JOURNAL OF BACTERIOLOGY, Oct. 2003, p. 6067–6075 Vol. 185, No. 200021-9193/03/$08.00�0 DOI: 10.1128/JB.185.20.6067–6075.2003Copyright © 2003, American Society for Microbiology. All Rights Reserved.

Regulation of Expression of Cellulosomal Cellulase and HemicellulaseGenes in Clostridium cellulovorans

Sung Ok Han,1 Hideaki Yukawa,2 Masayuki Inui,2 and Roy H. Doi1*Section of Molecular and Cellular Biology, University of California, Davis, California 95616,1 and Research Institute

of Innovative Technology for the Earth, Kyoto 619-0292, Japan2

Received 12 March 2003/Accepted 3 July 2003

The regulation of expression of the genes encoding the cellulases and hemicellulases of Clostridium cellulo-vorans was studied at the mRNA level with cells grown under various culture conditions. A basic pattern of geneexpression and of relative expression levels was obtained from cells grown in media containing poly-, di- ormonomeric sugars. The cellulase (cbpA and engE) and hemicellulase (xynA) genes were coordinately expressedin medium containing cellobiose or cellulose. Growth in the presence of cellulose, xylan, and pectin gave riseto abundant expression of most genes (cbpA-exgS, engH, hbpA, manA, engM, engE, xynA, and/or pelA) studied.Moderate expression of cbpA, engH, manA, engE, and xynA was observed when cellobiose or fructose was usedas the carbon source. Low levels of mRNA from cbpA, manA, engE, and xynA were observed with cells grown inlactose, mannose, and locust bean gum, and very little or no expression of cbpA, engH, manA, engE, and xynAwas detected in glucose-, galactose-, maltose-, and sucrose-grown cells. The cbpA-exgS and engE genes weremost frequently expressed under all conditions studied, whereas expression of xynA and pelA was morespecifically induced at higher levels in xylan- or pectin-containing medium, respectively. Expression of thegenes (cbpA, hbpA, manA, engM, and engE) was not observed in the presence of most soluble di- or monosac-charides such as glucose. These results support the hypotheses that there is coordinate expression of somecellulases and hemicellulases, that a catabolite repression type of mechanism regulates cellulase expression inrapidly growing cells, and that the presence of hemicelluloses has an effect on cellulose utilization by the cell.

The major components of plant cell walls are cellulose,hemicellulose, and lignin, with cellulose being the most abun-dant component, followed by hemicelluloses. Cellulose con-sists of long polymers of �-1,4-linked glucose units and formsa crystalline structure, whereas the structure of hemicellulosesis more variable. Hemicelluloses include xylan consisting of�-1,4-linked xylose units, glucomannans consisting of �-1,4-linked glucose and mannose units, and arabinans and galactansin which the main chain sugars include arabinose and galac-tose, respectively. The cellulolytic bacteria produce a set ofenzymes (called cellulosomes) which synergistically hydrolyzecrystalline cellulose and hemicelluloses to smaller oligosaccha-rides and finally to monosaccharides (6, 7, 13, 15, 16, 18, 33,36).

Clostridium cellulovorans, an anaerobic, mesophilic, andspore-forming bacterium, is one of the most efficient cellulo-lytic organisms (30). The cellulases and hemicellulases [we willabbreviate these two terms together as (hemi-)cellulases] pro-duced by C. cellulovorans have been studied extensively. Sev-eral cellulases (family 5 and 9 endoglucanases and a family 48exoglucanase), a mannanase, a xylanase, and a pectate lyasehave been characterized (6, 16, 18, 33). The genes encoding acluster of cellulosomal subunits, i.e., the gene cbpA encoding ascaffolding protein, the gene exgS encoding exoglucanase (18),the genes engH, engK, and engM encoding endoglucanases, thegene hbpA encoding a hydrophilic domain and a cohesin (31),and the gene manA encoding a mannanase (28), have been

cloned and sequenced. The gene engE encoding an endoglu-canase (34), the gene xynA encoding a xylanase (16), and thegene pelA encoding a pectate lyase are not linked to the genecluster (6, 29, 32), although they are cellulosomal enzymes.

Since plant polysaccharides are the most abundant renew-able biomass, cellulolytic microorganisms play a very majorrole in carbon turnover in nature. It is important to understandhow bacteria regulate expression of the various hydrolytic en-zymes in order to produce optimal enzyme mixtures for thedegradation of different plant materials. Expression of thecellulase genes of C. cellulovorans has been studied at theprotein level (8, 17, 22). Only a few studies concerning regu-lation of the (hemi-)cellulases of C. cellulovorans have beencarried out (1, 9). Therefore, many fundamental questions stillremain to be answered at the transcriptional level, such aswhether the expression of the different (hemi-)cellulases iscoordinately regulated by a shared mechanism and whether alow level of constitutive expression of (hemi-)cellulases occursunder all conditions. Preliminary evidence indicated that con-stitutive synthesis of cellulosome components occurred whencells were grown in the presence of glucose (22). Mechanismsof true induction or repression have not been studied in depth.For these reasons, we have addressed some of the questionsrelated to (hemi-)cellulase gene expression in C. cellulovoransin this paper.

MATERIALS AND METHODS

Bacterial strain and growth conditions. C. cellulovorans ATCC 35296 was usedas the source of genomic DNA and total RNA. The organism was grown understrictly anaerobic conditions at 37°C in round-bottom flasks containing a previ-ously described medium (28, 30), which included either di- or monomeric sugars(fructose, glucose, mannose and galactose, lactose, maltose, sucrose, and cello-

* Corresponding author. Mailing address: Section of Molecular andCellular Biology, University of California, One Shields Avenue, Davis,CA 95616. Phone: (530) 752-3191. Fax: (530) 752-3085. E-mail:rhdoi@ucdavis.edu.

6067

on March 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

biose; 0.5%, wt/vol) or polymeric sugars (microcrystalline cellulose [Avicel],locust bean gum, xylan, and pectin; 1%, wt/vol). Avicel was purchased from FMCCorporation. Locust bean gum, xylan (birch wood), and pectin (apples) werepurchased from Sigma.

Bacterial protein determination. The determination of cell mass in culturesgrown with cellobiose, cellulose, locust bean gum, pectin, and xylan was based onbacterial-protein estimation as described by Bensadoun and Weinstein (3; seealso reference 5). A 500-�l aliquot was centrifuged for 10 min at 13,000 � g. Thepellets were washed with 500 �l of sodium phosphate buffer (50 mM, pH 7.0) andincubated with 400 �l of sodium deoxycholate (2%) for 20 min. One hundredmicroliters of trichloroacetic acid (24%) was added to the suspension, which wascentrifuged at 13,000 � g for 10 min. The protein concentration was measured byusing the BCA Compat-Able protein assay kit (Pierce) with bovine serum albu-min as the standard.

Nucleic acid isolation. Chromosomal DNA of C. cellulovorans was isolated byusing a genomic DNA purification kit (Promega) according to the manufactur-er’s instructions. Total RNA was extracted from C. cellulovorans broth culturesby using an RNeasy kit (QIAGEN) with the additional step of treatment withRNAlater RNA stabilization reagent (Ambion), and RNase-free DNase (Pro-mega) according to the manufacturers’ instructions.

Northern blot analysis. RNA samples (up to 20 �g) were denatured in RNAsample buffer (250 �l of formamide, 83 �l of 37% [wt/vol] formaldehyde, 83 �lof 6� loading dye [Promega], 50 �l of 10� morpholine propanesulfonic acid[MOPS] buffer [20 mM MOPS, 5 mM sodium acetate, 1 mM EDTA {pH 7.0}],and 34 �l of distilled water) at 65°C for 10 min and separated through 1%agarose gels in MOPS buffer with 17% (vol/vol) formaldehyde. DNA probeswere synthesized by PCR by using specific oligonucleotides derived from the C.cellulovorans sequence as a template (Table 1). The probes were nonradioac-tively labeled by random priming by using digoxigenin (DIG) High Prime(Roche). To add the correct amount of probe to a hybridization, serial dilutions(0.05 to 10 pg) of each probe were spotted on a nylon membrane and labelingsensitivity (amount of labeled DNA per spot) was determined. RNA was trans-ferred overnight to a positively charged nylon membrane (Roche) by capillarytransfer by using 20� standard saline/citrate (0.3 M NaCl plus 0.03 M sodiumcitrate, pH 7). Hybridization was carried out for 16 to 20 h at 50°C in DIG EazyHyb buffer solution (Roche). Washing of the membrane and detection of specifictranscripts on the blots were carried out by using the DIG luminescent detectionkit (Roche) and its protocol.

RNA slot blot analysis. Total RNAs were diluted into appropriate concentra-tions with water, followed by the addition of two times the volume of the RNAsample buffer. After being incubated for 10 min at 65°C to denature the RNA,the samples were applied to a positively charged nylon membrane (Roche) byusing a Hybri-slot apparatus (Gibco-BRL) and the membrane was baked for 30min at 120°C under vacuum. Filters were hybridized with specific probes asdescribed for the Northern blot analyses.

RT-PCR analysis. Reverse transcriptase (RT) reactions were performed withtotal RNA by using a commercially available reverse transcription system (Pro-mega) with slight modifications to the recommended protocol. RT reactionswere performed in a final volume of 20 �l, which contained 5 mM MgCl2, 1� RTbuffer (10 mM Tris-HCl [pH 9. 0], 50 mM KCl, and 0.1% Triton X-100), 1 mM(each) deoxynucleoside triphosphates, 1 U of recombinant RNasin RNase in-hibitor, 15 U of avian myeloblastosis virus reverse transcriptase, 0.25 �M oligo-nucleotide primer, and 10 �g of substrate RNA. The reaction mixtures wereincubated at 42°C for 60 min, and reactions were terminated by heating themixtures at 95°C for 5 min, followed by incubation on ice for 5 min. The cDNA

products were then amplified in 25-�l PCR mixtures by using 2.5 �l of the RTreaction mixture as the template.

RESULTS

Relative expression levels of (hemi-)cellulase genes at dif-ferent growth phases. To determine whether the (hemi-)cellu-lase genes are regulated coordinately, changes in the expres-sion levels of several (hemi-)cellulase genes were monitoredduring the cultivation of C. cellulovorans on either cellulose orcellobiose as the sole carbon source. RNA was prepared fromthe culture at different stages of growth. The RNA was sub-jected to Northern blot analyses using probes that were specificto the cbpA-exgS, engE, xynA, pelA, arfA, and engF genes. Thesegenes represent major subunits (cbpA-exgS and engE) (8, 20,29, 34), one of the hemicellulase genes (xynA) (18), and thepectate lyase gene (pelA) (33). Two noncellulosomal genes,engF (27) and the �-L-arabinofuranosidase gene arfA (16),were also tested as control-endoglucanase genes.

A semiquantitative measure of the level of cbpA mRNA,using DIG-labeled probes and RNA isolated at different timesduring cell growth, was obtained by Northern blot analysis(Fig. 1A and B). The intensities of the bands were close ap-proximations of their relative abundance. The levels of cbpAmRNA increased simultaneously from early to middle expo-nential phase and dramatically decreased during the early sta-tionary phase when the cells were grown on cellobiose (Fig. 1Aand B, lanes 1 through 5). As with cbpA gene expression, thecells contained high levels of engE, xynA, and engF mRNAsduring most of the exponential growth phase, with the levelbeing the highest at the middle of the exponential phase (Fig.1A and B). Cellobiose clearly induced the expression of the(hemi-)cellulase genes after a rather short lag period. Reducedexpression was observed at a later stage of growth, but the arfAmRNA level increased when the cells reached the stationarygrowth phase (Fig. 1A and B, lane 5). The arfA transcript wasalso observed at the end of the stationary phase (Fig. 1A andB, lane 8). Cellobiose did not induce the expression of pelAduring the entire growth phase (Fig. 1B).

In addition, a similar analysis of Northern blots of culturesgrown on 1% cellulose showed the highest levels of expressionof cbpA, engE, xynA, and engF at the middle of the exponentialphase (Fig. 2A and B). Thereafter, the expression levels de-creased markedly as the cells entered the stationary phase.After further incubation (190 h), the transcripts were hardly

TABLE 1. PCR primers used for amplification of reverse transcripts and synthesis of gene-specific probes

Gene Enzyme encoded 5� primer 3� primer GenBank accession no.(reference[s])

cbpA Cellulose binding protein ATGCAAAAAAAGAAATCGCTG GGTTGATGTTGGGCTTGCTGTTTC M73817 (29)engH Endoglucanase H GGTGAAACAACAGCGACTCCAACA GCCCCAAGAATCCATCCAAGCTAA U34793 (20, 35)hbpA Hydrophobic protein A AGTATTGGCGTAGTAGTTGCAGGC GTGCGTTATCGGTGAAAGCTCCAA AF132735 (32, 35)manA Mannanase A AGATGCTGAATTGAAGGCGGCAGA CTCCACTCCACTTCATACTTGCAC AF132735 (32, 35)engM Endoglucanase M ATGATGGAGTAGAGGGAAGATGGG GCGTTCAGCATAAGGCATCGTT AF132735 (32, 35)engE Endoglucanase E TACTGATGACTGGGCTTGGATGAG GTTGCTTTCGCTGCTGC AF105331 (34)xynA Xylanase TGTTAGCCTCTTCTGC GATTCCAAGTGCCATAGC AF435978 (18)pelA Pectate lyase A TGATGCACCAAAAACAGCGC CAGTAGAAGAGCATCAAGCC AF105330 (33)engF Endoglucanase F TGGTCTACAATGGTTTCCTGGG GCATCATTCGTTACTCCACC U37056 (27)arfA �-L-arabinofurasnosidase ATGGAGGATTTTGGGTTGGG TCGGTGACTCTCCATC AY128945 (16)

6068 HAN ET AL. J. BACTERIOL.

on March 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

detectable. However, for arfA expression, the highest level oftranscripts was observed from late-exponential-phase cells(Fig. 2A and B, lane 6). In RNA from the early stationary andlate stationary growth phases, the arfA transcript was clearly

observed (Fig. 2A and B, lanes 7 and 8). With the pelA probe,very weak hybridization signals were observed only during thelate exponential phase (Fig. 2A and B, lane 6). Furthermore,when cultured on pectin and xylan, the growth pattern wassimilar to that of cells grown on cellulose (data not shown).The close correlation between the time course of transcriptionof cbpA, engE, and xynA supports the idea that cellulase andhemicellulase genes are produced simultaneously for plant cellwall degradation.

Induction of (hemi-)cellulase genes in response to di- ormonomeric sugars. To determine whether carbon sources ac-tivate (hemi-)cellulase gene transcription, C. cellulovorans cellswere grown to exponential phase in medium containing either0.5% monosaccharide (fructose, glucose, mannose, and galac-tose), 0.5% disaccharide (lactose, maltose, sucrose, and cello-biose), or 1% polysaccharide (cellulose, locust bean gum, pec-

FIG. 1. Relative levels of (hemi-)cellulase transcripts at differentgrowth phases on cellobiose culture. (A) C. cellulovorans growth curve.(B) Northern blot analyses were conducted with 5-�g concentrationsof RNA isolated from C. cellulovorans cultures grown on 0.5% cello-biose as the sole carbon source. The numbers of the lanes correspondto the numbers over the growth curve points shown in panel A.Ethidium bromide staining of rRNA is shown as a loading control.(C) The different DIG-labeled probes were prepared (each from 1 �gof template) by random primed labeling (see Materials and Methods).Dilutions (0.05 to 5 pg) of each probe were spotted on a nylon mem-brane, and labeling sensitivity (amount of labeled DNA per spot) wasdetermined in order to use similar amounts of each probe.

FIG. 2. Relative levels of (hemi-)cellulase transcripts at differentgrowth phases on cellulose culture. (A) Growth curve of C. cellulo-vorans. (B) Northern blot analyses were conducted with 5-�g concen-trations of RNA isolated from C. cellulovorans cultures grown on 1%cellulose as the sole carbon source. The numbers of the lanes corre-spond to the numbers over the growth curve points in panel A.Ethidium bromide staining of rRNA is shown as a loading control. Theprobes were labeled to a similar sensitivity, and the labeling sensitivitymethod corresponds to that described in Fig. 1C.

VOL. 185, 2003 REGULATION OF CELLULASE AND HEMICELLULASE GENES 6069

on March 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

tin, and xylan) as the sole carbon source. RT-PCR analysis wasperformed using various primer pairs specific for the cbpA,engH, manA, engE, and xynA transcripts. To ensure that theresulting PCR products were amplified from cDNA instead ofcontaminating chromosomal DNA, control experiments wereperformed in which RT was omitted. In these controls, no PCRfragments were detected (data not shown). RT-PCR withRNA from fructose-, lactose-, mannose-, cellobiose- and cel-lulose-grown cells revealed very similar patterns with all(hemi-)cellulase mRNAs tested (Fig. 3). Fructose and lactosemoderately induced expression of cbpA, engH, manA, engE,and xynA (Fig. 3, lanes 2 and 5), while mannose weakly inducedcellulase gene transcription (Fig. 3, lane 4). When celluloseand cellobiose were contained in the medium as positive con-trols, all genes (cbpA, engH, manA, engE, and xynA) wereclearly expressed (Fig. 3, lanes 8 and 9). No cDNA (cbpA,engH, manA, engE, and xynA) was detected with RNA isolatedfrom glucose-, galactose-, maltose-, and sucrose-grown cells.The pelA mRNA was not induced from any di- or monosac-charide (Fig. 3).

Induction of (hemi-)cellulase genes in response to polymericsugars. The media used to produce cellulosomal (hemi-)cellu-lases in C. cellulovorans are based on mixtures of plant mate-rials or cellulose (24). The levels of certain (hemi-)cellulaseshave been reported to be dependent on the growth substrate(6, 17, 22). These materials, particularly the insoluble sub-strates, tend to interfere with the estimation of cell mass andthe isolation of RNA and are often undefined in nature. Tostudy the effect of polymeric substrates on the relative expres-sion of (hemi-)cellulases in greater detail, RNAs from C. cel-lulovorans grown on four different insoluble polysaccharideswere selected for study on the basis of their gene products andtheir target substrates (i.e., cellulose and CbpA, locust beangum and ManA, pectin and PelA, and xylan and XynA). Theprobe fragments were amplified by PCR with specific primersto obtain fragments of similar length from each gene, and thesewere labeled to provide a similar sensitivity in order to allowcomparison of the relative expression levels of the (hemi-)cel-lulase genes studied (Fig. 4C).

Significant expression of many genes (cbpA, manA, xynA,

and/or pelA) was observed when cells grew on the polymericsubstrates (Fig. 4A). In the presence of cellulose, cbpA andmanA transcripts were strongly induced (Fig. 4A, lane C).Although cellulose induced the xynA gene, the level of expres-sion of xynA was higher in cells cultured on xylan (Fig. 4A, laneX). Very few pelA transcripts were found with the cellulose-grown culture. Although the cell mass on locust bean gum(galactomannan)-containing medium was the highest (Fig. 4B,lane M), the transcription of all tested genes was low (Fig. 4A,lane M). Pectin was clearly observed to stimulate the expres-sion of pelA transcripts (Fig. 4A, lane P). Pectin is also aninducer for cellulase genes such as cbpA and engE (Fig. 5).Both the 8-kb (cbpA-exgS) and 12-kb (cbpA-exgS-engH-engK)transcripts from the cbpA gene cluster were present (Fig. 5,lane 1) (9). Relatively, many fewer pelA transcripts were de-tected, however, when the bacterium was cultured on cellulose,locust bean gum, or xylan (Fig. 4A, lane P). These resultsindicated that polymeric substrates generally induced polymer-specific degrading enzymes, but interestingly, pectin and xylandid induce a number of cellulase genes.

Glucose repression of (hemi-)cellulase gene expression. Itwas reported previously that synthesis of cellulosomes wasreduced when glucose was used as the growth substrate (4, 22).To determine whether glucose acts by repressing transcriptionof (hemi-)cellulase genes, we performed Northern blot analysiswith intragenic probes derived from cbpA, manA, and engE.Glucose (final concentration, 0.25%) was added to activelygrowing cellulose (final concentration, 0.5%) cultures at 68 hof cultivation, when the cellulolytic system of this bacteriumwas being actively produced. Total cellular RNA was isolatedfrom cells grown in cellulose-containing medium before andafter addition of glucose. An increase in the cell mass wasobserved immediately after addition of the glucose, indicatingthat the bacteria started growing at the expense of glucose(Fig. 6A). cbpA, manA, and engE gene expression was almostcompletely repressed in cellulose cultures after addition ofglucose until the culture was incubated for 10 h at 37°C (Fig.6B). However, the cbpA, engE, and manA transcripts weredetected at low levels again after 10 h of incubation (Fig. 6B,lane 5). Thus, cellulase transcription was repressed when glu-cose was present but was derepressed upon exhaustion of glu-cose in the medium. This is the first transcriptional analysisthat reports that soluble carbohydrates cause a rapid repres-sion of cellulose-inducible systems of C. cellulovorans.

Di- or monomeric sugar repression of (hemi-)cellulase geneexpression. In order to confirm the existence of a repressiveeffect of additional di- or monomeric sugars other than glucoseon the (hemi-)cellulolytic system of C. cellulovorans, the cellswere grown to exponential phase in 1% cellulose-containingmedium. Subsequently, 0.5% monosaccharides (fructose, glu-cose, mannose, and galactose) and 0.5% disaccharides (lac-tose, maltose, sucrose, and cellobiose) were added to fullyinduced cellulose-based cultures (final concentration, celluloseto added carbon, 2:1 [wt/wt], 0.5 to 0.25% [wt/wt]). We alsotested 1 mM sophorose and 1% cellulose as controls. A sig-nificant reduction in the cellulase transcripts was observedafter the sugar addition in most cases, in contrast to the tran-scription level in the cellulose control (Fig. 7B and C). Whencellulose media were supplemented with glucose, lactose, orsucrose, the level of cellulase gene expression (i.e., cbpA,

FIG. 3. RT-PCR analysis of (hemi-)cellulase transcripts producedin C. cellulovorans grown on different sugars. Total RNA (1 �g) wasisolated from C. cellulovorans cultivated on media containing 0.5%monosaccharides (lanes 1 through 4; glucose, fructose, galactose, andmannose), 0.5% disaccharides (lanes 5 through 8; lactose, maltose,sucrose, and cellobiose), or 1% cellulose (lane 9) as the sole carbonsource. Primers specific for the cbpA, engH, engE, or xynA genes wereused to amplify fragments by PCR. In the negative controls, the reac-tions were performed in the absence of RT or RNA templates (datanot shown).

6070 HAN ET AL. J. BACTERIOL.

on March 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

manA, and engE) was greatly reduced after 1 h of incubation,although the inclusion of cellobiose, fructose, galactose, mal-tose, mannose, or sophorose in cellulose medium did not re-press the synthesis of these mRNAs as much (Fig. 7C, leftpanel). In addition, the cbpA and engE transcripts were de-tected at low levels at 1 h of incubation and increased after 7 hof incubation (Fig. 7C, right panel). A similar analysis using

RT-PCR, a less quantitative measure but with relatively highsensitivity on low levels of mRNAs, showed the same patternof (hemi-)cellulase transcription after 7 h of incubation (Fig.7B). No significant change in xynA and pelA transcription wasdetected before or after addition of carbon compounds underall conditions tested (Fig. 7B and C). The total cell proteinconcentration in the culture on all carbon sources tested wasnot significantly different (Fig. 7A).

Polymeric carbon compound repression of (hemi-)cellulasegene expression. These studies were carried out to determinethe effects of hemicellulose polymers and pectin on cellulosedegradation. The repressing or inducing action of additionalpolysaccharides on (hemi-)cellulase gene expression wasshown by addition of polysaccharides to fully induced cellu-lose-based mid-log-phase cultures (final concentration, cellu-lose to additional carbon, 1:1 [wt/wt], 0.5 to 0.5%). The re-sponses of cbpA, manA, xynA, and pelA transcriptions to addedcarbon sources on cellulose culture were analyzed by RNA slotblotting (Fig. 8A) after an additional 18 and 32 h of incubation(Fig. 8B). The growth rates indicate that cells grew most rap-idly on mannan, pectin, and xylan in decreasing fashion uponaddition of these polymers to the cellulose culture (Fig. 8B).Upon addition of locust bean gum (mannan) or xylan to cel-lulose cultures, cbpA transcription was repressed upon an ad-ditional 18 h of incubation and derepressed after 36 h ofincubation (Fig. 8A, lanes M and X). Addition of pectin re-

FIG. 4. RNA slot blot analysis of the expression of cbpA, manA, xynA, and pelA genes in C. cellulovorans grown on the different polysaccharides.Total RNA (50 and 500 ng) (A) and total protein (B) were isolated from C. cellulovorans cultivated on media containing 1% cellulose (48-h culture,panel A, lane C), 1% locust bean gum (18-h culture, panel A, lane M), 1% pectin (18-h culture, panel A, lane P), or 1% xylan (18-h culture, panelA, lane X) as the sole carbon source. The gene-specific probes used are indicated on the left of panel A. The different DIG-labeled probes wereprepared (each from 1 �g of template) by random primed labeling (see Materials and Methods). Dilutions (0.1 to 10 pg) of each probe werespotted on a nylon membrane, and labeling sensitivity (amount of labeled DNA per spot) was determined in order to use similar amounts of theprobes (C).

FIG. 5. Northern hybridization of C. cellulovorans RNA. TotalRNA was isolated from cells grown in the presence of 1% pectin as thesole carbon source. RNA (10 �g) was subjected to electrophoresisthrough 1.5% formaldehyde gels and transferred to nylon membranes,which were subsequently hybridized with the DIG-labeled cbpA (lane1-2), engH (lane 2)-, engE (lane 3)-, and pelA (lane 4)-specific probes.The ovals represent full-length specific transcripts. The sizes of theRNA markers (M) are indicated at the left in bases.

VOL. 185, 2003 REGULATION OF CELLULASE AND HEMICELLULASE GENES 6071

on March 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

pressed cbpA transcription only slightly. Although the locustbean gum (galactomannan) is a favored substrate for mannan-ase (ManA) (32), the addition of mannan depressed manAtranscription after an additional 18 h of incubation (Fig. 8A,lane M). However, manA transcription was clearly detectedagain after 36 h of incubation. On the other hand, the datashowed that the supplemented xylan induced xynA after both18- and 36-h incubation periods (Fig. 8A, lane X). Low pelAtranscription was detected in medium supplemented with pec-tin at 18 and 36 h (Fig. 8A, lane P). These results indicated thatthe presence of hemicellulosic polymers could differentiallyaffect genes for the utilization of cellulose.

DISCUSSION

For investigation of the expression pattern of (hemi-)cellu-lase genes, mRNA was isolated from cells from continuouscultures taken at various time points. The data demonstrategeneral and specific regulatory patterns in expression of(hemi-)cellulase genes by C. cellulovorans, including featuresof their relative expression levels under different culture con-ditions, i.e., various carbon sources and growth phases. Forinstance, cellulose and cellobiose induced the transcription ofmost of the (hemi-)cellulase genes (i.e., cbpA-exgS-engH-engK,

manA, engE, and xynA) and the time course of each (hemi-)cellulase gene transcription was approximately the same in allcases. This is the first report at the transcriptional level that the(hemi-)cellulase genes in a clostridial (hemi-)cellulolytic sys-tem that included cbpA-exgS (9), engEi, and xynA are coordi-nately expressed when various substrates such as cellobioseand cellulose are used (Fig. 1 and 2). It was also found that theexpression of the noncellulosomal cellulase gene, engF, wasregulated just as other cellulosomal cellulase genes were reg-ulated (Fig. 1 and 2). However, through visual inspection ofNorthern blot analysis (Fig. 1 and 2), the time courses ofendoglucanase gene expression (engE and engF) were thoughtto be greatly different from those of the noncellulosomal �-L-arabinofurasnosidase gene arfA. Cellulose and hemicelluloseare closely associated in nature, and it appears that C. cellulo-vorans has a mechanism(s) to ensure efficient utilization ofboth types of polymers. These results suggest that a commonregulatory mechanism may exist at the transcriptional level for(hemi-)cellulase induction by cellulose and cellobiose. A cel-lulose metabolite such as cellobiose or a derivative of cellobi-ose may act as an inducer and may bind to a receptor proteinin a signal transduction pathway, and this pathway may thenlead to cellulase induction.

Significant expression of most of the genes was observedwith polysaccharide substrates such as cellulose, pectin, andxylan, followed by moderate levels with other substrates suchas cellobiose and fructose. Low levels of (hemi-)cellulasemRNAs derived from cells grown with lactose, mannose, andlocust bean gum (mannan) were observed, and little or noexpression was detected with cells grown on glucose, galactose,maltose, and sucrose. These results give a general picture ofthe potential for (hemi-)cellulase expression when cells aregrown on different carbon sources. It was thought that cellulaseexpression would not occur on carbon sources that promotedrapid growth but would be stimulated by polysaccharides thatwere difficult to degrade (14, 24). It is noteworthy that expres-sion of cbpA-exgS and engE was especially strong under allconditions tested. The relative transcript levels of the differentcellulase genes were comparable to the amounts of the specificproteins produced in the culture medium. This finding is inaccordance with previous data on optimization of enzyme pro-duction, which showed that the highest CbpA, ExgS, and EngEactivity levels were present when cells were grown on cellulose(8, 20, 22). In the general model for the induction of cellulaseand hemicellulase expression, a sensor enzyme is constitutivelyexpressed which hydrolyzes cellulose and/or hemicellulose intooligosaccharides that enter the bacterium and activate the ex-pression of the (hemi-)cellulase genes (31, 35). The presentobservations indicate that (hemi-)cellulase genes in C. cellulo-vorans are expressed constitutively at low levels but are in-duced to express at higher levels in the presence of certainpolysaccharides, such as cellulose. It has been reported that abasal constitutive level of cellulosomal proteins was synthe-sized when the cells were grown with glucose or cellobiose (4,22). These cellulases were secreted into the extracellular cul-ture medium at a very low rate over a long period of incubation(2, 6, 22). These results are not contradictory to our presenttranscriptional analyses, since it is difficult to analyze the ex-tremely low levels of transcripts (e.g., fewer than 10 strands ofmRNA per cell) by methods such as Northern blotting or

FIG. 6. Growth curve (A) showing the time course of cbpA, manA,and engE transcription during growth of C. cellulovorans on cellulosemedium with subsequent glucose supplementation ([graphic]) andwithout glucose supplementation ([graphic]), as determined by North-ern blot analysis (B). Total RNA (5 �g) was isolated from cells grownon 1% cellulose medium without additional glucose (lanes 1 and 2) orsupplemented with 0.5% glucose (lanes 3 through 5) and hybridized tospecific probes. The number of the band in panel B corresponds to thenumber on the C. cellulovorans growth curve in panel A. The probeswere labeled to a similar sensitivity, and the labeling sensitivity methodcorresponds to that described in Fig. 4C.

6072 HAN ET AL. J. BACTERIOL.

on March 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

RT-PCR. The constitutive level of (hemi-)cellulase expressionis therefore very low. This type of result has also been reportedfor other glucose catabolite-repressed systems where proteinswere detected but their mRNAs could not be detected (12, 21).Our results indicated that certain carbon sources induced highlevels of expression of one gene or a set of genes, whereas theeffect on expression of other genes was weak or insignificant.This pattern varied depending on the carbon source. Althoughbeing a general inducing compound for (hemi-)cellulases, cel-lulose induced expression of the cellulase genes, such as cbpAand engE, most strongly. This may be an effect caused bycellobiose, other oligosaccharides, or derivatives of cellobiosethat are formed in the cell. The expression of hemicellulasegenes (manA and xynA) in cellulose-based medium could beinduced by cellulose or by certain contaminants in the com-mercial preparations of cellulose (10). Xylan especially causedexpression of the hemicellulase genes, such as xynA and manA,and was the most potent carbon source for induction of thexylanase gene (xynA). On the other hand, although celluloseand xylan did not act as inducers for the pectate lyase A gene,pectin definitely induced pelA gene expression. Thus, theseresults indicated that certain polymeric substrates were capa-ble of activating specific genes.

The induction and repression of cellulases by mixed sub-strates of cellulosic and hemicellulosic sugars reported here is

interesting. In the degradation of lignocellulosic substrates bymicroorganisms, it has been established that the first growthphase is developed at the expense of hemicelluloses and thatthe cellulase system is developed in a second stage (11, 19). Itis feasible that the products of certain hemicellulose degrada-tion, especially locust bean gum, could act as repressors of thecellulolytic system at high concentrations and that as theirconcentrations drop to low levels, the cellulolytic system isderepressed. This could explain the rapid pattern of growth onhemicelluloses and the sequence of enzyme production ofhemicellulases and cellulases. This fact also supports the ideaof an interrelationship between the systems regulating(hemi-)cellulases in this bacterium. The observed repression ofcellulases by high glucose and cellobiose concentrations is sim-ilar to that found for other cellulolytic bacteria (23, 26). How-ever, hemicellulose repression of the cellulolytic activity ofcellulose cultures has not been reported previously. This mightindicate a hierarchical relationship between the systems regu-lating cellulases and hemicellulases which would be particu-larly important in the degradation of complex lignocellulosicmaterials in nature.

Certain di- or monosaccharides (i.e., fructose, lactose, andcellobiose) induced expression of (hemi-)cellulase genes in C.cellulovorans. The cellulolytic bacteria like C. thermocellumwere reported to produce cellulases when grown with soluble

FIG. 7. Effect of added di- or monosaccharides on (hemi-)cellulase transcription in cellulose medium, as determined by RT-PCR analysis andRNA dot blot analysis. The numbers of the bars on the graph (A) correspond to different sugars (1 through 10: glucose, fructose, galactose,mannose lactose, maltose, sucrose, cellobiose, sophorose, and cellulose) and to the numbers of the bands in RT-PCR analysis (B) and RNA dotblot analysis (C). Total protein and total RNA (1 �g) was isolated from cells grown in 1% cellulose medium at 1 and 7 h of incubation aftersupplementation with 0.5% monosaccharides (panels B and C, lanes 1 through 4: glucose, fructose, galactose, and mannose), 0.5% disaccharides(panels B and C, lanes 5 through 8: lactose, maltose, sucrose, and cellobiose), 1 mM sophorose (panels B and C, lane 9) and 1% cellulose (panelsB and C, lane 10). The probes for RNA dot blot analysis (C) were labeled to a similar sensitivity, and the labeling sensitivity method correspondsto that described in Fig. 4C.

VOL. 185, 2003 REGULATION OF CELLULASE AND HEMICELLULASE GENES 6073

on March 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

carbon sources such as fructose, glucose, and cellobiose (25).Nevertheless, the usual pattern observed with C. cellulovoranswas the lack of expression of (hemi-)cellulases in the presenceof the easily metabolizable mono- or disaccharides such asglucose, and a catabolite repression-type mechanism seems toexist which mediates control of expression of various genesencoding different extracellular hydrolases as well as the scaf-folding protein.

ACKNOWLEDGMENTS

We are grateful to Helen Chan for skillful technical assistance andfor preparation of the media.

This research was supported in part by the Research Institute ofInnovative Technology for the Earth (RITE), Japanese Ministry ofEconomy, Trade, and Industry (METI), and by grant DE-DDF03-92ER20069 from the U.S. Department of Energy.

REFERENCES

1. Attwood, G. T., H. P. Blaschek, and B. A. White. 1994. Transcriptionalanalysis of the Clostridium cellulovorans endoglucanase gene, engB. FEMSMicrobiol. Lett. 124:277–284.

2. Bayer, E. A., Y. Shoham, and R. Lamed. 2000. Cellulose-decomposing bac-teria and their enzyme systems, p. 1–74. In M. Dworkin, S. Falkow, E.Rosenberg, K.-H. Schleifer, and E. Stackebrandt (ed.), The prokaryotes: anevolving electronic resource for the microbiological community, 3rd ed.Springer-Verlag, New York, N.Y.

3. Bensadoun, A., and D. Weinstein. 1976. Assay of proteins in the presence ofinterfering materials. Anal. Biochem. 70:241–250.

4. Blair, B. G., and K. L. Anderson. 1999. Regulation of cellulose-induciblestructures of Clostridium cellulovorans. Can. J. Microbiol. 45:242–249.

5. Brown, R. E., K. L. Jarvis, and K. J. Hyland. 1989. Protein measurementusing bicinchoninic acid: elimination of interfering substances. Anal. Bio-chem. 180:136–139.

6. Doi, R. H., and Y. Tamaru. 2001. The Clostridium cellulovorans cellulosome:an enzyme complex with plant cell wall degrading activity. Chem. Rec.1:24–32.

7. Fierobe, H. P., E. A. Bayer, C. Tardif, M. Czjzek, A. Mechaly, A. Belaich, R.Lamed, Y. Shoham, and J. P. Belaich. 2002. Degradation of cellulose sub-strates by cellulosome chimeras. Substrate targeting versus proximity ofenzyme components. J. Biol. Chem. 277:49621–49630.

8. Goldstein, M. A., M. Takagi, S. Hashida, O. Shoseyov, R. H. Doi, and I. H.Segel. 1993. Characterization of the cellulose-binding domain of the Clos-tridium cellulovorans cellulose-binding protein A. J. Bacteriol. 175:5762–5768.

9. Han, S. O., H. Yukawa, M. Inui, and R. H. Doi. 2003. Transcription ofClostridium cellulovorans cellulosomal cellulase and hemicellulase genes. J.Bacteriol. 185:2520–2527.

10. Hrmova, M., P. Biely, and M. Vrsanska. 1986. Specificity of cellulase and�-xylanase induction in Trichoderma reesei QM 9414. Arch. Microbiol. 144:307–311.

11. Hrmova, M., E. Petrakova, and P. Biely. 1991. Induction of cellulose- andxylan-degrading enzyme systems in Aspergillus terreus by homo- and hetero-disaccharides composed of glucose and xylose. J. Gen. Microbiol. 137:541–547.

12. Ilmen, M., A. Saloheimo, M. L. Onnela, and M. E. Penttila. 1997. Regulationof cellulase gene expression in the filamentous fungus Trichoderma reesei.Appl. Environ. Microbiol. 63:1298–1306.

13. Jindou, S., S. Karita, E. Fujino, T. Fujino, H. Hayashi, T. Kimura, K. Sakka,and K. Ohmiya. 2002. �-Galactosidase Aga27A, an enzymatic component ofthe Clostridium josui cellulosome. J. Bacteriol. 184:600–604.

14. Kalra, M. K., M. S. Sidhu, D. K. Sandhu, and R. S. Sandhu. 1984. Produc-tion and regulation of cellulases in Trichoderma reesei. Appl. Microbiol.Biotechnol. 20:427–429.

15. Kataeva, I. A., R. D. Seidel III, X. L. Li, and L. G. Ljungdahl. 2001. Prop-erties and mutation analysis of the CelK cellulose-binding domain from theClostridium thermocellum cellulosome. J. Bacteriol. 183:1552–1559.

16. Kosugi, A., K. Murashima, and R. H. Doi. 2002. Characterization of twononcellulosomal subunits, ArfA and BgaA, from Clostridium cellulovoransthat cooperate with the cellulosome in plant cell wall degradation. J. Bacte-riol. 184:6859–6865.

17. Kosugi, A., K. Murashima, and R. H. Doi. 2001. Characterization of xylano-lytic enzymes in Clostridium cellulovorans: expression of xylanase activitydependent on growth substrates. J. Bacteriol. 183:7037–7043.

18. Kosugi, A., K. Murashima, and R. H. Doi. 2002. Xylanase and acetyl xylanesterase activities of XynA, a key subunit of the Clostridium cellulovoranscellulosome for xylan degradation. Appl. Environ. Microbiol. 68:6399–6402.

FIG. 8. Effect of added polysaccharides on (hemi-)cellulase transcription in cellulose medium, as determined by RNA dot blot analysis. TotalRNA (1 �g) (A) and total protein (B) were isolated from cells grown in 20 ml of 1% cellulose medium supplemented with 20 ml of 1% cellulose(panels A and B, lane C), locust bean gum (panels A and B, lane M), pectin (panels A and B, lane P), or xylan (panels A and B, lane X). Theprobes were labeled to a similar sensitivity, and the labeling sensitivity method corresponds to that described in Fig. 4C.

6074 HAN ET AL. J. BACTERIOL.

on March 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

19. Kulkarni, N., A. Shendye, and M. Rao. 1999. Molecular and biotechnologicalaspects of xylanases. FEMS Microbiol. Rev. 23:411–456.

20. Liu, C. C., and R. H. Doi. 1998. Properties of exgS, a gene for a major subunitof the Clostridium cellulovorans cellulosome. Gene 211:39–47.

21. Margolles-Clark, E., M. Ilmen, and M. Penttila. 1997. Expression patterns often hemicellulase genes of the filamentous fungus Trichoderma reesei onvarious carbon sources. J. Biotechnol. 57:167–179.

22. Matano, Y., J. S. Park, M. A. Goldstein, and R. H. Doi. 1994. Cellulosepromotes extracellular assembly of Clostridium cellulovorans cellulosomes. J.Bacteriol. 176:6952–6956.

23. Mishra, S., P. Beguin, and J. P. Aubert. 1991. Transcription of Clostridiumthermocellum endoglucanase genes celF and celD. J. Bacteriol. 173:80–85.

24. Murashima, K., A. Kosugi, and R. H. Doi. 2002. Determination of subunitcomposition of Clostridium cellulovorans cellulosomes that degrade plant cellwalls. Appl. Environ. Microbiol. 68:1610–1615.

25. Nochur, S. V., M. F. Robert, and A. L. Demain. 1993. True cellulase pro-duction by Clostridium thermocellum grown on different carbon sources.Biotechnol. Lett. 15:641–646.

26. Robson, L. M., and G. H. Chambliss. 1989. Cellulases of bacterial origin.Enzyme Microb. Technol. 11:626–644.

27. Sheweita, S. A., A. Ichi-ishi, J.-S. Park, C. Liu, L. M. Malburg, Jr., and R. H.Doi. 1996. Characterization of engF, a gene for a non-cellulosomal Clostrid-ium cellulovorans endoglucanase. Gene 182:163–167.

28. Shoseyov, O., and R. H. Doi. 1990. Essential 170-kDa subunit for degrada-

tion of crystalline cellulose by Clostridium cellulovorans cellulase. Proc. Natl.Acad. Sci. USA 87:2192–2195.

29. Shoseyov, O., M. Takagi, M. A. Goldstein, and R. H. Doi. 1992. Primarysequence analysis of Clostridium cellulovorans cellulose binding protein A.Proc. Natl. Acad. Sci. USA 89:3483–3487.

30. Sleat, R., R. A. Mah, and R. Robinson. 1984. Isolation and characterizationof an anaerobic, celluloytic bacterium, Clostridium cellulovorans sp. nov.Appl. Environ. Microbiol. 48:88–93.

31. Suto, M., and F. Tomita. 2001. Induction and catabolite repression mecha-nisms of cellulase in fungi. J. Biosci. Bioeng. 92:305–311.

32. Tamaru, Y., and R. H. Doi. 2000. The engL gene cluster of Clostridiumcellulovorans contains a gene for cellulosomal manA. J. Bacteriol. 182:244–247.

33. Tamaru, Y., and R. H. Doi. 2001. Pectate lyase A, an enzymatic subunit ofthe Clostridium cellulovorans cellulosome. Proc. Natl. Acad. Sci. USA 98:4125–4129.

34. Tamaru, Y., and R. H. Doi. 1999. Three surface layer homology domains atthe N terminus of the Clostridium cellulovorans major cellulosomal subunitEngE. J. Bacteriol. 181:3270–3276.

35. Zeilinger, S., R. L. Mach, M. Schindler, P. Herzog, and C. P. Kubicek. 1996.Different inducibility of expression of the two xylanase genes xyn1 and xyn2in Trichoderma reesei. J. Biol. Chem. 271:25624–25629.

36. Zverlov, V. V., K.-P. Fuchs, and W. H. Schwarz. 2002. Chi18A, the endo-chitinase in the cellulosome of the thermophilic, cellulolytic bacterium Clos-tridium thermocellum. Appl. Environ. Microbiol. 68:3176–3179.

VOL. 185, 2003 REGULATION OF CELLULASE AND HEMICELLULASE GENES 6075

on March 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from