Insect Biochemistry and Molecular Biology 33 (2003) 609–622www.elsevier.com/locate/ibmb

Multiple acyl-CoA desaturase-encoding transcripts in pheromoneglands ofHelicoverpa assulta, the oriental tobacco budworm�

Seong Eun Jeongb, Claire-Lise Rosenfielda, Patricia Marsella-Herricka,Kyung Man Youa,1, Douglas C. Knipplea,∗

a Department of Entomology, Cornell University, New York State Agricultural Experiment Station, 630 W. North Street, Geneva,NY 14456, USA

b Department of Biological Sciences, Hannam University, 133 Ojung-Dong, Taeduk-Ku, Taejon 300-791, South Korea

Received 2 October 2002; received in revised form 25 February 2003; accepted 3 March 2003

Abstract

Seven desaturase cDNAs were isolated from pheromone glands ofHelicoverpa assulta, a moth producing a sex pheromone blendwith high Z9-16:Ald and lowZ11-16:Ald, opposite to what is found in other heliothine moths such asHelicoverpa zea. Six of theseven sequences map onto recently defined lepidopteran desaturase sequence lineages and the other is orthologous to a desaturasesequence previously reported only inH. zea. The levels of desaturase-encoding transcripts in pheromone glands were determinedand the three most abundant ones were functionally expressed in a desaturase-deficient mutant strain ofSaccharomyces cerevisiae.The HassNPVE transcript, shown to encode a�9 desaturase producing moreZ9-18:Acid thanZ9-16:Acid, was the most abundant,followed by theHassKPSE transcript, shown to encode a�9 desaturase producing moreZ9-16:Acid thanZ9-18:Acid, and by theHassLPAQ transcript, shown to encode a�11 desaturase producing onlyZ11-16:Acid. Thus, the relative amounts of transcriptsencoding two�9 desaturases and a single�11 desaturase inH. assulta pheromone glands were consistent with the relative amountsof unsaturated fatty acid precursors required to produce the major and minor sex pheromone components of this species. Desaturasetranscript levels in pheromone glands were also found to be as high during scotophase as during light phase, when pheromoneproduction ceases. The other four transcripts were present at extremely low levels inH. assulta pheromone glands and the functionalroles of their encoded desaturase-homologous proteins could not be determined. 2003 Elsevier Science Ltd. All rights reserved.

Keywords: acyl-CoA desaturase; Gene family; Evolution; Sex pheromone biosynthesis; Lepidoptera; Noctuidae; Heliothinae;Helicoverpa assulta;Helicoverpa zea; cDNA; Functional expression;Saccharomyces cerevisiae; RT-PCR

1. Introduction

The oriental tobacco budworm,Helicoverpa assulta,found throughout East Asia, Africa, and Australasia, isa polyphagous pest of such crops as peppers, tobaccoand onions. Since larvae feed within the fruits of host

∗ Corresponding author. Tel.:+1-315-787-2363; fax:+1-315-787-2326.

E-mail address: dck2@cornell.edu (D.C. Knipple).� Database Accession Numbers: HassGATD (AF482905), HassK-

PSE (AF482906), HassKSVE (AF482907), HassLPAQ (AF482908),HassNPVE (AF482909), HassPPAE (AF482910), HassQPGE(AF482911).

1 Present address: Department of Molecular and Life Sciences,Pohang University of Science and Technology, Pohong 790-784,South Korea.

0965-1748/03/$ - see front matter 2003 Elsevier Science Ltd. All rights reserved.doi:10.1016/S0965-1748(03)00043-2

plants and are increasingly resistant to insecticides (Corket al., 1992), the control of this pest is an important chal-lenge. Recently, an alternative strategy for the control ofH. assulta infestations has been demonstrated by usingsynthetic sex pheromones to trap adult males, eliminat-ing them before they have a chance to mate (Park etal., 1999).

Moth sex pheromones are species-specific blends ofunsaturated fatty acid (UFA) derivatives that differ interminal functional group and in the number, positionand configuration (Z or E) of their double bond(s), whichare produced by various acyl-CoA desaturases (reviewedin Roelofs, 1995; Tillman et al., 1999; Knipple and Roel-ofs, 2003). The pheromone produced byH. assultafemales consists of a mixture ofZ9-16:Ald and Z11-16:Ald in a 20:1 ratio, as opposed to the pheromone

610 S.E. Jeong et al. / Insect Biochemistry and Molecular Biology 33 (2003) 609–622

blends of the other nine members of the heliothine fam-ily that have been studied so far, where the Z11-16:Aldcomponent predominates (Arn et al., 1992). In the cornearworm Helicoverpa zea, which produces a pheromonemixture of Z11-16:Ald and Z9-16:Ald in a 30:1 ratio,the most abundant desaturase-encoding transcript isHzeaLPAQ (called HzDs1 in Rosenfield et al., 2001),which encodes a �11 desaturase. Although two �9 desa-turase-encoding transcripts are also present at about 5–10 fold lesser abundance in H. zea pheromone glands,they do not appear to contribute to the production ofunsaturated pheromone components in proportion totheir abundance (Rosenfield et al., 2001). Furthermore,analysis of pheromone biosynthesis in H. zea pheromoneglands by stable isotope labeling experiments showedthat more Z9-16:Ald is produced via �11 desaturationof stearic acid (C18:Acid) followed by β-oxidation, thanby direct �9 desaturation of palmitic acid (C16:Acid)(Choi et al., 2002). In contrast, in H. assulta pheromoneglands, no β-oxidation activity was detected and all ofthe Z9-16:Acid precursor of the major pheromonecomponent of this species appears to be produced by �9desaturation of C16:Acid (Choi et al., 2002).

H. assulta also differs from H. zea in the pheromonedetection system in male antennae. The antennae of heli-othine moths are sexually dimorphic in the sensillarregion of their flagellar subsegments, differing in totalnumber and subtypes of sensilla trichodea (Koh et al.,1995). In H. assulta, females have only one type of sen-silla trichodea, which contain receptor neurons that arepresumed to respond principally to compounds producedby host plants, whereas males have a larger number ofmorphologically and functionally distinct subtypes ofsensilla trichodea containing receptor neurons tuned tospecific components of sex pheromone blends (Massonand Mustaparta, 1990). Electrophysiological recordingsfrom single receptor neurons have shown that theantennae of H. assulta males contain three types ofreceptor neurons tuned to specific compounds, i.e., thepheromone components Z9-16:Al (45% of receptorneurons) and Z11-16:Al (10% of receptor neurons), and,in the same region as those tuned to Z9-16:Al, neuronsof a different type tuned to the behavioral antagonist Z9-14:Al (45% of receptor neurons); in contrast, themajority of neurons in the antennae of H. zea males arespecifically tuned to the major pheromone componentZ11-16:Al, none are specifically tuned to Z9-16:Al, andthose tuned to Z9-14:Al, which is not a component ofthe H. zea pheromone blend, also respond to a lesserdegree to the minor pheromone component Z9-16:Al(Berg and Mustaparta, 1995).

In this paper, we describe the isolation of seven desat-urase-encoding sequences in H. assulta pheromonegland RNA by modification of a previously describedRT/PCR screen employing degenerate primers targetingconserved desaturase sequence motifs (Knipple et al.,

2002). We present the amino acid sequences encoded bythese cDNAs and describe their homology relationshipsto the major insect desaturase sequence lineages definedin a recent study (Knipple et al., 2002). We report theresults of experiments to quantify specific desaturasetranscript levels in pheromone glands during light anddark phase and describe the functional expression ofthree cDNAs corresponding to the most abundant ones ina desaturase-deficient strain of the yeast Saccharomycescerevisiae. We discuss the roles of specific desaturasesin the production of pheromone components of this spec-ies and the closely related species H. zea.

2. Materials and methods

2.1. Insect rearing and pheromone glands dissection

Insects were reared at 25 ± 1 °C, 60% RH under 16:8(L:D) on an artificial diet (Jeong et al., 2001). Pupaewere identified by their sexually dimorphic pigmentationpattern, and females were transferred to individual vials.The moths that emerged in scotophase were collectedand the tips of their lower abdomen (including the phero-mone gland, a ring structure located in the upper part ofthe posterior segment) were dissected as described byTang et al. (1991) either 4 h after lights-off in the nextscotophase or 14 h after lights-on in the following light-phase.

2.2. RNA extraction and cDNA synthesis

Thirty-five pheromone glands (about 35 mg) wereimmersed in 700 µl of TRIzol (GIBCO/BRL) immedi-ately after dissection and kept at �80 °C until total RNAwas extracted according to the manufacturer’s instruc-tions. Double stranded cDNA was synthesized from totalRNA using the Promega Universal Riboclone cDNASynthesis System according to the manufacturer’sinstructions.

2.3. PCR screening for desaturase-encoding cDNAs

PCR experiments were performed in a Perkin-ElmerModel 480 Thermal Cycler. Double stranded cDNA wasused as template in PCR reactions to screen for the �550bp highly conserved central fragment of acyl-CoA desat-urases, by a modification of a previously described hom-ology probing procedure (Rosenfield et al., 2001). Allnucleotide sequences encoding the desaturaseGAHRLW(A/T/S) target sequence motif were synthe-sized as 16 discrete degenerate primer pools (5� GAHRprimers) and used in PCR reactions in combination witha single 3� degenerate primer containing all nucleotidesequences of the strand opposite that encoding theEGFHNYH target sequence motif (3� EGFH primer).

611S.E. Jeong et al. / Insect Biochemistry and Molecular Biology 33 (2003) 609–622

Each 50 µl PCR screening reaction contained 100 ngcDNA template, 0.2 mM dNTP’s, 2 mM MgCl2, 0.8 µM5� GAHR and 4.0 uM 3� EGFH primers. A hot start (5min preincubation at 95 °C, then addition at 80 °C of2U Taq polymerase (Promega H1661) to start thereaction) was followed by 35 PCR cycles (one screenusing 95 °C, 1 min; 60 °C, 1 min; 72 °C, 2 min and asecond using 95 °C, 1 min; 55 °C, 1 min; 72 °C, 2 min).Amplification products were analyzed by electrophoresisin agarose gels. Specific amplification products wereexcised and eluted from gels, ligated into plasmidpCR2.1, cloned in E. coli using a TA cloning kit(Invitrogen), and sequenced by a variation of the Sangermethod (Sanger et al., 1977) using a Model 377 AppliedBiosystems Automated DNA Sequencer and the Prismdye-labeled dideoxynucleotide terminator chemistryaccording to the manufacturer’s protocol.

2.4. Protein sequence analysis

Deduced amino acid sequences were analyzed withtools contained in the MacVector 7.0 software suite(Oxford Molecular Ltd). Multiple sequence alignmentsof deduced amino acid sequences were obtained by usingthe ClustalW (1.4) algorithm (Thompson et al., 1994).Hydropathy plots (Kyte and Doolittle, 1982) obtainedwith the McVector protein analysis toolbox were scaledwith Adobe Illustrator 9.0 to allow direct comparisonsbetween full length and partial protein sequences.

2.5. Measurement of desaturase transcript levels inpheromone glands

A pair of gene-specific primers (GSP) was designedfor each of the seven H. assulta desaturase sequences(Table 1), synthesized with a Perkin-Elmer Model 392DNA Synthesizer, desalted, and tested for specificityagainst high concentrations of plasmids bearing the other(non-target) acyl-CoA desaturase-encoding cDNAs.PCR reactions were performed as above, using 0.4 µMeach of GSP1 and GSP2, and dilutions of plasmid con-taining the target template to be quantified. Once the sel-ectivity of the reaction and the linear range of productformation were established for each combination ofGSPs and cDNA template, PCR reactions were perfor-med for 30 cycles (39 cycles for weakly transcribedsequences) using 10-fold serial dilutions of the cDNApreparations over three or more orders of magnitude.The abundance of each specific H. assulta acyl-CoAdesaturase-encoding transcript in total RNA isolatedfrom pheromone glands was determined by comparingthe product yields from serial dilutions of the cDNA andthe target template. Primers developed to amplify H. zeaβ-actin (Rosenfield et al., 2001) were used in PCR reac-tions with dilutions of H. assulta cDNA template to nor-malize template concentrations.

2.6. Rapid amplification of cDNA ends (RACE)

To obtain the full length sequences of the desaturaseclones, pairs of PCR primers were designed with highcalculated melting temperature (Tm) (Table 1) to ampl-ify the 5� and 3� ends of each specific desaturase cDNAby the RACE procedure (Frohman et al., 1988) using aMarathon cDNA amplification kit (Clontech) andAdvantage cDNA polymerase (Clontech) according tothe manufacturer’s protocols. Amplification productswere gel-purified, cloned into pCR2.1 and sequencedas above.

2.7. Plasmid constructions

The full length coding regions of HassNPVE, HassK-PSE, HassLPAQ, and HassGATD were cloned into theexpression vector YEpOLEX as follows. In a first set ofPCR reactions, two overlapping portions of each openreading frame (ORF) were produced by PCR. The 5�-ORF was amplified using the BamHI-ATG and GSP2primers (Table 1) and a pCR2.1 plasmid template con-taining a 5�-RACE fragment with consensus sequence.Similarly, the 3�-ORF was produced with the GSP1 andSacI-end primers and a template containing a 3�-RACEfragment. PCR mixtures (50 µl) containing 5 ng plasmidtemplate, 0.2 mM dNTP’s, 0.2 µM primers and 1 ul highfidelity Advantage 2 Polymerase Mix (ClonTech #8430-1). The polymerase was activated by a 1 min incubationat 94 °C and 15 PCR cycles (94 °C, 1 min; primers’calculated annealing temperature, 1 min; 72 °C, 2 min)were performed.

In a second PCR reaction, 5 ng gel-purified 5�-ORFand 3�-ORF fragments were combined, heated at 94 °Cfor 2 min, cooled quickly on ice, and used as comp-lementary overlapping templates. The polymerase, acti-vated by 1 min at 94 °C, extended (without primers) theoverlapping templates for 7 min at 72 °C; then, after 1min at 94 °C and addition at 80 °C of the primersBamHI-ATG and SacI-end (Table 1), 15 PCR cycleswere performed as above.

The ORF, delineated by BamHI and SacI restrictionsites, was gel-purified, cloned into pCR2.1, and verifiedby DNA sequencing. A plasmid with the correctsequence was digested with BamHI and SacI enzymesand the ORF was directionally inserted by replacementcloning into the corresponding sites of YEpOLEX-CLR7(Knipple et al., 1998). The predicted sequences of thefinal expression plasmids, designated YEpOLEX-HassNPVE, YEpOLEX-HassKPSE, YEpOLEX-HassLPAQ and YEpOLEX-HassGATD, were verifiedby DNA sequencing.

2.8. Functional expression of H. assulta pheromonegland-specific desaturases

Cells of the L8-14C desaturase-deficient ole1 strainof S. cerevisieae (MATa, ole1�:LEU2, leu2-3, leu2-112,

612 S.E. Jeong et al. / Insect Biochemistry and Molecular Biology 33 (2003) 609–622

Table 1Gene-specific primers used in this study

Desaturase Primer Primer sequence from 5� to 3�

HassKPSE GSP1 TCCGAGTTATCCTGGTCATCGSP2 CTTGTCGTAGGGCTTGTCTC3�RACE GGGACTCGACATGAGCGACTTACTTG5�RACE GGATAAACGCGTAGCGGAACATGGCAGBamHI-ATG GGGGGATCCATGGCTCCAAATATATCGGASacI-end CCCGAGCTCTTAATCGTCCTTAGGATTGA

HassNPVE GSP1 GCTGACTCTCTTCAACACCCGSP2 GCACGACGAGGGAGACGG3�RACE CCTCAGATCAAGGCTAAGGGACACAC5�RACE CATTGAGCACGTACACGTAGCGGAAGABamHI-ATG GGGGGATCCATGCCTCCTCAAGGGCAAACTSacI-end GGGGAGCTCTTATTCAGTTTTCTCTGGGTTAATA

HassLPAQ GSP1 TGAGGGACCATCGTCTCCATGSP2 CACTGCTACATTTTGGGCAG3�RACE CCAGAAGTCAAAAAGCGTGGGAAAGAAC5�RACE GTGGCTACTGACACTGCTACATTTTGGGBamHI-ATG GGGGGATCCATGGCCCAAAGCTATCAATCSacI-end GGGGAGCTCTTAACTTAATTTATCGTTTACAC

HassGATD GSP1 GCTTTCCAGAACCACATCTACGSP2 AGCCACCATCTTGTTGTCAG3�RACE TCACCAGATACACAATAGCCCTTCACTT5�RACE CACAGATAGCCACCATCTTGTTGTCAGTBglII-ATGa GGGAGATCTATGGCGGCGATGTCGAGCACTSacI-end GGGGAGCTCCTATTCAGTTTTTCTGTGAAC

HassKSVE GSP1 CTATTACTGGACAGCGGGACGSP2 AAGCTATGTTCAGCGTGATGG

HassQPGE GSP1 AACTACCACTCCGCATTCTACGSP2 GTGAGCTACTGAGTTGACGAC

HassPPAE GSP1 GTGCTAGTCAGACATCAGTCGSP2 AGACACGTACCAGGATTCTG

a A BglII recognition site was chosen because of an internal BamHI recognition site in the GATD ORF. The cohesive end produced by BglIIdigestion is complementary to that produced by BamHI.

trp1-1, ura3-52, his4) (Stukey et al., 1990) were trans-formed with the above YEpOLEX plasmids by using analkali-cation yeast transformation kit (Bio 101 Inc.,Vista, CA) and selecting for URA+ phenotype. Individualtransformant colonies were then tested for comp-lementation of the UFA auxotrophy of the host strain asdescribed previously (Knipple et al., 1998). UFA+ trans-formants were grown in 10 ml complete (yeastextract/peptone/dextrose) liquid medium in 50 ml Erlen-meyer flasks shaken at 300 rpm in a New BrunswickG76 shaker at 15 °C (at which maximum induction ofthe OLE1 promotor used to drive the transcription of thecDNA expression constructs occurs) and at 30 °C. Cellswere grown to a density of approximately 2.0 × 107

cells/ml, pelleted and washed with water, after whichlipids were extracted with chloroform/methanol (2:1,vol/vol), and then methylated with 0.5 MKOH/methanol. The fatty acid methyl esters were ana-lyzed by GC/MS, using a Hewlett Packard 5890 gaschromatograph (HP1 capillary column, 25 × 0.2 mm ID,0.5 µm film thickness) coupled to an HP 5970 seriesMass Selective Detector. The oven temperature was pro-

grammed at 125 °C for 3 min, then 5 °C/min until 245°C and held for 32 min. The mass amounts of individualfatty acid methyl esters were estimated relative to knownmass quantities of nonadecanoic acid methyl ester(C19:Me) that were added to the washed cell pelletsprior to the extraction procedure. The amounts of 16-and 18-carbon UFAs produced in transformants express-ing moth desaturases were measured by integration oftheir peak areas in GC chromatograms and standardizedrelative to the peak areas of the 19:Me internal standard.The double bond positions of unsaturated fatty acidmethyl esters were verified by mass spectral analysis ofdimethyl disulfide adducts (Buser et al., 1983).

3. Results

3.1. Amplification of cDNAs encoding members of theacyl-CoA desaturase sequence family of H. assulta

In the initial screen using cDNA template synthesizedfrom H. assulta pheromone gland RNA, PCR reactions

613S.E. Jeong et al. / Insect Biochemistry and Molecular Biology 33 (2003) 609–622

containing 12 of the 16 5� GAHR primers paired withthe 3� EGFH primer were found to produce the predicted�550 bp amplification product. The PCR products fromeach reaction were cloned, sequenced, and the deducedamino acid sequences were aligned with those of pre-viously characterized desaturases. The initial screeningof more than 100 clones did not yield any clone withhigh sequence identity to a predicted orthologue of theHzeaDs1 cDNA (here called HzeaLPAQ) encoding a�11 desaturase (Rosenfield et al., 2001). Consequently,a second screening was performed under reduced strin-gency by lowering the temperature of the annealingphase of the PCR reactions from 60 to 55 °C.

The sequences of 68 clones isolated in the secondscreening were found to comprise seven categoriesnamed according to a recently formalized nomenclature(Knipple et al., 2002): HassNPVE (22 clones), HassK-PSE (23 clones), HassLPAQ (7 clones), HassQPGE (7clones), HassGATD (5 clones), HassKSVE (3 clones),and HassPPAE (1 clone). Sequences such as HassNPVEand HassKPSE were obtained from every reaction thatproduced amplification products, whereas othersequences were found in only a few reactions: e.g., allthree HassKSVE clones were obtained from reactionscontaining the GAHR15 primer, and the sevenHassQPGE clones were obtained from reactions contain-ing the primers GAHR11 and 13. The four reactions thatfailed to produce an amplification product in the initialscreen also failed in the second screen and had in com-mon 5� GAHR primers with 3� AG, consistent with theabsence of desaturase-encoding sequences in H. assultapheromone glands containing GAHRLWS, where theserine is encoded by AG(A or G).

Fig. 1 presents an alignment of the seven H. assultasequences together with four H. zea orthologues. Thecentral region, obtained in the screening with the GAHRand EGFHNYH primers, is highly conserved and hastypical features of desaturase protein sequences: threehistidine boxes H1, H2 and H3, containing eight histi-dine residues essential for catalysis (Shanklin et al.,1994) and two of four transmembrane domains (TM3and TM4). An alignment gap at positions 170 and 171is caused by the presence of two additional amino acidsin HassKSVE and is extended two amino acids upstream(positions 168, 169) in both QPGE sequences. Other sin-gle amino acid gaps occur in the HassNPVE andHassQPGE sequences (position 211) and in theHassLPAQ sequences (position 225). The orthology ofHassNPVE, HassKPSE, and HassLPAQ to the threedesaturase sequence lineages designated NPVE, KPSE,and xxxQ (Knipple et al., 2002) is indicated by the boot-strapped tree in Fig. 2, as well as by the extremely highsequence identity scores with members of those lineagesisolated from H. zea (Fig. 1, Table 2). Interestingly, theHassQPGE and HzeaQPGE sequences define a discretelineage that, to date, has no known orthologue in any

other species (Knipple et al., 2002). Differences in theamino sequences between HassKPSE and HzeaKPSEoccur at only one position (Fig. 1: Q58L); betweenHassNPVE and HzeaNPVE at three positions (Fig. 1:T27A, H192Y, E361A); between HassLPAQ andHzeaLPAQ at 11 positions (Fig. 1: P26L, T29Q, G241S,F302L, and seven within the last 13 amino acids of the3� end); and, between HassQPGE and HzeaQPGE atthree positions (Fig. 1: N191K, L263V, and L265F).Inspections of the corresponding nucleotide sequencealignments identified 30 synonymous substitutionsbetween HassKPSE and HzeaKPSE, 21 betweenHassNPVE and HzeaNPVE, 43 between HassLPAQ andHzeaLPAQ, and 19 between HassQPGE and HzeaQPGE(data not shown).

3.2. Desaturase transcript levels in H. assultapheromone glands

The steady state levels of the seven H. assulta desatur-ase mRNAs in pheromone glands of sexually maturefemales were estimated by RT-PCR over a range of tem-plate concentrations using gene-specific primers (Table1; Fig. 3). In parallel PCR reactions we showed that thecDNAs synthesized from pheromone gland RNAsampled in scotophase and in light phase contain similarlevels of Hassb-actin transcript, thus permitting directcomparisons of specific transcript levels between the twotime periods. The panels in Fig. 3 were ordered fromtop to bottom from highest to lowest transcript levels(sequences in Fig. 1 are also listed in this order). Thus,the most abundant of the seven acyl-CoA desaturasetranscripts found in the H. assulta pheromone gland wasHassNPVE (panel 1). Its diagnostic 463 bp cDNAamplification product was obtained in the linear rangeof product yield by using 10�5 and 10�6 cDNA dilutions,whereas the 433 bp product HassKPSE (panel 2) andthe 400 bp HassLPAQ (panel 3) were both detectedusing 10�3 and 10�4 dilutions of pheromone glandcDNA. The level of HassNPVE transcripts was about12 times higher than that of HassKPSE, and the latterwas about three times more abundant than that ofHassLPAQ. The inferred levels of the other four desatur-ase-encoding transcripts were quite low. In fact, thosesequences were only detected using higher amounts ofpheromone gland cDNA (10�2 and 10�3 dilutions).Compared to the transcript level of HassKPSE, that ofHassGATD (panel 4, 434 bp) was 15 times lower; thoseof HassKSVE (panel 5, 362 bp), HassQPGE (panel 6,406 bp), and HassPPAE (panel 7; 436 bp) were 60, 600,and 3000 times lower, respectively. Each of the sevendesaturase transcripts were found in similar amounts inpheromone glands sampled 4 h after the onset of the firstscotophase (lanes 5 and 6) and 14 h into the followinglight phase (lanes 7 and 8).

614 S.E. Jeong et al. / Insect Biochemistry and Molecular Biology 33 (2003) 609–622

Fig. 1. Alignment of deduced amino acid sequences of seven cDNAs from H. assulta (this study) and four from H. zea (Rosenfield et al., 2001).The sites used to design degenerate primers for PCR are shown in bold. Gray backgrounds highlight identities (dark) or conservative substitutions(light). Amino acids of predicted transmembrane domains (TM1, TM2, TM3, and TM4) are labeled underneath the alignment. Three essentialhistidine boxes (H1, H2, and H3) as well as the signature motif at positions 256–259 are framed.

615S.E. Jeong et al. / Insect Biochemistry and Molecular Biology 33 (2003) 609–622

3.3. Isolation and characterization of full lengthcDNAs of the four most abundant desaturase-encodingtranscripts in H. assulta pheromone gland

The 5� and 3� ends of HassNPVE, HassKPSE,HassLPAQ and HassGATD clones were obtained byRACE (Fig. 4), but efforts to obtain those of the otherthree sequences (HassKSVE, HassQPGE andHassPPAE) by this procedure were unsuccessful due tothe extremely low levels of their transcripts. The ORFsof HassNPVE, HassKPSE, and HassLPAQ were thesame lengths as their H. zea orthologues: 1059 nucleo-tides (353 amino acids) for the former two, and 1014nucleotides (338 amino acids) for the latter. HassNPVEand HassLPAQ had short 3�UTRs (about 400 bp),whereas HassKPSE and HassGATD had 3�UTRs longerthan 1 kb. HassGATD had the longest ORF (1116nucleotides, 372 amino acids), with 16 more amino acidsthan the others near its 3�-end (Fig. 1, positions 331–346).

3.4. Determination of protein hydropathy with Kyte–Doolittle plots

Hydropathy plots (Kyte and Doolittle, 1982) of thefull length and partial protein sequences deduced fromthe cDNAs isolated in this study were determined andaligned (Fig. 5). At a gross level, the locations of fourpostulated transmembrane domains of the full lengthsequences are conserved. However, a closer examinationof all seven sequences reveals significant differences inthe precise location of TM4, resulting from variations inthe length of the hydrophilic interval separating TM3and TM4, which has been identified as a functional

Fig. 2. Consensus bootstrap tree showing homology relationships ofseven H. assulta desaturase amino acid sequences of the present study(highlighted) and other characterized lepidopteran desaturasesequences from the corn earworm Helicoverpa zea (family Noctuidae)(Rosenfield et al., 2001), the cabbage looper Trichoplusia ni (familyNoctuidae) (Knipple et al., 1998; Liu et al., 1999), the cabbage army-worm Mamestra brassica (family Noctuidae), the common armywormPseudaletia separata (family Noctuidae), the beet armyworm Spodop-tera exigua (family Noctuidae), the common cutworm Spodoptera lit-ura (family Noctuidae), the Indian meal moth Plodia interpunctella(family Pyralidae), the pink bollworm Pectinophora gossypiella(family Gelechiidae) (Knipple et al., 2002), the silk moth Bombyx mori(family Bombycidae) (Yoshiga et al., 2000), the codling moth Cydiapomonella (family Tortricidae) (Knipple et al., 2002), a New Zealandleafroller species, Planotortrix octo (family Tortricidae) (Hao et al.,2002), the European corn borer Ostrinia nubilalis (family Tortricidae),the Asian corn borer Ostrinia furnacalis (family Tortricidae) (Roelofset al., 2002), the light brown apple moth Epiphyas postvittana (familyTortricidae) (Liu et al., 2002a), and the redbanded leafroller Argyrota-enia velutinana (family Tortricidae) (Liu et al., 2002b). Branch points(internal nodes) are retained if they occur in �50% of resampling trees(1000 × resampling); all other nodes are collapsed. Lineage andsequence names are according to the nomenclature of Knipple et al.(2002).

616 S.E. Jeong et al. / Insect Biochemistry and Molecular Biology 33 (2003) 609–622

Table 2Comparisons of partial amino acid sequences of seven H. assulta desaturases and seven other lepidopteran desaturasesa

Desaturase Hzea NPVEb Tni NPVEc Hzea KPSEb Epos KPSEd Hzea LPAQb Tni LPAQc Hzea QPGEb

HassNPVE 98 (98) 93 (96) 71 (85) 74 (86) 60 (77) 61 (79) 56 (76)HassKPSE 72 (86) 73 (86) 100 (100) 89 (96) 62 (82) 65 (83) 54 (72)HassLPAQ 60 (77) 61 (78) 64 (82) 66 (81) 99 (99) 79 (90) 51 (71)HassGATD 57 (77) 57 (77) 62 (78) 59 (77) 51 (74) 51 (72) 49 (66)HassKSVE 48 (68) 48 (68) 47 (67) 48 (65) 48 (65) 45 (65) 45 (64)HassQPGE 54 (76) 55 (76) 52 (72) 55 (72) 49 (70) 51 (71) 97 (98)HassPPAE 50 (66) 50 (66) 48 (68) 50 (68) 50 (65) 50 (65) 49 (66)

a Amino acid sequences delimited by the GAHR and EGFHNYH motifs were compared in a McVector Clustal W(1.4) multiple sequencealignment. Percent identities are followed by percent similarities (sum of identities and similarities) in parentheses. Values shown in bold indicatecomparisons of orthologous sequences.

b Sequences from H. zea: HzeaNPVE, HzeaKPSE, HzeaLPAQ and HzeaQPGE (Rosenfield et al., 2001).c Sequences from T. ni: TniLPAQ (Knipple et al., 1998) and TniNPVE (Liu et al., 1999).d Sequences from E. postvittana: EposKPSE (Liu et al., 2002a).

class-specifying feature of insect desaturases (Knipple etal., 2002).

3.5. Functional expression of H. assulta desaturase-encoding cDNAs

Desaturase-deficient yeast cells (ole1) transformedwith YEpOLEX plasmids containing HassNPVE,HassKPSE, and HassLPAQ cDNAs were able to growon media lacking supplemental UFAs, indicating that theplasmids were expressing desaturases that comp-lemented the UFA auxotrophy of the ole1 host cells; onthe other hand, YEpOLEX-HassGATD transformantswere unable to grow on media lacking supplementalUFAs. The identities and ratios of 16- and 18-carbonUFAs produced by the functionally expressed H. assultadesaturases were determined by GC-MS analysis, alongwith those produced by desaturases isolated previouslyfrom T. ni (Knipple et al., 1998; Liu et al., 1999), H. zea(Rosenfield et al., 2001), and the yeast OLE1 desaturase(Stukey et al., 1990) (data not shown; summarized inTable 3). GC chromatograms of methylated total lipidextracts of ole1 yeast cells expressing HassNPVE,HzeaNPVE, and TniNPVE were similar to each other,having two peaks with retention times corresponding toZ9-16:Me and Z9-18:Me, with the latter predominating.GC chromatograms of methylated total lipid extracts ofole1 yeast cells expressing HassKPSE were similar tothose obtained from cells expressing HzeaKPSE, i.e.,they had two peaks corresponding to Z9-16:Me and Z9-18:Me, with the former predominating. GC-chromatog-rams of methylated total lipid extracts of ole1 yeast cellsexpressing HassLPAQ and HzeaLPAQ were nearlyidentical, containing one peak with retention time corre-sponding to Z11-16:Me. In contrast, an additional peakcorresponding to Z11-18:Me was found in yeast cellsexpressing TniLPAQ, as reported previously (Knipple etal., 1998).

4. Discussion

In this paper we describe the isolation and characteriz-ation of cDNAs corresponding to seven members of thedesaturase sequence family expressed in the pheromonegland and surrounding tissues of H. assulta. In namingthese sequences (and others that we discuss here), weuse a recently formalized nomenclature (Knipple et al.,2002), which designates each inferred desaturase proteinsequence with an abbreviated Linnean name in conjunc-tion with a “signature motif” of four amino acids (aminoacid coordinates 256–259 in Fig. 1). This nomenclatureavoids ambiguity associated with the use of abbrevi-ations based on nonsystematic common names of speciesand emphasizes sequence homology relationships, whichare correlated with conserved biochemical properties insome lineages, e.g., the substrate chain length prefer-ences of the �9 desaturases of the NPVE and KPSE lin-eages (Fig. 2; Table 4). It also dispenses with the use ofabbreviations designating the tissue source from whichdesaturase-encoding cDNAs were isolated, which can bemisleading. In regard to the latter point, the tissue distri-butions of specific desaturase-encoding transcripts havebeen examined in only a few insect species and theseshow complex spatial and temporal patterns ofexpression (Liu et al., 1999; Yoshiga et al., 2000; Rosen-field et al., 2001). Furthermore, whereas the presence ofa desaturase transcript is a prerequisite for the functionalactivity of its encoded enzyme, the presence of a tran-script in any given tissue is not in itself proof of a func-tional role for its encoded protein in that tissue (Knippleand Roelofs, 2003). Thus, the nomenclature that we useand advocate here provides significant practical and con-ceptual advantages over earlier ad hoc naming schemes.

Six of the seven sequences isolated in this study aremembers of recently described lepidopteran desaturasesequence lineages (Knipple et al., 2002) (Fig. 2). Four,i.e., HassNPVE, HassKPSE, HassGATD, and

617S.E. Jeong et al. / Insect Biochemistry and Molecular Biology 33 (2003) 609–622

HassKSVE, belong to highly conserved lineages,whereas two others, HassLPAQ and HassPPAE, belongto more divergent lineages. The seventh sequence,HassQPGE, has only one known lepidopteranorthologue, i.e., HzeaQPGE from H. zea (previouslydesignated HzPGDs3) (Rosenfield et al., 2001).Extremely high amino acid sequence identities betweenH. zea desaturases (Rosenfield et al., 2001) and their H.assulta orthologues (Fig. 1; Table 2), and identical β-actin protein sequences (encoded by cDNAs differing byonly two synonymous substitutions; data not shown)strongly suggest a very recent divergence of these spec-ies from a common ancestor. In fact, basal heliothinedivergence is estimated to have occurred no more than20 million years ago (Cho et al., 1995) and phylogeneticanalyses of heliothine species using comparisons ofrDNA (Friedlander et al., 1995) and a number of singlecopy genes (Friedlander et al., 1995; Cho et al., 1995;Fang et al., 1997) have failed to resolve some relation-ships in this large subfamily due to extremely highsequence conservation.

Orthologues of HassGATD, HassKSVE, andHassPPAE were isolated from pheromone glands of sev-eral other lepidopteran species (Knipple et al., 2002), butnone was obtained from pheromone glands of H. zea(Rosenfield et al., 2001), strongly suggesting that thesewere not detected in this earlier study. The differentresults obtained in the high and low stringency screen-ings of the present study demonstrate that the nature ofthe degenerate primers used for homology probing andthe temperature of the annealing phase of the PCR reac-tions can profoundly influence the results of this pro-cedure. On the one hand, GAHR primers can act pro-miscuously when a particular desaturase transcript isabundant, yielding amplification products from severaldegenerate primers that are mismatched with the desatur-ase DNA target sequence, e.g., in the case of HassNPVEand HassKPSE; on the other hand, primer specificityappears to increase at lower transcript levels, e.g., asillustrated by the selectivity of GAHR15 for HassQPGE.Other results indicate that some of the GAHR primersare less effective than others in annealing to complemen-tary sequences. For example, in our initial high strin-

Fig. 3. Estimation of desaturase transcript levels by RT-PCR. PCRproducts derived from specific desaturase-encoding transcripts andHassb-actin transcripts were separated by agarose gel electrophoresis(panels 1 through 7 and panel 8, respectively). In lanes 1–4 of eachpanel a linear range of amplification was obtained by using sequential10-fold dilutions of known amounts of each desaturase- or Hassb-actin-encoding target sequences. Lanes 5 and 6, and lanes 7 and 8,respectively, show the amplification products obtained using dilutionsof cDNA templates prepared from pheromone gland RNA sampledeither 4 h after the onset of the first scotophase or 14 h into the follow-ing light phase. The estimated mass of each sequence in 1 µl of cDNAsolution (corresponding to about 100 ng total RNA) is indicated at thelower left of each panel.

618 S.E. Jeong et al. / Insect Biochemistry and Molecular Biology 33 (2003) 609–622

Fig. 4. Diagram of the full-length cDNA clones of HassNPVE, HassKPSE, HassLPAQ, and HassGATD. The sizes of the 5� untranslated ends,open reading frames, and 3� untranslated ends are indicated, as well as the locations and 5�-to-3� orientations of all primers used in PCR (Table 1)

gency screening, reactions using the GAHR2 primer,which contains sequences with perfect complementarityto target sequences of HassLPAQ and HassNPVE,yielded no clones of the former sequence and a muchlower number of clones of the latter than would beexpected on the basis of its high relative abundance inH. assulta pheromone glands (Fig. 3).

Transcript quantitation and functional expressionexperiments indicate that the three most abundant desat-urase transcripts in the H. assulta pheromone gland,HassNPVE, HassKPSE, and HassLPAQ, encode func-tional enzymes with properties consistent with their rolein the synthesis of the major and minor UFA precursorsof this species’ pheromone blend. Transcript levels ofthese (as well as the other four less abundant) desaturase-encoding sequences in RNA extracts of pheromoneglands collected 14 h into light phase and in mid-scoto-phase did not differ significantly (Fig. 3). Although fluc-tuations in transcript levels between these two timepoints cannot be ruled out, the results are consistent withsustained desaturase transcript levels throughout themoth diurnal cycle. If this is the case, the hiatus in phero-mone production during light phase (Park et al., 1996)is not regulated at the level of desaturase gene transcrip-tion, but rather must occur downstream of this process,consistent with the postulated regulation of pheromoneproduction at the level of the last enzymatic steps ofpheromone biosynthesis in this species (Choi et al.,1998).

Analysis of UFAs in the yeast functional expressionassay showed that the HassNPVE and HassKPSE desat-urases had substrate chain length preferences consistentwith the conserved biochemical properties of the desat-urases of their respective sequence lineages (Knipple etal., 2002; Knipple and Roelofs, 2003) (Fig. 2; Tables 3and 4). However, the magnitude of the differencesobserved in the C16/C18 UFA ratios of yeast cells

expressing orthologous �9 desaturases of H. zea and H.assulta (Table 3) was unexpected in light of theirextremely high amino acid sequence identities. Further-more, we observed that HassKPSE transformants grewsignificantly faster at 15 °C than cells expressingHzeaKPSE (data not shown), although these two desat-urases differ by only a single amino acid. These pheno-typic differences between transformant strains express-ing nearly identical proteins could reflect the effects ofsynonymous substitutions, resulting in some mRNAsthat are suboptimal in the context of the translationalprocesses of S. cerevisiae. Although the transcript levelof HassNPVE was found to be an order of magnitudegreater than that of HassKPSE in RNA extracted fromthe abdominal tip containing the pheromone gland (Fig.3), the substrate chain length preference of the HassK-PSE desaturase (Table 3) is more congruent with a pri-mary role in the production of the Z9-16:Acid phero-mone precursor. Thus, it cannot be ascertained fromthese results whether HassNPVE, HassKPSE, or bothare involved in the production of the Z9-16:Acid precur-sor of the H. assulta major pheromone component. Thethird most abundant desaturase transcript in the H.assulta pheromone gland, HassLPAQ, encodes a �11desaturase that catalyzes the formation of the Z11-16:Acid precursor of the minor component of the H.assulta pheromone blend. Its orthologue in H. zea,HzeaLPAQ, has indistinguishable enzymatic propertiesand is encoded by the most abundant desaturase-enco-ding transcript in that species’ pheromone gland(Rosenfield et al., 2001).

Additional insights into the functional roles of �9 and�11 desaturases in pheromone biosynthesis in H. zeaand H. assulta are provided by a recent investigationusing pheromone gland explants supplemented with deu-terium-labeled UFAs (Choi et al., 2002). In H. assultapheromone glands the UFA precursors of the major

619S.E. Jeong et al. / Insect Biochemistry and Molecular Biology 33 (2003) 609–622

Fig. 5. Kyte-Doolittle hydropathy plots of HassNPVE, HassKPSE,HassLPAQ, and HassGATD (full length proteins), and HassKSVE,HassQPGE, and HassPPAE (partial amino acid sequences delimitedby GAHR and EGFHNYH motifs). Plots were scaled with Adobe Illus-trator 9.0 to allow direct comparison between full length and partialprotein sequences and aligned at the start of the third transmembranedomain. Vertical gray bands delineate the locations of the GAHR andEGFHNYH motifs and the region containing the transmembranedomains TM3 and TM4.

pheromone component, Z9-16:Ald, and the minor phero-mone component, Z11-16:Ald, were found to be pro-duced directly by �9 and �11 desaturation of C16:Acid,respectively, consistent with the finding of the presentinvestigation. In H. zea pheromone glands, the UFA pre-cursor of the major pheromone component Z11-16:Aldwas found to be produced by �11 desaturation ofC16:Acid, as in H. assulta, but the precursor of this spe-cies’ minor pheromone component, Z9-16:Ald, wasfound to be produced predominantly by �11 desaturationof C18:Acid followed by β-oxidation. This result is

Table 3C16/C18 ratios of Z9 or Z11 UFAs produced in genetically transfor-med ole1 yeast cellsa expressing �9 or �11 desaturases from H. assul-tab, H. zeac, and T. nid,e

Desaturase 15 °C 30 °C

Z9-16 Z11-1+ Z9-16 Z11-16Z9-18 Z11-18 Z9-18 Z11-18

HassNPVEb 0.45 – 0.21 –HzeaNPVEc 0.40 – 0.11 –TniNPVEd 0.38 – 0.39 –HassKPSEb 2.70 – 1.38 –HzeaKPSEc 3.04 – 2.88 –Ole 4.8 2.17 – 2.55 –HassLPAQb – � – �HzeaLPAQc – � – �TniLPAQe – 2.50 – 2.46

a Cultures were grown aerobically at 15 °C and 30 °C in completeliquid medium lacking UFAs. Cells were harvested at a density of2.0×107 cells/ml, lipids were extracted, and fatty acid methyl esterswere analyzed by GC/MS.

b This report.c Rosenfield et al. (2001).d Liu et al. (1999).e Knipple et al. (1998).

inconsistent with the finding that ole1 yeast cellsexpressing HzeaLPAQ produce no detectable Z11-18:Acid, although they contain significant quantities ofC18:Acid (Rosenfield et al., 2001). In H. assulta phero-mone glands, no β-oxidation activity was detected (Choiet al., 2002). However, the occurrence of Z7-16:Al andZ7-16:Ac as minor components of pheromone blends ofH. assulta from Japan (Sugie et al., 1991) suggests thatresidual β-oxidation of the Z9-18:Acid precursor prob-ably occurs in this species in some geographic regions.

The results of transcript quantitation experiments indi-cate that HassGATD, HassKSVE, HassPPAE, andHassQPGE are present in pheromone glands at very lowlevels (Fig. 3). A full length sequence of the most abun-dant of these, HassGATD, failed to complement theUFA auxotrophy of the ole1 mutant yeast strain and wewere unsuccessful in obtaining full length codingsequences of the other three for functional expressionanalysis. Despite the absence of evidence of functionfrom the present investigation, analyses of the ratios ofnonsynonymous and synonymous nucleotide substi-tutions (Yang and Nielsen, 1998) of the desaturase lin-eages to which these sequences belong indicate that allare evolving under purifying selection and therefore arelikely to encode functional proteins (Knipple et al.,2002). Thus, the rarity in pheromone glands of the tran-scripts encoding these inferred proteins indicates thatthey are unlikely to be involved in sex pheromonebiosynthesis in this species, although functional roles inother tissues or biochemical pathways cannot be pre-cluded on the basis of present evidence.

620 S.E. Jeong et al. / Insect Biochemistry and Molecular Biology 33 (2003) 609–622

Tab

le4

Com

pari

sons

ofal

igne

dam

ino

acid

sequ

ence

sde

duce

dfr

om�

9de

satu

rase

cDN

As

obta

ined

from

pher

omon

egl

and

(PG

)an

dfr

omfa

tbo

dy(F

B)

ofei

ght

lepi

dopt

eran

spec

ies

from

thre

efa

mili

esa

Sour

cena

me

Syst

emat

icSu

bstr

ate

%am

ino

acid

iden

titie

sbe

twee

nfu

llle

ngth

desa

tura

sese

quen

ces

nam

eipr

efer

ence

Hze

aH

ass

Ofu

rO

nub

Epo

sPo

ctA

vel

Tni

Hze

aH

ass

Epo

sO

fur

Onu

bK

PSE

KPS

EK

PSE

KPS

EK

PSE

KPS

EK

PSE

NPV

EN

PVE

NPV

EN

PVE

NPV

EN

PVE

HzP

GD

s2b

Hze

aKPS

EC

16�

C18

–64

6262

6564

6410

086

8685

8384

Has

sKPS

Ec

Has

sKPS

EC

16�

C18

–64

6262

6564

6410

086

8685

8384

Ofu

916

(FB

)dO

furK

PSE

C16

�C

18–

6563

6365

6363

8686

9986

8486

Onu

916

(FB

)dO

nubK

PSE

C16

�C

18–

6563

6365

6363

8686

9986

8485

LB

AM

-FB

Z9e

Epo

sKPS

EC

16�

C18

–66

6464

6763

6385

8586

8693

94P

.oc

to-F

BZ

9fP

octK

PSE

C16

�C

18–

6463

6366

6262

8383

8484

9393

RB

LR

-FB

Z9g

Ave

lKPS

EC

16�

C18

–65

6363

6763

6384

8486

8594

93T

.ni

FBZ

9hT

niN

PVE

C18

�C

1664

6465

6566

6465

–93

9281

7373

HzF

BZ

9bH

zeaN

PVE

C18

�C

1662

6263

6364

6363

–93

9880

7373

Has

sNPV

Ec

Has

sNPV

EC

18�

C16

6262

6363

6463

63–

9298

8073

73L

BA

M-P

GZ

9eE

posN

PVE

C18

�C

1665

6565

6567

6667

–81

8080

7272

Ofu

918

(PG

)dO

furN

PVE

C18

�C

1664

6463

6363

6263

–73

7373

7210

0O

nu91

8(P

G)d

Onu

bNPV

EC

18�

C16

6464

6363

6362

63–

7373

7372

100

aC

ompa

riso

nsof

orth

olog

ous

sequ

ence

sar

esh

aded

;th

ose

ofH

.as

sult

aan

dH

.ze

aar

ebo

ld.

Com

pari

sons

ofpa

ralo

gous

sequ

ence

sar

eun

shad

ed.

bR

osen

field

etal

.(2

001)

.c

Thi

sre

port

.d

Roe

lofs

etal

.(2

002)

.e

Liu

etal

.(2

002a

).f

Hao

etal

.(2

002)

.g

Liu

etal

.(2

002b

).h

Liu

etal

.(1

999)

.i

Aft

erK

nipp

leet

al.

(200

2):

Ave

l=A

rgyr

otae

nia

velu

tina

na(T

ortr

icid

ae);

Epo

s=E

piph

yas

post

vitt

ana

(Tor

tric

idae

);H

ass=

Hel

icov

erpa

assu

lta

(Noc

tuid

ae);

Hze

a=H

elic

over

paze

a(N

octu

idae

);O

fur=

Ost

rini

afu

rnac

alis

(Pyr

alid

ae);

Onu

b=O

stri

nia

nubi

lali

s(P

yral

idae

);P

oct=

Pla

noto

rtri

xoc

to(T

ortr

icid

ae);

Tni

=Tri

chop

lusi

ani

(Noc

tuid

ae).

621S.E. Jeong et al. / Insect Biochemistry and Molecular Biology 33 (2003) 609–622

In conclusion, we show that the three most abundantdesaturase-encoding transcripts in the H. assulta phero-mone gland encode two �9 desaturases (HassKPSE andHassNPVE) capable of producing the Z9-16:Acid pre-cursor of the major pheromone component and a �11desaturase (HassLPAQ) capable of producing the Z11-16:Acid precursor of the minor pheromone componentof this species. The biochemical properties of these desa-turases, as revealed by functional expression in ole1yeast cells, are similar to those of their highly conservedorthologues in H. zea. Furthermore, the relative amountsof transcripts encoding the two �9 desaturases and the�11 desaturase in the pheromone glands of these twospecies are consistent with the levels of the specific �9and �11 UFA precursors required to produce their respe-citve major and minor pheromone components, i.e., in H.assulta, where Z9-16:Acid is the precursor of the majorpheromone component, HassNPVE and HassKPSE areapproximately 40 and three times more abundant,respectively, than HassLPAQ (this report), whereas, inH. zea, where Z11-16:Acid is the precursor of the majorpheromone component, HzeaLPAQ is about five andeight times more abundant than HzeaKPSE andHzeaNPVE, respectively (Rosenfield et al., 2001). Thefinal blends containing Z9-16:Ald and Z11-16:Ald inspecific ratios in these two species, i.e., 20:1 in H.assulta and 1:30 in H. zea, are probably also influencedby the flux of specific UFA precursors as well as by thespecies-specific regulation of downstream steps in therespective pheromone biosynthetic pathways.

The results of the present investigation, of course, donot explain how H. assulta came to produce Z9-16:Alas its major sex pheromone component, in contrast toall other characterized heliothine species, which use Z11-16:Al as the major component; nor do they address howthe antennae of H. assulta males acquired a type of sen-silla trichodea specifically tuned to Z9-16:Al, which doesnot exist in H. zea (Berg and Mustaparta, 1995). It ispossible that the replacement of Z11-16:Al with Z9-16:Al as the major pheromone attractant in the line ofheliothine moths leading to the present day H. assultawas preceded by a gain-of-function mutation resultingin the expression of Z9-16:Al-specific sensilla trichodeain male antennae. Alternatively, gain-of-function in themale sensory system need not be invoked in a scenariowhere mutational loss of Z9-16:Al-specific sensilla trich-odea in an ancestor of other heliothine species permittedselection for increased production of the Z11-16:Al pher-omone component in female moths and increased num-bers of sensilla trichodea specifically tuned to thiscomponent in male antennae.

Acknowledgements

We gratefully acknowledge financial support for thisstudy by Hannam University to SEJ and by USDA

Grants 97-35302-4345 and 2001-35302-09926 andEPA/NSF Grant BES-9728367 to DCK.

References

Arn, H., Toth, M., Priessner, E., 1992. List of sex pheromones of Lepi-doptera and related attractants, second ed. International Organiza-tion for Biological Control, Montfavet, France.

Berg, B.G., Mustaparta, H., 1995. The significance of major phero-mone components and interspecific signals as expressed by receptorneurons in the oriental tobacco budworm Helicoverpa assulta. J.Comp. Physiol. A 177, 683–694.

Buser, H.R., Arn, H., Guerin, P., Rauscher, S., 1983. Determinationof double bond position in mono-unsaturated acetates by massspectrometry of dimethyl disulfide adducts. Anal. Chem. 55,818–822.

Cho, S., Mitchell, A., Regier, J.C., Mitter, C., Poole, R.W., Fried-lander, T.P., Zhao, S., 1995. A highly conserved nuclear gene forlow-level phylogenetics: elongation factor-1α recovers mor-phology-based tree for Heliothine moths. Mol. Biol. Evol. 12,650–656.

Choi, M.-Y., Han, K.S., Boo, K.S., Jurenka, R.A., 2002. Pheromonebiosynthetic pathways in Helicoverpa zea and Helicoverpa assulta.Insect Biochem. Mol. Biol. 32, 1353–1359.

Choi, M.Y., Tatsuki, S., Boo, K.S., 1998. Regulation of sex pheromonebiosynthesis in the oriental tobacco budworm, Helicoverpa assulta(Lepidoptera: Noctuidae). J. Insect Physiol. 44, 653–658.

Cork, A., Boo, K.S., Dunkelblum, E., Hall, D.R., Jee-Rajunga, K.,Kehat, M., Kong Jie, E., Park, K.C., Tepgidagarn, P., Xun, L.,1992. Female sex pheromone of oriental tobacco budworm, Helico-verpa assulta (Guenee) (Lepidoptera: Noctuidae): Identificationand field testing. J. Chem. Ecol. 18, 403–418.

Fang, Q.Q., Cho, S., Regier, J.C., Mitter, C., Matthews, M., Poole,R.W., Friedlander, T.P., Zhao, S., 1997. A new nuclear gene forinsect phylogenetics: DOPA decarboxylase is informative ofrelationships within Heliothinae (Lepidoptera: Noctuidae). Syst.Biol. 46, 269–283.

Friedlander, T.P., Regier, J.C., Mitter, C., 1995. A nuclear sequencefor higher level phylogenetic analysis: Phosphoenolpyruvate car-boxykinase tracks Mesosoic age splits within Lepidoptera. Mol.Biol. Evol. 13, 594–604.

Frohman, M.A., Dush, M.K., Martin, G.R., 1988. Rapid production offull-length cDNAs from rare transcripts: amplification using a sin-gle specific oligonucleotide primer. Proc. Natl. Acad. Sci. USA 85,8998–9002.

Hao, G., Liu, W., O’Connor, M., Roelofs, W.L., 2002. Acyl-CoA Z9-and Z10-desaturase genes from a New Zealand leafroller mothspecies, Planotortrix octo. Insect Biochem. Molec. Biol. 32,961–966.

Jeong, S.E., Lee, Y., Hwang, J.H., Knipple, D.C., 2001. Effects of thesap of the common oleander Nerium indicum (Apocyanaceae) onmale fertility and spermatogenesis in the oriental tobacco budwormHelicoverpa assulta (Lepidoptera, Noctuidae). J. Exp. Biol. 204,3935–3942.

Knipple, D.C., Roelofs, W.L., 2003. Molecular biological investi-gations of pheromone desaturases, in: Blomquist, G.J., Vogt, R.(Eds.), Insect Pheromone Biochemistry and Molecular Biology.Academic Press, New York (in press).

Knipple, D.C., Rosenfield, C.-L., Miller, S.J., Liu, W., Tang, J., Ma,P.W.K., Roelofs, W.L., 1998. Cloning and functional expressionof a cDNA encoding a pheromone gland-specific acyl-CoA (11-desaturase of the cabbage looper moth, Trichoplusi ni. Proc. Natl.Acad. Sci. USA 95, 15287–15292.

Knipple, D.C., Rosenfield, C.-L., You, K.M., Jeong, S.E., 2002. Evol-

622 S.E. Jeong et al. / Insect Biochemistry and Molecular Biology 33 (2003) 609–622

ution of the integral membrane desaturase gene family in mothsand flies. Genetics 162, 1737–1752.

Koh, Y.H., Park, K.C., Boo, K.S., 1995. Antennal sensilla in adultHelicoverpa assulta (Lepidoptera: Noctuidae): morphology, distri-bution and ultrastructure. Ann. Entomol. Soc. Amer. 88, 519–530.

Kyte, J., Doolittle, R.F., 1982. A simple method for displaying thehydropathic character of a protein. J. Mol. Biol. 157, 105–132.

Liu, W., Ma, P.W.K., Marsella-Herrick, P., Rosenfield, C.-L., Knipple,D.C., Roelofs, W., 1999. Cloning and functional expression of acDNA encoding a metabolic acyl-CoA �9-desaturase of the cab-bage looper moth, Trichoplusis ni. Insect Biochem. Molec. Biol.29, 435–443.

Liu, W., Jiao, H., Murray, N.C., O’Conner, M., Roelofs, W.L., 2002a.Gene characterized for membrane desaturase that produces (E)-11isomers of mono- and diunsaturated fatty acids. Proc. Natl. Acad.Sci. USA 99, 620–624.

Liu, W., Jiao, H., O’Conner, M., Roelofs, W.L., 2002b. Moth desatur-ase characterized that produces both Z and E isomers of �11-tetra-decenoic acids. Insect Biochem. Mol. Biol. 32, 1489–1495.

Masson, C., Mustaparta, H., 1990. Chemical information processingin the olfactory system of insects. Physiol. Rev. 70, 199–245.

Park, K.C., Boo, K.S., Cork, A., Hall, D.R., 1999. Monitoring Helicov-erpa assulta (Guenee) (Lepidoptera: Noctuidae) with synthetic sexpheromone and small scale disruption trial with controlled-releasePVC resin formulation. Appl. Entomol. Zool. 34, 19–22.

Park, K.C., Cork, A., Boo, K.S., 1996. Intrapopulation changes in sexpheromone composition during scotophase in oriental tobacco bud-worm, Helicoverpa assulta (Guenee) (Lepidoptera: Noctuidae). J.Chem. Ecol. 22, 1201–1210.

Roelofs, W.L., 1995. Chemistry of sex attraction. Proc. Natl. Acad.Sci. USA 92, 44–49.

Roelofs, W.L., Liu, W., Hao, G., Jiao, H., Rooney, A.P., Linn, C.E.,2002. Evolution of moth sex pheromones via ancestral genes. Proc.Natl. Acad. Sci. USA 99, 13621–13626.

Rosenfield, C.-L., You, K.M., Marsella-Herrick, P., Roelofs, W.L.,

Knipple, D.C., 2001. Structural and functional conservation anddivergence among acyl-CoA desaturase-encoding genes of twoNoctuid species, the corn earworm, Helicoverpa zea, and the cab-bage looper, Trichoplusia ni. Insect Biochem. Mol. Biol. 31,949–964.

Sanger, F., Nicklen, S., Coulsen, A.R., 1977. DNA sequencing withchain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74,5463–5467.

Shanklin, J., Whittle, E., Fox, B.G., 1994. Eight histidine residues arecatalytically essential in a membrane-associated iron enzyme, stea-royl-CoA desaturase, and are conserved in alkane hydroxylase andxylene monooxygenase. Biochem 33, 12787–12794.

Stukey, J.E., McDonough, V.M., Martin, C.E., 1990. The OLE1 geneof Saccharomyces cerevisiae encodes the �9 fatty acid desaturaseand can be functionally replaced by the rat stearoyl-CoA desaturasegene. J. Biol. Chem. 265, 20144–20149.

Sugie, H., Tatsuki, S., Nakagaki, S., Rao, C.B.J., Yamamoto, A., 1991.Identification of the sex pheromone of the oriental tobacco bud-worm, Heliothis assulta (Guenee) (Lepidoptera: Noctuidae). Appl.Entomol. Zool. 26, 151–153.

Tang, J.D., Wolf, W.A., Roelofs, W.L., Knipple, D.C., 1991. Develop-ment of functionally competent cabbage looper moth sex phero-mone glands. Insect Biochem. 21, 573–581.

Thompson, J.D., Higgins, D.G., Gibson, T.J., 1994. Improving the sen-sitivity of progressive multiple sequence alignment throughsequence weighting, positions-specific gap penalties and weightmatrix choice. Nucl. Acids Res. 22, 4673–4680.

Tillman, J.A., Seybold, S.J., Jurenka, R.A., Blomquist, G.J., 1999.Insect pheromones—an overview of biosynthesis and endocrineregulation. Insect Biochem. Mol. Biol. 29, 481–514.

Yang, Z., Nielsen, R., 1998. Synonymous and nonsynonymous ratevariation in nuclear genes of mammals. J. Mol. Evol. 46, 409–418.

Yoshiga, T., Okano, K., Mita, K., Shimada, T., Matsomoto, S., 2000.cDNA cloning of acyl-CoA desaturase homologs in the silkworm,Bombyx mori. Gene 246, 339–345.