|Home | About | Journals | Submit | Contact Us | Français|
Cellulosilyticum ruminicola H1 is a newly described bacterium isolated from yak (Bos grunniens) rumen and is characterized by its ability to grow on a variety of hemicelluloses and degrade cellulosic materials. In this study, we performed the whole-genome sequencing of C. ruminicola H1 and observed a comprehensive set of genes encoding the enzymes essential for hydrolyzing plant cell wall. The corresponding enzymatic activities were also determined in strain H1; these included endoglucanases, cellobiohydrolases, xylanases, mannanase, pectinases, and feruloyl esterases and acetyl esterases to break the interbridge cross-link, as well as the enzymes that degrade the glycosidic bonds. This bacterium appears to produce polymer hydrolases that act on both soluble and crystal celluloses. Approximately half of the cellulytic activities, including cellobiohydrolase (50%), feruloyl esterase (45%), and one third of xylanase (31%) and endoglucanase (36%) activities were bound to cellulosic fibers. However, only a minority of mannase (6.78%) and pectinase (1.76%) activities were fiber associated. Strain H1 seems to degrade the plant-derived polysaccharides by producing individual fibrolytic enzymes, whereas the majority of polysaccharide hydrolases contain carbohydrate-binding module. Cellulosome or cellulosomelike protein complex was never isolated from this bacterium. Thus, the fibrolytic enzyme production of strain H1 may represent a different strategy in cellulase organization used by most of other ruminal microbes, but it applies the fungal mode of cellulose production.
The ruminant rumens are long believed to be the anaerobic environments efficiently degrading the plant-derived polysaccharides, which is attributed to the inhabited abundant rumen microorganisms. They implement the fibrolytic degradation by the combination of the enzymes comprising of cellulases, hemicellulases, and to a lesser extent pectinases and ligninases (12). The rumen bacteria are outnumbered of the other rumen microbes; however, only a few of cellulolytic bacteria have been isolated from rumens. Ruminococcus flavefaciens, Ruminococcus albus, and Fibrobacter succinogenes are considered to be the most important cellulose-degrading bacteria in the rumen (18), and they produce a set of cellulolytic enzymes, including endoglucanases, exoglucanases (generally cellobiohydrolase), and β-glucosidases, as well as hemicellulases. In addition, the predominant ruminal hemicellulose-digesting bacteria such as Butyrivibrio fibrisolvens and Prevotella ruminicola lack the ability to digest cellulose but degrade xylan and pectin and utilize the degraded soluble sugars as substrates (10, 14). Although the robust cellulolytic species F. succinogenes degrades xylan, it cannot use the pentose product as a carbon source (24). Culture-independent approaches indicate that the three cellulolytic bacterial species represent only ~2% of the ruminal bacterial 16S rRNA (43). Therefore, many varieties of rumen microbes remain uncultured (2). In recent years, rumen metagenomics studies have revealed the vast diversity of fibrolytic enzymes, multiple domain proteins, and the complexity of microbial composition in the ecosystem (9, 17). Hence, it is likely that the entire microbial community is necessary for the implementation of an efficient fibrolytic process in the rumen, including the uncultured species.
In the rumen and other fibrolytic ecosystems, cellulolytic bacteria have to cope with the structural complexity of lignocelluloses and the interspecies competition; thus, not only a variety of plant polymer-degrading enzymes but also a noncatalytic assistant strategy, such as including adhesion of cells to substrates by a variety of anchoring domains, is required (8, 33, 38, 39). The (hemi)cellulolytic enzyme systems have been intensively studied for nonrumen anaerobic bacteria, including Clostridium thermocellum (19, 40), Clostridium cellulolyticum (6), Clostridium cellulovorans (13), and Clostridium stercorarium (47), as well as the rumen species, Rumicoccocus albus (35), Ruminococcus flavefaciens (32), and Fibrobacter succinogenes (4). The results indicate that most of them, except for Fibrobacter succinogenes, produce multiple cellulolytic enzymes integrated in a complex, cellulosome, and free individual proteins.
The yak (Bos grunniens) is a large ruminant (~1,000 kg) in the bovine family that lives mainly on the Qinghai-Tibetan Plateau in China at an altitude of 3,000 m above sea level. It is a local species that lives mainly on the world's highest plateau. Yaks live in a full-grazing style with grasses, straws, and lichens as their exclusive feed, so the yak rumen can harbor a microbial flora distinct from those of other ruminants due to their fiber-component diet, since diet can be a powerful factor in regulating mammalian gut microbiome (27). A very different prokaryote community structure was revealed for yak rumen in our previous work based on the 16S rRNA diversity, which showed fewer phyla than for cattle but that a higher ratio of sequences was related to uncultured bacteria (2).
We previously isolated a novel anaerobic fibrolytic bacterium, Cellulosilyticum ruminicola H1, from the rumen of a domesticated yak (11). Strain H1 grew robustly on natural plant fibers such as corn cob, alfalfa, and ryegrass as the sole carbon and energy sources, as well as on a variety of polysaccharides, including cellulose, xylan, mannan, and pectin, but not monosaccharides such as glucose, which is preferred by most ruminal bacteria. In the present study, using a draft of its genome and enzymatic characterization, we analyzed the enzymatic activities and the structures of the polymer hydrolases of strain H1 that were involved in the hydrolysis of complex polysaccharides.
C. ruminicola (CGMCC 1.5065T) was cultured at 38°C under 1.01 × 105 Pa of CO2 gas phase in RC medium as described previously (11), using cellobiose as a carbon source. For enzyme induction experiments, we used the following as alternative carbon sources: 5% (wt/vol) filter paper (Whatman I), 5% (wt/vol) corn cob powder, or 1% (wt/vol) xylan (birch wood).
Strain H1 was cultured in 200 ml of RC medium for 11 days, after which a crude enzyme preparation was made by the method of Shoseyov and Doi (42) with slight modifications. The spent culture was centrifuged at 12,000 × g for 10 min to remove bacterial cells and the remaining substrate debris. The proteins in supernatant were precipitated with 80% (NH4)2SO4, dissolved in 2 ml of PC buffer (50 mM phosphate, 12 mM citrate, 1 mM sodium azide [pH 7.0]), and then dialyzed against the same buffer. The PC solution was used as the crude enzyme preparation.
Protein concentrations were determined by using BCA protein assay kit (Thermo Scientific, Rockford, IL) with bovine serum albumin as the calibration standard.
Endoglucanase, xylanase, mannanase, and pectinase activities were assayed according to the method of Murashima et al. (34), except that dinitrosalicylic acid was used for determination of reducing sugars (31). One unit of enzyme activity was defined as the amount of enzyme that liberated 1 μmol of reducing sugars per min. Cellobiohydrolase, β-glucosidase, β-xylosidase, arabinofuranosidase, and acetyl esterase were assayed according to the method of Adelsberger et al. (1). One unit of enzyme activity was defined as the amount of enzyme that liberated 1μmol of p-nitrophenol per min. Feruloyl esterase was assayed according to the method of Blum et al. (7) except that the reaction performed at 37°C.
Avicel (crystalline cellulose), xylan (birch wood), locust bean gum, pectin, pNP-cellobioside, pNP-glucopyranoside, pNP-xylopyranoside, pNP-acetate, and pNP-arabinofuranoside were purchased from Sigma (Beijing, China). Carboxymethyl cellulose (CMC sodium salt) and birch wood xylan were purchased from Fluka (Beijing, China).
Avicel was added to the crude enzyme preparation at a final concentration of 2% (wt/vol). After incubation at 37°C for 30 min, the enzyme-Avicel slurry was centrifuged at 8,000 × g for 5 min and then washed three times with PC buffer to remove the nonspecifically adhered proteins. The ratios of Avicel-bound activity of each enzyme were calculated according to the procedure described by Pason et al. (36).
Avicel-bound proteins were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) with 10% polyacrylamide gels to determine the protein composition. The zymograms of endoglucanase and xylanase were then determined inside the gel as follows. SDS was removed by incubating the gel in a renaturation buffer (25 mM Tris-HCl, 0.1% Triton X-100 [pH 7.0]) at room temperature overnight, and then overlaid with a thin agarose layer containing 0.1% of CMC-Na or xylan on the renatured gel. After 1 h of incubation at 37°C, the agarose layer was stained with Congo red solution (1%) for 5 min and washed with 1 M NaCl to visualize the decolored zone as an indication of the corresponding enzyme activities.
The stained SDS-PAGE gels were decolored with a destaining solution (methanol-H2O-acetic acid [45:45:10]) and subjected to in-gel digestion as previously described (25). The digested peptide mixtures from each sample were desalted, dissolved in 20 μl of 0.1% formic acid, and loaded onto a reverse phase column (BioBasic C18, 300 Å, 5-μm silica, 180 μm by 10 cm; ThermoHypersil, Allentown, PA). The flow rate was maintained at 100 μl/min before splitting and at 1.0 μl/min after the flow split. The gradient was started at 5% acetonitrile in 0.1% formic acid for 20 min, ramped to 50% acetonitrile for 80 min, and finally ramped to 95% acetonitrile for an additional 20 min. The resolved peptides were subjected to MS/MS analysis with LCQ Deca XP Plus ion-trap mass spectrometer (ThermoFinnigan, San Jose, CA) equipped with a nanospray source.
High molecular genomic DNA was isolated from C. ruminicola H1 and used to support whole-genome sequencing using the 454 LifeSciences GS20 pyrosequencing system (29) located at the Center for Genomic Sciences. We performed two complete sequencing runs that provided 691,130 reads with an average read length of 107 bases which were initially assembled using the Newbler assembler into 114 scaffolds; subsequent reprocessing of the data with a later version of the software reduced the number of scaffolds to 66. A nearly complete genomic sequence of 4,033,077 bp was assembled from the 26× average genomic coverage. Sequencing particulars and DNA sequence analyses were performed principally as described previously (23). Lander-Waterman statistics predicted that >99.9% of the genome was sequenced. Regions of duplicated sequence caused most of the assembly gaps. Informal comparisons between high-quality Sanger reads and 454 data suggested an error rate of less than 1 in 10,000 bases. Most base call errors are single base insertions or deletions in homonucleotide repeats that can result in frameshift artifacts.
All MS/MS spectra were searched by using Thermo Finnigan Bioworks 3.1 against the genome database of C. ruminicola H1, with a static modification of +57.0215 Da on cysteine residue and a differential modification of 15.9994 on methionine. The precursor ion mass tolerance was 1.4 Da, and the fragment ion mass tolerance was 1.5 Da. The following steps were performed for the data processing: (i) SEQUEST criteria were used to perform an initial filtration (DeltaCn ≥ 0.1; Rsp = 1; Xcorr ≥ 1.9 for singly charged fragments; Xcorr ≥ 2.2 for doubly charged; and Xcorr ≥ 3.75 for triply charged) and (ii) the AMASSv188.8.131.52 program developed by Sun et al. (44) and Li et al. (28) was used to further filter the SEQUEST results based on three parameters (MatchPct ≥ 60, Cont ≥ 40, and Rscore <2.6). Proteins with two or more spectra approved by AMASS would be accepted as positive identifications. For proteins with multiple isoforms or multiple entries in the databases, only the major form of the protein was accepted unless a specific peptide pointed specifically to a region of the protein, which existed only in one of the isoforms.
The genomic DNA of C. ruminicola H1 was extracted as previously described (11). The primers listed in Table Table11 were used to amplify the genes corresponding to the cellulose-bound fibrolytic enzymes Cel48A, Cel9A, and Fae1A and to provide restriction enzyme (RE) digestion sites for cloning. PCR amplifications were carried out with Pfu DNA polymerase (Promega, Madison, WI) for 30 cycles, with each cycle consisting of denaturation at 95°C for 1 min, annealing at 51°C for 1 min, and elongation at 72°C for 6 min (elongation of 3 min for Fae1). The PCR fragments were digested with the appropriate REs (Table (Table1)1) and cloned into the His tag expression vectors pET-15b, pET-28a, and pET-15b, respectively. The three plasmids containing the fibrolytic genes were then, respectively, transformed into E. coli BL21(DE3) cells. After growth at 37°C in LB medium supplemented with 100 μg of ampicillin/ml (for Cel48A and Fae1A) and 50 μg of kanamycin/ml (for Cel9A) to an optical density of 0.4 to 0.6, overproduction was induced by addition of 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside). Cells were collected after an additional 3-h culture, and the recombinant proteins were purified by nickel-affinity chromatography as described previously (45).
Since C. ruminicola H1 grew robustly on natural plant fibers such as corn cob, alfalfa, and ryegrass, its genome was sequenced to identify the genes pertinent for lignocellulose degradation. A draft genome sequence was obtained using the 454 LifeSciences GS20 platform. The genome was determined to contain ~4,033,707 bp which were assembled into 66 scaffolds. Automated annotation using PGAAP (http://www.ncbi.nlm.nih.gov/genomes/static/Pipeline.html) revealed 4012 open reading frames (ORFs) and showed that 65 ORFs with very high reliabilities encoded a variety of cellulases, hemicellulases, and pectinases (see Table S1 in the supplemental material). Similar to other cellulolytic bacteria, strain H1 possessed only one cellobiohydrolase-encoding gene (Cel48A, ORF 04820). Although strain H1 could not ferment starch, four amylase-encoding genes were identified by annotation. In addition, another 11 ORFs were predicted to be involved in polysaccharide degradation but without defined functions (see Table S2 in the supplemental material).
By using the cultures of strain H1 growing on filter paper, corn cob, or birch wood xylan as growth substrates, respectively, enzymatic activities, including cellulase, hemicellulase, and pectinase were determined (Table (Table2).2). In addition, acetyl and feruloyl esterase activities that are associated with lignocellulose degradation were determined. Generally, the specific enzymatic activity levels were correlated with the substrate on which the strain grew. Thus, the levels of xylanase, xylosidase, and acetyl esterase were all enhanced in birch wood xylan culture compared to the filter paper culture. Similarly, β-glucosidase and cellobiohydrolase activities were higher in the filter paper culture. However, we unexpectedly observed higher arabinofuranosidase and feruloyl esterase activity in filter paper culture than in xylan culture, while higher endoglucanase activity was observed in birch wood culture. These findings suggest that the regulation of some of these enzymes may be linked through specific regulons. The mannanase activity level seemed to be not affected by the substrate type. All of the fibrolytic enzymatic activities, except that of pectinase, were detected at lower levels in corn cob culture compared to the other two culture systems. Xylanase activity was the highest of all of the hydrolytic activities, suggesting that strain H1 preferentially used hemicellulose.
It is generally believed that bacteria degrade cellulosic fibers through the adhesion of their fibrolytic enzymes to their substrates. To determine the cellulose-bound fibrolytic enzymatic activities of strain H1, the spent corn cob culture was precipitated with 80% ammonium sulfate, and the precipitated proteins were used to perform the Avicel (crystalline cellulose) adhesive process. The fibro-polysaccharide hydrolytic enzymatic activities were determined by (i) taking the total ammonium sulfate-precipitated protein preparation as the total activity and (ii) taking the protein preparation that remained after removing Avicel by centrifugation as the nonbound activity, and (iii) the bound fibrolytic enzyme activities were calculated by subtracting the nonbound portion from the total. All of the fibrolytic enzymes in the study were found adhered to Avicel, but at various ratios relative to the total activity.
Cellobiohydrolase exhibited the highest binding ratio (51%), followed by feruloyl esterase (45%), endoglucanase (36%), and xylanase (32%) (Table (Table3).3). Mannanase and pectinase activities presented largely in the nonbound protein fractions, with only 7 and 2% bound, respectively.
To characterize the Avicel-bound fiber-polysaccharide hydrolases produced by strain H1, zymograms were produced and analyzed by subjecting the Avicel-bound proteins to SDS-PAGE and then performing in situ enzymatic activity assays for the renatured proteins inside the gel. Ten protein bands with molecular masses ranging from ~40 to 200 kDa were visualized (Fig. (Fig.1A).1A). Bands 2 and 4 displayed endoglucanase activity, with band 2 showing much higher activity (Fig. (Fig.1B)1B) although less protein visualized, indicating its major contribution to the bound endoglucanase activity of this bacterium. Bands 2, 5, and 6 all showed xylanase activity at moderately low levels. The remaining six protein bands did not show either of the two enzymatic activities. Since it was not feasible to determine other polymer hydrolase activities inside of the gel, their activities were characterized by using purified proteins in the experiments described below.
To identify the proteins that adhered to Avicel, the protein bands were subjected to in-gel trypsin digestion and analyzed by LC-MS/MS. Protein sequence data from the MS experiments were evaluated by using BLAST against the translated draft genome of strain H1. Based on these analyses, proteins from 6 of the 10 bands were annotated as being highly homologous to the known endoglucanases, a cellobiohydrolase, a xylanase, and an esterase (Table (Table4).4). All of the bioinformatic predictions were consistent with the Avicel-bound fibrolytic enzyme activities. The Avicel-bound proteins belonged to glycoside hydrolase families 5, 9, 10, 30, 43, and 48 and carbohydrate esterase family 1 (http://www.cazy.org/index.html), respectively. As expected, most of these proteins possessed carbohydrate-binding modules (CBM), which would enable them to bind to Avicel or other structured polysaccharides, and signal peptide, which would assist the proteins secreted extracellularly. Band 3 (ORF CGSCsYakCAS 04820) and band 4 (ORF CGSCsYakCAS 04825), the most abundant Avicel-bound proteins, were annotated as cellobiohydrolase GH48 and endoglucanase GH9, respectively. Surprisingly, no known CBM motif was identified in the protein encoded by CGSCsYakCAS 04820, although it displayed extraordinary affinity to Avicel. The proteins in band 6 were identified as three different glycoside hydrolases with similar molecular masses (69.4, 73.4, and 77.4 kDa, respectively). One was annotated as an endoglucanase (Cel9B, CGSCsYakCAS 07788), the other two were annotated as endo-1,4-β-xylanases (Xyn10A, CGSCsYakCAS 10610; Xyn30A, CGSCsYakCAS 04919), which is consistent with the xylanase activity identified in the zymogram assay. Although the protein in band 2 had the highest endoglucanase activity by the zymogram assay, xylanase activity was detected in this band as well. The protein in band 7 (ORF CGSCsYakCAS 18248) was annotated as a member of carbohydrate esterase family 1 (Fae1A). The remaining four bands contained proteins that were not annotated as polysaccharide hydrolases but as copper amine oxidase domain protein/surface-layer glycoprotein precursor, ATP-dependent exo-DNase, protease inhibitor precursor, and aldehyde-alcohol dehydrogenase (Table (Table44).
The Avicel-bound proteins in bands 3 and 7 of Fig. Fig.11 were annotated as cellobiohydrolase and feruloyl esterase, respectively. To determine whether the proteins encoded by the genes with expected enzymatic activities, the two genes (CGSCsYakCAS 04820 and CGSCsYakCAS 18248) were cloned with His tags and overexpressed in Escherichia coli. In addition, a gene annotated as encoding an endoglucanase GH9 (CGSCsYakCAS 04825) was also cloned. The three His-tagged proteins purified on Ni columns did show the predicted molecular masses on SDS-PAGE (Fig. (Fig.22).
Enzymatic assays of the purified proteins all showed the predicted activities. Both cellobiohydrolase and Avicelase activities were detected with protein CGSCsYakCAS 04820, but not endoglucanase activity. Feruloyl esterase activity was detected for CGSCsYakCAS 18248 protein. Surprisingly, CGSCsYakCAS 04825 protein showed not only endoglucanase activity (28.24 U/mg of protein) but also pectinase activity. Enzymatic characteristics of the three proteins listed in Tables Tables55 and and66 indicated that they were all mesophilic proteins.
C. ruminicola H1, a recently described yak rumen anaerobe, has been demonstrated previously to be a cellulolytic bacterium that can utilize a variety of natural celluloses as sole carbon sources, including corn fiber and grasses, as well as cellulose (e.g., Avicel and filter paper), hemicellulose (e.g., xylan and mannan), or pectin, the three principal polysaccharides in forages (11). In the present study the genomic sequence of strain H1 was determined, and annotation of the genome data identified a substantial number of genes encoding candidate polysaccharidases. Independently, an enzymatic activity assay also determined that this strain produced the entire spectrum of enzymes essential for natural (hemi)cellulose degradation, including those that remove the modified side groups and feruloyl esterase which release hemicellulose from lignin. These observations suggest that C. ruminicola H1 plays a key role in the fodder-rich feed digestion in the yak rumen.
Acetyl xylan esterase and phenolic acid (ferulic and p-coumaric acid) esterase work on the side groups of xylan and play important roles in hemicellulose degradation (3). Although acetyl xylan esterases are produced extensively by the ruminal bacteria and fungi, only a very limited number of described rumen bacterial species produce phenolic acid esterases. To date, feruloyl esterase activity associated with ruminal bacteria has only been confirmed for F. succinogenes, B. proteoclasticus, and P. ruminicola (15, 20, 30). Importantly, strain H1 produces ferulic acid esterase and three homologue genes are screened in the genome, making it one of the rare phenolic acid esterase-producing rumen bacterium identified. However, unlike the two catalytic domains of ferulic acid esterases in other rumen bacteria, the three homologue genes of strain H1 contain just one esterase domain, and the enzymatic assay confirm its mono-activity (Table (Table66).
Previous researches with other cellulolytic bacteria have indicated that the (hemi)cellulolytic enzymes are induced by the substrate celluloses, accordingly these enzymes are inhibited by the feasible carbohydrates (21, 26). However, strain H1 grew faster on xylan than on filter paper (data not shown), while it upregulated both endoglucanase and xylanase activities (2.0- and 3.5-fold, respectively) in the presence of xylan. As far as we know, this is the first example of a hemicellulose inducing endoglucanase production, as previous studies had shown that cellulolytic ability was inhibited by the presence of hemicellulose (22, 37).
Unlike some of the fibrolytic clostridia (e.g., Clostridium thermocellum and C. cellulyticum) (16) isolated from nonrumen niches that produce their multiple cellulases as a protein complex (cellulosome), strain H1 produced individual polysaccharide hydrolases essential for natural cellulose degradation. Most of the hydrolases possess CBMs that should provide the individual enzymes with the capability to adhere to structured cellulose, thereby enhancing degradation efficiency. Although an ORF (CGSCsYakCAS 16184, see Table S2 in the supplemental material) was predicted as a cellulosomal scaffoldin precursor, the noncatalytic protein in cellulosome for anchoring a variety of catalytic proteins by its multiple adhesive domain (cohesin) and several dockerin-borne proteins were identified in strain H1 genome, the characteristics of ORF 16184 encoding a short peptide (414 amino acids) and only one cohesin do not indicate it will act as a “scaffoldin.” In addition, we failed to isolate a fibrolytic protein complex from this strain (data not shown).
The binding ability and specificity of CBMs provide the carrier enzymes the ability to attach to the fibrous substrates. Among the Avicel-bound (hemi)cellulases of strain H1, Cel5A possessed a CBM2 domain, while Cel9A and Cel9B carried a CBM3 domain. Two of the three endo-1,4-β-xylanases (Xyn10A and Xyn30A) and the ferulic acid esterase Fae1 possessed a CBM2 domain, whereas endo-1,4-β-xylanases bore a CBM6 domain (Xyn43A) and displayed a different activity level. It is believed that CBM families 2 and 3 are able to bind crystalline cellulose, whereas family 6 CBM interacts with individual glycan chains rather than crystalline surfaces (8), which always occur in hemicellulases (http://www.cazy.org). Most of the Avicel-bound xylanases of strain H1 bore the cellulose-binding CBMs rather than the xylan-binding CBM, suggesting that the cellulose-binding ability is also essential for xylanases in targeting their natural substrates inside a complex mixture containing lignocelluloses.
Among the bound fibrolytic enzymatic proteins, the only cellobiohydrolase, Cel48A (CGSCsYakCAS 04820), together with a protein annotated as Cel9A (CGSCsYakCAS 04825) are the most abundant proteins. This finding is in agreement with the fact that cellulases of the GH families 48 and 9 play important roles in the cellulase enzyme systems of most fibrolytic bacteria (5, 41, 46). Therefore, the cellulases encoded by CGSCsYakCAS 04820 and CGSCsYakCAS 04825 genes can function in the cellulolytic processes of strain H1. ORF 04820 and ORF 04825 link as a cluster in strain H1 genome separated by a 119-bp fragment, whereas both a TATA box and the Shine-Delgarno (SD) motif exist in the upstream ORF 04820; only the SD motif is seen in the downstream ORF 04825, and this implies that both genes would be cotranscribed for maintaining the balanced level of the two key digesting enzymes. No regulative gene adjacent was observed. While CGSCsYakCAS 04825 gene product (GH 9 endoglucanase) bore a CBM 3, no known CBMs were observed in the only identified cellobiohydrolase, Cel48A (CGSCsYakCAS 04820), of this anaerobic rumen bacterium, in spite of the fact that it exhibits the highest binding ratio to crystalline cellulose. A novel domain without a previously identified function was identified in the C terminus of Cel48A to be responsible for cellulose binding (data not shown).
There were an additional four Avicel-bound proteins (CBP1 to CBP4, Table Table4)4) that could not be annotated as any known (hemi)cellulolytic enzymes, the functions of which remain to be determined. They may be involved in the adhesion of the cells to cellulose. It was also found that higher ratio (21%) of the noncarbohydrate catalytic proteins possessing a dockerin in the genome of Ruminococcus flavefaciens (32), while the leucine-rich repeats (LRR) modules exist in these ORFs, implying a role for LRR in mediating protein-protein interaction. All Avicel-bound glycanases of strain H1 possess a CBM, except for Cel48A and Cel5B; therefore, we hypothesize that this strain binds the individual enzymatic protein to their substrates instead of producing a cellulosome produced by some anaerobic bacteria. This arrangement could provide the bacterium with a more flexible means of regulating the expression of different enzymatic proteins.
This study was supported by CAS grant KSCX1-YW-11B1, the National High Technology Program (#863) of China (grant 2007AA021301), and the Allegheny General Hospital/Allegheny-Singer Research Institute.
Published ahead of print on 16 April 2010.
†Supplemental material for this article may be found at http://aem.asm.org/.