Salivary gland transcriptome analysis duringPlasmodium infection in malaria vectorAnopheles stephensi

Rajnikant Dixit a,*, Arun Sharma b, Devendra T. Mourya c,Raghavendra Kamaraju b, Millind S. Patole a, Yogesh S. Shouche a

aMolecular Biology Unit, National Center for Cell Science, Ganeshkhind, Pune, IndiabNational Institute of Malaria Research, Delhi, IndiacMicrobial Containment Complex, National Institute of Virology, Pashan, Pune, India

Received 14 January 2008; received in revised form 19 June 2008; accepted 12 July 2008Corresponding Editor: William Cameron, Ottawa, Canada

International Journal of Infectious Diseases (2009) 13, 636—646

http://intl.elsevierhealth.com/journals/ijid

KEYWORDSAnopheles;Mosquito;Salivary glands;Malaria;Plasmodium

Summary

Background: Understanding the tissue-specific molecular cross-talk mechanism during the mos-quito—parasite interaction is of prime importance in the design of new strategies for malariacontrol. Because mosquito salivary glands are the final destination for the parasite maturationand transmission of vector-borne diseases, identification and characterization of salivary genesand their products are equally important in order to access their effect on the infectivity of theparasite. During the last five years there have been several studies on the sialomes of Anophelesmosquitoes, however very limited information is available on the changes in the salivary glandtranscriptome in the presence of Plasmodium, and this information is limited to the mosquitoAnopheles gambiae.Methods: In this study we aimed to explore and identify parasite-induced transcripts from thesalivary glands of Anopheles stephensi, using a subtractive hybridization protocol.Results: Ninety-four percent of expressed sequence tags (ESTs) showed close homology topreviously known families ofmosquito salivary gland secretary proteins, representing the inducedexpression of alternative splicing and/or additional new members of the protein family. Theremaining 6% of ESTs did not yield significant homology to any known proteins in the non-redundant database and thus may represent a class of unknown/novel salivary proteins. Primaryanalysis of the ESTs also revealed identification of several novel immune-related transcripts,including defensin and cecropins, probably involved in counter-activation of the antagonisticdefense system. A comprehensive description of each family of proteins has been discussed inrelation to the tissue-specific mosquito—parasite interaction.

* Corresponding author. Current address: Laboratory of Malaria and Vector Research, National Institute of Allergy and Infectious Diseases,National Institute of Health, Rockville, MD 20852, USA. Tel.: +1 301 435 3182; fax: +1 301 480 1337.

E-mail address: dixit2k@yahoo.com (R. Dixit).

1201-9712/$36.00 # 2008 International Society for Infectious Diseases. Published by Elsevier Ltd. All rights reserved.

doi:10.1016/j.ijid.2008.07.027

Conclusion: This is the first report on the identification of new putative salivary genes, pre-sumably activated during parasite infection.# 2008 International Society for Infectious Diseases. Published by Elsevier Ltd. All rights reserved.

Salivary gland transcriptome, Anopheles stephensi 637

Introduction

Mosquito-borne diseases including malaria, dengue, Japa-nese encephalitis, yellow fever, and filariasis, remain amajor health problem in humans and the veterinary sector.Among these diseases, malaria alone affects more than200 million people worldwide, with 1.5 million deaths,annually.1 This disease is caused by the Plasmodium para-site transmitted by obligatory blood sucking anophelinemosquito species. During its complex life cycle in themosquito, the Plasmodium parasite undergoes severaldevelopmental transitions, traverse the midgut and salivarygland epithelia, and withstand the host’s immuneresponse.2 Several studies have shown that parasite block-age can occur during passage through the midgut3,4 or thesalivary glands,5 leading to the possibility of designing newstrategies for malaria control.6

Mosquito saliva and salivary glands are central to theinteraction between the parasite, vector, and mammalianhost.7 The salivary glands and their diversified gene pro-ducts are essential in overcoming the challenges posedduring blood feeding and parasite transmission. There isconvincing evidence that the pharmacological activity ofarthropod salivary gland molecules affects pathogen trans-mission, e.g., salivary gland lysate from the sandfly Lutzo-myia longipalpis facilitates the infection of mice by theprotozoan parasite Leishmania major.8 During the last fiveyears there have been several studies on the sialome ofAnopheles mosquitoes,9—13 however, very little work hasbeen done on the changes in the salivary gland transcrip-tome and/or proteome in the presence of Plasmodium, andthis too has been limited to the mosquito Anopheles gam-biae.14,15

Anopheles stephensi is an important vector of humanmalaria throughout the Middle East and South Asia regions,including Indo-Pakistan, and belongs to the same (Celia)subgenus as the most efficient African vector A. gambiae.16

A. stephensi is susceptible to both human and rodent malariaspecies such as Plasmodium berghei; for this reason it iswidely used as a laboratory model to study host—parasiteinteraction biology. Recent gene discovery projects based onsubtraction cDNA libraries have identified numerous newtranscripts of mosquito midgut origin that are crucial forinteractions between A. stephensi and P. berghei.17—19 How-ever, to date no attempts have been made to study thetranscriptome of the A. stephensi salivary gland in the pre-sence of the malaria parasite.

As a first step towards deriving information on thetranscriptional changes in the salivary glands during Plas-modium infection, we generated and analyzed a small-scalesubtraction cDNA library from the salivary gland of themosquito A. stephensi. Remarkably, several new putativegenes were identified including immune responsive tran-scripts and others with unknown function, which may playan important role in parasite transmission. Together, a set

of A. stephensi salivary expressed sequence tags (ESTs) andthe Plasmodium genome currently available may provideadditional tools for the comprehensive analysis of mole-cules that may play an active role in the transmission ofmalaria.

Materials and methods

Mosquito infection

A. stephensi mosquitoes were fed on P. berghei-infectedmice and maintained at 21 8C. Sporozoite numbers of 10—15 salivary glands dissected on day 15 after an infectiousblood meal were counted to assess the infection.

Total RNA preparation and cDNA synthesis

Total RNA was extracted from 50 dissected salivary glandsusing the TRIZOL reagent (Invitrogen) and treated withRNase-free DNase I to remove any genomic DNA contamina-tion. DNase-treated total RNA (approximately 1 mg) was thenreverse-transcribed, and the quality of cDNA synthesis wasassessed using ribosomal protein S7 gene-specific primers asdescribed previously.20 Single-stranded cDNAs (ss-cDNAs)were amplified to make double-stranded cDNA (ds-cDNA)through a PCR-based protocol using SMART cDNA synthesiskit (catalogue No. 634902; BD Clontech, Palo Alto, CA, USA)following the manufacturer’s instructions.

Construction of the subtracted cDNA library

To generate a subtracted cDNA library, a suppression sub-tractive hybridization protocol was performed using the PCR-Select cDNA subtraction kit (catalogue No. 637401; BD Clon-tech), following the manufacturer’s protocol. Briefly, 1 mgtotal RNA from infected salivary glands (tester) and unin-fected/control (driver) were used to prepare ds-cDNAs asdescribed above. Then ds-cDNA was purified and digestedwith RSaI enzyme. The tester cDNA was divided into twoaliquots, each of which was ligated to one of two differentadaptors, adaptor 1 or 2R. Both adaptor-ligated sampleswere denatured at 98 8C for 90 s and then hybridized at68 8C for 8 h with an excess of driver cDNA. These twoprimary hybridized samples were mixed together withoutdenaturing, and fresh denatured driver cDNA was added tofurther enrich differentially expressed genes by hybridizingat 68 8C for 10 h. The entire population of cDNAmolecule wasthen subjected to two rounds of suppression PCR for selectiveamplification of the differential transcripts. The resultingsecondary PCR products were cloned into pGEM-T vector(Promega) and transformed to Escherichia coli (DH5a) com-petent cells. The cloned inserts were PCR-amplified usingvector-specific M13 primers and sequenced with M13F primerin an ABI3730 sequencer.

638 R. Dixit et al.

Bioinformatic analysis of the databases

Each individual sequence chromatogramwas examinedmanu-ally. Vector- and end-specific low quality sequences wereremoved computationally from all sequences, using vectordatabase-specific BLAST. Then each individual sequence(FASTA FORMAT) was searched against non-redundant andother EST databases available at the National Center forBiotechnology and Information (NCBI) using BLASTN andBLASTX algorithms. ESTs were assembled into contigs usingCAP assembly program (BioEdit, version 7.0.8). Based on thepredicted function of the homologous protein and significantBLASTX similarities, contigs were grouped into a relativefamily of salivary proteins. Sequences with no significantBLASTX similarity were grouped based on their BLASTN simila-rities. ClustalX and GenDoc software were used for multiplesequence alignment and sequence identity analysis. All ESTshave been submitted to the NCBI GenBank (accession numbersEL595671 to EL595866) for public domain use.

Results and discussion

Primary characterization of the subtractive cDNAlibrary

A subtractive hybridization protocol was performed to iden-tify A. stephensi genes that are expressed in salivary glands inresponse to P. berghei sporozoite invasion. The blood-fedsalivary gland ds-cDNAs were subtracted from the cDNAs ofsporozoite-infected salivary glands using a PCR-based sub-tractive hybridization protocol. To minimize the risk of PCRartifacts, we optimized the PCR cycle numbers to a minimumto amplify initial cDNA populations. We were unable to seeany amplification of the RpS7 gene in subtracted cDNA poolsas compared to non-subtracted cDNA samples, even after 40PCR cycles, indicating a significant depletion of the commontranscripts and enrichment of rare transcripts. A total ofapproximately 106 primary clones were obtained for thesubtracted cDNA library. About 210 randomly selected posi-tive recombinant clones (white colonies) were subjected tohigh throughput sequencing and nearly 197 good qualitysequences were recovered. These sequences were furtheranalyzed to generate an EST catalogue for salivary gland-specific genes presumably activated in response to parasiteinfection. Table 1 shows the primary features of the subtrac-tion cDNA library.

Table 1 Characteristics of the subtraction cDNA library.

Total number of positive recombinant clonessequenced

210

Good quality ESTs analyzed 197Total number of clusters assembled 19Singletons/unassembled sequences 13ESTs homologous to known anophelinesalivary gland proteins

88%

New ESTs of salivary gland origin relatedto immunity

5.5%

ESTs with ribosomal protein origin 0.5%ESTs with parasite origin 0.5%ESTs with unknown function 6%

Description of the mosquito salivarytranscriptome

All ESTs could be assembled into 32 contigs, out of which 13sequences remained unassembled and grouped as singleton.Most of these (88%) sequences showed highest homology topreviously known salivary gland sequences of the anophelinemosquito. However, of the remaining 12% of sequences, 6%ESTs showed homology to immune-related genes and 6% ESTsdid not yield significant similarity to any known proteinsequences in the database, and may represent a class ofnew ESTs. With this information, the assembled contigs wereassigned putative functions based on close relatedness to afamily of known proteins in the available database (Table 2).Further, based on bioinformatics and comparative analysis, acomprehensive description was presented for the individualfamily of salivary gland proteins.

D7-related proteins

The D7 family of salivary proteins is expressed abundantly inthe blood feeding Diptera and is known to be distantly relatedto the odorant-binding protein super-family. In mosquitoes,two sub-families have been recognized, a short family havinga molecular mass of 15—30 kDa and a long family of 27—30kDa.21 Earlier studies analyzing mosquito salivary proteinsusing sodium dodecyl sulfate—polyacrylamide gel electro-phoresis (SDS-PAGE) and Edman degradation, have alsorevealed that the culicine and aedine mosquitoes have highexpression of the long form, whereas anophelines have a poorexpression of the long form and abundant expression of theshort form.22 Consistent with these studies, we observed ahigh abundance of ESTs (65 ESTs in total) belonging to the D7family of proteins. Out of these, only one contig (contig_32with one EST) was homologous to the long form of D7 protein,while the remaining 64 ESTs, assembled in six differentcontigs, belonged to the short form of D7 family proteins.10

Table 2 shows the seven contigs coding for proteins of the D7family. Three contigs showed 100% identity to the reportedD7 family proteins, while the other four contigs showed 94—97% identity and thus may represent either allelic splicingvariants or additional putative members of the D7 familyproteins in A. stephensi; these require further analysis.

Some of the D7-related proteins have been associatedwith binding and probably scavenging of biogenic amines suchas serotonin, histamine, and norepinephrine, which couldantagonize vasoconstrictor, platelet-aggregation, and pain-inducing properties.23 Recent analysis of the crystal structureof one member, D7r4, suggests that the D7 fold consists of anarrangement of eight alpha-helices that is stabilized by threedisulfide bonds, resulting in the formation of a ligand bindingpocket.24 Additionally, one short D7 protein from the mos-quito A. stephensi, named hamadarin, has been shown toprevent kallikrein activation by factor XIIa.25

Antithrombin-related proteins

Anophelin, a low molecular weight peptide, secreted in thesalivary glands of the anopheline mosquitoes, appears to playa role in the inhibition of thrombin-induced platelet aggre-gation during blood feeding. The first novel anophelin cDNA

Table 2 Molecular catalogue of salivary gland transcripts differentially induced in response to Plasmodium infection in Anopheles stephensi.

Contig Length EST/contig

Best match to NR E-value BLASTX seq. ID % Identity Species Best matchto Pfam

E-value Comments

Contig_1 396 3 ENSANGP00000015621 4.00E�40 gbjEAA05234.2 78 A. gambiaestr. PEST

Defensin_2 3.00E�11 Cysteine-richantibacterial peptide

Contig_2 538 44 Short form D7clu5 salivaryprotein

5.00E�79 gbjAAL16045.1 97 A. stephensi PBP_GOBP 2.00E�07 Odorant binding protein

Contig_3 403 12 Putative 19.1 kDa salivaryprotein SG3

2.00E�63 gbjAAO06833.1 97 A. stephensi AF-4 6.10E�02 Cytoskeleton andmorphogenesis

Contig_4 363 11 Putative salivary proteinD7B

2.00E�46 gbjAAO06831.1 100 A. stephensi PBP_GOBP 4.60E�01 Odorant binding protein

Contig_5 431 18 Salivary anti-thrombinanophelin

4.00E�32 gbjAAO06834.1 98 A. stephensi BAP 2.90E�03 Basal layer antifungalpeptide

Contig_6 248 30 Salivary anti-thrombinanophelin

4.00E�13 gbjAAO06834.1 97 A. stephensi TT_ORF1 1.20E�01 TT-viral ORF1/hepatitisetiology

Contig_7 256 1 Unknown Cnd1 6.40E�01 1,3-beta-glucan synthasecomponent

Contig_8 309 5 Putative salivary protein SG2B 6.00E�21 gbjAAO06827.1 98 A. stephensi Pyridoxal_deC 3.30E�01 Carboxylic acidmetabolism

Contig_9 481 7 Putative salivary protein SG2B 1.00E�63 gbjAAO06827.1 97 A. stephensi CD2 2.00E�03 Proline-rich sequencerecognition

Contig_10 435 4 ENSANGP00000011018 5.00E�39 gbjEAA05902.2 94 A. gambiaestr. PEST

Ribosomal_L9 5.00E�12 Structural constituentof ribosome

Contig_11 281 1 Unknown DAG1 1.00E+00 Dystrophin-associatedglycoprotein

Contig_12 415 23 gSG6 salivary gland proteinprecursor

3.00E�55 gbjAAO74842.1 100 A. stephensi NS - -

Contig_13 135 2 Unknown NS - -Contig_14 123 2 Unknown UPF0240 1.70E+00 Unknown functionContig_15 170 1 ENSANGP00000011995 1.30E�01 gbjEAA06859.2 94 A. gambiae

str. PESTCbiX 7.00E�01 Cobaltochelatase

activityContig_16 360 5 Putative 19.1 kDa salivary

protein SG36.00E�39 gbjAAO06833.1 98 A. stephensi Mucin 4.00E�09 Host—parasite

interaction/ invasionContig_17 371 1 Putative salivary protein hyp12 2.00E�44 gbjAAO06820.1 96 A. stephensi MCM 8.30E�02 DNA-dependent ATPasesContig_18 182 5 Anti-platelet aggregation protein 1.00E�06 dbjjBAE53440.1 100 A. stephensi UPF0120 1.00E�02 Ribosome nucleus exportContig_19 275 1 Putative salivary protein gSG7 4.00E�41 gbjAAO06828.1 96 A. stephensi DiHfolate_red 4.00E�03 Nucleotide biosynthesisContig_20 179 4 Short form D7 salivary protein 2.00E�28 gbjAAL16042.1 100 A. stephensi PBP_GOBP 2.00E�04 Odorant binding proteinContig_21 255 2 Short form D7 salivary protein 2.00E�22 gbjAAL16042.1 94 A. stephensi DUF457 1.00E�01 Predicted trans-membrane

proteinContig_22 101 2 Putative salivary protein D7B 2.00E�10 gbjAAO06831.1 100 A. stephensi PAD 3.30E+00 Protein arginine deiminaseContig_23 475 2 GE-rich salivary gland protein

precursor8.00E�79 gbjAAO74840.1 89 A. stephensi FAP 2.00E�10 Fibronectin attachment

proteinContig_24 196 3 Unknown WzyE 7.00E�02 Unknown function

Salivaryglan

dtran

scriptome,Anophelesste

phensi

639

Table

2(Continued

)

Contig

Length

EST

/co

ntig

Best

match

toNR

E-value

BLA

STXseq.ID

%Identity

Species

Best

match

toPfam

E-value

Comments

Contig_

25

192

1Anophelesga

mbiae

str.PEST

CRA_x

9P1GAUVCRS

8.00

E�34

NZ_A

AAB02

0692

59.

94A.ga

mbiae

str.PEST

Ribosomal_2

8S8.00

E�34

Structuralco

nstituent

Contig_

26

149

1Unkn

own

DUF1

484

2.80

E+0

0Unkn

ownfunction

Contig_

2717

81

Unkn

own

Phytoreo_P

ns

3.00

E�01

Unkn

ownfunction

Contig_

28

231

1Sh

ort

D7-4

saliva

ryprotein

precu

rsor

1.00

E�17

gbjAAO74

838.1

97A.stephensi

Chordopox_

G3

1.20

E�01

Chordopoxvirus

G3protein

Contig_

29

191

1ENSA

NGP00

000

011

995

1.00

E�15

gbjEAA06

859.2

97A.ga

mbiaestr.PEST

Mos_Cecropin

4.00

E�11

Mosquitoce

cropin

Contig_

30

361

15kD

asaliva

rypeptide

hyp

159.00

E�38

gbjAAO06

830.1

100

A.stephensi

DUF6

864.40

E�01

Unkn

ownfunction

Contig_

31

503

1Lo

ngform

D7clu2

saliva

ryprotein

3.00

E�88

gbjAAL1

6043

.196

A.stephensi

PBP_G

OBP

4.00

E�10

Odorantbinding

protein

Contig_

32

96

1Unkn

own

Unkn

ownfunction

EST,exp

ressedsequence

tags;NR,non-redundan

tdatab

ase;Pfam,protein

familiesdatab

ase.

640 R. Dixit et al.

was cloned and characterized from the salivary glands of themosquito Anopheles albimanus.26 One member of the ano-phelin family of proteins from A. stephensi has also beenreported on the NCBI database and has a similarity to A.gambiae putative protein sequences cE5 and F1.27 Table 2shows the identification of an additional two putative mem-bers (two contigs from 48 total transcripts) of the anophelinfamily of proteins with close similarities to the previouslyreported anophelin protein in A. stephensi. Interestingly, aPfam database search showed that one of the anophelincontigs (contig_5) codes a protein carrying the basal anti-fungal protein (BAP) domain, known to exhibit potent broad-range activity against a range of filamentous fungi in plants.28

Therefore, it will be interesting to verify whether anophelinpeptides from old world mosquitoes have the same affinity tothrombin as A. albimanus anophelin,29 or have some addi-tional role during parasite infection.

Putative immunity-related proteins

Although the exact biological processes implicated in themosquito response to the presence of parasite are still largelyunknown, an understanding of the mechanisms involved invector—parasite interactions and the means by which para-sites avoid the defensive response of the host is essential forthe design of new strategies to control vector-borne dis-eases.30 Recent studies have suggested that several putativeimmune responsive marker genes, including small antimicro-bial peptides (AMPs) like gambicin, defensins, and cecropins,presumably involved in immune surveillance, are activated inthe midgut and salivary glands of the mosquito A. gambiaeupon parasite infection.31,32 In the case ofA. stephensi, so farno any antimicrobial immune genes have been described.Although we have recently reported lysozyme activity in thesalivary glands,33 and a putative cDNA has also been identi-fied in the sialome of themosquito A. stephensi,10 whether ornot lysozyme plays a role in digestion and/or immunity is stilldisputed.34,35 In the present study, during BLASTX analysis,we recognized two novel ESTs encoding defensin- and cecro-pin-like peptides that appear to respond to the Plasmodiuminfection.

DefensinContig_1, with three identical cDNA sequences, showed 78%similarity to A. gambiae defensin. Subsequent analysis recog-nized it as full-length (396 bp) carrying a 291-bp long openreading frame (ORF), with 45 bp 50 UTR and 60 bp long 30 UTRsequences. Using SignalIP 3.0 online software and compara-tive analysis, we predicted three distinct regions: 24 residuesof signal peptide, 32 residues of prodefensin, and 40 residuesof mature defensin from the deduced amino acid sequence ofA. stephensi defensin; 96 residues in total, encoding a pro-tein with similarity to the antimicrobial peptide of themosquito defensin family. In A. gambiae, in vitro assessmentof the antiparasitic activities of defensin strongly suggeststheir important role in the immune defense against Plasmo-dium; however, the actual effect of defensin in antimalarialdefense is still disputed.36,37 A detailed sequence alignmentanalysis also showed significant diversity in the signal peptideand prodefensin region, where amino acid substitution,insertion, and deletions were seen amongst the differentisoforms of different species of insects in comparison to

Figure 1 Clustal alignment of known mosquito defensins. Highly conserved and similar residues are shown in dark black and gray,respectively. The pre- and pro-regions of defensin are shown by the double-headed arrows, while themature region is shown by a singledark line. The open circles mark the cysteine residues, while identical residues are marked with an asterisk. Species names include:Aara: Anopheles arabiensis; Agam: Anopheles gambiae; Aagy: Aedes aegypti; Aalb: Aedes albopictus.

Salivary gland transcriptome, Anopheles stephensi 641

mature functional defensin of these species (Figure 1).However, whether this variation affects the mosquito’ssusceptibility to the microbial infections remains to be elu-cidated.38,39

CecropinAnother two contigs (contig_15 and contig_29; singleton),encode proteins similar to the predicted A. gambiae cecropin(ENSANGP00000011995). Cecropins are a family of inducibleantibacterial peptides of 4 kDa, devoid of cysteine residues,constituted of two a-helices linked by a short hinge, with astrong basic N-terminal region and a long hydrophobicdomain in the C-terminal half. Cecropins integrate intothe bacterial cell membrane and cause lysis and cell deathby the loss of cations.40 Cecropins have been characterizedfrom diverse insects including mosquitoes.41,42 In A. gam-biae, the cecropin gene family are clustered on X chromo-somes.43 The expression of one member of this family, CEC1,is induced by Plasmodium as well as by bacterial infections,and over-expression of CEC1 in a transgenic mosquito line hasbeen observed to reduce the number of oocysts developing inthe midgut of A. gambiae.44 Multiple amino acid sequenceanalysis of one of the larger A. stephensi salivary cecropinsshowed a higher degree of conservation towards the C-ter-mini among all aligned sequences (Figure 2). However,whether this protein plays any role during Plasmodium inva-sion and development in the salivary gland needs furtherstudy.

In mosquito A. gambiae both defensin (DEF1) and cecropin(CEC2) display stronger responses at the onset of the salivarygland invasion.14 The A. stephensi salivary defensin andcecropin immune proteins that were enriched during thesubtraction protocol, which appear to be induced due toparasite infection to delimit the toxic effect of infectionstress, should be further investigated.

Figure 2 Clustal alignment of the knownmosquito cecropins. SpeciAe.alb: Aedes albopictus; Ae.agy: Aedes aegypti; A.gam: Anophele

GE-rich proteins

This class of protein family has been recognized as highlyimmunogenic in nature and belongs to one of the mostabundant classes of mRNA expressed in glands of differentmosquito species.45,46 The predicted protein molecularweight of this family varies from 30 to 35 kDa with a highcontent of glycine and glutamine amino acids, hence theterm GE-rich proteins. In A. stephensi two contigs have beendescribed that code for GE-rich proteins. In the presentstudy, BLASTX analysis of two independent contigs (contig_18with five ESTs and contig_23 with two ESTs) showed theclosest homology to the GE-rich family of salivary proteinsof anopheline mosquitoes. Multiple sequence alignment ofthe deduced amino acid sequence of the longest ESTshowed ahigher degree of conservation towards the C-terminal ascompared to the N-terminal and 89% shared identity withthe previously described A. stephensi ortholog (Figure 3).

Although, the specific function of the GE-rich proteins inblood feeding insects remains to be determined, their highquantity and acidic nature suggests some kind of role inbinding, neutralizing, and/or lubrication, rather than anycatalytic function associated with enzyme. A member of thisprotein family, named aegyptin, has recently been isolatedfrom the salivary glands of the mosquito Aedes aegypti,which specifically binds to the glycoproteins VI and preventscollagen interactions and thus inhibits platelet aggregationduring blood feeding.47

Interestingly, a Pfam database search analysis showedthat GE-rich family proteins may also contain the fibronec-tin-attachment protein (FAP) domain, rich in alanine andproline, and a fibronectin-binding motif that binds to fibro-nectin in the extracellular matrix (Table 2). These familymembers seem to be restricted to Mycobacterium spp andplay an important role in attachment and invasion of epithe-

es names include: A.s: Anopheles stephensi; C.pip: Culex pipiens;s gambiae.

Figure 3 Clustal alignment of mosquito salivary gland GE-rich family proteins showing a high degree of conservation towards the C-terminus. Species names include: A.s: Anopheles stephensi; Afun: Anopheles funestus; Aalb: Anopheles albimanus; Adirus: Anophelesdirus; Ae_al: Aedes albopictus.

642 R. Dixit et al.

lial cells in ruminants during the development of Johne’sdisease.48 Therefore, it is important to determine their rolein the cellular attachment and/or protein interaction withintracellular proteins during invasion of the salivary glands.

The SG family proteins

This family of proteins was first described in A. gambiae(termed SG or gSG proteins),49 and does not yield significantsimilarities to other proteins in the database, except amongits own members. The biological function of these proteins isstill completely unknown; however their distribution isrestricted to anopheline mosquito species and they areabundantly expressed in the female salivary glands, indicat-ing that they have some distinctive properties to carry outsome highly specialized function in the salivary glands; thisfunction needs to be elucidated.

SG2 family proteinsIn A. stephensi, two genes code for glycine SG2a- and proline-rich proteins SG2b-related proteins. We identified two con-tigs (contig_8 with five ESTs and contig_9 with seven ESTs)that share 94—96% identity to the previously known salivarySG2 proteins.10 Their glycine-rich composition is reminiscentof some antimicrobial peptides.50

An alignment of 10 sequences shows that A. stephensi SG2family proteins share an overall 61% identity, and A. gambiaeSG2a acquires an additional 30-aa insertion which separatesthe members of A. gambiae into two independent clusters inthe phylogram (Figure 4). A. stephensi family members showthe closest homology to the SG2b orthologs, suggesting that

only A. gambiae have two class members in their salivaryglands.

SG6 family proteinsThe gSG6 peptide was first described in A. gambiae49 andfound to be a unique protein coding sequence for a maturepeptide of approximately 10 kDa. The spacing of the 10-cysteine residue is unique to this family, as a BLAST searchwith the NR database shows homology to its own membersonly and it is known to be expressed specifically in the adultfemale salivary glands.12 One contig, contig_12 (with 23 ESTs)yielded 100% identity to the previously known gSG6 familyproteins. Although the function of this peptide is not known,high representation of this class of ESTs even after an effec-tive subtraction and Pfam search analysis suggests that it mayplay an important role in connection to carbohydrate meta-bolism and parasite development.

SG7 family proteinsLastly, a 275-bp long contig_19 (singleton) showed highersimilarity to that of the previously known salivary glandgSG7 family proteins. In A. gambiae two genes coding for thisunique salivary gland family of proteins are known. One mem-ber of this family has also been reported inA. stephensi, whichshares 96% identity to the contig_19. The biological function ofthis familyofproteins is still unknown.Recently anovelproteinnamed anophensin has been described in the salivary gland ofA. stephensi, which appears to play an important role in theattenuation of the host’s acute inflammatory response tomosquito bites. Although the function of this protein is verysimilar to thatof thepreviouslydescribedproteinhamadarin (a

Figure 4 Clustal alignment and phylogenetic analysis of the gSG2b family of proteins of anopheline mosquitoes. This family ofproteins is restricted to anopheline mosquitoes only (see text). A.s: Anopheles stephensi; A.f: Anopheles funestus; A.g: Anophelesgambiae.

Salivary gland transcriptome, Anopheles stephensi 643

member of the D7 family) involved in inhibiting bradykininrelease during blood feeding,25 anophensin lacks sequencehomologytohamadarin,and interestingly it sharesasignificantsimilarity to that of the gSG7 family of proteins.51

Together with previous reports on the multiple alignmentand phylogenetic analysis of gSG7/anophensin protein familymembers (Figure 5), this suggests that this family of proteins isabundantly enriched in the salivary glands of anophelinemosquitoes, and evolved to preserve their multi-dimensionalfunctions for the smooth and efficient uptake of a bloodmeal.51

19.1 kDa/mucin-related proteins

This family of proteins is characterized by a high degree ofamino acid conservation towards both N and C termini, ascompared to the central region, and contains a large numberof threonine- and serine-rich residues with a repeat ofTTTEEA in its C-terminal region, similar to the syndecanPfam domain of vertebrate mucins.52 Based on BLASTX ana-lysis, two contigs (contig_3 and contig_16) clustered from 17ESTs, showed the greatest similarities to the previouslyknown 19.1 kDa putative SG3 salivary protein of mosquito

Figure 5 Clustal alignment and phylogenetic analysis of the gSG7 family of proteins of anopheline mosquitoes. This family of proteinsis restricted to anopheline mosquitoes only (see text). A.s: Anopheles stephensi; A.f: Anopheles funestus; A.g: Anopheles gambiae.

644 R. Dixit et al.

A. stephensi (AAO06833), a member of the mucin family ofproteins (Table 2). Although, the function of this family is notknown, indirect evidence suggests that these may encodeglycoproteins, may act as lubricants of salivary canals, andmay be involved in the interaction with and invasion ofmammalian host cells.53

Hypothetical salivary proteins

BLASTX analysis of two contigs (contig_17 and contig_30),both singleton cDNA sequences, showed closest similaritiesto the previously described salivary gland hypothetical pro-teins hyp12 and hyp15, respectively. These families of pro-teins are unique to the anophelines and do not yieldsignificant matches to any other known proteins in the NRdatabase.49 The function of these proteins is completely

Figure 6 Clustal alignment of insect rpL9 protein. A unique extensiolipid attachment domains has been underlined with dots (see text).Drosophila yakuba; D.pseu: Drosophila pseudoobscura; A.mel: Apis mAe_agy: Aedes aegypti; C.sono: Culicoides sonorensis.

unknown, however identification of the additional contigsof hyp15 represent either new and/or allelic variants of thepreviously described hypothetical proteins in mosquito A.stephensi (Table 2).

Unique ribosomal protein family

One contig (contig_10 with four ESTs) yielded a 94% similarityto that of a predicted protein from the A. gambiae genome(ENSANGP00000011018) and another insect ribosomal pro-tein L9 (rpL9), while another singleton contig (contig_25)encoding a ribosomal protein showed a 74% identity to therp28S family protein of mosquito A. gambiae. The deducedamino acid sequence alignment analysis of both the proteinswith other known insect respective ribosomal proteins, indi-cated a high degree of conservation throughout. Interest-

n of 27 amino acids related to prokaryotic membrane lipoproteinSpecies names include: D.mel: Drosophila melanogaster; D.yak:ellifera; M.cori:Meladema coriacea; A.gam: Anopheles gambiae;

Salivary gland transcriptome, Anopheles stephensi 645

ingly, a unique extension of 27 basic amino acids containing aprokaryotic membrane lipoprotein lipid attachment-likedomain towards the C-terminal, was observed in salivaryrpL9 protein (Figure 6). In prokaryotes, membrane lipopro-teins are synthesized as a precursor signal peptide, which iscleaved by a specific lipoprotein signal peptidase (signalpeptidase II). The peptidase recognizes a conservedsequence and cleaves at a cysteine residue to which aglyceride-fatty acid lipid is attached.54

Although the ribosomal proteins belong to constitutivelyexpressed gene families, selective enrichment of rpL9-spe-cific protein during the subtraction protocol suggests thatthis may have some additional role during infection. Forexample, in vertebrates it has been shown that a pointmutation in ribosomal protein rpL9 gives rise to immunodo-minant CD4+ T cell-recognized tumor-specific antigens.55

Additionally, a homozygous deletion of rpL9 is known to belethal in Drosophila.56 Presumably, this membrane lipopro-tein domain may play a role in electrostatic binding of theparasite membrane during infection; however furtherdetailed analysis should be undertaken to confirm thesepresumptions.

Unknown family proteins

Eight contigs representing 12 ESTs that code for proteins ofunknown function did not yield any significant matches toknown proteins, even when the low complexity filter of theBLAST program was turned on (Table 2). In silico BLASTXanalysis with the A. gambiae genome also did not yield anysignificant homology; these represent a novel class of salivarygland family proteins that are specific to mosquito A. ste-phensi.

Conclusions

After genome sequence availability for the mosquito A.gambiae, large-scale EST projects are underway to identifypotential genes expressed in the different mosquito tissues.Although, these EST projects are descriptive in nature, theyalso generate hypotheses on the evolution of blood feeding ingeneral and in the discovery of novel genes that may play animportant role in malaria transmission. Here we havedescribed the salivary gland transcriptome of Plasmodium-infected A. stephensi and compared it with those of othermosquitoes and insects. Results indicate that salivary pro-teins are under intense selection and can be used as robustmarkers for closely related species. Roles are proposed forsome of the transcripts in relation to blood feeding andPlasmodium development. Together with ongoing highthroughput proteomic and current genomic projects, thisshould provide further insight into the understanding ofthe molecular complexity of mosquito salivary glands.

Acknowledgments

This work was financially supported by the Department ofBiotechnology, New Delhi, Government of India. The authorsare grateful to Sanjeev Kumar and J.M.C. Ribeiro, LMVR,NIAID, National Institute of Health, USA for their helpfulsuggestions and comments on the manuscript. We are also

thankful to Sarang Satoor for technical support duringsequencing of the ESTs.

Conflict of interest: No conflict of interest to declare.

References

1. Fauci AS. Infectious diseases: considerations for the 21st cen-tury. Clin Infect Dis 2001;32:675—85.

2. Ghosh A, Edwards MJ, Jacobs-Lorena M. The journey of malariaparasite in the mosquito: hopes for the new century. ParasitolToday 2000;16:196—201.

3. Ito J, Ghosh A, Moreira LA, Wimmer EA, Jacobs-Lorena M.Transgenic anopheline mosquitoes impaired in transmission ofa malaria parasite. Nature 2002;417:452—5.

4. Moreira LA, Wang J, Collins FH, Jacobs-Lorena M. Fitness ofanopheline mosquitoes expressing transgenes that inhibit Plas-modium development. Genetics 2004;166:1337—41.

5. Ghosh AK, Ribolla PE, Jacobs-Lorena M. Targeting Plasmodiumligands on mosquito salivary glands and midgut with a phagedisplay peptide library. Proc Natl Acad Sci U S A 2004;98:13278—81.

6. James AA. Blocking malaria parasite invasion of mosquito sali-vary glands. J Exp Biol 2003;206:3817—21.

7. Paul RE, Doerig C, Brey PT. Erythropoiesis andmolecularmechan-isms for sexual determination in malaria parasites. IUBMB Life2000;49:245—8.

8. Belkaid Y, Valenzuela JG, Kamhawi S, Rowton E, Sacks DL, RibeiroJM. Delayed-type hypersensitivity to Phlebotomus papatasi sandfly bite: an adaptive response induced by the fly? Proc Natl AcadSci U S A 2000;6:6704—9.

9. Francischetti IM, Valenzuela JG, Pham VM, Garfield MK, RibeiroJM. Toward a catalog for the transcripts and proteins (sialome)from the salivary gland of the malaria vector Anopheles gam-biae. J Exp Biol 2002;205:2429—31.

10. Valenzuela JG, Francischetti IM, Pham VM, Garfield MK, RibeiroJM. Exploring the salivary gland transcriptome and proteome ofthe A. stephensi mosquito. Insect Biochem Mol Biol 2003;33:717—32.

11. Calvo E, Andersen J, Francischetti IM, Capurro ML, Bianchi AG,James AA, et al. The transcriptome of adult female Anophelesdarlingi salivary glands. Insect Mol Biol 2004;13:73—88.

12. Arca B, Lombardo F, Valenzuela JG, Francischetti IM, Marinotti O,Colluzi M, et al. An updated catalogue of salivary gland tran-scripts in the adult female mosquito, Anopheles gambiae. J ExpBiol 2005;208:3971—86.

13. Calvo E, Tokumasu F, Marinotti O, Villeval JL, Ribeiro JM,Francischetti IM, Aegyptin. a novel mosquito salivary glandprotein, specifically binds to collagen and prevents itsinteraction with platelet glycoprotein VI, integrin alpha2-beta1, and von Willebrand factor. J Biol Chem 2007;282:26928—38.

14. Rosinski-Chupin I, Briolay J, Brouilly P, Perrot S, Gomez SM,Chertemps T, et al. SAGE analysis of mosquito salivary glandtranscriptomes during Plasmodium invasion. Cell Microbiol2007;9:708—24.

15. Choumet V, Carmi-Leroy A, Laurent C, Lenormand P, RousselleJC, Namane A, et al. The salivary glands and saliva of Anophelesgambiae as an essential step in the Plasmodium life cycle: aglobal proteomic study. Proteomics 2007;7:3384—94.

16. Hati AK. Urban malaria vector biology. Ind J Med Res1997;106:149—63.

17. Abraham EG, Islam S, Srinivasan P, Ghosh AK, Valenzuela JG,Ribeiro JM, et al. Analysis of the Plasmodium and Anophelestranscriptional repertoire during ookinete development andmidgut invasion. J Biol Chem 2004;279:5573—80.

18. Srinivasan P, Abraham EG, Ghosh AK, Valenzuela J, Ribeiro JM,Dimopoulos G, et al. Analysis of the Plasmodium and Anopheles

646 R. Dixit et al.

transcriptomes during oocyst differentiation. J Biol Chem2004;279:5581—7.

19. Xu X, Dong Y, Abraham EG, Kocan A, Srinivasan P, Ghosh AK, et al.Transcriptome analysis of Anopheles stephensi—Plasmodiumberghei interactions. Mol Biochem Parasitol 2005;142:76—87.

20. Dixit R, Dixit S, Roy U, Shouche YS, Gakhar S. Partial genomicorganization of ribosomal protein S7 gene frommalaria vector A.stephensi. Insect Sci 2007;14:101—6.

21. Valenzuela JG, Pham VM, Garfield MK, Francischetti IM, RibeiroJM. Toward a description of the sialome of the adult femalemosquito Aedes aegypti. Insect Biochem Mol Biol 2002;32:1101—22.

22. Ribeiro JM, Charlab R, Pham VM, Garfield M, Valenzuela JG. Aninsight into the salivary transcriptome and proteome of the adultfemale mosquito Culex pipiens quinquefasciatus. Insect Bio-chem Mol Biol 2004;34:543—63.

23. Calvo E, Dao A, Pham VM, Ribeiro JM. An insight into the sialomeof Anopheles funestus reveals an emerging pattern in anophelinesalivary protein families. Insect Biochem Mol Biol 2007;37:164—75.

24. Mans BJ, Calvo E, Ribeiro JM, Andersen JF. The crystal structureof D7r4, a salivary biogenic amine-binding protein from themalaria mosquito Anopheles gambiae. J Biol Chem 2007;282:36626—33.

25. Isawa M, Yuda M, Orito Y, Chinzei Y. A mosquito salivary proteininhibits activation of plasma contact system by binding to factorXII and high molecular weight kininogen. J Biol Chem 2002;277:27651—8.

26. Valenzuela JG, Francischetti IM, Ribeiro JM. Purification, clon-ing, and synthesis of a novel salivary anti-thrombin from themosquito Anopheles albimanus. Biochem 1999;38:11209—15.

27. Arca B, Lombardo F, Capurro LM, Torre A, Spanos L, DimopoulosG, et al. Trapping cDNAs encoding secreted proteins from thesalivary glands of the malaria vector Anopheles gambiae. ProcNatl Acad Sci U S A 1999;96:1516—21.

28. Serna A, Maitz M, O’Connell T, Santandrea G, Thevissen K,Tienens K, et al. Maize endosperm secretes a novel antifungalprotein into adjacent maternal tissue. Plant J 2001;25:687—98.

29. Francischetti IM, Valenzuela JG, Ribeiro JM. Anophelin: kineticsand mechanism of thrombin inhibition. Biochem 1999;38:16678—85.

30. Collins FH, James AA. Genetic modification of mosquitoes. SciMed 1996;3:52—61.

31. Dimopoulos G, Seeley D, Wolf A, Kafatos FC. Malaria infection ofthe mosquito Anopheles gambiae activates immune-responsivegenes during critical transition stages of the parasite life cycle.EMBO J 1998;17:6115—23.

32. Oduol F, Xu J, Niare O, Natarajanet R, Vernick KD. Genesidentified by an expression screen of the vector mosquito Ano-pheles gambiae display differential molecular immune responseto malaria parasites and bacteria. Proc Natl Acad Sci U S A2000;97:11397—402.

33. Dixit R, Dixit S, Gakhar SK. Identification and characterization ofa new putative c-type lysozyme from malaria vector A. ste-phensi. Ind J Biochem Biophys 2006;43:15—9.

34. Fujita AI. Lysozymes in insects: what role do they play in nitrogenmetabolism? Physiol Entomol 2004;29:305—10.

35. Li B, Calvo E, Marinotti O, James AA, Paskewitz SM. Character-ization of the c-type lysozyme gene family in Anopheles gam-biae. Gene 2005;360:131—9.

36. Shahabuddin M, Fields I, Bulet P, Hoffmann JA, Miller LH. Plas-modium gallinaceum: differential killing of some mosquitostages of the parasite by insect defensin. Exp Parasitol1998;89:103—12.

37. Blandin S, Moita LF, Kocher T, Wilm M, Kafatos FC, Levashina EA.Reverse genetics in the mosquito Anopheles gambiae: targeteddisruption of the defensin gene. EMBO Rep 2002;3:852—6.

38. Simard F, Licht M, Besansky NJ, Lehmann T. Polymorphism at thedefensin gene in the Anopheles gambiae complex: testingdifferent selection hypotheses. Infect Genet Evol 2007;7:285—92.

39. Dixit R, Sharma A, Patole MS, Shouche YS. Molecular and phy-logenetic analysis of a novel salivary defensin cDNA frommalariavector A. stephensi. Acta Trop 2008;106:75—9.

40. Durell SR, Raghunathan G, Guy HR. Modeling the ion channelstructure of cecropin. Biophys J 1992;3:1623—31.

41. Vizioli J, Bulet P, Charlet M, Lowenberger C, Blass C, Muller HM,et al. Cloning and analysis of a cecropin gene from the malariavector mosquito, Anopheles gambiae. Insect Mol Biol 2000;9:75—84.

42. Sun D, Fallon AM. Characterization of genomic DNA encodingcecropins from an Aedes albopictus mosquito cell line. InsectMol Biol 2002;11:21—30.

43. Zheng XL, Zheng AL. Genomic organization and regulation ofthree cecropin genes in Anopheles gambiae. Insect Mol Biol2002;11:517—25.

44. Kim W, Koo H, Richman AM, Seeley D, Vizioli J, Klocko AD, et al.Ectopic expression of a cecropin transgene in the human malariavector mosquito Anopheles gambiae (Diptera: Culicidae):effects on susceptibility to Plasmodium. J Med Entomol2004;41:447—55.

45. Simons FE, Peng Z. Mosquito allergy: recombinant mosquitosalivary antigens for new diagnostic tests. Int Arch Aller Immu-nol 2001;124:403—5.

46. Jariyapan N, Choochote W, Jitpakdi A, Harnnoi T, Siriyasatein P,Wilkinson MC, et al. A glycine- and glutamate-rich protein isfemale salivary gland-specific and abundant in the malariavector Anopheles dirus B (Diptera: Culicidae). J Med Entomol2006;43:867—74.

47. Calvo E, Mans BJ, Andersen JF, Ribeiro JM. Function and evolu-tion of a mosquito salivary protein family. J Biol Chem2006;281:1935—42.

48. Secott TE, Lin TL, Wu CC. Fibronectin attachment protein isnecessary for efficient attachment and invasion of epithelialcells by Mycobacterium avium subsp. paratuberculosis. InfectImmun 2002;70:2670—5.

49. Lanfrancotti A, Lombardo F, Santolamazza F, Veneri M, Cas-trignano T, Coluzzi M, et al. Novel cDNAs encoding salivaryproteins from the malaria vector Anopheles gambiae. FEBS Lett2002;517:67—71.

50. Otvos Jr L. Antibacterial peptides isolated from insects. J Pep-tide Sci 2000;6:497—511.

51. Isawa H, Orito Y, Iwanaga S, Jingushi N, Morita A, Chinzei Y, YudaM. Identification and characterization of a new kallikrein—kininsystem inhibitor from the salivary glands of the malaria vectormosquito Anopheles stephensi. Insect Biochem Mol Biol2007;37:466—77.

52. Hansen JE, Lund O, Tolstrup N, Googley AA, Williams KL, BrunakS. NetOglyc: prediction ofmucin type O-glycosylation sites basedon sequence context and surface accessibility. Glycoconj J1998;15:115—30.

53. Ropert C, Ferreira LR, Campos MA, Procopio DO, Travassos LR,Ferguson MA, et al. Macrophage signaling by glycosylphos-phatidylinositol-anchoredmucin-like glycoproteins derived fromTrypanosoma cruzi trypomastigotes. Microbes Infect 2002;4:1015—25.

54. Hayashi S, Wu HC. Lipoproteins in bacteria. J Bioenerg Bio-membr 1990;22:451—71.

55. Monach PA, Meredith SC, Siegel CT, Scheriber H. A unique tumorantigen produced by a single amino acid substitution. Immunity1990;2:45—59.

56. Schmidt A, Hollmann M, Schafer U. A newly identified minutelocus, M(2)32D, encodes the ribosomal protein L9 in Drosophilamelanogaster. Mol Gen Genet 1996;251:381—7.