HindIII to generate a 5' overhang. Nested sets of deletions were obtained by exonuclease III digestion for set time periods. Deletions from the EcoRI end were performed by first linearizing pRCZ- with XhoI (5' overhang), end filling with ot-phosphorothioate deoxynucleoside triphosphates (39), restricting with EcoRI, and sequentially digesting with exonuclease III. Southern hybridization. Cosmid clones were prepared from E. coli DH1 by using Qiagen plasmid preparation columns (Qiagen Inc., Studio City, Calif.). A small amount of E. coli genomic DNA was also extracted on the Qiagen columns because of the size (>20 kb) of the cosmid recombinants. DNA hybridization was performed essentially as described by Southern (42). Cosmids and F. succinogenes AR1 genomic DNA were digested to completion with HindIII. These digests, together with an EcoRI-Hindlll fragment from pRCEH and an approximately 6-kb Hindlll fragment encoding cellulase activity from pDA84 (8), were loaded onto 1% agarose gels and electrophoresed for 2 h at 50 V. Lambda DNA Hindlll and EcoRI-HindIII digests were included as molecular weight markers. Gels were capillary blotted onto N+-nylon filters (Amersham, Sydney, Australia) according to the directions of the manufacturer and hybridized by using the insert from pRCEH which was labeled with [32P]dATP by using a random priming kit (Bresa, Adelaide, Australia). High- and low-stringency hybridization and washing procedures were performed as outlined in the Amersham blotting and hybridization protocols for Hybond membranes. Cellular fractionation. Cells in the late logarithmic phase were harvested at 11,000 x g for 10 min, and the supernatant was taken as the extracellular phase. Periplasmic enzymes were prepared by the osmotic shock procedure (19), with modifications. Cells were washed in 10 mM Tris hydrochloride (pH 8), resuspended in 25% sucrose-10 mM Tris hydrochloride (pH 8), and shaken gently at 22°C for 10 min. Cells were pelleted (11,000 x g for 10 min) and resuspended in ice
APPL. ENVIRON. MICROBIOL.
water with further gentle shaking at 4°C for 10 min. Cells were pelleted, and the supernatant that was recovered was used as the periplasmic fraction. Cytoplasmic fractionation was performed by sonicating the remaining cell pellet (resuspended in 50 mM sodium citrate buffer [pH 5.7]) with a Branson sonifier fitted with a B-30 microtip at 70% maximum power and 50% duty cycle. The cytoplasmic fraction was the supernatant after centrifugation at 16,000 x g for 1 h at 40C. ,B-Galactosidase and alkaline phosphatase were assayed to determine the activities of these respective cytoplasmic (36) and periplasmic (19) marker enzymes. Activity was determined by a spectrophotometric method (34) by using paranitrophenyl (pNP) derivatives. Endoglucanase activity and molecular weight determination. Endoglucanase activity was assayed by measuring the release of reducing sugars (35) on 0.5% CMC, phosphoric acid-swollen cellulose (Avicel PH101 prepared by the method of Wood [48]), oatspelt xylan (batch no. X-0376; Sigma), lichenan (batch no. 124F-0106; Sigma), laminarin (batch no. 78F-3798; Sigma), ball-milled cellulose (Whatman no. 1; Whatman, Maidstone, England), and 5% Avicel PH101 (38-,um average particle size; Honeywill and Stein, Poole, United Kingdom), filter paper (Whatman No. 1), and sliva cotton buffered in 50 mM citrate phosphate buffer (pH 6). Activity was measured as (i) pNP liberated from pNPglucoside, pNP-xyloside, pNP-lactoside, and pNP-cellobioside (34) and (ii) glucose liberated from maltose, cellobiose, lactose, sucrose, melibiose, and melezitose with a glucose oxidase-peroxidase kit (Sigma). For evaluating the effect of pH, HCI-KCI (100 mM) was used for pH 1.0 to 2.2, citratephosphate buffer (various concentrations of citrate and phosphate were used, depending on the pH) was used for pH 2.6 to 7.0, 50 mM Tris hydrochloride was used for pH 7.2 to 9.0, and 50 mM glycine-NaOH was used for pH 9.0 to 12.5 (15). The pH of the final endoglucanase assay mixture (including added enzyme fractions) was measured directly, as the buffering capacities of final solutions were found to vary markedly. One unit of activity was defined as that which released 1 ,umol of glucose equivalent as reducing sugar or 1 ,umol of pNP per min at 39°C. Protein concentration was determined by the Lowry procedure (28). Thermal inactivation of endoglucanase activity was determined by measuring reducing sugar from CMC at 390C, after preincubation in 50 mM citrate phosphate (pH 6) at 4, 39, and 50°C. The mode of cellulase action was determined in an Ubbelohde viscometer by quantitating the decrease in viscosity with time caused by enzyme action in a solution of 1% CMC-50 mM citrate phosphate (pH 6). The molecular weight of the endoglucanase was determined from periplasmic extracts by gel filtration (3) by using Sephadex G100 (Pharmacia, Sydney, Australia). The column was calibrated with bovine serum albumin, ovalbumin, pepsin, and lysozyme, and the void volume was calculated with blue dextran (Sigma). The column was eluted with 50 mM citrate buffer (pH 6), and fractions were assayed for endoglucanase activity by measuring the release of reducing sugar by the para-hydroxybenzoic acid hydrazide method (26), with CMC used as the substrate.
RESULTS Isolation of endoglucanase genes from F. succinogenes AR1 genomic DNA. Two thousand cosmid clones were screened in E. coli DH1 for endoglucanase activity. Eight clones exhibited LCZs on CMC plates, and 25 clones exhibited SCZs. Plasmid pRC093 (Table 1), which was representative
flS
VOL. 57,
1991
Plasmid
F. SUCCINOGENES ENDOGLUCANASE GENES
361
TABLE 1. Recombinant plasmids and cosmids used in this study
or
Endoglucanase activity
Vector and construction
cosmid
pRC093 Cosmids 2 to 8 pDA84 pRCE PRCEH pRCEH2 pRCZ+ pRCZpRCZHA+ pRCZES+ pRCZEM+ pRCZEX+ pRCZEB+
pHC79 cosmid gene bank pHC79 cosmid gene bank pHC79 cosmid gene bank pGEM-1, approx. 3.3-kb Hindlll insert from pRC093 pGEM-1, 2.1-kb EcoRI-HindIII insert from pRC093 pGEM-2, 2.1-kb EcoRI-Hindlll insert from pRC093 pGEM7Zf+, 2.1-kb EcoRI-HindIII insert from pRC093 pGEM7Zf-, 2.1-kb EcoRI-HindIII insert from pRC093 pGEM7Zf+, 1.55-kb HindIII-ApaI insert from pRCEHa pGEM7Zf+, 1.35-kb EcoRI-SphI insert from pRCEH pGEM7Zf+, 1.38-kb EcoRI-Sau3AI insert from pRCEH pGEM7Zf+, 1.62-kb EcoRI-XbaI insert from pRCEH pGEM7Zf+, 1.86-kb EcoRI-BgIII insert from pRCEH
LCZ LCZ SCZ LCZ LCZ LCZ LCZ LCZ h
Exonuclease deletions pRCZ-1 pRCZ-2 pRCZ+1 pRCZ+2
pRCZ-, deletion from HindIll terminus, 1.97-kb insert pRCZ-, deletion from HindIll terminus, 1.74-kb insert pRCZ+, deletion from EcoRI terminus, 1.93-kb insert pRCZ+, deletion from EcoRI terminus, 1.58-kb insert
LCZ
LCZ
LCZ
a Restriction enzyme sites and exonuclease deletion sites have been mapped by DNA sequence analysis (7b). b-, No clearing zone and no detectable activity by reducing sugar assay.
Cf
LCZ, was isolated and completely digested with HindlIl,
-?coRI, or BamHI. Digests were ligated into appropriately roestricted pGem-1, transformed into
E. coli HB101, and gcreened for endoglucanase activity on CMC plates. A lumber of transformants derived from each digest gave 'MC-positive clones, indicating that all three restriction e'nzyme sites were absent from the functional endoglucanase ;ene. The EcoRI-derived clone pRCE harbored an insert of a pproximately 3.3 kb. HindlIl digestion of this insert reulted in a 2.1-kb EcoRI-HindIII fragment which was cloned iin pGem-1, generating pRCEH. Restriction map analysis of tihe pRCEH insert (Fig. 1) indicated an absence of sites for latII, NsiI, PstI, Sacl, Sall, SmaI, and XhoI. Fragments firom digestion with Sau3AI, which cut three times, and tpaI, BglII, SphI, and XbaI, which cut only once each, were Ii igated into pGem7Zf', transformed into E. coli HB101, and ,creened onto CMC plates. Only clone pRCEB' (EcoRI3glII insert) exhibited cellulase activity. However, activity vas not stably maintained in cells harboring this recombiiant, and the 2.1-kb EcoRI-HindIII fragment was considered ti o etemniu-ze
retito.nym
loubende . edogluc
v
act
iviy.
rgetta
Iwasmnoted
tihat cloning of DNA derived from F. succinogenes ARi in hGem-, pGem7Zf+n pGem7Zfr pHC79, 6 Gem1 pHC79'and pKK232-8 pKK2328 (6) pGem7Zf' pGem7Zf. Vvas generally successful, whereas cloning of recombinants C.onstructed in pTZ18U and pTZ19U (Biorad, Sydney, Aus-
tralia) and PBS- (Stratagene, Sydney, Australia) were frequently unstable. pGem7Zf+ and pGem7Zf- were most frequently used for subcloning and exonuclease deletion analysis because of the large number of suitable restriction enzyme sites in the multiple cloning site. Endoglucanase activity from the 2.1-kb fragment in pRCEH and pREH2 was independent of orientation, indicating that the gene was expressed from its own promoter in E. coli. Deletion analysis (Table 1) indicated that the functional gene was encoded within a fragment of approximately 1.7-kb delimited at the EcoRI end by pRCZ+1 and at the HindlIl end by pRCZ-1. Southern blot analysis. Seven of the recombinant cosmids producing LCZ, pRC093, and cosmids 2 to 7 (see Table 1 for description) showed equivalent bands of approximately 3.4 kb when they were hybridized with 32P-labeled probe synthesized from the pRCEH insert (Fig. 2). By contrast, cosmid 8 (LCZ) and pDA84 (SCZ) revealed an absence of hybridization under both high- and low-stringency condi-
hybridization and washing. Moreover, when the filter was stripped and then reprobed with a 6-kb HindIII fragment (cellulase encoding) derived from pDA84, only the corresponding 6-kb band in lanes B and N (Fig. 2a) indicated (data notgenes shown). These observations indihybridization and are encoded that2 three cated (i) pRC093 cosmids to 7, distinct which produce LCZs, (ii) by cosmid 8 (LCZ), tions of
and (iii) pDA84 (SCZ). Southern blot analysis of the EcoRI
HindlIl fragment from pRCEH hybridized to F. succinogenes AR1 genomic DNA digested with Hindlll showed a single band of hybridization under high- and low-stringency conditions. This indicated that F. succinogenes AR1 probably encodes only a single copy of the gene and that there is an absence of other genes showing a high degree of similarity 500 bp to it. It was noted that the genomic band (Fig. 2a, lane L) . to be smaller than the corresponding Restriction mapofthe21kbenoglucanasenappeared FIG. FIG. 1. Restriction map of the 2.1-kb endoglucanase-encoding cosmid bands (Fig.slightly 2a, lanes D through J). This may have
EV
A
V
Shl
mx
M B
H
iinsert from pRCEH. Restriction enzyme abbreviations: E, EcoRI,
X, ApaI; V, AvaI; S, SphI; M, Sau3Al; X, XbaI; B, BglII; H, HindIII. The insert was restricted with HindIll and ApaI to generate I)RCZHA' and with EcoRI combined with SphI, Sau3AI, XbaI, and gglIl to generate pRCZES', pRCZEM+, pRCZEX+, and pRCZEB respectively.
.
'
overloading. Determination of the cellular distribution of endoglucanase activity, been due to a gel
artifact,
perhaps caused by
Cellular localization of endoglucanase in E. coli.
as estimated by reducing sugar analysis by using the dinitrosalicylic acid method (32) or the procedure of Nelson (35)
362
APPL. ENVIRON. MICROBIOL.
CAVICCHIOLI AND WATSON
a ABC DE F GH I J K L MN
b ABC DE F G HI J K L MN
8~~~~~~U
'
FIG. 2. Southern blot analysis. (a) Hybridization of a 32P-labeled probe (2.1-kb EcoRI-Hindlll fragment from pRCEH) to Hindlll-digested cosmid clones (lanes B through J), F. succinogenes ARi genomic DNA (lane L), and 2.1- and 6-kb cellulase-encoding fragments from pRCO93 (lane M) and pDA84 (lane N), respectively. X Hindlll-EcoRl (lane A) and X Hindlll (lane K) were used as molecular weight markers. Sizes of bands marked by arrows for X Hindlll-EcoRI are as follows, from the top: 21.2, 4.3, 2.0, 1.4, and 0.95 kb. (b) Ethidium bromide-stained agarose gel of the gel used in panel a.
(see Materials and Methods), led to some inconsistencies in activity determinations, particularly in the extracellular fraction, in which high and variable blank values were obtained. The Nelson method (35), however, provided the most consistent results and was used throughout this study. Because of its relative purity, the periplasmic fraction was used for all subsequent enzyme activity determinations, including substrate utilization studies (see below). The cellular distribution for each fraction was confirmed by comparing the rates of fluidity change in an Ubbelohde viscometer. This method allowed estimation of the enzyme distribution by assessing endoglucanase activity by a purely physical method rather than by a chemical assay. Samples were appropriately diluted until a constant rate of change in fluidity was obtained for each fraction. E. coli cells harboring plasmids with either a >20-kb insert in pRC093 or a 2.1-kb insert in pRCEH produced clearing zones of equivalent sizes on CMC plates, and this was reflected by the endoglucanase activity (Table 2). The cellular distribution of endoglucanase for E. coli harboring either recombinant was mainly periplasmic (approximately 80%). A low but significant proportion (16%) was, however, found in the extracellular medium. Marker enzymes indicated the integrity of each fraction. Extracellular activity was also monitored, and endoglucaTABLE 2. Cellular distribution of endoglucanase in E. coli HB101 harboring pRCEH and E. coli DH1 harboring pRC093 Activity (mU/ml) of original culture Fraction
Endoglucanase"
pNP-
pNP-alkaline
galactosidase
phosphatase
NDc ND 0.4 (100) 0.4 (100)
ND 0.6 (100) ND 0.6 (100)
nase levels were found to increase proportionally with respect to the cell growth phase (data not shown). The activities of the cell-bound fractions (pelleted cellular membranes following sonication) were less than 1% (data not shown). Physiochemical properties of endoglucanase activity. It was noteworthy that expression of endoglucanase activity of E. coli clones was observed under aerobic or anaerobic growth conditions. Cellobiose or glucose (100 mM) had no effect on the expression of endoglucanase activity, as determined by clearing zone size on plates containing CMC and glucose or cellobiose. Endoglucanase pH profiles were constructed, and activity was high (>50% of maximum activity) between pH 4.5 and 8.0, peaking at about pH 5.0 for E. coli HB101 harboring pRCEH (Fig. 3). Endoglucanase activity showed a sharp decrease below pH 4.5 and a more gradual decrease at alkaline pH values. Endoglucanase activity was maximum at
100
.
w
o
Extracellular Periplasmic Cytoplasmic Total
6
(16)b
30 (78) 2 (6) 38 (100)
pRCEH
7 32 2 41
(17) (78) (5) (100)
a Endoglucanase activity was determined by the method of Nelson (35), and cellular distribution was confirmed by a viscometric assay. b Values in parentheses are percentages of total activity. ' ND, Not detectable (0.1 mU/ml or less).
0
50
z w w 0-
pRC093
75
25r
3
4
5
6
7
8
9
10 11 12
pH
FIG. 3. Effect of pH on the endoglucanase activity of E. coli HB101 harboring pRCEH.
VOL. 57, 1991
F. SUCCINOGENES ENDOGLUCANASE GENES
TABLE 3. Hydrolysis of glycan substratesa after 24 h of incubation with periplasmic enzymes from E. coli HB101 harboring pRCEH
1001
Substrate
Reducing sugar
% Activity relative to CMC
26.8b 10.9 8.5 2.4 46.2 165.4 NDc 13.4 15.8
100
Avicel PH101 Filter paper Cotton Ball-milled cellulose Acid-swollen cellulose Laminarin Lichenan Xylan
751 CMC CD
363
501
z w
0
251
I
1
2
3
4
5
6
7
8
9
TIME (HR) FIG. 4. Effect of preincubation at 4°C (A), 39°C (0), and 50°C (A) on the endoglucanase activity of E. coli HB101 harboring pRCEH.
37 to 39°C, decreasing markedly at elevated temperatures. This was reflected in thermal stability at 39°C compared with that at 50°C (Fig. 4). At 39°C, the activities of the enzyme fractions, which were buffered in 50 mM citrate phosphate (pH 6), remained high (75 to 100%) for 8 h, whereas endoglucanase was completely lost after 3 h when it was stored at 50°C. By comparison, endoglucanase activity was retained for several months at 0 to 4°C. The molecular weight of the endoglucanase was estimated by gel filtration to be approximately 46,500. Mode of action and substrate specificity. As measured with an Ubbelohde viscometer, a rapid increase in CMC fluidity was observed with periplasmic enzymes from E. coli harboring pRCEH, and this was indicative of an endo-acting enzyme. Activity on natural substrate (cellobiose) or substituted substrate (pNP-glucoside) was not detectable. These substrates are diagnostic for P-glucosidase activity. Low activity on pNP-cellobioside, 0.72 nmol/min/mg of protein, indicated that exoglucanase activity was also minimal. Similarly, activity was not detected on other natural or substituted short-chain glycans (maltose, lactose, sucrose, melibiose, melezitose, pNP-xyloside, pNP-lactoside, and pNPglucoside). It was noted that trehalose was actively hydrolyzed by periplasmic enzymes from E. coli HB101 and DH1 (7a). Activity on recalcitrant substrates such as filter paper, cotton, and Avicel was significant (Table 3). Less crystalline celluloses were more extensively hydrolyzed, with the highest activity detected on acid-swollen cellulose. It was noted that hydrolysis of CMC resulted in the rapid production of reducing sugar, with >70% maximum production observed after 1 h and essentially 100% observed by 4 h. Specific activity on CMC was about 70 nmol/min/mg of protein (30-min assay). Activity on lichenan (1,3-1,4-4-D-glucan), but not on laminarin (1,3-3-D-glucan), indicated that it was probably the 1,4-p-D-glucan component of lichenan that was being hydrolyzed. It was noteworthy that the endoglucanase was also active on xylan.
(p.g of glucose)
41 32 9 173 618 ND 50 59
' Reducing sugars were determined by the method of Nelson (35) after incubation of periplasmic enzymes in 50 mM sodium citrate buffer (pH 6) for 24 h. Each substrate was 5 mg/ml, except for Avicel, filter paper, and cotton, which were 50 mg/ml. b Production of reducing sugar on CMC at 1 h was 72% of the level at 24 h and was essentially 100% by 4 h. ND, Not detectable (1 jLg or less in 24 h).
DISCUSSION
Genomic DNA from F. succinogenes AR1 was successfully cloned into E. coli by using cosmid pHC79, and a number of recombinants expressed endoglucanase activity. Distinct LCZs and SCZs on CMC plates were features of endoglucanase expression in E. coli. A similar phenomenon has been reported by other authors during the construction of gene libraries of Cellulomonas (14), Clostridium thermocellum (40), Ruminococcus albus (21), and F. succinogenes (8, 9, 24). In all cases a correlation between clearing zone diameter and endoglucanase activity was observed. Southern blot analysis revealed the isolation of three separate genes encoding endoglucanase (Fig. 2). It is possible that other unique endoglucanase genes are represented in the family of 25 clones that produced SCZs. The fact that two separate genes were found for the eight clones that produced LCZs indicates that this possibility is likely. Multiple endoglucanase genes have been identified for other cellulolytic bacteria, and in C. thermocellum, for example, 15 distinct endoglucanase genes have been isolated (18). In F. succinogenes S85, a number of endoglucanases have been identified, and three have been characterized in particular (30, 31). It will thus be useful to isolate and characterize all the genes encoding endoglucanases from our gene bank, because this will lead to a comprehensive understanding of the endoglucanase requirements for this microorganism. In view of the recent reclassification of Fibrobacter as a genus distinct from the genus Bacteroides and, furthermore, to the suggestion that it may form a new family distinct from the family Bacteroidaceae (33), it was significant that the endoglucanase genes in pRCEH and pRCEH2 were expressed in E. coli, indicating that promoters from F. succinogenes may be recognized by E. coli RNA polymerase. Data from DNA sequencing and primer extension analysis to map transcription start sites would be useful to confirm the presence of functional promoters within the cloned insert. Delimitation of the boundaries of the functional gene in pRCEH was undertaken by subcloning restriction fragments in pGem7Zf+ and constructing nested deletions from pRCZ+ and pRCZ-. As a result, a 1.7-kb fragment was identified as being important for the expression of endoglucanase. This indicated that the possible molecular weight of the encoded protein could be up to 62,000. However, preliminary data
364
indicated that the molecular weight of the periplasmic enwas 46,500, which suggested that regions in addition to the coding sequence are required for activity. In F. succinogenus ARI, 80% of endoglucanase activity was found in the extracellular medium (7a), and a similar level was found in strain S85 (16). In E. coli harboring pRCEH and pRC093, most activity (approximately 80%) was localized in the periplasm (Table 2). E. coli is not noted for its ability to secrete cloned products into the extracellular medium; however, secretion to the periplasm is normally efficient (38). Most cloned cellulases are secreted to the periplasm, and DNA sequence analysis has revealed that all cellulases to date have signal peptides predicted for their amino acid sequences (4). Maximum endoglucanase activity was at pH 5.0, which is around the range (pH 5.5 to 7.0) reported for cloned endoglucanases from F. succinogenes (44), R. albus (25, 37, 47), and Bacteroides ruminicola (48). Maximum activity of extracellular endoglucanase from F. succinogenes ARI is observed at pH 5.5 (7a), which is similar to that for the single gene product cloned in E. coli. F. succinogenes exhibits a diverse range of enzyme activities with the ability to solubilize lignocellulosic materials (23). Some of these properties were reflected in the broad range of substrates hydrolyzed by the gene product of
doglucanase
pRCEH (Table 3). Most endoglucanases that have been cloned from ruminal bacteria show broad-spectrum activity on cellulosic compounds. Kinetic studies were not under-
taken; however, it was noteworthy that activities on acidswollen cellulose and ball-milled cellulose were higher than that on CMC. Initial production of reducing sugars on CMC was, in fact, high, but within 4 h, hydrolytic action essentially ceased. This can be explained by an endoglucanase attack on the portions of CMC that have at least three or more consecutive unsubstituted glucose units, resulting in only about 2% of CMC that can be readily hydrolyzed (11, 27). ACKNOWLEDGMENTS We are grateful to Peter East, John Pemberton, and coworkers Keith Gregg, Anil Lachke, Philip Vercoe, and Cheryl Ware for valuable advice and discussion; Tom Bauchop for providing cultures of F. succinogenes AR1; Michelle Deegenaars for technical expertise; and Jean Hansford for typing the manuscript. We also acknowledge in particular Ronald Teather, Geoff Hazlewood, Bryan White, and David Wilson for preprints and recent publications and David Stahl and Chuzhao Lin for communication of unpublished data.
REFERENCES 1. Amann, R. I., L. Krumholz, and D. A. Stahl. 1990. Fluorescent
oligonucleotide probing of whole cells for determinative, phylogenetic, and environmental studies in microbiology. J. Bacteriol. 172:762-770. 2. Anderson, D. G., and L. L. McKay. 1983. Simple and rapid method for isolating large plasmid DNA from lactic streptococci. Appl. Environ. Microbiol. 46:549-552. 3. Andrews, P. 1964. Estimation of molecular weights of proteins
by Sephadex gel filtration. Biochem. J. 91:222-233. Beguin, P., J. Millet, S. Chauvaux, E. Yague, P. Tomme, and J.-P. Aubert. 1989. Genetics of bacterial cellulases, p. 57-72. In M. P. Coughlan (ed.), Enzyme systems for lignocellulose degradation. Elsevier Applied Science Publications, London. 5. Birnboim, H. C., and J. Doly. 1979. A rapid alkaline extraction procedure for screening recombinant plasmid DNA. Nucleic Acids Res. 12:4411-4427. 6. Brosius, J. 1984. Plasmid vectors for the selection of promoters. Gene 27:151-160. 7. Bryant, M. P., and R. N. Doetsch. 1954. A study of actively 4.
APPL. ENVIRON. MICROBIOL.
CAVICCHIOLI AND WATSON
cellulolytic rod shaped bacteria of the ovine rumen. J. Dairy Sci. 37:1176-1183. 7a.Cavicchioli, R. Unpublished data. 7b.Cavicchioli, R., P. D. East, and K. Watson. Submitted for
publication.
8. Cavicchioli, R., D. H.-L. Lai, and K. Watson. 1989. Cloning and expression of cellulase genes from Bacteroides succinogenes, p. 153-156. In M. Sleigh (ed.), Eighth Australian biotechnology conference. Australian Biotechnology Association, Sydney, Australia. 9. Crosby, B., B. Collier, D. Y. Thomas, R. M. Teather, and J. D. Erfle. 1984. Cloning and expression in Escherichia coli of cellulase genes from Bacteroides succinogenes, p. 573-576. In S. Hasain (ed.), Fifth Canadian Bioenergy R & D Seminar. Elsevier Applied Science Publications, Amsterdam. 10. Erfle, J. D., R. M. Teather, P. J. Wood, and J. E. Irvin. 1988. Purification and properties of a 1,3-1,4-3-D-glucanase (lichenase, 1,3-1,4-13-D-glucan 4-glucanohydrolase, EC 3.2.1.73) from Bacteroides succinogenes cloned in Escherichia coli. Biochem. J. 255:833-841. 11. Eriksson, K.-E., and B. H. Hollmark. 1969. Kinetic studies of the action of cellulase upon sodium carboxymethyl cellulose. Arch. Biochem. Biophys. 133:233-237. 12. Flint, H. J., C. A. McPherson, G. Avgustin, and C. S. Stewart. 1990. Use of a cellulase encoding gene probe to reveal restriction fragment length polymorphisms among ruminal strains of Bacteroides succinogenes. Curr. Microbiol. 20:63-67. 13. Gilardi, G. L. 1971. Characterization of nonfermentive nonfastidious gram negative bacteria encountered in medical bacteriology. J. Appl. Bacteriol. 34:623-644. 14. Gilkes, N. R., M. L. Langsford, D. G. Kilburn, R. C. Miller Jr., and R. A. J. Warren. 1984. Mode of action and substrate specificities of cellulases from cloned bacterial genes. J. Biol. Chem. 259:10455-10459. 15. Gomori, G. 1955. Preparation of buffers for use in enzyme studies. Methods Enzymol. 1:138-146. 16. Groleau, D., and C. W. Forsberg. 1981. Cellulolytic activity of the rumen bacterium Bacteroides succinogenes. Can. J. Microbiol. 27:517-530. 17. Hanahan, D. 1983. Studies on transformation of Escherichia coli with plasmids. J. Mol. Biol. 166:557-580. 18. Hazlewood, G. P., M. P. M. Romaniec, K. Davidson, 0. Grepinet, P. Beguin, J. Millet, 0. Raynaud, and J.-P. Aubert. 1988. A catalogue of Clostridium thermocellum endoglucanase, j-glucosidase and xylanase genes located in Escherichia coli. FEMS Microbiol. Lett. 51:231-236. 19. Heppel, L. A. 1971. The concept of periplasmic enzymes, p. 223-347. In L. I. Rothfield (ed.), Structure and function of biological membranes. Academic Press, Inc., New York. 20. Hohn, B., and J. Collins. 1980. A small cosmid for efficient cloning of large DNA fragments. Gene 11:291-298. 21. Howard, G. T., and B. A. White. 1988. Molecular cloning and expression of cellulase genes from Ruminococcus albus 8 in Escherichia coli bacteriophage X. Appl. Environ. Microbiol. 54:1752-1755. 22. Huang, L., and C. W. Forsberg. 1987. Isolation of a cellodextrinase from Bacteroides succinogenes. Appl. Environ. Microbiol. 53:1034-1041. 23. Huang, L., M. McGavin, C. W. Forsberg, J. S. Lam, and K. J. Cheng. 1990. Antigenic nature of the chloride stimulated cellobiosidase and other cellulases of Fibrobacter succinogenes subsp. succinogenes S85 and related fresh isolates. Appl. Environ. Microbiol. 56:1229-1234. 24. Irvin, J. E., and R. M. Teather. 1988. Cloning and expression of a Bacteroides succinogenes mixed linkage 3-glucanase (1,3-1,4,-D-glucan 4-glucanohydrolase) gene in Escherichia coli. Appl. Environ. Microbiol. 54:2672-2676. 25. Kawai, S., H. Honda, T. Tanase, M. Taya, S. IiJima, and T. Kobayashi. 1987. Molecular cloning of Ruminococcus albus cellulase gene. Agric. Biol. Chem. 51:59-63. 26. Koziol, M. J. 1981. An evaluation of the alkaline para-hydroxybenzoic acid hydrazide procedure for the determination of reducing sugars. Anal. Chim. Acta 128:195-205.
VOL. 57, 1991
27. Kyriacou, A., R. C. MacKenzie, and R. J. Neufeld. 1987. Detection and characterization of the specific and nonspecific endoglucanases of Trichoderma reesei: evidence demonstrating endoglucanase activity by cellobiohydrolase II. Enzyme Microb. Technol. 9:25-32. 28. Layne, E. 1957. Spectrophotometric and turbidometric methods for measuring proteins. Methods Enzymol. 3:447-454. 29. McDermid, K. P., R. MacKenzie, and C. W. Forsberg. 1990. Esterase activities of Fibrobacter succinogenes subsp. succinogenes S85. Appl. Environ. Microbiol. 56:127-132. 30. McGavin, M., and C. W. Forsberg. 1988. Isolation and characterization of endoglucanases 1 and 2 from Bacteroides succinogenes S85. J. Bacteriol. 170:2914-2922. 31. McGavin, M., C. W. Forsberg, B. Crosby, A. W. Bell, D. Dignard., and D. Y. Thomas. 1989. Structure of the cel-3 gene from Fibrobacter succinogenes S85 and characteristics of the encoded gene product, endoglucanase 3. J. Bacteriol. 171:55875595. 32. Miller, G. L. 1959. Use of dinitrosalicylic acid reagent for determination of reducing sugar. Anal. Biochem. 31:426-428. 33. Montgomery, L., B. Flesher, and D. Stahl. 1988. Transfer of Bacteroides succinogenes (Hungate) to Fibrobacter gen. nov. as Fibrobacter succinogenes comb. nov. and description of Fibrobacter intestinalis sp. nov. Int. J. Syst. Bacteriol. 38:430435. 34. Nath, R. L., and N. Rydon. 1954. The influence of structure on the hydrolysis of substituted phenyl-p-D-glucosides by emulsin. Biochem. J. 57:1-10. 35. Nelson, N. 1944. A photometric adaptation of the Somogyi method for the determination of glucose. J. Biol. Chem. 153: 375-380. 36. Neu, H. C., and L. A. Heppel. 1965. The release of enzymes from Escherichia coli by osmotic shock and during formation of spheroplasts. J. Biol. Chem. 240:3685-3692. 37. Ohmiya, K., K. Nagashima, T. Kajino, E. Goto, A. Tsukada, and S. Shimizu. 1988. Cloning of the cellulase gene from Ruminococcus albus and its expression in Escherichia coli. Appl. Environ. Microbiol. 54:1511-1515. 38. Oliver, D. B. 1987. Periplasm and protein secretion, p. 56-69. In F. C. Neidhardt, J. L. Ingraham, K. B. Low, B. Magasanik, M. Schaechter, and H. E. Umbarger (ed.), Escherichia coli and
F. SUCCINOGENES ENDOGLUCANASE GENES
365
Salmonella typhimurium: cellular and molecular biology. American Society for Microbiology, Washington, D.C. 39. Putney, S. D., S. J. Benkovic, and P. R. Schimmel. 1981. A DNA fragment with an alpha-phosphorothioate nucleotide at one end is asymmetrically blocked from digestion by exonuclease III and can be replicated in vivo. Proc. Natl. Acad. Sci. USA 78:7350-7454. 40. Romaniec, M. P. M., N. G. Clarke, and G. P. Hazlewood. 1987. Molecular cloning of Clostridium thermocellum DNA and the expression of further novel endo-P-1,4-glucanase genes in Escherichia coli. J. Gen. Microbiol. 133:1297-1307. 41. Sipat, A., K. A. Taylor, R. Y. C. Lo, C. W. Forsberg, and P. J. Krell. 1987. Molecular cloning of a xylanase gene from Bacteroides succinogenes and its expression in Escherichia coli. Appl. Environ. Microbiol. 53:477-481. 42. Southern, E. M. 1975. Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Mol. Biol. 98:503-517. 42a.Stahl, D. A., and C. Lin. Personal communication. 43. Stewart, C. S., C. Paniagua, D. Dinsdale, K.-J. Cheng, and S. H. Garrow. 1981. Selective isolation and characteristics of Bacteroides succinogenes from the rumen of a cow. Appl. Environ. Microbiol. 41:504-510. 44. Taylor, K. A., B. Crosby, M. McGavin, C. W. Forsberg, and D. Y. Thomas. 1987. Characteristics of the endoglucanase encoded by a cel gene from Bacteroides succinogenes expressed in Escherichia coli. Appl. Environ. Microbiol. 53:4146. 45. Teather, R. M., and P. J. Wood. 1982. Use of Congo redpolysaccharide interactions in enumeration and characterization of cellulolytic bacteria from the bovine rumen. Appl. Environ. Microbiol. 43:777-780. 46. Van Gylswyk, N. O., and J. J. T. K. Van der Toorn. 1986. Enumeration of Bacteroides succinogenes in the rumen of sheep fed maize-straw diets. FEMS Microbiol. Lett. 38:205-209. 47. Ware, C. E., T. Bauchop, and K. Gregg. 1989. The isolation and comparison of cellulase genes from two strains of Ruminococcus albus. J. Gen. Microbiol. 135:921-930. 48. Wood, T. M. 1970. The cellulase of Fusarium solani: purification and specificity of the P-(1,4)-glucanase and the P-D-glucosidase components. Biochem. J. 121:353-362.
Vol. 57, No. 2
0099-2240/91/020359-07$02.00/0 Copyright C 1991, American Society for Microbiology
Molecular Cloning, Expression, and Characterization of Endoglucanase Genes from Fibrobacter succinogenes ARI RICARDO CAVICCHIOLI*
AND
KENNETH WATSON
Department of Biochemistry, Microbiology, and Nutrition, University of New England, Armidale 2351, New South Wales, Australia Received 10 September 1990/Accepted 9 November 1990
A cosmid gene library was constructed in Escherichia coli from genomic DNA isolated from the ruminal anaerobe Fibrobacter succinogenes ARL. Clones were screened on carboxymethyl cellulose, and 8 colonies that produced large clearing zones and 25 colonies that produced small clearing zones were identified. Southern blot hybridization revealed the existence of at least three separate genes encoding cellulase activity. pRC093, which is representative of cosmid clones that produce large clearing zones, was subcloned in pGem-1, and the resulting hybrid pRCEH directed synthesis of endoglucanase activity localized on a 2.1-kb EcoRI-HindIII insert. Activity was expressed from this fragment when it was cloned in both orientations in pGem-1 and pGem-2, indicating that F. succinogenes promoters functioned successfully in E. coli. A high level of endoglucanase activity was detected on acid-swollen cellulose, ball-milled cellulose, and carboxymethyl cellulose; and a moderate level was detected on filter paper, Avicel, lichenan, and xylan. Most activity (80%) was localized in the periplasm of E. coli, with low but significant levels (16%) being detected in the extracellular medium. The periplasmic endoglucanase had an estimated molecular weight of 46,500, had an optimum temperature of 39°C, and exhibited activity over a broad pH range, with a maximum at pH 5.0.
Fibrobacter succinogenes is one of the most prolific cellulolytic microorganisms in the rumen, possessing a highly active cellulase system (16, 46). However, difficulty in isolating F. succinogenes (43) has resulted in a large proportion of studies concentrating on strain S85, which was originally isolated by Bryant and Doetsch in 1954 (7). Extensive research has focused on the characterization of various enzymes involved in the saccharification of lignocellulosic compounds. Enzyme activities associated with the breakdown of cellulose and hemicellulose from S85 include endoglucanase (30, 31), cellodextrinase (22), esterase (29), cellobiosidase (23), ,B-glucosidase (16), xylanase (41), and lichenase (10). Recently, Bacteroides succinogenes was reclassified to Fibrobacter succinogenes with two subspecies, F. succinogenes subsp. succinogenes and F. succinogenes subsp. elongata (33). Strain AR1 has not been extensively characterized; however, it appears to belong to F. succinogenes subsp. elongata (type strain HM2). Other strains belonging to F. succinogenes subsp. elongata include MM4 and MB4 (1). Evidence that AR1 may belong to F. succinogenes subsp. elongata is suggested by the fact that genomic DNA isolated from strains MM4, MB4, and HM2 (42a), like AR1 (7a), is not successfully digested with the restriction enzyme Sau3Al, whose recognition sequence is GATC. BamHI (whose core recognition sequence is GATC) also fails to digest AR1 DNA (7a). By comparison, genomic DNAs from strains BL2 and S85 are able to be restricted by BamHI (12). This property may be helpful in distinguishing the two subspecies. The cloning of separate genes or operons would facilitate the analysis of individual cellulase enzymes, allowing a better understanding of their role in lignocellulose breakdown in the rumen. Elucidation of the mechanisms involved in gene expression would allow an informed attempt at *
genetic manipulation of these bacteria. Furthermore, it would provide important knowledge that is required before the release of genetically engineered microorganisms can be considered responsible. In this study, a gene library from strain AR1 was constructed to allow analysis of the genetic and enzymatic components responsible for the organism's ability to actively degrade lignocellulose. The isolation from F. succinogenes AR1 of three distinct endoglucanase genes, and in particular, the characterization of one clone with a high level of activity, is reported. The nucleotide sequence and novel aspects of gene control in the endoglucanase gene from AR1 with a high level of activity are presented in a separate report (7b). MATERIALS AND METHODS
Bacterial strains and cloning vectors. F. succinogenes AR1 (formerly B. succinogenes [33]) was provided by Tom Bauchop. Escherichia coli HB101 and DH1 (17) were grown in 2 x YT medium containing (per liter) 16 g of tryptone, 10 g of yeast extract, and 5 g of sodium chloride. Incubations were conducted at 37°C in flasks placed in an orbital shaker operating at 200 opm (orbits per minute). Ampicillin was added at 100 ,ug/ml. For plate media, agar was added at 2%. Cosmid vector pHC79 (20) was used for cosmid cloning, and plasmids pGem-1, pGem-2, pGem7Zf+, and pGem7ZF (Promega, Sydney, Australia) were used for subcloning and exonuclease deletions. Gene library and screening for endoglucanase activity. Genomic DNA was isolated essentially by the procedure of Anderson and McKay (2). Genomic DNA was partially digested with HindlIl, and following separation on a 0.7% agarose gel, appropriately sized fractions were pooled and ligated into the HindIII site of pHC79. Recombinant cosmids were packaged and infected into E. coli DH1 by using the Promega Packagene kit. Cells were plated onto ampicillin (100 ,ug/ml)-containing plates, and 2,000 colonies were screened onto plates containing (per liter) 2 g of carboxy-
Corresponding author. 359
360
CAVICCHIOLI AND WATSON
methyl cellulose (CMC; low viscosity; Sigma Chemical Co., St. Louis, Mo.), 5 g of yeast extract, the minimal salt of Gilardi (13), 100 mg of ampicillin, and 50 mg of cycloheximide. Plates were incubated for 3 days at 37°C and then flooded with 1 mg of Congo red per ml for 15 min. Plates were drained and then washed with 1 M NaCl. Endoglucanase-positive clones were identified by the clearing zones surrounding the colonies (45). Thirty-three clones were identified as endoglucanase positive, with 8 large clearing zones (LCZs; 3 cm in diameter after 2 days of growth) and 25 small clearing zones (SCZs; 1 cm in diameter after 2 days of growth). Recombinants pRC093 and pDA84 were chosen as representatives of cosmid clones, producing large and small clearing zones, respectively (see Table 1). Subcloning and exonuclease deletions. Plasmid DNA was isolated by the alkaline lysis procedure (5). Plasmid pRCE was constructed by cloning an approximately 3.3-kb EcoRI fragment from pRC093 into pGem-1 (Table 1). A 2,075-bp EcoRI-HindIII fragment encoding FSendA (7b) was subcloned from pRCE into pGem-1, pGem-2, pGem7Zf+, and pGem7Zf-, generating pRCEH, pRCEH2, pRCZ+, and pRCZ-, respectively. The 1.55-kb HindIII-ApaI, 1.35-kb EcoRI-SphI, 1.38-kb EcoRI-Sau3AI, 1.62-kb EcoRI-XbaI, and 1.86-kb EcoRI-BglIll restriction enzyme fragments were cloned from pRCEH into pGem7Zf+. Exonuclease deletions of pRCZ+ and pRCZ- were generated by using the Erasea-base kit (Promega). Deletions from the HindIII end of pRCZ+ were obtained by first digesting it with Sacl, to produce a 3' overhang (to inhibit exonuclease III digestion), followed by digestion with HindIII to generate a 5' overhang. Nested sets of deletions were obtained by exonuclease III digestion for set time periods. Deletions from the EcoRI end were performed by first linearizing pRCZ- with XhoI (5' overhang), end filling with ot-phosphorothioate deoxynucleoside triphosphates (39), restricting with EcoRI, and sequentially digesting with exonuclease III. Southern hybridization. Cosmid clones were prepared from E. coli DH1 by using Qiagen plasmid preparation columns (Qiagen Inc., Studio City, Calif.). A small amount of E. coli genomic DNA was also extracted on the Qiagen columns because of the size (>20 kb) of the cosmid recombinants. DNA hybridization was performed essentially as described by Southern (42). Cosmids and F. succinogenes AR1 genomic DNA were digested to completion with HindIII. These digests, together with an EcoRI-Hindlll fragment from pRCEH and an approximately 6-kb Hindlll fragment encoding cellulase activity from pDA84 (8), were loaded onto 1% agarose gels and electrophoresed for 2 h at 50 V. Lambda DNA Hindlll and EcoRI-HindIII digests were included as molecular weight markers. Gels were capillary blotted onto N+-nylon filters (Amersham, Sydney, Australia) according to the directions of the manufacturer and hybridized by using the insert from pRCEH which was labeled with [32P]dATP by using a random priming kit (Bresa, Adelaide, Australia). High- and low-stringency hybridization and washing procedures were performed as outlined in the Amersham blotting and hybridization protocols for Hybond membranes. Cellular fractionation. Cells in the late logarithmic phase were harvested at 11,000 x g for 10 min, and the supernatant was taken as the extracellular phase. Periplasmic enzymes were prepared by the osmotic shock procedure (19), with modifications. Cells were washed in 10 mM Tris hydrochloride (pH 8), resuspended in 25% sucrose-10 mM Tris hydrochloride (pH 8), and shaken gently at 22°C for 10 min. Cells were pelleted (11,000 x g for 10 min) and resuspended in ice
APPL. ENVIRON. MICROBIOL.
water with further gentle shaking at 4°C for 10 min. Cells were pelleted, and the supernatant that was recovered was used as the periplasmic fraction. Cytoplasmic fractionation was performed by sonicating the remaining cell pellet (resuspended in 50 mM sodium citrate buffer [pH 5.7]) with a Branson sonifier fitted with a B-30 microtip at 70% maximum power and 50% duty cycle. The cytoplasmic fraction was the supernatant after centrifugation at 16,000 x g for 1 h at 40C. ,B-Galactosidase and alkaline phosphatase were assayed to determine the activities of these respective cytoplasmic (36) and periplasmic (19) marker enzymes. Activity was determined by a spectrophotometric method (34) by using paranitrophenyl (pNP) derivatives. Endoglucanase activity and molecular weight determination. Endoglucanase activity was assayed by measuring the release of reducing sugars (35) on 0.5% CMC, phosphoric acid-swollen cellulose (Avicel PH101 prepared by the method of Wood [48]), oatspelt xylan (batch no. X-0376; Sigma), lichenan (batch no. 124F-0106; Sigma), laminarin (batch no. 78F-3798; Sigma), ball-milled cellulose (Whatman no. 1; Whatman, Maidstone, England), and 5% Avicel PH101 (38-,um average particle size; Honeywill and Stein, Poole, United Kingdom), filter paper (Whatman No. 1), and sliva cotton buffered in 50 mM citrate phosphate buffer (pH 6). Activity was measured as (i) pNP liberated from pNPglucoside, pNP-xyloside, pNP-lactoside, and pNP-cellobioside (34) and (ii) glucose liberated from maltose, cellobiose, lactose, sucrose, melibiose, and melezitose with a glucose oxidase-peroxidase kit (Sigma). For evaluating the effect of pH, HCI-KCI (100 mM) was used for pH 1.0 to 2.2, citratephosphate buffer (various concentrations of citrate and phosphate were used, depending on the pH) was used for pH 2.6 to 7.0, 50 mM Tris hydrochloride was used for pH 7.2 to 9.0, and 50 mM glycine-NaOH was used for pH 9.0 to 12.5 (15). The pH of the final endoglucanase assay mixture (including added enzyme fractions) was measured directly, as the buffering capacities of final solutions were found to vary markedly. One unit of activity was defined as that which released 1 ,umol of glucose equivalent as reducing sugar or 1 ,umol of pNP per min at 39°C. Protein concentration was determined by the Lowry procedure (28). Thermal inactivation of endoglucanase activity was determined by measuring reducing sugar from CMC at 390C, after preincubation in 50 mM citrate phosphate (pH 6) at 4, 39, and 50°C. The mode of cellulase action was determined in an Ubbelohde viscometer by quantitating the decrease in viscosity with time caused by enzyme action in a solution of 1% CMC-50 mM citrate phosphate (pH 6). The molecular weight of the endoglucanase was determined from periplasmic extracts by gel filtration (3) by using Sephadex G100 (Pharmacia, Sydney, Australia). The column was calibrated with bovine serum albumin, ovalbumin, pepsin, and lysozyme, and the void volume was calculated with blue dextran (Sigma). The column was eluted with 50 mM citrate buffer (pH 6), and fractions were assayed for endoglucanase activity by measuring the release of reducing sugar by the para-hydroxybenzoic acid hydrazide method (26), with CMC used as the substrate.
RESULTS Isolation of endoglucanase genes from F. succinogenes AR1 genomic DNA. Two thousand cosmid clones were screened in E. coli DH1 for endoglucanase activity. Eight clones exhibited LCZs on CMC plates, and 25 clones exhibited SCZs. Plasmid pRC093 (Table 1), which was representative
flS
VOL. 57,
1991
Plasmid
F. SUCCINOGENES ENDOGLUCANASE GENES
361
TABLE 1. Recombinant plasmids and cosmids used in this study
or
Endoglucanase activity
Vector and construction
cosmid
pRC093 Cosmids 2 to 8 pDA84 pRCE PRCEH pRCEH2 pRCZ+ pRCZpRCZHA+ pRCZES+ pRCZEM+ pRCZEX+ pRCZEB+
pHC79 cosmid gene bank pHC79 cosmid gene bank pHC79 cosmid gene bank pGEM-1, approx. 3.3-kb Hindlll insert from pRC093 pGEM-1, 2.1-kb EcoRI-HindIII insert from pRC093 pGEM-2, 2.1-kb EcoRI-Hindlll insert from pRC093 pGEM7Zf+, 2.1-kb EcoRI-HindIII insert from pRC093 pGEM7Zf-, 2.1-kb EcoRI-HindIII insert from pRC093 pGEM7Zf+, 1.55-kb HindIII-ApaI insert from pRCEHa pGEM7Zf+, 1.35-kb EcoRI-SphI insert from pRCEH pGEM7Zf+, 1.38-kb EcoRI-Sau3AI insert from pRCEH pGEM7Zf+, 1.62-kb EcoRI-XbaI insert from pRCEH pGEM7Zf+, 1.86-kb EcoRI-BgIII insert from pRCEH
LCZ LCZ SCZ LCZ LCZ LCZ LCZ LCZ h
Exonuclease deletions pRCZ-1 pRCZ-2 pRCZ+1 pRCZ+2
pRCZ-, deletion from HindIll terminus, 1.97-kb insert pRCZ-, deletion from HindIll terminus, 1.74-kb insert pRCZ+, deletion from EcoRI terminus, 1.93-kb insert pRCZ+, deletion from EcoRI terminus, 1.58-kb insert
LCZ
LCZ
LCZ
a Restriction enzyme sites and exonuclease deletion sites have been mapped by DNA sequence analysis (7b). b-, No clearing zone and no detectable activity by reducing sugar assay.
Cf
LCZ, was isolated and completely digested with HindlIl,
-?coRI, or BamHI. Digests were ligated into appropriately roestricted pGem-1, transformed into
E. coli HB101, and gcreened for endoglucanase activity on CMC plates. A lumber of transformants derived from each digest gave 'MC-positive clones, indicating that all three restriction e'nzyme sites were absent from the functional endoglucanase ;ene. The EcoRI-derived clone pRCE harbored an insert of a pproximately 3.3 kb. HindlIl digestion of this insert reulted in a 2.1-kb EcoRI-HindIII fragment which was cloned iin pGem-1, generating pRCEH. Restriction map analysis of tihe pRCEH insert (Fig. 1) indicated an absence of sites for latII, NsiI, PstI, Sacl, Sall, SmaI, and XhoI. Fragments firom digestion with Sau3AI, which cut three times, and tpaI, BglII, SphI, and XbaI, which cut only once each, were Ii igated into pGem7Zf', transformed into E. coli HB101, and ,creened onto CMC plates. Only clone pRCEB' (EcoRI3glII insert) exhibited cellulase activity. However, activity vas not stably maintained in cells harboring this recombiiant, and the 2.1-kb EcoRI-HindIII fragment was considered ti o etemniu-ze
retito.nym
loubende . edogluc
v
act
iviy.
rgetta
Iwasmnoted
tihat cloning of DNA derived from F. succinogenes ARi in hGem-, pGem7Zf+n pGem7Zfr pHC79, 6 Gem1 pHC79'and pKK232-8 pKK2328 (6) pGem7Zf' pGem7Zf. Vvas generally successful, whereas cloning of recombinants C.onstructed in pTZ18U and pTZ19U (Biorad, Sydney, Aus-
tralia) and PBS- (Stratagene, Sydney, Australia) were frequently unstable. pGem7Zf+ and pGem7Zf- were most frequently used for subcloning and exonuclease deletion analysis because of the large number of suitable restriction enzyme sites in the multiple cloning site. Endoglucanase activity from the 2.1-kb fragment in pRCEH and pREH2 was independent of orientation, indicating that the gene was expressed from its own promoter in E. coli. Deletion analysis (Table 1) indicated that the functional gene was encoded within a fragment of approximately 1.7-kb delimited at the EcoRI end by pRCZ+1 and at the HindlIl end by pRCZ-1. Southern blot analysis. Seven of the recombinant cosmids producing LCZ, pRC093, and cosmids 2 to 7 (see Table 1 for description) showed equivalent bands of approximately 3.4 kb when they were hybridized with 32P-labeled probe synthesized from the pRCEH insert (Fig. 2). By contrast, cosmid 8 (LCZ) and pDA84 (SCZ) revealed an absence of hybridization under both high- and low-stringency condi-
hybridization and washing. Moreover, when the filter was stripped and then reprobed with a 6-kb HindIII fragment (cellulase encoding) derived from pDA84, only the corresponding 6-kb band in lanes B and N (Fig. 2a) indicated (data notgenes shown). These observations indihybridization and are encoded that2 three cated (i) pRC093 cosmids to 7, distinct which produce LCZs, (ii) by cosmid 8 (LCZ), tions of
and (iii) pDA84 (SCZ). Southern blot analysis of the EcoRI
HindlIl fragment from pRCEH hybridized to F. succinogenes AR1 genomic DNA digested with Hindlll showed a single band of hybridization under high- and low-stringency conditions. This indicated that F. succinogenes AR1 probably encodes only a single copy of the gene and that there is an absence of other genes showing a high degree of similarity 500 bp to it. It was noted that the genomic band (Fig. 2a, lane L) . to be smaller than the corresponding Restriction mapofthe21kbenoglucanasenappeared FIG. FIG. 1. Restriction map of the 2.1-kb endoglucanase-encoding cosmid bands (Fig.slightly 2a, lanes D through J). This may have
EV
A
V
Shl
mx
M B
H
iinsert from pRCEH. Restriction enzyme abbreviations: E, EcoRI,
X, ApaI; V, AvaI; S, SphI; M, Sau3Al; X, XbaI; B, BglII; H, HindIII. The insert was restricted with HindIll and ApaI to generate I)RCZHA' and with EcoRI combined with SphI, Sau3AI, XbaI, and gglIl to generate pRCZES', pRCZEM+, pRCZEX+, and pRCZEB respectively.
.
'
overloading. Determination of the cellular distribution of endoglucanase activity, been due to a gel
artifact,
perhaps caused by
Cellular localization of endoglucanase in E. coli.
as estimated by reducing sugar analysis by using the dinitrosalicylic acid method (32) or the procedure of Nelson (35)
362
APPL. ENVIRON. MICROBIOL.
CAVICCHIOLI AND WATSON
a ABC DE F GH I J K L MN
b ABC DE F G HI J K L MN
8~~~~~~U
'
FIG. 2. Southern blot analysis. (a) Hybridization of a 32P-labeled probe (2.1-kb EcoRI-Hindlll fragment from pRCEH) to Hindlll-digested cosmid clones (lanes B through J), F. succinogenes ARi genomic DNA (lane L), and 2.1- and 6-kb cellulase-encoding fragments from pRCO93 (lane M) and pDA84 (lane N), respectively. X Hindlll-EcoRl (lane A) and X Hindlll (lane K) were used as molecular weight markers. Sizes of bands marked by arrows for X Hindlll-EcoRI are as follows, from the top: 21.2, 4.3, 2.0, 1.4, and 0.95 kb. (b) Ethidium bromide-stained agarose gel of the gel used in panel a.
(see Materials and Methods), led to some inconsistencies in activity determinations, particularly in the extracellular fraction, in which high and variable blank values were obtained. The Nelson method (35), however, provided the most consistent results and was used throughout this study. Because of its relative purity, the periplasmic fraction was used for all subsequent enzyme activity determinations, including substrate utilization studies (see below). The cellular distribution for each fraction was confirmed by comparing the rates of fluidity change in an Ubbelohde viscometer. This method allowed estimation of the enzyme distribution by assessing endoglucanase activity by a purely physical method rather than by a chemical assay. Samples were appropriately diluted until a constant rate of change in fluidity was obtained for each fraction. E. coli cells harboring plasmids with either a >20-kb insert in pRC093 or a 2.1-kb insert in pRCEH produced clearing zones of equivalent sizes on CMC plates, and this was reflected by the endoglucanase activity (Table 2). The cellular distribution of endoglucanase for E. coli harboring either recombinant was mainly periplasmic (approximately 80%). A low but significant proportion (16%) was, however, found in the extracellular medium. Marker enzymes indicated the integrity of each fraction. Extracellular activity was also monitored, and endoglucaTABLE 2. Cellular distribution of endoglucanase in E. coli HB101 harboring pRCEH and E. coli DH1 harboring pRC093 Activity (mU/ml) of original culture Fraction
Endoglucanase"
pNP-
pNP-alkaline
galactosidase
phosphatase
NDc ND 0.4 (100) 0.4 (100)
ND 0.6 (100) ND 0.6 (100)
nase levels were found to increase proportionally with respect to the cell growth phase (data not shown). The activities of the cell-bound fractions (pelleted cellular membranes following sonication) were less than 1% (data not shown). Physiochemical properties of endoglucanase activity. It was noteworthy that expression of endoglucanase activity of E. coli clones was observed under aerobic or anaerobic growth conditions. Cellobiose or glucose (100 mM) had no effect on the expression of endoglucanase activity, as determined by clearing zone size on plates containing CMC and glucose or cellobiose. Endoglucanase pH profiles were constructed, and activity was high (>50% of maximum activity) between pH 4.5 and 8.0, peaking at about pH 5.0 for E. coli HB101 harboring pRCEH (Fig. 3). Endoglucanase activity showed a sharp decrease below pH 4.5 and a more gradual decrease at alkaline pH values. Endoglucanase activity was maximum at
100
.
w
o
Extracellular Periplasmic Cytoplasmic Total
6
(16)b
30 (78) 2 (6) 38 (100)
pRCEH
7 32 2 41
(17) (78) (5) (100)
a Endoglucanase activity was determined by the method of Nelson (35), and cellular distribution was confirmed by a viscometric assay. b Values in parentheses are percentages of total activity. ' ND, Not detectable (0.1 mU/ml or less).
0
50
z w w 0-
pRC093
75
25r
3
4
5
6
7
8
9
10 11 12
pH
FIG. 3. Effect of pH on the endoglucanase activity of E. coli HB101 harboring pRCEH.
VOL. 57, 1991
F. SUCCINOGENES ENDOGLUCANASE GENES
TABLE 3. Hydrolysis of glycan substratesa after 24 h of incubation with periplasmic enzymes from E. coli HB101 harboring pRCEH
1001
Substrate
Reducing sugar
% Activity relative to CMC
26.8b 10.9 8.5 2.4 46.2 165.4 NDc 13.4 15.8
100
Avicel PH101 Filter paper Cotton Ball-milled cellulose Acid-swollen cellulose Laminarin Lichenan Xylan
751 CMC CD
363
501
z w
0
251
I
1
2
3
4
5
6
7
8
9
TIME (HR) FIG. 4. Effect of preincubation at 4°C (A), 39°C (0), and 50°C (A) on the endoglucanase activity of E. coli HB101 harboring pRCEH.
37 to 39°C, decreasing markedly at elevated temperatures. This was reflected in thermal stability at 39°C compared with that at 50°C (Fig. 4). At 39°C, the activities of the enzyme fractions, which were buffered in 50 mM citrate phosphate (pH 6), remained high (75 to 100%) for 8 h, whereas endoglucanase was completely lost after 3 h when it was stored at 50°C. By comparison, endoglucanase activity was retained for several months at 0 to 4°C. The molecular weight of the endoglucanase was estimated by gel filtration to be approximately 46,500. Mode of action and substrate specificity. As measured with an Ubbelohde viscometer, a rapid increase in CMC fluidity was observed with periplasmic enzymes from E. coli harboring pRCEH, and this was indicative of an endo-acting enzyme. Activity on natural substrate (cellobiose) or substituted substrate (pNP-glucoside) was not detectable. These substrates are diagnostic for P-glucosidase activity. Low activity on pNP-cellobioside, 0.72 nmol/min/mg of protein, indicated that exoglucanase activity was also minimal. Similarly, activity was not detected on other natural or substituted short-chain glycans (maltose, lactose, sucrose, melibiose, melezitose, pNP-xyloside, pNP-lactoside, and pNPglucoside). It was noted that trehalose was actively hydrolyzed by periplasmic enzymes from E. coli HB101 and DH1 (7a). Activity on recalcitrant substrates such as filter paper, cotton, and Avicel was significant (Table 3). Less crystalline celluloses were more extensively hydrolyzed, with the highest activity detected on acid-swollen cellulose. It was noted that hydrolysis of CMC resulted in the rapid production of reducing sugar, with >70% maximum production observed after 1 h and essentially 100% observed by 4 h. Specific activity on CMC was about 70 nmol/min/mg of protein (30-min assay). Activity on lichenan (1,3-1,4-4-D-glucan), but not on laminarin (1,3-3-D-glucan), indicated that it was probably the 1,4-p-D-glucan component of lichenan that was being hydrolyzed. It was noteworthy that the endoglucanase was also active on xylan.
(p.g of glucose)
41 32 9 173 618 ND 50 59
' Reducing sugars were determined by the method of Nelson (35) after incubation of periplasmic enzymes in 50 mM sodium citrate buffer (pH 6) for 24 h. Each substrate was 5 mg/ml, except for Avicel, filter paper, and cotton, which were 50 mg/ml. b Production of reducing sugar on CMC at 1 h was 72% of the level at 24 h and was essentially 100% by 4 h. ND, Not detectable (1 jLg or less in 24 h).
DISCUSSION
Genomic DNA from F. succinogenes AR1 was successfully cloned into E. coli by using cosmid pHC79, and a number of recombinants expressed endoglucanase activity. Distinct LCZs and SCZs on CMC plates were features of endoglucanase expression in E. coli. A similar phenomenon has been reported by other authors during the construction of gene libraries of Cellulomonas (14), Clostridium thermocellum (40), Ruminococcus albus (21), and F. succinogenes (8, 9, 24). In all cases a correlation between clearing zone diameter and endoglucanase activity was observed. Southern blot analysis revealed the isolation of three separate genes encoding endoglucanase (Fig. 2). It is possible that other unique endoglucanase genes are represented in the family of 25 clones that produced SCZs. The fact that two separate genes were found for the eight clones that produced LCZs indicates that this possibility is likely. Multiple endoglucanase genes have been identified for other cellulolytic bacteria, and in C. thermocellum, for example, 15 distinct endoglucanase genes have been isolated (18). In F. succinogenes S85, a number of endoglucanases have been identified, and three have been characterized in particular (30, 31). It will thus be useful to isolate and characterize all the genes encoding endoglucanases from our gene bank, because this will lead to a comprehensive understanding of the endoglucanase requirements for this microorganism. In view of the recent reclassification of Fibrobacter as a genus distinct from the genus Bacteroides and, furthermore, to the suggestion that it may form a new family distinct from the family Bacteroidaceae (33), it was significant that the endoglucanase genes in pRCEH and pRCEH2 were expressed in E. coli, indicating that promoters from F. succinogenes may be recognized by E. coli RNA polymerase. Data from DNA sequencing and primer extension analysis to map transcription start sites would be useful to confirm the presence of functional promoters within the cloned insert. Delimitation of the boundaries of the functional gene in pRCEH was undertaken by subcloning restriction fragments in pGem7Zf+ and constructing nested deletions from pRCZ+ and pRCZ-. As a result, a 1.7-kb fragment was identified as being important for the expression of endoglucanase. This indicated that the possible molecular weight of the encoded protein could be up to 62,000. However, preliminary data
364
indicated that the molecular weight of the periplasmic enwas 46,500, which suggested that regions in addition to the coding sequence are required for activity. In F. succinogenus ARI, 80% of endoglucanase activity was found in the extracellular medium (7a), and a similar level was found in strain S85 (16). In E. coli harboring pRCEH and pRC093, most activity (approximately 80%) was localized in the periplasm (Table 2). E. coli is not noted for its ability to secrete cloned products into the extracellular medium; however, secretion to the periplasm is normally efficient (38). Most cloned cellulases are secreted to the periplasm, and DNA sequence analysis has revealed that all cellulases to date have signal peptides predicted for their amino acid sequences (4). Maximum endoglucanase activity was at pH 5.0, which is around the range (pH 5.5 to 7.0) reported for cloned endoglucanases from F. succinogenes (44), R. albus (25, 37, 47), and Bacteroides ruminicola (48). Maximum activity of extracellular endoglucanase from F. succinogenes ARI is observed at pH 5.5 (7a), which is similar to that for the single gene product cloned in E. coli. F. succinogenes exhibits a diverse range of enzyme activities with the ability to solubilize lignocellulosic materials (23). Some of these properties were reflected in the broad range of substrates hydrolyzed by the gene product of
doglucanase
pRCEH (Table 3). Most endoglucanases that have been cloned from ruminal bacteria show broad-spectrum activity on cellulosic compounds. Kinetic studies were not under-
taken; however, it was noteworthy that activities on acidswollen cellulose and ball-milled cellulose were higher than that on CMC. Initial production of reducing sugars on CMC was, in fact, high, but within 4 h, hydrolytic action essentially ceased. This can be explained by an endoglucanase attack on the portions of CMC that have at least three or more consecutive unsubstituted glucose units, resulting in only about 2% of CMC that can be readily hydrolyzed (11, 27). ACKNOWLEDGMENTS We are grateful to Peter East, John Pemberton, and coworkers Keith Gregg, Anil Lachke, Philip Vercoe, and Cheryl Ware for valuable advice and discussion; Tom Bauchop for providing cultures of F. succinogenes AR1; Michelle Deegenaars for technical expertise; and Jean Hansford for typing the manuscript. We also acknowledge in particular Ronald Teather, Geoff Hazlewood, Bryan White, and David Wilson for preprints and recent publications and David Stahl and Chuzhao Lin for communication of unpublished data.
REFERENCES 1. Amann, R. I., L. Krumholz, and D. A. Stahl. 1990. Fluorescent
oligonucleotide probing of whole cells for determinative, phylogenetic, and environmental studies in microbiology. J. Bacteriol. 172:762-770. 2. Anderson, D. G., and L. L. McKay. 1983. Simple and rapid method for isolating large plasmid DNA from lactic streptococci. Appl. Environ. Microbiol. 46:549-552. 3. Andrews, P. 1964. Estimation of molecular weights of proteins
by Sephadex gel filtration. Biochem. J. 91:222-233. Beguin, P., J. Millet, S. Chauvaux, E. Yague, P. Tomme, and J.-P. Aubert. 1989. Genetics of bacterial cellulases, p. 57-72. In M. P. Coughlan (ed.), Enzyme systems for lignocellulose degradation. Elsevier Applied Science Publications, London. 5. Birnboim, H. C., and J. Doly. 1979. A rapid alkaline extraction procedure for screening recombinant plasmid DNA. Nucleic Acids Res. 12:4411-4427. 6. Brosius, J. 1984. Plasmid vectors for the selection of promoters. Gene 27:151-160. 7. Bryant, M. P., and R. N. Doetsch. 1954. A study of actively 4.
APPL. ENVIRON. MICROBIOL.
CAVICCHIOLI AND WATSON
cellulolytic rod shaped bacteria of the ovine rumen. J. Dairy Sci. 37:1176-1183. 7a.Cavicchioli, R. Unpublished data. 7b.Cavicchioli, R., P. D. East, and K. Watson. Submitted for
publication.
8. Cavicchioli, R., D. H.-L. Lai, and K. Watson. 1989. Cloning and expression of cellulase genes from Bacteroides succinogenes, p. 153-156. In M. Sleigh (ed.), Eighth Australian biotechnology conference. Australian Biotechnology Association, Sydney, Australia. 9. Crosby, B., B. Collier, D. Y. Thomas, R. M. Teather, and J. D. Erfle. 1984. Cloning and expression in Escherichia coli of cellulase genes from Bacteroides succinogenes, p. 573-576. In S. Hasain (ed.), Fifth Canadian Bioenergy R & D Seminar. Elsevier Applied Science Publications, Amsterdam. 10. Erfle, J. D., R. M. Teather, P. J. Wood, and J. E. Irvin. 1988. Purification and properties of a 1,3-1,4-3-D-glucanase (lichenase, 1,3-1,4-13-D-glucan 4-glucanohydrolase, EC 3.2.1.73) from Bacteroides succinogenes cloned in Escherichia coli. Biochem. J. 255:833-841. 11. Eriksson, K.-E., and B. H. Hollmark. 1969. Kinetic studies of the action of cellulase upon sodium carboxymethyl cellulose. Arch. Biochem. Biophys. 133:233-237. 12. Flint, H. J., C. A. McPherson, G. Avgustin, and C. S. Stewart. 1990. Use of a cellulase encoding gene probe to reveal restriction fragment length polymorphisms among ruminal strains of Bacteroides succinogenes. Curr. Microbiol. 20:63-67. 13. Gilardi, G. L. 1971. Characterization of nonfermentive nonfastidious gram negative bacteria encountered in medical bacteriology. J. Appl. Bacteriol. 34:623-644. 14. Gilkes, N. R., M. L. Langsford, D. G. Kilburn, R. C. Miller Jr., and R. A. J. Warren. 1984. Mode of action and substrate specificities of cellulases from cloned bacterial genes. J. Biol. Chem. 259:10455-10459. 15. Gomori, G. 1955. Preparation of buffers for use in enzyme studies. Methods Enzymol. 1:138-146. 16. Groleau, D., and C. W. Forsberg. 1981. Cellulolytic activity of the rumen bacterium Bacteroides succinogenes. Can. J. Microbiol. 27:517-530. 17. Hanahan, D. 1983. Studies on transformation of Escherichia coli with plasmids. J. Mol. Biol. 166:557-580. 18. Hazlewood, G. P., M. P. M. Romaniec, K. Davidson, 0. Grepinet, P. Beguin, J. Millet, 0. Raynaud, and J.-P. Aubert. 1988. A catalogue of Clostridium thermocellum endoglucanase, j-glucosidase and xylanase genes located in Escherichia coli. FEMS Microbiol. Lett. 51:231-236. 19. Heppel, L. A. 1971. The concept of periplasmic enzymes, p. 223-347. In L. I. Rothfield (ed.), Structure and function of biological membranes. Academic Press, Inc., New York. 20. Hohn, B., and J. Collins. 1980. A small cosmid for efficient cloning of large DNA fragments. Gene 11:291-298. 21. Howard, G. T., and B. A. White. 1988. Molecular cloning and expression of cellulase genes from Ruminococcus albus 8 in Escherichia coli bacteriophage X. Appl. Environ. Microbiol. 54:1752-1755. 22. Huang, L., and C. W. Forsberg. 1987. Isolation of a cellodextrinase from Bacteroides succinogenes. Appl. Environ. Microbiol. 53:1034-1041. 23. Huang, L., M. McGavin, C. W. Forsberg, J. S. Lam, and K. J. Cheng. 1990. Antigenic nature of the chloride stimulated cellobiosidase and other cellulases of Fibrobacter succinogenes subsp. succinogenes S85 and related fresh isolates. Appl. Environ. Microbiol. 56:1229-1234. 24. Irvin, J. E., and R. M. Teather. 1988. Cloning and expression of a Bacteroides succinogenes mixed linkage 3-glucanase (1,3-1,4,-D-glucan 4-glucanohydrolase) gene in Escherichia coli. Appl. Environ. Microbiol. 54:2672-2676. 25. Kawai, S., H. Honda, T. Tanase, M. Taya, S. IiJima, and T. Kobayashi. 1987. Molecular cloning of Ruminococcus albus cellulase gene. Agric. Biol. Chem. 51:59-63. 26. Koziol, M. J. 1981. An evaluation of the alkaline para-hydroxybenzoic acid hydrazide procedure for the determination of reducing sugars. Anal. Chim. Acta 128:195-205.
VOL. 57, 1991
27. Kyriacou, A., R. C. MacKenzie, and R. J. Neufeld. 1987. Detection and characterization of the specific and nonspecific endoglucanases of Trichoderma reesei: evidence demonstrating endoglucanase activity by cellobiohydrolase II. Enzyme Microb. Technol. 9:25-32. 28. Layne, E. 1957. Spectrophotometric and turbidometric methods for measuring proteins. Methods Enzymol. 3:447-454. 29. McDermid, K. P., R. MacKenzie, and C. W. Forsberg. 1990. Esterase activities of Fibrobacter succinogenes subsp. succinogenes S85. Appl. Environ. Microbiol. 56:127-132. 30. McGavin, M., and C. W. Forsberg. 1988. Isolation and characterization of endoglucanases 1 and 2 from Bacteroides succinogenes S85. J. Bacteriol. 170:2914-2922. 31. McGavin, M., C. W. Forsberg, B. Crosby, A. W. Bell, D. Dignard., and D. Y. Thomas. 1989. Structure of the cel-3 gene from Fibrobacter succinogenes S85 and characteristics of the encoded gene product, endoglucanase 3. J. Bacteriol. 171:55875595. 32. Miller, G. L. 1959. Use of dinitrosalicylic acid reagent for determination of reducing sugar. Anal. Biochem. 31:426-428. 33. Montgomery, L., B. Flesher, and D. Stahl. 1988. Transfer of Bacteroides succinogenes (Hungate) to Fibrobacter gen. nov. as Fibrobacter succinogenes comb. nov. and description of Fibrobacter intestinalis sp. nov. Int. J. Syst. Bacteriol. 38:430435. 34. Nath, R. L., and N. Rydon. 1954. The influence of structure on the hydrolysis of substituted phenyl-p-D-glucosides by emulsin. Biochem. J. 57:1-10. 35. Nelson, N. 1944. A photometric adaptation of the Somogyi method for the determination of glucose. J. Biol. Chem. 153: 375-380. 36. Neu, H. C., and L. A. Heppel. 1965. The release of enzymes from Escherichia coli by osmotic shock and during formation of spheroplasts. J. Biol. Chem. 240:3685-3692. 37. Ohmiya, K., K. Nagashima, T. Kajino, E. Goto, A. Tsukada, and S. Shimizu. 1988. Cloning of the cellulase gene from Ruminococcus albus and its expression in Escherichia coli. Appl. Environ. Microbiol. 54:1511-1515. 38. Oliver, D. B. 1987. Periplasm and protein secretion, p. 56-69. In F. C. Neidhardt, J. L. Ingraham, K. B. Low, B. Magasanik, M. Schaechter, and H. E. Umbarger (ed.), Escherichia coli and
F. SUCCINOGENES ENDOGLUCANASE GENES
365
Salmonella typhimurium: cellular and molecular biology. American Society for Microbiology, Washington, D.C. 39. Putney, S. D., S. J. Benkovic, and P. R. Schimmel. 1981. A DNA fragment with an alpha-phosphorothioate nucleotide at one end is asymmetrically blocked from digestion by exonuclease III and can be replicated in vivo. Proc. Natl. Acad. Sci. USA 78:7350-7454. 40. Romaniec, M. P. M., N. G. Clarke, and G. P. Hazlewood. 1987. Molecular cloning of Clostridium thermocellum DNA and the expression of further novel endo-P-1,4-glucanase genes in Escherichia coli. J. Gen. Microbiol. 133:1297-1307. 41. Sipat, A., K. A. Taylor, R. Y. C. Lo, C. W. Forsberg, and P. J. Krell. 1987. Molecular cloning of a xylanase gene from Bacteroides succinogenes and its expression in Escherichia coli. Appl. Environ. Microbiol. 53:477-481. 42. Southern, E. M. 1975. Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Mol. Biol. 98:503-517. 42a.Stahl, D. A., and C. Lin. Personal communication. 43. Stewart, C. S., C. Paniagua, D. Dinsdale, K.-J. Cheng, and S. H. Garrow. 1981. Selective isolation and characteristics of Bacteroides succinogenes from the rumen of a cow. Appl. Environ. Microbiol. 41:504-510. 44. Taylor, K. A., B. Crosby, M. McGavin, C. W. Forsberg, and D. Y. Thomas. 1987. Characteristics of the endoglucanase encoded by a cel gene from Bacteroides succinogenes expressed in Escherichia coli. Appl. Environ. Microbiol. 53:4146. 45. Teather, R. M., and P. J. Wood. 1982. Use of Congo redpolysaccharide interactions in enumeration and characterization of cellulolytic bacteria from the bovine rumen. Appl. Environ. Microbiol. 43:777-780. 46. Van Gylswyk, N. O., and J. J. T. K. Van der Toorn. 1986. Enumeration of Bacteroides succinogenes in the rumen of sheep fed maize-straw diets. FEMS Microbiol. Lett. 38:205-209. 47. Ware, C. E., T. Bauchop, and K. Gregg. 1989. The isolation and comparison of cellulase genes from two strains of Ruminococcus albus. J. Gen. Microbiol. 135:921-930. 48. Wood, T. M. 1970. The cellulase of Fusarium solani: purification and specificity of the P-(1,4)-glucanase and the P-D-glucosidase components. Biochem. J. 121:353-362.
