Chi18A, the Endochitinase in the Cellulosome of the Thermophilic, Cellulolytic Bacterium Clostridium thermocellum

Abstract

The chitinase gene chiA was identified on the Clostridium thermocellum genome downstream of the endoglucanase gene celA. It contains a catalytic module of glycosyl hydrolase family 18 and a cellulosomal dockerin module. Chi18A hydrolyzes aryl-acetyl-chito-oligosaccharides preferentially. In denaturing electrophoresis of purified cellulosomes, a single chitinase activity band was identified in zymograms and Western blots, indicating that Chi18A is the only chitinase in the cellulosome.

Clostridium thermocellum is a saccharolytic, thermophilic bacterium that occurs in hot springs and self-heated, rotting biomass. It takes part in the carbon cycle by mineralizing cellulosic biomass, following the aerobic cellulose degraders (like fungi and actinomycetes) in degradation and in competition with them. To hydrolyze its highly degradation-resistant substrate, it produces large multienzyme complexes, i.e., the cellulosomes, each consisting of various catalytic and noncatalytic modules, that are held together on the scaffoldin by protein-protein interaction via specific binding modules (18, 19, 20).

Despite its specialization for the hydrolysis of crystalline cellulose, the cellulosomes contain, in addition to multiple cellulases, hemicellulolytic enzymes like mixed-linkage glucanases and xylanases, an arabinoxylan-esterase, and a mannanase (4, 7, 9, 25). These components are believed to increase the accessibility of the crystalline cellulose, which is embedded in a network of hemicelluloses and other organic compounds.

We describe the identification and localization of a new cellulosome component, endochitinase ChiA. Clone pWS01 of a genomic DNA library from C. thermocellum type strain DSM1237 (ATCC 27405) expressed chitinase activity in addition to the cellulosomal endoglucanase celA (2, 14, 15). A 3,097-bp sequence between the restriction endonuclease recognition sites EcoRV and EcoRI of plasmid pWS01 was determined from both strands (Fig. 1). In addition to the C-terminal part of endoglucanase gene celA, the sequence contained two complete reading frames encoded on the same DNA strand, chiA and orfZ. The orfZ frame from bp 1833 to a TAG stop codon (bp 2770) showed similarity to cellulose binding modules. These are usually connected to a second, catalytic module. OrfZ, however, consists of a module of unknown function (amino acids [aa] 1 to 104), a so-called PTS box (aa 105 to 175), and a cellulose binding module with high similarity to carbohydrate binding module type IIIb. This structure is unusual and hitherto found only in the eucaryotic slime mold Dictyostelium discoideum with a carbohydrate binding module, type IIa (22).

FIG. 1.

Arrangement of genes in the chiA cluster and modular structure of the proteins. (A) The extension of clones, sequences, and genes is denoted. GenBank accession numbers are K03088 and Z68924 from C. thermocellum ATCC 27405 and AJ420770 from C. thermocellum F-7. The bar indicates 1 kb. (B) The modular structure of Chi18A is indicated by shading: LP, leader peptide; GHF 18, GHF18 module; and DD, dockerin module.

To exclude the possibility that DNA rearrangement occurred, e.g., during the cloning procedure (a possible reason for unusual structures), the integrity of the cloned DNA in the chiA-orfZ “intergenic” region and the orfZ region was verified by PCR on chromosomal DNA from C. thermocellum DSM1237 and the closely related C. thermocellum strain F-7 (strain VKMB2203 from the All-Russian Collection of Microorganisms, WDCM no. 342) using oligonucleotide primers chi1500n (GGCCATAAGT GATTTTCCGA TT-3′), chi2000n (CGGGAAAGTT CCGTCAGCAT-3′), and chi2500c (CCGTCCCTTG TAAACTCATT A-3′). The resulting amplificates showed the expected size and thus corroborated the location of the binding sites and the integrity of the cloned sequence in both strains (data not shown).

The amplified DNA from C. thermocellum strain F-7 was cloned and sequenced (Fig. 1A). Its sequence was identical to that derived from the type strain with the exception of seven bases: four exchanges or insertions and/or deletions in the intergenic region (136 bp) and three silent exchanges in third codon positions in the 224 N-terminal codons of orfZ. No difference was found in the C-terminal 45 codons of chiA. An overall mutation rate of less than 0.5% was observed. There was no rearrangement observed in the C-terminal region of chiA, the intergenic region, or orfZ, and the unusual structure of OrfZ was not the result of a singular event in the type strain.

A putative translation start for the chiA sequence was identified 145 bp downstream of celA and 6 bp downstream of a purine-rich Shine-Dalgarno sequence, located within an AT-rich region around bp 174 (GGAGG) and with sufficient base pairing capacity to the 16S rRNA of clostridia (24). The codon usage was similar to that of other cellulolytic genes of thermophilic clostridia. Palindromic structures resembling transcription terminators were detected downstream of celA (bp 66 to 110; ΔG = −79 kJ/mol), chiA (bp 1797 to 1833; ΔG = −75 kJ/mol) and orfZ (bp 2824 to 2868; ΔG = −115 kJ/mol). At least the terminator downstream of celA is active, precluding cotranscription of celA and chiA (3).

Start (ATG) and stop (TGA) codons for the chiA gene at bp 185 and 1631, respectively, mark a reading frame for a protein of 482 amino acid residues with a calculated M_r of 55,025. Three modules were detected in the amino acid sequence: a signal peptide with a putative processing site between Q26 and D27 (13), a catalytic module (aa 35 to 391), and a C-terminally duplicated dockerin module (Fig. 1B). The catalytic module was similar to the bacterial chitinases grouped in glycosyl hydrolase family 18 subfamily A (GHF18A) and was accordingly called Chi18A (11). The ₁₁₉SXGG and ₁₅₆DXXDXDXE motifs located at the ends of the third and fourth β-strands of GHF18 modules were perfectly conserved (21). This suggests that Glu-163 is involved as the proton donor in the catalytic double-displacement mechanism during hydrolysis (11, 21, 23). However, there are no tryptophan residues in homologous positions or chitin binding modules as in most other chitinases of GHF18. These traits are believed to be a prerequisite for the effective hydrolysis of insoluble chitin, and their absence could account for the low activity on the insoluble substrate (21, 23).

The C terminus of Chi18A consists of a repeated 24-aa module, a dockerin domain type I (Fig. 1B). The highest scores in a BLAST search (1) were obtained with dockerins from XynC and ManA, other hemicellulolytic components of the C. thermocellum cellulosome. It can be speculated that these minor accessory hemicellulolytic components compete with ChiA for the same cohesin binding site(s) in the cellulosome.

The complete chiA gene was specifically amplified by PCR from chromosomal DNA with the oligonucleotide primers chiA1n (CACAGGATCC CTCTCTTCGA CAAAAAGG-3′) and orfZ2c (CTGAGAGCTC CTACGGTTCC TTTCCGTATA CC-3′) and the truncated Chi18Acat with the primers chiA1n and chi1420c (CCGTGAGCTC TCATCTTTGA ATGATG-3′). The latter protein was used for propagation of antibodies as described by Zverlov et al. (25) to avoid an unspecific reaction with the dockerins of other cellulosomal components. Both DNA fragments were cloned by fusing the Ser-29 codon with the N-terminal hexahistidyl-tag linker peptide of vector pQE31 (Qiagen).

Isopropyl-β-d-thiogalactopyranoside (IPTG) induction of chiA in Escherichia coli led to the formation of inclusion bodies. IPTG-treated cells were collected by centrifugation and cracked in a French press (Aminco; SLM Instruments). The inclusion bodies were purified by high-speed centrifugation and two times' washing with 150 mM NaCl. The resulting pellet was dissolved in 5 M urea in 100 mM Tris-HCl (pH 8.0) and slowly dialyzed against 20 mM Tris-HCl (pH 8.0) by four buffer changes within 30 h for renaturation. The protein was further purified by His-tag affinity chromatography on Ni-nitrilotriacetic acid columns using the instructions of the supplier (Qiagen).

The purified Chi18A protein contained a major fraction that showed a molecular mass of 53 kDa in denaturing sodium dodecyl sulfate gel electrophoresis, as predicted from the sequence (467 aa; molecular mass, 53,429 Da) (Fig. 2, lane 3). A minor fraction was C-terminally truncated Chi18A, since (i) it was active on 4-methylumbelliferyl-β-d-N,N′,N″-triacetylchitotriose in zymogram stainings and (ii) the N-terminal amino acid residues were determined unambiguously by Edman degradation sequencing using a Procise 492 protein sequencer (Applied Biosystems); the sequence MFGSHHHHHHTDP-SLP corresponded exactly to the N terminus expected from the cloning procedure (hyphen between vector and insert sequence). The loss of the dockerin module had no negative effect on the enzyme activity.

FIG. 2.

Denaturing protein electrophoresis, zymogram staining, and Western blot of cellulosome and purified chitinase. Track 1, molecular size markers (in kilodaltons); track 2, denatured cellulosomal proteins and zymogram staining of chitinase activity in CM-chitin-RBV; track 3, purified Chi18A; and track 4, Western blot of denatured cellulosomal proteins with anti-Chi18Acat antibodies.

The enzymatic activity of purified Chi18A was quantitatively determined with the substrates p-nitrophenyl-β-d-N,N′-diacetylchitobiose (pNP-chitobiose) and p-nitrophenyl-β-d-N,N′,N″-triacetylchitotriose (pNP-chitotriose) at 60°C in 0.1 M phosphate-citrate buffer, pH 6.2, by photometric determination of released p-nitrophenol (λ = 395 nm) (substrates from Sigma-Aldrich). The reaction was stopped by addition of 2 volumes of 1 M sodium carbonate. Activity was calculated photometrically (ɛ₃₉₅ = 13,500 [mM · cm]⁻¹; 1 U = 1 μmol of p-nitrophenol · min⁻¹). The protein concentration was determined with the Protein Assay Kit (Bio-Rad) using bovine serum albumin as the standard. All measurements were done at least in triplicate.

The purified ChiA preparation had a specific activity of 26 ± 0.3 U mg of protein⁻¹ on pNP-chitotriose and 1.7 ± 0.1 U mg⁻¹ on pNP-chitobiose. Activity on insoluble chitin was detected but was too low to be quantified by measuring the reducing sugars with the dinitrosalicylic acid method (12). Activity on chitosan, microcrystalline cellulose, or the aryl-β-glycosides p-nitrophenyl-cellobioside and p-nitrophenyl-β-glucoside was not observed. The enzyme had a broad optimum of activity between pH 4.5 and 6.5 and was stable at elevated temperature with a half-life of more than 10 h at 60°C.

Chi18A readily hydrolyzed derivatized chitin (carboxymethyl-chitin-remazol-brilliant-violet [CM-chitin-RBV] from Loewe Biochemica GmbH, Sauerlach, Germany). Substrate limitation, which would be expected for an exochitinase or chitobiosidase, was not observed. The activity on pNP-chitobiose was 15 times lower than on pNP-chitotriose. Thin-layer chromatography of hydrolysis products from CM-chitin revealed the initial production of large oligomers and the presence of chitobiose as the major end product after extended incubation (data not shown). It therefore can be assumed that Chi18A is an endochitinase (EC 3.2.1.14).

Chi18A is the only chitinase gene in the databases with a C-terminal dockerin module, the binding partner of the cohesin modules on the cellulosomal scaffoldin (20). The expression of ChiA and its localization in cellulosomes of C. thermocellum were shown by Western blotting and in zymograms of cellulosomal proteins separated by denaturing sodium dodecyl sulfate-polyacrylamide gel electrophoresis (16). Cellulosomes were purified from cellulose-grown cells by gel filtration as described previously (26). Antibodies against purified Chi18Acat detected a single band of identical molecular mass in a Western blot (Fig. 2, lane 4) (25). A band of chitinase activity was visible in zymograms as a band of decoloration of CM-chitin-RBV, 0.05% (wt/vol) (Fig. 2, lane 2). No protein band was detectable within the clearing zone, indicating a minor protein. Similarly, a single band of identical M_r was obtained with 4-methylumbelliferyl-β-d-N,N′,N″-triacetylchitotriose (not shown). Chi18A thus is a cellulosomal component with a molecular mass of 51 kDa, the size expected for the mature Chi18A protein. It appears as the only chitinase present in the cellulosome of C. thermocellum. This conclusion is corroborated by the similarity of the enzymatic activity pattern of recombinant Chi18A and the cellulosomal preparation.

The clearly defined enzymatic activity of Chi18A, with only a single component responsible for it, opens the possibility of roughly estimating the number of molecules present in individual cellulosomal particles. The specific activities of recombinant Chi18A protein and the purified cellulosome preparation on pNP-chitotriose were compared (26 ± 0.3 versus 0.07 ± 0.01 U mg⁻¹). Assuming an average mass of 0.9 MDa for a single cellulosome unit (containing one scaffoldin molecule and nine of the major catalytic components) and applying the formula (SpA_ChiA/SpA_cellulosome) × (M_ChiA/M_cellulosome), where SpA is the specific activity and M is the molecular mass, it can be concluded that one Chi18A molecule is present in ca. 20 cellulosomal units. If we assume that the specific activity of the recombinant Chi18A renatured from inclusion bodies was smaller than that of the native protein, this ratio would be even higher. The scarcity of cellulosomes containing Chi18A sheds light on the fact that each unit has a fixed number of components (nine) but has an individual composition. Obviously, Chi18A is not well expressed, at least in cellulose-grown C. thermocellum cells. As chitinase is not a main component of the cellulosome and does not have an obvious function in the hydrolysis of cellulose, its rare occurrence is not surprising.

The gene chiA is the first chitinase gene isolated from a thermophilic clostridium and is the first in the saccharolytic clostridia besides two functionally uncharacterized similar sequences in Clostridium acetobutylicum (10). The role of a chitinase for a bacterium specialized on cellulose hydrolysis remains obscure: chitin is not present in plant biomass itself and is not used as a fermentation substrate by C. thermocellum. However, the first colonizers of dead biomass in the initial phase of the rotting process are fungi. Fungal hyphae could block access of the bacteria to the cellulose substrate, the hydrolysis products of fungal cell walls could be used for biosynthetic purposes, or fungi, competing for cellulose degradation in the habitat, could be fought off by the combined action of chitinase and the β-1,3-glucanase also produced by C. thermocellum (5, 6, 8, 17).

Nucleotide sequence accession numbers.

The sequences described in the text were deposited at GenBank under accession numbers Z68924 and AJ420770.