Exoproteome Profiles of Clostridium cellulovorans Grown on Various Carbon Sources

Abstract

The cellulosome is a complex of cellulosomal proteins bound to scaffolding proteins. This complex is considered the most efficient system for cellulose degradation. Clostridium cellulovorans, which is known to produce cellulosomes, changes the composition of its cellulosomes depending on the growth substrates. However, studies have investigated only cellulosomal proteins; profile changes in noncellulosomal proteins have rarely been examined. In this study, we performed a quantitative proteome analysis of the whole exoproteome of C. cellulovorans, including cellulosomal and noncellulosomal proteins, to illustrate how various substrates are efficiently degraded. C. cellulovorans was cultured with cellobiose, xylan, pectin, or phosphoric acid-swollen cellulose (PASC) as the sole carbon source. PASC was used as a cellulose substrate for more accurate quantitative analysis. Using an isobaric tag method and a liquid chromatography mass spectrometer equipped with a long monolithic silica capillary column, 639 proteins were identified and quantified in all 4 samples. Among these, 79 proteins were involved in saccharification, including 35 cellulosomal and 44 noncellulosomal proteins. We compared protein abundance by spectral count and found that cellulosomal proteins were more abundant than noncellulosomal proteins. Next, we focused on the fold change of the proteins depending on the growth substrates. Drastic changes were observed mainly among the noncellulosomal proteins. These results indicate that cellulosomal proteins were primarily produced to efficiently degrade any substrate and that noncellulosomal proteins were specifically produced to optimize the degradation of a particular substrate. This study highlights the importance of noncellulosomal proteins as well as cellulosomes for the efficient degradation of various substrates.

INTRODUCTION

The cellulosome, a protein complex produced by many cellulolytic Gram-positive anaerobic bacteria, such as Clostridium, efficiently degrades plant cell wall polysaccharides (1–3). The cellulosome was first found in Clostridium thermocellum (4–6), and much progress has been made in understanding the characteristics of the highly cellulolytic protein complex. However, the molecular mechanism of the formation of the cellulosome has not been characterized (7). Previously, we sequenced the entire genome of Clostridium cellulovorans and identified all genes, including those that encode proteins of known and unknown functions, related to cellulosomal composition (8). Genome analysis of C. cellulovorans indicated the presence of 57 cellulosomal protein-encoding genes and 168 noncellulosomal (hemi)cellulolytic protein-encoding genes. The noncellulosomal (hemi)cellulolytic proteins (hereafter called noncellulosomal proteins) have an N-terminal signal peptide but do not have dockerin domains. The 57 cellulosomal protein-encoding genes included 4 scaffold protein-encoding genes and 53 cellulosomal protein-encoding genes with dockerin domains. The major scaffold protein, CbpA, is comprised of 9 cohesin domains that bind to various cellulosomal proteins via cohesin-dockerin interactions (9). Interestingly, even though most of the cellulosomal proteins were glycoside hydrolases (GHs), proteins such as proteases, protease inhibitors (36), and unknown proteins were also included. These proteins, seemingly not related to saccharification, may be important in the degradation of various resources. However, we have only general information regarding the proteins that actively degrade biomass, although genome analysis has provided many interesting insights into the characteristics of C. cellulovorans. Additionally, C. cellulovorans has a larger number of noncellulosomal proteins (10) that have polysaccharide degradation ability but do not have dockerin domains in their amino acid sequences than other cellulosome-producing clostridia.

In almost all previous studies of the cellulosome, cellulosomal proteins have been the focus (11, 12). However, it is important to note that C. cellulovorans must utilize noncellulosomal proteins for polysaccharide degradation under native conditions (13). There have been several studies about noncellulosomal proteins from either C. thermocellum or C. cellulovorans using gel-based methods, but they were limited to only a small fraction of the noncellulosomal proteins (14, 15). Without understanding the contribution of noncellulosomal proteins comprehensively, it has been impossible to determine the strategy by which cellulosome-producing bacteria degrade many kinds of substrates.

In this study, we performed quantitative proteomic analysis using tandem mass tag (TMT)-labeled samples on a mass spectrometer in order to characterize the quantitative changes in cellulosomal and noncellulosomal proteins secreted by C. cellulovorans (16). Using a unique long (480-cm) monolithic silica capillary column-based liquid chromatography-tandem mass spectrometry (LC-MS/MS) approach, we successfully identified and quantified the exoproteome of C. cellulovorans, including both cellulosomal and noncellulosomal proteins, without cellulosome isolation and prefractionation. Through this approach, we attempted to clarify how C. cellulovorans optimizes cellulosomal and noncellulosomal proteins to degrade different substrates.

MATERIALS AND METHODS

Growth substrates.

Cellobiose (Sigma, MO, USA), xylan (Sigma), and pectin (Sigma) were used in the growth experiments. Phosphoric acid-swollen cellulose (PASC) was prepared from a form of microcrystalline cellulose (Merck, Darmstadt, Germany) (18) and was used in the growth experiments.

Cell culture and medium.

C. cellulovorans 743B (ATCC 35296) was grown anaerobically as described previously (17), except for the carbon source, which was 0.3% (wt/vol) cellobiose, 0.3% (wt/vol) xylan, 0.3% (wt/vol) pectin, or 0.3% (wt/vol) PASC.

Estimating the growth of the anaerobic bacteria.

The growth curves of anaerobic C. cellulovorans on cellobiose, xylan, pectin, and PASC were determined by bacterial protein estimation, as described by Bensadoun and Weinstein (19). Five hundred microliters of cell culture was centrifuged for 10 min at 13,000 × g. The cell pellets were washed with 500 μl of phosphate-buffered saline (pH 7.4) and incubated with 400 μl of sodium deoxycholate (2%) for 20 min at 37°C. One hundred microliters of trichloroacetic acid (24%) was added to the suspension, which was centrifuged at 13,000 × g for 10 min. Fifty microliters of resolubilization solution (5% SDS, 2 N NaOH) was added to the suspension and vigorously mixed. The protein concentration was measured using a bicinchoninate protein assay kit (Nacalai Tesque, Kyoto, Japan), with bovine serum albumin as the standard.

Preparation of extracellular proteins for quantitative proteome analysis.

Proteome samples were prepared from C. cellulovorans culture medium. The culture (50 ml) was centrifuged (6,000 × g, 25°C), and the supernatant was subjected to ultrafiltration using Amicon Ultra YM-10 (Millipore, Bedford, MA) to obtain the extracellular proteins. The concentrated samples from each culture were dissolved in 100 μl of triethylammonium hydrogen carbonate buffer (200 mM). Five microliters of tris(2-carboxyethyl)phosphine (200 mM) was added. The reaction was performed for 60 min at 55°C. Five microliters of iodoacetamide (375 mM) was then added, and the mixtures were reacted for 30 min, protected from light, at room temperature. Two microliters of freshly prepared sequence-grade modified trypsin (1 μg/μl; Promega, WI, USA) was added, and the proteins were digested overnight at 37°C. The 4 proteome samples (cellobiose, xylan, pectin, and PASC) were labeled using a tandem mass tag (TMT) 6-plex labeling kit (Thermo Fisher Scientific, MA, USA) with reporters at m/z 126, 127, 128, and 129, respectively, in 41 μl of CH₃CN. After 60 min of reaction at room temperature, 8 μl of 5% (wt/vol) hydroxylamine was added to each tube and mixed for 15 min. In addition, a mixture of tryptic fragments from all substrates was reacted with TMT-131 as an internal standard for quantification. The aliquots were then combined, and the pooled sample was evaporated under vacuum. The samples were then dissolved in 100 μl of trifluoroacetic acid (0.1%) before LC-MS/MS analysis.

LC-MS/MS analysis.

Proteome analyses were performed using a liquid chromatography-mass spectrometry system equipped with a long monolithic silica capillary column, as described in our previous study (13). A monolithic silica capillary column was prepared from a mixture of tetramethoxysilane and methyltrimethoxysilane (20). Tryptic digests were separated by reversed-phase chromatography using a monolithic silica capillary column (480 cm long, 0.1-mm inner diameter [ID]) at a flow rate of 500 nl/min. The gradient was provided by changing the mixing ratio of the 2 eluents: A, 0.1% (vol/vol) formic acid, and B, 80% acetonitrile containing 0.1% (vol/vol) formic acid. The gradient was started with 5% B, increased to 45% B for 600 min, further increased to 95% B to wash the column, returned to the initial condition, and held for reequilibration. The separated analytes were detected on a mass spectrometer with a full scan range of m/z 350 to 1500 (resolution 60000), followed by 10 data-dependent higher-energy c-trap dissociation (HCD) MS/MS scans acquired for TMT reporter ions. A normalized collision energy of 80% in HCD with a 0.1-ms activation time was used. An electrospray ionization (ESI) voltage of 2.4 kV was applied directly to the LC buffer distal to the chromatography column using a microtee. The ion transfer tube temperature on the LTQ Velos ion trap was set to 300°C. Duplicate analyses were performed for each sample in 3 independent experiments. Blank runs were inserted between runs for different samples.

Data analysis.

The mass spectrometry data files were analyzed for protein identification and quantification using a WF_LTQ_Orbitrap_Mascot_HCD_ReporterQuantitation template in Protein Discoverer software (Thermo Fisher Scientific). Protein identification was performed using the Mascot algorithm against the C. cellulovorans protein database (4,254 sequences) from NCBI (http://www.ncbi.nlm.nih.gov/) with a precursor mass tolerance of 20 ppm and a fragment ion mass tolerance of 20 milli-mass units (mmu). Carbamidomethylation of cysteine and the TMT 6-plex at the N terminus were set as fixed modifications. Protein quantification was performed using the Reporter Ions Quantifier with the TMT 6-plex method. The data were then filtered with a cutoff criterion of a q value of ≤0.05, corresponding to a 5% false-discovery rate (FDR) on a spectral level. The detected spectral counts of the identified peptides were used to estimate the abundances of cellulosomal and noncellulosomal proteins. Proteins with no missing values in 3 replicates were accepted in the protein quantitation analysis. Global median normalization was carried out to normalize the amount of tryptic digests injected into the mass spectrometer. The heat map was constructed using Cluster 3.0 (21), which can perform hierarchical cluster analysis. Euclidean distance was used to measure the similarities of the protein profile patterns within the clustering analysis. To visualize the clustering results from Cluster 3.0, Java TreeView (22) software was used.

RESULTS

In order to elucidate the quantitative changes in cellulosomal and noncellulosomal proteins in the exoproteome of C. cellulovorans on different growth substrates, C. cellulovorans was cultured under anaerobic conditions with cellobiose, xylan, pectin, or PASC as the sole carbon source. Proteome samples were prepared from the supernatants without cellulosome purification and prefractionation. Using a quantitative proteomic approach based on an isobaric tag method, we obtained the quantitative proteomic profiles of 4 kinds of C. cellulovorans exoproteomes, as illustrated in Fig. 1.

Fig 1.

Schematic representation of the work flow used to quantify proteins in the exoproteome. The proteins in culture supernatants of C. cellulovorans grown with cellobiose, xylan, pectin, or PASC as a sole carbon source were individually digested with trypsin, and each peptide sample and the internal standard sample were labeled with tandem mass tags (TMTs) with various reporters at m/z 126, 127, 128, 129, and 131, respectively. The mixture of labeled samples was applied to LC-MS/MS with a long monolithic silica capillary column to analyze the proteins secreted on each growth substrate. Statistical analyses of the proteome data were performed to elucidate the quantitative changes in the exoproteome profile on 4 different growth substrates.

Growth confirmation.

To determine the growth of C. cellulovorans with the 4 different substrates, we performed bacterial protein estimation (23) (Fig. 2). C. cellulovorans can grow on pectin and xylan as a sole carbon source, whereas other cellulosome-producing bacteria, such as Clostridium thermocellum, cannot (17). PASC was used as a cellulose substrate instead of Avicel for faster growth of C. cellulovorans and larger amounts of produced proteins, comparable to when the other substrates were used (i.e., cellobiose, xylan, and pectin), which leads to more convenient and accurate quantitative analysis. Moreover, even though it is not crystalline cellulose, we can see the difference between when the polymer and the disaccharide were used as the substrates. Although there was an initial delay of growth in pectin and PASC cultures compared with that in cellobiose and xylan cultures, 12-day supernatants from cellobiose, xylan, pectin, and PASC cultures had comparable amounts of protein. Hence, we determined that C. cellulovorans cultured for 12 days was in stationary phase and that we could obtain the appropriate amounts of protein for proteome analysis at this time point. We performed quantitative proteome analysis of C. cellulovorans supernatant cultured with the 4 kinds of substrates for 12 days.

Fig 2.

Confirmation of growth of C. cellulovorans cultured with 4 different growth substrates. The growth curve was determined by measuring the amount of protein inside the cells. Cells cultured for 12 days (in stationary phase in all samples) were used for the proteome analysis.

Quantitative proteome analysis of C. cellulovorans culture supernatants.

We used tandem mass tag (TMT) 6-plex isobaric tags and nano-LC-MS/MS with a long monolithic silica capillary column (480 cm). Each TMT-labeled peptide sample was mixed at a specific ratio such that the mixed samples were estimated to contain the same amount of TMT-labeled peptides from each sample. The peptide concentrations of each sample were measured using the bicinchoninic acid (BCA) assay. Principal-component analysis (PCA) was performed to examine the similarity of the protein production profiles between 3 biological replicates of culture supernatant for each substrate. A score plot of PCA showed high similarity between each of the biological replicates, and the plots of each substrate formed individual groups, indicating that the protein profiles of the exoproteomes reflected the difference in substrate (see Fig. S1 in the supplemental material).

C. cellulovorans has 4,254 protein-coding genes in its genome. The mass spectrometry data were used for protein identification with the protein database built from the genome analysis of C. cellulovorans. In total, with the cutoff criteria, we identified 639 proteins in each sample from 3 biological replicates. All identified proteins are summarized in Table S1 in the supplemental material. The quantitative data were combined and standardized using a median approach to normalize the amount of tryptic digest injected into the mass spectrometer. We then assessed the variance between all biological replicates. Scatter plots of all comparisons were created using the data from the combination of the biological replicates (see Fig. S2 in the supplemental material).

Quantitative exoproteome analysis focused on saccharification.

The work flow for the data analysis and a summary of the data are shown in Fig. 3. From the 639 identified proteins, cellulosomal and noncellulosomal proteins (see Table S2 in the supplemental material) involved in saccharification were selected for further study, and the number of these proteins was 79. Of these, 35 were cellulosomal proteins and 44 were noncellulosomal proteins. Our previous genomic study identified 57 and 168 genes encoding cellulosomal and noncellulosomal proteins, respectively. Thus, our results indicate that 61.4% (35 of 57) of the cellulosomal proteins and 26.1% (44 of 168) of the noncellulosomal proteins were produced under the conditions analyzed in this study.

Fig 3.

Work flow and summary of the data analysis of the C. cellulovorans exoproteome. The total number of proteins identified and the number of selected proteins involved in saccharification are shown. The numbers of cellulosomal and noncellulosomal proteins in the total data set and among the substrate-specific extracellular proteins are also shown. The number of substrate-specific extracellular proteins was 26, and 81% of them were noncellulosomal proteins. Among the 57 cellulosomal genes that have been reported (10), 35 cellulosomal proteins were produced in all of the samples in this study. The numbers of pectin- and PASC-specific cellulosomal proteins were 0, which means that 35 cellulosomal proteins were produced but substrate-specific cellulosomal proteins with significant fold changes among them were not observed.

Using the standardized relative quantitative data for 79 cellulosomal and noncellulosomal proteins, hierarchical cluster analysis was performed to generate a heat map and to observe the trends of extracellular proteins (Fig. 4). In the hierarchical cluster analysis, the 3 biological replicates clearly grouped into small clusters. Within each cluster, generally, many xylanases from the xylan samples, many pectate lyases from the pectin samples, and many kinds of glucanases from the cellobiose and PASC samples were most commonly observed. These results indicate that C. cellulovorans recognized the extracellular environment (in this case, the substrate) and secreted the optimal enzymes necessary for degrading each substrate (24).

Fig 4.

Exoproteome profiles of C. cellulovorans cultivated with 4 different growth substrates shown in a heat map after hierarchical cluster analysis. The quantitative proteome data from 3 biological replicates of each substrate sample, 12 samples in total, were used for the hierarchical cluster analysis. The tree structure represents the similarity between the data of each of the 12 samples and shows that the biological replicates grouped together, indicating that their exoproteome profiles were similar and the profiles were reproducible. Distinctive differences in the exoproteome profiles depending on the growth substrate were confirmed, and most of the proteins produced in response to a given growth substrate were proteins able to degrade that substrate.

Abundance of cellulosomal and noncellulosomal proteins.

Among the selected 79 cellulosomal and noncellulosomal proteins, the spectral counts of each protein were used to compare the abundances of the cellulosomal and noncellulosomal proteins. The average of the spectral counts of the cellulosomal proteins was much larger than that of the noncellulosomal proteins (Table 1). In addition, a scaffolding protein of the cellulosome (Clocel_2824) showed the highest counts, and basic cellulosomal proteins (Clocel_2823 and Clocel_3359), which are reported to be constitutively produced (13), and another cellulosomal protein produced in xylan culture (Clocel_2821) also showed high counts (third, fourth, and second, respectively) (see Table S2 in the supplemental material). The results indicate that the cellulosomal proteins were more highly produced than the noncellulosomal proteins.

Table 1.

Protein abundance index of the exoproteome

Protein type	No. of proteinsa	No. of substrate-specific proteins	ΣSpCb	ΣSpC/no. of proteins
Cellulosomal	35	5	7,937	227
Noncellulosomal	44	21	3,924	89

^aNumber of selected proteins involved in saccharification among 639 identified proteins.

^bΣSpC, sum of spectral count using data analysis for each protein.

Substrate-specific extracellular proteins.

To determine the proteins that showed a significant fold change between different growth substrates, an empirical Bayes moderated t test was performed. P values were adjusted with the Benjamini-Hochberg method to avoid the problem of multiple testing. Volcano plots were generated to visualize extracellular proteins present in different amounts. The criteria that we adopted for the extracellular proteins present in different amounts were an FDR-adjusted P value of <0.01 and a fold change of protein ratio of >2. Extracellular proteins defined as present in different amounts were identified in all comparisons, and they are shown as blue dots in the volcano plots (see Fig. S3 in the supplemental material).

Pectin/cellobiose and xylan/cellobiose volcano plots are shown in Fig. 5, as examples of the analysis. We identified proteins differentially produced in pectin medium compared to cellobiose medium. Proteins classified as polysaccharide lyases (PLs) by CAZy (http://www.cazy.org/) are putative pectin-degrading proteins. The PLs among the proteins differentially produced in pectin medium are indicated in pink font with an underline. In total, 9 PLs were identified in the group of 79 selected proteins (see Table S2 in the supplemental material), and 7 of these PLs were differentially produced in pectin medium compared to cellobiose medium (Fig. 5a). Among these proteins is Clocel_1623 (PelA), a well-known cellulosomal pectate lyase that was classified as a member of the PL1 and PL9 families; it is reported to degrade polygalacturonic acid into digalacturonic acid or trigalacturonic acid (25). Moreover, 4 of the 7 PLs were also highly produced in xylan medium (Fig. 5b), possibly because xylan and pectin have similar structures, including pentose sugar moieties.

Fig 5.

Two representative volcano plots, pectin/cellobiose (a) and xylan/cellobiose (b), show extracellular proteins present in different amounts. The extracellular proteins present in different amounts were defined as the proteins secreted on a specific substrate that showed an FDR-adjusted P value of <0.01 and a fold change of protein ratio of >2 compared to secretion on a different substrate; they are shown as blue dots. The names in pink font with underlines are known or putative pectin-degrading enzymes which are categorized into the polysaccharide lyase (PL) family by CAZy.

We defined the proteins that were present in a different amount on one specific substrate compared to all of the 3 other substrates as “substrate-specific extracellular proteins” (i.e., cellobiose-specific proteins, xylan-specific proteins, pectin-specific proteins, and PASC-specific proteins). The list of these substrate-specific extracellular proteins is shown in Table 2, and the details are discussed in the following sections. The number of substrate-specific extracellular proteins was 26 (Fig. 3), and 53 other proteins were basal proteins that did not show a significant fold change on different growth substrates.

Table 2.

Substrate-specific extracellular proteins

Substrate specificity	Protein type	Locus	Name	CAZy family(ies)a	vs cellobiose		vs xylan		vs pectin		vs PASC
					Log2 fold change	FDR-adjusted P valueb	Log2 fold change	FDR-adjusted P valueb	Log2 fold change	FDR-adjusted P valueb	Log2 fold change	FDR-adjusted P valueb
Cellobiose	Cellulosomal	Clocel_1624	EngY	CBM30, GH9			1.82	2.54E−07	1.65	4.94E−07	2.28	3.57E−08
		Clocel_2575		CBM35,GH26			1.13	7.86E−05	2.09	4.94E−07	2.11	3.72E−07
		Clocel_3112		NA			2.1	1.14E−08	2.25	1.11E−08	2.71	3.30E−09
	Noncellulosomal	Clocel_1011		PL1, CBM13			1.72	5.38E−06	1.27	4.54E−05	1.53	9.41E−06
		Clocel_1420		CBM30, GH9			1.46	2.91E−04	2.07	1.46E−05	2.23	6.88E−06
		Clocel_1478	EngO	CBM4, GH9			1.31	1.86E−05	2.11	4.32E−07	1.75	1.43E−06
		Clocel_2815	EngN	NA			2.56	7.13E−08	1.57	3.65E−06	2.09	3.17E−07
		Clocel_3196		GH130			2.24	3.59E−06	2.03	6.96E−06	1.38	9.36E−05
		Clocel_3197		GH130			1.68	6.86E−07	1.46	2.01E−06	1.03	1.93E−05
		Clocel_3242	EngD	GH5, CBM2			1.52	5.90E−07	1.39	1.04E−06	1.82	9.82E−08
		Clocel_3662	EngF	GH5, CBM17			4.13	2.54E−06	2.47	1.04E−04	3.6	5.15E−06
Xylan	Cellulosomal	Clocel_1432		GH5	2.72	3.94E−05			2.7	6.34E−05	2.99	1.47E−05
		Clocel_2295	XynA	GH11, CE4	4.5	5.49E−08			4.38	1.20E−07	4.51	5.41E−08
	Noncellulosomal	Clocel_0912		GH5	4.16	2.10E−09			3.77	1.16E−08	2.89	5.02E−08
		Clocel_1430		GH31	1.3	7.07E−06			1.98	4.23E−07	2.44	5.02E−08
		Clocel_2595		GH43	5.5	4.91E−09			5.08	1.67E−08	4.62	4.41E−08
Pectin	Noncellulosomal	Clocel_0873		PL9	2.32	3.64E−07	1.49	1.92E−05			3.91	6.07E−09
		Clocel_1172		PL1	2.93	3.65E−06	2.01	1.08E−04			3.79	4.21E−07
		Clocel_1243		NA	2.13	1.14E−07	1.64	1.05E−06			1.03	3.79E−05
		Clocel_2256		GH105	3.23	1.46E−06	3.47	1.24E−06			3.92	3.22E−07
		Clocel_2507		GH42	3.19	1.75E−05	2.6	1.43E−04			2.64	1.23E−04
		Clocel_3380		PL9	3.24	1.14E−07	2.76	4.23E−07			3.62	3.44E−08
PASC	Noncellulosomal	Clocel_0032		GH94	1.02	1.65E−05	1.15	8.02E−06	1.48	1.22E−06
		Clocel_2606		GH5, CBM46	2.11	1.68E−08	1.55	1.47E−07	2.33	6.07E−09
		Clocel_3657		GH43	1.92	1.69E−05	1.03	2.08E−03	1.67	1.13E−04
		Clocel_4124		GH26, CBM59	1.56	6.88E−06	2.36	2.57E−07	2.54	2.39E−07

^aSee http://www.cazy.org/. NA, not annotated.

^bP values were adjusted for multiple testing with the Benjamini-Hochberg method.

Substrate-specific extracellular proteins for cellobiose.

Eleven cellobiose-specific proteins were identified; 3 were cellulosomal proteins, and 8 were noncellulosomal proteins (Table 2). Five of the 11 cellobiose-specific proteins were endoglucanases: Clocel_1624 (EngY) (26), Clocel_2815 (EngN) (27), Clocel_1478 (EngO) (28), Clocel_3242 (EngD) (29), and Clocel_3662 (EngF) (30). These proteins have been previously reported; however, they are not enzymes known to be involved in cellobiose degradation. Therefore, cellobiose might be a trigger for the detection of extracellular cellulose. Among the cellobiose-specific proteins, Clocel_3662 (EngF) showed the highest fold change (Fig. 5 and Table 2).

Substrate-specific extracellular proteins for xylan.

Xylan-specific proteins comprised 2 cellulosomal proteins and 3 noncellulosomal proteins (Table 2). Among the identified cellulosomal proteins, Clocel_1432 was classified as a member of the GH5 family by CAZy. Generally, GH5 enzymes have various substrate specificities. Because endo-β-1,4-xylanase activity has been reported for GH5 enzymes, this protein, which was produced in xylan, could have endo-β-1,4-xylanase activity. In the previous study, Clocel_2295 (XynA) was classified as a member of the GH11 and carbohydrate esterase 4 (CE4) families, and endoxylanase activity and deacetylase activity were confirmed (31). The primary degradation products from xylan by this enzyme were xylobiose and xylotriose. Among the noncellulosomal proteins, Clocel_0912, classified as a GH5 enzyme, is thought to have endo-β-1,4-xylanase activity. Clocel_2595, classified as a GH43 enzyme, is thought to have β-xylosidase activity and to produce xylose by degrading xylan. Clocel_1430, classified as a GH31 enzyme, is thought to have α-xylosidase activity. In short, C. cellulovorans can first degrade xylan into xylobiose or xylotriose through the activity of cellulosomal XynA or other endo-β-1,4-xylanases and then degrade xylobiose to xylose through the activity of Clocel_2595 and Clocel_1430. The supernatant from cultures with xylan as a substrate contained xylose, but xylobiose or xylotriose was not detected (data not shown), consistent with their conversion to xylose.

Substrate-specific extracellular proteins for pectin.

Pectin-specific proteins comprised 6 noncellulosomal proteins (Table 2). Clocel_2507, classified as a GH42 enzyme, is thought to have β-galactosidase activity. Clocel_0873 and Clocel_3380 were classified as PL9 enzymes that show activity to degrade pectate polymer into disaccharide. Clocel_2256, classified as a GH105 enzyme, is thought to have unsaturated rhamnogalacturonyl hydrolase activity. Clocel_1172 was classified as a PL1. In summary, C. cellulovorans can first degrade pectate into pectate disaccharide through the activity of pectate lyases and then degrade pectate disaccharide into monomers through the activity of Clocel_2507, which is similar to xylan degradation.

Substrate-specific extracellular proteins for PASC.

PASC-specific proteins comprised 4 noncellulosomal proteins (Table 2). Clocel_4124 was especially enriched, and a domain of Clocel_4124 was classified as a CBM59 domain, which was reported to have the ability to bind to cellulose. Clocel_2606, classified as a GH5 and a CBM46 enzyme, is thought to have cellulose-binding ability and to be able to degrade cellulose. Clocel_0032, classified as a GH94, is expected to have cellodextrin phosphorylase activity.

DISCUSSION

In this study, we identified a strategy by which C. cellulovorans degrades several different substrates via the optimization of cellulosomal and noncellulosomal proteins through quantitative exoproteome analysis, and we identified the importance of noncellulosomal proteins for the optimal degradation of various biomasses.

We focused on 79 cellulosomal and noncellulosomal proteins, out of the 639 extracellular proteins identified. The 79 proteins could be divided into two types: (i) basal proteins that were produced rather abundantly and consistently irrespective of the type of growth substrates and (ii) substrate-specific extracellular proteins that were differentially produced with one specific substrate. First, we considered the abundances of cellulosomal and noncellulosomal proteins in the exoproteome. Although it is necessary to perform an exact quantification using antibody and a synthetic compound to precisely compare the amounts of each protein, the spectral counts, shown in Table 1 and in Table S2 in the supplemental material, correlate with the abundance (32). The comparison of spectral counts of cellulosomal and noncellulosomal proteins (Table 1) showed that the cellulosomal proteins were more abundant, and therefore, the highly abundant cellulosomal proteins might play a major role in the degradation of biomass.

The analysis of substrate-specific extracellular proteins showed that 81% (21 proteins) of 26 substrate-specific extracellular proteins were noncellulosomal proteins (Fig. 3 and Table 2). C. cellulovorans has a larger number of noncellulosomal protein genes than the other cellulosome-forming clostridia (10), judging from the absence of dockerin domains and the presence of N-terminal signal peptides. We confirmed that noncellulosomal proteins encoded in the genome were actually produced at the protein level. In previous reports, quantitative changes in cellulosomal proteins were generally thought to be important, and the cellulosome was purified in most other proteomic studies of the cellulosome (23, 33). However, our results indicate the possibility that most of the cellulosomal proteins, including Clocel_2824, Clocel_2823, and Clocel_3359, were basal proteins and that noncellulosomal proteins were produced with drastic change for the optimized degradation of various biomasses as substrate-specific extracellular proteins, which provides new insights into the importance of noncellulosomal proteins. Interestingly, in a recent study of C. cellulolyticum (34), transcriptomic analysis revealed that an “accessory” set of CAZymes specifically expressed for each of the noncellulose substrates consisted mostly of noncellulosomal protein-encoding genes, which supports the results in this study.

There has been a study that demonstrated the synergistic activity of the cellulosomes and noncellulosomal proteins (14) by comparing the sum of each activity of cellulosomes and noncellulosomal proteins and the measured activity of mixtures of cellulosomes and noncellulosomal proteins. Furthermore, Ding et al. (35) compared the mechanisms by which noncomplexed fungal secreted cellulases and multienzyme complexes (cellulosomes) deconstruct plant cell walls. Whereas cellulosomes peeled off individual cellulose microfibrils from the cell wall surface, the fungal secreted cellulases penetrated inside the cellulose microfibril network and dissolved the entire wall in a uniform manner, which resulted in faster degradation of cellulose compared to that by cellulosomes (Fig. 6). Not only the proximity effect of cellulosomes but also the penetrating effect of noncomplexed secreted cellulases was important, which supports the significance of noncellulosomal proteins. Noncellulosomal proteins are noncomplexed proteins like fungal secreted cellulases (Fig. 6). The penetrating effects of noncellulosomal proteins could explain why noncellulosomal proteins accounted for such a large fraction of the substrate-specific extracellular proteins and how cellulosomes and noncellulosomal proteins show synergistic activity.

Fig 6.

Schematic model of the process of biomass degradation by cellulosomal and noncellulosomal proteins secreted by C. cellulovorans.

In conclusion, this study has quantitatively characterized the exoproteome profiles of C. cellulovorans using LC-MS/MS with a long monolithic silica capillary column. This study reports the whole exoproteome of C. cellulovorans, including cellulosomal and noncellulosomal proteins, from cell culture supernatant without a purification and prefractionation process. We selected 79 cellulosomal and noncellulosomal proteins out of 639 identified proteins and found that the abundance of cellulosomal proteins was higher than that of noncellulosomal proteins, based on spectral counts of the proteins. Subsequently, we further identified 26 substrate-specific extracellular proteins (Fig. 3). The substrate-specific extracellular proteins for each of the 4 substrates consisted of proteins expected to degrade the particular substrate, e.g., xylanases for xylan and pectin lyase for pectin. Surprisingly, among the substrate-specific extracellular proteins, the number of noncellulosomal proteins identified was almost 4 times the number of cellulosomal proteins. The results indicate that cellulosomal proteins, produced in large amounts, play a major role in substrate degradation and that noncellulosomal proteins, rather specifically produced, optimize the components of the secretome for the efficient degradation of various components of a biomass. We propose that noncellulosomal proteins penetrate the surface of a biomass and dissolve the entire cell wall in a uniform manner and that cellulosomes vigorously degrade a biomass from its surface; noncellulosomal proteins and cellulosomal proteins could thus work synergistically (Fig. 6).