And Larval-Specific Gene Expression in the Horn Fly (Diptera

MOLECULAR BIOLOGY/GENOMICS Microarray Analysis of Female- and Larval-Speciﬁc Gene Expression in the Horn Fly (Diptera: Muscidae)

FELIX. D. GUERRERO,1,2 SCOT. E. DOWD,3 YAN SUN,3 LEONEL SALDIVAR,1 GRAHAM B. WILEY,4 SIMONE L. MACMIL,4 FARES NAJAR,4 4 5 BRUCE A. ROE, AND LANE D. FOIL

J. Med. Entomol. 46(2): 257Ð270 (2009) ABSTRACT The horn ßy, Haematobia irritans L., is an obligate blood-feeding parasite of cattle, and control of this pest is a continuing problem because the ßy is becoming resistant to pesticides. Dominant conditional lethal gene systems are being studied as population control technologies against agricultural pests. One of the components of these systems is a female-speciÞc gene promoter that drives expression of a lethality-inducing gene. To identify candidate genes to supply this promoter, microarrays were designed from a horn ßy expressed sequence tag (EST) database and probed to identify female-speciÞc and larval-speciÞc gene expression. Analysis of dye swap experiments found 432 and 417 transcripts whose expression levels were higher or lower in adult female ßies, respectively, compared with adult male ßies. Additionally, 419 and 871 transcripts were identiÞed whose expression levels were higher or lower in Þrst-instar larvae compared with adult ßies, respectively. Three transcripts were expressed more highly in adult females ßies compared with adult males and also higher in the Þrst-instar larval lifestage compared with adult ßies. One of these transcripts, a putative nanos ortholog, has a high female-to-male expression ratio, a moderate expression level in Þrst-instar larvae, and has been well characterized in Drosophila. melanogaster (Meigen). In conclusion, we used microarray technology, veriÞed by reverse transcriptase-polymerase chain reaction and massively parallel pyrosequencing, to study life stageÐ and sex-speciÞc gene expression in the horn ßy and identiÞed three gene candidates for detailed evaluation as a gene promoter source for the development of a female-speciÞc conditional lethality system.

KEY WORDS Haematobia irritans, microarrays, gene expression analysis, pyrosequencing

The horn ßy, Haematobia irritans L., is an obligate and Brazil, respectively (Kunz et al. 1991, Grisi et al. blood-feeding parasite of cattle, and control of this 2002). The primary means of controlling the horn ßy pest is a continuing problem for cattle producers in the is through the application of insecticides, primarily United States and other parts of the world. Feeding from the pyrethroid and organophosphate classes. several times per day, infestations of several thousand However, populations of horn ßies that exhibit sig- ßies per animal have been reported (Bruce 1964). niÞcant resistance to these pesticides are becoming Severe infestations interfere with normal feeding ac- common (Kunz and Schmidt 1985, Kunz et al. 1995, tivity, and economic losses to producers result from Guerrero and Barros 2006), and novel control meth- the reduced weight gain on cattle being prepared for odologies would be an important contribution to the market during the ßy season. Total economic losses cattle industry. attributable to the horn ßy have been estimated at An interesting approach to insect population con- $876 and $150 million annually in the United States trol using a dominant, repressible lethal gene system in Drosophila melanogaster (Meigen)was reported by Thomas et al. (2000). This system requires a sex-spe- Mention of trade names or commercial products in this publication is solely for the purpose of providing speciÞc information and does not ciÞc gene promoter to drive the expression of a re- imply recommendation or endorsement by the U.S. Department of pressible transcription factor. The transcription factor Agriculture. controls the expression of a toxic gene product. This 1 USDAÐARS Knipling-Bushland U.S. Livestock Insects Research control approach and variants on the general theme Laboratory, 2700 Fredericksburg Rd., Kerrville, TX 78028. 2 Corresponding author, e-mail: [email protected]. are also being studied in the agricultural pests Ceratitis 3 USDAÐARS Livestock Issues Research Unit, 1604 E. FM 1294, capitata (Wiedemann) (Fu et al. 2007) and Lucilia Lubbock, TX 79403. cuprina (Wiedemann) (Scott et al. 2004). With the 4 Department of Chemistry and Biochemistry, University of Okla- problems insecticide-resistant horn ßy populations homa, 620 Parrington Oval, Norman, OK 73019. 5 Louisiana State University, Department of Entomology, 404 Life pose to current control programs, we initiated a Sciences Building, Baton Rouge, LA 70803. project aimed at designing a dominant, repressible 258 JOURNAL OF MEDICAL ENTOMOLOGY Vol. 46, no. 2 female-speciÞc lethal gene system tailored for the (NanoDrop Technologies, Wilmington, DE), and horn ßy. Rather than use genetic components from the quality was conÞrmed by electrophoresis. D. melanogaster laboratory strainÐderived system Microarray Design. Using an expressed sequence (Thomas et al. 2000), we sought to identify speciÞc tag (EST) collection described previously (Guerrero orthologues and gene promoters from a natural pop- et al. 2008), a unigene set for horn ßy was generated ulation of the horn ßy. Initially, it was necessary to by assembly of the ESTs at 96% similarity. This unigene acquire an adequate database of expressed gene se- set was used for BlastX search against SwissProt/ quences, because there were very few GenBank en- Tremble to identify those genes that were in forward tries for the horn ßy (Guerrero et al. 2008). A critical or reverse mRNA orientation based on the largest component of this system is the sex-speciÞc gene pro- ORF with e-value top hits Ͻ1 ϫ 10Ϫ112. Those unigene moter that must drive the expression of the lethality- set members with adequate information to determine inducing gene product. An obvious choice for a pro- the mRNA orientation were included into the array moter source would be the yolk protein genes, of design after putting the sequencing into mRNA sense which there are at least three in Musca domestica L. orientation. Unigene sequences without adequate in- (White and Bownes 1997). Sequences with signiÞcant formation were included in the array design in both BlastX similarity to yolk proteins 2 and 3 have been orientations. A total of 6,505 horn ßy tentative con- reported from the horn ßy (Guerrero et al. 2004). sensus (hfTC) sequences were entered into a custom However, at least one of the M. domestica yolk protein script and used to design 60-mer oligonucleotide genes, yp3, has detectable expression in male ßies probes with two probes sought for each hfTC. A total (White and Bownes 1997). Thus, microarrays were of 12,806 probes were identiÞed. These probes were designed from the recently reported horn ßy se- uploaded to E-array 4.5 in an 8 by 15K high-density quence database (Guerrero et al. 2008) and probed to format (Agilent Technologies, Santa Clara, CA). A characterize female-speciÞc gene expression and total of 536 control genes (including positive and neg- identify several candidate genes that could supply ative control probes) were also included in the design promoters with tight female-speciÞc expression. An of the array, and the Agilent two color spike-in kit was additional desirable trait we sought in the gene pro- used for technical normalizations. moter activity proÞle for the repressible female-spe- Microarray Protocol. For each sample, 10 ␮g of total ciÞc lethal gene system was expression at an early RNA and a 50-fold dilution of the Agilent two color stage in the life cycle. Ideally, lethality would be in- spike-in control RNA kit (Agilent Technologies) were duced in the embryonic or early larval stage to prevent labeled with either CyDye3-dCTP or CyDye5-dCTP devoting resources to mass rear individuals destined to (Amersham Biosciences, Piscataway, NJ) using the become females only to be killed later in the process. LabelStar kit (Qiagen, Valencia, CA) and oligo-dT and To identify early gene expression, the microarrays random nonamers (Sigma-Aldrich, St. Louis, MO). In were probed with labeled cDNA from Þrst-instar horn short, RNA from female horn ßies was labeled with ßy larvae and larval expression data compared with CyDye3-dCTP and hybridized against CyDye5- that from labeled cDNA from adult ßies. dCTPÐlabeled RNA from male horn ßies. As a second replicate to control for gene-speciÞc dye bias, the female RNA was labeled with CyDye5-dCTP and hybridized against CyDye3-dCTPÐlabeled RNA from Materials and Methods male ßies. Similarly, RNA from Þrst-instar larvae was Insect Material. Adult ßies were collected on a sin- labeled and hybridized against a labeled 50:50 mix of gle collection date from pastured cattle at the Loui- RNA from adult male and female ßies using the dye siana State University Agricultural Center, St. Gabriel swap design. All labeling, hybridization, and washing Research Station (St. Gabriel, LA) by aerial nets and procedures were performed according to the respec- transferred to Erlenmeyer ßasks at 30ЊC and total tive manufacturersÕ protocols. Labeled cDNA was hy- darkness for 1.5 h to facilitate egg collection (Lysyk bridized to the Agilent microarrays using the Gene 1991). Collected eggs were divided into two samples, Expression Hybridization kit (Agilent Technologies) and adult ßies were frozen at Ϫ80ЊC and sexed while following the manufacturerÕs protocols. Arrays were on dry ice. One egg sample was frozen at Ϫ80ЊC, and washed with Gene Expression Wash Buffer kit (Agi- the second egg sample was transferred to moist Þlter lent Technologies). A total of four arrays (two inde- paper in petri plates and maintained at room temper- pendent RNA extractions per ßy tissue type and one ature to induce egg hatch. The following day, the dye swap per extraction) were used in the male to wandering Þrst-instar larvae were collected and fro- female comparison, and four arrays in the larval to zen at Ϫ80ЊC. adult comparison were used to obtain genes that were RNA Extraction. All procedures were performed consistently regulated while limiting false discovery according to respective manufacturer protocols. RNA rate (FDR) to Benjamini and Hochberg 1995). was extracted from Ϸ1 g of adult ßies and 0.1 g larvae Microarray Analysis. Microarray images were cap- using the TotallyRNA Kit (Ambion, Austin, TX). DNA tured using a Genepix 4000B (Molecular Devices, was removed using the Turbo DNA Free Kit (Am- Union City, CA) laser scanner, and images were pro- bion). Two RNA preparations were made from inde- cessed using Genepix 6.0 software (Molecular De- pendent samplings of each ßy material sample. RNA vices). Microarray data analyses were performed us- was quantiÞed using a NanoDrop ND-1000 device ing Acuity 4.0 software. Slides were normalized using March 2009 GUERRERO ET AL.: MICROARRAY ANALYSIS OF HORN FLY GENE EXPRESSION 259 standard ratio-based methods. Data were analyzed entailed shearing the DNA through nebulization and based on log ratio (635/532) values. Genes were in- subsequent end repair, as described (Roe 2004), fol- cluded in the Þnal dataset that exhibit signiÞcance lowed by ligation of adapter sequences and a second based on a very conservative FDR Ͻ2.5% (Tsai et al. round of end repair to yield a blunt ended DNA library 2003) and at least 2.0-fold comparative regulation. The that is quantiÞed and diluted before ampliÞcation biological materials collected from these Þeld popu- through emPCR (Margulies et al. 2005). After emPCR lations of horn ßies were limited in quantity because enrichment, the DNAs were loaded onto a 454 Roche they were used as directly comparable matched sam- GS FLX for massively parallel pyrosequencing. The ples for cDNA library synthesis (Guerrero et al. 2008), resulting sequence data from the male and female 454 pyrosequencing, and microarray analysis. Limited samples was separately analyzed by BlastN (Altschul quantities of RNA were available for microarray ex- et al. 1990) to tally the occurrences of sequence frag- perimental replicates. Thus, in our statistical analysis, ments possessing identity to a member of the horn ßy we treated each of the four arrays, including the dye Unigene set (Guerrero et al. 2008). As seen in Table swaps, as biological replicates, a designation not sta- 1, fold-change was determined by dividing the tally tistically accurate because we had only two indepen- result from the test dataset by the tally from the other dent RNA extractions for each sample. As such, we dataset. If a 0 was encountered in the control dataset veriÞed the microarray gene expression results using (denominator of the equation fold-change ϭ test/ quantitative polymerase chain reaction (PCR) and control), a one was substituted, and the symbol Ͼ 454 pyrosequencing for a total of three independent inserted in the table column. methodologies quantifying expression in our studies of Bioinformatics. The differentially regulated genes sex-speciÞc and larval-speciÞc gene expression. from the microarray study were mapped to functional Quantitative PCR. Quantitative PCR results for dif- classiÞcations schemes such as Gene Ontology terms ferentially expressed transcripts normalized to glyc- (Ashburner et al. 2000), Kyoto Encyclopedia of Genes eraldehyde 3-phosphate dehydrogenase (GAPDH) and Genomes (KEGG) pathways (Ogata et al. 1999), were used as one methodology for validation of the and Swiss Prot Protein Keywords (Bairoch et al. 2005), microarray results. Measurements of relative tran- through the use of the High-Throughput Gene On- script amounts were performed by quantitative re- tology and Functional Annotation Toolkit (HTGO- verse transcription-PCR (qRT-PCR) with the Quanti- FAT; http://liru.ars.usda.gov), which is a functional Tect SYBR Green RT-PCR kit (Qiagen) according to annotation engine based on WND-BLAST (Dowd et the manufacturerÕs instructions. Cycling conditions al. 2005) and the Database for Annotation, Visualiza- were based on the standard universal settings recom- tion and Integrated Discovery (DAVID) (Dennis et mended for the 7500 Sequence Detection System (PE al. 2003). To accomplish this, regulated genes (FDR Ͻ Applied Biosystems, Foster City, CA). One quantita- 5%) were mapped to UniProt accession numbers and tive PCR validation run was performed for each se- Gene Index numbers using the built in functions of lected differentially expressed transcript. Three rep- HTGOFAT. These UniProt accessions were entered licates for each transcript were used to allow into DAVID to evaluate functional category enrich- correlation analysis for comparison to microarray ex- ment statistics with Sus scrofa as the background ge- pression data. SpeciÞc primer pairs were designed nome. using Applied Biosystems PrimerExpress Software and Statistical Analyses of Microarray Data. Statistics standard settings (PE Applied Biosystems). The algorithms built into Acuity 4.0 were used for analyses GAPDH gene was analyzed for both control and treat- related to microarrays. Algorithms in the ABI Prism ment samples to normalize the qRT-PCR results of 7500 Sequence Detection system software (PE Ap- selected genes. The reactions were performed on an plied Biosystems) were used for all calculations re- ABI Prism 7500 Sequence Detection system (PE Ap- lated to quantitative RT-PCR. Algorithms in the plied Biosystems). Six differentially expressed tran- DAVID (Dennis et al. 2003) functional annotation scripts were selected for validation by quantitative tool were used to evaluate categorization statistics. PCR. The difference (fold) in the initial concentration Correlation analyses were performed using multivar- of each transcript with respect to the control samples iate analyses functions of JMP 6.0 (SAS Institute). The was calculated according to the comparative Ct microarray datasets have been submitted to the GEO method using the built in functions of the 7500 system database (www.ncbi.nlm.nih.gov/geo/; GEO Plat- Sequence Detection Software version 1.3 (Applied form GPL6818, Series GSE11828). Biosystems). 454 Pyrosequencing. PolyA RNA was isolated from Results and Discussion two 400-␮g aliquots of total RNA from adult males and adult females using the MicroPolyA Purist Kit (Am- Sex-speciﬁc Gene Expression. RNA from female bion). Five micrograms of polyA RNA was used to horn ßies was labeled and hybridized against labeled synthesize blunt-ended double-stranded cDNA using RNA from male horn ßies in a dye swap design using the Just cDNA Double-Stranded cDNA Synthesis Kit a total of four arrays incorporating the two RNA iso- (Stratagene, LA Jolla, CA) and the supplied random lation replicates. A total of 849 genes showed relative primers. DNA sample preparation for sequencing on transcript ratios with Ͼ2.0-fold expression difference the 454/Roche GS FLX was performed as described by with statistical signiÞcance set at a FDR Ͻ2.5% (see the manufacturer (Margulies et al. 2005). Brießy, this Supplementary Materials). With these parameters, 260 JOURNAL OF MEDICAL ENTOMOLOGY Vol. 46, no. 2

Table 1. Transcripts showing >10-fold sex-speciﬁc differential expression in adult horn ﬂies

b Unigene FC BLAST annotation a ID Array 454 Gene ID Species Accession no. e-value Female Ͼ males transcripts 670 42.3 5.5 NSc ÑÑÑ 1117 31.6 Ͼ5d Nanos D. melanogaster P25724 2.8eϪ25 3 21.1 2.25 NS Ñ Ñ Ñ 794 20.3 Ͼ2NS Ñ Ñ Ñ 2531 18.1 Ñe MAD2-like Mus musculus Q9Z1B5 3.7eϪ51 7 17.5 19.0 Ribonucleoside diphos. reductase D. melanogaster P48592 4.4eϪ173 sm. sub. 4208 15.9 1.0 NS Ñ Ñ Ñ 4174 15.0 4.0 Sarcoplasmic calcium-binding protein Neanthes diversicolor P04571 9.4eϪ13 4195 14.7 Ͼ1 DNA primase large subunit D. melanogaster Q9VPH2 7.0eϪ73 767 13.9 Ͼ5 Hoi-Polloi D. melanogaster Q9U3Z7 6.2eϪ47 1148 13.3 Ͼ2 Flap endonuclease I Xenopus laevis P70054 1.4eϪ125 402 13.0 Ñ NS Ñ Ñ Ñ 19 12.6 5.0 10 kDa mitochondrial Heat shock Oryzias latipes Q9W6X3 3.0eϪ24 protein 1418 12.5 3.0 NS Ñ Ñ Ñ 1423 12.4 1.7 RAC GTPase activating protein Homo sapiens Q9H0H5 2.2eϪ74 507 12.4 0.0 Aquaporin Aedes aegypti Q9NHW7 5.0eϪ22 477 11.7 3.0 Hypothetical protein D. melanogaster Q9W3C2 4.1eϪ61 551 11.6 2.6 High mobility group protein D D. melanogaster Q05783 1.2eϪ28 2745 11.6 Ͼ6 Microtubule-assoc. protein RP/EB Coturnix japonica Q66T82 1.2eϪ42 3755 11.1 Ͼ4NS Ñ Ñ Ñ 616 11.1 8.0 Methyltransferase-like ribonucleoprotein Rattus norvegicus Q63009 1.5eϪ133 106 11.0 2.0 Nucleolar protein family A D. melanogaster Q9V3U2 5.7eϪ56 506 10.8 0.0 Aquaporin 1 M. musculus Q02013 3.9eϪ41 1408 10.6 Ͼ4 Disc proliferation abnormal D. melanogaster Q26454 0 1664 10.5 7.0 Maternal protein exuperantia Drosophila virilis Q24747 6.0eϪ50 4200 10.5 Ͼ5 Ribonucleoside diphosphate reductase D. melanogaster P48591 1.0eϪ48 2481 10.3 3.6 Heat shock protein 27 D. melanogaster P02518 2.6eϪ58 4234 10.3 Ͼ1 Thyroid hormone receptor interactor H. sapiens Q15645 6.2eϪ40 Male Ͼ female transcripts 2258 12.4 Ͼ2 Dual speciÞcity phosphatase 18 M. musculus Q8VE01 5.8eϪ30 3098 12.4 4.0 Glycerol-3-phosphate dehydrog. mitoch. H. sapiens P43304 8.8eϪ77 3501 11.9 8.0 Glycerol-3-phosphate dehydrog. mitoch. H. sapiens P43304 2.9eϪ96 2294 11.1 Ͻ1f Calcium-dependent secretion activator D. melanogaster Q9NHE5 2.7eϪ125 2965 11.0 7.0 Glycerol-3-phosphate dehydrog. mitoch. R. norvegicus P35571 5.2eϪ10 444 10.8 1.4 Myosin light polypeptide kinase Gallus gallus P11799 4.7eϪ42 2631 10.1 2.2 Allergen Plo I Plodia interpunctella Q95PM9 4.2eϪ20

a Unigene ID represents the identiÞcation no. from the horn ßy combined assembled EST database (Guerrero et al. 2008). b FC is the fold-change for each analyzed transcript derived by dividing the test expression level value by the control expression level value. c No signiÞcant BLAST hit. d Ͼ signiÞes the Unigene transcript was not found in the control expression dataset, so control value was set as ϭ 1 to avoid division by zero in the fold-change calculation. e Ñ signiÞes neither the female nor male 454 datasets contained sequences corresponding to this Unigene set transcript. f Ͻ signiÞes the Unigene transcript was not found in the test expression dataset, so test value was set as ϭ 1. there were 432 and 417 transcripts whose expression 2008), life stages occurring before adult ßy sexual was higher and lower, respectively, in females com- differentiation. Figure 1 shows the percentages of the pared with males. Because the horn ßy Unigene da- differentially expressed horn ßy sequences as they tabase (Guerrero et al. 2008) contains 4,408 unique match sequences from various taxa. The larval and transcripts, the 849 sex-speciÞc expressed transcripts adult comparisons will be discussed below. However, represent 19% of the horn ßy Unigene database. By there is a major difference between the “no hit” per- comparison, a study by Ranz et al. (2003) found Ϸ60% centages of the adult female expression Ͼ adult male of the genes in D. melanogaster were differentially expression (2%) and the adult male expression Ͼ adult expressed in a sex-speciÞc manner. Ellegren and Par- female expression genes (29%). This difference is pri- sch (2007) consider that sex-biased gene expression marily reßected in a reduction in the male Ͼ female might be more prevalent after sexual differentiation. hits to D. melanogaster (40Ð23%) and mouse (15Ð9%) This might explain the relatively lower percentage of genes. These horn ßy data are consistent with the sex-speciÞc gene expression in our horn ßy array study information reported in the review by Ellegren and compared with that reported by Ranz et al. (2003) Parsch (2007) that noted the rates of evolution of because the horn ßy Unigene dataset used to create male-biased genes in both D. melanogaster and Cae- our horn ßy microarrays was derived from normalized norhabditis elegans were faster than female-biased embryonic and larval cDNA libraries (Guerrero et al. genes of those species. A faster rate of evolution of March 2009 GUERRERO ET AL.: MICROARRAY ANALYSIS OF HORN FLY GENE EXPRESSION 261

Female > Male Male > Female

No Hit Mouse Other Misc. 2% Mous e 9% 13% 15% Rat 4% No Hit Other Insect 29% 3% Rat 6%

Other Vertebrate Drosophila 10% melanogaster 23%

Human 11% Other Misc. 13% Drosophila Ot her In sect Human melanogaster 3% 11% 40% Other Vertebrate 8% Adult > Larval Larval > Adult

Mouse Mouse 7% 11% No Hit Rat 20% Rat 2% 6% Drosophila No Hit melanogas ter 38% 19% Ot her Misc. 9% Drosophila melanogaster Human Ot her Insect 25% 6% 6% Other Vertebrate 4% Other Vertebrate 11% Human Other Misc. Other Insect 12% 13% 11% Fig. 1. Percentage of horn ßy sequences from various differentially expressed categories that match known sequences of different taxa. male-biased horn ßy genes could evidence itself in a notation. Perhaps these are transcripts with functions higher percentage of genes without Blast hits in the unique to female horn ßies such as aspects of the life adult male expression Ͼ adult female expression data- cycle related to host Þnding, mating, or location of set compared with data from the adult female Ͼ adult oviposition site. A lack of Blast hits could also result if male dataset. This is what Fig. 1 details. However, an EST was composed largely of untranslated region there is avian data from Mank et al. (2008) that shows with poor conservation among orthologs, rather than female-biased genes have faster rates of evolution than a more conserved protein-coding region of the gene. male-biased genes, so generalizations over distantly A number of the regulated genes in Table 1 have been related taxa might be questionable. linked to differential expression based on sex. Among Table 1 lists the members of the Unigene set whose the 21 transcripts in Table 1 whose expression was transcripts showed Ͼ10-fold differential expression higher in females compared with males that could be under our experimental conditions. There were 28 and assigned putative function were Nanos, Mad2, Flap 7 transcripts expressed Ͼ10-fold higher and lower, Endonuclease I, Aquaporin, and Exuperantia. In ad- respectively, in females compared with males. Of the dition to its role in early embryonic pattern formation, transcripts whose expression was Ͼ10-fold higher in Nanos is required in the adult ovary for the mainte- females compared with males, seven had no signiÞcant nance of germline stem cells and in the adult germline BlastX hits (E-value Ͻ0.001). In fact, three of the top for early oogenesis (Forbes and Lehmann 1998). four transcripts in the female expression Ͼ male ex- MAD2, as a spindle checkpoint protein, plays a role in pression category could not be assigned putative func- regulation of the Þrst meiotic division of mammalian tions, a bit surprising in light of having the well-an- oocytes and chromosome segregation (Wang et al. notated D. melanogaster genome as a good model. A 2007). Aquaporins are a family of small water channels BlastN analysis also failed to discover informative an- that facilitate water movement across membranes and 262 JOURNAL OF MEDICAL ENTOMOLOGY Vol. 46, no. 2

Table 2. Most abundant functional categorization terms in transcripts differentially overexpressed in females compared with males (P value < e؊05)

Categorya Termb Countc %d P value SP PIR Keywords Nuclear protein 59 17.6 1.2eϪ16 SP PIR Keywords DNA replication 13 3.9 1.7eϪ14 SP PIR Keywords RNA binding 21 6.3 3.4eϪ08 SP PIR Keywords Initiation factor 7 2.1 3.1eϪ06 SP PIR Keywords rRNA processing 6 1.8 3.9eϪ06 GO Term MF Nucleic acid binding 61 18.2 1.2eϪ15 GO Term MF RNA binding 24 7.2 6.4eϪ11 GO Term MF Binding 112 33.4 1.5eϪ09 GO Term MF Protein binding 82 24.5 3.6eϪ07 GO Term MF mRNA binding 16 4.8 4.1eϪ07 GO Term CC Intracellular 107 31.9 1.6eϪ19 GO Term CC Nucleus 63 18.8 9.2eϪ11 GO Term CC Protein complex 59 17.6 2.4eϪ10 GO Term CC Intracellular organelle 82 24.5 2.4eϪ09 GO Term CC Organelle 82 24.5 3.1eϪ09 GO Term CC Intracellular membrane-bound organelle 74 22.1 9.1eϪ09 GO Term CC Membrane-bound organelle 74 22.1 1.2eϪ08 GO Term CC Nuclear lumen 22 6.6 3.8eϪ08 GO Term CC Cell 109 32.5 5.9eϪ07 GO Term CC Organelle lumen 24 7.2 1.2eϪ06 GO Term CC Membrane-enclosed lumen 24 7.2 1.2eϪ06 GO Term BP Nucleobase/nucleoside/nucleotide/nucleic acid metabolism 66 19.7 9.7eϪ15 GO Term BP Primary metabolism 106 31.6 8.2eϪ13 GO Term BP Macromolecule metabolism 87 26 1.7eϪ12 GO Term BP DNA replication 17 5.1 6.9eϪ11 GO Term BP DNA metabolism 28 8.4 7.6eϪ11 GO Term BP Cellular metabolism 104 31 8.9eϪ11 GO Term BP Biopolymer metabolism 57 17 3.5eϪ10 GO Term BP Cellular physiological process 120 35.8 7.1eϪ10 GO Term BP Metabolism 107 31.9 1.7eϪ09 GO Term BP DNA-dependent DNA replication 11 3.3 1.1eϪ08 GO Term BP Cell organization and biogenesis 46 13.7 1.3eϪ08 GO Term BP RNA metabolism 22 6.6 4.5eϪ08 GO Term BP Cellular process 122 36.4 7.0eϪ08 GO Term BP DNA strand elongation 5 1.5 6.0eϪ07 GO Term BP Chromosome organization and biogenesis 16 4.8 9.2eϪ07 GO Term BP Organelle organization and biogenesis 29 8.7 1.2eϪ06 GO Term BP RNA processing 17 5.1 1.3eϪ06 GO Term BP Translation 13 3.9 4.6eϪ06 GO Term BP Physiological process 120 35.8 5.4eϪ06 GO Term BP mRNA processing 14 4.2 6.9eϪ06

a ClassiÞcation system either from Swiss-Prot Protein Knowledgebase or Gene Ontology (GO) Project. b Swiss-Prot Protein Knowledgebase Keywords or GO Molecular Function (MF), Cellular Component (CC), or Biological Process (BP) terms. c Number of transcripts containing the speciÞc term in the categorization annotation of female expression Ͼ male expression transcript list. d Percent of transcripts in female expression Ͼ male expression transcript list that contain the speciÞc term. members of the family have roles in both male and family of protein tyrosine phosphatases that have ex- female reproductive systems (Huang et al. 2006). The pression proÞles that vary according to family mem- exuperantia gene in D. melanogaster encodes overlap- ber. Human DUSP18 mRNA was found in 16 human ping sex-speciÞc transcripts that are subjects of sex- tissues examined but showed highest levels in liver, speciÞc RNA processing. The sex-speciÞc expression brain, testis, and ovary (Wu et al. 2003). This horn ßy of exuperantia is critical to both oogenesis and sper- transcript may be a DSP family member that is male matogenesis (Hazelrigg and Tu 1994). The Flap En- speciÞc in its expression pattern. The mitochondrial donuclease I transcript encodes a DNA repair enzyme GPD expression pattern is interesting in that three that interacts with mammalian estrogen receptor. In different horn ßy Unigene sequences showed signif- this system, estrogen exerts great effects on expression icant similarity to GPDs. Horn Fly Unigene Contig of the Flap Endonuclease I in speciÞc uterine cell 3098 and 3501 both showed signiÞcant sequence sim- types (Schultz-Norton et al. 2007). ilarity to the identical human mitochondrial GPD (Ta- Among the genes showing the greatest sex-speciÞc ble 1; GenBank: P43304), whereas the top BlastX hit differential expression, those with greater expression for Contig 2965 was the mitochondrial GDP from in males compared with females include dual speci- Rattus norvegicus (GenBank: P35571). However, the Þcity phosphatase (DSP) 18, mitochondrial glycerol- three horn ßy sequences that were all derived from a 3-phosphate dehydrogenase (GPD), calcium-depen- Þrst-instar larval cDNA library (Guerrero et al. 2008) dent secretion activator, myosin light polypeptide have very little nucleotide similarity between them- kinase, and allergen Plo I (Table 1). DSPs comprise a selves (data not shown). Mitochondrial GPD is known March 2009 GUERRERO ET AL.: MICROARRAY ANALYSIS OF HORN FLY GENE EXPRESSION 263 to show sex- and tissue-speciÞc differential expression Table 3. Most abundant functional categorization terms in transcripts differentially under-expressed in females compared Weitzel et al. 2003) and perhaps the three horn ßy ؊) with males (P value < e 05) larval transcripts are derived from differential splicing of transcripts from male-speciÞc larval tissues at various Categorya Termb Countc %d P value stages of development. The lower expression of calcium- Ϫ SP PIR Keywords Alternative splicing 19 9.9 7.2e 09 dependent secretion activator in females compared with SP PIR Keywords Mitochondrion 11 5.8 4.3eϪ07 males is difÞcult to explain, because it is reported to be SP PIR Keywords Transit peptide 8 4.2 2.8eϪ06 expressed preferentially in neural tissues, but expression GO Term CC Cytoplasm 27 14.1 2.6eϪ06 Ϫ07 in mouse testis and ovary is reported to be similar in those GO Term BP Organismal physiological 19 9.9 4.2e process organs (Su et al. 2002; http://smd.stanford.edu/cgi-bin/ GO Term BP Transmission of nerve 12 6.3 1.9eϪ06 source/expressionSearch?optionϭcluster&criteriaϭ impulse ϭ ϭ GO Term BP CellÐcell signaling 13 6.8 2.0eϪ06 Mm.260881&dataset 11&organism Mm). Myosin Ϫ06 light polypeptide kinase, a calmodulin-regulated ki- GO Term BP Synaptic transmission 10 5.2 2.3e nase that targets the myosin II molecular motor sys- a ClassiÞcation system either from Swiss-Prot Protein Knowledge- tem, has been shown to be expressed in a tissue- base or Gene Ontology Project. speciÞc manner (Birukov et al. 1998). Finally, allergen b Swiss-Prot Protein Knowledgebase Keywords or GO Molecular Plo I is an arginine kinase from the indian mealmoth, Function (MF), Cellular Component (CC), or Biological Process (BP) terms. and these enzymes show ecdysone-responsive expres- c Number of transcripts containing the speciÞc term in the cate- sion (James and Collier 1992), although information gorization annotation of female expression Ͻ male expression tran- on the sex-speciÞc expression of arginine kinase from script list. d Percent of transcripts in female expression Ͻ male expression indian mealmoth was not found. transcript list that contain the speciÞc term. Based on functional category enrichment in the gene set with female expression Ͼ male expression, it is notable that the SP PIR Keywords nuclear proteins mon theme of the biological process GO Terms in and DNA replication were the two most abundant Table 3. terms (Table 2). The most abundant molecular func- On a broader scale, it is interesting to note the tion GO Terms in Table 2 are all related to binding studies of Mank et al. (2008), who reported sex-biased gene expression in gonads from embryonic chickens. processes, including nucleic acid, RNA, mRNA, and In their study, male-biased GO Terms were generally protein binding. In addition, the GO Term sequence- related to membrane-associated activities or cellular speciÞc DNA binding was just above the P value cut- components (e.g., 0016020 membrane, 0005244 volt- off used in creating Table 2 (see Supplementary Ma- age-gated ion channel activity, and 0031224 intrinsic to terials). The most abundant cellular component GO membrane), whereas female-biased GO Terms were Terms in Table 2 were intracellular, nucleus, and pro- more related to transport-associated processes and tein complex, whereas the most abundant biological components (e.g., 0030705 cytoskeletal-dependent in- process GO Terms generally related to various aspects tracellular transport, 0030136 clathrin-coated vesicle, of metabolism. These female over-represented classi- and 0006828 iron ion transport). GO Term 005576 Þcation terms showed good consistency with the list of extracellular region was also reported as a female- female overexpressed transcripts (Table 1). For ex- biased cellular component term. Our study using the ample, Nanos and Exuperantia are associated with embryonic and larval horn ßy tissue-derived Unigene oogenesis, whereas the proteins MAD2, DNA primase, set showed the cellular component GO Term intra- Flap endonuclease, GTPase activating protein, and cellular as a female-biased term along with several nucleolar protein are associated with nuclear binding other intracellular-associated terms (Table 2). Addi- or metabolic processes of various types. tionally, several membrane-associated GO terms re- Functional category enrichment of the transcripts ex- lated to organelles were found in the female-biased pressed higher in males compared with females showed terms of Table 2. The membrane-associated GO terms that genes were highly enriched to the cytoplasmic cel- in the embryonic chick gonad study were found in the lular component and the primary biological processes male-biased category of Mank et al. (2008). Whether were related to cellÐcell signaling and neurological pro- these differences would stand if gene expression in ßy cesses (Table 3). The SP PIR Keywords are alternative gonad tissues were compared with the chick gonad splicing, mitochondrion, and transit peptide. Both SP PIR gene expression study is not known, because Mank et and GO Terms categories are consistent with the female al. (2007) cited studies that suggest the evolution of underexpressed transcripts (male Ͼ female category) sex-biased genes differs between D. melanogaster, a listed in Table 1. Three of the seven transcripts in this dipteran as is the horn ßy, and birds. category have signiÞcant similarity to mitochondrial Four transcripts whose expression was greater in GPD, an enzyme with a number of alternative transcrip- females than males and two whose expression was tional pathways (Weitzel et al. 2003). The calcium-de- greater in males than females were selected for veri- pendent secretion activator, another transcript showing Þcation of the microarray results by qRT-PCR. The greater expression in males compared with females, is results of the qRT-PCR (fold-change) correlated with expressedinneuraltissues(http://smd.stanford.edu/cgi- microarray expression at a correlation coefÞcient of bin/source/expressionSearch?optionϭcluster&criteriaϭ 0.98 (P Ͻ 0.001; Fig. 2). The female expression Ͼ male Mm.260881&datasetϭ11&organismϭMm), the com- expression Contigs 2531 (MAD2-like, 18.1 microarray 264 JOURNAL OF MEDICAL ENTOMOLOGY Vol. 46, no. 2

22.0

20.0 MAD2

18.0

16.0 PCR FC

14.0 Aquaporin

12.0

Myosin 10.0 GPD

8.0 Exuperantia

10.5 11.5 12.5 13.5 14.5 15.5 16.5 17.5 Microarray FC

Fig. 2. Plot of microarray fold-change versus fold-change from quantitative RT-PCR. The MAD2, Aquaporin, GPD, Myosin, and Exuperantia transcript PCR and microarray fold-change (FC) results were plotted with a resultant line of y ϭ 1.64(x) Ϫ 8.61, r2 ϭ 0.985, P ϭ 0.0008. fold-change), 507 (aquaporin, 12.4 microarray fold- results. The absolute sex-speciÞc expression levels change), and 1664 (maternal protein exuperantia, 10.5 showed variation, with only 7 of the 35 analyzed con- microarray fold-change) showed 21.0-, 12.6-, and 8.2- tigs (Table 1) having microarray and 454 fold-change fold-change, respectively, in qRT-PCR. The male ex- values within a 2.5-fold difference of each other. This pression Ͼ female expression Contigs 3501 (GPD, 11.9 variation is likely caused by the lack of true biological microarray fold-change) and 444 (myosin light replicates in the microarray, although more 454 se- polypeptide kinase, 10.8 microarray fold-change) quencing might reduce the discrepancies. The 454 showed 10.2- and 9.5-fold-change, respectively, in data from each dataset was obtained from a one-quar- qRT-PCR. Contig 1117 (nanos, 31.6 microarray fold- ter section from two independent 454 sequencing change) was also analyzed by qRT-PCR. Experimental plates, yielding a total of 74,750 and 57,438 sequences difÞculties prevented quantitative results to be deter- comprising the female and male datasets, respectively, mined; however, the fold-change for this transcript with average read lengths of 217 bp (data not shown). was Ͼ100-fold higher in females compared with males Nevertheless, the consistency of the directional trends (data not shown). of the microarray and 454 data were encouraging and Because of the limited amount of biological mate- led us to evaluate the data for larval-speciÞc and com- rial, we were not able to incorporate a powerful sta- bined sex-speciÞc and larval-speciÞc gene expression. tistical design with true biological replications into our Larval-speciﬁc Gene Expression. In a dye-swap de- microarray protocol. Recognizing that dye swaps are sign, RNA from Þrst-instar larvae was labeled and not true replications, for an overall determination of hybridized against a labeled 50:50 mix of RNA from sex-speciÞc transcript expression patterns, we per- adult male and female ßies to identify differential gene formed statistical analysis with the dye swaps entered expression in the early larval stages of the horn ßy life into the statistical program as true independent rep- cycle. As in the sex-speciÞc expression microarray lications. To have more conÞdence in the resultant design, limited biological material prompted us to des- microarray expression data, we used an independent ignate the dye swaps as biological replicates for sta- veriÞcation methodology, quantitative analysis of 454 tistical analysis. As expected, there were more gene pyrosequencing data, to validate the results from those expression differences in this larval study compared transcripts showing Ͼ10-fold sex-speciÞc differential with the sex-speciÞc study above. A total of 1,290 expression (Table 1). We tallied the number of times differentially expressed transcripts were found with each contig in the horn ßy Unigene Set (Guerrero et 419 and 871 transcripts expressed at higher and lower al. 2008) occurred in the female and male 454 se- levels, respectively, in larvae compared with adults quence datasets. Comparing the female and male tal- (FDR Ͻ 2.5%; see Supplementary Materials). Figure lies, we found that 23 of the 28 female Ͼ male tran- 1 shows the distributions of the horn ßy adult expres- scripts and 6 of the 7 male Ͼ female transcripts of sion Ͼ larval expression and larval expression Ͼ adult Table 1 showed consistency in the direction of sex- expression BLAST sequence matches to gene se- speciÞc expression between the microarray and 454 quences from various taxa. Despite having about one March 2009 GUERRERO ET AL.: MICROARRAY ANALYSIS OF HORN FLY GENE EXPRESSION 265

Table 4. Transcripts showing >20-fold stage-speciﬁc differential expression

Unigene b BLAST annotation a FC set ID Gene ID Species Accession no e-value Larval Ͼ adult transcripts 1014 396 AGAPER-1 Anopheles gambiae O76217 7.2eϪ13 3981 388 Larval cuticle protein VIII D. melanogaster P92201 1.7eϪ28 871 199 1,4-␤-N-acetylmuramidase 1 M. domestica Q7YT16 1.9eϪ72 1925 140 Proline-rich Extensin-related protein Daucus carota P06600 2.8eϪ6 314 139 Endocuticle structural glycoprot. SGABD-9 Schistocerca gregaria Q7M4F0 0.9eϪ3 2946 128 NSc ÑÑÑ 2085 123 Peritrophin-48 Lucilia cuprina P91745 6.5eϪ17 2199 110 Jonah 99CII D. melanogaster P17205 4.6eϪ47 2333 91 Mite allergen DER F 3- Dermatophagoides farinae P49275 5.2eϪ26 1124 89 NS Ñ Ñ Ñ 2591 89 NS Ñ Ñ Ñ 2677 88 NS Ñ Ñ Ñ 335 75 NS Ñ Ñ Ñ 527 66 NS Ñ Ñ Ñ 568 57 Serine protease SP24D A. gambiae Q17004 5.8eϪ39 2383 53 NS Ñ Ñ Ñ 2101 50 Trypsin ␤ precursor Drosophila erecta P54625 1.2eϪ75 2446 49 Troponin C D. melanogaster P47948 9.1eϪ60 2500 46 NS Ñ Ñ Ñ 2323 46 NS Ñ Ñ Ñ 445 45 Amyloid ␤ (A4) precursor-like H. sapiens Q06481 4.0eϪ12 2181 40 NS Ñ Ñ Ñ 308 38 Peritrophin-55 L. cuprina Q95UE8 3.1eϪ6 1044 34 NS Ñ Ñ Ñ 2488 32 NS Ñ Ñ Ñ 530 31 NS Ñ Ñ Ñ 1080 29 Proline hydroxylase Caenorhabditis elegans Q20065 8.2eϪ69 1060 29 Glutactin D. melanogaster P33438 3.7eϪ24 2579 29 NS Ñ Ñ Ñ 872 25 Ecdysone-inducible gene L3 D. melanogaster Q95028 2eϪ163 3112 24 NS Ñ Ñ Ñ 3405 24 NS Ñ Ñ Ñ 3143 24 Zinc carboxypeptidase A1 Drosophila pseudoobscura Q29NC4 9eϪ123 2452 23 Chymotrypsin 1 A. gambiae Q27289 1.6eϪ44 2420 23 Sporozoite surface protein 2 Plasmodium yoelii Q01443 7.3eϪ5 1884 23 NS Ñ Ñ Ñ 3043 22 E-selectin Sus scrofa P98110 1.3eϪ17 769 22 Actin Mayetiola destructor O16808 6.1eϪ61 Adult Ͼ larval transcripts 2449 42 Bifunctional purine biosynthesis protein G. gallus P31335 0 3772 40 3-Phosphoglycerate dehydrogenase R. novegicus O08651 7.5eϪ36 654 30 Lysosomal aspartic protease A. aegypti Q03168 1eϪ158 757 29 Glutathione S-transferase M. domestica P46437 3eϪ113 1325 24 Serine protease F56F10.1 C. elegans P90893 6.6eϪ73 1664 22 Maternal protein exuperantia D. virilis Q24747 6.0eϪ50 1335 21 Yolkless D. melanogaster P98163 2.9eϪ50 2965 20 Glycerol-3-phosphate dehydrogenase 2 R. norvegicus P35571 5.2eϪ10 3410 20 Manganese (II) purple acid phosphatase Glycine max Q09131 1.0eϪ16 69 20 Multifunctional protein ADE2 D. melanogaster Q9I7S8 0

Table 5. Most abundant functional categorization terms in transcripts differentially overexpressed in larvae compared with adults (P (value < e؊05

Categorya Termb Countc %d P value SP PIR Keywords Alternative splicing 53 8.2 9.7eϪ20 SP PIR Keywords Mitochondrion 27 4.2 1.1eϪ13 SP PIR Keywords Proteasome 15 2.3 1.1eϪ11 SP PIR Keywords Transit peptide 18 2.8 2.8eϪ11 SP PIR Keywords Phosphorylation 19 2.9 1.0eϪ07 SP PIR Keywords Threonine protease 9 1.4 1.8eϪ07 GO Term BP Protein catabolism 17 2.6 2.6eϪ11 GO Term BP ATP-dependent proteolysis 10 1.5 3.6eϪ11 GO Term BP Macromolecule catabolism 23 3.5 4.4eϪ11 GO Term BP Biopolymer catabolism 17 2.6 2.2eϪ10 GO Term BP Catabolism 29 4.5 9.8eϪ10 GO Term BP Primary metabolism 144 22.2 1.2eϪ09 GO Term BP Cellular metabolism 146 22.5 1.2eϪ09 GO Term BP Ubiquitin-dependent protein catabolism 14 2.2 1.3eϪ09 GO Term BP ModiÞcation-dependent protein catabolism 14 2.2 1.8eϪ09 GO Term BP Cellular macromolecule catabolism 20 3.1 4.0eϪ09 GO Term BP Cellular protein catabolism 14 2.2 4.3eϪ09 GO Term BP Proteolysis during cellular protein catabolism 14 2.2 4.3eϪ09 GO Term BP Cellular catabolism 26 4 5.3eϪ09 GO Term BP Metabolism 153 23.6 5.6eϪ09 GO Term BP Cell organization and biogenesis 62 9.6 7.4eϪ09 GO Term BP Chromosome organization and biogenesis 21 3.2 1.3eϪ07 GO Term BP Cellular physiological process 171 26.4 3.9eϪ07 GO Term BP Biopolymer metabolism 68 10.5 8.7eϪ07 GO Term BP Macromolecule metabolism 105 16.2 1.7eϪ06 GO Term BP Development 66 10.2 2.7eϪ06 GO Term BP Nucleobase/nucleoside/nucleotide/nucleic acid metab. 68 10.5 2.9eϪ06 GO Term BP Germ cell development 13 2 3.5eϪ06 GO Term BP Mitotic chromosome condensation 6 0.9 6.2eϪ06 GO Term BP Sexual reproduction 30 4.6 7.2eϪ06 GO Term BP Organelle organization and biogenesis 36 5.6 7.5eϪ06 GO Term BP Reproduction 32 4.9 7.8eϪ06 GO Term BP Chromosome condensation 7 1.1 7.9eϪ06 GO Term CC Intracellular 147 22.7 2.4eϪ19 GO Term CC Cytoplasm 85 13.1 1.4eϪ12 GO Term CC Proteasome complex (sensu Eukaryota) 15 2.3 5.5eϪ12 GO Term CC Protein complex 73 11.3 2.7eϪ08 GO Term CC Proteasome core complex (sensu Eukaryota) 9 1.4 3.1eϪ08 GO Term CC Intracellular membrane-bound organelle 99 15.3 3.6eϪ08 GO Term CC Intracellular organelle 109 16.8 4.3eϪ08 GO Term CC Membrane-bound organelle 99 15.3 5.0eϪ08 GO Term CC Organelle 109 16.8 5.7eϪ08 GO Term CC Cell 156 24.1 3.9eϪ07 GO Term CC Proton-transporting ATP synthase complex (sensu Eukaryota) 7 1.1 5.1eϪ06 GO Term CC Proton-transporting ATP synthase complex 7 1.1 5.1eϪ06 GO Term CC Mitochondrion 29 4.5 6.1eϪ06 GO Term MF Obsolete molecular function 21 3.2 1.6eϪ19 GO Term MF Proteasome endopeptidase activity 9 1.4 4.6eϪ12 GO Term MF Catalytic activity 113 17.4 1.1eϪ10 GO Term MF Protein binding 118 18.2 2.0eϪ08 GO Term MF Pyrophosphatase activity 27 4.2 3.3eϪ07 GO Term MF Binding 154 23.8 4.1eϪ07 GO Term MF Hydrolase act., acting on acid anhyd., in phosph.-containing anhyd. 27 4.2 5.5eϪ07 GO Term MF Hydrolase activity, acting on acid anhydrides 27 4.2 5.7eϪ07 GO Term MF Mrna binding 19 2.9 8.2eϪ07 GO Term MF Nucleoside-triphosphatase activity 26 4 8.5eϪ07 GO Term MF Hydrogen-transporting two-sector atpase act. 7 1.1 1.2eϪ06 GO Term MF Atpase activity 21 3.2 2.6eϪ06 GO Term MF RNA binding 22 3.4 2.8eϪ06 GO Term MF Hydrolase activity 58 9 3.8eϪ06

Table 6. Most abundant functional categorization terms in transcripts differentially underexpressed in larvae compared with adults (P value < e؊05)

Categorya Termb Countc %d P value SP PIR Keywords Ribosomal protein 32 13.7 1.5eϪ38 SP PIR Keywords Ribonucleoprotein 33 14.1 1.5eϪ37 SP PIR Keywords Ribosome 12 5.1 5.4eϪ16 SP PIR Keywords Protein biosynthesis 12 5.1 2.5eϪ10 GO Term BP Protein biosynthesis 33 14.1 5.6eϪ20 GO Term BP Macromolecule biosynthesis 33 14.1 3.1eϪ19 GO Term BP Biosynthesis 36 15.4 6.1eϪ16 GO Term BP Cellular biosynthesis 35 15 6.5eϪ16 GO Term BP Cellular macromolecule metabolism 39 16.7 8.5eϪ08 GO Term BP Cellular protein metabolism 38 16.2 9.2eϪ08 GO Term BP Protein metabolism 39 16.7 1.0eϪ07 GO Term BP Macromolecule metabolism 46 19.7 7.5eϪ07 GO Term CC Cytosolic ribosome (sensu Eukaryota) 31 13.2 4.1eϪ44 GO Term CC Ribosome 31 13.2 2.7eϪ31 GO Term CC Cytosol 32 13.7 2.7eϪ25 GO Term CC Cytosolic small ribosomal subunit (sensu Eukaryota) 17 7.3 7.6eϪ25 GO Term CC Eukaryotic 48S initiation complex 17 7.3 7.6eϪ25 GO Term CC Ribonucleoprotein complex 32 13.7 1.8eϪ24 GO Term CC Eukaryotic 43S preinitiation complex 17 7.3 2.7eϪ21 GO Term CC Small ribosomal subunit 17 7.3 4.2eϪ20 GO Term CC Cytosolic large ribosomal subunit (sensu Eukaryota) 14 6 4.0eϪ17 GO Term CC Intracellular non-membrane-bound organelle 34 14.5 3.3eϪ16 GO Term CC Non-membrane-bound organelle 34 14.5 3.3eϪ16 GO Term CC Large ribosomal subunit 14 6 2.3eϪ13 GO Term CC Cytoplasm 41 17.5 2.1eϪ11 GO Term CC Protein complex 39 16.7 1.4eϪ10 GO Term CC Intracellular organelle 45 19.2 5.9eϪ06 GO Term CC Organelle 45 19.2 6.8eϪ06 GO Term MF Structural constituent of ribosome 31 13.2 2.2eϪ35 GO Term MF Structural molecule activity 37 15.8 5.1eϪ25 GO Term MF Nucleic acid binding 36 15.4 6.9eϪ10

a ClassiÞcation system either from Swiss-Prot Protein Knowledgebase or Gene Ontology Project. b Swiss-Prot Protein Knowledgebase Keywords or GO Molecular Function (MF), Cellular Component (CC), or Biological Process (BP) terms. c Number of transcripts containing the speciÞc term in the categorization annotation of larval expression Ͻ adult expression transcript list. d Percent of transcripts in larval expression Ͻ adult expression transcript list that contain the speciÞc term. expression Ͼ adult expression transcripts with Ͼ50- fold differential expression, 7 did not have a GenBank match, 4 (Contigs 2199, 2333, 568, and 2101) were proteases probably involved in digestive processes, 1 and 2 each were associated with the larval cuticle FemFem LLarar (Contigs 3981 and 314) or peritrophic membrane + + (Contigs 1014 and 2085). These protein types are ((432)432) ((419)419) consistent with the results of the functional category enrichment analysis of the transcripts with greater expression in larvae compared with adults (Table 5), 1144 5 which had proteasome, transit peptide, and threonine protease among the top SP PIR Keywords. Addition- ally, the top 14 biological process GO Terms were LLarar FFemem catabolism/metabolism-related, among the top Þve - - cellular component GO Terms were proteasome com- ((871)871) ((417)417) plex, protein complex, and proteasome core complex, 1111 and proteasome endopeptidase activity is among the top molecular function GO Terms. Clearly, relative to adult ßies, the Þrst-instar larval stage is highly invested Fig. 3. Diagrammatic representation of sex-speciÞc and in the expression of genes involved in catabolic pro- larval-speciÞc microarray datasets. The 432, 417, 419, and 871 female Ͼmale (Fem ϩ), female Ͻmale (Fem Ϫ), larval Ͼadult cesses, particularly protein digestion. The ant, C. fes- ϩ Ͻ Ϫ tinatus, also showed GO Terms in the larval-biased (Lar ), and larval adult (Lar ) transcripts, respectively, were examined to determine whether transcripts differentially genes, which were involved with processes associated expressed in the sex-speciÞc experiments were differentially with protein metabolism (Goodisman et al. 2005). expressed in the larval-speciÞc experiments and vice versa. The Anopheles gambiae shows a similar enrichment of GO numerals in the intersecting regions indicate the number of Terms associated with catabolism in larval-biased transcripts with a minimum fold-change ratio of Ͼ4.0 that occur genes (Koutsos et al. 2007). This preponderance of in both datasets. SpeciÞc transcript data are shown in Table 7. 268 JOURNAL OF MEDICAL ENTOMOLOGY Vol. 46, no. 2

Table 7. Contigs with both female- and larval-speciﬁc > 4.0-fold differential expression

b Unigene set FC BLAST annotation a ID Sex Stage Gene ID Species Accession no. Female Ͼ male and larval Ͼ adult transcripts 387 4.0 4.4 ␣-Tubulin 84D D. melanogaster P06605 Female Ͼ male and adult Ͼ larval transcripts 3 21.1 8.5 NSc ÑÑ 7 17.5 6.4 Ribonucleoside diphosphate reductase SS D. melanogaster P48592 1148 13.3 5.8 Flap endonuclease I X. laevis P70054 551 11.6 5.2 High mobility protein group D D. melanogaster Q05783 702 9.1 8.1 HVA22-like protein Arabidopsis thaliana Q8LEM6 194 9.0 8.1 Histone RNA hairpin binding protein D. melanogaster Q9VAN6 1252 8.7 5.5 NS Ñ Ñ 564 8.6 7.6 Minichromosome maintenance 2 D. melanogaster P49735 229 8.0 5.2 Minichromosome maintenance 7 X. laevis Q91876 1479 7.0 4.5 SET protein D. melanogaster P53997 2522 6.6 7.2 NS Ñ Ñ 265 5.6 4.2 Small ubiquitin-related modiÞer 2 R. norvegicus P61959 3856 5.1 6.1 Translin M. musculus Q62348 1698 5.0 4.1 Leucine aminopeptidase Helicobacter pylori O25294 2642 4.8 4.7 Reg. of chromosome. condens. 2 homolog Danio rerio Q6NYE2 Male Ͼ female and larval Ͼ adult transcripts 3857 7.2 4.7 Serine/threonine prot. kinase minibrain D. melanogaster P49657 3137 6.5 5.8 Pre-collagen alpha-1 R. norvegicus P05539 2848 6.2 10.4 Lysosomal membrane glycoprotein1 M. musculus P11438 446 4.5 10.2 NS Ñ Ñ 1898 4.4 4.6 NS Ñ Ñ Male Ͼ female and adult Ͼ larval transcripts 3098 12.4 7.3 Glycerol-3-phosphate dehydrogenase H. sapiens P43304 2631 10.1 7.4 Arginine kinase Plodia interpunctella Q95PM9 2969 10.0 4.8 NS Ñ Ñ 1656 9.1 4.2 Troponin T D. melanogaster P19351 1974 7.9 4.0 Titin D. melanogaster Q9I7U4 3730 7.2 4.9 NS Ñ Ñ 1807 5.4 5.0 NS Ñ Ñ 2611 5.4 4.8 Troponin C D. melanogaster P47947 3167 5.3 7.2 NADP-dependent malic enzyme Columga livia P40927 111 5.2 4.3 Ubiquitin thioesterase 64E D. melanogaster Q24574 1196 4.1 5.9 NS Ñ Ñ

a Unigene ID represents the identiÞcation no. from the horn ßy combined assembled EST database (Guerrero et al. 2008). b FC is the fold-change ratio from the sex-speciÞc and stage-speciÞc array experimental designs. c No signiÞcant BLAST hit. catabolic activity in the larval stage is not unexpected, cating the expression of transcript Contig 2965 is prob- because metamorphic processes require the break- ably restricted to the adult male life stage (Table 1). down of larval tissues in preparation for pupal devel- Two adult expression Ͼ larval expression transcripts opment. The early larval stage of the horn ßy is char- (Contigs 2449 and 69) were related to the de novo acterized by continual feeding in relatively refractory pathway for biosynthesis of purines, a pathway re- yet microbially active cattle manure. The larval stage quired during D. melanogaster metamorphosis and em- is also characterized by rapid development to the bryogenesis, whereas the larval stages survive on ex- pupal stage, because the volume increase from the ogenous source of purines (Clark and MacAfee 2000). Þrst-instar larval stage to the pupal stage is Ϸ50-fold, Several of these genes with adult expression Ͼ larval occurring over a 3- to 4-d period (Bruce 1964). expression seem to be related to functions associated Only 10 transcripts were found to be expressed at with adult females, such as blood feeding (the iron- Ͼ20-fold higher levels in adults compared with larvae binding manganese purple acid phosphatase-encod- (Table 4). Of these 10 transcripts, the maternal protein ing Contig 3410), vitellogenesis (Yolkless-encoding exuperantia (Contig 1664) was also listed as a female Contig 1335), and embryogenesis (lysosomal aspartic expression Ͼ male expression transcript, whereas the protease-encoding Contig 654). Table 6 shows the SP mitochondrial GPD2 (Contig 2965) was listed as a PIR Keywords and GO Terms that are associated with male expression Ͼ female expression transcript, indi- the transcripts expressed at lower levels in larvae com- March 2009 GUERRERO ET AL.: MICROARRAY ANALYSIS OF HORN FLY GENE EXPRESSION 269 pared with adults. All of the four SP PIR Keywords signiÞcance (see Materials and Methods). Thus, in associated with this group of transcripts are clearly Contigs 387, 670, and 1117, we identiÞed three can- related to ribosomal and protein biosynthetic activi- didates for detailed evaluation as a gene promoter ties. This could be because of the adult femaleÕs de- source for the development of a female-speciÞc con- votion of resources to oogenesis, embryogenesis, and ditional lethality system. These three gene promoters maternal biosynthesis of RNAs and proteins essential will be cloned from horn ßy genomic DNA and ligated to the survival of embryos before hatch. The top GO to reporter gene constructs to identify promoter se- Term in the biological process and molecular function quence regions that control sex- and life stageÐspeciÞc categories relate to the ribosome or protein biosyn- expression. Contig 1117, the putative nanos ortholog, thesis, whereas eight of the top nine cellular compo- seems to be the a very good candidate for a gene nent GO Terms relate to ribosomes or complexes that promoter source for our female-speciÞc conditional include ribosomes. lethality system, because it has a high female-to-male Combined Female- and Larval-speciﬁc Analysis. expression ratio, a moderate expression level in Þrst- With our primary goal of identifying a candidate gene instar larvae, and has been characterized in D. mela- promoter that could drive a female-speciÞc lethality nogaster. system with lethality expressed in the early larval stage, we analyzed the microarray experimental datasets for all combinations of female versus male and Acknowledgments larvae versus adult expression. The results are expressed diagrammatically in Fig. 3 and speciÞc tran- We thank C. Younger for collection of horn ßies, K. Ben- script identiÞcation data are tabulated in Table 7 for dele for RNA isolations, C. Wu for array hybridizations, and Ͼ C. Qui and P. Wang for help in preparing the libraries for transcripts with a minimum fold-change ratio of 4.0. pyrosequencing. Only one transcript showed a female Ͼ male and larval Ͼ adult expression pattern, Contig 387, which possesses sequence similarity to a-tubulin 84D from D. References Cited melanogaster (GenBank: P06605). Reportedly, this member of the tubulin family is expressed constitu- Altschul, S. F., W. Gish, E. W. Myers, and D. J. Lipman. 1990. tively in D. melanogaster (Theurkauf et al. 1986); how- Basic local alignment search tool. J. Mol. Biol. 215: 403Ð ever, in horn ßy, Contig 387 was expressed 4.0- and 410. 4.4-fold in females compared with males and Þrst- Ashburner, M., C. A. Ball, J. A. Blake, D. Botstein, H. Butler, J. M. Cherry, A. P. Davis, K. Dolinski, J. T., J. T. Eppig, instar larvae compared with adults, respectively. Sev- M. A. Harris, et al. 2000. Gene ontology: tool for the eral transcripts that seem to be adult female speciÞc uniÞcation of biology. The Gene Ontology consortium. (i.e., female Ͼ male and adults Ͼ larvae) were found, Nat. Genet. 25: 25Ð29. especially Contigs 3, 7, 1148, and 551, which had fe- Bairoch, A., R. Apweiler, C. H. Wu, W. C. Barker, B. Bo- male Ͼ male fold-changes Ͼ10. Two transcripts, Con- eckmann, S. Ferro, E. Gasteiger, H. Huang, R. Lopez, M. tigs 2848 and 446, were found whose expression was Magrane, et al. 2005. The universal protein resource Ͼ10-fold higher in larvae compared with adults but (UniProt). Nucleic Acids Res. 33: D154ÐD159. were preferentially expressed in males. Three tran- Benjamini, Y., and Y. Hochberg. 1995. Controlling the false scripts, Contigs 3098, 2631, and 2969, were preferen- discovery rate: a practical and powerful approach to multiple testing. J. R. Stat. Soc. B 57: 289Ð300. tially expressed Ͼ10-fold higher in adult males than Ͼ Ͼ Birukov, K. G., J. P. Schavocky, V. P. Shirinsky, M. V. adult females (i.e., males female and adult lar- Chibalina, L. J. Van Eldik, and D. M. Watterson. 1998. vae). Organization of the genetic locus for chicken myosin light Despite our identiÞcation of Contig 387 as a possible chain kinase is complex: multiple proteins are encoded source for a gene promoter that is active in the early and exhibit differential expression and localization. larval stage of females, the difference in the expression J. Cell. Biochem. 70: 402Ð413. of the Contig 387 transcript in females compared with Bruce, W. G. 1964. The history and biology of the horn ßy, males is only 4.0-fold. A number of transcripts in our Haematobia irritans (Linnaeus); with comments on con- database had higher female versus male expression trol. U.S. Department of Agriculture, Agricultural Re- search Service Tech. Bull. 157. fold changes (Table 1). Thus, we examined the raw Clark, D. V., and N. MacAfee. 2000. The purine biosynthesis intensity data for the top two female overexpressed enzyme PRAT detected in proenzyme and mature forms transcripts, Contigs 670 (no GenBank match) and during development of Drosophila melanogaster. Insect 1117 (putative nanos ortholog), to verify that their Biochem. Mol. Biol. 30: 315Ð323. expression in larvae would be high enough to warrant Dennis G., Jr., B. T. Sherman, D. A. Hosack J. Yang, W. Gao, study of their gene promoter for use in the female- H. C. Lane, and R. A. Lempicki. 2003. DAVID: database speciÞc lethality system with desired lethality ex- for annotation, visualization, and integrated discovery. pressed in the early larval stage. Both Contigs 670 and Genome Biol. 4: R60. 1117 had moderate expression levels in the larval sam- Dowd, S. E., J. Zaragoza, J. R. Rodriguez, M. J. Oliver, and P. R. Payton. 2005. Windows. NET network distributed ple (data not shown). In fact, Contig 1117 was ex- Ϸ basic local alignment search toolkit (W.ND-BLAST). pressed 10-fold higher in the larvae compared with BMC Bioinform. 6: 93. the mixed adult pool, although Contig 1117 did not fall Ellegren, H., and J. Parsch. 2007. The evolution of sex-bi- out as a signiÞcant larval expression Ͼ adult expression ased genes and sex-biased gene expression. Nat. Rev. transcript based on our selected cut-offs for statistical Genet. 8: 689Ð698. 270 JOURNAL OF MEDICAL ENTOMOLOGY Vol. 46, no. 2

Forbes, A., and R. Lehmann. 1998. Nanos and Pumilio have biased genes: insights from avian microarray data. BMC critical roles in the development and function of Dro- Genom. 9: 148. sophila germline stem cells. Development 125: 679Ð690. Margulies, M., M. Egholm, W. E. Altman, S. Attiya, J. B. Fu, G., K. C. Condon, M. J. Epton, P. Gong, L. Jin, G. C. Bader, L. A. Bemben, J. Berka, M. S. Braverman, Condon, N. I. Morrison, T., H. Dafa’alla, and L. Alphey. Y-J. Chen, Z. Chen, et al. 2005. Genome sequencing in 2007. Female-speciÞc insect lethality engineered using microfabricated high-density picolitre reactors. Nature alternative splicing. Nature Biotech. 25: 353Ð357. (Lond.) 437: 376Ð380. Goodisman, M.A.D., J. Isoe, D. E. Wheeler, and M. A. Wells. Ogata, H., S. Goto, K. Sato, W. Fujibuchi, H. Bono, and M. 2005. Evolution of insect metamorphosis: a microarray- Kanehisa. 1999. KEGG: Kyoto encyclopedia of genes based study of larval and adult gene expression in the ant, and genomes. Nucleic Acids Res. 27: 29Ð34. Camponotus festinatus. Evolution. 59: 858Ð870. Ranz, J. M., C. I. Castillo-Davis, C. D. Meiklejohn, and D. L. Grisi, L., C. L. Massard, G. E. Moya Borja, and J. B. Pereira. Hartl. 2003. Sex-dependent gene expression and evolu- 2002. Impacto econoˆmico das principais ectoparasitoses tion of the Drosophila transcriptome. Science 300: 1742Ð em bovinos no Brasil. Hora Vet. 125: 8Ð10. 1744. Guerrero, F. D., and A.T.M. Barros. 2006. Role of kdr and Roe, B. 2004. Shotgun library construction for DNA se- esterase-mediated metabolism in pyrethroid-resistant quencing, pp. 171Ð187. In S. Zhao and M. Stodolsky (eds.), populations of Haematobia irritans irritans (Diptera: Methods in molecular biology, vol. 255, bacterial artiÞcial Muscidae) in Brazil. J. Med. Entomol. 43: 896Ð901. chromosomes, volume 1: library construction, physical Guerrero, F. D., T. J. Lysyk, and L. Kalischuk-Tymensen. mapping, and sequencing. Humana Press, Totowa, NJ. 2004. Expressed sequence tags and new gene coding re- Schultz-Norton, J. R., K. A. Walt, Y. S. Ziegler, I. X. McLeod, gions from the horn ßy. Southwest. Entomol. 29: 193Ð208. J. R. Yates, L. T. Raetzman, and A. M. Nardulli. 2007. The Guerrero, F. D., S. E. Dowd, V. M. Nene, and L. D. Foil. deoxyribonucleic acid repair protein ßap endonuclease-1 2008. Expressed cDNAs from embryonic and larval modulates estrogen-responsive gene expression. Mol. En- stages of the horn ßy (Diptera: Muscidae). J. Med. En- docrinol. 21: 1569Ð1580. tomol. 45: 686Ð692. Scott, M. J., J. C. Heinrich, and X. Li. 2004. Progress towards Hazelrigg, T., and C. Tu. 1994. Sex-speciÞc processing of the development of a transgenic strain of the Australian the Drosophila exuperantia transcript is regulated in male sheep blowßy (Lucilia cuprina) suitable for a male-only germ cells by the tra-2 gene. Proc. Natl. Acad. Sci. U.S.A. sterile release program. Insect Biochem. Mol. Biol. 34: 91: 10752Ð10756. 185Ð192. Huang, H. F., R. H. He, C. C. Sun, Y. Zhang, Q-X. Meng, and Su, A. I., M. P. Cooke, K. A. Ching, Y. Hakak, J. R. Walker, Y-Y. Ma. 2006. Function of aquaporins in female and T. Wiltshire, A. P. Orth, R. G. Vega, L. M. Sapinoso, A. male reproductive systems. Hum. Reprod. Update 12: Moqrich, et al. 2002. Large-scale analysis of the human 785Ð795. and mouse transcriptomes. Proc. Natl. Acad. Sci. U.S.A. James, J. M., and G. E. Collier. 1992. Early gene interaction 99: 4465Ð4470. during prepupal expression of Drosophila arginine kinase. Theurkauf, W. E., H. Baum, J. Bo, and P. C. Wensink. 1986. Dev. Genet. 13: 302Ð305. Tissue-speciÞc and constitutive a-tubulin genes of Dro- Koutsos, A. C., C. Blass, S. Meister, S. Schmidt, R. M. Mac- sophila melanogaster code for structurally distinct pro- Callum, M. B. Soares, F. H. Collins, V. Benes, E. Zdobnov, teins. Proc. Natl. Acad. Sci. U.S.A. 83: 8477Ð8481. F. C. Kafatos, and G. K. Christophides. 2007. Life cycle Thomas, D. D., C. A. Donnelly, R. J. Wood, and L. S. Alphey. transcriptome of the malaria mosquito Anopheles gambiae 2000. Insect population control using a dominant, re- and comparison with the fruitßy Drosophila melanogaster. pressible, lethal genetic system. Science 287: 2474Ð2476. Proc. Natl. Acad. Sci. U.S.A. 104: 11304Ð11309. Tsai, C. A., H. M. Hsueh, and J. J. Chen. 2003. Estimation of Kunz, S. E., and C. D. Schmidt. 1985. The pyrethroid resis- false discovery rates in multiple testing: application to tance problem in the horn ßy. J. Agric. Entomol. 2: 358Ð gene microarray data. Biometrics 59: 1071Ð1081. 363. Wang, J-Y, Z-L. Lei, C-L. Nan, S. Yin, J. Liu, Y. Hou, Y-L. Li, Kunz, S. E., K. D. Murrell, G. Lambert, L. F. James, and C. E. D-Y. Chen, and Q-Y. Sun. 2007. RNA interference as a Terrill. 1991. Estimated losses of livestock to pests, pp. tool to study the function of MAD2 in mouse oocyte 69Ð98. In Pimentel D (ed.), CRC handbook of pest man- meiotic maturation. Mol. Reprod. Dev. 74: 116Ð124. agement, vol. 1, 2nd ed. CRC Press, Boca Raton, FL. Weitzel, J. M., N. B. Shiryaeva, R. Middendorff, M. Balvers, Kunz, S. E., M. O. Estrada, and H. F. Sanchez. 1995. Status C. Radtke, R. Ivell, and H. J. Seitz. 2003. Testis-speciÞc of Haematobia irritans (Diptera: Muscidae) insecticide expression of rat mitochondrial glycerol-3-phosphate de- resistance in northeastern Mexico. J. Med. Entomol. 32: hydrogenase in haploid male germ cells. Biol. Reprod. 68: 726Ð729. 699Ð707. Lysyk, T. J. 1991. Use of life history parameters to improve White, N. M., and M. Bownes. 1997. Cloning and charac- a rearing method for horn ßy, Haematobia irritans irritans terization of three Musca domestica yolk protein genes. (L.) (Diptera: Muscidae), on bovine hosts. Can. Entomol. Insect Mol. Biol. 6: 329Ð341. 123: 1199Ð1207. Wu, Q., S. Gu, J. Dai, J. Dai, L. Wang, Y. Li, L. Zeng, J. Xu, Mank, J. E., L. Hultin-Rosenberg, E. Axelsson, and H. El- X. Ye, W. Zhao, C. Ji, Y. Xie, and Y. Mao. 2003. Molecular legren. 2007. Rapid evolution of female-biased, but not cloning and characterization of a novel dual-speciÞcity male-biased, genes expressed in avian brain. Molec. Biol. phosphatase18 gene from human fetal brain. Biochem. Evol. 24: 2698Ð2706. Biophys. Acta 1625: 296Ð304. Mank, J. E., L. Hultin-Rosenberg, M. T. Webster, and H. Ellegren. 2008. The unique genomic properties of sex- Received 24 October 2008; accepted 12 December 2008.