lable at ScienceDirect

Insect Biochemistry and Molecular Biology 58 (2015) 46e54

Contents lists avai

Insect Biochemistry and Molecular Biology

journal homepage: www.elsevier .com/locate/ ibmb

Two chitinase 5 genes from Locusta migratoria: Molecularcharacteristics and functional differentiation

Daqi Li a, Jianqin Zhang a, Yan Wang a, Xiaojian Liu a, Enbo Ma a, Yi Sun a, b, Sheng Li c,Kun Yan Zhu d, Jianzhen Zhang a, *

a Research Institute of Applied Biology, Shanxi University, Taiyuan, Shanxi 030006, Chinab Biotechnology Research Center, Shanxi Academy of Agricultural Sciences, Taiyuan, Shanxi 030031, Chinac Key Laboratory of Insect Developmental and Evolutionary Biology, Institute of Plant Physiology and Ecology, Shanghai Institutes for Biological Sciences,Chinese Academy of Sciences, Shanghai 200032, Chinad Department of Entomology, 123 Waters Hall, Kansas State University, Manhattan, KS 66506, USA

a r t i c l e i n f o

Article history:Received 7 October 2014Received in revised form7 January 2015Accepted 8 January 2015Available online 24 January 2015

Keywords:Chitinase 5Gene duplication20E regulationLocust migratoriaRNA interference

* Corresponding author.E-mail address: zjz@sxu.edu.cn (J. Zhang).

http://dx.doi.org/10.1016/j.ibmb.2015.01.0040965-1748/© 2015 Elsevier Ltd. All rights reserved.

a b s t r a c t

The duplication of chitinase 5 (Cht5) into two to five different genes has been reported only in mosquitospecies to date. Here, we report the duplication of Cht5 genes (LmCht5-1 and LmCht5-2) in the migratorylocust (Locusta migratoria). Both LmCht5-1 (505 aa) and LmCht5-2 (492 aa) possess a signal peptide and acatalytic domain with four conserved motifs, but only LmCht5-1 contains a chitin-binding domain.Structural and phylogenetic analyses suggest that LmCht5-1 is orthologous to other insect Cht5 genes,whereas LmCht5-2 might be newly duplicated. Both LmCht5 genes were expressed in all tested tissueswith LmCht5-1 highly expressed in hindgut and LmCht5-2 highly expressed in integument, foregut,hindgut and fat bodies. From the fourth-instar nymphs to the adults, LmCht5-1 and LmCht5-2 showedsimilar developmental expression patterns with transcript peaks prior to each nymphal molting, sug-gesting that their expression levels are similarly regulated. Treatment with 20-hydroxyecdysone (20E;the most active molting hormone) and reducing expression of EcR (ecdysone receptor gene) by RNAiincreased and decreased expression of both LmCht5 genes, respectively, indicating that both genes areresponsive to 20E. Although transcript level of LmCht5-2 is generally 10-fold higher than that of LmCht5-1, RNAi-mediated suppression of LmCht5-1 transcript led to severe molting defects and lethality, but sucheffects were not seen with RNAi of LmCht5-2, suggesting that the newly duplicated LmCht5-2 is notessential for development and survivorship of the locust.

© 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Chitin is a b- (1,4)-linked polymer of N-aceytylglucosaminemoieties, and found in a vast variety of taxonomic groups includingalgae, fungi, protists, sponges, rotifers, nematodes, arthropods,cuttlefish, brachiopods, and mollusks (Merzendorfer, 2013). In in-sects, chitin is extensively distributed in ectodermal epithelial tis-sues, including cuticles, trachea, foregut and hindgut; and alsoserves as an important constituent of intestinal peritrophicmatrices (PMs) (Moussian, 2010). During the development, insectsmust undergo periodically molting to allow for continued growthwith increased body size, wherein the old cuticle is degraded and

replaced by new synthesized one (Kramer et al., 1993). Digestion ofthe old cuticle is necessary prior to ecdysis and is mainly accom-plished by several hydrolytic enzymes called chitinase, proteinaseand lipase.

Insect chitinases belong to family 18 of glycoside hydrolase(GH18) and hydrolyze chitin by an endo-type of cleavage that retainthe anomeric b-(1,4) configuration of products (Kramer andMuthukrishnan, 2005). They are highly diverse enzymes encodedby many different genes as implicated in all sequenced insecttranscriptoms or genomes (Zhu et al., 2004, 2008a; Nakabachi et al.,2010; Pan et al., 2012). In Drosophila melanogaster, Tribolium cas-taneum and Anopheles gambiae, 16, 22 and 20 chitinase andchitinase-like genes have been reported, respectively (Zhang et al.,2011a). These genes differ significantly in size, and developmentaland tissue expression patterns, and their deduced proteins alsodiffer in their primary structures and domain architectures

D. Li et al. / Insect Biochemistry and Molecular Biology 58 (2015) 46e54 47

(Arakane and Muthukrishnan, 2009; Zhang et al., 2011a). Insectchitinases appear to play roles in cuticle turnover, regulatingabdominal contraction and wing expansion, digestion, immunityand natural defense (Zhu et al., 2008b; Arakane andMuthukrishnan, 2009).

Currently, insect chitinases and chitinase-like proteins areclassified into eight groups based on phylogenetic analyses of theircatalytic domains. Among the eight groups, Group I has relativelybeen well characterized. The first cDNA of MsCht5 belonging toGroup I was isolated from Manduca sexta in 1993 (Kramer et al.,1993). Since then, the orthologous cDNAs of MsCht5 have beenidentified or sequenced from at least 15 different insect speciesincluding Bombyx mori, Hyphantria cunea (Kim et al., 1998), Chiro-nomus tentans (Feix et al., 2000), Spodoptera litura (Shinoda et al.,2001), Choristoneura fumiferana (Zheng et al., 2002), Helicoverpaarmigera (Ahmad et al., 2003), D.melanogaster (Zhu et al., 2004), An.gambiae, Aedes aegypti, Culex quinquefasciatus, T. castaneum (Zhuet al., 2008a; Zhang et al., 2011a, 2011b), Lacanobia oleracea(Fitches et al., 2004), Spodoptera exigua (Zhang et al., 2012),Mamestra brassicae (Paek et al., 2012) and Ostrinia furnacalis (Wuet al., 2013).

The transcripts of Cht5 are mainly detected in the epidermis andthe guts (Kramer et al., 1993; Ahmad et al., 2003), which suggest theCht5 may be involved in chitin turnover associated tissues such ascuticular exoskeleton and peritrophic membrane. The supposedfunction of Cht5 was confirmed in T. castaneum and S. exigua byusing RNAi technology (Zhu et al., 2008b; Zhang et al., 2012).Heterologous expression of insect Cht5 has been successfully per-formed in the Hi5 and Sf9 cell lines and the yeast Pichia pastoris, andthe recombinant protein showed high levels of chitinolytic activity(Gopalakrishnan et al., 1995; Shinoda et al., 2001; Zhu et al., 2008c;Wu et al., 2013). The crystal structure of OfCht5 from O. furnacalishas been determined (Chen et al., 2014a), and a series of fullydeacetylated chitooligosaccharides (GlcN)2e7 has been demon-strated as inhibitors of OfCht5 (Chen et al., 2014b).

Only single Cht5 gene had been reported in various insect spe-cies until 2010 with five different Cht5 genes, presumably evolvedby gene duplication events, were identified in An. gambiae (Zhanget al., 2011b). Multiple Cht5 genes were also found in other twomosquito species including Ae. aegypti with four Cht5 genes and C.quinquefasciatus with three Cht5 genes (Zhang et al., 2011b).Although these duplicated Cht5s in An. gambiae were distinct ingenome structures, domain architectures and expression profiles(Zhang et al., 2011b), their roles in chitin metabolism remainsunknown.

In this paper, we reported two Cht5 genes possibly originatedfor a duplication event in Locusta migratoria, a serious agriculturalpest in many regions of the world. Specifically, we sequenced theircDNAs, profiled their developmental and tissue expression pat-terns, and examined their transcription responses to 20-hydroecdysone (20E) treatment and RNAi-mediated suppressionof EcR (the 20E receptor gene). We further examined their biolog-ical functions by using RNAi. Our results suggested that LmCht5-1was essential for molting, whereas LmCht5-2 was not.

2. Materials and methods

2.1. Insect

The migratory locusts (L. migratoria) were maintained in thelaboratory at the Research Institute of Applied Biology, ShanxiUniversity, Taiyuan. They were reared with fresh wheat seedlingsand wheat bran at 30 ± 2 �C and 40% humidity with 14:10 h light:dark cycle. The nymphs were reared in cages (25 � 25 � 25 cm)with density of approximately 150e200 insects per cage. Newly

molted individuals were synchronized and transferred into glassbeakers covered with net screen for dsRNA injection.

2.2. Search of chitinase 5 genes in locust transcriptome and genomedatabases

The candidate cDNA sequences of L. migratoria Cht5 were firstsearched against the locust transcriptome database, which wasgenerated from the mixed mRNA samples prepared from eggs,nymphs and adults. BLASTx was performed for Cht5 homologysearch. We identified two distinct cDNA fragments putativelyencoding two different LmCht 5 proteins based on an E-value cut-off of 8e�121. To confirm that the locust possessed only two LmCht5genes, we extensively searched the L.migratoria's genome database(Wang et al., 2014).

2.3. Sequencing of two chitinase 5 cDNAs

Primers were designed with the software of Primer Premier 5,and used for amplification of 5'-end, 3'-end or full-length cDNAsequences. The sequences of primers and the expected sizes of PCRproducts were shown in Table S1. Total RNA was isolated fromintegument of the sixth day of the 5th-instar nymphs using TrizolPlus reagent (TaKaRa, Dalian, China). The mRNA was isolated usingPolyATtract mRNA isolation systems (Promega, Madison, WI, USA).The first-strand cDNA was synthesized from 1 mg mRNA using theSmart RACE cDNA Amplification kit (Clontech, Mountain View, CA,USA). All PCR amplifications were performed using Advantage® 2Polymerase (Clontech). For RACE-PCR, the program was used asfollows: initial 94 �C for 3 min; 5 cycles of 94 �C for 30 s, 70 �C for30 s and 72 �C for 1 min; 35 cycles of 94 �C for 30 s, 68 �C for 30 sand 72 �C for 1 min; and 72 �C for 10 min for final extension. Foramplification of full-length cDNA sequences, the programwas usedas follows: 94 �C for 3 min; 40 cycles of 94 �C for 30 s, 62 �C for 30 sand 72 �C for 2 min; and 72 �C for 10 min. The PCR products werepurified using E.Z.N.A® Gel Extraction kit (Omega Bio-Tek, Norcross,GA, USA), subcloned into pEASY-Blunt Zero plasmid (TransGen,Beijing, China), and sequenced from both directions by Life Tech-nologies Company (Beijing, China).

2.4. Deduced amino acid sequence analysis of two chitinase 5 genes

The amino acid sequences were deduced using Translate tool onthe ExPASy Proteomics website (http://web.expasy.org/translate/).The molecular mass (MM) and isoelectric point (pI) for each Cht5were predicted by using Compute pI/Mw tool (http://web.expasy.org/compute_pi/). SMART domain analysis (http://smart.embl-heidelberg.de/) and SignalP 4.1 Server (http://www.cbs.dtu.dk/services/SignalP/) were used to predict the domain architectureand signal peptide.

The known insect Cht5 genes deposited in GenBank were usedfor constructing phylogenetic trees. These insects include hyme-nopteran (Apis mellifera), dipteran (Ae. aegypti, An. gambiae, C.quinquefasciatus and D. melanogaster), lepidopteran (Bombyxmandarina, B. mori, C. fumiferana, H. armigera, H. cunea, L. oleracea,M. sexta, O. furnacalis, S. exigua, Spodoptera frugiperda and S. litura),coleopteran (T. castaneum), and orthopteran (L. migratoria) species.Their catalytic domains were aligned using ClustalW software(http://www.ebi.ac.uk/clustalw/), and used for constructing treesby the neighbor-joining algorithm using Mega 6.06 software.

To compare the amino acid sequences and catalytic domainsof two LmCht5 genes along with those from five different insectorders, Genedoc software (http://www.psc.edu/biomen/genedoc)was used for multiple sequence alignments. Four conservedmotifs were identified based on the references previously

D. Li et al. / Insect Biochemistry and Molecular Biology 58 (2015) 46e5448

described by Arakane and Muthukrishnan (2009) and Zhang et al.(2011a).

2.5. The tissue and development-dependent expression of twoLmCht5 genes

For tissue-dependent expression profiling of two LmCht5 genes,seven different tissues, including integument, foregut, midgut,gastric caeca, hindgut, Malpighian tubules and fat bodies, weredissected from 6-day-old fifth-instar nymphs (N5D6). For devel-opmental expression profiling, samples were collected from theabdominal integument of insects at 21 different developmentalstages ranging from 1-day-old fourth-instar nymph (N4D1) to 7-day-old adult (AD7). Each sample, containing at least four locusts,was used to isolate total RNA, and the collection of each tissue typeor developmental stage was repeated three times. The first-strandcDNA was synthesized using RevertAid™ H Minnus M-Mulvreverse transcriptase (Fermentas, Glen Burnie, MD, USA).

The expression profiles of two LmCht5 genes were determinedusing reverse transcription quantitative PCR (RT-qPCR) and b-actinas a reference gene with ABI 7300 Real-Time PCR detection system(Applied Biosystems Inc, Foster City, CA, USA). The primers and theexpected size of each PCR product are summarized in Table S1. Each20-mL PCR mixture consisted of 10 mL SYBR® Green Real-time PCRMaster Mix (TOYOBO, Osaka, Japan), 2 mL of 20-fold diluted tem-plate cDNA, 0.8 mL of 0.4 mM of each primer and 6.4 mL of deionizedwater. The PCR program included 94 �C for 60 s, followed by 40cycles of 94 �C for 5 s and 60 �C for 31 s. At the end of the PCR,amplification specificity was verified by obtaining the dissociationcurve, in which the samples were cooled to 60 �C after denaturingand then the melting curves were obtained by auto increment 1 �C/15 s for each cycle until reaching 94 �C. The data (Ct value) auto-matically recorded by 7300 system SDS software were analyzedaccording to the 2�DDCt method (Pfaffl, 2001). The relative expres-sion (fold) was calculated based on the sample with the lowestexpression arbitrarily set as 1. The statistical differences among thesamples were determined using TukeyeKramer test (P < 0.05).

2.6. 20-hydroxyecdysone treatment

The 2-day-old fifth-instar nymphs were used for 20-hydroxyecdysone (20E) treatments based on the distribution of20E titer (Baehr et al., 1978). To determine an appropriate dosage of20E for the experiments, each nymph was injected with 2, 5, 10 and20 mg 20E (Sigma, St. Louis, MO, USA), which was dissolved in 10%ethanol. Insects were collected and used for mRNA expressionanalysis at 2 h after the treatment. We found that the injection of10 mg 20E per nymph could improve the expression of LmCht5genes, effectively (data were not shown), and therefore, this dosewas used in the following experiments. The 2-day-old fifth-instarnymphs (N5D2) were divided into two groups, each with 45 in-sects. One group was injected with 10 mg 20E in each nymph,whereas the other was injectedwith 10% ethanol as a control group.All the treated nymphs were reared in beakers as described above.For transcriptional analysis of each gene, the third and fourthabdominal segments of insects were dissected at 2, 6 and 12 h afterthe treatment, and four locusts were used as a biological replica-tion. Total RNA was isolated and RT-qPCR was carried out for geneexpression analysis following the above steps.

To inhibit biosynthesis of proteins in the nymphs, 10 mg cyclo-heximide (CHX, Sigma, St. Louis, MO, USA), dissolved in dime-thylsulfoxide (DMSO, Fengfan, Tianjin, China) at 2 mg/mL, wasinjected at 2 h before 20E treatment. The insects were divided intofour groups, each with 15 nymphs. Each group was individuallyinjected with 10% ethanol and DMSO (CK), 10% ethanol and CHX

(CHX), 20E and DMSO (20E), or 20E and CHX (20Eþ CHX). Six hourslater, the samples were collected for LmCht5 expression analyses asdescribed above.

2.7. RNAi of ecdysone receptor gene (LmEcR)

Double-stranded RNAs (dsRNAs) were prepared in vitro using T7RiboMAX Express RNAi System (Promega, Madison, WI, USA). Theprimers for dsRNA synthesis of ecdysone receptor (EcR) and greenfluorescent protein (GFP) were designed by E-RNAi webservice(http://www.dkfz.de/signaling/e-rnai3/). The primers used fordsRNA synthesis are shown in Table S1. The templates, whichcontained the sequence of T7 promoter at both ends, were syn-thesized using the PCR program of 94 �C for 3 min; 5 cycles of 94 �Cfor 30 s, 57 �C for 30 s and 72 �C for 45 s; 35 cycles of 94 �C for 30 s,68 �C for 30 s and 72 �C for 45 s; and a final extension of 72 �C for10 min. The PCR products were then purified using E.Z.N.A® GelExtraction Kit (Omega Bio-Tek, Norcross, GA, USA). About 2 mg ofeach purified PCR product was used to synthesize dsRNA, and finalconcentration of the dsRNA was adjusted to 2 mg/mL.

For RNAi of LmEcR, 10 mg dsRNA of LmEcR (dsEcR) was injectedinto the abdomen between the second and third abdominal seg-ments of each 5th-instar nymph (3-day old) by using a micro-syringe. Nymphs in the control group were injected with dsGFP.Each group consisted of 45 nymphs. The injected nymphs werereared in the same condition as previously described. After thetreatment, the integument of the third and fourth abdominal seg-ments of the fifth-instar nymphs at day 5, 6 and 7 (N5D5-N5D7)were collected for expression analyses for LmCht5s by using RT-qPCR and the primers listed in Table S1.

2.8. Functional analysis of LmCht5s by RNAi

To explore the roles of two LmCht5 genes in locust development,10 mg of dsLmCht5-1 or dsLmCht5-2were injected into each of the 2-day-old fifth-instar nymph (N5D2). In addition, we injected 20 mg ofthemixture of dsLmCht5-1 and dsLmCht5-2 (10 mg each dsRNA) intoeach of the N5D2 nymphs to investigate possible effect of RNAi forone gene on another. The methods for dsRNA synthesis and injec-tion were the same as described above. Each RNAi experiment wasrepeated three times, each with 16 nymphs. At 24 h after the dsRNAinjection, integument of four nymphs were collected from eachreplication and used to analyze the suppression of each gene byusing RT-qPCR. The remaining nymphs in each replication weremaintained under the same conditions for observing theirphenotypes.

3. Results

3.1. Identification and sequencing of two LmCht5 cDNAs

We identified two distinct cDNA fragments putatively encodingtwo different LmCht5 proteins from the L.migratoria transcriptomedatabase. These two cDNA fragments were obviously derived fromtwo different genes (LmCht5-1 and LmCht5-2). Further search in theL. migratoria genome database did not yield either additionalLmCht5 gene or its pseudogene. Thus, L. migratoria possesses twoLmCht5 genes.

3.2. Full-length cDNAs and the deduced amino acid sequences oftwo LmCht5 genes

Two cDNA fragments of putative LmCht5 genes were identifiedfrom L. migratoria transcriptome. Blastx analysis suggested thatthey were novel chitinase 5 genes from L. migratoriawith different

D. Li et al. / Insect Biochemistry and Molecular Biology 58 (2015) 46e54 49

molecular characteristics, which were named LmCht5-1 andLmCht5-2, respectively. Alignments with other insect Cht5 se-quences suggested a full-length cDNA of LmCht5-2, but lacks of boththe 50 and 30 ends of LmCht5-1. The full-length cDNA sequence ofLmCht5-1 was obtained from RACE-PCR using the cDNA templateprepared from the insect integument. The full-length cDNA se-quences were deposited at NCBI with the accession numbers ofKM397371 for LmCht5-1 and KM397372 for LmCht5-2.

The complete cDNA sequence of LmCht5-1 contains 1650 nu-cleotides, consisting of an open reading frame (ORF) of 1515 bp thatencoded 505 amino acid residues, and 188- and 67-bp non-codingregions at the 50- and 30-end, respectively. The full-length cDNAsequence of LmCht5-2 contains 2010 nucleotides, consisting of anORF of 1479 bp that encoded 493 amino acid residues, and 57- and477-bp non-coding regions at the 50- and 30-end of the cDNA,respectively (Fig. S1). The calculated MM and pI of LmCht5-1 andLmCht5-2 are 56.1 and 55.0 kDa, and 6.13 and 5.24, respectively.

The domain analysis showed that both LmCht5-1 and LmCht5-2had a signal peptide (26 and 21 amino acid residues, respectively)and a chitinase catalytic domain. However, only LmCht5-1 had achitin-binding domain (Fig. 1A). The regions of two chitinase cat-alytic domain contained 348 and 343 amino acid residues,respectively, and showed 73% sequence identity.

3.3. Gene structure of two chitinase 5 genes

Both LmCht5-1 and LmCht5-2 were located on the scaffold 728,and the gene sizes were 38342 bp and 112334 bp, respectively(Fig. 1B). The distance between two genes was 47103 bp, and eachof the two genes had 10 exons and 9 introns (Fig. 1C). However, theintrons of LmCht5-2 were much larger than that of LmCht5-1. Thesizes of the exons and introns are shown in Table S2.

3.4. Multiple sequence alignment of insect Cht5 catalytic domains

Comparisons with other insect Cht5 catalytic domains showedthat both LmCht5-1 and LmCht5-2 possessed the four conservedmotifs (Fig. 2). In the most important motif region II, all the aminoacid residues including a key residue E are highly conserved be-tween the two duplicated LmCht5s. However, 1e3 amino acid

Fig. 1. Deduced domain architectures, genome structure and exon/intron organizations ofLmCht5-2. Yellow boxes represent catalytic domains and green hexagon represents chitin-bin1 and LmCht5-2. The green box and black line represent DNA sequences of each gene and theExons are shown by green boxes whereas introns are shown by black lines. (For interpretaversion of this article.)

substitutions were found in the other three conserved motifs ofLmCht5-2. There was a conspicuous deletion in each of the catalyticdomains. Specifically, LmCht5-1 missed the consensus sequence(G/S)X(K/N)YS in Region 1 (R1), where X is a non-specific aminoacid residue, and LmCht5-2 missed the consensus sequence (S/T)(F/Y)INKEAGGG in Region 2 (R2).

3.5. Phylogenetic analysis of two LmCht5 protein sequences

A phylogenetic tree was constructed based on the catalytic do-mains of LmCht5-1 LmCht5-2 and Cht5s from other insect species(Fig. 3). These Cht5s were divided into two groups supported by abootstrap value of 100 after 5000 replications. The group with thegray background included LmCht5-1, LmCht5-2, Cht5-1s of threemosquito species and other Cht5s from the insect species with onlyone Cht5 identified. These Cht5s clustered in a unique node inde-pendent of the node of Cht5-2s, Cht5-3s, Cht5-4s and Cht5-5 fromthree mosquito species, suggesting that the event of LmCht5duplication is independent of that of Cht5 duplications in mosqui-toes. Lepidopteran Cht5 proteins formed themselves a node withbootstrap value of 100. Sequence conservation was indicated bybranch length. The branch length of LmCht5-2 was much longerthan LmCht5-1, implying that LmCht5-2 and LmCht5-1 werediverged long time ago.

The ratio of nonsynonymous substitution rate (Ka) to synony-mous substitution rate (Ks) has beenwidely used for measuring theselective pressure on a protein-coding sequence (Yang and Nielsen,2000). The LmCht5-1 and LmCht5-2 pairwise Ka value was 0.384, Ksvalue was 4.13, and the Ka/Ks ratio was 0.093. Because the Ka/Ksratio was much smaller than 1, our results indicate that both theLmCht5 genes are under strong negative selection.

3.6. Tissue and developmental expression profiles of two LmCht5genes

To determine whether each of the two LmCht5 genes wasdifferentially expressed, we analyzed transcript in various tissues ofN5D6 nymphs and in different developmental stages by RT-qPCR(Fig. 4). The MIQE checklist was shown in Table S3. The expres-sion of LmCht5-1 was relatively consistent in seven tested tissues

two chitinase 5 genes. A: Schematic diagram of deduced domains of LmCht5-1 andding domains. Blue triangles show the signal peptides. B: Genome structure of LmCht5-linker region. C: Schematic diagram of the exon and intron organization of two genes.

tion of the references to colour in this figure legend, the reader is referred to the web

Fig. 2. Multiple sequence alignments of catalytic domains of the deduced chitinase 5 proteins in insects. The four conserved motifs (CM) sequences are boxed in red and denoted asCM1, CM2, CM3 and CM4. The different region (R) of LmCht5-1 and LmCht5-2 are boxed in green and denoted as R1 and R2. Fully conserved amino acid sequences are shaded inblack. Anopheles gambiae, Ag, HQ456129; Apis mellifera, Am, XP_623995.1; Drosophila melanogaster, Dm, CG9307; Manduca sexta, Ms, P36362; Tribolium castaneum, Tc, AY675073;Locusta migratoria, Lm, for LmCht5-1, for LmCht5-2. Catalytic domains of chitinase 5s were analyzed by SMART software. (For interpretation of the references to colour in this figurelegend, the reader is referred to the web version of this article.)

D. Li et al. / Insect Biochemistry and Molecular Biology 58 (2015) 46e5450

including integument (IN), foregut (FG), midgut (MG), gastric caeca(GC), hindgut (HG), Malpighian tubules (MT) and fat bodies (FB),with relatively higher expression in the hindgut. The expression ofLmCht5-2wasmuch higher in IN, FG, HG and FB than inMG, GC andMT (Fig. 4A).

During the development from fourth-instar nymphs to adults,the expression patterns of the two genes were similar. However,both genes showed rapid increases in expression level before eachmolting. Furthermore, LmCht5-1 maintained a relatively highexpression level during the adult stage (Fig. 4B).

3.7. Regulation of two LmCht5 genes by 20E

Based on the expression profiles of the two LmCht5 genes anddistribution of 20E titers (Baehr et al., 1978), the N5D2 nymphswere treated with 20E and the dsRNA of EcR (dsLmEcR) to inves-tigate whether LmCht5-1 and LmCht5-2 were regulated by 20E.The mRNA levels of both genes were remarkably induced by 20Eat 2 h and the peak of expression was observed at 6 h with a 27-fold increase for LmCht5-1 and a 410-fold increase for LmCht5-2.The increased expressions maintained at 12 h after 20E injection(Fig. 5A). To detect whether the expressions of the two genes weredirectly regulated by 20E, cycloheximide (CHX), a proteinbiosynthesis inhibitor, was used for injection prior to 20E treat-ment. The results showed that CHX had no effect on the 20E-induced up-regulation of LmCht5-2, but reduced the mRNA level ofLmCht5-1 by half (Fig. 5B). After LmEcR was silenced by injectingdsLmEcR in N5D2 nymphs, the transcript levels of both LmCht5s

were significantly suppressed at days 5e6 of the fifth-instarnymphs.

3.8. Functional analysis of LmCht5s by RNAi

To determine whether LmCht5-1 and LmCht5-2 displayed anydifference in their functions, each of sequence-specific dsRNAs forLmCht5-1 and LmCht5-2 was injected into the N5D2 nymphs. At24 h after the injection, we found that injections of dsLmCht5-1 anddsLmCht5-2 selectively suppressed their target genes (Fig. 6A andB).

We observed three distinct phenotypes (p1-p3) after the in-jection of dsLmCht5-1 (Fig. 6C and D). In the first phenotype (p1),dsRNA for LmCht5-1 had no effect on molting. The second pheno-type (p2) was associated with the failure of breaking up the oldcuticle on the head and abdomen even though the old cuticle ofpronotum was split open. In this case, the nymph body wasemboweled in the old cuticle leading to death. The third phenotype(p3) was associated with the inability of shedding the old cuticle ofthe appendages although the old cuticle of the remaining bodyparts could be successfully shed. In this case, the insect walkedslowly and could not jump to search for food. Eventually theseinsects also died. The proportion of three phenotypes were shownin Fig. 6C. In contrast, no abnormal phenotypes were observed inthe nymphs after the LmCht5-2 transcript was suppressed by RNAi.Injection of the dsRNA mixture for the two genes also resulted inthe same phenotypes as observed with the injection of dsLmCht5-1alone and no synergistic effects were observed.

Fig. 3. Phylogenetic tree of deduced chitinase 5 proteins from 19 insect species basedon their catalytic domains. ClustalW program was used to perform multiple sequencealignments prior to phylogenetic analysis. The phylogenetic tree was constructed byMEGA 6.06 software using neighbor-joining method. Aedes aegypti, Aa,XP_001656234.1 for AaCht5-1, XP_001656233.1 for AaCht5-2, XP_001656232.1 forAaCht5-3, and XP_001656231.1 for AaCht5-4; Apis mellifera, Am, XP_623995.1;Anopheles gambiae, Ag, HQ456129 for AgCht5-1, HQ456130 for AgCht5-2, HQ456131for AgCht5-3, HQ456132 for AgCht5-4, and HQ456133 for AgCht5-5; Bombyx man-darina, Bma, AAG48700.1; Bombyx mori, Bmo, AAB47538; Choristoneura fumiferana, Cf,AAM43792; Culex quinquefasciatus, Cq, XP_001863384.1 for CqCht5-1, XP_001863385.1for CqCht5-2 and XP_001863386.1 for CqCht5-3; Drosophila melanogaster, Dm,CG9307; Helicoverpa armigera, Ha, AAQ91786; Hyphantria cunea, Hc, AAB47537;Lacanobia oleracea, Lo, CAF05663; Mamestra brassicae, Mb, JN558350; Manduca sexta,Ms, P36362; Ostrinia furnacalis, Of, AAW50396.1; Spodoptera exigua, Se, AEV22117.1;Spodoptera frugiperda, Sf, AAS18266; Spodoptera litura, Sl, AB032107; Tribolium casta-neum, Tc, AY675073; Locusta migratoria, Lm, KM397371 for LmCht5-1 and KM397372for LmCht5-2.

D. Li et al. / Insect Biochemistry and Molecular Biology 58 (2015) 46e54 51

4. Discussion

Gene duplication is a common phenomenon in organisms and isa major mechanism contributing to new gene generation. Themultiple chitinase 5 genes have been identified in three mosquitospecies, including five in An. gambiae, four in Ae. aegypti and threein C. quinquefasciatus (Zhang et al., 2011b). Because the geneduplication of chitinase 5 has not been observed in T. castaneum, D.melanogaster and any other insect species examined, it was

assumed that the gene duplication of Cht5 might be unique to themosquito lineage (Zhang et al., 2011b). In this study, we uncovereda similar phenomenon of gene duplication for LmCht5 in L. migra-toria. Two cDNA sequences putatively encoding two different Cht5swere obtained and designated as LmCht5-1 and LmCht5-2. Ouranalysis of the domain structures suggests that LmCht5-1 isorthologous tomosquito Cht5-1s and other Cht5 proteins identifiedin other insect species. Although LmCht5-2 lost the consensussequence (S/T) (F/Y)INKEAGGG in Region 2 (R2) of the catalyticdomain and a cysteine-rich chitin-binding domain in c-terminalregion, our analysis of a phylogenetic tree still supports a closerelationship of LmCht5-2 with insect Cht5s (bootstrap value: 100),but it is far away from all the duplicated Cht5s in the mosquitoes(Cht5-2s to Cht5-5) derived from Cht5-1. Amino acid sequencealignment showed that all the duplicated Cht5s with the exceptionof LmCht5-2 lack the catalytically critical glutamate (E) residue inconserved motif region II. It has been postulated that the substi-tution of E in DWEYP could result in a total loss of catalytic activityin some of these proteins (Lu et al., 2002). As expected, LmCht5-1 isclosely related to Cht5-1 of the insects with multiple Cht5 genesand Cht5 of the insects with only single Cht5 gene (Fig. 4). However,LmCht5-2 is more closely related to LmCht5-1 than to Cht5-2s,Cht5-3s, Cht5-4s and Cht5-5 of the mosquito species due to over-all low sequence similarities of Cht5s between the locust andmosquito species, which explains why LmCht5-2 is out-groupedfrom mosquito Cht5-2s, Cht5-3s, Cht5-4s and Cht5-5. BecauseLmCht5-1 is more closely related to other insect Cht5s with shorterbranch compared with LmCht5-2, it is likely that LmCht5-2 hasbeen derived from a more recent gene duplication event.

Our analyses of the development-dependent expression pat-terns of LmCht5s generated some very interesting results. Thedramatically increased expressions of both LmCht5-1 and LmCht5-2from fourth-instar to fifth-instar nymphs and then to adults wereobserved during a specific time prior to each molting. These resultsstrongly suggest that both LmCht5-1and LmCht5-2 are responsiblefor chitin metabolism in the cuticle turnover. Furthermore, such anexpression pattern appears to coincide with the 20 ecdysteroidstiters (Baehr et al., 1978), suggesting that the expressions ofLmCht5s are under endocrine regulation. For the high expressionlevel of LmCht5-1 in adults, we hypothesized that LmCht5-1 mightbe involved in early adult development or fecundity of locusts.Indeed high transcripts of chitin synthase 1 have been found in L.migratoria (Zhang et al., 2010) and T. castaneum (Arakane et al.,2008) during the early adult stage. These results suggest thatchitin metabolic pathways could be still very active in the earlyadult stage.

To investigate whether the expressions of LmCht5s were regu-lated by 20E, we injected 20E into the 2-day-old fifth-instarnymphs. As expected, both the LmCht5 transcripts sharplyincreased at 2 h with a dramatic peak expression (27-fold forLmCht5-1 and 410-fold for LmCht5-2) at 6 h after the injection of20E. These data clearly showed that the expression of both LmCht5genes in the integument can be effectively up-regulated by 20E.Conversely, after the injection of dsLmEcR in the 2-day-old fifth-instar nymphs to suppress the expression of LmEcR, we found 2.5and 8-fold down-regulations of LmCht5-1 in the 5 and 6-day-oldfifth-instar nymphs, and 15 and 9.2-fold down-regulations ofLmCht5-2 in the 5 and 6-day-old fifth-instar nymphs, respectively.Because insect EcR is a main component of the ecdysone receptorheterodimer, the suppression of LmEcR expression by RNAi is ex-pected to prevent the 20E-induced expressions of both LmCht5-1and LmCht5-2. These results support our notion that the expres-sions of both LmCht5-1and LmCht5-2 are regulated by 20E.

To determine whether or not the 20E-induced up-regulations ofLmCht5-1 and LmCht5-2 were due to direct action of 20E on the

Fig. 4. Relative expression levels of LmCht5-1 and LmCht5-2 in different tissues of N5D6 nymphs (A) by and integument of N4D1-AD7 insects (B) by RT-qPCR. Integument (IN),foregut (FG), midgut (MG), gastric caeca (GC), hindgut (HG), Malpighian tubules (MT) and fat bodies (FB). N5D6: the sixth day of the 5th instar nymphs. N4D1-AD7: the first day ofthe 4th instar nymphs to the seventh day of adults. Different letters on the bars in the same histogram indicate significant difference of the expression among different samplesbased on three biological replications (P < 0.05, TukeyeKramer test, n ¼ 3).

D. Li et al. / Insect Biochemistry and Molecular Biology 58 (2015) 46e5452

regulatory region of these genes, or indirect action through theinteractions with transcription factors, we assessed the effect of theprotein biosynthesis inhibitor cycloheximide (CHX) on the 20E-induced up-regulations of LmCht5-1 and LmCht5-2. When CHX wasinjected to the 2-day-old fifth-instar nymphs prior to the 20E in-jection, we found that CHX had no effect on the 20E-induced up-regulation of LmCht5-2, but restrained the transcript level ofLmCht5-1 under the same conditions, These results strongly sug-gest that LmCht5-2 is a 20E primary-response gene, while LmCht5-1shows both primary and secondary response to 20E. Indeed, 20Ecan up-regulate gene expression ofMsCht5 in both integument andgut of M. sexta (Kramer et al., 1993), and also enhanced proteinexpression of CfCht5 in C. fumiferana (Zheng et al., 2003). Our re-sults not only supported their findings but also extended the reg-ulatory effect of 20E on duplicated LmCht5-2. Recently, miR-24 hadbeen identified to regulate the expression of HaCht5 both in vitroand in vivo in H. armigera (Agrawal et al., 2013). Therefore, it is mostlikely that the expression of insect Cht5 is regulated by multiplefactors.

Although the duplicated Cht5 genes were previously identi-fied in three mosquito species (Zhang et al., 2011b), the role of

each of these duplicated genes has not been examined. In thisstudy, we took the advantage of robust RNAi response in L.migratoria to reveal biological role of LmCht5-1 and LmCht5-2.Our study showed disrupted molting from nymphs to adults afterthe injection of dsLmCht5-1. In contrast, no visible phenotype wasobserved after the injection of dsLmCht5-2. These results clearlyindicate functional difference of the two duplicated LmCht5genes.

Because the catalytic and chitin-binding domains are twoimportant structural components of chitinases, we compareddeduced amino acid sequences between LmCht5-1 and LmCht5-2.Apparently, LmCht5-2 lost a chitin-binding domain, but retainedthe critical residue E in the conserved region (II) of catalyticdomain. Thus, LmCht5-2 might still be catalytically active (Lu et al.,2002; Zhu et al., 2008c). However, further studies are needed toclarify whether or not LmCht5-2 pays any significant role in thedevelopment of L. migratoria. Nevertheless, our study suggestedthat LmCht5-1 and LmCht5-2 had diverged via sub-functionalization. LmCht5-1 had an important role in locust molt-ing, but the divergent function of LmCht5-2 remained to be clarifiedin the future.

Fig. 5. Relative expression levels of LmCht5-1 and LmCht5-2 after 20E treatment and RNAi of LmEcR or GFP. A: Effect of 20E on LmCht5s mRNA expression. mRNAs were extracted at2, 6 and 12 h after 20E treatment. Control: 10% ethanol injection; 20E: 20E injection; **, highly significant (P < 0.01, Student's t test, n ¼ 3). B: Effect of cycloheximide on 20E-inducedLmCht5s mRNA up-regulation. CK: 10% ethanol and DMSO injection; CHX: 10% ethanol and cycloheximide injection; 20E: 20E and DMSO injection; 20E þ CHX: 20E and cyclo-heximide injection. Different letters on the bars in the same histogram indicate significant difference of the expression among different samples based on three biological repli-cations (P < 0.05, TukeyeKramer test, n ¼ 3). C: The expression of LmCht5s in N5D5-N5D7 locusts after dsGFP or dsEcR injection (P < 0.01, Student's t test, n ¼ 3).

Fig. 6. Effects of dsLmCht5 injection in the 5th-instar nymphs on the LmCht5 transcript levels and development of L. migratoria. A and B: Relative expression levels of two LmCht5genes in integuments after the dsRNA injection as evaluated by RT-qPCR. The relative expression (fold) was calculated based on the value of the lowest expression which wasascribed an arbitrary value of 1. b-actinwas used as an internal reference gene (P < 0.01, TukeyeKramer test, n ¼ 3). C: Percentages of three phenotypes of the locusts injected withdsLmCht5-1. p1, p2, p3: Three types of locusts phenotypes. D: The phenotypes of the nymphs after the LmCht5-1, LmCht5-2 or GFP dsRNA injection (Scale bar: 5 mm). p1: the nymphdeveloped into adults successfully; p2: the nymph body was trapped in the old cuticle leading to death; p3: the old cuticle was retained the appendages, and adult was unable tomove freely.

D. Li et al. / Insect Biochemistry and Molecular Biology 58 (2015) 46e54 53

D. Li et al. / Insect Biochemistry and Molecular Biology 58 (2015) 46e5454

Acknowledgments

This work was supported by National Basic Research Program ofChina (2012CB114102), National Natural Science Foundation ofChina (Grant No. 31272380), Program for Top Young AcademicLeaders of Higher Learning Institutions of Shanxi (TYAL), andResearch Funds for the Doctoral Program of Higher Education ofChina (20121401110008). The authors would like to give specialthanks to Prof. Le Kang at the Institute of Zoology, Chinese Academyof Sciences, for sharing the resource of the locust EST and genomedatabase platform.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.ibmb.2015.01.004.

References

Agrawal, N., Sachdev, B., Rodrigues, J., Sree, K.S., Bhatnagar, R.K., 2013. Developmentassociated profiling of chitinase and microRNA of Helicoverpa armigera identi-fied chitinase repressive microRNA. Sci. Rep. 3, 2292.

Ahmad, T., Rajagopal, R., Bhatnagar, R.K., 2003. Molecular characterization of chi-tinase from polyphagous pest Helicoverpa armigera. Biochem. Biophys. Res.Commun. 310, 188e195.

Arakane, Y., Muthukrishnan, S., 2009. Insect chitinase and chitinase-like proteins.Cell. Mol. Life Sci. 67, 201e216.

Arakane, Y., Specht, C.A., Kramer, K.J., Muthukrishnan, S., Beeman, R.W., 2008. Chitinsynthases are required for survival, fecundity and egg hatch in the red flourbeetle, Tribolium castaneum. Insect Biochem. Mol. Biol. 38, 959e962.

Baehr, J.C., Porcheron, P., Papillon, M., Dray, F., 1978. Heamolymph levels of juvenilehormone ecdysiteroid and protein during the last two larval instars of Locustamigratoria. J. Insect Physiol. 25, 415e421.

Chen, L., Liu, T., Zhou, Y., Chen, Q., Shen, X., Yang, Q., 2014a. Structural characteristicsof an insect group I chitinase, an enzyme indispensable to moulting. Acta.Crystallogr. D. Biol. Crystallogr. 70 (Pt 4), 932e942.

Chen, L., Zhou, Y., Qu, M., Zhao, Y., Yang, Q., 2014b. Fully deacetylated chitooligo-saccharides act as efficient glycoside hydrolase family18 chitinase inhibitors.J. Biol. Chem. 289, 17932e17940.

Feix, M., Gl€oggler, S., Londershausen, M., Weidemann, W., Spindler, K.D., Spindler-Barth, M., 2000. A cDNA encoding a chitinase from the epithelial cell line ofchironomus tentans (Insecta, diptera) and its functional expression. Arch. InsectBiochem. Physiol. 45, 24e36.

Fitches, E., Wilkinson, H., Bell, H., Bown, D.P., Gatehouse, J.A., Edwards, J.P., 2004.Cloning, expression and functional characterization of chitinase from larvae oftomato moth (Lacanobia oleracea): a demonstration of the insecticidal activityof insect chitinase. Insect Biochem. Mol. Biol. 34, 1037e1050.

Gopalakrishnan, B., Muthukrishnan, S., Kramer, K.J., 1995. Baculovirus-mediatedexpression of a Manduca sexta chitinase gene: properties of the recombinantprotein. Insect Biochem. Mol. Biol. 25, 255e265.

Kim, M.G., Shin, S.W., Bae, K.S., Kim, S.C., Park, H.Y., 1998. Molecular cloning ofchitinase cDNAs from the silkworm, Bombyx mori and the fall webworm,Hyphantria cunea. Insect Biochem. Mol. Biol. 28, 163e171.

Kramer, K.J., Corpuz, L., Choi, H.K., Muthukrishnan, S., 1993. Sequence of a cDNA andexpression of the gene encoding epidermal and gut chitinases ofManduca sexta.Insect Biochem. Mol. Biol. 23, 691e701.

Kramer, K.J., Muthukrishnan, S., 2005. Chitin metabolism in insects. In: Gilbert, L.I.,Iatrou, K., Gill, S.S. (Eds.), Comprehensive Molecular Insect Science, vol. 4.Elsevier, New York, pp. 111e144.

Lu, Y., Zen, K.C., Muthukrishnan, S., Kramer, K.J., 2002. Site-directed mutagenesisand functional analysis of active site acidic amino acid residues D142, D144 andE146 in Manduca sexta (tobacco hornworm) chitinase. Insect Biochem. Mol.Biol. 32, 1369e1382.

Merzendorfer, H., 2013. Insect-derived chitinases. In: Vilcinskas, A. (Ed.), YellowBiotechnology II Insect Biotechnology in Plant Protection and Industry. Adv.Biochem. Eng. Biotechnol., vol. 136. Springer-Verlag, Berlin Heidelberg,pp. 19e50.

Moussian, B., 2010. Recent advances in understanding mechanisms of insect cuticledifferentiation. Insect Biochem. Mol. Biol. 40, 363e375.

Nakabachi, A., Shigenobu, S., Miyagishima, S., 2010. Chitinase-like proteins encodedin the genome of the pea aphid, Acyrthosiphon pisum. Insect Mol. Biol. 19(Suppl. 2), 175e185.

Paek, A., Park, H.Y., Eun, Jeong S., 2012. Molecular cloning and functional expressionof chitinase-encoding cDNA from the cabbage moth, Mamestra brassicae. Mol.Cells 33, 439e447.

Pan, Y., Lü, P., Wang, Y., Yin, L., Ma, H., Ma, G., Chen, K., He, Y., 2012. In silico iden-tification of novel chitinase-like proteins in the silkworm, Bombyx mori,genome. J. Insect Sci. 12, 150.

Pfaffl, M.W., 2001. A new mathematical model for relative quantification in real-time RT-PCR. Nucleic. acids. Res. 29, e45.

Shinoda, T., Kobayashi, J., Matsui, M., Chinzei, Y., 2001. Cloning and functionalexpression of a chitinase cDNA from the common cutworm, Spodoptera litura,using a recombinant baculovirus lacking the virus-encoded chitinase gene.Insect Biochem. Mol. Biol. 31, 521e532.

Wang, X., Fang, X., Yang, P., Jiang, X., Jiang, F., Zhao, D., Li, B., Cui, F., Wei, J., Ma, C.,Wang, Y., He, J., Luo, Y., Wang, Z., Guo, X., Guo, W., Wang, X., Zhang, Y., Yang, M.,Hao, S., Chen, B., Ma, Z., Yu, D., Xiong, Z., Zhu, Y., Fan, D., Han, L., Wang, B.,Chen, Y., Wang, J., Yang, L., Zhao, W., Feng, Y., Chen, G., Lian, J., Li, Q., Huang, Z.,Yao, X., Lv, N., Zhang, G., Li, Y., Wang, J., Wang, J., Zhu, B., Kang, L., 2014. Thelocust genome provides insight into swarm formation and long-distance flight.Nat. Commun. 5, 2957.

Wu, Q., Liu, T., Yang, Q., 2013. Cloning, expression and biocharacterization of OfCht5,the chitinase from the insect Ostrinia furnacalis. Insect Sci. 20, 147e157.

Yang, Z., Nielsen, R., 2000. Estimating synonymous and nonsynonymous substitu-tion rates under realistic evolutionary models. Mol. Biol. E 17, 32e43.

Zhang, J.Z., Zhang, X., Arakane, Y., Muthukrishnan, S., Kramer, K.J., Ma, E.B., Zhu, K.Y.,2011a. Comparative genomic analysis of chitinase and chitinase-like family inthe African Malaria Mosquito (Anopheles gambiae). PLoS ONE 6, e19899.

Zhang, J.Z., Zhang, X., Arakane, Y., Muthukrishnan, S., Kramer, K.J., Ma, E.B., Zhu, K.Y.,2011b. Identification and characterization of a novel chitinase gene cluster(AgCht5) possibly derived from tandem duplications in the African malariamosquito, Anopheles gambiae. Insect Biochem. Mol. Biol. 41, 521e528.

Zhang, D., Chen, J., Yao, Q., Pan, Z., Chen, J., Zhang, W., 2012. Functional analysis oftwo chitinase genes during the pupation and eclosion stages of the beetarmyworm Spodoptera exigua by RNA interference. Arch. Insect Biochem.Physiol. 79, 220e234.

Zhang, J.Z., Liu, X.J., Zhang, J.Q., Li, D.Q., Sun, Y., Guo, Y.P., Ma, E.B., Zhu, K.Y., 2010.Silencing of two alternative splicing-derived mRNA variants of chitin synthase 1gene by RNAi is lethal to the oriental migratory locust. Insect Biochem. Mol.Biol. 40, 824e833.

Zheng, Y.P., Zheng, S., Cheng, X., Ladd, T., Lingohr, E.J., Krell, P.J., Arif, B.M.,Retnakaran, A., Feng, Q.L., 2002. A molt-associated chitinase cDNA from thespruce budworm, Choristoneura fumiferana. Insect Biochem. Mol. Biol. 32,1813e1823.

Zheng, Y.P., Retnakaran, A., Krell, P.J., Arif, B.M., Primavera, M., Feng, Q.L., 2003.Temporal, spatial and induced expression of chitinase in the spruce budworm,Choristoneura fumiferana. J. Insect Physiol. 49, 241e247.

Zhu, Q., Arakane, Y., Banerjee, D., Beeman, R.W., Kramer, K.J., Muthukrishnan, S.,2008a. Domain organization and phylogenetic analysis of the chitinase likefamily of proteins in three species of insects. Insect Biochem. Mol. Biol. 38,452e466.

Zhu, Q., Arakane, Y., Beeman, R.W., Kramer, K.J., Muthukrishnan, S., 2008b. Func-tional specialization among insect chitinase family genes revealed by RNAinterference. Proc. Natl. Acad. Sci. U. S. A. 105, 6650e6655.

Zhu, Q., Arakane, Y., Beeman, R.W., Kramer, K.J., Muthukrishnan, S., 2008c. Char-acterization of recombinant chitinase-like proteins of Drosophila melanogasterand Tribolium castaneum. Insect Biochem. Mol. Biol. 38, 467e477.

Zhu, Q., Deng, Y., Vanka, P., Brown, S.J., Muthukrishnan, S., Kramer, K.J., 2004.Computational identification of novel chitinase-like proteins in the Drosophilamelanogaster genome. Bioinformatics 20, 161e169.