Structural and Functional Definition of the Human Chitinase Chitin-binding Domain*

Mammalian chitinase, a chitinolytic enzyme expressed by macrophages, has been detected in atherosclerotic plaques and is elevated in blood and tissues of guinea pigs infected withAspergillus. Its normal physiological function is unknown. To understand how the enzyme interacts with its substrate, we have characterized the chitin-binding domain. The C-terminal 49 amino acids make up the minimal sequence required for chitin binding activity. The absence of this domain does not affect the ability of the enzyme to hydrolyze the soluble substrate, triacetylchitotriose, but abolishes hydrolysis of insoluble chitin. Within the minimal chitin-binding domain are six cysteines; mutation of any one of these to serine results in complete loss of chitin binding activity. Analysis of purified recombinant chitin-binding domain revealed the presence of three disulfide linkages. The recombinant domain binds specifically to chitin but does not bind chitosan, cellulose, xylan, β-1,3-glucan, β-1,3–1,4-glucan, or mannan. Fluorescently tagged chitin-binding domain was used to demonstrate chitin-specific binding toSaccharomyces cerevisiae, Candida albicans, Mucor rouxii, and Neurospora crassa. These experiments define structural features of the minimal domain of human chitinase required for both specifically binding to and hydrolyzing insoluble chitin and demonstrate relevant binding within the context of the fungal cell wall.

Chitin, a ␤-1,4-linked polymer of N-acetylglucosamine, is one of the most abundant biopolymers in nature. It serves a structural role in arthropods, including crustaceans and insects, as well as mollusks, nematodes, and worms. It is also found in fungi, making up from less than 1% to more than 40% of the cell wall, depending on the organism. Chitin has not been found, however, in mammals. Nevertheless, several mammalian proteins with homology to fungal, bacterial, or plant chitinases have been identified (1)(2)(3)(4)(5)(6)(7). The only one of these proteins with demonstrated chitinolytic activity was identified as a tetra-Nacetylchitotetraoside hydrolase that is specific for the ␤-1,4linked N-acetylglucosamines of chitin (8,9). The enzyme was subsequently identified as a chitotriosidase that is dramatically elevated in serum from patients with Gaucher disease (10). Cloning of the cognate cDNA revealed several regions with high sequence homology to the catalytic domains of family 18 glycosyl hydrolases, including conservation of the putative active site (11).
The physiological purpose for a mammalian chitinase remains unproven. Because chitin is an integral component of fungal cell walls, a defensive function has been suggested for the enzyme (9,12). This has been supported by the observations that chitinase levels are strikingly elevated in both serum and tissues of guinea pigs infected with Aspergillus fumigatus (13,14). Another possibility is suggested by the observation that a distinct subset of lipid-laden macrophages in atherosclerotic lesions expresses chitinase (15). The authors speculate that the enzyme, along with other macrophage gene products, may modulate the extracellular matrix in the vessel wall and thus affect the downstream tissue remodeling events associated with atherogenesis.
Our laboratory has a long-standing interest in macrophage biology. As part of an effort to analyze gene expression patterns in macrophages, we determined the nucleotide sequences of nearly 3500 random cDNAs from monocyte-derived macrophages (16). Interestingly, the chitinase cDNA was the sixth most common transcript found, appearing at a frequency of 0.5%. The only transcripts appearing more frequently were those of the ferritin heavy and light chains, collagenase, ␤-actin, and annexin II, all serving either housekeeping or tissue remodeling and repair functions. Comparison of the predicted protein sequence of the chitinase open reading frame with chitinases from other species revealed that the protein is likely composed of two distinct domains. The putative catalytic domain, predicted on the basis of its extensive homology with the catalytic domains of the other chitinases (11), is contained within the N-terminal 75% of the protein. The remaining onequarter of the protein has limited homology with the chitinbinding domains of the Manduca sexta and Brugia malayi chitinases (11). Consistent with the possibility that this C terminus is a chitin-binding domain, Renkema et al. (17) describe a C-terminally truncated form of the human chitinase that can hydrolyze chitotriose but is unable to bind chitin.
The possibility of distinct autonomous domains in human chitinase provides an opportunity to dissect out the functional constraints that define the enzyme. In this report, we focus on the chitin-binding domain, describing structural features required for activity and the interaction of the domain with the fungal cell wall.

EXPERIMENTAL PROCEDURES
Generation of Expression Plasmids-A full-length chitinase clone was isolated from a human macrophage cDNA plasmid library (16). This clone was used as a template for generation of all other expression constructs. A truncated form in which the region encoding the Cterminal 72 amino acids is deleted was generated by the polymerase chain reaction (PCR) 1 using primers anchored on the translation initiation codon (5Ј-CGCAAGCTTGAGAGCTCCGTTCCGCCACATGGTGC-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ‡ To whom correspondence should be addressed: ICOS Corp., 22021 20th Ave. S.E., Bothell, WA 98021. Tel.: 206-485-1900; Fax: 206-486-0300; E-mail: ltjoelker@icos.com. GGTCTGTGGCCTGGG) and terminating on the codon encoding Thr-394 (5Ј-GACTCTAGACTAGGTGCCTGAAGGCAAGTATGG). This PCR product was digested with HindIII and XbaI and cloned into the corresponding sites of the mammalian expression plasmid, pcDNA3 (Invitrogen, Carlsbad, CA).
Mammalian Expression-Expression plasmids were introduced into COS-7 cells grown to 50% confluency in 60-mm tissue culture dishes by incubation in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) containing 0.5 mg/ml DEAE dextran, 0.1 mM chloroquine, and 10 g of plasmid DNA for 1.5 h. The cells then were treated with 10% dimethyl sulfoxide in phosphate-buffered saline (PBS) for 45 s, washed with serum-free medium, and incubated in Dulbecco's modified Eagle's medium supplemented with 1 mM L-glutamine, 100 units/ml penicillin, 100 g/ml streptomycin, and 10% fetal calf serum. After 4 days, culture media were harvested, and nonadherent cells were removed by centrifugation.
Yeast Expression-To produce recombinant chitin-binding domain, an expression construct was designed in which the nucleotides encoding the C-terminal 54 amino acids of the chitinase were fused to the 3Ј terminus of sequence encoding the Saccharomyces cerevisiae ␣-factor pre-pro sequence (18). The PCR was used to generate a chitin-binding domain-encoding fragment (sense primer, 5Ј-AAGAGGTACCTTTGGA-TAAAAGAAGCCCTGGACAAGACACGTTCTGCC; antisense primer, 5Ј-GAGAGCGGCCGCGACTTAATTCCAGGTGCAGCATTTGCAGG), which was digested with Asp718 I and NotI, ligated into an expression cassette, and cloned into pSAC/VEC, a modified version of the disintegration vector pSAC35 (19). This shuttle vector is composed of a complete two micron plasmid with a LEU2d selectable marker and two repeated sequences flanking the pUC-derived E. coli origin of replication and ␤-lactamase resistance marker. The resulting pSAC2/AF/ chitin-binding domain plasmid was transformed into the S. cerevisiae host strain IE41 (19) (cir o leu2 pep4::URA3 L261) and selected by growth on leucine-deficient media.
A selected clone was grown overnight at 30°C in 10 ml of synthetic complete (BioWhittaker)-leu-ura medium containing 2% glucose. One ml of this culture was seeded into 100 ml of the same medium and incubated at 30°C with shaking. After 19 h, 65 ml of the culture was seeded into 1.2 liters of medium in a 3-liter fermentor. The fermentation was carried out at 30°C, pH 5.5, with 1200 rpm agitation and an airflow rate of 3 liters/min. After 93 h, cells were harvested by centrifugation. The supernatant was filtered through a prefilter and a 0.2 m filter and stored at Ϫ20°C until purification of the recombinant protein.
Purification of Recombinant Chitin-binding Domain-The recombinant chitin-binding domain was purified by passing filtered culture medium through a chitin bead column. A 25-ml bed volume of chitin beads (New England Biolabs, Beverly, MA) was prepared in a 50-ml column (Amicon, Beverly, MA), prewashed with 250 ml of 1% SDS and equilibrated with 250 ml of Buffer A (20 mM Tris, pH 8, 500 mM NaCl) at a flow rate of 2 ml/min. Clarified medium was passed through the chitin bead column at a rate of 1 ml/min. Following a 250-ml wash with Buffer A, protein was eluted from the beads with 50% acetonitrile, 0.1% trifluoroactetate and collected into 4-ml fractions. Acetonitrile was evaporated from the eluate by vacuum centrifugation. Analysis of the purified protein by matrix-assisted laser desorption ionization time-offlight mass spectrometry showed a single peak with molecular mass of 5911.9 Da, a value that corresponds favorably to the predicted mass of 5909.6 Da. Chitin binding activity was confirmed by readsorbing the purified protein to chitin beads and eluting with 1% SDS.
Antibodies-Monoclonal antibodies were generated using standard protocols (20). Hybridomas generated from BALB/c mice immunized with recombinant full-length chitinase were screened for chitinasespecific antibodies by enzyme-linked immunosorbent assay. Antibody preparations from positive clones were further tested for specificity using Western analysis of complex tissue culture media from cells expressing recombinant chitinase. Binding sites of specific antibodies were mapped to either the N-terminal three-fourths of the molecule or the chitin-binding domain by Western blotting of recombinant truncated chitinase or chitin-binding domain.
Alkaline Phosphatase Activity Assay-Tissue culture media from cells transfected with the SEAP-chitin-binding domain fusion expression constructs were assayed for alkaline phosphatase activity as described previously (21,22). Briefly, samples were incubated at 65°C for 10 min to inactivate serum-derived phosphatases. Serial 10-fold dilutions were mixed with an equal volume of a 2ϫ solution containing 2 M diethanolamine, pH 9.8, 1 mM MgCl 2 , 1 mg/ml bovine serum albumin, 20 mM L-homoarginine, and 24 mM p-nitrophenylphosphate. After a 30-min incubation at 37°C, substrate metabolism was assessed by determination of the A 410 .
Substrate Binding Activity Assays-To determine chitin binding activity of COS-7 cell-produced fusion proteins, crab shell chitin (Sigma) was ground to a fine powder using a mortar and pestle, washed three times with Buffer A, and resuspended to 100 mg/ml. One hundred l of the chitin suspension was added to 1 ml of transfected COS-7 cell medium, and the mixture was incubated 4 h at 4°C with continuous end-over-end mixing. Following incubation, the chitin was pelleted by centrifugation (5 min at 12,000 ϫ g). Supernatants were either assayed for SEAP activity (in the case of SEAP fusion proteins) or evaluated by polyacrylamide gel electrophoresis followed by Western blotting. For Western blotting, equivalent volumes of supernatant were supplemented with Laemmli buffer and 50 mM dithiothreitol, boiled, and electrophoresed through a 12% polyacrylamide gel (Novex, San Diego, CA). Contents of the gel were subsequently transferred to polyvinylidene difluoride membrane (Novex) and detected using a monoclonal antibody specific for the N-terminal portion of chitinase.
Binding specificity of the chitin-binding domain was determined by incubating the purified, recombinant material with the following insoluble polysaccharides: chitin beads (New England Biolabs), crab shell chitin, chitosan, carboxymethyl cellulose, curdlan, lichenan, mannan, and xylan (all from Sigma). Five hundred l of recombinant chitinbinding domain at 200 g/ml in Buffer A was added to 100 g of washed matrix. After incubation at 23°C for 30 min on a rocking platform, the insoluble material was removed by centrifugation. Chitin-binding domain remaining in solution was detected by electrophoresing equivalent volumes of the supernatants through a 4 -20% gradient polyacrylamide gel (Novex) and staining the gel with Coomassie Blue stain (Gelcode, Pierce). Pelleted polysaccharides were washed twice with Buffer A and boiled in Laemmli buffer with 50 mM dithiothreitol. The resulting material in solution was electrophoresed and stained in parallel to the supernatants described above.
Chitinase Activity Assays-Chitotriosidase activity was assayed as described (10) using 4-methylumbelliferyl-␤-D-N,NЈ,NЉ-triacetylchitotriose (Sigma) as the substrate. To assay for hydrolytic activity on insoluble chitin, ground and washed crab shell chitin (see above) was suspended to 0.4% in 1.5% molten agarose, 20 mM sodium phosphate, pH 6. The mixture was poured to a thickness of 2 mm in a 10-cm Petri dish and allowed to solidify. Wells 3 mm in diameter were cut into the agarose/chitin matrix. Prior to loading wells, the recombinant proteins in media from transfected COS-7 cells were concentrated 70-fold over a 30,000 molecular weight cutoff filter device (Amicon). Equivalent quantities of full-length and C-terminally truncated recombinant proteins were loaded into adjacent wells. A third well was loaded with concentrated medium from mock-transfected COS-7 cells. Plates were incubated for 48 h at 37°C in a humidified chamber. After incubation, zones of clearing around the well indicate hydrolysis of the chitin.
Synthesis of Rhodamine-labeled Chitin-binding Domain-The chitin-binding domain (C-terminal 54 amino acids of human chitinase) was synthesized using in situ neutralization/HBTU activation protocols for Boc chemistry (23). After synthesis, the final Boc group on the N terminus was removed, and 6-carboxytetramethylrhodamine, succinimidyl ester (Molecular Probes, Eugene, OR) was coupled selectively to the N-terminal amine group. Removal of side chain protecting groups and cleavage of polypeptide from resin was achieved by treatment with hydrogen fluoride. The labeled polypeptide was purified and analyzed by reverse phase high pressure liquid chromatography and electrospray mass spectrometry. The protein was folded and then purified again by reverse phase high pressure liquid chromatography.
Detection of Fungal Chitin-A colony of S. cerevisiae (ATCC catalog no. 9763) was inoculated into 3 ml of YEPD medium containing 2% glucose and grown at 37°C for 16 h. Twenty l of an overnight culture derived from a single colony of Candida albicans (strain 32033) was used to innoculate 3 ml of YEPD medium containing 2% glucose and 10% fetal bovine serum (BioWhittaker) and was incubated at 37°C for 16 h. Mucor rouxii (ATCC catalog no. 44260) was grown as yeast by inoculating spores into YPG medium (24) and incubating with stirring at 23°C for 3 days under continuous CO 2 flow. Neurospora crassa hyphal walls were generously provided by Dr. Claude Selitrennikoff (University of Colorado). Cells of all four organisms were pelleted by centrifugation, washed with PBS, and resuspended to high turbidity in PBS. To detect chitin, 30 l of washed cells was mixed with either no fluorophore or with calcofluor white (BactiDrop, Remel, Lenexa, KS), 2 g/ml of wheat germ agglutinin-FITC (Sigma), or 2 g/ml of synthetic chitin-binding domain-rhodamine. Final volume was adjusted to 50 l with PBS. To demonstrate that any labeling seen with chitin-binding domain-rhodamine was due to the chitin-binding domain, we included control tubes in which only the rhodamine component was added to cells at 20 g/ml. For staining with calcofluor white, 30 l of washed cells was pelleted in a microcentrifuge tube then resuspended in 30 l of BactiDrop. All reactions were incubated on a rocking platform at 23°C for 2 h in the dark. Following incubation, cells were pelleted, washed twice with PBS, and resuspended in ProLong mounting medium (Molecular Probes). A drop of suspended cells was applied to a glass slide and overlaid with a coverslip, and staining patterns were observed with fluorescence microscopy. Autofluorescence was negligible in control cells incubated in the absence of fluorescent reagents. Fig. 1 illustrates the proposed domain structure of human chitinase and presents the sequence of the C-terminal 72 amino acids. To assess the functional importance of the C-terminal portion of the enzyme, full-length chitinase and a truncated form lacking the C-terminal 72 amino acids were expressed in COS-7 cells. Chitin was incubated with media from the transfected cells then removed by centrifugation. As shown in Fig. 2, full-length chitinase was removed from solution by chitin, but the truncate was not. This is consistent with the prior observation that the C terminus is required for binding of the enzyme to native chitin (17).

Functional Definition of the Chitin-binding Domain-
The enzymatic activities of the two forms of chitinase were also compared. Using equivalent molar quantities of each form, we found equivalent substrate turnover when the soluble substrate, 4-methylumbelliferyl-␤-D-N,NЈ,NЉ-triacetylchitotriose was used (Fig. 3A). This was not the case, however, with native chitin as substrate. In an agar diffusion assay in which particulate crab shell chitin was incorporated into agarose, the fulllength chitinase generated a zone of clearing around the well of origin (Fig. 3B). In contrast, no clearing of the substrate could be detected around the well containing the truncated enzyme. These observations suggest that although the chitin-binding domain does not appear to affect the catalytic site of the enzyme, its presence is essential for the enzyme to act upon insoluble chitin.
Definition of the Minimal Chitin-binding Domain-The above experiments demonstrate that the C-terminal 72 amino acids of chitinase are sufficient to mediate all of its binding to native chitin. To determine the minimal number of residues necessary to accomplish this binding, we designed a series of expression constructs in which chitin-binding domain derivatives were fused to the C terminus of the human placental SEAP (21). Media from COS-7 cells transfected with the constructs were assayed for SEAP activity both before and after incubation with crab shell chitin. Depletion of SEAP activity from the media after incubation with chitin indicates a chitinbinding domain-mediated precipitation of the fusion protein.
As presented in Fig. 4, 80% of the fusion protein containing the full 72-residue chitin-binding domain (construct 395-466) was depleted from the medium by chitin. This level of depletion was retained with the removal of up to 23 amino acids from the N terminus of the domain (construct 418 -466). Removal of three additional residues, however, completely abolished the ability of chitin to precipitate the fusion protein from the medium (construct 421-466). Interestingly, the C terminus of the domain is critical; deletion of as few as two amino acids completely abolished binding activity (construct 413-464). Thus, the minimal chitin-binding domain is composed of the most C-terminal 49 amino acids of chitinase. Production and Activity of Recombinant Chitin-binding Domain-In order to generate pure chitin-binding domain for further characterization, we constructed a S. cerevisiae expression plasmid that encodes the C-terminal 54 amino acids of chitinase fused to the yeast ␣-factor pre-pro sequence (18). Cells transfected with this construct routinely produced 0.85-1.0 mg/ml of mature chitin-binding domain. The recombinant material was purified to Ͼ95% homogeneity from culture medium with a single passage over chitin beads. Interestingly, elution from the beads required harsh conditions, such as 1% SDS or 50% acetonitrile (Table I). This indicates a strong interaction between the peptide and chitin.
Once removed from the eluting agent, the purified chitinbinding domain retained its chitin binding activity. To determine whether the domain is specific for chitin, we incubated various insoluble polysaccharides with the recombinant material. After chitin-binding domain-polysaccharide complexes were removed by centrifugation, supernatants, as well as precipitated protein, were electrophoresed and stained (Fig. 5). The chitin-binding domain was largely specific for chitin, binding both chitin beads and raw crab shell chitin. Little or no precipitation of the binding domain was affected by chitosan, cellulose, mannan, xylan, or the ␤-glucans curdlan and lichenan.
Role of Cysteines in Chitin Binding Activity-Many substrate-binding domains of proteins that metabolize polysaccharides are rich in cysteine residues (25). Interestingly, the 49residue minimal chitin-binding domain of human chitinase contains all six of the cysteine residues present in the Cterminal half of the protein (Fig. 1). To test whether these residues are required for chitin binding activity, we constructed six additional SEAP-chitin-binding domain fusion constructs in which each cysteine was individually mutated to a serine. Fig. 6 shows that, unlike the control fusion, none of the mutants could be precipitated from transfected COS-7 media with crab shell chitin. Therefore, each cysteine is critical for the chitin binding activity of the domain.
The indispensability of each cysteine suggests that each may be engaged in disulfide bond formation. In support of this, the average mass of the recombinant chitin-binding domain observed by electrospray ionization mass spectroscopy was 5908.2 Ϯ 1 Da. This is consistent with the predicted mass of 5909.6 Da for the unreduced polypeptide. Furthermore, chemical analysis of the recombinant polypeptide using DTNB revealed that there are no free sulfhydryl groups. Preliminary characterization of disulfide linkages using proteolytic cleavage of the domain followed by purification, mass determination, and sequencing of the fragments revealed that Cys-420 and Cys-440 are linked to each other (not shown). However, because of their close proximity, exact assignment of linkages to the remaining cysteines has not been possible.
Binding of Chitin-binding Domain to Fungal Cells-To test whether the chitin-binding domain is able to bind chitin in the context of the fungal cell wall, we compared fungal labeling patterns of chemically synthesized, rhodamine-tagged chitinbinding domain with two other chitin-binding molecules, calcofluor white and FITC-tagged wheat germ agglutinin. Fig. 7 shows labeling of C. albicans and M. rouxii by the three reagents. In general, all three reagents detected the same structures, although there were subtle differences. In C. albicans, the bud scars and chitin rings of budding yeast were readily detected. These structures were also labeled in S. cerevisiae yeast, although a diffuse background staining of the entire cell by chitin-binding domain-rhodamine tended to obscure the chitin-rich structures (not shown). In C. albicans hyphae, chitin-binding domain-rhodamine brightly labeled both hyphal  1 and 2) and chitinase with the C-terminal 72 amino acids deleted (lanes 3 and 4) (see Fig. 1) were expressed in COS-7 cells. Culture media from transfected COS-7 cells were incubated without (lanes 1 and 3) or with (lanes 2 and 4)  tips and septa. Calcofluor white produced a similar labeling pattern, but wheat germ agglutinin-FITC labeled few if any hyphal tips or septa. Similar labeling patterns were observed with another hyphal fungus, N. crassa (not shown). M. rouxii normally grows as branched, nonseptate hyphae, but when incubated under CO 2 , the yeast form predominates (24). Chitin-binding domain-rhodamine brightly labeled a thick layer surrounding the yeast cell. This layer was contiguous with a thinner labeled layer on the occasional hyphae emanating from a yeast cell. Incubation of these cells with calcofluor resulted in bright labeling of the entire cell with a slight preference for the outer cell wall structure. Wheat germ agglutinin-FITC labeling of the cell was similar to that of the chitinbinding domain except that it also labeled bud scars. These results demonstrate that in general, the chitin-binding domain binds fungal cell wall chitin in patterns that are consistent with those exhibited by other chitin-binding molecules. That the binding is chitin-specific is also supported by a complete lack of labeling of non-chitin-containing COS-7 and E. coli cells (not shown). DISCUSSION Human chitinase, one of the most abundant proteins produced by monocyte-derived macrophages, was identified as a Culture media from transfected cells were incubated with or without crab shell chitin, centrifuged, and then assayed for SEAP activity. SEAP activity remaining in culture media after incubation with and subsequent removal of chitin is expressed as a percentage of the total activity in the same medium without precipitation with chitin.

TABLE I Elution of recombinant chitin-binding domain from chitin beads
Recombinant chitin-binding domain was adsorbed to chitin beads. Starting material and eluates were adjusted to equivalence based upon assumption of 100% recovery and then electrophoresed and stained with Coomassie Blue. Percentage of recovery was estimated according to relative band intensity.  6. Each of the six cysteines within the chitin-binding domain is essential for binding activity. Each cysteine (see Fig. 1) was independently mutated to a serine, and the resulting modified chitinbinding domains (413-466) were expressed as fusions with SEAP and assayed for chitin binding activity (see Fig. 4 legend). Cysteines are numbered according to position within the full-length chitinase polypeptide.
protein highly elevated in the serum of patients with Gaucher disease (10). The protein is composed of two discrete domains. The N-terminal three-fourths is homologous to a variety of well defined chitinases and has chitotriosidase activity (11). The remainder of the protein was proposed to mediate chitin binding activity because its absence rendered the protein impervious to precipitation by chitin (17). However, despite the elegant investigative efforts applied to chitinase, its physiological function remains unknown.
In order to understand more fully the interaction of the human chitinase with its substrate, we set out to define the chitin-binding domain. Using both a positive and a negative approach, we have confirmed that the C-terminal portion of human chitinase mediates binding to chitin. A recombinant truncate lacking the C-terminal 72 amino acids does not bind chitin. The same C-terminal 72 residue fragment, expressed as a recombinant protein, does bind chitin. By creating a series of nested deletions, we found that the C-terminal 49 amino acids are necessary and sufficient for binding. The idea that the chitinase is composed of two discrete domains is supported by the observation that the C-terminal region is not required for hydrolysis of a soluble substrate (Ref. 17 and this report). This is also true of a number of cellulases in which enzymatic activity is not compromised by the removal of the cellulose-binding domain (reviewed in Ref. 26). The substrate-binding domain does appear to be more critical for hydrolysis of insoluble substrates, however. For example, removal of the chitin-binding domain from a rye seed chitinase reduces its activity against colloidal chitin (27). This is also the case for the human chitinase; the C-terminal truncate showed no hydrolytic activity against insoluble chitin in an agar diffusion assay. In vivo, the chitin-binding domain is likely to be critical for targeting the enzyme to its substrate, whether it be chitin in the fungal cell wall or some as yet undefined polysaccharide associated with the mammalian cell membrane or interstitium.
Minimally 49 amino acids in length, the chitin-binding domain depends upon the structural stability imparted by three internal disulfide linkages for function. In fact, the N-terminal boundary of the domain is likely defined by Cys-420, because deletions N-terminal to it are tolerated, but any deletions in which it is removed are crippling. Furthermore, we found that mutation of Cys-420 or any of the other five cysteines within the binding domain to serine renders it inactive. The cysteinerich character of this region is reminiscent of the substratebinding domains of the invertebrate and plant families of chitin-binding proteins (25). These domains have six or eight cysteines, respectively, all of which are linked by disulfide bonds. Apart from the structural constraints imposed by the disulfides, there are no other secondary structural features. Preliminary circular dichroism analysis of the chitinase chitinbinding domain produced a spectrum consistent with a similar lack of secondary structure. At the C terminus of the domain, it was surprising to find that deletion of just two amino acids completely abolished binding activity. Because this deletion does not remove any of the cysteines, little effect on the structural integrity of the domain would be expected. It is possible that the penultimate residue, a tryptophan, is required for interaction with the substrate. Aromatic residues are highly conserved in polysaccharide-binding proteins of a wide variety of invertebrates (25,26). Conserved tryptophans are also found in the starch-binding domains of a microbial family of enzymes that degrade starch. At least two of these conserved residues are required for binding activity and are predicted to participate in binding to glucosyl units through hydrophobic or hydrogen bonding (28).
We found that once bound, the chitin-binding domain requires harsh conditions such as 1% SDS or 50% acetonitrile to elute it from chitin. The kinetics of this interaction are likely to be complex, as suggested by the fact that we found it virtually impossible to achieve saturable binding of the domain to chitin beads (not shown). Others have reported similar binding properties of a chitin-binding domain from an Alteromonas species (29). Efforts to more fully define the mechanisms and kinetics of binding to chitin are ongoing. In addition to its strong affinity for chitin, the chitin-binding domain also appears to be highly specific. We could precipitate it from solution only with chitin, not with the deacetylated form of chitin, chitosan, or with two other polysaccharides with ␤-1,4-glycosidic bonds, cellulose and xylan. Additionally, neither the ␤-1,3nor ␤-1,3-1,4-glucans nor mannan was able to substantially precipitate the binding domain. This suggests that the binding domain is tightly constrained to a specific ligand composition and conformation. Assuming that the natural ligand is an insoluble polysaccharide, such strict ligand specificity is relevant to the ongoing discussion regarding the physiological function of the human chitinase. If, as has been suggested (15), the enzyme participates in tissue remodeling during embryogenesis or certain pathophysiological states, it is probable that the substrate is chitin or is very similar to chitin. The chitin-binding domain may be a useful probe for identifying such a polysaccharide in vertebrate tissue.
Alternatively, an exogenous substrate may be the main target of human chitinase. Because the macrophage serves as one of the primary defense mechanisms against invading fungal pathogens (30), it is reasonable to conclude that chitinase, copiously produced by macrophages, contributes to the defensive activity of macrophages by degrading fungal chitin (9,12). Consistent with this is the high affinity and specificity of the chitin-binding domain for insoluble chitin. To test whether the domain is able to bind chitin enmeshed within the complex milieu of the fungal cell wall (31)(32)(33), we compared fungal labeling patterns of fluorescently tagged chitin-binding domain with those of calcofluor and wheat germ agglutinin. In all four organisms tested, the chitin-binding domain labeled the same chitin-rich structures detected by the two chitin-staining control reagents. Because we have demonstrated that the chitinbinding domain is required for hydrolysis of insoluble chitin, it follows that binding to substrate is a prerequisite for its degradation. These labeling experiments demonstrate that the binding requirement is met, thus supporting the hypothesis that the enzyme hydrolyzes fungal chitin. Whether this would ultimately result in fungistasis or fungal death remains to be tested. However, as a general observation, defense mechanisms against fungal pathogens or competitors in plant, fungal, and bacterial models often invoke combinations of effectors. For example, in the pea, Pisum sativum, a chitinase and a ␤-1,3glucanase synergize in fungicidal activity (34). Similarly, the yeast lytic system of Arthrobacter sp. relies upon both protease and glucanase activities, neither of which has lytic activity alone (35,36). Therefore, it is reasonable to consider the macrophage chitinase an important component of a defense system that may include other enzymes and/or immunoglobulin, complement, and reactive oxygen species. Another intriguing possibility to consider is a potential fungistatic activity of the chitin-binding portion of the enzyme independent of any enzymatic activity. It has been reported that the stinging nettle lectin, Urtica dioica aglutinin, is capable of inhibiting phytopathogenic fungal growth, possibly by cross-linking chitin at the hyphal apex (37). A homologous chitin-binding domain from the chestnut seed endochitinase, Ch3, appears to interfere with hyphal growth independent of the chitinase activity of the enzyme (38). On the other hand, a chitin-binding lectin from Streptomyces olivaceoviridis, CHB1, bound to fungal hyphae but exhibited no antifungal activity (39). Additional structure-function analyses are needed to determine whether the human chitin-binding domain may exert antifungal activity independent of the catalytic activity of the enzyme.