Stomach Chitinase from Japanese Sardine Sardinops melanostictus: Purification, Characterization, and Molecular Cloning of Chitinase Isozymes with a Long Linker

Fish express two different chitinases, acidic fish chitinase-1 (AFCase-1) and acidic fish chitinase-2 (AFCase-2), in the stomach. AFCase-1 and AFCase-2 have different degradation patterns, as fish efficiently degrade chitin ingested as food. For a comparison with the enzymatic properties and the primary structures of chitinase isozymes obtained previously from the stomach of demersal fish, in this study, we purified chitinase isozymes from the stomach of Japanese sardine Sardinops melanostictus, a surface fish that feeds on plankton, characterized the properties of these isozymes, and cloned the cDNAs encoding chitinases. We also predicted 3D structure models using the primary structures of S. melanostictus stomach chitinases. Two chitinase isozymes, SmeChiA (45 kDa) and SmeChiB (56 kDa), were purified from the stomach of S. melanostictus. Moreover, two cDNAs, SmeChi-1 encoding SmeChiA, and SmeChi-2 encoding SmeChiB were cloned. The linker regions of the deduced amino acid sequences of SmeChi-1 and SmeChi-2 (SmeChi-1 and SmeChi-2) are the longest among the fish stomach chitinases. In the cleavage pattern groups toward short substrates and the phylogenetic tree analysis, SmeChi-1 and SmeChi-2 were classified into AFCase-1 and AFCase-2, respectively. SmeChi-1 and SmeChi-2 had catalytic domains that consisted of a TIM-barrel (β/α)8–fold structure and a deep substrate-binding cleft. This is the first study showing the 3D structure models of fish stomach chitinases.

Regarding fish stomach chitinase, although the primary structure and the enzymatic properties of mainly fish stomach chitinases of demersal fish that feed on shrimp, crab, and squid have been elucidated, there is little information concerning the primary structure and the enzymatic properties of surface fish chitinases.There is also no information concerning 3D structure models of fish stomach chitinases.
In the present study, we aimed to purify chitinase isozymes from the stomach of Japanese sardine Sardinops melanostictus, a surface fish that feeds on plankton, and we characterized the properties of these isozymes.For a comparison with the primary structures of chitinase cDNAs obtained previously from the stomach of demersal fish [16,17,21,22], we cloned the cDNAs encoding chitinases from the stomach of S. melanostictus and determined their primary structures.We also predicted the 3D structure models using the primary structures of S. melanostictus stomach chitinases, and we compared them with those of other previously reported chitinases.To our knowledge, this is the first study showing the 3D structure models of fish stomach chitinases.

Purification of SmeChiA and SmeChiB
After ammonium sulfate fractionation, the enzyme solution was applied to a Chitin EX column, and chitinase activity was detected in both non-adsorbed and adsorbed fractions.The adsorbed fraction was further purified by cation exchange chromatography using a TOYOPEARL CM-650S column and obtained chitinase active fraction (SmeChiA).The non-adsorbed fraction was fractionated by anion exchange chromatography using a TOYOPEARL DEAE-650S column and further purified a TOYOPEARL CM-650S column and obtained chitinase active fraction (SmeChiB).
When pNp-(GlcNAc) 3 was used as a substrate, SmeChiB showed maximum activity at pH 5.0, whereas SmeChiA showed 34% or less of the maximum activity.As shown in Figure 2, SmeChiA worked well in the acidic pH range compared to SmeChiB.The relative activity toward pNp-(GlcNAc) 2 and pNp-(GlcNAc) 3 of both chitinases exhibited 30% or less of the maximum activity at pH 7.0.
We used HPLC to analyze the hydrolysis products of (GlcNAc) 5 by SmeChiA and SmeChiB and compared them with those of other reported fish chitinases (Table 2).Both chitinases hydrolyzed (GlcNAc) 5 to produce (GlcNAc) 2 + (GlcNAc) 3 , and SmeChiA hydrolyzed the second and third glycosidic bonds at the rates of 80.8% and 19.2%, respectively.In contrast, SmeChiB degraded the second and third glycosidic bonds at the rates of 29.6% and 70.4%, respectively.

Substrate Specificities of SmeChiA toward Insoluble Substrates
We measured the substrate specificity of SmeChiA toward insoluble substrates by using crystalline α-chitin (crab shell chitin and shrimp shell chitin), crystalline β-chitin (squid pen chitin), chitin nanofibers of crystalline chitin, and non-crystalline colloidal chitin (Table 3).SmeChiA exhibited markedly high activity (33.4 U/mg) toward chitin nanofiber, which was 3-to 4.3-times higher than the activities of PtChiA [16], SmChiA, and SmChiB [17].The second highest activity of SmeChiA was toward squid pen β-chitin (1.45 U/mg).The activities of SmeChiA toward α-chitin, which has the most rigid crystalline structure, were as follows in descending order: crab shell chitin (0.922 U/mg) > shrimp shell chitin (0.303 U/mg).When crab shell chitin was used as a substrate, the SmeChiA activities were second highest among the other fish stomach chitinases after PaChiB [15].SmeChiA also efficiently degraded non-crystalline colloidal chitin (Table 3).These results indicate that SmeChiA exhibits wide substrate specificity toward crystalline chitin.

Molecular Cloning of the Two Chitinase cDNAs
We cloned the two cDNAs SmeChi-1 encoding SmeChiA purified from the stomach of S. melanostictus, and SmeChi-2 encoding SmeChiB.The sequences determined for the two cDNAs encoding the chitinases SmeChi-1 and SmeChi-2 were registered with the DNA Data Bank of Japan (DDBJ) database (accession nos.AB985610 for SmeChi-1 and AB985611 for SmeChi-2).The SmeChi-1 cDNA was cloned up to 1,543 bp and contains an ORF of 1,515 bp encoding 505 amino acids.The SmeChi-2 cDNA was cloned up to 1,622 bp and contains an ORF of 1,584 bp encoding 528 amino acids (Figure 3).
SmeChi-1 and SmeChi-2 contained N-terminal signal peptides, GH18 catalytic domains, linker regions, and C-terminal chitin-binding domains (Figure 3), and the linker regions of SmeChi-1 and SmeChi-2 are the longest among the fish stomach chitinases identified thus far.Parts of the amino acid sequences of the GH18 catalytic domains of both enzymes were detected as the sequence of the "DXDXE" active-site motif, which conserves sequences of GH family 18 chitinases including other fish stomach chitinases (Figure 3).In addition, the 25-residue N-terminal amino acid sequence of the GH18 catalytic domain of SmeChi-1 was completely concordant with those of purified SmeChiA and that of SmeChi-2 was also completely concordant with that of purified SmeChiB (Figure 3).

Tissue Expressions of SmeChi-1 and SmeChi-2
We analyzed the expressions of SmeChi-1 and SmeChi-2 in different S. melanostictus tissues by RT-PCR using the housekeeping gene, β-actin, as a control.SmeChi-1 mRNAs were found in the stomach, intestine, testis, and ovary (Figure 4).The fChi1 mRNA of Japanese flounder Paralichthys olivaceus was found predominantly in the stomach, and at a lower level, in the testis and ovary [21], and the distribution of SmeChi-1 mRNA was similar to that of fChi1.SmeChi-2 mRNA was found in the stomach, which is concordant with the distributions of PtChi-1 and PtChi-2 mRNA in P. trilineatum [16].As shown in Figure 5, SmeChi-1 was expressed more strongly than SmeChi-2 as well as PtChi-1 [16].

Prediction of the 3D Structural Models of SmeChi-1 and SmeChi-2
We predicted the 3D structures of SmeChi-1 and SmeChi-2 by using the structure of the catalytic domain of acidic mammalian chitinase from Homo sapiens (Protein Data Bank [PDB] ID: 3FY1) and those of chitin-binding proteins from Tachypleus tridentatus (PDB ID: 1DQC) as a template.SmeChi-1 and SmeChi-2 had catalytic domains that consisted of a TIM-barrel (β/α) 8 -fold [8,9] structure.It is thought that this structure is a characteristic of GH18 family chitinases and that it hydrolyzes the β-1, 4 linkage of chitin by its own catalytic mechanism (Figure 6).In addition, the 3D structures of SmeChi-1 and SmeChi-2 revealed a deep substrate-binding cleft similar to that of Bacillus circulans chitinase A1 [32], and this cleft is needed to important in processive hydrolysis of the chitin chain [33].The present results showed that SmeChiA efficiently degraded crystalline chitin and exhibited especially high activity toward chitin nanofibers (Table 3).
SmeChi-1 and SmeChi-2 have the longest linker regions among the fish stomach chitinases identified to date (Figure 3).SmeChiA was adsorbed by affinity columns when chitin was used as a carrier.In light of these results, it seems that SmeChiA has chitin binding ability hydrolyze chitin with moving on the surface of chitin.Our present findings also suggest that the catalytic domain of SmeChiA can hydrolyze a wide area of chitin by using the longest linker region and the chitin binding domain.This mechanism may underlie the high chitinase activity toward crystalline chitins.

Chitinase Activity Assay
We assayed the chitinase activity using various substrates.First, pNp-(GlcNAc) 2 and pNp-(GlcNAc) 3 were used as a substrate to measure enzyme activity during the purification of chitinases.When pNp-(GlcNAc) n (n = 1-4) was used as the substrate, the enzyme activity was assayed by the method of Ohtakara [35].Briefly, 25 µL of enzyme solution and 10 µL of 4 mM pNp-(GlcNAc) n were added to 25 µL of 0.2 M phosphate-0.1 M citrate buffer (pH 4.5) and incubated for 10 min at 37 ˝C.After incubation, 100 µL of 0.2 M sodium carbonate solution was added, and the absorbance of the released p-nitrophenol was measured at 420 nm.One unit of enzyme activity was defined as the amount of enzyme releasing 1 µmol of p-nitrophenol per min at 37 ˝C.
The hydrolysis products of (GlcNAc) 5 produced by SmeChiA and SmeChiB and their anomer formation ratios were analyzed according to the method of Koga et al. [36].Briefly, 5 µL enzyme solution and 25 µL 0.22 mM (GlcNAc) n were added to 25 µL 0.1 M sodium acetate buffer (pH 4.0), and the mixture was incubated for 10 min at 25 ˝C.The reaction was stopped by cooling to 0 ˝C in an ice bath.The reaction solution was analyzed at 25 ˝C by high-performance liquid chromatography (HPLC) using a TSK-GEL Amide-80 column (4.6 mm dia.ˆ250 mm, Tosoh, Tokyo, Japan).(GlcNAc) 5 was eluted with 70% acetonitrile solution at a flow rate of 0.8 mL/min, and the absorbance was measured at 210 nm.
When 0.5% colloidal chitin, 0.5% αor β-chitin, or 1% chitin nanofiber was used, the enzyme activity was assayed by the method of Ohtakara [35].Briefly, 250 µL of enzyme solution and 250 µL of substrate solution were added to 500 µL of 0.2 M phosphate-0.1 M citrate buffer (pH 4.5), and the mixture was incubated for 2 h at 37 ˝C with shaking.After the incubation, the reaction was stopped by boiling for 3 min.The reaction solution was centrifuged, and 375 µL of the supernatant was sampled.To measure the amount of reducing sugar produced by the enzymatic reaction, 500 µL of Schales' reagent was added to the collected solution, and the absorbance was measured at 420 nm.The solution was then boiled for 15 min and cooled in running water.The absorbance was then measured again at 420 nm.The standard curve was prepared using authentic GlcNAc, and the absorbance was then converted into the amount of GlcNAc.One unit of enzyme activity was defined as the amount of enzyme required to degrade substrates at 37 ˝C and produce reducing sugars corresponding to 1 µmol GlcNAc per min.

Effect of pH on Chitinase Activity
When pNp-(GlcNAc) 2 and pNp-(GlcNAc) 3 were used as a substrate, the optimum pH was determined by assaying the enzyme activity.Briefly, the solution was incubated for 10 min at 37 ˝C in 0.1 M sodium acetate-0.1 M HCl buffer (pH 1.0-2.0)or 0.2 M phosphate-0.1 M citrate buffer (pH 2.5-8.0) for SmeChiA and 0.2 M phosphate-0.1 M citrate buffer (pH 2.5-8.0) for SmeChiB.

Protein Measurement
Protein concentrations were measured by the method of Bradford using bovine serum album as the standard [37].

Gel Electrophoresis
SDS-PAGE was carried out by the method of Laemmli [38] with 12.5% polyacrylamide gel (e-PAGEL, Atto, Tokyo, Japan).A sample was mixed with Ez Apply (Atto), and the mixture was heated for 5 min at 100 ˝C.The proteins in the gels were stained with Coomassie Brilliant Blue R-250.

N-Terminal Amino Acid Sequence Analysis
The N-terminal amino acid sequences were analyzed using a protein sequencer (PE Applied Biosystems 447/120A, Foster City, CA, USA).
3.8.Cloning of S. melanostictus cDNA sequences of all primers are presented in Supplementary Table S1.Total RNA was extracted from the stomach of S. melanostictus using ISOGEN II reagent (Nippon Gene, Tokyo, Japan) according to the manufacturer's instructions.First-strand cDNA was synthesized using 500 ng total RNA and oligo dT primers with PrimeScript II Reverse Transcriptase (RNase H-free) (Takara Bio, Shiga, Japan) according to the manufacturer's instructions.The reverse transcriptase-polymerase chain reaction (RT-PCR) was performed with primers Chi-a, Chi-b and Chi-c for SmeChi-1 and Chi-d, Chi-e, and Chi-f for SmeChi-2, respectively.The first PCR was carried out with S. melanostictus cDNA as a template, with Chi-a and Chi-c for SmeChi-1 and Chi-d and Chi-f for SmeChi-2 as primers.Nested PCR was performed with the products of the first PCR as the templates, and with Chi-b and Chi-c for SmeChi-1 and Chi-e and Chi-f for SmeChi-2 as primers (Table S1, Figure S1).
To obtain the full-length cDNA of SmeChi-1 and SmeChi-2, we performed both 3 1 -and 5 1 -RACE (rapid amplification of cDNA ends) using a gene-specific primer based on the sequence of the cDNA fragment obtained from the RT-PCR.cDNA fragments encoding the 3 1 region of SmeChi-1 and SmeChi-2 were amplified with S. melanostictus cDNA as the template and the primer pairs SmeChi-1-1 and 3R, and SmeChi-2-1 and 3R, respectively.

Nucleotide Sequence Analysis
We subcloned the RT-PCR, 3 1 RACE, and 5 1 RACE amplification products into pGEM-T ® Easy vector (Promega, Madison, WI) according to the respective manufacturer's instructions.We carried out A-tailing with the full-length amplification products of SmeChi-1, and these products were subcloned into pGEM-T Easy vector.The full-length amplification products of SmeChi-2 were subcloned into pCR-Blunt II-TOPO ® vector (Life Technologies, Carlsbad, CA, USA).Sequences were determined on an ABI PRISM 3130 genetic analyzer (Applied Biosystems, Foster City, CA, USA) using the Big Dye Terminator v3.1 cycle sequencing kit (Applied Biosystems, Foster City, CA, USA).

Phylogenetic Analysis of SmeChi-1 and SmeChi-2
To classify the chitinases from the stomach of S. melanostictus among vertebrate chitinases, we constructed a phylogenetic tree based on the sequences of enzyme precursors by the neighbor-joining method using the ClustalW2 program (http://www.ebi.ac.uk/Tools/msa/clustalw2/). A bacterial chitinase (GenBank: X03657) was used as an outgroup.

Prediction of the 3D Structure Models of SmeChi-1 and SmeChi-2
The 3D structure models of SmeChi-1 and SmeChi-2 were predicted using the SWISS-MODEL program (http://swissmodel.expasy.org/),based on template the acidic mammalian chitinase catalytic domain in complex with methylallosamidin (PDB ID: 3fy1) for the structure of the catalytic domain the solution structure of tachycitin, an antimicrobial protein with chitin-binding function (PDB ID: 1dqc) for the chitin-binding domain.
3.12.Tissue-Specific Gene Expressions of SmeChi-1 and SmeChi-2 Total RNA was prepared from the stomach, intestine, hepatopancreas, pyloric appendage, kidneys, testis, and ovary of S. melanostictus as described in the RNA isolation section above.First-strand cDNA was prepared from the RNA isolated from each tissue as described in the RT-PCR section above.SmeChi-1 and SmeChi-2 were amplified using the first-strand cDNA as the template and the primer pairs SmeChi-1-a and SmeChi-1-b for SmeChi-1, and SmeChi-2-a and SmeChi-2-b for SmeChi-2 (Table S1).
To determine the amount of total RNA in each tissue, β-actin mRNA fragments were amplified using specific primer pairs (Table S1).

Conclusions
The linker regions of SmeChi-1 and SmeChi-2 are the longest among the fish stomach chitinases identified thus far.In addition, the 3D structures of SmeChi-1 and SmeChi-2 revealed a deep substrate-binding cleft that is needed to degrade insoluble substrates.The present results showed that SmeChiA efficiently degraded crystalline chitin and SmeChiA was adsorbed by affinity columns when chitin was used as a carrier.Our present findings also suggest that the catalytic domain of SmeChiA can hydrolyze a wide area of chitin by using the longest linker region and the chitin binding domain.This mechanism may underlie the high chitinase activity toward crystalline chitins.

Figure 5 .
Figure 5. Phylogenetic tree analysis of chitinase amino acid sequences by the neighbor-joining method using ClustalW2.A bacterial chitinase, Serratia marcescens chitinase, was used as an outgroup.The scale bar indicates the substitution rate per residue.The black circles show SmeChi-1 and SmeChi-2 obtained in the present study.

Figure 6 .
Figure 6.3D structure prediction models of SmeChi-1 (a) and SmeChi-2 (b).The 3D structure models were predicted using the SWISS-MODEL program (http://swissmodel.expasy.org/).The structure of the catalytic domain of acidic mammalian chitinase from Homo sapiens (Protein Data Bank [PDB] ID: 3FY1) and those of chitin-binding proteins from Tachypleus tridentatus (PDB ID: 1DQC) were used as a template.

Table 2 .
Reaction patterns and cleavage patterns of (GlcNAc) 5 by SmeChiA and SmeChiB, and other fish stomach chitinases.