The cellulosomes from Clostridium cellulolyticum Identification of new components and synergies between complexes Imen Fendri 1 , Chantal Tardif 1,2 , Henri-Pierre Fierobe 1 , Sabrina Lignon 3 , Odile Valette 1 , Sandrine Page ` s 1,2 and Ste ´ phanie Perret 1,2 1 Laboratoire de Chimie Bacte ´ rienne, CNRS, UPR9043, IMM, Marseille, France 2 Universite ´ Aix Marseille, France 3 Centre de microse ´ quencage et d’analyse prote ´ omique, IMM, Marseille, France Biomass from plant cell walls contains large quantities of structural polysaccharides. Cellulose, the most abundant polysaccharide, is a linear glucose polymer forming fibrils with a regular crystalline arrangement [1–3]. In plant cell walls, cellulose fibrils are sur- rounded by a complex matrix of polysaccharides such as hemicellulose and pectin [4], which make plant cellulose resistant to enzymatic hydrolysis. Some microorganisms secrete diverse cellulases, hemicellu- lases (xylanases, mannanases, etc.) and pectinases that have various and complementary modes of action (endo, exo and processive) [5]. These plant cell-wall- degrading enzymes, which include glycoside hydrolases (GH), polysaccharide lyases and carbohydrate ester- ases, have been classified into families based on their sequence homologies (Carbohydrate Active Enzyme Database; http://www.cazy.org) [6]. In cellulose-rich anaerobic biotopes, bacteria such as Ruminococ- cus flavefaciens [7,8], Bacteroides cellulosolvens [9], Clostridium cellulolyticum [10], Clostridium thermocel- lum [11], Clostridium cellulovorans [12] and Clostrid- ium papyrosolvens [13] secrete multienzyme complexes called cellulosomes which degrade plant cell walls effi- ciently. In general, cellulosomes are composed of a scaffolding protein devoid of enzymatic activity which binds the complexes to the substrate via its carbo- hydrate-binding module (CBM). This protein contains several cohesin modules that serve as anchoring points Keywords cellulosome; Clostridium cellulolyticum; diversity; new components; synergy Correspondence S. Perret, Laboratoire de Chimie Bacte ´ rienne, CNRS, UPR9043, 31 chemin Joseph Aiguier 13009, Marseille, France Fax: +33 4 91 71 33 21 Tel: +33 4 91 16 43 40 E-mail: perret@ifr88.cnrs-mrs.fr (Received 18 January 2009, revised 24 March 2009, accepted 27 March 2009) doi:10.1111/j.1742-4658.2009.07025.x Cellulosomes produced by Clostridium cellulolyticum grown on cellulose were purified and separated using anion-exchange chromatography. SDS ⁄ PAGE analysis of six fractions showed variations in their celluloso- mal protein composition. Hydrolytic activity on carboxymethyl cellulose, xylan, crystalline cellulose and hatched straw differed from one fraction to another. Fraction F1 showed a high level of activity on xylan, whereas fractions F5 and F6 were most active on crystalline cellulose and carb- oxymethyl cellulose, respectively. Several cellulosomal components specific to fractions F1, F5 and F6 were investigated using MS analysis. Several hemicellulases were identified, including three xylanases in F1, and several cellulases belonging to glycoside hydrolase families 9 and 5 and, a cystein protease inhibitor were identified in F5 and F6. Synergies were observed when two or three fractions were combined. A mixture containing fractions F1, F3 and F6 showed the most divergent cellulosomal composition, the most synergistic effects and the highest level of activity on straw (the most heterogeneous substrate tested). These findings show that on complex sub- strates such as straw, synergies occur between differently composed cellulo- somes and the degradation efficiency of the cellulosomes is correlated with their enzyme diversity. Abbreviations CBM, carbohydrate-binding module; CipC, cellulosome-integrating protein C; CMC, carboxymethyl cellulose; GH, glycoside hydrolases. 3076 FEBS Journal 276 (2009) 3076–3086 ª 2009 The Authors Journal compilation ª 2009 FEBS for the enzymes via a strong interaction with enzyme- borne dockerin modules. The cellulosomes produced by C. cellulolyticum grown on cellulose contain  30 dockerin-containing proteins [14]. The majority of these proteins are GHs belonging to families 2, 5, 8, 9, 10, 11, 18, 26, 27, 44, 48, 53, 62, 74 and 95. In addition, 62 ORFs that potentially encode dockerin-containing proteins have been found in the genome sequence of C. cellulolyticum (http://www. ncbi.nlm.nih.gov; GI:220927459), (J. C. Blouzard, per- sonal communication). Twelve genes were found to be gathered in a large operon called the cip-cel cluster, beginning with the gene encoding the scaffolding pro- tein named cellulosome-integrating protein C (CipC), followed by genes that code for the major cellulosomal cellulase Cel48F and nine other dockerin-containing enzymes [15]. Cellulases encoded by the cip-cel cluster are essential for the formation of efficient cellulosomes to degrade crystalline cellulose [16], in particular the processsive cellulase Cel48F [17]. In C. cellulolyticum, the scaffolding protein contains eight cohesin modules that potentially bind to 62 dock- erin-bearing proteins. Previous studies have suggested that any CipC cohesin can bind to any enzyme docker- in: Cel5A binds to the most divergent cohesins with similar affinities [18] and cohesin 1 shows a similar affinity for Cel5A and Cel48F [19]. In addition, over- production of a minor cellulosomal enzyme, the man- nanase Man5K, resulted in mannanase-enriched cellulosomes [20]. The data strongly suggest that the composition of C. cellulolyticum cellulosomes is hetero- geneous and may depend on the relative amounts of dockerin-containing enzymes available. The hydrolytic efficiency of cellulosomes has also been studied in mini-cellulosomes assembled in vitro. These mini-cellulosomes had a strictly controlled enzyme composition and contained two or three engi- neered cellulases [21,22]. Enzyme binding to scaffoldin was found to enhance the activity of the enzymatic components, particularly on recalcitrant substrates. This enhancement was attributed to the physical prox- imity of the enzymes in the mini-cellulosomes and to cellulose targeting of the complexes via the CBM of the mini-scaffoldin [21]. The most active mini-cellulo- some on microcrystalline cellulose was composed of the processive cellulase Cel48F combined with endo- glucanase Cel9G. Adding the C. thermocellum bifunc- tional esterase ⁄ xylanase Xyn10Z to this cellulase pair yielded the most active mini-cellulosome on hatched straw [22]. Compared with naturally occurring cellulo- somes, however, the most active mini-cellulosomes are fivefold less active on crystalline cellulose and 3.5-fold less active on straw. Additional factors present in naturally occurring cellulosomes may therefore account for their high efficiency. Cellulosomes produced by C. papyrosolvens and C. cellulovorans grown on cellulose have been split into several peaks using anion-exchange chromatography [13,23]. The subpopulations had diverse enzymatic compositions and patterns of activity. However, puta- tive synergistic activities between several subpopula- tions were not examined. In this study, we separated cellulosomes produced by C. cellulolyticum. The acti- vity and composition of the complexes present in each fraction were analysed to identify new active compo- nents and ⁄ or an efficient association of components. The possible occurrence of synergies between cellulo- somal fractions which might account for the efficiency of the cellulosomes was investigated. Results Fractionation of cellulosomes To separate the various cellulosomes of C. cellulolyti- cum, we first extracted cellulose-bound proteins from the residual cellulose in a 6-day culture. Cellulosomes (500–900 kDa) were purified using gel-filtration chro- matography. The cellulosomal fraction was subse- quently subjected to anion-exchange chromatography. The elution profile showed that the cellulosomes were eluted in a single peak with a long tail (Fig. 1). This elution profile was systematically obtained with cellulo- somes originating from several clostridial cultures on cellulose. Using different NaCl gradients or performing elution with a pH gradient also yielded a single peak (data not shown). Cellulosome composition was analy- sed from the beginning to the end of the large peak; the peak was arbitrarily divided into six fractions numbered F1–F6 (Fig. 1) and the protein composition Fig. 1. Anion-exchange chromatography of the cellulosomal fraction purified by gel filtration. Three milligrams of protein were loaded onto the column. F1–F6 are the arbitrarily separated fractions. The grey line gives the continuous NaCl gradient. I. Fendri et al. Diversity of cellulosomes and their synergies FEBS Journal 276 (2009) 3076–3086 ª 2009 The Authors Journal compilation ª 2009 FEBS 3077 and enzymatic properties of the cellulosomes present in each fraction were analysed. Protein analyses of the fractionated cellulosomes The six fractions, separated as described above, were first analysed using Native PAGE. In all fractions, a single major diffuse band was observed, which showed that the proteins present in all the fractions were in a ‘cellulosome state’ [20] and that anion-exchange chro- matography did not dissociate the complexes (data not shown). The subunit composition of these complexes was therefore analysed further by SDS ⁄ PAGE (Fig. 2). A control sample (C) was formed by pooling the elu- tion fractions (F1–F6) corresponding to the entire peak obtained using anion-exchange chromatography. In this control sample, the proportions were those of naturally occurring cellulosomes and the sample was subjected to the same chromatographic procedures as each of the fractions analysed separately. Each of the fractions obtained by anion-exchange chromatography showed numerous proteins, most of which had molecu- lar masses in the range 30–160 kDa. As expected, the scaffolding protein CipC (160 kDa) was detected as a major protein in all the fractions. The protein composition of the fractions was found to differ, particularly F1 and F6 which corresponded to the extremities of the peak (Fig. 2). In each fraction, Cel48F and Cel9E were found to be abundant. How- ever, the distribution patterns of the four cellulases Cel5A [24], Cel9G [25], Cel8C [26] and Cel9M [27] were quite different (Fig. 3). Cel5A and Cel5M were present almost exclusively at the end of the peak (in fractions F4–F6), whereas Cel8C was detected in only the first two fractions. Cel9G was present in all the fractions except the first. A complementary analysis was then carried out using silver-stained SDS ⁄ PAGE. The components showing the greatest variation in rela- tive amounts were numbered 1–14 (Fig. 2). Fraction 1 contained high amounts of proteins 6, 9, 12 and 14, whereas proteins 1, 2, 3, 4, 5, 7, 8, 10, 11 and 13 were present in fractions F5 and F6 but absent or barely detectable in F1–F4. The middle fractions, F3 and F4, which account for most of the complexes in naturally occurring proportions, were found to have a fairly sim- ilar composition, midway between those of F1 and F6. As expected, the cellulosomes produced during a 6-day period of growth on cellulose showed considerable heterogeneity and were partly separated using anion- exchange chromatography. Enzymatic properties of the various fractions The activities of the six fractions and the control sample were compared on noncrystalline substrates such as carboxymethyl cellulose (CMC) and xylan, Fig. 2. Composition of the cellulosome fractions (F1–F6). Five micrograms of protein were separated on a 10% SDS ⁄ PAGE and silver stained. C, control sample containing the unfractionated mix- ture of cellulosomes; F1–F6, fractions separated by anion-exchange chromatography. Major components CipC, Cel9E and Cel48F are indicated; bands analysed using MS methods are numbered; the asterisks indicate a band containing a nonsecreted protein identified by ion-trap MS ⁄ MS as a ketol-acid reductoisomerase, which is not a cellulosomal component (data not shown). Fig. 3. Identification of several components in cellulosomes from fractions F1–F6. Proteins (5 lg) were separated on 10% SDS ⁄ PAGE, transferred to a polyvinylidene fluoride membrane and probed with anti-CelA, anti-CelC, anti-CelM and anti-CelG serum. C, control sample corresponding to the unfractionated mixture of cellulosomes; F1–F6, fractions separated by anion-exchange chromatography. Diversity of cellulosomes and their synergies I. Fendri et al. 3078 FEBS Journal 276 (2009) 3076–3086 ª 2009 The Authors Journal compilation ª 2009 FEBS microcrystalline cellulose (Avicel) and hatched straw, a heterogeneous natural substrate. As shown in Fig. 4, fractions F1–F6 showed different patterns of activity. On CMC, the most active fraction was F6, which was 45% more active than F5, and 70–90% more active than the control sample and fractions F1–F4. On xylan, rather weak activity was measured with the con- trol sample and all the fractions, except for F1, which was found to be approximately fivefold more active than the others. The cellulosomes present in fractions F1 and F6 therefore have the most efficient enzymes for degrading xylan and CMC, respectively. On straw, and to a lesser extent on Avicel (Fig. 4), the differences in activity between the fractions were less pronounced than on CMC and xylan. On straw, all fractions showed a substantial level of activity that was never less than half that of the most active frac- tion, F1. On Avicel, F5 was the most efficient fraction, and the least active fractions, F1 and F2, showed less than half the activity of F5. All in all, these results indicate that the differences in the protein composition of the cellulosomes are related to different enzymatic properties. Identification of new cellulosomal components MS analysis was performed to identify the specific components of fractions F1, F5 and F6 (Table 1). Complete MS data such as the spectra and the corre- sponding annotation table can be found in Figs S1– S11 and Table S1. Two of the four components, which were found to be abundant in F1, were identified as xylanases belonging to GH family 10 (protein 12), named Xyn10A [14], and to GH family 11 (protein 14), renamed Xyn11B. In addition a hypothetical xylanase (protein 9) was detected. The catalytic domain of this protein showed 29% identity with the xylanase XynA from Erwinia chrysanthemi (accession no. AAB53151.1) [28,29]. The latter enzyme contains a GH catalytic domain that has been reported to be intermediate between families 5 and 30 [29]. The abun- dance of proteins 9, 12 and 14 in fraction F1 is consis- tent with the high xylanase activity seen in this fraction. However, GH10 (protein 12) is also present in noticeable quantities in the other five fractions. All the proteins present in F5 were also present in F6. Among these, we detected protein 2 (Cel9P) [14] and proteins 5a (Cel9G) [25] and 5b (Cel9H) [28]. Cel9P and Cel9H show the same modular organization as the endoglucanase Cel9G (GH9-CBM3c-Doc) char- acterized previously [25]. Although Cel9P and Cel9H have not yet been characterized, they are expected to show enzymatic properties similar to those of Cel9G. Four enzymes belonging to the GH5 family were also identified: proteins 7a and 11 correspond to the endo- cellulases Cel5D [30] and Cel5A [31], respectively, and protein 8 corresponds to the carboxymethyl cellulase Cel5N [14], and protein 4, in which the GH5 catalytic domain shows 33% identity with a mannanase from Bacillus circulans (accession no. BAA25878.1) [31]. Protein 7b was identified as the mannanase Man26A [14]. Lastly, protein 13 was identified as an N-terminal dockerin-borne chagasin domain. Chagasin belongs to Fig. 4. Enzymatic activities of the cellulosome fractions on various substrates. Specific activities were measured at 37 °C after 30 min on 0.8% CMC and xylan, and after 24 h on 0.35% microcrystalline cellulose Avicel and hatched straw at final protein concentrations of 2, 3, 20 and 6 lgÆmL )1 , respectively. I. Fendri et al. Diversity of cellulosomes and their synergies FEBS Journal 276 (2009) 3076–3086 ª 2009 The Authors Journal compilation ª 2009 FEBS 3079 a family of cystein protease inhibitors found in lower eukaryotes and prokaryotes [32]. The specific proteins in F6 either show a GH9- CBM3c modular organization (protein 1 was renamed Cel9R and protein 3 was identified as Cel9J) or belong to the GH18 family (protein 11). Protein 11 is a puta- tive chitinase in which the catalytic module shows 43% identity with a chitinase from Bacillus pumilus (acces- sion no. ABI15082.1) and 39% identity with Chi18A found in C. thermocellum cellulosomes [33]. Synergistic activities between cellulosomes from different fractions The possibility that synergies between various fractions of cellulosomes might account for the efficiency of the cellulosome mixture was investigated. It was previously established that the association of an endo-processive cellulase active on crystalline cellulose with an endo- cellulase active on CMC leads to the most efficient in vitro reconstituted mini-cellulosomes on microcrys- talline cellulose [19]. First, we studied the hydrolysis of Avicel, combining the most complementary fractions F5 and F6, which are most active on Avicel and CMC, respectively. We further studied the synergies combining these fractions with F3, which accounts for most of the cellulosomes. No synergies were measured with the F3 ⁄ F5 pair (Fig. 5). However, the F5 ⁄ F6 and F3 ⁄ F6 pairs showed synergies of 1.2 and 1.3, respec- tively, and a synergy of 1.2 was also obtained with the combination F3 ⁄ F5 ⁄ F6. On straw, the activity of the control sample was higher than that of any of the fractions (Fig. 4). This indicates that on this natural substrate, this combina- tion of different fractions results in synergistic activity. Because straw is a complex substrate composed mostly of cellulose (40% w ⁄ w) and hemicelluloses (15% w ⁄ w), we analysed its degradation using the following combi- nation of fractions showing complementary activities: the xylanase F1 fraction combined with either the Table 1. Identification of specific components detected in fractions F1, F5 and F6 using MS analysis. All identifications were based on pep- tide mass fingerprint analyses using the MALDI-TOF technique, except for the proteins 6, 7a and 7b, and 10 which were identified using the MS ⁄ MS technique. The modular structure of new proteins was determined by performing PFAM and BLAST analyses. S, signal sequence; GH, glycoside hydrolase; CBM, carbohydrate binding module, GH and CBM numbers are those of the carbohydrate active enzyme database classification (http://www.cazy.org); Doc, dockerin domain; Ig, immunoglobulin-like domain of cellulase; X2, Ig-like module of unknown func- tion; M r , theoretical molecular mass of the mature protein. The cleavage site was determined using http://www.cbsdtu.dk/services/SignalP/; Cov, percentage of amino acid coverage in the matched proteins; M pep , the number of unique matched peptides; U pep , the number of unmatched peptides in the MALDI-TOF experiments. The function of new proteins is based on the GH family of the catalytic module and the modular organization of the protein, or on the identity of the catalytic domain with another characterized protein (see text). Protein GI number a Modular structure M r (kDa) M pep ⁄ U pep Cov (%) Score Protein name and ⁄ or description Reference F1 6 220928204 S-Ig-GH9-doc 66.1 6 15.9 60.2 b Cellulase Cel9S This study 9 220928101 S-GH5 ⁄ GH30-doc 55.6 12 ⁄ 27 19.0 89.0 c Putative xylanase This study 12 110588916 S-GH10-doc 44.3 17 ⁄ 37 36.0 135.0 c Xylanase Xyn10A 9 14 220928199 S-GH11-doc 29.6 5 ⁄ 4 19.0 71.0 c Xylanase Xyn11B This study F5 ⁄ F6 2 110588925 S-GH9-CBM3-doc 83.3 9 ⁄ 36 15.0 72.0 c Cellulase Cel9P 9 4 220927835 S-GH5-CBM32-X2-X2-Doc 78.6 9 ⁄ 25 14.0 77.0 c Putative mannanase This study 5a 585234 S-GH9-CBM3-doc 76.1 10 ⁄ 48 15.0 5.4 · 10 5d Cellulase Cel9G 20 5b 12007365 S-GH9-CBM3-doc 78.7 11 ⁄ 47 14.0 5.1 · 10 3d Cellulase Cel9H 22 7a 121824 S-GH5-doc 63.4 5 9.8 50.2 b Cellulase Cel5D 25 7b 110588924 S-GH26-doc 61.8 2 10.5 20.2 b Mannanase Man26A 9 8 220928189 S-GH5-doc 56.6 6 ⁄ 7 10.0 68.0 c Cellulase Cel5N 9 11 121802 S-GH5-doc 50.7 29 ⁄ 39 45.0 249.0 c Cellulase Cel5A 19 13 220929230 S-doc-Chagasin 31.0 5 ⁄ 18 16.0 60.0 c Unknown function This study F6 1 220929070 S-GH9-CBM3-doc 102.3 12 ⁄ 23 13.0 74.0 c Cellulase Cel9R This study 3 220928185 S-GH9-CBM3-doc 81.3 23 ⁄ 28 26.0 165.0 c Cellulase Cel9J 22 10 220928973 S-GH18-doc 51.1 4 9.5 40.2 b Putative chitinase This study a Accession numbers of new components are those of the newly released complete genome (http://www.ncbi.nlm.nih.gov; GI:220927459). b Scores obtained using using BioworksBrowser search engine (MS ⁄ MS data). c Scores obtained using MASCOT search engine (MALDI- TOF data). d Scores obtained using MS-Fit (MALDI-TOF data). For this latter analysis the top nonhomologous protein shows a score of 94.8. Diversity of cellulosomes and their synergies I. Fendri et al. 3080 FEBS Journal 276 (2009) 3076–3086 ª 2009 The Authors Journal compilation ª 2009 FEBS Avicelase F5 and ⁄ or the carboxymethyl cellulase F6 fractions, with and without fraction F3, which accounts for most of the cellulosomes. The F1 ⁄ F5 and F1 ⁄ F6 pairs did not exhibit important synergies (1.1 and 1.16, respectively) and released 30% fewer soluble sugars than the control sample. However, the combi- nation of fractions F1 ⁄ F5 ⁄ F6 induced greater synergies (1.3) and released an amount of soluble sugars similar to that seen with the control. The highest synergy (1.4) was measured when fractions F1, F3 and F6 were combined, which resulted in a larger quantity of released soluble sugars than with the control sample. Each individual fraction was therefore less active than the naturally occurring cellulosome mixture, but mixing fractions combining the most diverse cellulo- somes induces important synergies. Discussion Cellulosomes from C. papyrosolvens and C. cellulovo- rans were separated using anion-exchange chromatog- raphy, giving seven and four elution peaks, respectively. These different cellulosome subpopula- tions have distinct protein compositions and patterns of activity [13,23]. It has been suggested that the pres- ence of several well-separated peaks in the case of cell- ulosomes from C. cellulovorans may be partly due to the existence of various categories of cohesins and dockerins which determine the composition of the cell- ulosomes [23,34]. Despite the enzymatic diversity of the cellulosomes from C. cellulolyticum (there are 62 ORFs encoding dockerin-containing protein versus 8 enzymatic units per cellulosome), anion-exchange chro- matography gave a single peak followed by a long tail. This suggests a random assembly of enzymes on the scaffoldin, leading to a large number of enzyme combi- nations. In this study, a GH11 xylanase and a GH5 ⁄ 30 puta- tive xylanase were identified. In the genome sequence of C. cellulolyticum, only one gene encoding a cellulosomal GH11 was found. To date, GH11 modules have been found in modular bifunctional cellulosomal enzymes, such as XynA from C. cellulovorans (GH11-Doc- CE4) [35] and XynA from C. thermocellum strain ATCC27405 (GH11-CBM6-Doc-CE4) [33], or associ- ated with a CBM6 module such as in XynB (GH11- CBM6-Doc) from C. thermocellum strain F1 ⁄ YS [36]. In the C. cellulolyticum enzyme, the GH11 catalytic module had no such catalytic or CBM partner, which suggests that the catalytic behaviour of the enzyme may differ from that of previously described enzymes con- taining GH11. To date, no GH5 ⁄ GH30 enzymes have been found in C. cellulovorans cellulosomes, whereas a bifunctional GH30-a-l-arabinofuranosidase B has been detected in C. thermocellum cellulosomes [33]. The cata- lytic domain of C. cellulolyticum GH5⁄ GH30 shows 25% identity with the C. thermocellum GH30 module, 28% identity with the E. chrysanthemi XynA catalytic module and 24% identity with the B. subtilis XynC catalytic module. XynC has been characterized as an endoxylanase cleaving the methylglucuronoxylan chain in close proximity to a methylglucuronosyl-substituted xylose residue [37]. The functional role of C. cellulolyti- cum GH5 ⁄ GH30 enzyme remains to be identified. Interestingly, a nonenzymatic protein (protein 13) was detected in substantial quantities in the cellulo- somes. This dockerin-bearing chagasin (MEROPS peptidase database identification number I42) is a putative cystein protease C1A inhibitor (http://merops. sanger.ac.uk) [38]. A gene encoding a dockerin-con- Fig. 5. Synergies between cellulosomes. Light grey bars indicate activity measured for the fraction mixture at total protein concentra- tions of 20 and 6 lgÆmL )1 on crystalline cellulose (A) and straw (B), respectively. White bars indicate the theoretical sum of the activi- ties of the individual fractions measured independently (at half or one third of the protein concentration). The dark grey bars indicate the activity of the control. Synergy values are indicated on the light grey bars. I. Fendri et al. Diversity of cellulosomes and their synergies FEBS Journal 276 (2009) 3076–3086 ª 2009 The Authors Journal compilation ª 2009 FEBS 3081 taining protein related to cystein protease C1A was found (GI:220929842) in the genome sequence of C. cellulolyticum. This cellulosomal chagasin ⁄ cystein protease system is reminiscent of the serpins ⁄ serine protease cellulosomal system reported in C. thermocel- lum [39,40]. A cellulosomal protease inhibitor ⁄ protease system may, therefore, be more widespread than expected and have a common and important function in cellulosome regulation, displacement from the cell surface, degradation and ⁄ or protection of the cellulo- somes [40]. All the fractions showed a substantial level of activity on crystalline cellulose. In previous studies, mini-cellulo- somes reconstituted in vitro, in which the endocellulase Cel9G (GH9-CBM3c-Doc) was combined with the processive enzyme Cel48F, were found to hydrolyse crystalline cellulose the most efficiently [21,22]. In this study, all the cellulosome fractions contained Cel48F and several GH9-CBM3c-Doc (Cel9P, Cel9G, Cel9H, Cel9J). This enzyme combination may be essential for efficient degradation of crystalline cellulose by the cellulosome. The most active fraction on Avicel (F5) was found to contain a small amount of Cel9J, but large amounts of Cel9P and Cel9G ⁄ Cel9H. Because the least active fractions, F3 and F4, contained large amounts of Cel9G ⁄ Cel9H, but lower amounts of Cel9P, it seems likely that Cel9P might contribute to the high level of activity on crystalline cellulose seen in F5. Although F6 contained all the proteins present in F5, it showed lower levels of activity on Avicel and higher levels on CMC than F5. This may be because of the presence of addi- tional proteins such as proteins 1 and 3 (which were identified as GH9-CBM3c-Doc enzymes) and protein 10 (which was identified as a chitinase), and ⁄ or to varia- tions in the enzyme ratios. On straw, the activity of the mini-cellulosomes containing Cel9G ⁄ Cel48F was greatly enhanced by adding the C. thermocellum bifunctional xylanase (XynZ) [22]. It is worth noting that all the naturally occurring cellulosome fractions studied here contained at least one xylanase (GH10 protein 12) and showed high levels of activity on straw. Individual fractions displayed less specific activity on straw than the control (consisting of a combination of all fractions in naturally occurring proportions). This indicated that synergies occur in the naturally occurring control mixture. The activity of each fraction on straw probably resulted from synergies between different cellulosomes. This explains why only low synergies were observed when two or three fractions were mixed. However, combinations of these cellulosome fractions in equal proportions, which differ from the naturally occurring proportions, resulted in levels of activity on Avicel and straw higher than those seen with the control mixture, highlighting cellulosome synergies. On straw, a mixture combining the most complementary fractions, i.e. the most active fractions on xylan (F1), Avicel (F5) and CMC (F6), showed lower levels of activity and synergy than a mixture consisting of F1, F3 and F6, which was the most diverse combination of cellulosomes. This finding strongly suggests that, on complex substrates, the diversity of the combined cellu- losomes has a greater impact on the final activity than do the enzymatic properties of the combined fractions. In a previous study,  30 dockerin-containing enzymes were detected by performing proteomic analyses on cellulosomes produced by C. cellulolyticum on cellulose [14]. The enzyme diversity they contain and their heter- ogeneous composition are inherent characteristics of cellulosomes. Our data suggest that these characteristics give rise to synergistic effects between diverse com- plexes, which may account for the great efficiency of plant cell-wall degradation processes. Experimental procedures Bacterial strain and cell culture conditions C. cellulolyticum ATCC35319 [41] was grown anaerobically at 32 °C on basal medium [42] supplemented with cellobi- ose (4 gÆL )1 ; Sigma-Aldrich, St Louis, MO, USA) or MN300 cellulose (5 gÆL )1 ; Serva, Heidelberg, Germany). Purification of the cellulose-adsorbed cellulolytic system from C. cellulolyticum C. cellulolyticum cultures were inoculated with a cellobiose culture at D 450 = 0.7, and grown in 800 mL of cellulose- supplemented basal medium for 6 days. The cell culture was filtered through a 3-lm pore size GF ⁄ D glass filter (Whatman, Maidstone, UK). The residual cellulose was subsequently washed with 50 and 12.5 mm Na 2 HPO 4 / NaH 2 PO 4 (pH 7.0). The cellulosome-containing fraction was eluted from the residual cellulose with water, dialysed and concentrated in 20 mm Tris ⁄ HCl buffer (pH 8.0), 150 mm NaCl and 2 mm CaCl 2 by ultrafiltration. Chromatography Liquid chromatography was performed at 4 °C using a fast protein purification liquid chromatography system (A ¨ kta Explorer Ô ; Amersham Biosciences, Uppsala, Sweden). Gel-filtration chromatography was performed using a HiLoad 26 ⁄ 60 Superdex 200 column (Amersham Bio- sciences) equilibrated with 20 mm Tris ⁄ HCl buffer (pH 8.0), 150 mm NaCl and 2 mm CaCl 2 . Fractions of interest were pooled and dialysed against 20 mm Tris ⁄ HCl (pH 8.0) and Diversity of cellulosomes and their synergies I. Fendri et al. 3082 FEBS Journal 276 (2009) 3076–3086 ª 2009 The Authors Journal compilation ª 2009 FEBS 2mm CaCl 2 buffer before loading into a Resource Q column (6 mL) (Amersham Biosciences) equilibrated with 20 mm Tris ⁄ HCl (pH 8.0) and 2 mm CaCl 2 buffer. Elution was per- formed with a linear NaCl gradient of 0–1 m, in the same buffer. Fractions were concentrated using microconcentators (30 kDa cut-off; Vivaspin, Vivasciences, Palaiseau, France). Protein concentration was determined as described by Lowry et al. [43], using bovine serum albumin as the standard. Enzyme activity Avicel microcrystalline cellulose (PH101; Fluka, Buchs, Switzerland), CMC (medium viscosity; Sigma, St Louis, MO, USA), oat spelt xylan (Sigma) and hatched straw (Valagro, Poitiers, France) were used as substrates. Hatched straw was prepared as described by Fierobe et al. [22]. Insoluble xylan was washed four times in distilled water and the concentration of the residual material was estimated from the dry weight. Enzymatic assays were per- formed in 20 mm Tris-maleate (pH 6.0) at 37 °C. A suitable amount of protein (see legend to Figs 5 and 6) was mixed with the substrate preparation at a final substrate concen- tration of 0.8% (CMC or xylan) or 0.35% (Avicel or straw). After incubating for 30 min (CMC and xylan) or 24 h (Avicel and straw), aliquots were analysed to deter- mine the soluble reducing sugar content using the method of Park & Johnson [44] with d-glucose as the standard. SDS ⁄ PAGE and western blot analysis SDS ⁄ PAGE was performed using Prosieve 50 gel solution (Lonza, Rockland, ME, USA). Native PAGE was per- formed with precast 4–15% polyacrylamide gradient gels using a Phast-System apparatus (Amersham Biosciences). Gels were either silver stained using the Plus one silver-stain- ing kit (Amersham Biosciences) or electrotransferred onto nitrocellulose BA83 membranes (Schleicher & Schuell, Das- sel, Germany). After saturation, membranes were probed with polyclonal rabbit antibodies raised against Cel9G, Cel5A, Cel8C or Cel9M. Antibodies were detected using an anti-rabbit horseradish peroxidase conjugate (Promega, Madison, WI, USA) and chemiluminescent substrate kit (ECL plus; GE Healthcare, Little Chalfont, UK). The same membrane was stripped and sequentially probed with several antibodies, in line with the manufacturer’s instructions. In-gel trypsin digestion of proteins MS analysis was performed to identify proteins that differed between the various cellulosome fractions. Proteins of interest were excised from the silver-stained gel and prepared on a robotic workstation (freedom EVO 100; TECAN, Ma ¨ nnedorf, Switzerland). The automated prepa- ration process included destaining steps (ProteoSliver TM ; Sigma), washing, reduction and alkylation, digestion by trypsin (proteomics grade; Sigma), extraction and drying of mixed peptides, as described previously [45]. MALDI-TOF MS analyses Complete experimental procedures of MALDI-TOF MS analysis are described in Doc. S1. Digested peptides were treated using MALDI-TOF Voyager DE-RP apparatus (Applied Biosystems, Foster City, CA, USA) in the positive reflectron mode. Contaminant peaks were removed prior to a peptide mass fingerprint search against the nonredundant NCBI database (20080210), restricted to ‘Other Firmicutes’ (445 464 sequences) using the freely available MASCOT search engine (http://www.matrixscience.com). Searches were performed using a maximum peptide mass tolerance of 150 p.p.m., one missing cleavage allowed, a fixed modifi- cation of cysteines by iodoacetamide (carbamidomethyl), a variable modification of methionines (oxidation) and N-term glutamine (pyro-glutamine). Proteins were taken to have been identified only when they had at least five matching peptides and scores > 60 (P < 0.05). When identification scores < 60 were obtained, we assessed their reliability using the search engine MS-FIT v4.27.2Basic (http://prospector.ucsf.edu). In the case of peptides matching multiple members of a protein family, the proteins selected were those with the largest number of matching peptides. When several proteins were identified with equal numbers of matching peptides we checked that they corresponded to the same gene product and selected the database entry that was the best annotated. Ion-trap MS ⁄ MS analyses Complete experimental procedures of ion-trap MS ⁄ MS anal- yses are described in Doc. S1. Samples which did not produce a sufficiently clear signal in the MS analyses were studied using 2D liquid chromatography in a tandem mass spectro- meter. Peptides were loaded onto a strong cation-exchange column and eluted in salt steps with an increasing ammonium acetate molarity, before being separated in a reversed-phase PicoFrit Ô column (New Objective, Woburn, MA, USA). An ion trap LCQ-DECA XP mass spectrometer (Thermo Finni- gan, Waltham, MA, USA) was used for the data acquisition. Maximum coverage identification was carried out using the big three program included in the data acquisition Xcali- bur Ô Finnigan proteomex 2.0 software program. Protein identification was performed using the Sequest (v28 rev12) algorithm in the bioworksbrowser 3.3 software program (Thermo Electron Corp., Waltham, MA, USA) using both the nonredundant NCBI database (20071113) (http://www. ncbi.nlm.nih.gov) and C. cellulolyticum extract containing 6641 entries. The following search parameters were adopted: two missed cleavage sites allowed, variable methionine I. Fendri et al. Diversity of cellulosomes and their synergies FEBS Journal 276 (2009) 3076–3086 ª 2009 The Authors Journal compilation ª 2009 FEBS 3083 oxidation, cysteine carbamidomethylation and no fixed modification, and 1.5 and 1.0 Da as the maximum precursor and fragment tolerance. Positive identification of peptides was assessed by a cross-correlation number (Xcorr) versus charge state, as follows: Xcorr > 1.5 for singly charged ions, Xcorr > 2.0 for doubly charged ions and Xcorr > 2.5 for triply charged ions, peptide probability was £ 5 · 10 )3 . Pro- tein identification required maximum coverage or at least two rank one unique peptides. Protein sequence analyses The amino acid sequences of the new proteins were com- pared with those in the NCBI sequence databases using the blast program [46]. Protein domain compositions were analysed using the PFAM database (http://pfam.sanger. ac.uk) [47]. Signal peptide position was determined using the server http://www.cbs.dtu.dk/services/SignalP [48]. Acknowledgements Imen Fendri received a doctoral fellowship from the Tunisian Ministry of Higher Education and Scientific Research. We are very grateful to Danielle Moinier and Re ´ gine Lebrun (Centre de microse ´ quencage et d’analyse prote ´ omique, IMM, Marseille, France) for performing the MS analysis. Financial support from the Marseille- Nice Ge ´ nopole and the ANR (contracts PNRB – HYPAB and ‘non the ´ matique BioH 2 ’) is acknowledged. We thank Jessica Blanc for correcting the English. The genomic sequence data were provided by the US Department of Energy’s Joint Genome Institute (http:// www.jgi.doe.gov). References 1 Nishiyama Y, Langan P & Chanzy H (2002) Crystal structure and hydrogen-bonding system in cellulose 1b from synchrotron X-ray and neutron fiber diffraction. J Am Chem Soc 124, 9074–9082. 2 Nishiyama Y, Sugiyama J, Chanzy H & Langan P (2003) Crystal structure and hydrogen bonding system in cellulose 1a from synchrotron X-ray and neutron fiber diffraction. J Am Chem Soc 125, 14300–14306. 3 Hanley SJ, Revol JF, Godbout L & Gray DG (1997) Atomic force microscopy and transmission electron microscopy of cellulose from Micrasterias dentriculata; evidence for a chiral helical microfibril twist. Cellulose 4, 209–220. 4 Mc Cann MC & Carpita NC (2008) Designing the deconstruction of plant cell walls. Curr Opin Plant Biol 11, 314–320. 5 Gilbert HJ, Sta ˚ lbrand H & Brumer H (2008) How the walls come crumbling down: recent structural biochemistry of plant polysaccharide degradation. Curr Opin Plant Biol 11, 338–348. 6 Coutinho PM & Henrissat B (1999) Carbohydrate- active enzymes: an integrated database approach. In Recent Advances in Carbohydrate Bioengineering (Gilbert HJ, Davies G, Henrissat B & Svensson B, eds), pp. 3–12. Royal Society of Chemistry, Cambridge. 7 Rincon MT, Ding SY, McCrae I, Martin JC, Aurilia V, Lamed R, Shoham Y, Bayer EA & Flint HJ (2003) Novel organization and divergent dockerin specificities in the cellulosome system of Ruminococcus flavefaciens. J Bacteriol 185, 703–713. 8 Rincon MT, Cepeljnik T, Martin JC, Lamed R, Barak Y, Bayer EA & Flint HJ (2005) Unconventional mode of attachment of the Ruminococcus flavefaciens cellulo- some to the cell surface. J Bacteriol 187, 7569–7578. 9 Ding SY, Bayer EA, Steiner D, Shoham Y & Lamed R (2000) A scaffoldin of the Bacteroides cellulosolvens cell- ulosome that contains 11 type II cohesins. J Bacteriol 182, 4915–4925. 10 Tardif C, Be ´ laı ¨ ch A, Fierobe HP, Page ` s S, de Philip P & Be ´ laı ¨ ch JP (2006) Clostridium cellulolyticum: cellulosomes and cellulolysis. In Cellulosome (Kataeva I & Hauppauge NY, eds), pp. 221–259. Nova Sciences, New York. 11 Bayer EA, Be ´ laı ¨ ch JP, Shoham Y & Lamed R (2004) The cellulosomes: multienzyme machines for degrada- tion of plant cell wall polysaccharides. Annu Rev Micro- biol 58 , 521–554. 12 Doi RH, Park JS, Liu CC, Malburg LM Jr, Tamaru Y, Ichiishi A & Ibrahim A (1998) Cellulosome and non cellulosomal cellulases of Clostridium cellulovorans. Extremophiles 2, 53–60. 13 Pohlschroder M, Leschine SB & Canale-Parola E (1994) Multicomplex cellulose–xylanase system of Clostrid- ium papyrosolvens C7. J Bacteriol 176, 70–76. 14 Blouzard JC, Bourgeois C, de Philip P, Valette O, Be ´ laı ¨ ch A, Tardif C, Be ´ laı ¨ ch JP & Page ` s S (2007) Enzyme diversity of the cellulolytic system produced by Clostridium cellulolyticum explored by two-dimensional analysis: identification of seven genes encoding new dockerin-containing proteins. J Bacteriol 189, 2300– 2309. 15 Maamar H, Abdou L, Boileau C, Valette O & Tardif C (2006) Transcriptional analysis of the cip-cel gene clus- ter from Clostridium cellulolyticum. J Bacteriol 188, 2614–2624. 16 Maamar H, Valette O, Fierobe HP, Be ´ laı ¨ ch A, Be ´ laı ¨ ch JP & Tardif C (2004) Cellulolysis is severely affected in Clostridium cellulolyticum strain cipCMut1. Mol Micro- biol 51 , 589–598. 17 Perret S, Maamar H, Be ´ laı ¨ ch JP & Tardif C (2004) Use of antisense RNA to modify the composition of cellulo- somes produced by Clostridium cellulolyticum. Mol Microbiol 51, 599–607. Diversity of cellulosomes and their synergies I. Fendri et al. 3084 FEBS Journal 276 (2009) 3076–3086 ª 2009 The Authors Journal compilation ª 2009 FEBS 18 Page ` sS,Be ´ laı ¨ ch A, Fierobe HP, Tardif C, Gaudin C & Be ´ laı ¨ ch JP (1999) Sequence analysis of scaffolding protein CipC and ORFXp, a new cohesin-containing protein in Clostridium cellulolyticum: comparison of various cohesin domains and subcellular localization of ORFXp. J Bacteriol 181, 1801–1810. 19 Fierobe HP, Bayer EA, Tardif C, Czjzek M, Mechaly A, Be ´ laı ¨ ch A, Lamed R, Shoham Y & Be ´ laı ¨ ch JP (2002) Degradation of cellulose substrates by cellulo- some chimeras. Substrate targeting versus proximity of enzyme components. J Biol Chem 277, 49621–49630. 20 Perret S, Be ´ laı ¨ ch A, Fierobe HP, Be ´ laı ¨ ch JP & Tardif C (2004) Towards designer cellulosomes in Clostridia: mannanase enrichment of the cellulosomes produced by Clostridium cellulolyticum. J Bacteriol 186, 6544–6552. 21 Fierobe HP, Mechaly A, Tardif C, Be ´ laı ¨ ch A, Lamed R, Shoham RY, Be ´ laı ¨ ch JP & Bayer EA (2001) Design and production of active cellulosome chimeras. Selective incorporation of dockerin-containing enzymes into defined functional complexes. J Biol Chem 276, 21257– 21261. 22 Fierobe HP, Mingardon F, Mechaly A, Be ´ laı ¨ ch A, Rincon MT, Page ` s S, Lamed R, Tardif C, Be ´ laı ¨ ch JP & Bayer EA (2005) Action of designer cellulosomes on homogeneous versus complex substrates: controlled incorporation of three distinct enzymes into a defined trifunctional scaffoldin. J Biol Chem 280, 16325–16334. 23 Han SO, Yukawa H, Inui M & Doi RH (2005) Effect of carbon source on the cellulosomal subpopulations of Clostridium cellulovorans. Microbiology 151, 1491–1497. 24 Fierobe HP, Gaudin C, Be ´ laı ¨ ch A, Loutfi M, Faure E, Bagnara C, Baty D & Be ´ laı ¨ ch JP (1991) Characteriza- tion of endoglucanase A from Clostridium cellulolyti- cum. J Bacteriol 173, 7956–7962. 25 Gal L, Gaudin C, Be ´ laı ¨ ch A, Page ` s S, Tardif C & Be ´ laı ¨ ch JP (1997) CelG from Clostridium cellulolyticum: a multidomain endoglucanase acting efficiently on crys- talline cellulose. J Bacteriol 179, 6595–6601. 26 Fierobe HP, Bagnara-Tardif C, Gaudin C, Guerlesquin F, Sauve P, Be ´ laı ¨ ch A & Be ´ laı ¨ ch JP (1993) Purification and characterization of endoglucanase C from Clostrid- ium cellulolyticum. Catalytic comparison with endoglu- canase A. Eur J Biochem 217, 557–565. 27 Be ´ laı ¨ ch A, Parsiegla G, Gal L, Villard C, Haser R & Be ´ laı ¨ ch JP (2002) Cel9M, a new family 9 cellulase of the Clostridium cellulolyticum cellulosome. J Bacteriol 184, 1378–1384. 28 Braun EJ & Rodrigues CA (1993) Purification and properties of an endoxylanase from a corn stalk rot strain of Erwinia chrysanthemi. Phytopathology 83, 332–338. 29 Keen NT, Boyd C & Henrissat B (1996) Cloning and characterization of a xylanase gene from corn strains of Erwinia chrysanthemi. Mol Plant Microbe Interact 9, 651–657. 30 Shima S, Igarashi Y & Kodama T (1993) Purification and properties of two truncated endoglucanases produced in Escherichia coli harbouring Clostridium cel- lulolyticum endoglucanase gene celCCD. Appl Microbiol Biotechnol 38, 750–754. 31 Yoshida S, Sako Y & Uchida A (1998) Cloning, sequence analysis, and expression in Escherichia coli of a gene coding for an enzyme from Bacillus circulans K-1 that degrades guar gum. Biosci Biotechnol Biochem 62, 514–520. 32 Sanderson SJ, Westrop GD, Scharfstein J, Mottram JC & Coombs GH (2003) Functional conservation of a natural cystein peptidase inhibitor in protozoan and bacterial pathogens. FEBS Lett 542, 12–16. 33 Gold ND & Martin VG (2007) Global view of the Clostridium thermocellum cellulosome revealed by quantitative proteomic analysis. J Bacteriol 189, 6787– 6795. 34 Park JS, Matano Y & Doi RH (2001) Cohesin–docker- in interactions of cellulosomal subunits of Clostrid- ium cellulovorans. J Bacteriol 183, 5431–5435. 35 Kosugi A, Murashima K & Doi RH (2002) Xylanase and acetyl xylan esterase activities of XynA, a key sub- unit of the Clostridium cellulovorans cellulosome for xylan degradation. Appl Environ Microbiol 68, 6399– 6402. 36 Hayashi H, Takehara M, Hattori T, Kimura T, Karita S, Sakka K & Ohmiya K (1999) Nucleotide sequences of two contiguous and highly homologous xylanase genes xynA and xynB and characterization of XynA from Clostridium thermocellum. Appl Microbiol Biotech- nol 51, 348–357. 37 St John FJ, Rice JD & Preston JF (2006) Characteriza- tion of XynC from Bacillus subtilis subsp. subtilis strain 168 and analysis of its role in depolymerisation of glucuronoxylan. J Bacteriol 188, 8617–8626. 38 Rawlings ND, Morton FR, Kok CY, Kong J & Barrett AJ (2008) MEROPS: the peptidase database. Nucleic Acids Res 36 (database issue), D320–D325. 39 Kang S, Barak Y, Lamed R, Bayer EA & Morrison M (2006) The functional repertoire of prokaryote cellulo- somes includes the serpin superfamily of serine protein- ase inhibitors. Mol Microbiol 60, 1344–1354. 40 Schwarz WH & Zverlov VV (2006) Protease inhibitors in bacteria: an emerging concept for the regulation of bacterial protein complexes? Mol Microbiol 60, 1323– 1326. 41 Petitdemange E, Caillet F, Giallo J & Gaudin C (1984) Clostridium cellulolyticum sp. nov, a cellulolytic, meso- philic species from decayed grass. Int J Sys Bacteriol 34, 155–159. 42 Giallo J, Gaudin C, Be ´ laı ¨ ch JP, Petitdemange E & Caillet-Mangin F (1983) Metabolism of glucose and cellobiose by cellulolytic mesophilic Clostridium sp. strain H10. Appl Environ Microbiol 45, 843–849. I. Fendri et al. Diversity of cellulosomes and their synergies FEBS Journal 276 (2009) 3076–3086 ª 2009 The Authors Journal compilation ª 2009 FEBS 3085 [...]... in the online version of this article Please note: Wiley-Blackwell is not responsible for the content or functionality of any supplementary materials supplied by the authors Any queries (other than missing material) should be directed to the corresponding author for the article Supporting information The following supplementary material is available: 3086 FEBS Journal 276 (2009) 3076–3086 ª 2009 The. .. MS spectrum of protein 5, fraction 5 ⁄ fraction 6 Fig S7 MS spectrum of protein 8, fraction 5 ⁄ fraction 6 Fig S8 MS spectrum of protein 11, fraction 5 ⁄ fraction 6 Fig S9 MS spectrum of protein 13, fraction 5 ⁄ fraction 6 Fig S10 MS spectrum of protein 1, fraction 6 Fig S11 MS spectrum of protein 3, fraction 6 Table S1 Masses and peptide assignments Doc S1 Complete experimental procedures of mass spectrometry...Diversity of cellulosomes and their synergies I Fendri et al 43 Lowry OH, Rosebrough NJ, Farr AL & Randall RJ (1951) Protein measurement with the Folin phenol reagent J Biol Chem 193, 265–275 44 Park JT & Johnson MJ (1949) A submicrodetermination of glucose J Biol Chem 181, 149–151 45 Brugna-Guiral M, Tron P, Nitschke W, Stetter... Guigliarelli B, Bruschi M & Guidici-Orticoni MT (2003) [NiFe] hydrogenases from the hyperthermophilic bacterium Aquifex aeolicus: properties function, and phylogenetics Extremophiles 7, 145–157 46 Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W & Lipman DJ (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs Nucleic Acids Res 25, 3389–3402 47 Finn... tools and services Nucleic Acids Res 34, 247–251 48 Dyrlov Bendtsen J, Nielsen H, von Heijne G & Brunak S (2004) Improved prediction of signal peptides: SignalP 3.0 J Mol Biol 340, 783–795 Fig S1 MS spectrum of protein 9, fraction 1 Fig S2 MS spectrum of protein 12, fraction 1 Fig S3 MS spectrum of protein 14, fraction 1 Fig S4 MS spectrum of protein 2, fraction 5 ⁄ fraction 6 Fig S5 MS spectrum of protein . The cellulosomes from Clostridium cellulolyticum Identification of new components and synergies between complexes Imen Fendri 1 ,. experiments. The function of new proteins is based on the GH family of the catalytic module and the modular organization of the protein, or on the identity of the