Unconventional Mode of Attachment of the Ruminococcus flavefaciens Cellulosome to the Cell Surface

Abstract

Sequence extension of the scaffoldin gene cluster from Ruminococcus flavefaciens revealed a new gene (scaE) that encodes a protein with an N-terminal cohesin domain and a C terminus with a typical gram-positive anchoring signal for sortase-mediated attachment to the bacterial cell wall. The recombinant cohesin of ScaE was recovered after expression in Escherichia coli and was shown to bind to the C-terminal domain of the cellulosomal structural protein ScaB, as well as to three unknown polypeptides derived from native cellulose-bound Ruminococcus flavefaciens protein extracts. The ScaB C terminus includes a cryptic dockerin domain that is unusual in its sequence, and considerably larger than conventional dockerins. The ScaB dockerin binds to ScaE, suggesting that this interaction occurs through a novel cohesin-dockerin pairing. The novel ScaB dockerin was expressed as a xylanase fusion protein, which was shown to bind tenaciously and selectively to a recombinant form of the ScaE cohesin. Thus, ScaE appears to play a role in anchoring the cellulosomal complex to the bacterial cell envelope via its interaction with ScaB. This sortase-mediated mechanism for covalent cell-wall anchoring of the cellulosome in R. flavefaciens differs from those reported thus far for any other cellulosome system.

Cellulolytic microorganisms vary in the organization of their hydrolytic enzymes, from free, soluble enzyme systems to high-molecular-weight protein complexes present on the cell surface (6, 27, 44). Cell-associated cellulosome complexes are known to play a key role in the degradation of plant cell walls by anaerobic bacteria and fungi (3, 5, 7, 11, 14, 16, 21, 35). The organization of polypeptides in the cellulosome occurs via interaction of the conserved dockerin domains present in the catalytic subunits with reiterated cohesin domains of the noncatalytic scaffoldin subunit.

Early biochemical and microscopic evidence on cellulosome structure in anaerobic cellulolytic bacteria revealed that the multicatalytic enzyme complex is tightly associated with the bacterial cell surface (4, 21, 22). The molecular basis for anchoring the cellulosome complex to the bacterial envelope was elucidated in Clostridium thermocellum by the identification of cell surface proteins interacting with typical C-terminal SLH (S-layer homology) repeats that show affinity for components of the peptidoglycan cell wall (17, 26). Thus, anchoring proteins SdbA, Orf2p, and OlpB all exhibit cohesin domains (one, two, and four copies, respectively) that interact specifically with the C-terminal type-II dockerin of the main scaffoldin protein CipA (23-25). In mesophilic clostridia, hydrophilic domains present in the scaffoldin proteins themselves have been proposed to fulfil a similar surface-anchoring role (15).

R. flavefaciens strain 17 produces a cellulosome complex that is known to involve the cohesin-containing structural components ScaA, ScaB, and ScaC, together with interacting enzymes and unidentified proteins that carry dockerin domains (1, 36-38). The assembly of the different components in the R. flavefaciens cellulosome differs from the proposed molecular architecture in the clostridial cellulosomes. In R. flavefaciens the primary (enzyme-incorporating) scaffolding protein ScaA is capable of binding a group of dockerin-containing enzymes to its three resident cohesin repeats (36, 38). In addition, ScaC, a small dockerin-bearing protein that also possesses a single divergent cohesin domain, binds both to ScaA via its dockerin and to a range of thus-far-unidentified polypeptides via its cohesin. ScaC has therefore been proposed to serve as an adaptor protein that enhances the repertoire of subunits present in the cellulosome (37). ScaA binds to any of the seven cohesin repeats of ScaB via a specific cohesin-dockerin interaction (13). Previous biochemical evidence has indicated that all three scaffoldins—ScaC, ScaA, and ScaB—are associated with the cell surface, and it was proposed that the C-terminal domain of ScaB might have a role in attachment (13, 36). This domain, however, shows no significant sequence homology with the SLH or hydrophilic domains previously described in cellulolytic clostridia. One of the major questions that has remained unanswered, therefore, is how the R. flavefaciens cellulosome complex is retained on the bacterial cell surface.

We report here that an additional component, ScaE, encoded by the sca gene cluster, interacts with the ScaB C-terminal domain via a novel cohesin-dockerin interaction. The C terminus of ScaB includes a cryptic dockerin domain, which contains several inserts that conceal its resemblance to conventional dockerin sequences. ScaE includes a cohesin domain at its N terminus, and the sequence at its C terminus contains an LPXTG-like motif, which suggests that it is positioned covalently on the cell surface via proteolytic cleavage and transfer of the peptide to the cell wall in a well documented sortase-mediated attachment mechanism common in gram-positive bacteria (32-33, 40, 43).

MATERIALS AND METHODS

Strains and growth conditions.

R. flavefaciens 17 was grown anaerobically in modified M2 medium (29) or in modified Hungate-Stack medium (18) with the addition of 5% clarified rumen fluid. Modified M2 medium was supplemented with 1% microcrystalline cellulose (Avicel PH105; Honeywill & Stein, London, United Kingdom), 0.4% birchwood xylan, or 0.4% cellobiose as an energy source, as described previously (36). Modified Hungate-Stack medium was supplemented with 1% (wt/vol) crystalline cellulose as an energy source for the preparation of the cellulose-bound protein fraction. Escherichia coli Solopack Gold XL10 and BL21(DE3) (Stratagene, La Jolla, CA) were used as hosts for transformation and protein expression, respectively, with constructs made in pET30Ek/LIC (Novagen, Madison, WI). E. coli strains were routinely grown on Luria-Bertani (LB) medium (Oxoid, Basingstoke, Hampshire, United Kingdom) with appropriate antibiotic selection. All reagents were purchased from Sigma Chemical Co. (St. Louis, MO) unless otherwise specified.

Sequence of scaE.

The scaE coding sequence was obtained from PCR extension of the scaB gene by a combination of genome walking and PCR extension. The genome-walking procedure was performed using GenomeWalker Kit (Clontech/BD Bioscience, Oxford, United Kingdom) according to the manufacturer's guidelines. In addition to the PCR products obtained with the GenomeWalker kit, PCR extension of some fragments of DNA was performed by amplification from a pUC13 plasmid library constructed previously (36) or from a Lambda phage library constructed by ligating DNA partially digested with EcoRI* into EcoRI-cut phage arms (λZAPII; Stratagene). Amplification was achieved with a combination of M13 primers (specific for the vector) and specific primers designed to extend the DNA sequence downstream of the scaB gene.

Cloning and expression of His₆-tagged protein constructs.

The putative cohesin of ScaE (ScaE-Coh), after removal of the signal peptide (amino acids [aa] 32 to 211; AAGQAYD → QVESTTS), was cloned into the pET30Ek/LIC expression vector (Novagen) after PCR amplification of the DNA fragment with primers designed to contain 14 or 15 bp of sequence specific for the ligation-independent cloning site of the vector (forward sequence, 5′-GACGACGACAAGATGGCTGCT-GGTCAGGCTTATG; reverse, 5′-GAGGAGAAGCCCGGT-TTA-AGATGTAGTACTCTCAACCTG; underlined sequences are specific for the LIC site in the vector, and stop codons added to the antisense primer are indicated in italics). After amplification, the DNA product was purified with a QIAquick PCR purification kit (QIAGEN, Hilden, Germany), used according to the manufacturer's instructions. The DNA fragment was then treated with 5 U of T4 DNA polymerase (Roche Diagnostics GmbH, Mannheim, Germany) in a 70-μl (total volume) mixture containing 2.5 mM dATP, 1 mM dithiothreitol, 50 mM Tris-HCl (pH 8.8), 15 mM (NH₄)SO₄, 7 mM MgCl₂, 0.1 mM EDTA, 10 mM 2-mercaptoethanol, and 20 μg of bovine serum albumin/ml. The reaction mixture was incubated at 37°C for 1 h and heat inactivated at 70°C for 15 min. The T4 DNA polymerase-treated DNA fragment was then purified with a QIAquick PCR purification kit (QIAGEN) and concentrated by using DNA concentrator at 3,600 rpm and 45°C under vacuum to approximately 10 μl of total volume. Ligation was carried out by using the pET30Ek/LIC-linearized T4 DNA polymerase-treated vector as recommended by the manufacturer with only one modification: 1 U of T4 DNA Ligase (Roche) was added to the ligation mix, which enhanced the cloning efficiency. The resulting recombinant plasmid (pETScaE-Coh) was transformed in E. coli SoloPack Gold XL-10 chemically competent cells (Stratagene), plated on LB agar supplemented with 30 μg of kanamycin.ml⁻¹, and incubated at 37°C for 16 h.

The dockerin domain of ScaB (aa 1510 to 1752; TTDANY → DARFGK) was cloned into pET28 in such a way that a His₆ tag was added to the C terminus of the construct. The dockerin domain was amplified with specific primers designed to contain extra 10 bases coding for NcoI and XhoI restriction sites for the forward and reverse primers, respectively (5′-ATTACCATGGACAACAGATGCTA-ACTACGAT and 5′-TTATCTCGAGTTTACCGAATCTTGCGTCTC, respectively). The PCR product was double digested with NcoI/XhoI and ligated into an NcoI/XhoI-digested pET28a(+) vector. The plasmid construct (pETScaB-Doc) was amplified in E. coli XL1-Blue (Stratagene).

Double His₆-S-tagged protein from the pETScaE-Coh construct was overexpressed after transformation in E. coli BL21(DE3) incubated in 50 ml of LB broth supplemented with 30 μg of kanamycin ml⁻¹, 1% (wt/vol) glucose, and 1% (vol/vol) glycerol. The induction phase was carried out by adding 0.5 mM isopropyl-β-d-1-thiogalactopyranoside (IPTG) after the culture reached an optical density at 600 nm of 1.5 to 2.0, and the cells were incubated for an additional 4 h. Cells were recovered and lysed by adding 2.5 ml of BugBuster lysis buffer (Novagen). The protein construct was purified from the soluble fraction by nickel affinity chromatography using Ni-NTA Spin Kit (QIAGEN) according to the manufacturer's instruction. Recombinant protein from pETScaB-Doc was overexpressed in E. coli BL21(DE3)/pLysS (Stratagene) incubated in 1 liter of LB broth in a similar way as described above, except that the induction was carried out at 16°C for 16 h. Cells were recovered by centrifugation, and lysis was carried out by sonication, followed by metal affinity purification using a column of Cu²⁺-NTA (Talon; BD Bioscience, Oxford, United Kingdom) according to the manufacturer's instructions. Molecular weight and His₆ tag reactivity of the protein construct was assessed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis and blotting onto polyvinylidene difluoride (PVDF) membrane, respectively. After blotting the PVDF membrane was incubated with nickel-conjugated peroxidase (INDIA HisProbe-HRP; Pierce Biotechnology, Rockford, IL). His₆ tag reactivity was revealed by enhanced chemiluminescence using the Supersignal West Pico Chemiluminescence Substrate (Pierce Biotechnology) for detection of peroxidase-conjugated (HRP) antibodies according to the manufacturer's guidelines.

Affinity analysis of the cohesin-dockerin interaction.

The cohesin-dockerin interaction was assessed in microtiter plates by affinity analysis using matching fusion-proteins systems (2). For this purpose, XynDoc fusion proteins, consisting of a His₆-tagged Geobacillus stearothermophilus xylanase T-6 cloned upstream of the desired dockerin domain, i.e., derived from Rf Cel44A (previously known as EndB) (38), ScaA (36), or the C-terminal domain of ScaB (13), were produced by using the appropriate plasmid cassette (pETXynDoc) with CGCTGGTACCTGCTAACTACG-ATCACTCCTAC and GGACAGATCTTATTTACCGAATCTTGCGTC as the respective forward and reverse primers. The complementary CBM-Coh fusion protein, comprising the family-3 CBM from the C. thermocellum scaffoldin CipA upstream of the ScaE cohesin, was prepared likewise by using pETCBMCoh, with GCTAGGATCCGGTCAG-GCTTATGATGC and GGCCGCTCGAGTTAAGATG-TAGTACTCTC as forward and reverse primers, respectively. The various proteins were expressed in an E. coli host cell system. The His₆-tagged dockerin-borne constructs were purified by using a Ni-NTA column and the CBM-tagged cohesin-containing constructs using a cellulose resin. Wells of microplates were coated with 0.3 μM concentrations of the desired CBM-Coh construct, and incremental concentrations of the desired XynDoc constructs were added. The washed plates were incubated sequentially with primary anti-xylanase T-6 antibody and HRP-labeled anti-rabbit antibody (2).

Preparation of native protein extracts from R. flavefaciens 17.

Native protein extracts were purified by sequential fractionation from cultures of R. flavefaciens 17. Cellulose-bound proteins were recovered from residual undigested cellulose from a culture of R. flavefaciens 17 grown in 800 ml of modified Hungate-Stack medium containing 1% (wt/vol) of microcrystalline cellulose. The cells were grown statically for several days (usually 5 to 7 days) at 37°C. Residual cellulose was collected after careful aspiration of the supernatant with a vacuum pump. The residual Avicel was then washed five times with phosphate-buffered saline (PBS; 25 mM Na-PO₄ buffer [pH 6.8] and 150 mM NaCl) containing 0.05% Tween 20. Proteins were recovered by incubating the residual cellulose with 2% CHAPS {3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate} and heating at 70°C for 1 h. The suspension was centrifuged (5,000 × g 5 min), the supernatant was retained, and the excess CHAPS removed with an ultrafiltration device (Amicon Ultra, 10,000 MWCO; Millipore, Bedford, MA). The retentate from the last step, containing high-molecular-weight proteins, was resuspended in deionized water. The ultrafiltration step was repeated three times to ensure removal of the CHAPS. Concentrated proteins were stored at −20°C.

Supernatant protein fraction was prepared by freeze-drying the cell-free supernatant of the above culture. The freeze-dried proteins were then resuspended in deionized water until saturation and precipitated using 5 volumes of cold methanol containing 100 mM ammonium acetate. The suspension was incubated overnight at −20°C, and proteins were precipitated by centrifugation (12,000 × g, 20 min, 4°C). The pellet was washed with cold acetone and finally resuspended in deionized water and stored at −20°C.

Cell-wall-attached and cell-wall-associated proteins, representing covalently attached and noncovalently attached proteins, respectively, were isolated from the cell pellet of the same culture used above. After centrifugation to obtain the pellet, it was washed twice in PBS buffer. In order to recover cell-wall-associated proteins, cells were resuspended in 2% N-lauryl sarcosine (sarcosyl) in PBS buffer and incubated for 1 h on ice with occasional vortexing. The suspension was then centrifuged (15,000 × g, 20 min, 4°C). The supernatant fraction from above, containing cell-wall-associated proteins, was collected and ultracentrifuged (60,000 × g, 1 h, 4°C). Proteins were concentrated by using a centrifugal filter device (Amicom Ultra, 10,000 MWCO) and washed twice in PBS to remove excess of sarcosyl.

Cell-wall-attached proteins were then isolated from the cell pellet after removal of cell-wall-associated proteins by using sarcosyl. The pellet was resuspended in 30 ml of PBS and cells were disrupted by using a French press (Thermo Electron, Milford, MA) set to work at 2,500 lb/in² for a total of five times. Unbroken cells were spun down (5,000 × g, 5 min, 4°C), and the supernatant was centrifuged further (18,000 × g, 20 min, 4°C) to precipitate the bacterial cell wall-enriched fraction. The cell wall fraction was resuspended in 3 ml of boiling 8% SDS in PBS, and 3 ml of boiling PBS was further added carefully. The suspension was incubated at 60°C for 1 h with an occasional 5-min incubation in a sonicating bath. The suspension was finally boiled for 15 min at 100°C. The cell wall fraction was recovered by ultracentrifugation (200,000 × g, 30 min, 25°C). The cell wall pellet was resuspended in 1 ml of PBS containing 5 mg of lysostaphin, 500 U of mutanolysin, and 4 mg of lysozyme. The suspension was incubated at 37°C for 18 h, followed by inactivation of the muraminidases by heat inactivation at 100°C for 20 min.

Western blot analysis.

Native proteins of R. flavefaciens 17 were separated in SDS-PAGE and blotted onto a PVDF membrane. The membrane was then blocked for 1 h with TBS-Ca-T buffer (25 mM Tris [pH 7.0], 150 mM NaCl, 1 mM CaCl₂ and 0.05% [vol/vol] Tween 20) containing 3% powdered skimmed milk. Recombinant His₆-tagged ScaE-Coh or ScaB-Doc was added to the blocking buffer to serve as a protein probe. After an additional 1-h incubation, the membrane was washed three times with TBS-Ca-T buffer. Nickel-conjugated peroxidase (India Probe-HRP; Pierce Biotechnology), diluted 1:5,000 in TBS-Ca-T buffer containing 1% powdered skimmed milk, was used as a second probe to specifically recognize His₆-tagged proteins bound to the target proteins on the membrane. After 1 h of incubation and three additional washes of the membranes, cohesin-dockerin interactions were revealed by enhanced chemiluminescence as detailed above.

Biotin labeling of proteins.

The ScaC-Coh protein construct was biotinylated in order to incorporate a different label onto this probe for use in protein-protein interaction experiments where other recombinant products with alternative type of tags (i.e., His₆ tag and S tag) were used. The incorporation of biotin and the detection of biotinylated proteins was carried out as detailed previously (36).

Multiple sequence alignment and phylogenetic analysis.

Phylogenetic trees were generated by using the CLUSTAL W program (http://www2.ebi.ac.uk/clustalw/). The resulting phylogenetic tree was edited by using the program TreeView (http://taxonomy.zoology.gla.ac.uk/rod/rod.html). Dockerin sequences were obtained from the GenBank Web site (http://www.ncbi.nlm.nih.gov/) or via the Carbohydrate-Active Enzymes server (CAZy; http://afmb.cnrs-mrs.fr/∼pedro/CAZY/db.html), designed by Coutinho and Henrissat (8, 10). Sources for the cohesins and enzyme-borne dockerin sequences used in the present study were as reported earlier (13, 36-38, 45, 46). Cell wall anchoring domain sequences were obtained from the Protein families database (http://www.sanger.ac.-uk/cgi-bin/Pfam/getacc?PF00746) and aligned by using CLUSTAL X. Sequences were edited by using the program GeneDoc (http://www.psc.edu/biomed/genedoc/).

Nucleotide sequence accession number.

The DNA sequence for scaE has been deposited in the EMBL Nucleotide Sequence Database under the accession number AJ810899.

RESULTS

ScaE sequence and domain structure.

The three previously characterized cohesin-containing proteins from R. flavefaciens 17 (ScaA, ScaB, and ScaC) are encoded by linked genes within the sca gene cluster (13, 36, 37). The sequence of this cluster was extended downstream from scaB by genome-walking PCR (see Materials and Methods), revealing two additional open reading frames (Fig. 1A). The second of these, which will be designated scaE, encodes a predicted product of 306 aa with an estimated molecular mass of 31.82 kDa. The start codon of the gene is preceded by a putative ribosome-binding site at position −17 with the sequence GAAAGGA, similar to that preceding the scaA, scaB, and scaC genes. There is also a predicted stem-loop secondary structure (T_m = 95.1°C, ΔG = −15.2 kcal/mol) which is located at position +31 downstream of the stop codon and presumably acts as a rho-independent terminator for the transcription of this gene (Fig. 1B). The gene product is estimated to harbor a gram-positive signal peptide sequence with a predicted position for the signal peptidase I cleavage site at 29 to 30 (aa AFA-GP, Fig. 1B), as calculated by the program SignalP (http://www.cbs.dtu.dk/services/SignalP/). BLAST search analysis of the 179-amino-acid N-terminal region of the mature ScaE revealed no significant hits in the database. However, multiple sequence alignment analysis of the region with cohesin domains from R. flavefaciens 17 ScaA, ScaB, and ScaC indicated a distant but significant homology. Furthermore, secondary structure prediction analysis (not shown) of the 179-aa region of ScaE showed a pattern of predicted protein folding similar to that of representative cohesins from ScaA and ScaB.

FIG. 1.

Schematic diagram of the R. flavefaciens Sca cluster and sequence of ScaE. (A) Region of 17 kb from R. flavefaciens 17, showing the sca cluster comprising scaC, scaA, scaB, and scaE. (B) Sequence of scaE with the deduced amino acid sequence on top. The region of the protein highlighted in dark gray represents the predicted signal peptide leader sequence. The region highlighted in black represents the cohesin domain. The LPXTG-like canonical sequence is highlighted in light gray. Arrows indicate position of the primers for cloning of the recombinant protein in the expression vector. The presence of a stem-loop downstream of the stop codon (T_m = 95.1°C, ΔG = −15.2 kcal/mol)⁻¹, calculated by using the M. Zuker algorithm (http://www.bioinfo.rpi.edu/applications/mfold/old/dna/form1.cgi), is also indicated.

The putative cohesin domain of ScaE is followed by a 30-residue threonine-rich linker segment that separates the cohesin from a C-terminal domain. BLAST analysis of the C terminus of ScaE revealed significant hits with surface-anchored proteins in gram-positive bacteria. This region shows several features common to sortase-catalyzed surface-anchored proteins that are found widely in many gram-positive bacterial species (32, 40, 42). It bears the sequence SPKTG, which closely resembles the canonical sortase signal motif LPXTG, preceding a stretch of hydrophobic residues, and terminates in a tail of mostly positively charged amino acids (Fig. 2). This organization suggests strongly that ScaE is a cohesin-containing protein that is anchored covalently to the bacterial cell wall envelope via a presumptive sortase-catalyzed mechanism.

FIG. 2.

Multiple sequence alignment of the sortase signal motif from gram-positive cell-surface proteins. Accession numbers for the sequences used are as follows: STRPN_SUBT (Streptococcus pneumonia Q97RY6), STRSU_MRP (Streptococcus suis P32653), STRMU_GBP (Streptococcus mutans P97082), STAAU_FNBA (Staphylococcus aureus P14738), STRAG_BAG (Streptococcus agalactiae P27951), STRMU_PAC (Streptococcus mutans P11657), STRMU_WAPA (Streptococcus mutans P11000), STAEP_Q6UV37 (Staphylococcus epidermidis Q6UV37), and RUMFL_ScaE (R. flavefaciens 17 AJ810899). Numbers in the right side of the alignment refer to the position of the amino acid within the protein sequence.

Upstream of scaE there is a gene (orf3) coding for an unknown protein (Fig. 1A), while downstream there are two unrelated open reading frames, orf4 and orf5, coding for a predicted ATPase and cystathionine gamma synthase, respectively (Fig. 1A).

Western blotting analysis of cell extracts.

The putative cohesin domain of ScaE was cloned into the pET30 Ek/LIC system in order to generate a recombinant protein (ScaE-Coh) harboring a His₆ tag followed by an S tag at the N terminus. The overexpressed protein was used as a probe in a Western blotting experiment against native protein fractions of R. flavefaciens 17. Three different protein fractions from a R. flavefaciens 17 culture grown on crystalline cellulose as a sole carbon source were probed with ScaE-Coh in the present study. Four protein bands (>300, 150, 75, and ∼70 kDa) were visualized after probing the cellulose-bound protein fraction with ScaE-Coh (Fig. 3A), and a major reactive band of ca. 75 kDa was observed in the cell-wall associated protein fraction (Fig. 3B). The supernatant fraction (Fig. 3C) revealed a positive interaction between the ScaE-Coh probe and a high-molecular-weight protein (>300 kDa). The > 300-kDa native protein band detected in R. flavefaciens 17 by SDS-PAGE appears to represent the presumed cellulosomal anchoring protein ScaB (13). In order to verify the identity of ScaB as the protein band interacting with ScaE-Coh probe, a second immunoblotting experiment was carried out in which native proteins from the supernatant fraction were probed with a rabbit polyclonal anti-ScaB antibody raised against ScaB. The results (Fig. 3D) revealed a positive interaction of the polyclonal antibody with the high-molecular-weight band that interacted specifically with the ScaE-Coh probe. The identity of 150-, 75-, and 70-kDa bands observed in the cellulose-bound protein and cell-wall-associated protein fractions remains unknown.

FIG. 3.

Western blot analysis on native R. flavefaciens 17 proteins. Native protein from cellulose-bound protein fraction (A), cell-wall associated protein fraction (B), and culture supernatant protein fraction (C) were separated by SDS-PAGE and blotted to a PVDF membrane. The membrane was then probed with the recombinant His₆-tagged cohesin of ScaE and cohesin-dockerin interactions were revealed by enhanced chemiluminescence after incubation with nickel-conjugated peroxidase. (D) As in lane C, except the blotted proteins were incubated with rabbit polyclonal antibody against the C terminus of ScaB. Peroxidase-conjugated goat anti-rabbit antibody was used as a secondary antibody followed by enhanced chemiluminescence. Numbers at the left indicate the position of the protein molecular weight markers.

Sequence analysis of the cryptic ScaB dockerin.

The unique C terminus of ScaB has previously been proposed to have a role in cell surface attachment (13). Closer examination of this C-terminal sequence, in this work, indeed revealed a suspected cryptic dockerin sequence that is unusual in character (Fig. 4). One conventional calcium-binding loop (the first loop) is clearly present, but the second repeat, typically found in dockerin sequences, is not obvious. The first loop exhibits a threonine and a serine at the second and third calcium-coordinating positions, respectively; both of these can serve as rare substitutes for the usual aspartate or asparagine residues (the other three conserved calcium-coordinating residues are indeed aspartic acids). Otherwise, the first repeated sequence can be considered a conventional type of dockerin sequence. The second loop, however, is obscured by the inclusion of an apparent 13-residue insertion, the absence of which would yield a standard type of calcium-binding loop, a finding consistent with other known dockerins. Additional 26- and 16-residue insertions are, respectively, located in the linking segment that connects the first “F-helix” with the second calcium-binding loop and at the terminal portion of the ScaB dockerin. These long insertions are unprecedented in the accumulated (>100) dockerin sequences currently in the available public databases. The consequences of these apparent insertions to the action of the putative dockerin are currently unknown. Until now, single residue insertions have been reported in other R. flavefaciens dockerins, and the ScaA dockerin of Acetivibrio cellulolyticus (12) also includes a comparatively small (four-residue) insertion at an alternative position in its first calcium-binding loop (Fig. 4). The ScaB sequence also includes a predicted immunoglobulin-like folding pattern (residues 1512 to 1626, referred as the X-domain hereafter) that precedes the dockerin-like region. The X-domain domain of ScaB shares some similarities with the X-domains that precede the C-terminal dockerins in CipA of C. thermocellum and in ScaA of A. cellulolyticus.

FIG. 4.

Deduced amino acid sequence alignment of the putative C-terminal dockerin domain from R. flavefaciens (Rumfl) ScaB and its relationship to selected type I and type II dockerins. Sequences ins1, ins2, and ins3 represent the 26-, 13-, and 16-residue inserts in the newly identified ScaB dockerin that previously masked its identification. A 116-residue X-domain with a predicted immunoglobulin-like folding pattern precedes the sequence shown in the figure. The degree of conservation of each position within the repeated ScaB sequence is indicated as follows: vertical lines denote identity, colons indicate that the residues are conserved, and dots indicate that the residues are semiconserved, as defined by the EBI Server (http://www2.ebi.ac.uk/clustalw/). The presumed calcium-coordinating residues are shown in large font, highlighted in gray, and suspected specificity residues are shown in bold font, labeled at top with an asterisk. A four-residue insert (ins4) is also present in the A. cellulolyticus (Acece) ScaA sequence. The GenBank accession codes for C. thermocellum (Clotm) Cel48A and CipA, C. cellulolyticum (Cloce) Cel5A, and A. cellulolyticus ScaA are L06942, L08665, M93096, and AF155197, respectively.

Interaction between ScaE-Coh and ScaB-Doc.

To investigate the possible interaction between ScaE-Coh and the ScaB C terminus, dot blot interaction analysis was carried out with recombinant proteins from both ScaE and ScaB. ScaB-Doc and other recombinant dockerin domains from R. flavefaciens 17 cellulosome components were spotted on a nylon membrane and probed with a biotinylated recombinant ScaE-Coh. The recombinant ScaE-Coh recognized specifically and exclusively the ScaB-Doc recombinant protein (Fig. 5).

FIG. 5.

Specificity of the R. flavefaciens 17 ScaE cohesin for the putative ScaB dockerin. Recombinant dockerins from various R. flavefaciens 17 cellulosomal proteins were spotted onto a nylon membrane. A biotinylated recombinant ScaE cohesin was used as a probe. Cohesin-dockerin interaction was revealed by incubation with peroxidase-conjugated streptavidin followed by enhanced chemiluminescence. Spots: 1, ScaB-dockerin; 2, Xyn11E-dockerin; 3, Cel44B-dockerin; 4, ScaA-dockerin; 5, HRP-streptavidin (positive control). (A) Polyhistidine-tagged recombinant proteins spotted onto the nylon membrane and revealed by nickel-conjugated peroxidase that recognize the polyhistidine tag. (B) Same spots as in panel A but after incubation with biotinylated recombinant ScaE-cohesin, which recognizes cognate dockerins, followed by peroxidase-conjugated streptavidin.

The interaction of the ScaE-Coh with ScaB-Doc was further analyzed on microtiter plates, using a highly efficient matching fusion-protein system, in conjunction with an enzyme-linked affinity assay. For this purpose, the relevant cohesins were expressed as fusion proteins downstream of the scaffoldin-borne C. thermocellum cellulose-binding module (CBM). The dockerins were expressed as fusion proteins downstream of the G. stearothermophilus xylanase T6. Both plasmids encoding for the CBM-Coh and XynDoc target proteins were generated with high-expression systems that provide relatively large quantities of the desired proteins. Both types of carrier proteins lend stability to the fusion partner while providing an appropriate standardized functional setting (2). As shown in Fig. 6, the CBM-based ScaE-Coh probe (ScaE-Coh) bound very strongly to the xylanase-borne ScaB-Doc (ScaB-Doc). The affinity of the binding was found to be even higher than that observed for the positive control, i.e., the interaction between the ScaA-based cohesin (ScaA-Coh) and the dockerin of the enzyme Cel44A (formerly known as EndB [38]) (Cel44A-Doc). The interaction between ScaE-Coh and ScaB-Doc is selective, since ScaB-Doc essentially fails to recognize a ScaA cohesin (ScaA-Coh) and ScaE-Coh fails to interact with the Cel44A dockerin (Cel44A-Doc, Fig. 6).

FIG. 6.

Quantitative analysis of the interaction between the ScaE cohesin and the ScaB dockerin. The interaction of R. flavefaciens XynDoc fusion proteins: dockerins from scaffoldin ScaB (ScaB-Doc) and the enzyme Cel44A (Cell44A-Doc) were examined by using CBM-Coh constructs derived from R. flavefaciens ScaE (ScaE-Coh) and ScaA (ScaA-Coh). Microtiter plates were coated with the desired CBM-Coh construct at a concentration of 0.3 μM, and the XynDoc constructs were examined at incremented concentrations. Cohesin-dockerin interactions were revealed by using sequential incubation with primary anti-xylanase T-6 antibody and HRP-labeled anti-rabbit antibody. Absorbance was recorded at 450 nm.

Localization of ScaE in the bacterial cell wall fraction.

In order to determine the location of ScaE on the bacterial surface of R. flavefaciens 17, a cell wall protein extract was prepared by isolating an enriched bacterial peptidoglycan cell wall preparation, followed by digestion with the muraminidases lysostaphin and mutanolysin. The proteins released after muraminidase digestion were separated by SDS-PAGE and subjected to Western blotting, using an overexpressed His₆-tagged dockerin domain from ScaB (ScaB-Doc). The results (Fig. 7) revealed a specific interaction of the ScaB-Doc probe with an ∼25-kDa protein band from the bacterial cell wall protein fraction. The band corresponds to the predicted molecular weight of ScaE after cleavage of the N-terminal signal peptide and the putative cleavage of the C-terminal LPXTG-like motif during translocation of ScaE to the cell wall.

FIG. 7.

Localization of native ScaE in the cell-wall fraction of R. flavefaciens 17. Proteins attached to the bacterial cell wall envelope were separated in SDS-PAGE and blotted onto a PVDF membrane. Recombinant His₆-tagged ScaB-Doc was used as a probe, and the cohesin-dockerin interaction was revealed after incubation with nickel-conjugated peroxidase followed by enhanced chemiluminescence. Lanes: M, protein molecular weight markers (numbers at left indicate the position of the markers in the gel); A, Coomassie blue staining of proteins separated by SDS-PAGE; B, Western blot analysis with recombinant His₆-tagged ScaB-Doc as a probe.

Phylogenetic analysis of the newly identified cohesin domain from ScaE.

Mapping of the ScaE-Coh sequence on a phylogenetic tree, together with other known cohesin sequences, reveals clearly that it comprises a type-III cohesin, related to those of ScaA and ScaB from the same species (Fig. 8). Nevertheless, it differs from the other type-III cohesins, since it occupies its own branch on the tree, whose deflection point is located near those of the ScaA and ScaB cohesins. This probably indicates that an ancestral cohesin was first acquired by the bacterium, likely by horizontal gene transfer. The ancestral cohesin then underwent a gene duplication event(s), resulting in three new genes. The sequences of the cohesins continued to diverge via conventional mutagenesis, followed by additional gene duplication events—first in ScaB (ultimately leading to seven cohesins) and later in ScaA (three cohesins). The single ScaE cohesin indicates that, subsequently, the gene apparently remained exempt from such events, as opposed to the cohesins of ScaA and ScaB. In contrast to the other known cohesins, the single ScaC cohesin occupies a unique branch of the tree, in the vicinity of the type I cohesins. This indicates that ScaC was acquired separately by the bacterium, probably by another horizontal transfer event.

FIG. 8.

Relationship of the R. flavefaciens ScaE-borne cohesin to previously described cellulosomal cohesins. ScaE-Coh maps as a type III cohesin, separated from the other type III cohesins of ScaA and ScaB. In contrast, the R. flavefaciens ScaC cohesin maps on a separate branch, closer to A. cellulolyticus ScaC and the type I cohesins of other species. (For a precise list of the protein sequences used to construct the cohesin tree and their respective accession numbers, see reference 37.) Scale bar indicates the percentage (0.1) of amino acid substitutions.

DISCUSSION

In C. thermocellum, the anchoring of the cellulosome to the bacterial cell wall occurs via one of three currently characterized type-II cohesin-bearing cell-surface proteins (SbdA, OlpB, and Orf2p), each of which carries a C-terminal tandem SLH (surface layer homology) repeat that interacts with the S-layer (17, 23, 24, 26). The scaffoldin protein itself, CipA, interacts with these surface-anchored proteins via a type II cohesin-dockerin interaction. In the mesophilic species C. cellulolyticum (30), C. josui (19), C. acetobutylicum (39), and C. cellulovorans (41), cellulosome attachment occurs by largely unknown mechanisms (30), and the scaffoldin proteins lack any recognized C-terminal dockerin domain. In these species, conserved hydrophilic domains (HLD) found in the scaffoldin sequences may play a role in binding to the cell surface and in assisting the anchoring function of a purported SLH domain present in the cellulosomal enzyme EngE (20). In A. cellulolyticus (46) and Bacteroides cellulosolvens (45), the mechanism of attachment of the cellulosome complex to the cell envelope appears to resemble the SLH-type of interaction reported in C. thermocellum, by virtue of SLH-bearing anchoring scaffoldins that have been described in these two cellulosome-producing species.

Many surface proteins of gram-positive bacteria are attached to the peptidoglycan envelope via an amide bond between the carboxyl-group of a threonine residue and the amino-group of the pentaglycine cell wall crossbridge by a mechanism catalyzed by a sortase (33, 40, 42, 43). The sorting of proteins to the cell surface is initiated by the recognition of a sorting signal in the form of an LPXTG motif (28) and the subsequent proteolytic cleavage and transfer of surface proteins to the cell wall (32, 43). The evidence presented here for sortase-mediated covalent linkage of ScaE to the cell surface provides the first example of this mode of attachment for a bacterial cellulosome complex. The size of the native ScaE as visualized after probing with the His₆-tagged ScaB-Doc was ca. 25 kDa, a finding consistent with cleavage of the hydrophobic and basic tail of the C-terminal sorting domain of the estimated 31.8-kDa ScaE gene product, as predicted in a sortase-mediated anchoring mechanism. Unlike the hydrophilic interactions of the SLH-mediated mechanism, the anchoring of proteins via sortase reflects a covalent attachment to the cell surface. In this context, ScaE could only be released from the cell surface by treatment with muraminidases. Thus, ScaE appears typical of the many gram-positive bacterial surface proteins (e.g., ClfB from Staphylococcus aureus [34], Ace from Enterococcus faecalis [31], and StrH from Streptococcus pneumoniae [9]), that have been reported to be anchored to the cell envelope via a transpeptidation mechanism that requires a sorting signal (28). Slight variations on the canonical LPXTG hexapeptide sorting signal are common, as in the SPKTG motif found for ScaE (28).

ScaE contains a single cohesin domain that shows considerable divergence from cohesins found in the other known R. flavefaciens structural proteins—ScaA, ScaB, and ScaC—encoded by the sca cluster (Fig. 8). The divergent type III cohesin of ScaE is shown here to interact specifically with the atypical C-terminal dockerin-containing region of the ScaB protein. Despite its divergent sequence, the calcium dependence of the binding interaction (2) suggests that the ScaB dockerin is responsible for binding to the ScaE cohesin.

The results of the protein-protein interaction experiments with the recombinant cohesin of ScaE indicate that in addition to the specific binding to ScaB via its C-terminal dockerin, ScaE also interacts with at least three other yet-unknown proteins from the cellulose-bound protein fraction. This fact indicates that these proteins must either bind cellulose individually or form a complex that includes at least one polypeptide that binds tightly to cellulose. Thus, ScaE may also play a key role in linking cells to their substrate by anchoring either unidentified cellulose-binding proteins or complexes containing cellulose-binding proteins to the cell surface.

We have inferred previously at least two distinct cohesin-dockerin interactions between enzyme subunits and scaffoldin proteins (36), together with two additional interactions: one between the ScaA dockerin and ScaB cohesins (13) and one between the ScaC cohesin and an uncharacterized set of dockerins (37). The interaction between ScaE and the ScaB C terminus therefore represents the fifth cohesin-dockerin specificity to be identified in this species. As shown in Fig. 9, ScaE is attached covalently to the cell surface via a sortase-mediated reaction, and its N-terminal cohesin binds to the ScaB dockerin. In turn, the seven ScaB cohesins can each interact with the dockerin of ScaA, whose cohesins can interact either directly with an array of components which bear Cel44A-like dockerins or indirectly, via the ScaC adaptor scaffoldin, with a series of as-yet-uncharacterized dockerin-containing components. Finally, the cohesins of at least one additional, hypothetical scaffoldin, provisionally designated ScaD, are proposed to interact with a set of components, which harbor a divergent type of dockerin exemplified by those of the enzymes CE3B and Xyn11E (formerly referred as CesA and XynE, respectively). In view of the above, the characteristics of the R. flavefaciens cellulosome appear to reflect its intricate and versatile nature, both in its modular and subunit arrangement and in its capacity to incorporate numerous types of cellulosomal components, presumably in response to environmental requirements. Future research will be devoted to unraveling the complex collection of intermodular interactions that dictate the particularly elaborate architecture of the R. flavefaciens cellulosome.

FIG. 9.

Schematic representation of our current view of cellulosome organization on the R. flavefaciens cell surface. Five different cohesin-dockerin specificities are shown. (a) The ScaE-Coh is implanted into the cell surface via a sortase-like signal motif and binds strongly to the atypical ScaB-Doc. (b) The 7 ScaB-Coh's bind to the ScaA-Docs. (c) The 3 ScaA-Cohs bind either to Cel44A-type dockerins or to the ScaC-Doc. (d) The ScaC-Coh binds to one of a number of currently unknown dockerin-containing proteins. (e [inset]) Cohesins of a putative scaffoldin (“ScaD”) bind to the dockerins of CE3B, Xyn11E, and other related dockerins. A dockerin for ScaD has not yet been described, and it is as yet unknown how ScaD and its complement of enzymes and other proteins are incorporated into the R. flavefaciens cellulosome.