Chitinase B of “Microbulbifer degradans” 2-40 Contains Two Catalytic Domains with Different Chitinolytic Activities

Abstract

Chitinase B of “Microbulbifer degradans” 2-40 is a modular protein that is predicted to contain two glycoside hydrolase family 18 (GH18) catalytic domains, two polyserine domains, and an acidic repeat domain. Each of the GH18 domains was shown to be catalytically active against chitin. Activity assays reveal that the amino-terminal catalytic domain (GH18_N) releases methylumbelliferone from 4′-methylumbelliferyl-N,N′-diacetylchitobiose 13.6-fold faster than the carboxy-terminal catalytic domain (GH18_C) and releases chitobiose from the nonreducing end of chitooligosaccharides, therefore functioning as an exochitinase. GH18_C releases methylumbelliferone from 4′-methylumbelliferyl-N,N′,N"-triacetylchitotriose 2.7-fold faster than GH18_N and cleaves chitooligosaccharides at multiple bonds, consistent with endochitinolytic activity. Each domain was maximally active from 30 to 37°C and from pH 7.2 to 8.0 and was not affected by Mg²⁺, Mn²⁺, Ca²⁺, K⁺, EDTA, EGTA, or 1.0 M NaCl. The activity of each domain was moderately inhibited by Ni²⁺, Sr²⁺, and Cu²⁺, while Hg²⁺ completely abolished activity. When the specific activities of various recombinant portions of ChiB were calculated by using native chitin as a substrate, the polypeptide containing the endo-acting domain was twofold more active on native chitin than the other containing the exo-acting domain. The presence of both domains in a single reaction increased the amount of reducing sugars released from native chitin to 140% above the theoretical combined rate, indicating that the domains function cooperatively to degrade chitin. These data demonstrate that the GH18 domains of ChiB have different activities on the same substrate and function cooperatively to enhance chitin depolymerization.

Chitin, a homopolymer of β-1,4-linked N-acetylglucosamine, is the second most abundant polymer in the biome (15). Chitin is a difficult substrate for microbial degradation because it is usually crystalline and complexed with protein, salts, and other carbohydrates. However, many microorganisms have developed efficient strategies for the depolymerization, transport, and metabolism of chitin and its derivatives. These systems involve multiple enzyme activities, usually encoded on separate polypeptides. Pseudoalteromonas sp. strain S91 (24), Serratia marcescens (22), and Streptomyces coelicolor A3 (2, 20), for example, secrete several chitin-depolymerizing enzymes in the presence of chitin. Endo- and exochitinases that function cooperatively to depolymerize chitin have been described (3, 6, 22). Endochitinases randomly cleave glycosidic linkages, generating free ends and long chitooligosaccharides. These are then acted upon by exochitinases that release chitobiose from the nonreducing ends of each. While exo- and endochitinases are able to depolymerize chitin alone, the presence of both activities significantly increases the efficiency of chitinolytic systems.

The glycoside hydrolase family 18 (GH18) domain is the most common catalytic domain of microbial chitin depolymerases (7). Despite sharing a consensus sequence and a conserved catalytic glutamic acid residue, GH18 domains may differ in their activity toward polymeric chitin and chitooligosaccharides (i.e., endo- versus exo- activity) (19). Chitodextrinases, which depolymerize chitooligosaccharides but not chitin, also contain GH18 domains (11). Chitinolytic enzymes with GH18 domains have been isolated from organisms as diverse as psychrophilic eubacteria (12) and hyperthermophilic archaeons (23), demonstrating the wide range of conditions to which these domains have adapted. Because conserved residues are found in GH18 domains with divergent optima and substrate specificities, sequence analysis is insufficient to determine the enzymatic specificities of newly discovered chitinases.

“Microbulbifer degradans” 2-40, a marine γ-subgroup proteobacterium isolated from the Chesapeake Bay watershed in coastal Virginia, is able to degrade 10 complex polysaccharides, including chitin (2). The chitinolytic system of 2-40 has recently been shown to include three chitin depolymerases (ChiA, ChiB, and ChiC), a noncatalytic chitin-binding protein (CbpA), a chitodextrinase (CdxA), and three N-acetylglucosaminidases (HexA, HexB, and HexC) (9). ChiA and ChiB include long polyserine domains that appear to separate functional groups. One of the chitin depolymerases, ChiB, was selected for further study in this work because of its unusual structural features.

ChiB is a modular, 1,271-amino-acid enzyme with a calculated molecular mass of 136.1 kDa (9). The amino terminus is predicted to contain a secretion signal that is separated from the remainder of the protein by a polyserine domain of 148 amino acids, 99 of which are serine residues. ChiB is predicted to include two complete GH18 domains (amino-terminal domain GH18_N and carboxy-terminal domain GH18_C) separated by a 180-amino-acid linker domain which includes an acidic region consisting of TE-(ET)₁₀ and another polyserine domain containing 39 serine residues. Here we report that both GH18 domains of ChiB are catalytically active but differentially cleave glycosidic linkages, depending on their location within the chitin polymer. In addition, it was shown that chitin depolymerization is enhanced by the presence of both domains. The implications and advantages of encoding two catalytic domains on a single polypeptide are discussed.

MATERIALS AND METHODS

Chemicals and reagents.

Standard reagents, chitooligosaccharides, methylumbelliferone (MUF) substrates, and chitin were obtained from Sigma (St. Louis, Mo.). Ethylene glycol chitin was purchased from Fisher Scientific (Pittsburgh, Pa.). Ni-nitrilotriacetic acid (Ni-NTA) agarose was obtained from Qiagen (Valencia, Calif.). Restriction enzymes and T4 DNA ligase were purchased from New England Biolabs (Beverly, Mass.). Bugbuster NT and pETBlue2 were obtained from Novagen (Madison, Wis.).

Cloning and expression of GH18_N and GH18_C.

Oligonucleotide primers were designed to amplify the nucleotide sequence corresponding to each catalytic domain by PCR with purified “M. degradans” genomic DNA as a template. Primer sequences were as follows: GH18_N-F (468), CTTGGCGCGCCATGGTGTAGATGCCGAATTG; GH18_N-R (1924), CCGGGTACCGTTGTCTTCGTAATTGCCTTC; GH18_C-F (2512), CTTGGCGCGCCATGGCGAAACAGATTTAG; and GH18_C-R (3800), CCGGGTACCCTGCTTTTCGTTGCCGAA (restriction sites are underlined, and the relative position of the 5′ nucleotide start of each primer within the chiB sequence is shown in parentheses). GH18_N+C was created by using primers GH18_N-F (468) and GH18_C-R (3800). Each amplified fragment was then digested with the appropriate restriction enzymes and ligated into the protein expression vector pETBlue2 by using T4 DNA ligase. Expression constructs were verified by sequencing and transformed into E. coli Tuner DE3(pLacI) cells. Protein expression was performed according to the manufacturer's protocol. Cells were lysed with Bugbuster NT lysis buffer and centrifuged, and the supernatant was collected. Supernatants containing recombinant enzymes were applied to an Ni-NTA agarose column and purified according to the manufacturer's protocol for native protein purification. Purified enzyme samples were quantified by using a bovine serum albumin protein quantification kit (Pierce, Rockford, Ill.).

Glycol chitin zymography.

Ethylene glycol chitin was incorporated into the separating portion of a sodium dodecyl sulfate-polyacrylamide gel to a final concentration of 0.01%. After fractionation of the proteins, the zymogram was incubated in refolding buffer (50 mM Tris-Cl, 1 mM EDTA, 5 mM 2-mercaptoethanol [pH 7.5]) overnight at 4°C and subsequently analyzed for chitin depolymerase activity as described elsewhere (8, 25).

Enzyme assays with chitin analogs.

Solutions of 4′-methylumbelliferyl-N,N′-diacetylchitobiose (MUF-diNAG) and 4′-methylumbelliferyl-N,N′,N"-triacetylchitotriose (MUF-triNAG) were prepared in 50 mM sodium phosphate buffer (pH 7.0). Reaction mixtures contained 2 μg of purified enzyme and 30 μM analog solution. After incubation for 5 to 10 min at 37°C for GH18_N or for 5 to 20 min at 30°C for GH18_C, reactions were stopped by submersion in an ice water bath. Liberated methylumbelliferone was detected with a Hoefer TKO-100 fluorometer. The reaction was measured at multiple time points between 5 and 20 min and was found to be linear, with less than 10% of the substrate being degraded.

Oligosaccharide electrophoresis.

Labeling and electrophoresis of chitooligosaccharides were performed as described previously (8). Briefly, the reactions were incubated with 2 volumes of labeling solution (1.0 M sodium cyanoborohydride, 0.2 M 2-aminobenzoic acid) and dried under vacuum. Each sample was mixed with standard 2× sodium dodecyl sulfate-polyacrylamide gel electrophoresis loading buffer and fractionated in a 15% polyacrylamide gel at a 45-mA constant current. Labeled oligosaccharides were visualized under UV light.

Determination of reaction optima for each domain.

MUF-diNAG or MUF-triNAG was added to 20 μg of purified enzyme and incubated at a given pH or temperature, and activity was detected as described above. The buffers used were sodium acetate (pH 4.0 to 5.5), MES (morpholineethanesulfonic acid) (pH 5.5 to 6.5), PIPES [piperazine-N,N′-bis(2-ethanesulfonic acid)] (pH 6.5 to 7.0), HEPES (pH 7.0 to 8.0), and Tris base (pH 8.0 to 9.5). For a given enzyme, the activity under reaction conditions that permitted maximum activity was assigned a value of 100%. Where indicated, EDTA, EGTA, KCl, NiCl₂, SrCl₂, MgCl₂, MnCl₂, CuCl₂, CaCl₂, or HgCl₂ was added to reaction mixtures to a final concentration of 10 mM; NaCl was added at concentrations of up to 1.0 M. Reaction mixtures containing metal ions contained 200 pmol of enzyme and were incubated for 10 min at 37°C for GH18_N or for 20 min at 30°C for GH18_C.

Enzyme assays with chitin and chitin derivatives.

Purified enzyme and substrate (2 mg of chitin or 10 nmol of chitooligosaccharide) were added to 50 mM HEPES (pH 7.5) and incubated at 30°C. The amount of reducing sugar generated was determined by the dinitrosalicylic acid assay as described elsewhere (14). Specific enzyme activity was estimated by comparison to a standard curve.

Protein sequence analysis.

Analysis of protein domains was performed with the Simple Modular Architecture Research Tool (21). Similarity between proteins and protein domains was determined by the BLAST algorithm (1). The lipoprotein-anchoring site within ChiB was identified by using the database of bacterial lipoproteins (13).

Nucleotide sequence accession number.

The nucleotide and protein sequences of ChiB have been placed in GenBank under accession number BK001042.

RESULTS

ChiB is predicted to contain two catalytic domains and a lipoprotein acylation site.

ChiB was previously predicted to contain two catalytic sites (9). The first catalytic site, GH18_N, was identified in the amino-terminal region of ChiB (residues 221 to 605) (Fig. 1). It consists of 385 amino acids and is most similar to the GH18 domain of the exochitinase ChiB of S. marcescens (S52422) (55% identity and 69% similarity). A second predicted catalytic site, GH18_C, was present in the carboxy-terminal domain of ChiB (residues 860 to 1254). This domain is composed of 395 amino acids and is most similar to a chitinase from Vibrio sp. strain 5SM-1 (AAL46648) (49% identity and 66% similarity). The two GH18 domains of ChiB share only 29% identity and 42% similarity when aligned at the amino acid level. GH18_N and GH18_C include the motifs SVGGWAESN-X₃₃-FDGIDIDWEYP and SIGGWTMSTPF-X₂₆-FDGVDIDWEYP, respectively. These sequences are nearly identical to the consensus sequence that characterizes a GH18 domain, and each also includes the key catalytic Glu residue (underlined) (19).

FIG. 1.

Comparison of the domain architectures of T. kodakaraensis KOD1 ChiA and “M. degradans” strain 2-40 ChiB. Gray boxes, type II secretion signal; black boxes, predicted lipobox; SSS, polyserine domains; hatched box, acidic repeat domain; crosshatched boxes, chitin binding domains. Black bars indicate the truncated portions of ChiB created for this work. GH18_N is located between amino acids 221 and 605 of ChiB. GH18_C is located between amino acids 860 and 1254.

ChiB was found to contain a predicted lipobox within amino acid residues 16 to 19, composed of L-S-A-C. In addition, two positively charged residues are found within the first five amino acids (N at position 2 and K at position 5) and are separated from the lipobox by a hydrophobic stretch of 10 amino acids. These characteristics satisfy the major criteria required for a lipoprotein secretion signal and acylation site (10, 13).

GH18_N and GH18_C independently depolymerize chitin.

To determine whether the GH18 domains of ChiB are catalytically active against chitin, the sequence corresponding to each domain (GH18_N, codons 156 to 641; GH18_C, codons 837 to 1266) was amplified by PCR and ligated into pETBlue2 to create carboxy-terminal His₆ fusions. The polypeptides were expressed in Escherichia coli and purified on Ni-NTA agarose columns. The chitinolytic activity of each GH18 domain was tested by using a glycol chitin zymogram. Consistent with their conserved sequence features, the ability of each catalytic domain to independently depolymerize chitin was apparent in zymograms (Fig. 2). Clear zones indicative of depolymerization were observed and corresponded to the predicted masses of the recombinant polypeptides (50.5 kDa for GH18_N and 47.7 kDa for GH18_C).

FIG. 2.

The predicted GH18 domains of ChiB are catalytically active against chitin. The nucleotide sequence corresponding to each predicted catalytic domain and some flanking sequence (GH18_N, codons 156 to 641; GH18_C, codons 837 to 1266) was amplified by PCR, ligated into the pETBlue2 expression vector, induced by IPTG (isopropyl-β-d-thiogalactopyranoside), and purified from cell lysates on an Ni-NTA agarose column. Equal amounts of each recombinant protein (20 μg) were fractionated by electrophoresis in a glycol chitin zymogram and refolded as described in Materials and Methods. After staining with Calcofluor, zones of activity appear as dark bands against a bright background. The sizes of the active bands are in good agreement with the predicted masses of the recombinant proteins.

GH18_N and GH18_C differentially degrade chitin analogs.

One possible explanation for the presence of two catalytic domains within ChiB is that each has a different role in the degradation of chitin, as was observed in an archaeal chitinase from Thermococcus kodakaraensis KOD1 (23). The chitin analogs MUF-diNAG and MUF-triNAG consist of chitobiose or chitotriose linked to an MUF moiety at the reducing end that fluoresces under UV light only when cleaved from the saccharide (16). In theory, both exochitinases and endochitinases will hydrolyze the second glycosidic linkage from the nonreducing end of MUF-diNAG, thus releasing fluorescent MUF. Exochitinase activity on MUF-triNAG will result in the formation of chitobiose and nonfluorescent MUF-GlcNAc, while endochitinolytic activity can hydrolyze both the second and third glycosidic linkages of MUF-triNAG, thus releasing MUF.

Purified enzyme samples were added to solutions of either analog, and the release of MUF was monitored fluorometrically during the period of linear accumulation of product. When incubated with MUF-diNAG, the rate of MUF release by GH18_N was 13.6-fold higher than that observed when GH18_C was utilized. However, when GH18_C was incubated with MUF-triNAG, the rate of MUF release was 2.7-fold higher than when it was incubated with GH18_N (Table 1). These results suggest that GH18_N may have exochitinase activity whereas GH18_C may have endochitinase activity.

TABLE 1.

Activities of polypeptides containing ChiB catalytic domains on different MUF analogs

Domain	Rate of MUF release (nmol/min/mg of enzyme)a on:		Ratio of ratesb
	MUF-diNAG	MUF-triNAG
GH18N	80 ± 1.04	5.9 ± 1.77	13.6
GH18C	6.9 ± 0.311	19 ± 0.429	0.363

^aPolypeptides containing each active domain were purified as described in Materials and Methods, and 2 μg was incubated with the chitin analogs MUF-diNAG and MUF-triNAG. Reactions were stopped after 5 to 20 min, and the fluorescence of the reaction was measured. The rate of reaction was linear for at least 20 min. The data are the means and standard errors from three replicates, and each experiment was repeated three times with similar results.

^bRate of MUF release from MUF-diNAG/rate of MUF release from MUF-triNAG.

The GH18 domains have similar reaction optima.

The presence of two catalytic domains for the same substrate within a single polypeptide of an enzyme is rare. If the dual domains of ChiB act together to degrade chitin, it would follow that these domains are most active under similar physical conditions. Ionic, pH, and temperature optima were determined for each domain. Purified samples of each enzyme were incubated with the optimal MUF substrate as identified above. GH18_N had a pH optimum of between 7.2 and 8.0, while GH18_C was most active from pH 7.2 to 7.8 (Fig. 3, top panel). The temperature optimum of GH18_N was determined to be 37°C, with retention of 80% of its activity at 30°C. GH18_C was most active at 30°C and retained only 67% of its activity at 37°C (Fig. 3, bottom panel). A significant loss of activity was observed for each domain at temperatures of above 40°C. Each domain was most active on its optimal MUF substrate regardless of temperature or pH (data not shown).

FIG. 3.

GH18_N and GH18_C have similar pH and temperature optima. To determine the pH optimum (top panel) and temperature optimum (bottom panel) for each catalytic domain, GH18_N (triangles) and GH18_C (squares) were purified as described in Materials and Methods and incubated with MUF-diNAG or MUF-triNAG, respectively, as described in footnote a of Table 1. The activity under reaction conditions that permitted maximum activity was assigned a value of 100%. The data are the means from three replicates, and each experiment was repeated three times with similar results. Standard errors are indicated by error bars.

To examine the effect of ionic conditions on each domain, various chloride salts were added to reaction mixtures. The addition of Mg²⁺, Mn²⁺, Ca²⁺, K⁺, EDTA, and EGTA to 10 mM, and NaCl up to 1.0 M had no effect on the activity of either domain. The activity of GH18N was reduced 36% by Ni²⁺, 8% by Sr²⁺, and 41% by Cu²⁺, while the activity of GH18C was reduced 14% by Ni²⁺, 5% by Sr²⁺, and 53% by Cu²⁺. Hg²⁺ completely inhibited the activities of both domains (data not shown).

The GH18 domains have different activities on chitooligosaccharides.

The products formed from the activities of GH18_N and GH18_C on native chitooligosaccharides were determined. Native chitooligosaccharides (GlcNAc₄, GlcNAc₅, and GlcNAc₆) were incubated with purified samples of each polypeptide, and degradation products were labeled with 2-aminobenzoic acid and fractionated by gel electrophoresis. Consistent with exochitinase activity, the sole degradation product of GH18_N activity on GlcNAc₄ was chitobiose (Fig. 4A). Further, GH18_N released chitobiose primarily from GlcNAc₅, and GlcNAc₄ was not observed (Fig. 4B). When incubated with GlcNAc₆, GH18_N produced chitobiose and GlcNAc₄ but did not produce GlcNAc₃ or GlcNAc₅ (Fig. 4C). In contrast, GH18_C produced a mixture of chitooligosaccharides when acting on GlcNAc_4, GlcNAc_5, and GlcNAc₆ (Fig. 4), consistent with the ability of an endochitinase to cleave a chitooligosaccharide at any glycosidic linkage after the first bond at the nonreducing end. When incubated with 2-aminobenzoic acid-labeled chitohexose, GH18_N produced an increasing amount of labeled chitobiose over time, consistent with degradation from the nonreducing end. GH18_C activity on prelabeled chitohexose produced labeled GlcNAc₂, GlcNAc₃, and GlcNAc₄ (data not shown). The absence of labeled GlcNAc₅ suggests that the first glycosidic bond at the nonreducing end cannot be cleaved by this enzyme.

FIG. 4.

GH18_N and GH18_C exhibit exo- and endochitinase activities, respectively. GH18_N and GH18_C were purified as described in Materials and Methods and incubated with GlcNAc₄ (A), GlcNAc₅ (B), and GlcNAc₆ (C). Standards and reaction products were labeled with 2-aminobenzoic acid and fractionated by electrophoresis in a 15% polyacrylamide gel (8). Labeled chitooligosaccharides were visualized by using a UV transilluminator. Standards, from bottom, are chitobiose, chitotriose, chitotetrose, and chitopentose.

GH18_N and GH18_C function cooperatively to degrade native chitin.

The impact of the differential activities in a single reaction mixture, both when the catalytic domains are linked on a single polypeptide and when they are expressed as separate enzymes, was examined. Equivalent amounts (250 pmol) of GH18_N or GH18_C were added individually to native chitin to determine the rate at which each could release reducing sugars, an indication of depolymerization. GH18_N released 0.0158 μmol of reducing sugar/min when added to native chitin, whereas GH18_C released 0.0340 μmol/min (Table 2). A similar rate was measured at multiple time points during the initial 30 min of each reaction.

TABLE 2.

Activities of polypeptides containing ChiB catalytic domains on native chitin

Polypeptide components of reaction (pmol)			Rate of depolymerizationa
GH18N	GH18C	GH18N+C
250			0.0158 ± 0.0015
	250		0.0340 ± 0.0017
250	250		0.1190 ± 0.0081b
		250c	0.0645 ± 0.0031

^aRates are given as micromoles of reducing sugar released per minute per 250 or 500 pmol of catalytic domain (depending on the reaction components) Data are means and standard errors.

^bThe theoretical calculated rate (0.0498 μmol/min/500 pmol of catalytic domain) was determined by adding the observed rate for each domain acting alone and assuming no cooperative interaction.

^cEach molecule of GH18_N+C contains two catalytic domains; therefore, 500 pmol of catalytic domain was present.

To determine if the active domains function cooperatively to degrade chitin, equivalent amounts of each polypeptide were added to native chitin in a single reaction. If the domains act independently of each other, the theoretical combined rate of degradation should be greater than the sum of the two independent activities calculated above, i.e., 0.0498 μmol of reducing sugar/min/500 pmol of total protein. Consistent with the proposed endo- and exo- activities of each domain, the actual rate was 140% higher than the theoretical rate (Table 2).

Because in their native state the domains are linked on a single polypeptide, the rate of depolymerization was also measured when both catalytic domains were present and attached with their native linkage (Fig. 1). Full-length enzyme could not be used in these experiments because of difficulties in expressing the complete protein, possibly due to the serine-rich, 150-residue linker region at the amino terminus. The truncated form of ChiB lacking the postulated lipoprotein-anchoring site and linker region, GH18_N+C, was used instead (residues 156 to 1266). GH18_N+C released 0.0645 μmol of reducing sugar/min when 250 pmol of polypeptide (and therefore 250 pmol of each active domain) was added, an increase of 23% over the theoretical combined rate.

DISCUSSION

“M. degradans” strain 2-40 is known to efficiently metabolize chitin and many other insoluble complex polysaccharides (4). Analysis of the chitinolytic system of 2-40 revealed an unusual chitin depolymerase, ChiB, which appeared to include two catalytic domains. One of the catalytic domains of ChiB was shown here to function as an endochitinase, while the other functions as an exochitinase. ChiB is the first eubacterial chitinase demonstrated to contain two functional GH18 catalytic domains (9). The lack of carbohydrate binding domains and typical accessory domains (e.g., fibronectin type III domains or polycystic kidney disease domains) (19), coupled with the discrete activities of each catalytic domain, emphasizes the novelty of this enzyme.

When expressed as separate polypeptides, each GH18 domain of ChiB was able to depolymerize chitin in zymograms and was most active under similar temperature, pH, and ionic conditions. GH18_N was more active on MUF-diNAG than on MUF-triNAG and displayed a pattern of activity typical of an exochitinase on chitooligosaccharides. Chitobiose was released from the nonreducing end of GlcNAc₄, GlcNAc₅, and GlcNAc₆. Conversely, GH18_C released MUF most rapidly from MUF-triNAG and was able to cleave chitooligosaccharides at multiple linkages, demonstrating endochitinase activity. GH18_C was more than twice as active on native chitin as GH18_N; because native chitin has a paucity of free, exposed ends, exochitinases have far fewer sites at which they can act than do random-cutting endochitinases, which can cleave virtually any glycosidic linkage in the polymer. The synergistic degradation of chitin observed when both domains were present further supports their proposed function. The presence of both domains on separate polypeptides increased the release of reducing sugars 140% over the theoretical combined rate calculated if the domains were only to act additively. This synergism would not be observed if both domains had the same activity.

Carbohydrases with two catalytic domains are rare among prokaryotes. Only a small number have been characterized, mostly from ruminants and thermophiles. For example, Ruminococcus flavefaciens 17 (5) and Fibrobacter succinogenes S85 (17), produce xylanases with two catalytic domains, although the latter appears to encode a xylanase with two domains of the same function. Two extreme thermophiles, Anaerocellum thermophilum (a γ-subgroup proteobacterium) and T. kodakaraensis KOD1 (an archaeon), produce enzymes with two catalytic domains (23, 26). A. thermophilum produces a cellulase with separate GH9 and GH48 domains that encode endo- and exoglucanase activities, respectively. A chitinase from T. kodakaraensis, Tk-ChiA, was shown to have an amino-terminal exochitinase domain, while the carboxy terminus contains an endochitinase domain. Unlike ChiB of “M. degradans,” this enzyme also contains chitin binding domains and is not predicted to anchor to the cell surface. Further, the exolytic domain of Tk-ChiA is able to weakly cleave the third glycosidic linkage from the nonreducing ends of free chitin chains (23), an activity not observed in experiments with GH18_N.

The dual catalytic domains of ChiB function cooperatively to degrade chitin to chitobiose. Although maximal depolymerization was achieved when the catalytic domains of ChiB were on separate polypeptides, there are clear benefits to their presence as a single unit. First, a single promoter region is able to regulate the expression of two enzymatic activities. This permits two essential components of the chitinolytic system to be simultaneously regulated from a single locus, much like an operon regulating genes encoding a polycistronic mRNA. However, unlike for an operon, where several individual proteins are produced, a single enzyme is encoded. The amount of energy and secretion machinery needed to deliver two enzymatic functions to the exterior of the cell is therefore decreased. Second, encoding both activities on a single polypeptide ensures the proximity of the two domains during the in situ depolymerization of chitin. This allows for a synergistic and focused degradation of the polymer. In the environment, secreted enzymes may diffuse away from their intended targets and not be available to assist other components of a degradative system. This is partially solved by the presence of carbohydrate binding domains (which appear to be lacking from ChiB), but there is no assurance that both endo- and exo-acting enzymes will bind to the same location and have the opportunity to act in concert to achieve the full potential of the system unless they are linked on a single polypeptide.

When both domains were present on the same polypeptide, the synergism between the domains was less obvious. The activity detected when the domains are joined was only modestly increased over the theoretical activity when compared to the activities of the two catalytic domains as separate entities. The decreased activity of the domains when linked may be the result of the domains then moving as a single protein as each encounters substrate. For example, as the exolytic domain is cleaving soluble chitooligosaccharides, perhaps away from the insoluble polymer, the endolytic domain is unable to contact, and therefore degrade, its primary substrate. One can envision that the amount of reducing sugars released would increase if the domains were free to act at different locations. However, such an arrangement may not be of benefit in nature, where substrate is much more limited and less often encountered than in a laboratory reaction.

Based upon the data presented in this work and on the known properties of chitinases, a model of ChiB activity can be proposed (Fig. 5). Each catalytic site has been shown to be independently active, so the linkage between the domains may prevent interference between them during the degradation of chitin. The significance of the repetitive sequence in this region is unclear. The processive cutting nature of exochitinases and random cutting behavior of endochitinases have been described (18) and can be applied to the activity model of ChiB. As GH18_C releases chitooligosaccharides from the polymer, they can be immediately acted upon by GH18_N, which processively cleaves chitobiose from the nonreducing end. The lipoprotein acylation site present at the amino terminus of ChiB likely functions to anchor the enzyme to the outer membrane. This notion is strengthened by the observation that chitinase activity has been associated with outer membrane preparations of “M. degradans” (L. A. Whitehead and R. M. Weiner, unpublished observations). The membrane anchorage would keep two critical enzymatic activities in close proximity to the cell and perhaps eliminate the necessity of chitin binding domains. If this is the case, the importance of the catalytic domain arrangement within ChiB becomes apparent; chitooligosaccharides released by the activity of the distal GH18_C can be transferred to the exo-acting domain, which is in close proximity to the outer membrane where newly formed chitobiose can be taken up by the cell. The outer membrane localization of ChiB and other carbohydrases produced by “M. degradans” is currently being evaluated.

FIG. 5.

Model of chitin depolymerization by ChiB. ChiB is likely to attach to the surface of the cell via a lipoprotein anchor (black box). Activity of the endochitinolytic GH18_C releases chitooligosaccharides from polymeric chitin (stippled box). Free chitooligosaccharides (circles) are then acted upon by the exochitinolytic GH18_N that processively releases chitobiose from the nonreducing end. Free chitobiose would then be taken up by the cell and metabolized. The polyserine linkers (SSS) may provide flexibility to the enzyme and optimize interaction with substrates.