Characterization of the CipA Scaffolding Protein and In Vivo Production of a Minicellulosome in Clostridium acetobutylicum

Abstract

The cipA gene encoding the Clostridium acetobutylicum scaffolding protein CipA was cloned and expressed in Escherichia coli. CipA contains an N-terminal signal peptide, a family 3a cellulose-binding domain (CBD), five type I cohesin domains, and six hydrophilic domains. The uniqueness of CipA lies in the enchainment of cohesin domains that are all separated by a hydrophilic domain. Affinity-purified CipA was used in equilibrium-binding experiments to characterize the interaction of CipA with crystalline cellulose. A K_d of 0.038 μM and a [C]_max of 0.43 μmol of CipA bound per g of Avicel were determined. A mini-CipA polypeptide consisting of a CBD3a and two cohesin domains was overexpressed in C. acetobutylicum, yielding the in vivo formation of a minicellulosome. This is to our knowledge the first demonstration of the in vivo assembly of a recombinant minicellulosome.

Clostridium acetobutylicum is a gram-positive, sporeforming anaerobic bacterium that converts various sugars and polysaccharides into acids and solvents (7). Although this bacterium is unable to degrade cellulose, the sequence analysis of its genome revealed the presence of a large cellulosomal gene cluster (12). In a previous work, we demonstrated that C. acetobutylicum produced an inactive cellulosome with an apparent molecular mass of 665 kDa (14). Biochemical and Western analysis revealed that the C. acetobutylicum cellulosome comprised four major subunits, including the scaffolding protein CipA and the cellobiohydrolases Cel48A, Cel9X, and Cel9C or Cel9E. The CipA scaffolding protein of C. acetobutylicum has never been purified, and its biochemical properties are unknown.

In this communication, we report the cloning and heterologous expression of cipA as well as the purification and cellulose-binding properties of the CipA protein. This is to our knowledge the first characterization of a full-length scaffolding protein. In addition, a gene encoding a mini-CipA scaffolding protein was constructed and expressed in C. acetobutylicum. For the first time, we will demonstrate that the expression of a mini-CipA scaffolding protein leads to the assembly of a minicellulosome in vivo.

Amino acid sequence and domain structure of CipA.

CipA contains 1,484 amino acids and has a molecular mass of 154.4 kDa. The overall modular architecture of the scaffolding protein CipA is shown in Fig. 1A. CipA is organized into five type I cohesin domains, a family 3a cellulose-binding domain (CBD) at the N terminus, and six hydrophilic domains. Each cohesin domain is spaced by a hydrophilic domain, which confers on CipA an organization that differs from that of all previously reported scaffolding proteins.

FIG. 1.

(A) Schematic diagram of the CipA scaffolding protein from C. acetobutylicum. CipA contains 1,845 amino acids and has a molecular mass of 154 kDa. (B) Structure-based sequence alignment of family 3a CBDs. Secondary-structure elements are indicated and enumerated. Proposed cellulose-binding residues are bpxed, as are the homologous residues in the other CBDs. Calcium-binding residues are shown on a dark shaded background. Highly conserved residues, which occupy the shallow groove of unknown function on the top face of the molecule, are shown on a solid background.

CBD3a displays 46% and 42% identity with the cellulose-binding domains from Clostridium thermocellum and Clostridium cellulolyticum, respectively. A lower degree of identity, 38%, was found with the Clostridium cellulovorans CBD. A structure-based alignment of selected CBD sequences is shown in Fig. 1B. Based on sequence homology, the CipA CBD structure appears to be very similar to that of the previously elucidated family 3a CBD (15, 17), including a nine-stranded jellyroll topology which exhibits distinctive structural elements consistent with family 3 CBDs. Mapping of conserved residues in the CipA CBD showed a well-conserved calcium-binding site, probably involved in stabilizing the protein fold, and a putative cellulose-binding surface.

As previously demonstrated from the structure of the family 3a cellulosomal CBD, the His61, Tyr70, and Trp123 residues which form a contact with the glucose rings of cellulose by aromatic stacking (15) are conserved in the C. acetobutylicum CipA CBD. Moreover, as shown in Fig. 1B, Arg117 and Asp60 are also conserved in the CipA CBD sequence. The homologous residues in the family 3 CBDs are believed to form a salt bridge, which would then stack against an additional glucose ring at the cellulose surface (15).

Cloning, expression, and purification of recombinant CipA.

The DNA region coding for CipA was amplified by PCR from genomic DNA with the primers CipA_dir (5′-CACCGTGAAAAAAAGAAATATTGCCATTCTAGGAATG-3′) and CipA_rev(5′-TTCAACAGTTATTTTTCCATTTGTTGATACATGATC-3′). The resulting blunt-end PCR product was cloned into the pET Directional Topo expression vector (Invitrogen) and sequenced. In this vector, CipA was expressed as a C-terminal His-tagged fusion protein under the control of the T7 RNA polymerase. The CipA protein was produced from the cytoplasmic fraction of E. coli BL21(DE3) cells harboring the resulting plasmid, pET-CipA. Expression of the CipA protein was obtained after induction by 40 μM IPTG (isopropyl-β-d-thiogalactopyranoside).

CipA was purified on an Ni-nitrilotriacetic acid column, and after sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), a major band consistent with the estimated molecular mass was obtained (Fig. 2). Four minor bands, probably corresponding to degradation products, were revealed by SDS-PAGE (Fig. 2) and Western blotting (data not shown) with an antibody directed against the cohesin 5 domain (14). Similar results were obtained when the native full-size CbpA of C. cellulovorans was purified from a culture supernatant by Avicel binding (2). When a similar Avicel-bound preparation of recombinant CbpA was examined by Western immunoblots with anti-CbpA immunoglobulin G, the ladder of fragments was essentially identical (16).

FIG. 2.

Purification of CipA protein. Lanes: 1, molecular mass markers; 2, purified CipA fraction. Arrow indicates full-length CipA scaffolding protein. Arrowheads show the bands corresponding to CipA degradation products.

One other scaffolding protein (from C. thermocellum) has been cloned and expressed in E. coli (1), although no data were presented regarding the purity of the protein and its biochemical properties. However, previous work (8) has demonstrated an anomalous behavior of the 210-kDa CipA of C. thermocellum, yielding a collection of bands by simply altering the environmental conditions (pH or ionic strength) prior to electrophoresis. The reasons for the degradation products observed in SDS-PAGE has recently been examined (5) and demonstrated to be due to chemically labile Asp-Pro sequences within the cohesin domain. This sequence appears quite frequently in the large cellulosomal proteins and is also prevalent in C. acetobutylicum CipA.

Affinity of CipA for crystalline cellulose.

The affinity constant of the purified CipA for microcrystalline cellulose was determined with a double-reciprocal plot of bound versus free CipA (Fig. 3). Various amounts of purified CipA (between 1.5 and 10.5 μg) were added to 0.5 mg of Avicel (Merck) in 50 mM Tris-HCl buffer, pH 7. The suspension was agitated with an orbital shaker for 1 h at room temperature. The samples were then centrifuged at 12,000 × g to sediment cellulose and the adsorbed CipA. The amount of protein that remained in the supernatant fluids (free CipA) was determined colorimetrically. The amount of adsorbed CipA was calculated by subtracting the amount of free CipA from the total amount added in the assay. Data were analyzed by double-reciprocal plots (Avicel-bound CipA versus free CipA).

FIG. 3.

Double-reciprocal plot to quantify the parameters of adsorption of CipA to Avicel.

From the linear plot that indicates a specific and single type of interaction, the dissociation constant K_d (0.038 μM) and the maximum binding capacity [C]_max (0.43 μmol of CipA bound per g of Avicel) were determined (Table 1). The calculated dissociation constant indicated that CipA had a higher affinity (≈10-fold) for Avicel than the CBDs of C. cellulovorans (4) and C. thermocellum (9), whereas the value was very close to that obtained with the mini-CipC1 protein from C. cellulolyticum (3). To our knowledge, this is the first determination of a dissociation constant for cellulose with a full-length scaffolding protein. From the CipA and the mini-CipC1 results, it could be argued that the modular arrangement around CBD positively influences the binding of CBD to cellulose.

TABLE 1.

Comparison of CipA cellulose-binding parameters

Species	Component	Kd (μM)	[Cmax] (μmol/g)	Reference
C. cellulolyticum	Mini-CipCl	0.14	0.40	3
	Cellulosome	0.014	0.0114	3
C. cellulovorans	CBD	0.6	2.10	4
C. thermocellum	CBD	0.4	0.54	9
C. acetobutylicum	CipA	0.038	0.43	This study

Expression of a minicellulosome in C. acetobutylicum.

Recent studies indicated that a minicellulosome containing a miniscaffolding protein and cellulosomal enzymatic subunits could be created and was functional in vitro (2, 11). Interestingly in C. cellulovorans, in vitro assembly of ExgS, EngH, and mini-CbpA into a minicellulosome increased the activity against insoluble cellulose 1.5- to 3-fold (10).

As our previous work demonstrated that a cellulosome is produced by C. acetobutylicum when grown on cellobiose (14), we investigated the possibility of assembling in vivo a recombinant minicellulosome by overexpressing a miniscaffolding protein in C. acetobutylicum. For this purpose, a mini-cipA gene was constructed by PCR to express the protein CBD-HD₁-HD₂-Coh₁-HD₃-Coh₂-HD₄. As preliminary results have shown that part of the CipA protein is not secreted, the original signal peptide was replaced by the cleavage site (Ala-Phe-Ala-Ala) of the C. cellulolyticum scaffolding protein CipC (13), which is well secreted. The recombinant mini-cipA gene was cloned into the pSOS95 expression vector (P. Soucaille, unpublished data), which allows the cloning of PCR-amplified structural genes between the constitutive promoter of the thiolase gene (thl) and the rho-independent transcriptional terminator of the acetoacetate decarboxylase gene (adc), both isolated from C. acetobutylicum ATCC 824.

The mini-cipA gene was constructed by cloning into the pSOS95 two PCR products. The first product was amplified with primers CipA1_dir (5′-CGCGGATCCGCGAATGGAGGGTTAAAACAGTG-3′) and CipA1_rev (5′-AACGACACCGGTACCTGCTGCGAAGACATGATTTTTAGAAATACC-3′) so that this 128-bp PCR product contained a 5′ BamHI site (underlined in the forward primer) and an AgeI site (underlined in the reverse primer). The second product was amplified with primers CipA2_dir (5′-GCAGGTACCGGTGTCGTTCAAATACAATTTGCTGATACAAATACTAGTACAAC-3′) and CipA2_rev (5′-GAATTCCATATGGAATTCCGGCGCCTCCTTTAACAACAACAGTAAATGTCCCTGCATTTCCTGCAC-3′) so that the 2,483 bp contained a 5′ AgeI site (underlined in the forward primer) and an EheI and an NdeI site (underlined in the reverse primer). Residues indicated in italic in the CipA1_rev and CipA2_dir primers correspond to the sequence encoding the new peptidase cutting site.

These PCR products were cloned into the pGEM-T Easy (Promega) to give the pCipA1 and pCipA2 plasmids, respectively, and the cloned DNA fragment was sequenced. After double digestion with AgeI and NdeI, the pCipA2 insert was ligated into similarly digested pCipA1. The resulting plasmid, pCipA3, was then double digested with BamHI and EheI, and the 2,565-bp fragment obtained was cloned into the pSOS95 vector digested with the same restriction enzymes. The resulting plasmid, prSCipA, encoding the mini-CipA protein with a theoretical molecular mass of 89.5 kDa was then methylated in vivo in Escherichia coli ER2275 carrying the pAN1 methylating plasmid and used to transform C. acetobutylicum by electroporation (6).

The recombinant C. acetobutylicum(prSCipA) strain was grown in synthetic medium with cellobiose (20 g liter⁻¹) and crystalline cellulose (10 g liter⁻¹) as carbon sources. After the cellobiose was totally consumed, the cellulose was washed, and the adsorbed proteins were eluted as previously described (14). A 5% native polyacrylamide gel was used to separate the concentrated fraction eluted from cellulose. After transfer onto a nitrocellulose membrane, recombinant and native CipAs were detected by Western blotting with antibodies directed against the Coh5 domain (Fig. 4A and B). In addition to the large (>665-kDa) enzyme complex described previously, a second signal with an apparent molecular mass of 250 kDa was detected.

FIG. 4.

Cellulosome characterization. Immunological detection of cellulolytic complex purified by native PAGE from wild-type C. acetobutylicum (A) and a recombinant strain transformed with the prSCipA plasmid (B). Detection was done with the anti-Coh5 antiserum. (C) SDS-PAGE analysis of the minicellulosome. Mini-CipA and Cel48A were identified with anti-Coh5 and C. cellulolyticum Cel48F antisera, respectively.

On a second native gel, the band corresponding to the minicellulosome was cut out, the proteins were denatured by boiling in 1% SDS plus mercaptoethanol, and the mixture was subjected to separation by SDS-PAGE. The gel revealed that the minicellulosome consisted of two major proteins (Fig. 4C) which had molecular masses of 122 and 84 kDa. These proteins were identified as mini-CipA and Cel48A by Western blot analysis (data not shown) with the Coh5-specific antibody from C. acetobutylicum CipA and the Cel48F antibody from C. cellulolyticum, as we have previously demonstrated the specificity of the cross-reaction. As shown in Fig. 4C, the molecular mass of Cel48A is in good agreement with the theoretical 80.9 kDa, whereas the mini-CipA protein with a theoretical molecular mass of 89.9 kDa migrated with a higher molecular mass of 122 kDa. Similar results were obtained with the CipA protein from the wild-type cellulosome (14), where the CipA protein with a theoretical mass of 154.4 kDa migrated at a molecular mass of 180 kDa. These observations probably result from an in vivo glycosylation of this protein.

The immunological evidence of the presence of Cel48A in this minicomplex was not surprising, as we have previously demonstrated that Cel48A is the major component of the cellulosome produced by C. acetobutylicum. This clearly demonstrated that by expressing a mini-CipA protein, a minicellulosome could be produced in vivo. This is to our knowledge the first report of a recombinant minicellulosome assembly. Moreover, no differences in the ratio of mini-CipA to the other protein band was obvious in the mini- compared to the wild-type cellulosome except for Cel48A, which was much more abundant.

Enzyme activities of the purified cellulolytic complex containing only the wild-type cellulosome or both the wild-type and the recombinant minicellulosome were measured on carboxymethyl cellulose, phosphoric acid-swollen cellulose (18), Avicel, and bacterial cellulose. Bacterial cellulose was a generous gift from H. P. Fierobe (BIP, Marseille, France). The results indicated that neither preparation showed detectable activity on Avicel or bacterial cellulose. This finding was in agreement with the fact that C. acetobutylicum is unable to grow on crystalline cellulose. Low and similar levels of activities were detected for both fractions on carboxymethyl cellulose and phosphoric acid-swollen cellulose (0.011 and 0.002 UI/mg, respectively, versus 0.015 and 0.0074 UI/mg, respectively), indicating that the minicellulosome was not more active than the wild-type cellulosome. On the other hand, when the recombinant C. acetobutylicum(prSCipA) strain was grown in synthetic medium with cellobiose (20 g liter⁻¹) and crystalline cellulose (10 g liter⁻¹) as carbon sources, sedimentation of cellulose after mixing occurred at a slower rate. As all the mini-CipA protein overexpressed was bound to cellulose (data not shown), we suggest that this observation resulted from an increase in the amount of CBDs (from free mini-CipA and/or complexed mini-CipA) bound to cellulose.

It has been observed that the CBD of CipC from C. cellulolyticum prevents the flocculation of bacterial cellulose and phosphoric acid-swollen cellulose (13), suggesting that the presence of the family 3a CBD leads to macroscopic changes in cellulose structure. This work now opens the possibility of studying the cohesin-dockerin interaction between a full-length scaffolding protein and purified cellulases. In this regard, work is currently under way in our laboratory to clone, overexpress, purify, and biochemically characterize the binding to CipA and the hydrolytic activity of the Cel48A, Cel9X, and Cel9C cellulases.