Two SusD-Like Proteins Encoded within a Polysaccharide Utilization Locus of an Uncultured Ruminant Bacteroidetes Phylotype Bind Strongly to Cellulose

Abstract

We demonstrate that two characteristic Sus-like proteins encoded within a polysaccharide utilization locus (PUL) bind strongly to cellulosic substrates and interact with plant primary cell walls. This shows associations between uncultured Bacteroidetes-affiliated lineages and cellulose in the rumen and thus presents new PUL-derived targets to pursue regarding plant biomass degradation.

TEXT

The Bacteroidetes are the most abundant Gram-negative bacteria in gut microbiomes and are commonly associated with degradation of xylan and other noncellulosic polysaccharides (4). However, cellulolytic Bacteroidetes isolates have been described (13), suggesting that these bacteria also contribute toward cellulose degradation in the gut, despite the lack of genes corresponding to common cellulases in glycoside hydrolase families 6 and 48 (GH6 and GH48) (3). Recent metagenomic analyses of rumen microbiomes have revealed the occurrence of polysaccharide utilization loci (PULs) linked to putative GH5 and GH9 cellulases in several uncultured Bacteroidetes phylotypes (12). Bacteroidetes-affiliated PULs are typified by gene clusters that encode lipoanchored glycoside hydrolases together with a set of outer membrane lipoproteins (referred to as Sus-like). The Sus-like proteins bear a resemblance to proteins of the starch utilization system (Sus), first identified in the human gut bacterium Bacteroides thetaiotaomicron (14). The so-called SusD-like proteins contribute to saccharide capture, as has been demonstrated for starch and fructan (8, 15). Aside from starch and fructan PULs, pectin and hemicellulose PULs have been detected in isolated gut bacteria (5, 9). Nothing is known about the association of PULs with cellulose.

Figure 1A shows a PUL from the dominating uncultured SRM-1 (for “Svalbard reindeer microorganism 1”) phylotype found in the Svalbard reindeer rumen, which exhibits only 91% sequence identity to its closest cultured relative, Bacteroidales genomosp. P1 (12). This PUL was found on a fosmid encoding activity for carboxymethylcellulose (CMC) and includes two putative GH5 cellulases, a cellobiose phosphorylase (GH94), and various Sus-like proteins, including SusD1 and SusD2. The SusD1 and SusD2 genes (accession numbers JQ755420 and JQ755421; 22% amino acid sequence identity at the protein level) were cloned into the pNIC-CH expression vector using ligation-independent cloning (LIC) (1) and primers SusD1_lic_NT (TTAAGAAGGAGATATACTATGGTGGACCGGCTCGCCATCGGCGACGCATTC), SusD1_lic_CT (AATGGTGGTGATGATGGTGCGCCCAACCGGGATTCTGCGTGAGGCCGTATCC), SusD2_lic_NT (TTAAGAAGGAGATATACTATGGTCGACCTCAACTATACGGAGGAGAACACA), and SusD2_lic_CT (AATGGTGGTGATGATGGTGCGCCCATCCTGCATTTTGGGTGAGGTTGGGGTT) (overhangs are underlined). Subsequently, recombinant proteins lacking the putative signal peptide and containing a C-terminal His₆ tag were overexpressed in Escherichia coli BL21, purified by immobilized metal affinity chromatography, and dialyzed and concentrated using Vivaspin concentrators. To analyze polysaccharide binding, purified proteins (1 mg/ml), 6% (wt/vol) Sigma cellulose, Avicel (Sigma-Aldrich), filter paper (Whatman), and the insoluble fractions of xylan (Carl Roth GmbH), mannan (Megazyme), or lichenin (Megazyme) were combined in the presence of MES buffer (20 mM; pH 6; final volume, 200 μl) and incubated at 37°C with vertical shaking at 1,000 rpm for 1 h. After centrifugation, the supernatant (referred to as flowthrough) was removed, and the insoluble substrate was resuspended in 200 μl of MES buffer and incubated for 15 min, after which the supernatant was removed by centrifugation (referred to as the wash step). Bound proteins were eluted with 100 μl of 50 mM bis-Tris-propane with 5% Triton (pH 10). Harsher denaturing conditions (100 μl of 8 M urea and boiling for 10 min) were also used (for lichenan, boiling was omitted).

Fig 1.

Gene arrangement of the SRM-1 putative GH5-linked PUL and binding of SusD proteins to cellulose. (A) The SRM-1 PUL gene cluster described with putative functional assignments in reference 12 consists of two SusC-like TonB-dependent receptors (blue), two SusD-like glycan-binding proteins (light brown), a hypothetical SusF-like outer membrane lipoprotein (green), a putative inner membrane-bound sugar-transporter (white), an acetyl xylan esterase (yellow), and an assortment of putative glycoside hydrolases (red). GH5, endoglucanase; GH26, mannanase; GH43, xylosidase-arabinanase-arabinofuranosidase; GH94, cellobiose phosphorylase; CE7, acetyl xylan esterase; +SP, signal peptide detected (indicates a putative outer membrane protein). Genes encoding SusD-like proteins that we describe in this study are indicated by arrows. (B to D) SDS-PAGE gels showing binding of SusD1 (D1) and SusD2 (D2) to various polysaccharides. Fractions are labeled as follows: C or Control, protein loaded without substrate; ft, flowthrough fraction containing unbound protein; w, wash fraction; and e, protein eluted from substrate with 5% Triton (B) or by treatment with 8 M urea and boiling (C and D). −, empty lane. The binding experiments whose results are shown in panels B and C were at pH 6; panel D shows experiments at varying pH. Sample volumes for the SDS-PAGE analysis were identical.

Analysis of the various fractions from the binding experiments by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (Fig. 1B to D) demonstrated that SusD1 and SusD2 bind to various forms of cellulose. Elution with Triton failed (Fig. 1B), whereas elution was achieved with 8 M urea and boiling (Fig. 1C). Elution of SusD2 was incomplete in all cases, suggesting that SusD2 exhibits a different binding mechanism than SusD1 (Fig. 1C). A further difference between SusD1 and SusD2 is that only the former binds to lichenan (β-1,3,β-1,4 β-glucan) (Fig. 1B and C). Studies with SusD1 showed that binding is pH dependent and strongest at pH <8.0 (Fig. 1D) (the pH of the Svalbard reindeer rumen ranges from 6 to 6.75 [11]). Interestingly, both proteins exhibited only weak binding to mannan or xylan. This result is notable because the presence of putative GH26 (mannanase), GH43 (xylosidase, arabinanase, arabinofuranosidase), and CE7 (acetyl xylan esterase) enzymes within the SRM-1 PUL (Fig. 1A) suggests such hemicellulosic substrates as potential targets.

To further explore their ability to recognize plant polysaccharides, we investigated binding of SusD1 and SusD2 to the cell wall of Arabidopsis thaliana (Fig. 2). Hand-cut sections through the stems of 4- to 5-week-old plants were labeled using a His₆ tag-based three-stage procedure essentially as previously described (10), in which binding was detected using a fluorescein isothiocyanate conjugated tertiary antibody. Cellulose-binding CBM3a from Clostridium thermocellum (2) was included as a positive control. The binding of SusD2 and that of the positive control, CBM3a, were similar in that both produced widespread labeling across diverse cell types and both produced a characteristic punctate labeling pattern. However, there were subtle differences in the binding of these probes. CBM3a bound predominantly to the adhered faces of adjacent pith parenchyma cell walls, whereas SusD2 binding was more apparent to regions of wall delineating the intercellular spaces. Also, CBM3a bound strongly to epidermal cell walls but weakly to the walls of underlying cortical cells, whereas the reverse was true for SusD2. SusD1 did not bind to equivalent sections (data not shown), confirming that SusD1 and SusD2 have different binding specificities. Interestingly, our data may be taken to indicate that SusD1 has greater binding affinity for lichenan (Fig. 1), a substrate that is scant in cell walls of dicotyledons such as A. thaliana.

Fig 2.

Indirect immunofluorescence microscopy of SusD2 and CBM3a showing binding to transverse sections of Arabidopsis stem sections. The images show binding of CBM3a (A and B) and SusD2 (C and D) to pith parenchyma (A, C) and cortical parenchyma (cp) plus epidermal cells (ep) (B, D). (E and F) Negative controls (experiments without addition of a binding protein). Insets (A and C) show regions near intercellular spaces (*). Binding was detected using a fluorescein isothiocyanate-conjugated tertiary antibody as previously described (10). Arrows indicate adherent faces of adjacent cell walls, and the double arrow indicates a region of wall delineating the intercellular space. Bars, 125 μm.

To our knowledge, these data provide the first experimental evidence linking Sus proteins and PULs to cellulose. Moreover, the difference in binding specificities suggests that SusD1 and SusD2 have complementary functions and are optimized to bind to distinct features of the microstructure of cell walls. The variety of putative glycoside hydrolases encoded within the SRM-1 PUL suggests activities against a broad range of hemicellulosic (GH5, GH26, GH43, CE7) and cellulosic (GH5, GH94) substrates. Preliminary activity data obtained with overexpressed enzymes show that the two GH5 enzymes cleave β-1,4-linked glucose units in various substrates, including Avicel, phosphoric acid-swollen cellulose, lichenan, and glucomannan, and that they produce cellobiose. It remains to be elucidated if insoluble cellulose is a target substrate for this PUL or if SusD binding to cellulose serves the purpose of positioning PUL-linked glycoside hydrolases close to other (hemicellulosic) polysaccharides intertwined with cellulose, i.e., a proximity effect similar to that shown for certain CBMs (6). Interestingly, as previously pointed out (12), one of the Avicel-degrading enzymes extracted from the cow rumen metagenome (7) is part of a PUL containing a SusD-like homologue. All in all, available data strengthen the hypothesis that the membrane anchored enzyme systems encoded by PULs are involved in cellulose degradation. Confirmation of this hypothesis would establish a third paradigm for cellulose degradation, next to cellulosomes and free enzyme systems.