Adherence of the Gram-Positive Bacterium Ruminococcus albus to Cellulose and Identification of a Novel Form of Cellulose-Binding Protein Which Belongs to the Pil Family of Proteins

Abstract

The adherence of Ruminococcus albus 8 to crystalline cellulose was studied, and an affinity-based assay was also used to identify candidate cellulose-binding protein(s). Bacterial adherence in cellulose-binding assays was significantly increased by the inclusion of either ruminal fluid or micromolar concentrations of both phenylacetic and phenylpropionic acids in the growth medium, and the addition of carboxymethylcellulose (CMC) to assays decreased the adherence of the bacterium to cellulose. A cellulose-binding protein with an estimated molecular mass following sodium dodecyl sulfate-polyacrylamide gel electrophoresis of ∼21 kDa, designated CbpC, was present in both cellobiose- and cellulose-grown cultures, and the relative abundance of this protein increased in response to growth on cellulose. Addition of 0.1% (wt/vol) CMC to the binding assays had an inhibitory effect on CbpC binding to cellulose, consistent with the notion that CbpC plays a role in bacterial attachment to cellulose. The nucleotide sequence of the cbpC gene was determined by a combination of reverse genetics and genomic walking procedures. The cbpC gene encodes a protein of 169 amino acids with a calculated molecular mass of 17,655 Da. The amino-terminal third of the CbpC protein possesses the motif characteristic of the Pil family of proteins, which are most commonly involved with the formation of type 4 fimbriae and other surface-associated protein complexes in gram-negative, pathogenic bacteria. The remainder of the predicted CbpC sequence was found to have significant identity with 72- and 75-amino-acid motifs tandemly repeated in the 190-kDa surface antigen protein of Rickettsia spp., as well as one of the major capsid glycoproteins of the Chlorella virus PBCV-1. Northern blot analysis showed that phenylpropionic acid and ruminal fluid increase cbpC mRNA abundance in cellobiose-grown cells. These results suggest that CbpC is a novel cellulose-binding protein that may be involved in adherence of R. albus to substrate and extends understanding of the distribution of the Pil family of proteins in gram-positive bacteria.

The cellulosome paradigm, developed largely from the study of Clostridium spp., is the most firmly established example of a stable, multienzyme complex specialized in the adherence to and degradation of crystalline cellulose (8, 12). Although high-molecular-mass cellulase complexes have also been identified in a variety of other anaerobic bacteria, including Bacteroides cellulosolvens, Fibrobacter succinogenes, and Ruminococcus albus (7, 18), it is still unclear to what extent the Clostridium spp. paradigm can be applied to these other bacteria (7). With specific reference to R. albus, micromolar concentrations of phenylacetic acid (PAA) and phenylpropionic acid (PPA) stimulate cellulase enzyme production (35) and cellulose digestion kinetics (26). The cell morphology of R. albus is also altered in response to micromolar concentrations of PAA and PPA, both vesicular and fimbrial structures are produced, and cellulases remain associated with the bacterial capsule (34, 35). Recent analysis of the endA gene from Ruminococcus flavefaciens revealed a distant relationship between regions of an 80-amino-acid sequence in EndA and the duplicated 23-amino-acid dockerin sequences found in Clostridium thermocellum cellulosomal enzymes (17). Even though such findings are consistent with the hypothesis that cellulosome-like structures are produced by Ruminococcus spp., the endoglucanase genes cloned to date lack other features characteristic of cellulosomal enzymes. In particular, no consensus cellulose-binding domains appear to have been identified and characterized (15, 36), although such domains are common in both cellulases and xylanases (14, 24).

Therefore, the mechanisms underpinning the adherence of R. albus cells to substrate, as well as cellulase secretion and anchoring to the cell surface (and cellulose), have yet to be elucidated. We report here the identification of two low-molecular-mass cellulose-binding polypeptides from R. albus 8, one of which (CbpC) possesses structural motifs typical of the Pil protein family, which is comprised largely of type 4 fimbrial proteins produced by gram-negative pathogenic bacteria.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

R. albus 8 was provided by B. A. White, University of Illinois, Urbana-Champaign; R. albus SY3 was provided by J. Miron, Volcani Research Institute, Bet Dagan, Israel. R. albus type strain 7 and the noncellulolytic strain B199, as well as R. flavefaciens FD-1 and type strain C-94, were obtained from M. A. Cotta, National Center for Agricultural Utilization Research, U.S. Department of Agriculture, Peoria, Ill. The cellobiose and cellulose (Sigmacell-19) used in growth media were obtained from Sigma Chemical Co., St. Louis, Mo. All bacterial strains were routinely cultured in EM-cellobiose medium (10). When necessary, strains were grown in a defined minimal medium (34) containing either 0.2% (wt/vol) cellobiose or cellulose and transferred at least three times prior to analysis. For cellulose-binding assays, the bacterium was cultivated in cellobiose-containing minimal medium, supplemented with either 5% (vol/vol) clarified ruminal fluid, 25 μM each PPA and PAA (Sigma), or no additions.

Adherence assay.

The methods used were similar to those described by Bayer et al. (6). Microcrystalline cellulose (Avicel type PH-101, lot no. 1647; FMC Corporation, Philadelphia, Pa.) was used as the substrate in all assays.

Cell fractionation procedures.

Cultures (500 ml) of R. albus 8 were grown in either EM-cellobiose or EM-cellulose medium; following overnight (EM-cellobiose) or 72-h (EM-cellulose) growth, the cells were harvested by low-speed centrifugation (10,000 × g, 10 min, 4°C) and washed twice with 50 mM phosphate buffer (pH 6.5)–200 mM NaCl (phosphate-buffered saline [PBS]). The washed cells were then resuspended in PBS prepared to contain 1 mM phenylmethylsulfonyl fluoride and broken by two passages through a French pressure cell (SLM Aminco Instruments, Urbana-Champaign, Ill.) set at 250,000 kPa. Unbroken cells and large debris were removed by low-speed centrifugation, and the membrane fragments were recovered by ultracentrifugation (250,000 × g, 60 min, 4°C). The supernatant was assumed to contain only cytoplasmic proteins and was stored frozen at −20°C until further analysis. The pellet was resuspended in PBS containing 0.02% (vol/vol) Triton X-100, incubated at 37°C for 15 min, and then subjected to low-speed centrifugation. The supernatant fraction was retained and assumed to contain solubilized membrane proteins. The sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) protein profile of the solubilized membrane proteins was indistinguishable from that of crude membrane fragments (data not shown), suggesting the solubilization procedure provided a representative sample of the membrane proteins produced by this bacterium.

Cell-free culture fluids were concentrated by ultrafiltration in an Amicon TCF-10 manifold fitted with a YM10 membrane (10,000-molecular-weight cutoff). The retentate was washed and resuspended in PBS containing 1 mM phenylmethylsulfonyl fluoride and then stored at −20°C until further analysis.

Identification of cellulose-binding proteins by an affinity-based assay.

Aliquots of the solubilized membrane proteins (∼2 mg) or proteins concentrated from cell-free culture fluids (100 μg) were mixed with 100 mg of Sigmacell-19 cellulose, suspended in a solution of PBS prepared to contain 4 mM CaCl₂ and 2 mM dithiothreitol, and adjusted to a final volume of 1 ml with the same buffer. This mixture was left at room temperature for 30 min with occasional agitation, and the cellulose particles were then harvested by low-speed centrifugation. The resulting pellet was subsequently washed with either sterilized water, PBS, various detergent solutions, or a 10% (wt/vol) filter-sterilized solution of cellobiose. Aliquots of each wash fraction and the cellulose particles were mixed with SDS-PAGE running buffer, boiled, and subjected to SDS-PAGE according to standard procedures. The stacking and resolving gels consisted of 4 and 10% (wt/vol) T, respectively, and proteins were visualized by silver staining.

CbpC cellulose binding in the presence or absence of CMC.

Aliquots (25 μg) of the CbpC protein concentrated from culture fluids were mixed in assay cocktails as described above, but with the amount of cellulose present ranging from 1.6 mg to 100 mg. To assess the effects of carboxymethylcellulose (CMC) on CbpC binding to cellulose, a solution of CMC prepared in PBS was added to some of the mixtures prior to the addition cellulose, to give a final concentration of 0.1% (wt/vol). The cellulose particles were recovered by centrifugation, washed once with PBS, and then centrifuged again. The cellulose particles with bound proteins were resuspended in 50 μl of SDS-PAGE running buffer and boiled, and the eluted proteins were subjected to SDS-PAGE.

Protein sequencing and DNA techniques.

The solubilized membrane proteins and the proteins in concentrated cell-free culture fluids were subjected to SDS-PAGE and electroblotted onto a polyvinylidene difluoride membrane by using a Mini Trans-Blot system (Bio-Rad Laboratories). The membrane was stained for 2 min in a solution of methanol-water (40:60) containing 0.005% (wt/vol) bromophenol blue and destained with a solution of methanol-water (50:50). The CbpC protein was cut from the membranes with a sterile razor blade. The intact CbpC protein as well as peptide fragments generated by trypsin digestion was subjected to Edman degradation, and the primary amino acid sequence was determined with a Procise-HT Microsequencer, made available through the University of Nebraska Protein Core Facility.

Standard recombinant DNA procedures were used (4, 33), and enzymes were obtained from either Promega or Gibco BRL. Chromosomal DNA from all R. albus strains was isolated according to standard procedures (4), following lysis of the cells in 10 ml of 50 mM phosphate buffer (pH 6.0) containing mutanolysin (200 U/ml), proteinase K (150 μg/ml), and 0.5% (wt/vol) SDS for 1 h at 55°C. The amino acid sequences obtained from CbpC amino terminus, and tryptic fragments, were back-translated to generate oligonucleotide primers suitable for PCR amplification of the cbpC gene. The PCR mixture (50 μl) contained 20 ng of R. albus 8 genomic DNA, 1 μM each primer A (5′-ATG ATG GGN TAY GTN AAG AA-3′; complementary to the antisense strand of the sequence 24-MMGYVKK-30 of the mature CbpC protein) and primer B (5′-GCY TCN GGR TAY TGN CCN AC-3′; complementary to the nucleotide sequence encoding amino acids 143-VGQYPEA-149 [see Fig. 5]), 200 μM deoxynucleoside triphosphates, 1.5 mM MgCl₂, 50 mM KCl, 10 mM Tris-HCl (pH 9.0), 0.1% (wt/vol) Triton X-100, and 2.5 U of Taq polymerase (Promega). The thermal cycling parameters were 30 cycles for 1 min at 95°C, 1 min at 40°C, and 1 min at 72°C, followed by 1 cycle for 5 min at 72°C. The resulting 350-bp PCR product was also used to prepare digoxigenin-labeled probe (Genius System; Boehringer Mannheim) in Southern blots of genomic DNA extracted from the R. albus strains listed above. Briefly, BglII-digested DNA was transferred to a charged-nylon membrane (Zeta-Probe GT; Bio-Rad Laboratories, Hercules, Calif.) by vacuum blotting and immobilized by UV cross-linking. The membrane was hybridized with the probe overnight at 65°C, washed twice at 43°C for 30 min in 40 mM disodium phosphate buffer (pH 7.0) containing 1 mM EDTA and 5% (wt/vol) SDS, and then washed once at 43°C for 30 min in 40 mM disodium phosphate buffer (pH 7.0) containing 1 mM EDTA and 1% (wt/vol) SDS. The hybridizations were visualized according to the manufacturer’s specifications.

FIG. 5.

Nucleotide sequence of the cbpC gene from R. albus 8. The predicted amino acid sequence is shown in single-letter code above the coding sequence; putative −10, −35, and ribosome-binding sites are underlined; the stop codon is indicated by an asterisk. Sequences also determined by Edman degradation of the purified CbpC protein are double underlined, and the presumed cleavage site to remove the leader sequence is denoted by the inverted triangle.

The nucleotide sequences of the 5′ and 3′ ends of the cbpC gene were amplified and isolated by genomic walking procedures (Universal Genome Walker kit; Clontech). The nested primary and secondary cbpC gene-specific primers used for genomic walking procedures were based on the 350-bp sequence obtained from the PCR product described above and were designed according to the manufacturer’s recommendations. Primary PCRs were carried out in 25-μl volumes containing approximately 10 ng of DraI-digested and adapter-ligated R. albus 8 genomic DNA, 200 μM each adapter primer (AP1) and either a forward or reverse gene-specific primer, 200 μM deoxynucleoside triphosphates, 1.1 mM magnesium acetate, 15 mM potassium acetate, 40 mM Tris-HCl (pH 9.3), and 1 U of rTth DNA polymerase XL (Perkin-Elmer). The thermal cycling parameters used were 7 cycles for 15 s at 94°C and 3 min at 72°C, 37 cycles for 15 s at 94°C and 3 min at 67°C, and then 1 cycle for 4 min at 67°C. Secondary PCRs were carried out in 50-μl volumes containing 1 μl of a 1:50 dilution of the appropriate primary PCR mixture, adapter primer AP2, and either a forward or reverse nested gene-specific primer. The other reaction components and thermal cycling parameters were the same as those used for primary PCR. The genomic walking was performed on two separate occasions; independent PCR products were sequenced by the DNA sequencing facility at Iowa State University and analyzed with the Wisconsin Package (Genetics Computer Group, Madison, Wis.).

Northern blot analysis.

Total RNA was isolated from R. albus 8 cultures grown in cellobiose minimal medium containing either 5% (vol/vol) clarified ruminal fluid, 25 μM each PAA and PPA, or no additions. Cells were harvested in mid-logarithmic phase of growth (optical density at 600 nm of ∼0.5) by anaerobic centrifugation, and resuspended in 5 ml of 25 mM sodium phosphate buffer (pH 7.0) containing 25 mM EDTA. Cells were lysed by the addition of mutanolysin (20 U/ml), proteinase K (100 μg/ml), and 0.5% (wt/vol) SDS and incubation at 55°C for 5 min. Total RNA was extracted from the cell lysate by the RNeasy purification system (Qiagen, Valencia, Calif.) according to the manufacturer’s recommendations. The RNA was concentrated by ethanol precipitation and stored at −80°C. For Northern blots, total RNA was fractionated in 1% (wt/vol) agarose gels in the presence of both glyoxal and dimethyl sulfoxide according to standard procedures (5, 33). The RNA was transferred to a Zeta-Probe GT membrane by vacuum blotting and immobilized by UV cross-linking. The cbpC-specific radiolabeled probe was prepared by random primer labeling (Boehringer Mannheim). Hybridization was performed overnight at 43°C, and the blot was washed as described above for Southern blots. To determine if there were differences in RNA loading, blots were stripped and reprobed with a ³²P-end-labeled 16-mer oligonucleotide complementary to virtually all known 16S rRNA sequences (5′-TAC CGC GGC TGC TGG CAC-3′).

Nucleotide sequence accession number.

The nucleotide sequence described in this article will be available in the GenBank/EMBL databases under accession no. AF089753.

RESULTS

Adherence of R. albus 8 to cellulose.

Although neither growth rate nor cell yield was affected by the addition of either ruminal fluid or PAA-PPA to cellobiose growth medium, the adherence of R. albus to cellulose was significantly increased by these additions. The adherence values (percentage) obtained for cells harvested during mid-log phase of growth were 65.6 ± 2.54, 73.0 ± 3.37, and 79.5 ± 3.74 for cells cultured in the presence of no supplements, PAA-PPA and ruminal fluid, respectively. When cells were harvested during early stationary phase of growth (optical density at 600 nm of 0.8 to 1.0), the relative number of adherent cells decreased but the effects of ruminal fluid or a combination of PAA-PPA were maintained (data not shown). Therefore, all subsequent assays were performed with cells cultivated in EM-cellobiose medium and harvested during mid-log phase of growth. Similar to earlier studies with another strain of R. albus (25), inclusion of CMC in the adherence assay mixtures decreased, but did not eliminate, the binding of cells to cellulose. The effect of CMC was concentration dependent, up to 0.1% (wt/vol), where adherence values had decreased to approximately 30% (data not shown). Taken together, the results of these experiments suggest the adherence of R. albus cells to cellulose is positively impacted by components present in ruminal fluid such as PPA and PAA and that the cellulose analog CMC effectively inhibits the adherence process.

Identification of two low-molecular-mass, cellulose-binding proteins from R. albus 8 cell fractions.

When membrane proteins extracted from cellulose-grown cells were assayed, two polypeptides of ∼21 kDa (CbpC) and ∼16 kDa (CbpD) appeared to bind completely to cellulose. Elution of both CbpC and CbpD from cellulose particles was achieved with a solution of 0.1% (vol/vol) Triton X-100, although 10% (wt/vol) cellobiose was also capable of removing lesser amounts (Fig. 1). Subsequent experiments showed that the CbpC protein could also be removed by boiling the cellulose in SDS-PAGE loading buffer. Therefore, the limited amount of CbpC protein present following this step in Fig. 1 (lane 8) indicates that virtually all of the CbpC protein was removed from cellulose by Triton X-100 and cellobiose. Further examination of the CbpC protein showed that unlike the low-affinity cellulose-binding domains of some clostridial proteins (reference 24, for example), sterile water did not result in CbpC elution from cellulose (data not shown). The results obtained with membrane proteins extracted from cellobiose-grown cells were somewhat different from those shown in Fig. 1. The relative amount of CbpC appeared to be less, and CbpD was not detectable. When the membrane, cytosolic, and cell-free supernatant fractions of cellulose- and cellobiose-grown cultures of R. albus 8 were compared by SDS-PAGE analysis, a greater abundance of CbpC was present in the membrane fractions of cellulose-grown cells, and a polypeptide of similar molecular mass was also clearly evident in the cell-free supernatant of the same cultures (data not shown). The cell-free supernatant of cellulose-grown cultures was subsequently concentrated by ultrafiltration and used in cellulose-binding assays, and the results were virtually identical to those observed for CbpC (Fig. 2). The protein bound tightly to cellulose in functional assays and was largely removed by washing the particles with 0.1% (vol/vol) Triton X-100; lesser amounts eluted with 10% (wt/vol) cellobiose. Protein sequence analysis ultimately confirmed that the cellulose-binding proteins in both membrane and culture fluid fractions were identical (see below), and all subsequent assays were performed with the CbpC protein obtained from culture fluids.

FIG. 1.

SDS-PAGE analysis of cellulose affinity (binding) assays using membrane proteins isolated from cellulose-grown cultures of R. albus 8. Lanes: 1, protein molecular mass standards; 2, aliquot of solubilized membrane proteins; 3, proteins unbound after incubation with cellulose; 4 and 5, aliquots of separate washes of the cellulose with 10 mM phosphate buffer; 6 and 7, proteins removed from cellulose by washing with 0.1% (vol/vol) Triton X-100 and 10% (wt/vol) cellobiose, respectively; 8, aliquot of SDS-PAGE running buffer mixed with the cellulose and boiled. CbpC and another low-molecular-mass cellulose-binding polypeptide are marked with arrows. Sizes are indicated in kilodaltons.

FIG. 2.

SDS-PAGE analysis of cellulose affinity assay using proteins concentrated from the cell-free supernatant of cellulose-grown cultures of R. albus 8. Lanes: 1, protein molecular mass standards; 2, aliquot of concentrated extracellular proteins; 3, proteins unbound following incubation with cellulose; 4 through 7, aliquots of four separate washes of the cellulose particles with 10 mM phosphate buffer (pH 6.5); 8 and 9, proteins removed from cellulose following treatment of the cellulose particles with 0.1% (vol/vol) Triton X-100 and 10% (wt/vol) cellobiose, respectively; 10, proteins removed from cellulose particles after resuspension and boiling in SDS-PAGE running buffer. Sizes are indicated in kilodaltons.

CbpC binding to cellulose is impaired in the presence of CMC.

In accordance with the observation that CMC blocked the adherence of R. albus whole cells to cellulose, we chose to examine whether CMC exerted a similar effect on the binding of the CbpC protein to cellulose. In the experiments shown in Fig. 3, the concentrations of inhibitor (CMC, 0.1% [wt/vol]) and protein (CbpC) remained constant, but the amount of substrate (cellulose) was varied. In assay mixtures containing 25 to 100 mg of cellulose, all of the CbpC protein was bound to cellulose, irrespective of the presence or absence of CMC (data not shown). However, with lesser amounts of cellulose (from 1.25 to 12.5 mg), the amount of CbpC bound was dependent on the amount of cellulose present, and in the presence of CMC, CbpC binding was further reduced, as is clearly illustrated in Fig. 3. This is analogous to CMC exerting its effect on the K_m (of CbpC for cellulose) but not the V_max (of CbpC bound to cellulose), reflective of CMC serving as a competitive inhibitor of CbpC binding to cellulose; the positional interactions of CMC and cellulose with the CbpC protein are similar.

FIG. 3.

Analysis of CbpC binding to increasing amounts of cellulose, in the presence or absence of 0.1% (wt/vol) CMC. (A) Relative amount of the CbpC protein preparation bound to either 12.5 mg (lane 1), 6.25 mg (lane 2), 3.13 mg (lane 3), or 1.56 mg (lane 4) of cellulose particles as described in Material and Methods; (B) results obtained when CMC was added prior to cellulose in the assay buffer. There were no visible differences in CbpC binding to cellulose in the presence or absence of CMC when the amount of cellulose in the assay was increased to 25 mg or above.

Nucleotide and amino acid sequence analysis of CbpC.

The 33-residue amino-terminal sequence of CbpC was XTLVELVVVIAIIGVLAAILVPSMMGYVKKARL, where X represents an unidentified amino acid residue. When primers A and B were used together in PCRs, a 350-bp amplification product was routinely produced. Southern blot analyses of BglII-digested R. albus 8 genomic DNA using the 350-bp PCR product as a probe showed that two fragments (∼1.8 and ∼6 kb) reacted strongly with the probe (Fig. 4). In addition, the PCR product cross-hybridized with genomic DNA prepared from the other R. albus strains (Fig. 4), indicating that there are regions of homology to the cbpC gene present in these other strains. Our initial attempts at cloning and propagating the DNA fragments from R. albus 8 proved unsuccessful, and the entire cbpC gene sequence was ultimately obtained by a combination of genomic walking procedures and PCR (Fig. 5). Using these methods, we identified an open reading frame (ORF) which would encode a protein of 169 amino acids with a calculated molecular mass of 17,655 Da. Although the molecular mass is slightly smaller than that of CbpC estimated from SDS-PAGE, amino acid residues 10 through 43 and 143 through 149 in the predicted amino acid sequence were identical to the peptide sequences obtained from purified CbpC. We therefore conclude that the DNA sequence obtained by PCR and genomic walking procedures contains the entire cbpC gene. Nucleotide sequence analysis also identified a BglII site within the cbpC gene, explaining why the PCR product hybridized strongly to two fragments of BglII-digested genomic DNA from R. albus 8.

FIG. 4.

Southern blot analysis of genomic DNA isolated from R. albus 8 (lane 1), 7 (lane 2), B199 (lane 3), and SY3 (lane 4), using the 350-bp PCR product of the cbpC gene as a probe. Genomic DNA from all strains was digested with the restriction enzyme BglII, and the migration distances of standard DNA fragments are shown on the left.

The ORF encoded by the cbpC gene can be subdivided into no fewer than two distinct domains, based on BLAST analysis of the predicted CbpC amino acid sequence with those currently in the databases. The amino-terminal third of the protein shows significant sequence identity with amino-terminal sequences of type 4 fimbrial precursors from the gram-negative pathogenic bacteria Dichelobacter (Bacteroides) nodosus, Moraxella bovis, Neisseria gonorrhoeae, and Pseudomonas aeruginosa (Fig. 6A). The ORF encodes an eight-amino-acid positively charged leader sequence which terminates with a G residue. This leader sequence is followed by an F residue and the sequence obtained from Edman degradation of the purified CbpC protein. In other bacterial species expressing type 4 fimbrial proteins, the GF dipeptide motif serves as the site of proteolytic cleavage, and the phenylalanine is methylated to become the first residue of the mature type 4 fimbrial protein (11). The canonical glutamate residue at position 5 of the mature protein, found in all type 4 fimbriae and Pil homologs, is also present in CbpC, in addition to the highly conserved hydrophobic sequence typical of this family of proteins. A G residue was also present at position 58 of the mature protein (Fig. 5), which has been suggested by previous analyses of type 4 fimbrial proteins to represent the border between the conserved amino-terminal third and variable carboxy-terminal two-thirds of the protein (11). Beyond this G residue, the predicted sequence was found to possess significant homology (33% amino acid identity) to 72 and 75 amino acid motifs tandemly repeated 13 times within a 190-kDa outer membrane protein of Rickettsia rickettsii (3) (Fig. 6B). A similar motif is also present in one of the large virion glycoproteins (Vp260) from Chlorella virus PBCV-1, where it is also repeated 13 times in tandem (32). Kyte-Doolittle plots of the CbpC protein and the type 4 fimbriae of the bacteria described above are shown in Fig. 7. For the sake of comparison, the ComG protein of Bacillus subtilis, which possesses the Pil-like amino-terminal domain but is a membrane-bound DNA-binding protein (1), is also included. Despite the lack of primary sequence identity among these proteins, CbpC and the type 4 fimbrial proteins all possess a reasonable degree of similarity in terms of the spatial distribution and length of hydrophobic and hydrophilic stretches of amino acids. Notably, CbpC and the fimbrial proteins all possess a fairly long stretch of relatively hydrophilic residues at the carboxy terminus, while the ComG protein does not.

FIG. 6.

(A) Sequence alignments of the CbpC amino terminus (Ralb) with the amino-terminal sequences from the type 4 fimbrial proteins of M. bovis (Mbov), D. nodosus (Dnod), N. gonorrhoeae (Ngon), and P. aeruginosa (Paer); (B) alignment of amino acid residues 43 through 128 of CbpC with the 72- and 75-amino-acid motifs present in the surface protein antigen of R. rickettsii (Rric72 and Rric75). Identical amino acid residues are highlighted by a dark background; similar amino acid residues are indicated by the shaded background.

FIG. 7.

Kyte-Doolittle plots of the deduced amino acid sequences for R. albus cellulose-binding protein CbpC and type 4 fimbrial proteins from M. bovis, D. nodosus, N. gonorrhoeae, and P. aeruginosa. For the sake of comparison, the Kyte-Doolittle plot of ComG protein from B. subtilis, a nonfimbrial Pil-like protein, is included.

cbpC transcript abundance is increased by the inclusion of either ruminal fluid or PAA-PPA in the growth medium.

Northern blot analysis showed that the cbpC transcript is approximately 700 nucleotides in length, consistent with the predicted size of the gene and the transcript being monocistronic. Additionally, cbpC transcript abundance was increased in cells cultivated in cellobiose minimal medium supplemented with either ruminal fluid or PPA-PAA. These findings could not be explained by differences in total RNA loading, because the 16S rRNA-specific probe hybridized with all preparations with similar intensities (Fig. 8). The increase in transcript abundance correlates well with the increases seen in bacterial adherence which were also observed following growth of R. albus 8 in medium containing either ruminal fluid or PPA-PAA.

FIG. 8.

Northern blot analysis of cbpC transcript abundance following growth in various media. Total RNA was harvested from cells following growth in cellobiose medium containing either ruminal fluid (lane 1), no additions (lane 2), or both PAA and PPA. (A) Results obtained with the cbpC-gene specific probe; (B) the same membrane probed with an oligonucleotide complementary to 16S rRNA.

DISCUSSION

Previous microscopic examination of the ultrastructure and adhesion of R. albus to cellulose revealed that the cells produce a compact mat of fibers surrounding the bacterial cell wall. These fibrous elements also were found to project outward from the surface by as much as 600 nm, and it was concluded that this material mediated the attachment of the bacterium to cellulose (30). A more quantitiative examination of the adherence process in selected ruminal bacteria later revealed that the adherence of R. albus to cellulose could be inhibited by CMC (25). However, despite the identification of a role for PPA and/or PAA in stimulating capsule production and cellulose digestion kinetics by R. albus (26, 34, 35), no mechanistic information about the adherence process has been obtained to date. The findings reported here provide some of the first molecular details to explain the observations made in these earlier studies. First, the inclusion of either PAA-PPA or ruminal fluid in the growth medium improved the adherence of R. albus 8 to cellulose, and these additions were also found to increase cbpC transcript abundance. Second, the inhibition of both R. albus whole cells and CbpC protein adherence to cellulose in the presence of CMC further implicates the CbpC protein in the adherence process. Third, Southern blot analysis confirmed that a number of R. albus strains possess genes with a high degree of sequence homology to the cbpC gene. Preliminary Western immunoblot analysis of these other strains with anti-CbpC antibodies identified in R. albus SY3 a protein of ∼25 kDa that is also a cellulose-binding protein (27). Therefore, it seems reasonable to suggest that the CbpC protein represents a novel strategy for the adherence of gram-positive bacteria to cellulose; the relative contribution of the CbpC protein to the adherence process in R. albus 8 and other R. albus strains requires further examination.

A conserved protein sequence implies that the domain or motif is involved with conserved structural-functional roles. The amino-terminal sequence of CbpC is homologous to similarly located motifs present in a variety of proteins from gram-negative bacteria, all of which are involved in the assembly of protein complexes at the cell surface. Many of these proteins are also implicated in pathogenesis, especially host cell colonization and degradation (16). However, the identification and characterization of Pil proteins and homologs in gram-positive bacteria appear to be limited, having previously been shown to exist in B. subtilis (1, 9), Streptococcus gordonii (19), and Clostridium perfringens (16). The Pil-like proteins in B. subtilis and S. gordonii are involved in competence factor-dependent DNA transformation, rather than attachment to surfaces and(or) protein secretion, whereas the function of Pil homologs in C. perfringens is unclear. The results from Southern and Western blot analyses suggest that homologs of the CbpC protein appear to exist in other strains of R. albus and that this family of proteins may be more widespread throughout gram-positive bacteria than currently recognized.

The tandemly repeated motifs within the 190-kDa surface protein of Rickettsia spp. are thought to have a role in rickettsia-host interaction (2, 20, 23), possibly through host cell recognition and attachment in a manner similar to that seen with repetitive peptides present in parasitic eukaryotes such as Plasmodium falciparum (22). Given the similarities in Kyte-Doolittle plots for CbpC and the type 4 fimbriae shown in Fig. 7, it is tempting to speculate that CbpC protein is involved with (i) the formation of a fimbria-like structure which possesses cellulose-binding properties and (ii) plant surface colonization. Indeed, R. albus 8 was previously shown to produce fimbria-like structures at the polar ends of the cell, especially when the bacterium was cultivated in the presence of PAA and PPA (35). In other bacteria known to produce type 4 fimbriae, a proportion of these fimbriae are also shed into the growth medium and remain there after the cells are removed by centrifugation (21, 28). This would explain the presence of CbpC with both membrane and cell-free supernatant fractions of cellulose-grown cultures. Type 4 fimbrial proteins are usually of relatively low molecular mass (20 to 25 kDa), and it was originally hypothesized that the hydrophobic amino-terminal motif formed the core of the fimbrial strand, facilitating a helical stacking of the subunits (13, 21). However, beyond this hydrophobic stretch of ∼50 amino acids, the proteins show limited if any sequence similarity (11, 16), although the variable regions do possess similarities in terms of protein chemistry. The carboxy-terminal two-thirds of many of the type 4 fimbrial subunit proteins consist of a series of short, fairly hydrophobic domains which have the potential to form β-sheet structures, around which hydrophilic sequences determining immunoreactivity are arranged (11). It is this region of the protein that also confers specificity in terms of binding to the cell surface. An O-linked disaccharide has been identified in this region of some type 4 fimbrial proteins (29) and is also thought to contribute to immunoreactivity and surface recognition. Although we have yet to demonstrate that CbpC is glycosylated, the presence of sugar moieties could explain the discrepancy in molecular mass estimated for CbpC from SDS-PAGE and predicted from cbpC sequence analysis.

Recent monomer assembly models following protein crystallography analysis have further resolved the structure of type 4 fimbriae (29). A right-handed helical arrangement of five monomers still results in the amino termini forming the core of the fimbrial strand, giving rise to a waffle cone morphology. These pentameric helical structures upon stacking bury the amino termini, and only the carboxy-terminal two-thirds is left exposed, as a smooth cylinder of high mechanical stability. Interestingly, the secondary and tertiary structures predicted for the repetitive motifs within the 190-kDa R. rickettsii surface protein antigen are similar in nature to those described above for the fimbrial subunit proteins (3). If the assembly of the CbpC protein on the surface of R. albus is similar to that described above for other type 4 fimbrial proteins (29), a complex of CbpC proteins could also resemble the characteristics associated with a larger protein comprised of tandemly repeated domains of the same sequence. However, even though the motifs identified within the CbpC protein appear to be more closely related to type 4 fimbriae, proteins with similar amino-terminal motifs that are not involved with type 4 fimbrial biosynthesis and function have been described for other bacterial species. Of these, the most notable examples include the pullulanase secretion system of Klebsiella oxytoca (31) and cellulase secretion systems of the plant pathogens Xanthomonas campestris and Erwinia carotovora (16, 31). Therefore, future studies dissecting the potential roles of Pil homologs in R. albus adherence, enzyme secretion, and cellulase complex assembly should prove to be a valuable comparative model of this family of proteins in gram-positive bacteria.