Gene 452 (2010) 63–71

Contents lists available at ScienceDirect

Gene

j ourna l homepage: www.e lsev ie r.com/ locate /gene

Evolutionary and biochemical differences between human and monkey acidicmammalian chitinases

Rustem Krykbaev a,⁎,1, Lori J. Fitz a,1, Padmalatha S. Reddy b, Aaron Winkler a, Dejun Xuan a, Xiaoke Yang a,Margaret Fleming a, Stanley F. Wolf a

a Pfizer Biotherapeutics Research and Development, Department of Inflammation and Immunology, 200 CambridgePark Drive, Cambridge, MA 02140, USAb Pfizer Biotherapeutics Research and Development, Department of Global Biotherapeutic Technologies, 87 CambridgePark Drive, Cambridge, MA 02140, USA

Abbreviations: MACase, Macaca fascicularis Acidic Mmethylallosamidin; chitobiose-4MU, 4-methylumbellifeside; RFU, relative fluorescence unit; qPCR, quantitativHARP, human acidic ribosomal protein; OVA, ovalbpolymorphism; HSP, high-scoring segment pair; UTR, ureading frame; GlcNAc, N-acetylglucosamine; AHR, airw⁎ Corresponding author. Fax: +1 617 6655584.

E-mail addresses: rkrykbaev@wyeth.com (R. Krykbaepreddy@wyeth.com (P.S. Reddy), awinkler@wyeth.comdxuan@wyeth.com (D. Xuan), xyang@wyeth.com (X. Ya(M. Fleming), swolf@wyeth.com (S.F. Wolf).

1 Both authors contributed equally to this work.

0378-1119/$ – see front matter © 2009 Elsevier B.V. Adoi:10.1016/j.gene.2009.12.005

a b s t r a c t

a r t i c l e i n f o

Article history:Received 25 August 2009Received in revised form 8 December 2009Accepted 16 December 2009Available online 28 December 2009

Received by J.G. Zhang

Keywords:Acidic Mammalian ChitinaseOrthologPseudoorthologPseudogeneMacaca geneEnzyme

Acidic mammalian chitinase (AMCase), an enzyme implicated in the pathology of asthma, is capable of chitincleavage at a low pH optimum. The corresponding gene (CHIA) can be found in genome databases of avariety of mammals, but the enzyme properties of only the human and mouse proteins were extensivelystudied. We wanted to compare enzymes of closely related species, such as humans and macaques. In ourattempt to study macaque AMCase, we searched for CHIA-like genes in human and macaque genomes. Wefound that both genomes contain several additional CHIA-like sequences. In humans, CHIA-L1 (hCHIA-L1) isan apparent pseudogene and has the highest homology to CHIA. To determine which of the two genes isfunctional in monkeys, we assessed their tissue expression levels. In our experiments, CHIA-L1 expressionwas not detected in human stomach tissue, while CHIA was expressed at high levels. However, in thecynomolgus macaque stomach tissue, the expression pattern of these two genes was reversed: CHIA-L1 wasexpressed at high levels and CHIA was undetectable. We hypothesized that in macaques CHIA-L1 (mCHIA-L1), and not CHIA, is a gene encoding an acidic chitinase, and cloned it, using the sequence of human CHIA-L1as a guide for the primer design. We named the new enzyme MACase (Macaca Acidic Chitinase) toemphasize its differences from AMCase. MACase shares a similar tissue expression pattern and pH optimumwith human AMCase, but is 50 times more active in our enzymatic activity assay. DNA sequence of themCHIA-L1 has higher percentage identity to the human pseudogene hCHIA-L1 (91.7%) than to hCHIA (84%).Our results suggest alternate evolutionary paths for human and monkey acidic chitinases.

ammalian Chitinase; MALLO,ryl β-D-N,N'-diacetylchitobio-e polymerase chain reaction;umin; SNP, single-nucleotidentranslated region; ORF, openay hyperresponsiveness.

v), lfitz@wyeth.com (L.J. Fitz),(A. Winkler),ng), mfleming@wyeth.com

ll rights reserved.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

Chitinases are endo-β-1,4-N-acetylglucosaminidases that hydro-lyze chitin. They have been identified in many organisms, includingthose that do not make chitin, such as mammals. There are twocatalytically active chitinases in humans, AMCase and chitotriosidase.Both belongs to the family of 18 glycosylhydrolases (van Aalten et al.,2001; Synstad et al., 2004). AMCase is most active at acidic pH (Boot etal., 2001; Chou et al., 2006) whereas chitotriosidase (Boot et al., 1995;

Renkema et al., 1995; Zheng et al., 2005) has a different pH optimumthan AMCase. Also a number of non-catalytic chitinase-like proteinshave been identified, such as CHI3L1 (Hcgp39) and CHI3L2 (YKL39),in humans, as well as CHI3L3 (Ym1), CHI3L4 (Ym2), Ym3, CHI3L1(gp39), inmice (Houston et al., 2003; Bussink et al., 2007; Funkhouserand Aronson, 2007).

The AMCase enzyme gene (CHIA) has 11 exons and is located onchromosome 1. Its translation results in a 50-kDa protein containingcatalytic and chitin-binding domains connected via a flexible linker.The catalytic domain of chitinases folds into a TIM-barrel structurewith DXXDXDXE as a consensus sequence motif at the active site(Henrissat and Bairoch, 1993; Coulson, 1994; Saito et al., 1999;Synstad et al., 2004). The chitin-binding domain is located at the C-terminus and is responsible for the anchoring of the enzyme to thechitin (Tjoelker et al., 2000). Catalytically inactive forms lackconserved glutamate at the active site and do not contain the C-terminal chitin-binding domain, although some retain an ability tobind carbohydrates (Sun et al., 2001; Fusetti et al., 2003).

The function of AMCase in mammals remains unclear. In humansand mice it is expressed at high levels in the stomach, and is alsoexpressed in the lung, where it is further induced upon antigen

64 R. Krykbaev et al. / Gene 452 (2010) 63–71

stimulation (Boot et al., 2001; Zhu et al., 2004; Follettie et al., 2006). Ina mouse model of asthma, blockade of AMCase prevents lunginflammation and airway hyperreactivity (AHR) (Zhu et al., 2004).However, in another study, mice over-expressing AMCase wereshown to have a reduced innate immune response caused by chitin(Reese et al., 2007). Recently, an isoform of human AMCase withincreased enzymatic activity has been shown to be associated withprotection from asthma in humans (Seibold et al., 2009).

So far, only human and mouse AMCase enzymes have beencharacterized. Non-human primates, being closer to humans in theirgenetic makeup than many other species, would be valuable inevaluating the involvement of AMCase in asthma. However, themonkey acidic chitinase has not as yet been studied.

In our effort to characterize the acidic mammalian chitinase inmacaques we searched for CHIA-like sequences in humans andcynomolgus monkeys (Macaca fascicularis). Our analyses of thehuman genome database showed that there are several CHIAparalogous genes in the 1p13 locus. Using the human CHIA and oneof the more closely related human CHIA paralogs, CHIA-L1, we clonedand characterized a monkey gene of the CHIA family. A study andcomparison of the gene structures and the biochemical properties ofmonkey and human acidic chitinases revealed unexpected differencesbetween the two species.

2. Materials and methods

2.1. Database search

To identify monkey orthologs and paralogs of CHIA, a TBLASTNsearch against the Rhesus macaque genome was carried out using thefull-length human AMCase protein sequence (AF290004). All relevantand significant high-scoring segment pairs (HSPs), aligning with thequery sequence from the TBLASTN search, were sorted on the basis oftheir chromosomal location, as chitinases are typically clustered in agiven chromosomal region. A sequence of contiguous HSPs, that spansmost of the query sequence, defined a hit sequence. Each hit was thensearched against a protein database for annotation purposes.

The chromosomal regions, containing the CHIA and CHIA-likesequences, were chosen for predicting human CHIA paralogs usingGeneWise program and the CHIA sequence (AF290004). Each of thesepredicted genes was searched against the EST database for supportingevidence.

2.2. qPCR methods

mRNA expression of CHIA and CHIA-L1 was quantified using real-time quantitative PCR (qPCR). Human tissue RNA's were obtainedfrom a commercial tissue panel (Clontech), while the cynomolgusmacaques tissues were harvested fresh from an animal euthanizedduring an Animal Care and Use Committee approved study at Wyeth.Total RNA was purified following the manufacturer's protocol withthe RNAeasy Midi kit (Qiagen) after homogenization in the providedRLT buffer. Primers and probes were designed with Primer Expresssoftware (Applied Biosystems) and synthesized by Eurogentec, as wasthe 2× qRT-PCR Mastermix. CHIA: 5′-CCCCGTGGCAAATAACAGAA-3′(forward), 5′-CCCGGCCTGGCAGTTC-3′ (reverse), 5′-ACTCCATTCACG-CAGTGCCAGAAGG-3′ (FAM-labeled probe ). CHIA-L1: 5′-CAGGGCCC-CAGCTGAGA-3′ (forward), 5′-GGGTTGCTCAGAAGGAAGGAG-3′(reverse), 5′-CTCTTGGTTGGATTCCCAGCCTATGGAC-3′ (FAM-labeledprobe). HARP: 5′-CGCTGCTGAACATGCTCAA-3′ (forward), 5′-TGTCG-AACACCTGCTGGATG-3′ (reverse), 5′-TCCCCCTTCTCCTTTGGGCTGG-3′(FAM-labeled probe).

The RT-PCR reaction and cycling were performed on an ABI 7700device (Applied Biosystems). A six-point standard curve was con-structed from ten-fold dilutions of 50 ng/reaction of stomach RNA(human for CHIA and monkey for CHIA-L1). 5 ng RNA samples in

duplicatewere tested for chitinase expression and normalized for RNAcontent using human acidic ribosomal protein (HARP) expression.Normalized qRT-PCR units were calculated by dividing the chitinaseRNA equivalents for each sample by the HARP RNA equivalents.

2.3. Gene cloning

2.3.1. Human AMCase cloningHuman AMCase that we used in this study was cloned as reported

previously (Chou et al., 2006). Our version of the gene was slightlydifferent from the variant (accession number AF290004) described inthe original publication (Boot et al., 2001). At the protein level itcontained N45D, D47N and R61 M substitutions which are knowngenetic variations (Seibold et al., 2009). The crystal structure of ourvariant has been solved recently (Olland et al., 2009).

2.3.2. Monkey CHIA-L1 cloningPolyA+mRNA from cynomolgus monkey stomach was obtained

from BioChain Institute, Inc., (cat. No. M1534248-CY). Total RNA waspurified from cynomolgus macaque stomach tissue at Wyeth. RT-PCRwas performed on both total and PolyA+RNAs with primersrepresenting 5′ and 3′ untranslated sequences of hCHIA-L1: 5′-TATAAATGGCAGGTTGGATGAGGG-3 ′ ( forward) and 5 ′ -CTGGGTGAGGTGATATCTGAAAAATG-3′ (reverse) and human CHIA:5′-TCAGAACATATAAAAAGCTCTGCGG-3′ (forward) and 5′-CTCTAGG-GAATATAGACCAGGTCAGGT-3′ (reverse). DNA bands of appropriatesize were excised from the agarose gel and sequenced. Aftersequencing the RT-PCR fragments, cloning primers were designedrepresenting the mCHIA-L1 ORF. 5′-GGGGACAAGTTTGTACAAAAAAG-CAGGCTTCCACCATGGCCAAGCTTACCCTCCTCACTG-3′ (forward) and5′-GGGGACCACTTTGTACAAGAAAGCTGGGTCCTAGTGATGGTGATGGT-GATGCCAGCTGCAGCAGGAGCAGGAGGCTTC-3′ (reverse) containgateway attB cloning sequences and add a coding sequence for ahis6-tag at the C-terminus of the protein. The purified PCR fragmentwas cloned into the expression vector under the control of the CMVpromoter for protein expression and purification.

2.3.3. Human CHIA-L1 cloningTo engineer a corrected full-length version of hCHIA-L1 an EST

LIFESEQ3645544 (Open Biosystems) was obtained. Site-directedmutagenesis was performed by the overlap extension method (Ho etal., 1989; Ge and Rudolph, 1997). Primers NHC-U: 5′-GCTTTCAATG-GCCTGAAAAACAAGAATAGTCAACTGAAAACTCTCTTGGCTATTGG-3′and NHC-L: 5′-CCAATAGCCAAGAGAGTTTTCAGTTGACTATTCTTGTTT-TTCAGGCCATTGAAAGC-3′ were used to delete the remaining partof the intron sequence; primers NHC-U1: 5′-CCCCAGCTGAGAAGCTC-TTGGTTGGATTCCCAGCCTATGGAC-3′ and NHC-L1: 5′-GTCCA-TAGGCTGGGAATCCAACCAAGAGCTTCTCAGCTGGGG-3′ were used toremove a one base-pair deletion in exon 8 by introducing an extra Gnucleotide as exists in hAMCase in this position and to correct theM262L mutation. Primers NHC-U2: 5′-GCGGTGTCAGCCACAGTGG-TAGCTCTGGGGGCCGCT-3′ and NHC-L2: 5′-AGCGGCCCCCAGAGCTAC-CACTGTGGCTGACACCGC-3′were used to correct the R486Gmutation.Full-length hCHIA-L1 cDNA was cloned into the expression vector bygateway recombination using primers 5′-GGGGACAAGTTTGTA-CAAAAAAGCAGGCTTCCACCATGGCCAAGCTCACCCTTCTCACTG-3′ and5′-GGGGACCACTTTGTACAAGAAAGCTGGGTCCTAGTGATGGTGATGGT-GATGCCAGCTGCAGCAGGAGCAGAAGGT-3′.

A truncated version of hCHIA-L1 was made by using the NHC-Uand NHC-L primers to delete the remaining part of the intronsequence. The single nucleotide deletion at exon 8 was not removed,thus resulting in premature termination of translation.

2.4. Enzyme expression and purification

Full-length AMCase was expressed in mammalian CHO cells andpurified as described previously (Chou et al., 2006). Full-length

Fig. 1. Chromosomal organization of human and monkey CHIA-like genes based onhuman and M. mulatta (Rhesus) genome sequences and in silico prediction of genestructure. Numbers represent distances between genes in kilobases. All underlinedgenes have chitinase active site signature sequences. Human CHIA is known to beencoding a functional enzyme, AMCase. Asterisks mark pseudogenes. Monkey genomedata became available after cloning and expression of a macaque acidic chitinase wasdone.

65R. Krykbaev et al. / Gene 452 (2010) 63–71

MACase was expressed in mammalian COS6 M cells by transienttransfection. It was purified to homogeneity from serum-freeconditioned media using Ni-NTA superflow resin (Qiagen Inc.,Valencia, CA). Conditioned medium (500 ml) was applied to 1 ml ofNi-NTA resin and allowed to bind at 4 °C overnight. Unbound proteinwas washed away with wash buffer (50 mM Tris pH 8.0, 0.3 M NaCl,20 mM imidazole). MACase was then eluted using the wash bufferwith 250 mM imidazole. The protein was dialyzed into 25 mM Tris pH8.0 and 0.4 M NaCl.

2.5. Chitobiose-4MU assays

A fluorogenic substrate, 4-methylumbelliferyl β-N,N'-diacetylchi-tobioside (chitobiose-4MU) was purchased from Sigma-Aldrich (St.Louis MO). Methylallosamidin was used as described previously(Chou et al., 2006). Chitinolytic activity was measured using achitobiose-4MU and monitored on a Saphire II fluorescence platereader with excitation at 345 nM, emission at 440 nM, and bandwidthof 5 nM (Molecular Devices, Sunnyvale, CA). Reactions were typicallymonitored for 60 min, with less than 10% substrate conversion. Assayswere performed in triplicate at 25 °C inMcIlvaine buffer at pH 5.0. Theconcentration of the fluorescent 4-MU produced was converted fromRFU using a 4MU standard curve. The initial rate of product formationdetermined from the progress curves at substrate concentrations,ranging from 1 μM to 150 μM chitobiose-4MU, were fit to a Michaelis–Menten model using nonlinear regression (GraphPad Prism version5.00 for Windows, Graph Pad Software Inc., San Diego, CaliforniaUSA).

To determine the pH profiles of the enzymes, reactions wereperformed in McIlvaine buffer at pH values of 2.0, 3.5, 5.0, 6.5 and 8.0.Activity was measured with 20 μM chitobiose-4MU and 1 nM AMCaseor 0.25 nM MACase.

The IC50 values of methylallosamidin were determined at pH 5.0by evaluating the rate of product formation at various concentrationsof the inhibitor and the chitobiose-4MU at 20 μM, and either AMCaseat 1 nM or MACase at 0.25 nM. The data were normalized to noinhibitor added. The IC50 values were derived from Eq. 1 usingGraphPad Prism version 5.00 for Windows (Graph Pad Software Inc.,San Diego, California USA).

Y¼ 100= 1þ 10^ X−LogIC50ð Þð Þ ð1Þ

2.6. Chitin RBV assay

Carboxymethyl-substituted chitin labeled covalentlywith RemazolBrilliant Violet (CM-Chitin-RBV) (Loewe Biochemica, Sauerlach,Germany) was used to measure chitinolytic activity (Wirth andWolf, 1992). Enzymatic reactions were performed in 0.2 M citratephosphate buffer, pH5.0 containing 10nMAMCase orMACase enzymeand 0.5mg/ml of CM-chitin-RBV. Thefinal reaction volumewas 200 μl.Reactions were performed in 96-well v-bottom plates (Costar, Acton,MA). The plates were sealed with low evaporation lids and incubatedat 37 °C for 24 h. The reaction was terminated by the addition of 1 NHCl, causing precipitation of the non-degraded high-polymericsubstrate. The plates were cooled on ice for 10 min and centrifuged(1400g, 10 min). The supernatants (100 μl) were transferred to a newset of 96-well plates (Costar) and the released soluble chitin oligomerswere measured at 550 nM (Molecular Devices).

3. Results

3.1. Database search for human and monkey genes

At the beginning of this study only the human genome databasewas available for analysis and we searched it for CHIA-like genes to

find potential leads for the cloning of the monkey acidic mammalianchitinase.

One of the CHIA-like sequences we identified in the humandatabase was CHIA-L1. CHIA-L1 shares a similar structure with CHIA,except for a stop codon in exon 8, resulting in a truncated hypotheticalprotein. Both genes are located at the 1p13 locus. In addition to CHIAand CHIA-L1, we have identified another partial CHIA-like sequence inthe vicinity of the CHIA gene and it is referred to as CHIA-L2(unpublished data). There is limited EST evidence in support of thisgene (data not shown). CHIA-L2 has an incomplete exon 3 sequence,and complete exons 4–11, with several stop codons in exons 7–10(data not shown). CHIA-L1 and CHIA-L2 are likely to correspond to theLOC149620 and LOC728204 pseudogenes (Bussink et al., 2007)derived by automated computational analysis.

The human CHIA-L1 sequence was used as template for the primerdesign in order to clone its cynomolgus monkey ortholog mCHIA-L1.Subsequently, the rhesus macaque (Macaca mulatta) genome data-base became available and we searched it for CHIA-like genes as well.As in the human chitinase family, three sequences with homology tothe human CHIA gene were identified in the monkey genome. Theseare referred to as mCHIA, mCHIA-L1 and mCHIA-L2. mCHIA andmCHIA-L1 share the same 11 exon gene structure as their humancounterparts. The genomic organization of the CHIA cluster forhumans and monkeys is shown in Fig. 1.

Based on the identities of the monkey and human CHIA clustergenes (Table 1), mCHIA is inferred to be the monkey ortholog ofhuman CHIA while mCHIA-L1 is the inferred monkey ortholog ofhuman CHIA-L1. However, mCHIA has a stop codon in exon 8 andencodes a hypothetical protein that is truncated. mCHIA-L1, on thecontrary, is full-length, and instead shares a higher percent identitywith the human CHIA-L1, which encodes a hypothetical truncatedprotein. Interestingly, mCHIA and hCHIA-L1 both have a stop codon inexon 8, albeit at different positions.

If spliced properly, the cDNA of hCHIA-L1 would consist of 11exons and 10 introns, just as CHIA, and would contain a sequenceDGLDFDWE, which is a characteristic chitinase active site signaturesequence DGXDXDXE (Saito et al., 1999; Synstad et al., 2004). hCHIA-L1 is 84.5% identical to CHIA at the nucleotide level. However, it has a

Table 1Identity comparison of human and monkey genes.

Gene hCHIA hCHIA-L1⁎

mCHIA⁎ 93% 81.5%mCHIA-L1 84% 91.7%

⁎ Indicates pseudogene.

66 R. Krykbaev et al. / Gene 452 (2010) 63–71

single nucleotide deletion at position 784 (exon 8) that would cause aframeshift in translation, resulting in premature termination ofprotein synthesis and therefore producing a truncated polypeptideof 272 amino acid residues. This suggests that hCHIA-L1 is apseudogene. However, as the deletion is located past the catalyticsite, the possibility remains that it could have catalytic activity.Another possibility is that this deletion comprises an SNP expressedby only some human subpopulations.

3.2. Tissue expression

To determine which CHIA family gene is expressed in humans andmonkeys, we designed qPCR primers to specifically differentiatebetween CHIA and CHIA-L1 mRNAs. Specificity was tested on RNAcollected from COS cells transiently transfected with hCHIA or hCHIA-L1 expression vectors (data not shown). CHIA mRNA was most highlyexpressed in human stomach, with lower, but detectable levels in thelung. Interestingly, similar analyses of the monkey tissue RNAsrevealed some CHIA expression in the spleen and bronchioles, butvery little expression in the stomach (Fig. 2). A distinct expressionpattern was observed when primers to CHIA-L1 were used: lowrelative expression in human lung and no expression in the humanstomach, but with a significant expression level in monkey stomach(Fig. 2). Thus, an unexpected, reciprocal pattern of CHIA and CHIA-L1

Fig. 2. Tissue expression of CHIA-like genes in humans and monkeys. qPCR primerswere designed to detect (A) exon 11 of human CHIA or (B) exon 9 of human CHIA-L1 invarious human or monkey tissues, as indicated. Data shown are TaqMan fluorescenceunits normalized to the HARP signal.

expression was revealed in two species: CHIA was highly expressed inhumans and not in monkeys, while CHIA-L1 was highly expressed inmonkeys and not in humans.

3.3. Cloning of human and monkey genes

An in silico search of the human genome database coupled withour gene prediction strategy produced a chitinase-like gene (CHIA-L1)with an 84.5% identity level to CHIA. A single base-pair deletion inexon 8 indicated that it was most likely a pseudogene or, if active, atruncated protein. To answer the question of whether or not it wasindeed a pseudogene, we attempted to clone it using RNA from thehuman lung and stomach. However, we were not able to obtain a RT-PCR fragment suitable for cloning (data not shown). Commerciallyavailable ESTs were another potential source of CHIA-L1 cDNA. Mostof thesewere incomplete and somewere incorrectly spliced since thatthey contained fragments from genomic intron sequences. Toconstruct a functional gene we purchased an EST that contained oneextra partial intron sequence and had a one base-pair deletion in exon8 (LIFESEQ3645544, Open Biosystems). We thenmodified it to encodea full-length protein, with and without the premature termination oftranslation, by removing the remaining intron sequence and insertinga G nucleotide to restore the reading frame in exon 8 caused by asingle nucleotide deletion. Despite these modifications, we were notable to detect any chitinase activity when variants were transientlyexpressed in COS cells (data not shown). In addition, we sequencedthe exon 8 genomic DNA of five donors. All had the same singlenucleotide deletion as in the genome database (data not shown).Altogether these experiments strongly suggest that hCHIA-L1 is apseudogene.

qPCR analysis revealed the reciprocal tissue expression pattern ofCHIA-L1 and CHIA genes in humans, and monkeys (Fig. 2). Wehypothesized that the human pseudogene CHIA-L1 could be an activegene in monkeys, whereas the monkey ortholog of human CHIA is apseudogene. To test this hypothesis we attempted to clone bothcounterparts of human CHIA-L1 and CHIA in cynomolgus macaques.

The macaque genome sequence database was not available at thattime. Instead, we used human sequences as a guide to primer design.Alignment of the hCHIA-L1 and hCHIA genes revealed that there wereenough differences in the 5′-UTR and the 3′-UTR sequences to designprimers specific to each gene. The cloning strategy involved usingprimers homologous to UTRs to perform RT-PCR reactions on stomachRNA, and then sequencing the resulting PCR fragment to deduce theentire ORF. Subsequently, primers were designed to clone the ORF toan expression vector for protein expression and purification.

Attempts to clone the monkey ortholog of human CHIA wereunsuccessful: the overall RT-PCR product yield was very low andsequencing of individual clones revealed a variety of incompletelyspliced gene variants (data not shown). However, using hCHIA-L1primers in a RT-PCR reaction resulted in a distinct DNA band amplifiedfrom monkey stomach RNA. Sequencing revealed an ORF 1446nucleotides in length encoding a protein of 474 amino acids, whichis slightly shorter than AMCase (476 amino acids) (Fig. 3). Thenucleotide sequence reported in this paper has been submitted toGenBank™ with accession number FJ685619. The calculated molec-ular weight for MACase was 52.2 kDa. The hydrophobic leadersequence was followed by a putative mature protein, starting attyrosine 22, the same as in AMCase (Boot et al., 2001). MACase had84% identity to hAMCase at the nucleotide level and 82.1% identity atthe protein level (Fig. 4). The total number of cysteines was twelveand their locations were identical for AMCase and MACase. Six ofthem were located in a putative chitin-binding domain at the C-terminus and are essential for the enzyme's capacity to bind chitin(Tjoelker et al., 2000). In addition, MACase contained a DGLDFDWEsequence in its catalytic site, which represents a consensusDXXDXDXE sequence for a catalytically active chitinase. An identical

Fig. 3. The cynomolgus monkey CHIA-L1 sequence. Corresponding protein sequence is shown above the nucleotide sequence and is represented by a single letter code. The putativesignal peptide sequence is underlined.

67R. Krykbaev et al. / Gene 452 (2010) 63–71

sequence is present in AMCase. Thus, the MACase sequence had anumber of features characteristic of a functional chitinase (Fig. 4).

3.4. Enzyme assays and pH optima comparison of AMCase and MACase

The two previously described mammalian chitinases, AMCase andchitotriosidase, differ in their pH optima. To better understand thesimilarities and differences between AMCase and MACase, thechitinolytic activity was measured in a fluorogenic assay using 4-methylumbelliferyl (4-MU) chitobioside as a substrate. The assaymonitored the generation of the fluorescent product, 4-MU, from thecleavage of the β-glycosidic linkage between 4MU and GlcNAc at pHvalues between 2 and 8. For both human AMCase and MACase theoptimal pH for activity was the same, pH 5.0 (Fig. 5), which agreeswith previous results (Chou et al., 2006). The enzymatic activity ofAMCasewas about 30% at pH 2, pH 3.5 and pH 6.5 versus 100% activityat pH 5.0, and no activity was observed at pH 8.0. The enzymaticactivity of MACase was 20% at pH 2, pH 3.5 and pH 6.5 versus 100%activity at pH 5.0. It also had 30% activity at pH 8.0.

3.5. Kinetic parameter comparison between AMCase and MACase

Initial experiments suggested that MACase was more active thanAMCase. To better understand the mechanism for the increasedspecific activity, we measured the Km and kcat for each enzyme. Theprogress of AMCase (1 nM) hydrolysis reactions for 4MU productformation was linear during the first 10 min at chitobioside-4MUsubstrate concentrations between 0 and 75 μM (data not shown). Therate of product formation was slower at the highest substrateconcentration tested, 150 μM, compared to 75 μM, reflecting apparentsubstrate inhibition due to transglycosylation, which is consistentwith a previous report (Chou et al., 2006). We confirmed that the highsubstrate concentration inhibition of the enzymatic reaction was notdue to the inner filter effect. In separate control experiments, the 4MUfluorescent signal was not quenched by chitobiose-4MU up to 600 μM.

For MACase (0.25 nM), the hydrolysis reactions for 4MU productformation were linear during the first 5 min at chitobioside-4MUsubstrate concentrations ranging from 0 to 37 μM. MACase exhibitsapparent substrate inhibition, indicating that transglycosylation may

Fig. 4.mCHIA-L1 is most homologous to the hCHIA-L1 pseudogene. The amino acid sequences for human CHIA, human CHIA-L1 pseudogene, and macaque CHIA-L1 are aligned. Thesignal peptide sequence is underlined with a single line. The conserved motif for the family 18 chitinases catalytic site is underlined with a dashed line. The chitin-binding domain isunderlined with a dotted line. Asterisks indicate cysteines in the chitin-binding domain. The sequence shown for hCHIA-L1 was engineered from an available EST to obtain a correctlyspliced sequence and eliminate a single base-pair frame-shifting deletion in exon 8.

68 R. Krykbaev et al. / Gene 452 (2010) 63–71

also occur for this enzyme. A fit of the linear initial rates to aMichaelis–Menten model gave Km values of 7+/−1 μM for MACaseand 32+/−1 μM for AMCase. MACase had a higher kcat (31.8+/−3.2 s−1) than AMCase (3.0+/−0.35 s−1), resulting in a highercatalytic efficiency for MACase compared to AMCase (kcat/Km ratio of4.5 s−1/μM−1 vs. 0.09 s−1/μM−1).

3.6. Inhibition by methylallosamidin

Methylallosamidin (MALLO) is a methylated form of allosamidin, apotent inhibitor of family 18 chitinases (Koga et al., 1999). Increasingconcentrations of MALLO were preincubated with 1 nM AMCase or0.25 nMMACase and the initial rate of chitinolytic activity of the 4-MU

Fig. 5. pH profile of human AMCase and MACase. The activity of AMCase (■) andMACase (□) was monitored by a fluorogenic assay after incubation with 5 μMchitobiose-4MU substrate in citrate phosphate buffer containing 0.005% Brij-35,prepared at various pH values, as indicated. Data shown are the relative activity levelsfor various pH values compared to the maximal activity at pH 5.0.

Fig. 7. AMCase and MACase hydrolyze chitin and can be inhibited by MALLO. (A) ChitinRBV was incubated with increasing concentrations of AMCase (●) or MACase (○) for24 h at 37 °C. The maximal soluble oligosaccharide release from 100 nM MACase wasplotted as 100% maximum activity. (B) Chitin RBV was incubated with 25 nM AMCase(■) or 5 nM MACase (□), each with a range of concentrations of methylallosamidin(MALLO), as indicated, for 24 h at 37 °C. The enzyme plus MALLO values werenormalized to enzyme with no inhibitor values for relative activity.

69R. Krykbaev et al. / Gene 452 (2010) 63–71

chitobioside for each concentration was measured. The data gener-ated was fit to a sigmoid curve using Eq. (1) from Materials andmethods. Methylallosamidin is a more potent inhibitor of AMCasewith IC50 of 5.7 nM, compared toMACase, with IC50 of 54 nM (Fig. 6).

3.7. Chitin hydrolysis

We employed the chitin RBV assay, a soluble, dye-labelled andacid-precipitable assay for endo-chitinases to determine if MACaseand AMCase hydrolyze chitin (Wirth and Wolf, 1992). Both AMCaseand MACase generated soluble chitin oligomers after incubation withthe Chitin-RBV at 37 °C for 24 h (Fig. 7A). The amount of soluble chitinoligomer generated was dependent on the concentration of enzyme.MALLO blocked the generation of cleaved oligosaccharides (Fig. 7B)indicating that this activity was dependent on the catalytic activity ofthese enzymes.

4. Discussion

There is no clear understanding at present of the physiological andpathological roles of chitinases in mammals. AMCase homologs inlower life forms have been shown to be a part of the response againstparasitic infections (Herrera-Estrella and Chet, 1999; Palli andRetnakaran, 1999; Oshima et al., 2002). However, there are contra-dictory opinions on the exact role of AMCase in asthma. AMCase geneexpression is increased during allergen challenge in ova-sensitizedmice (Zhu et al., 2004; Follettie et al., 2006). These observations led to

Fig. 6. Inhibition of AMCase and MACase by methylallosamidin. Various concentrationsof methylallosamidin, as indicated, were incubated with 1 nM AMCase (■) or 0.25 nMMACase (○) and the chitobiose-4MU substrate. Data shown are the percent enzymeactivity values compared to the maximum activity with no inhibitor.

the hypothesis that AMCase contributes to the pathogenesis of Th2immune responses (Zhu et al., 2004). If this is the case, it would bebeneficial to inhibit AMCase for asthma patients. According to anotherstudy, however, AMCase actually plays a protective role by degradingchitin, and thus attenuating the immune response induced by chitin inmouse lungs (Reese et al., 2007). This lack of clear understanding ofAMCase's function underscores the importance of further studies toelucidate the role of AMCase in lung inflammation.

So far mechanistic data has been obtained using only mousemodels, and the question remains as to how relevant these are tohuman biology. Monkeys are genetically closer to humans and arefrequently used in the drug discovery process to confirm drug safetyand efficiency before clinical trials (Coffman and Hessel, 2005).However, to validate particular animal usage as a model, the druginteraction with the target on the molecular level must be confirmed.To obtain a monkey enzyme for this purpose, the correct genesequence must be found first.

Based on the tissue expression data, we hypothesized that themCHIA-L1 gene, and not the mCHIA, could be a functional gene inmonkeys. Indeed, using a hCHIA-L1 sequence as a guide, we were ableto clone and characterize a fully active monkey acidic chitinase. Wecalled it MACase to emphasize its difference from AMCase. In ourexperimentsMACase shares a similar expression profile with AMCase,with the highest expression in the stomach. Just as in the case ofAMCase, most enzymatic activity was observed at an acidic pH rangewith optimum at pH 5. The calculated isoelectric point is even moreacidic for MACase, 4.97, then for AMCase, 5.42. Despite the similaritiesin expression and pH profile, we found MACase to be quite differentfrom human AMCase in terms of both gene sequence and enzymatic

70 R. Krykbaev et al. / Gene 452 (2010) 63–71

properties. mCHIA-L1 has more percentage identity to the hCHIA-L1pseudogene (91.7%) than to hCHIA (84%). At the protein level MACaseis 87.6% identical to a version of the hCHIA-L1 protein that is modifiedto remove a frameshifting deletion, and only 82.1% identical toAMCase. Moreover, MACase is much more catalytically efficient thanAMCase in our assay, using chitobiose-4MU as a substrate. Thecatalytic efficiency is about 50 times higher for MACase compared toAMCase at the optimal pH of 5.0. This finding fits into our originalhypothesis of the existence of two reciprocal pairs of genes: inhumans, CHIA is an active gene and CHIA-L1 is a pseudogene, while inmonkeys CHIA-L1 is a functional gene and CHIA is a pseudogene.

Once the Rhesus macaque genome data became available, weconfirmed our findings. Both cynomolgus (M. fascicularis) and rhesus(M. mulatta) macaques belong to the same Macaca genus of OldWorldMonkeys, indicating that they are genetically very close to eachother. Indeed, our in silico search revealed a pair of genes atchromosome 1: mCHIA, which had the highest percentage identityto human CHIA, and another one, mCHIA-L1, which had the highestpercentage identity to the hCHIA-L1 pseudogene. Rhesus CHIA had aone nucleotide deletion in exon 8, which was indicative of its being apseudogene, and rhesus CHIA-L1 was identical to the cynomolgusCHIA-L1 that we discovered. There is a correct automated geneprediction of a rhesus functional gene that can be found in Genebank(XM_001104487) (Bussink et al., 2007). However, Ensembl genomicdatabase contains an incorrect gene prediction representing mCHIA(Transcript ID ENSMMUT00000012389).

One possible explanation of our findings could be that aduplication of the CHIA ancestor gene probably occurred duringevolution some time before humans and macaques diverged about25 million years ago (Gibbs et al., 2007). In this case, human CHIA andCHIA-L1 are paralogs to each other in evolutionary terms, as are themonkey genes mCHIA and mCHIA-L1. After that speciation eventoccurred, only one copy of the duplicated gene, different in humansand monkeys, survived subsequent selection (Fig. 8). In that respect,human CHIA and monkey CHIA-L1 are actually outparalogs orpseudoorthologs, because each one represents a surviving copy of aparalogous gene (Koonin, 2005).

Whether the evolutionary differences correlate with functionaldifferences between the two genes remains to be seen. So far, we haveobserved a substantial increase in catalytic efficiency of MACasecompared to AMCase. In our study, we compared MACase to a variantof human AMCase that is catalytically more active than the originallydiscovered version (AF290004) and is shown to be associated withprotection from asthma (Seibold et al., 2009). It could be possible thatmacaques needed an even more active enzyme to efficiently remove

Fig. 8. Evolutionary relationships between CHIA and CHIA-L1 genes in primate species.(A) Proposed diagram of human and macaque gene evolution. Asterisk indicatespseudogene. CHIA-L1 gene indicated by shaded box.

chitin, and therefore prevent an allergic response (Reese et al., 2007).Since both enzymes are expressed most abundantly in the stomach,diet, which may include a chitin-rich food such as insects andcrustaceans, may be an important selective factor. Therefore, we canspeculate that in the course of evolution, the macaques' eating habitsmay have required the development of amore active enzyme. The roleof AMCase in T-helper 2 immunity responses, such as a responseagainst parasites, has not been studied, but may be an importantfactor as well. In case of allergic inflammation, further studies arerequired to better understand what MACase's role is compared toAMCase's. Our findings may also have implications in the usage ofmonkeys to model human diseases. Macaques and humans are veryclose genetically, yet they have enzymes with subtle variations insequence, which result in significant changes in their activity levels. Inlight of those differences in enzymatic activity, adjustments may berequired in strategies to evaluate drug effects.

References

Boot, R.G., Renkema, G.H., Strijland, A., van Zonneveld, A.J., Aerts, J.M., 1995. Cloningof a cDNA encoding chitotriosidase, a human chitinase produced by macrophages.J. Biol. Chem. 270, 26252–26256.

Boot, R.G., et al., 2001. Identification of a novel acidic mammalian chitinase distinctfrom chitotriosidase. J. Biol. Chem. 276, 6770–6778.

Bussink, A.P., Speijer, D., Aerts, J.M., Boot, R.G., 2007. Evolution of mammalian chitinase(-like) members of family 18 glycosyl hydrolases. Genetics 177, 959–970.

Chou, Y.T., et al., 2006. Kinetic characterization of recombinant human acidicmammalian chitinase. Biochemistry 45, 4444–4454.

Coffman, R.L., Hessel, E.M., 2005. Nonhuman primate models of asthma. J. Exp. Med.201, 1875–1879.

Coulson, A.F., 1994. A proposed structure for 'family 18′ chitinases. A possible functionfor narbonin. FEBS Lett. 354, 41–44.

Follettie, M.T., et al., 2006. Gene expression analysis in a murine model of allergicasthma reveals overlapping disease and therapy dependent pathways in the lung.Pharmacogenomics J 6, 141–152.

Funkhouser, J.D., Aronson Jr., N.N., 2007. Chitinase family GH18: evolutionary insightsfrom the genomic history of a diverse protein family. BMC Evol. Biol. 7, 96.

Fusetti, F., Pijning, T., Kalk, K.H., Bos, E., Dijkstra, B.W., 2003. Crystal structure andcarbohydrate-binding properties of the human cartilage glycoprotein-39. J. Biol.Chem. 278, 37753–37760.

Ge, L., Rudolph, P., 1997. Simultaneous introduction of multiple mutations usingoverlap extension PCR. BioTechniques 22, 28–30.

Gibbs, R.A., et al., 2007. Evolutionary and biomedical insights from the rhesus macaquegenome. Science 316, 222–234.

Henrissat, B., Bairoch, A., 1993. New families in the classification of glycosyl hydrolasesbased on amino acid sequence similarities. Biochem. J. 293 (Pt 3), 781–788.

Herrera-Estrella, A., Chet, I., 1999. Chitinases in biological control. EXS 87, 171–184.Ho, S.N., Hunt, H.D., Horton, R.M., Pullen, J.K., Pease, L.R., 1989. Site-directed

mutagenesis by overlap extension using the polymerase chain reaction. Gene 77,51–59.

Houston, D.R., Recklies, A.D., Krupa, J.C., van Aalten, D.M., 2003. Structure and ligand-induced conformational change of the 39-kDa glycoprotein from human articularchondrocytes. J. Biol. Chem. 278, 30206–30212.

Koga, D., Mitsutomi, M., Kono, M., Matsumiya, M., 1999. Biochemistry of chitinases. EXS87, 111–123.

Koonin, E.V., 2005. Orthologs, paralogs, and evolutionary genomics. Annu. Rev. Genet.39, 309–338.

Olland, A.M., et al., 2009. Triad of polar residues implicated in pH specificity of acidicmammalian chitinase. Protein Sci. 18, 569–578.

Oshima, H., Miyazaki, R., Ohe, Y., Hayashi, H., Kawamura, K., Kikuyama, S., 2002.Molecular cloning of a putative gastric chitinase in the toad Bufo japonicus. Zool. Sci.19, 293–297.

Palli, S.R., Retnakaran, A., 1999. Molecular and biochemical aspects of chitin synthesisinhibition. EXS 87, 85–98.

Reese, T.A., et al., 2007. Chitin induces accumulation in tissue of innate immune cellsassociated with allergy. Nature 447, 92–96.

Renkema, G.H., Boot, R.G., Muijsers, A.O., Donker-Koopman, W.E., Aerts, J.M., 1995.Purification and characterization of human chitotriosidase, a novel member of thechitinase family of proteins. J. Biol. Chem. 270, 2198–2202.

Saito, A., Ozaki, K., Fujiwara, T., Nakamura, Y., Tanigami, A., 1999. Isolation andmappingof a human lung-specific gene, TSA1902, encoding a novel chitinase familymember. Gene 239, 325–331.

Seibold, M.A., et al., 2009. Differential enzymatic activity of common haplotypicversions of the human acidic mammalian chitinase protein. J. Biol. Chem. 284,19650–19658.

Sun, Y.J., Chang, N.C., Hung, S.I., Chang, A.C., Chou, C.C., Hsiao, C.D., 2001. The crystalstructure of a novel mammalian lectin, Ym1, suggests a saccharide binding site.J. Biol. Chem. 276, 17507–17514.

Synstad, B., Gaseidnes, S., van Aalten, D.M.F., Vriend, G., Nielsen, J.E., Eijsink, V.G.H.,2004. Mutational and computational analysis of the role of conserved residues inthe active site of a family 18 chitinase. Eur. J. Biochem. 271, 253–262.

71R. Krykbaev et al. / Gene 452 (2010) 63–71

Tjoelker, L.W., et al., 2000. Structural and functional definition of the human chitinasechitin-binding domain. J. Biol. Chem. 275, 514–520.

van Aalten, D.M.F., Komander, D., Synstad, B., Gaseidnes, S., Peter, M.G., Eijsink, V.G.H.,2001. Structural insights into the catalytic mechanism of a family 18 exo-chitinase.Proc. Natl. Acad. Sci. 98, 8979–8984.

Wirth, S.J., Wolf, G.A., 1992. Micro-plate colourimetric assay for Endo-acting cellulase,

xylanase, chitinase, 1,3-[beta]-glucanase and amylase extracted from forest soilhorizons. Soil Biol. Biochem. 24, 511–519.

Zheng, T., et al., 2005. Molecular cloning and functional characterization of mousechitotriosidase. Gene 357, 37–46.

Zhu, Z., et al., 2004. Acidic mammalian chitinase in asthmatic Th2 inflammation andIL-13 pathway activation. Science 304, 1678–1682.