ORIGINAL ARTICLE

Si Hyeock Lee Æ Patricia J. Ingles Æ Douglas C. KnippleDavid M. Soderlund

Developmental regulation of alternative exon usagein the house fly Vssc1 sodium channel gene

Received in revised form: 27 September 2001 /Accepted: 29 September 2001 / Published online: 10 November 2001� Springer-Verlag 2001

Abstract Sequence analysis of cDNA clones amplifiedby PCR from house fly (Musca domestica L.) Vssc1voltage-sensitive sodium channel a subunit transcripttemplates identified 11 putative alternatively splicedexons. Nine of these corresponded to the 7 optionalexons (designated a, b, e, f, h, i, and j) and 2 mutuallyexclusive exons (designated c/d) identified previously inthe orthologous para sodium channel a subunit genes ofDrosophila melanogaster and Drosophila virilis, whereastwo segments represented new mutually exclusive exonsin Vssc1 (designated k/l) located in a region not previ-ously identified as a site of alternative splicing in para.Diagnostic PCR assays on individual Vssc1 cDNAtemplates detected the presence or absence of each pu-tative alternative exon in multiple partial cDNAs (42–96individual clones per cDNA pool) from newly emergedfirst instar larvae, pupae, day 1 adult heads, and day 1adult bodies. Exons h and i were present in all cDNAclones from all developmental stages. Exon d was alsopresent in all clones from all developmental stages thatencoded full-length amino acid sequences; however, 1 of42 clones from adult head contained an exon c-likesegment in which the coding sequence was terminated bya premature stop codon. In contrast, the frequencies ofexons a, b, e, f, j, k, and l differed between develop-mental stages and adult anatomical regions. Analysis of

the Vssc1 region containing alternative exons a, b, c/d, e,f, h, and i as a single amplified cDNA segment identifiednine Vssc1 splice variants involving these exons. Thesplice variant containing exons a, d, h, and i was themost abundant form in all cDNA pools examined, butthe observed patterns of splice variant expression werespecific to each developmental stage and adult anatom-ical region. Our results document the strong conserva-tion of alternative exon location and structure betweenthe Vssc1 gene of the house fly and the para gene ofD. melanogaster but identify marked differences in exonusage between these species.

Keywords Voltage-sensitive sodium channel ÆHouse fly Æ Musca domestica Æ Alternative exon usage

Introduction

Alternative RNA splicing is one of several mechanismsthat underlies the structural and functional diversity ofion channels in animal nervous systems (Harris-Warwick 2000). In Drosophila melanogaster, alternativesplicing of shaker transcripts produces at least sevenvoltage-sensitive potassium channel proteins that differin electrophysiological properties and spatial expressionpatterns (Schwarz et al. 1988). Similarly, alternativesplicing of five optional exons in the slowpoke gene ofD. melanogaster produces a family of calcium-activatedvoltage-sensitive potassium channels that differ in uni-tary conductance, calcium sensitivity, and gating (Lag-rutta et al. 1994). Diversity in the D. melanogastercalcium channel a1 subunit (Dmca1A) is also predictedfrom the existence of three alternative exons and severalpotential RNA editing sites (Peixoto et al. 1997).

Alternative RNA splicing is the principal determinantof molecular diversity of voltage-sensitive sodiumchannel a subunits in D. melanogaster and the relatedspecies, Drosophila virilis. The D. melanogaster genomecontains only two genes, para and DSC1, that encodepredicted proteins with sequence homology to vertebrate

Invert Neurosci (2002) 4: 125–133DOI 10.1007/s10158-001-0014-1

S.H. Lee Æ P.J. Ingles Æ D.C. Knipple Æ D.M. Soderlund (&)Department of Entomology, New York State AgriculturalExperiment Station, Cornell University, Geneva,NY 14456, USAE-mail: dms6@cornell.eduTel.: +1-315-7872364Fax: +1-315-7872326

Present address: S.H. LeeEntomology Program, Division of Agricultureand Life Sciences, Seoul National University,103 Seodun-Dong, Suwon, Korea 441-744

Present address: P.J. InglesDepartment of Biology,University of Leicester,Leicester LE1 7RH, UK

voltage-sensitive sodium channel a subunits (Littletonand Ganetzky 2000). Whereas products of the para geneare recognized as voltage-sensitive sodium channela subunits by genetic and functional criteria (Loughneyet al. 1989; Warmke et al. 1997), products of the DSC1locus have no known function. Sequence analyses of thepara genes of D. melanogaster and D. virilis and oftranscripts derived from them identified alternativesplicing at eight sites involving seven optional exons(designated a, b, e, f, h, i, and j) and two mutuallyexclusive exons (designated c/d; Loughney et al. 1989;O’Dowd et al. 1995; Thackeray and Ganetzky 1994,1995; Warmke et al. 1997). Alternative splicing at thesesites produces a heterogeneous, developmentally regu-lated family of sodium channel a subunit transcripts(Thackeray and Ganetzky 1994, 1995).

The Vssc1 sodium channel a subunit gene of thehouse fly (Musca domestica L.), the ortholog of para inthis species (Ingles et al. 1996; Williamson et al. 1996), isthe locus of point mutations that cause resistance topyrethroid insecticides (Soderlund and Knipple 1999).The initial sequence analyses of Vssc1 cDNA clonesamplified or isolated from adult heads did not identifyalternatively spliced transcripts (Ingles et al. 1996; Wil-liamson et al. 1996). The present study investigated theissue of alternative splicing at the Vssc1 locus in greaterdetail by examining the heterogeneity of multiple cDNAclones derived from Vssc1 transcripts in three differentdevelopmental stages and two different adult bodyregions. In this paper we identify multiple alternativeexons in the Vssc1 gene, including a new pair of mutu-ally exclusive exons at a splice site not previously knownfrom the analysis of para transcripts in Drosophila spe-cies. We also document the developmentally regulatedalternative splicing of Vssc1 transcripts and compare thestructures and patterns of expression of Vssc1 splicevariants with those described from the orthologous paragenes of Drosophila species.

Materials and methods

PCR amplification of Vssc1 fragments encompassingalternative exons

Four total RNA fractions (first instar larvae, 0–6 h old; pupae; andadult heads and bodies separated by sieving frozen flies, 0–24 h old)were extracted from house flies (insecticide-susceptible NAIDMstrain) using the acid guanidinium thiocyanate–phenol–chloroformextraction methods (Chomczynski and Sacchi 1987), and mRNAwas purified from 1 mg of each total RNA sample using thePolyATract mRNA isolation system (Promega, Madison, Wis.,USA). First strand cDNA was synthesized from 0.1 to approxi-mately 0.5 lg of each mRNA using Superscript II reverse tran-scriptase (Life Technologies, Grand Island, N.Y., USA) by primingwith oligo(dT) and recovered in a 50-ll volume (Centricon100;Amicon, Beverly, Mass., USA). A 2-ll aliquot of each sample wasused as the template in PCR reactions.

Vssc1 cDNA fragment A (ca 2.0 kb), containing seven alter-native splicing sites known to exist in the D. melanogaster para gene(exons a, b, c/d, e, f, h, and i), was amplified by PCR with forwardprimer A1 (5¢-GGAGGTACCGCGTCGAAGTGGAATCGGA-GT-3¢) and reverse primer A2 (5¢-TCGGCATGCAATGCTGG-

CAGTGTCATCGTC-3¢) using eLONGase (Life Technologies).The PCR reaction was comprised of 2 cycles of 94�C for 30 s, 59�Cfor 1 min, and 68�C for 4 min, followed by 33 cycles of 94�C for30 s, 65�C for 50 s, and 68�C for 4 min. Vssc1 cDNA fragment B(388–421 bp), containing exon j, was amplified with forwardprimer B1 (5¢-CCCATGACAGAAGATTCCGACTCGATATC-TG-3¢) and reverse primer B2 (5¢-GATTGAATGGATCGAGCA-GCC-3¢). Vssc1 cDNA fragment C (492 bp), containing mutuallyexclusive exons k and l, was amplified with forward primer C1(5¢-CAATGCCTGGTGTTGGCTGG-3¢) and reverse primer C2(5¢-GGCTGCTTGTCCACCTCTCG-3¢). PCR reactions for frag-ments B and C were performed using Taq polymerase (Promega)with the following thermal programs: 35 cycles of 94�C for 30 s,60�C for 40 s, and 72�C for 1 min followed by 1 cycle of 72�C for10 min.

Fragment A was digested with KpnI and SphI and ligated intoKpnI/SphI-digested pALTER-1 (Promega). The ligated DNA wastransformed into Stbl2 competent cells (Life Technologies). Excessprimers were removed from fragments B and C by solventreplacement (three passes) in a Microcon-100 concentrator(Amicon) and directly cloned into pCR2.1 (Invitrogen, Carlsbad,Calif., USA). Multiple transformants (42–96 colonies from eachcloning) were isolated for characterization by diagnostic PCR.

Diagnostic PCR analyses for the detection of alternative exons

Diagnostic PCR reactions were used to detect the presence of eachputative alternative exon in individual cDNA clones. Table 1 givesthe names and nucleotide sequences of the primers employed indiagnostic PCR reactions and the PCR products obtained in thepresence or absence of each exon. To detect exon h, the sizes of thePCR products generated by a set of primers flanking the alternativeexon region were compared. For alternative exons j and b, similardiagnostic assays based on the sizes of the PCR products were notfeasible due to the small size of the exons. Therefore one of theprimers in each amplification was designed to anneal to sequencewithin the exon so that the presence or absence of the exon wasdetermined by the presence or absence of expected specific PCRproduct. Adjacent exons (for example, i and a, e and f) weredetected using two sets of primers, each with one primer specific forone of the adjacent exons. The presence or absence of each exon-specific fragment unambiguously detected the presence of eachexon individually, whereas the simultaneous amplification of bothfragments, which could be resolved on agarose gels, identifiedcDNAs containing both exons. Mutually exclusive exons c/d and k/l were identified by the use of two exon-specific primers, one basedon the sense strand sequence of one exon and the other on theantisense strand of the other exon. These exon-specific primers werepaired with counterpart primers located on either side of thealternatively spliced segment to generate PCR products with un-ique sizes for each exon. Assays for pairs of adjacent (for example, iand a, e and f) and mutually exclusive (for example, c/d and k/l)exons were performed simultaneously, taking advantage of theunique sizes of exon-specific PCR products to identify the exonspresent in each clone and the amplification of larger fragments bypairs of flanking primers as a positive control for each PCR reaction.

All diagnostic PCR reactions were performed in a 30- or 50-llvolume using a 96-well format. A 1-ll aliquot of the overnightculture of the each Vssc1 fragment clone mixed with 14 ll1·standard reaction buffer was heated to 96�C for 10 min and usedas the template bottom solution (15 ll). The top solution (25 or35 ll) contained all other ingredients including appropriate primersets and Taq polymerase (Promega). The PCR reaction was initi-ated by mixing the top and bottom solutions. The thermal cyclesvaried only in annealing temperature depending on the Tm valuesof the primers: 35 cycles of 94�C for 30 s, 45–58�C for 30 s, and72�C for 60 s. Following PCR, an aliquot (6–8 ll) of each PCRreaction mixture was analyzed by electrophoresis in 2% NuSieve/0.5% SeaKem agarose (FMC BioProducts, Rockland, Me., USA)gels. Representative clones exhibiting each unique banding patternwere sequenced to confirm the reliability of diagnostic PCR assays.

126

Results

Identification and structure of putative alternativeexons in Vssc1 transcripts

To characterize the structural heterogeneity of Vssc1transcripts in the house fly, we analyzed individualcloned partial cDNAs obtained by PCR amplificationon first strand cDNA template pools. Our initialexperiments involved amplification of a ca 2.0 kb frag-ment (fragment A) that spanned the sites of alternativesplicing of the six optional exons (a, b, e, f, h, and i) andtwo mutually exclusive exons (c/d) identified previously

in para transcripts from D. melanogaster (see Fig. 1).During the course of these experiments, DNA sequenceanalysis of cDNA fragments employed in the assemblyof a full-length Vssc1 (Smith et al. 1997) and para(Warmke et al. 1997) cDNAs identified two additionalregions of transcript heterogeneity lying outside thesegment of Vssc1 sequence contained in fragment A. We

Table 1 Oligonucleotide prim-ers and diagnostic PCR prod-ucts used to detect alternativeexons in cloned Vssc1 cDNAs

Exon Primer sequence (5¢fi3¢)a Diagnostic PCR products

Positive Negative

a F: CCTGGTTCACCATTTAACCTA 303 bp NoneR: CATGCCACCCAATAAATCACC

b F: GTCTCCGTTTACTATTTTCCCAC 370 bp NoneR: GTAAATTCAGTGTGGGCCATG

c F: TTGACCTTCGTCTTGTGT 147 bp NoneR: CATGAAGCTGTGCATGAA

d F: GGTGCTGAAAAGTGGTAA 301 bp NoneR: TTATGGGCCGGACAATG

e F: GACTGCCGACAATGATACC 163 bp NoneR: GATCTGGTTGATCCTCTCA

f F: GGCAAAGGGGTTTGTCGTT 309 bp NoneR: TCGGCATGCAATGCTGGCAGTGTCATCGTC

h F: ATGAAGGGCGAGACCCAGCTG 210 bp 135 bpR: TCGGCATGCAATGCTGGCAGTGTCATCGTC

i F: GGAGGTACCGCGTCGAAGTGGAATCGGAGT 100 bp NoneR: CGTGCTAACTTTACGGACTTTAG

j F: CCGCGATATGGTCGCAAGAAA 262 bp NoneR: GATTGAATGGATCGAGCAGCC

k F: CAATGCCTGGTGTTGGCTGG 91 bp NoneR: CACGGCTATATCATTTAAG

l F: AGCTGGTGGTATACAAGCC 190 bp NoneR: AAGCTGTACTCCCATAATGG

aForward (F; sense strand) and reverse (R; antisense strand) primers for detection of each exon

Fig. 1 Top Schematic diagram of the Vssc1 sodium channel cDNAshowing the four homology domains (I–IV) with six putativetransmembrane helices (labeled S1–S6 in homology domain I) ineach homology domain. Bottom Expanded diagram of part of theVssc1 cDNA showing fragments A, B, and C and the locations andrelative lengths of optional exons (hatched bars) and mutuallyexclusive exons (filled bars)

127

therefore expanded our analysis to include PCR-ampli-fied partial cDNAs (fragments B and C; see Fig. 1)encompassing these regions. Fragments A, B, and Cspanned approximately 3 kb of the 6.3-kb Vssc1 tran-script and contained all of the sites of alternative splicingidentified to date in the para transcripts of Drosophilaspecies.

We sequenced multiple cDNA clones from larvae,pupae, and adults to identify the Vssc1 sequence seg-ments corresponding to alternative exons of para andidentify novel alternatively spliced elements unique toVssc1. The nucleotide and inferred amino acid sequencesof Vssc1 segments corresponding to alternative paraexons a, d, h, i, and k were described in our initial reportof a full-length Vssc1 sequence that contained theseexons (Ingles et al. 1996). Figure 2 shows the nucleotideand inferred amino acid sequences of six additionalalternative exons identified in the present study. Ofthese, five corresponded to optional exons b, c, e, f and jof para, whereas the sixth (exon l) formed a secondmutually exclusive exon pair with exon k in homologydomain III, a region not previously identified as a site ofalternative splicing in para. Exon j of Vssc1 exhibited asingle nucleotide polymorphism resulting in the substi-tution of Gln for Arg at the second amino acid residueof the exon in some of the clones examined.

Figure 3 compares the inferred amino acid se-quences of these Vssc1 segments to the correspondingsegments of the para gene product. Six of the seven

putative optional exons examined in this study (ex-ons a, b, e, f, h, and j) encoded amino acid sequencesthat either were identical to those encoded by the cor-responding segments of para or differed from the para-encoded sequences by a single conservative amino acidsubstitution. Exon i of Vssc1 differed from exon i ofpara at two residues, the non-conservative replacementof His by Pro and the deletion of the adjacent Glnresidue.

Unlike mutually exclusive exons c and d in para,which differ at only 2 of 55 amino acid residues(Loughney et al. 1989), mutually exclusive exons c and din Vssc1 differed substantially from each other at the 3¢termini. The Vssc1 segment originally described in thisregion (Ingles et al. 1996) was identical to exon d of paraat 54 of 55 amino acid residues and was thereforeidentified as exon d of Vssc1. The present study identi-fied a second segment, found in place of exon d in asingle cDNA clone, that was truncated by a stop codonafter 49 amino acid residues. The nucleotide sequenceupstream of the stop codon differed from that of exon dat 27 positions; of these, 24 were silent substitutionswhereas 1 resulted in the replacement of Met by Val thatdistinguishes para exons c and d in this region and theothers resulted in the substitution of Ala for Leu justahead of the stop codon. Downstream of the stopcodon, this fragment encoded an additional 6 aminoacid residues that were not conserved with respect toexon c of para or exon d of either para or Vssc1 followedby a second in-frame stop codon. We identify this novelsequence as exon c of Vssc1 based on its location relativeto exon d and the identity of the majority of its codingsequence with exon c of para.

The sequences of mutually exclusive exons k and l inVssc1 were also highly divergent, differing at 16 of 41amino acid residues. The inferred amino acid sequenceof exon l of Vssc1 differed from the corresponding seg-

Fig. 2 Nucleotide and inferred amino acid sequences of previouslyundescribed alternative exons in the house fly Vssc1 gene. The stopcodon in exon c is italicized in the nucleotide sequence andindicated by an asterisk in the inferred amino acid sequence. Thenucleotide and amino acid residues underlined in the exon jsequence are polymorphic (A in place of G encoding Gln in placeof Arg in some clones)

128

ment encoded by para by only 1 residue. Although noexon homologous to exon k of Vssc1 has been reportedin para, a BLAST (Altschul et al. 1997) search of theD. melanogaster genome database (Adams et al. 2000)using the exon k peptide as the query sequence yielded apreviously unrecognized sequence element, located up-stream of the segment encoding exon l, having a pre-dicted amino acid sequence similar to that of exon k inVssc1 (identical at 35 of 41 sequence positions; seeFig. 3). Despite the substantial sequence divergence ofexons k and l in both Vssc1 and para, these elementsretain the five positively charged amino acid residues intransmembrane domain IIIS4 that form part of thevoltage sensor of the sodium channel (Catterall 1992).

Developmental regulation of alternative exon usage

We used diagnostic PCR reactions for each putativealternative exon to determine the alternative exonstructure of fragments A–C amplified from four discretemRNA pools: newly emerged (0–6 h) larvae, pupae, 0–24 h adult heads, and 0–24 h adult bodies. Data were

collected from sets of 42–96 clones for each fragmentand developmental stage.

The frequencies of each putative alternative exon ineach of the four mRNA pools are summarized inFig. 4. Exons h and i were present in all cDNA clonesexamined. Exon d was also present in all clonesexamined from the larval, pupal, and adult bodycDNA pools and in 41 of the 42 clones analyzed fromadult head cDNA. Exons a and k were also present athigh frequencies in all cDNA pools (91–98% forexon a; 78–100% for exon k). In contrast, exon e wasnot detected in larvae or adult bodies but was presentat low frequencies (<10%) in pupae and adult heads.Only three optional exons (b, f, and j) and one pair ofmutually exclusive exons (k/l) exhibited marked devel-opmental or anatomical regulation of expression. Threeof the optional exons (e, f, and j) appeared to haverestricted anatomical expression in adults: exons e and jwere detected in some cDNA clones from adult headsbut not adult bodies, whereas exon f was found only inclones from adult bodies but not adult heads. Mutuallyexclusive exons k and l also were differentiallyexpressed in adult heads and bodies.

Fig. 3 Annotated align-ments of the predictedamino acid sequences forputative alternative exons inVssc1 and para. Amino acidresidues in para that areidentical to those in Vssc1are indicated by periods; thestop codon at residue 50 ofVssc1 exon c is indicated byan asterisk; the Met residuein Vssc1 exon d that ismutated to Thr in super-kdrhouse fly strains is under-lined; the gap introduced inVssc1 exon i to preservealignment is indicated by adash; the polymorphic ami-no acid residue in exon j isindicated by an asteriskabove the Vssc1 sequence;the amino acid residues inexons k and l that formtransmembrane do-main IIIS4 are underlined;the conserved positivelycharged (Arg or Lys) resi-dues in this domain aremarked with plus signs aboveeach Vssc1 sequence

129

Analysis of the Vssc1 region containing alternativeexons a, b, c/d, e, f, h, and i as a single amplified cDNAsegment (fragment A) detected nine Vssc1 splice vari-ants involving these exons. The relative abundance ofthese nine splice variants is shown in Fig. 5. Theobserved patterns of splice variant expression werespecific to each developmental stage and adult anatom-ical region. The splice variant containing exons a, d, h,and i (designated adhi) was the most abundant form inall four cDNA pools, and this variant plus the threerelated variants obtained by alternative splicing of ex-ons b and f (i.e., abdhi, adfhi, and abdfhi) accounted for88–98% of the splice variants in all four cDNA pools.The larval cDNA pool contained six splice variants,whereas the pupal pool contained seven variants andeach adult pool contained four variants. The pupalcDNA pool was most heterogeneous (none of the sevenvariants represented more than 35% of the total)whereas the adult body cDNA pool was least hetero-geneous (the adhi variant represented 84% of the total).Six of the nine variants were detected in at least two

cDNA pools, whereas the adefhi variant was found as aminor component only in the pupal cDNA pool and theadehi variant and the unusual truncated fragment con-taining exon c (designated abc*i) were found only in theadult head cDNA pool.

Discussion

Our results demonstrate that transcripts from the Vssc1sodium channel a subunit gene of the house fly, likethose from its orthologs in D. melanogaster and D. virilis(Thackeray and Ganetzky 1994, 1995), are alternativelyspliced at multiple sites. The locations and encodedamino acid sequences of alternative exons in Vssc1 arehighly conserved in comparison to the correspondingsegments of para. Eight of the 11 putative alternativeexons identified in Vssc1 were either identical in aminoacid sequence to the corresponding regions of para orcontained a single conservative amino acid substitution.The degree of conservation of the structure, location,

Fig. 4 Relative frequenciesof individual Vssc1 alterna-tive exons in house fly larvae(L), pupae (P), adult heads(AH), and adult bodies (AB).The alternative exons areshown in the linear order oftheir occurrence in the Vssc1cDNA (see Fig. 1). Optionalexons are indicated by blackbars; mutually exclusive ex-ons are represented byhatched bars (exons d and k)combined with open bars(exons c and l). The totalnumbers of clones investi-gated were 91 (larvae), 96(pupae), 42–78 (adult heads),and 61–72 (adult bodies)

Fig. 5 Frequencies of Vssc1splice variants generated byexons i, a, b, c/d, e, f, and hin different developmentalstages and anatomical re-gions of the house fly.Exon c was detected in onlyone clone in which the pres-ence or absence of the ex-ons e, f, and h was notdetermined because of thelack of downstream codingsequence following a pre-mature stop codon. The to-tal numbers of clonesinvestigated were 91 (larvae),96 (pupae), 42 (adult heads),and 61 (adult bodies)

130

and sequence of these segments between Vssc1 and parais greater than that found between some of these seg-ments in para and the corresponding segments of para-orthologous sodium channel a subunit genes of eitherHeliothis virescens (Park et al. 1999) or Leptinotarsadecemlineata (Lee et al. 1999).

The unique aspects of alternative splicing in Vssc1were most evident in the two regions of mutuallyexclusive exon usage involving the exon pairs c/d and k/l.In para, mutually exclusive exons c and d are very sim-ilar, being distinguished by only 2 out of 55 amino acidresidues (Loughney et al. 1989). The segment of theoriginal Vssc1 sequence (Ingles et al. 1996) that corre-sponds to the exon c/d region of para was slightly moresimilar to para exon d (54 identical amino acids out of55) than to exon c (53 identical amino acid out of 55)and was therefore designated exon d of Vssc1. In thepresent study, we detected only 1 out of a total of 292Vssc1 clones from four cDNA pools that was notidentical to the previously described exon d sequence.This single clone encoded a protein segment identical topara exon c at 48 of the first 49 amino acid residues,followed by an in-frame stop codon (see Fig. 2). Thesequence following the stop codon encoded six non-conserved amino acid residues followed by a second in-frame stop codon. The nucleotide sequence of this exondiffered from that of Vssc1 exon d not only downstreamof the stop codon but also at 27 sequence positionswithin the upstream coding region. The degree ofnucleotide sequence divergence between the coding regionof this exon and the corresponding region of exon d (non-identical at 27 of 147 positions) is similar to the divergencebetween the corresponding regions of exons c and d of thepara gene of D. melanogaster (non-identical at 26 of 147positions; Loughney et al. 1989). We consider this novelsequence element to be exon c of Vssc1.Vssc1 transcripts containing exon c would encode a

truncated, and presumably non-functional, sodiumchannel a subunit isoform. A precedent for sodiumchannel truncation by alternative splicing exists for theSCN8A sodium channel a subunit gene from mice,humans, and fish (Plummer et al. 1997). In the case ofthe SCN8A sodium channel, the truncated form iswidely expressed in non-neuronal tissues and is thoughteither to serve as a mechanism regulating channelexpression or to produce a novel protein with a functionspecific to non-neuronal tissues (Plummer et al. 1997). Incontrast, our results suggest that truncated Vssc1variants containing exon c are rare.

The unusual structure of Vssc1 exon c sheds addi-tional light on the molecular mechanisms that underlieknockdown resistance to pyrethroid insecticides in thehouse fly. Two Vssc1 gene mutations found in knock-down-resistant fly populations are known to reduce thesensitivity of fly sodium channels to pyrethroids: sub-stitution of Phe for Leu at position 1014 (designatedL1014F) in all resistant populations, and the additionalsubstitution of Thr for Met at position 918 (M918T) inhighly resistant populations having the super-kdr phe-

notype (Soderlund and Knipple 1999). It is of particularinterest in the context of the present study that theM918T mutation is located within alternative exon d(see Fig. 3). This result suggests that exon d is anobligatory component of toxicologically relevant sodi-um channels in the house fly and provides further con-firmation that channel variants containing exon c arenon-functional. Unlike the primary knockdown resis-tance mutation (L1014F), which is found not only inVssc1 but in the orthologous sodium channels ofpyrethroid-resistant strains of several insect species(Soderlund and Knipple 1999), the M918T mutation hasbeen identified to date only in the house fly and in highlyresistant populations of the horn fly (Haematobia irri-tans; Guerrero et al. 1997). It is possible that the M918Tmutation can only be selected in those insects in whichalternative splicing at the exon c/d site is lost or those inwhich functional channels are produced from only oneof these exons, as is inferred from the structure of exon cin the house fly.

The existence of a second pair of mutually exclusiveexons (k/l) in homology domain III of the Vssc1 genewas not predicted from previous studies of alternativeexon usage in the para gene. This pair of exons, whichencode part of domain IIIS3 and all of domain IIIS4,share only 16 of 41 amino acid residues but retain theessential structural features of these domains, includingthe positively charged amino acid residues indomain IIIS4 that form a component of the voltagesensor of the sodium channel (Catterall 1992). Exon l ofVssc1 differed from the reported sequence of the paragene in this region by only a single conservative aminoacid substitution. Although a segment of para corre-sponding to exon k was not detected previously, aBLAST search of the D. melanogaster genome databaselocated a segment of the para gene upstream from exon lthat encoded a predicted amino acid sequence identicalat 35 of 41 positions to exon k of Vssc1 (see Fig. 3).Whereas this segment appears to represent exon k of thepara gene, the extent and pattern of alternative splicingof exons k and l in para transcripts are not known.

The location of exons k and l corresponds exactly tothe location of mutually exclusive exons 18A and 18N inthe SCN8A sodium channel gene of vertebrate species(Plummer et al. 1997). Interestingly, some vertebratesodium channel genes also contain a pair of mutuallyexclusive exons located at the same relative position inhomology domain I (i.e., encoding part of domain IS3and all of domain IS4; Gustafson et al. 1993; Plummeret al. 1998; Sarao et al. 1991). The conservation of theposition of mutually exclusive exon splicing in homologydomains I and III of these genes suggests that duplica-tion of this exon preceded the formation of the four-domain sodium channel structure by duplication of anancestral two-domain gene (Plummer et al. 1997).Despite this implied ancient origin of mutually exclusiveexons at homologous positions in domains I and III,none of the studies of the transcript heterogeneity ofpara and its orthologs in insects has identified mutually

131

exclusive exons at this site in homology domain I. Theseobservations suggest that this site of alternative splicinghas been lost in insects.

Alternative splicing at all of the sites identified eitherin Vssc1 or in the para genes of D. melanogaster orD. virilis could theoretically generate up to 512 struc-turally unique sodium channel splice variants. However,the patterns of Vssc1 exon usage that we observed sug-gest that only a small subset of these variants is actuallyemployed in the house fly. Exons h and i, which arealternatively spliced in Drosophila species (O’Dowd et al.1995; Thackeray and Ganetzky 1994, 1995), were foundin all Vssc1 clones examined. Therefore, either thesesegments are absent in a very small proportion of Vssc1transcripts or they do not function as alternativelyspliced elements in the house fly. Similarly, alternativesplicing at the exon c/d site in Vssc1 appears to beextremely rare and would not, in any case, contribute tothe diversity of transcripts encoding full-length,presumably functional channel variants. Analysis of thesplice variants reflected in the fragment A cDNA poolsdetected only nine variants (see Fig. 5), four of whichcomprised more than 88% of the transcripts in eachpool. The adhi variant was the most abundant splicetype in each cDNA pool. Our analysis of splice variantabundance does not include exons j and k/l becausethese were analyzed in separate PCR-amplified cDNApools. However, considering the low frequencies ofexon j and the high frequencies of exon k in these pools(see Fig. 4) it is likely that the adhik variant is the mostabundant Vssc1 transcript in all of these pools. Thisfinding is consistent with the recovery of only the adhikvariant in the initial characterizations of the full-lengthcDNA sequence of Vssc1 (Ingles et al. 1996; Williamsonet al. 1996).

Our data on splice variant structure (Fig. 5) provide abasis for the comparison of alternative splicing patternsin the house fly with those in D. melanogaster (Thack-eray and Ganetzky 1994) and D. virilis (Thackeray andGanetzky 1995). Although the data sets for the twoDrosophila species consider only exons a–f, our datafrom fragment A are directly comparable becauseexons h and i were present in all clones examined andtherefore did not contribute to the diversity of Vssc1splice variants. Embryonic cDNA pools from eachDrosophila species contained 11 splice variants involvingexons a–f, whereas the adult pools contained 17(D. virilis) or 18 (D. melanogaster) variants. When datafor the abundance of exon h in D. virilis are included, thetotal number of variants detected in this speciesincreases to 28 (Thackeray and Ganetzky 1995). Incontrast, 0–6 h house fly larvae contained 6 splice vari-ants and 0–24 h house fly adults (combined head andbody pools to achieve a comparable data set) contained5 splice variants. These data suggest that the transcriptpool obtained by alternative splicing of Vssc1 is sub-stantially less complex than those obtained by alterna-tive splicing of the para genes of Drosophila specieswithin the transcript region analyzed in these studies.

The marked difference in the number of splice variantsrecovered is not likely to result from differences in thenumber of cDNAs examined because our study, whichanalyzed 96 larval cDNA clones and 102 adult cDNAclones, was similar in scope to those of the two Droso-phila studies (125–139 clones per cDNA pool; Thack-eray and Ganetzky 1994, 1995). However, differences inthe number of splice variants recovered could resultfrom our use of poly(A)+ RNA, rather than total RNA,as the template for first-strand cDNA synthesis.

The house fly also differs significantly from bothDrosophila species in the exon structure of the mostabundant splice variants. Considering first only the coreregion encompassing exons a–f for which directly com-parable data exist, the predominant splice variant in allVssc1 cDNA pools (with the alternative exon structureof ad for this region) corresponds to a minor splice typedetected in D. melanogaster adults and D. virilis embryosand adults (Thackeray and Ganetzky 1994, 1995).Among the other three most abundant Vssc1 variants,the abd variant corresponds to the most abundant splicevariant in Drosophila adults, the abdf variant corre-sponds to an abundant variant in Drosophila embryos,and the adf variant corresponds to a minor variant inboth developmental stages. Additional differences inexon usage patterns are evident when our data arecompared to those for D. virilis, which included ananalysis of exon h (Thackeray and Ganetzky 1995).Whereas all Vssc1 clones contained exon h, the mostabundant splice variants in D. virilis lacked this exon.Finally, our data show that most house fly Vssc1 tran-scripts contain exon k, whereas D. melanogaster cDNAs(Loughney et al. 1989), as well as the cDNAs for thepara-orthologous genes of H. virescens (Park et al. 1999)and Blattella germanica (Dong 1997; Miyazaki et al.1996), have been found so far to contain only exon l-likesequences. Taken together, these observations suggestthat despite the strong conservation of exon structurebetween the orthologous sodium channel genes of thehouse fly and these two Drosophila species, the patternof exon use in the house fly is distinctive.

The conservation of alternative exon structure andthe developmental and anatomical regulation of alter-native exon usage in the para sodium channel gene ofD. melanogaster and its orthologs in other insect speciesimply that alternative splicing may generate a family ofsodium channel proteins with differing functionalproperties, as has been found for other ion channels andreceptors (Harris-Warwick 2000). Among the putativealternative exons considered in this study, all of theoptional exons (a, b, e, f, h, i, and j) are located inintracellular domains of the sodium channel protein andmay therefore be involved in the regulation of sodiumchannel expression or function as the result of interac-tions with protein kinases or G proteins (Cukierman1996). Moreover, exons a and i contain consensus pro-tein kinase A phosphorylation sites (Ingles et al. 1996;O’Dowd et al. 1995). So far, there is little direct evidencebearing on the functional significance of the alternative

132

splicing of optional exons. In embryonic D. melanogas-ter neurons, functional sodium channels were detectedonly in those cells having para transcripts containingexon a, thereby implying a critical role for this exon inchannel function (O’Dowd et al. 1995). However, directcomparison of the bdeijl and abdeijl variants of para,which differ only by the presence or absence of exon a,in functional expression assays in Xenopus laevis oocytesdid not find any effects of exon a on sodium currentexpression or properties in this system (Warmke et al.1997). Our discovery of mutually exclusive exons k andl, which encode substantially divergent amino acidsequences in a region of the sodium channel proteinassociated with the voltage sensor mechanism, suggeststhat channels differing in their exon structure at this sitemight exhibit different voltage-dependent properties. Itwill therefore be of interest to examine directly thefunctional significance of these mutually exclusive andstructurally divergent exons in expression assays usingcloned cDNAs that differ only at the exon k/l splice site.

Acknowledgements We thank P. Adams and K. Nelson for tech-nical assistance. This study was supported by grants 94-37302-0408and 97-35302-4323 from the United States Department of Agri-culture.

References

Adams M, Celniker S, Holt R, et al (2000) The genome sequence ofDrosophila melanogaster. Science 287:2185–2195

Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, MillerW, Lipman DJ (1997) Gapped BLAST and PSI-BLAST: a newgeneration of protein database search programs. Nucleic AcidsRes 25:3389–3402

Catterall WA (1992) Cellular and molecular biology of voltage-gated sodium channels. Physiol Rev 72:S15–S48

Chomczynski P, Sacchi N (1987) Single-step method of RNA iso-lation by acid guanidinium thiocyanate-phenol-chloroform ex-traction. Anal Biochem 162:156–159

Cukierman S (1996) Regulation of voltage-dependent sodiumchannels. J Membrane Biol 151:203–214

Dong K (1997) A single amino acid change in the para sodiumchannel protein is associated with knockdown-resistance (kdr)to pyrethroid insecticides in the German cockroach. InsectBiochem Mol Biol 27:93–100

Guerrero FD, Jamroz RC, Kammlah D, Kunz SE (1997) Toxico-logical and molecular characterization of pyrethroid-resistanthorn flies, Haematobia irritans: identification of kdr and super-kdr point mutations. Insect Biochem Mol Biol 27:745–755

Gustafson TA, Clevinger EC, O’Neill TJ, Yarowsky PJ, KruegerBK (1993) Mutually exclusive exon splicing of type III brainsodium channel a subunit RNA generates developmentallyregulated isoforms in rat brain. J Biol Chem 268:18648–18653

Harris-Warwick RM (2000) Ion channels and receptors: moleculartargets for behavioral evolution. J Comp Physiol A 186:605–616

Ingles PJ, Adams PM, Knipple DC, Soderlund DM (1996) Char-acterization of voltage-sensitive sodium channel gene codingsequences from insecticide-susceptible and knockdown-resistanthouse fly strains. Insect Biochem Mol Biol 26:319–326

Lagrutta A, Shen KZ, North RA, Adelman JP (1994) Functionaldifferences among alternatively spliced variants of Slowpoke, aDrosophila calcium-activated potassium channel. J Biol Chem269:20347–20351

Lee SH, Dunn JB, Clark JM, Soderlund DM (1999) Molecularanalysis of kdr-like resistance in a permethrin-resistant strain ofColorado potato beetle. Pestic Biochem Physiol 63:63–75

Littleton JT, Ganetzky B (2000) Ion channels and synaptic orga-nization: analysis of the Drosophila genome. Neuron 26:35–43

Loughney K, Kreber R, Ganetzky B (1989) Molecular analysis ofthe para locus, a sodium channel gene in Drosophila. Cell58:1143–1154

Miyazaki M, Ohyama K, Dunlap DY, Matsumura F (1996)Cloning and sequencing of the para-type sodium channel genefrom susceptible and kdr-resistant German cockroaches (Blat-tella germanica) and house fly (Musca domestica). Mol GenGenet 252:61–68

O’Dowd DK, Gee JR, Smith MA (1995) Sodium current densitycorrelates with expression of specific alternatively spliced sodi-um channel mRNAs in single neurons. J Neurosci 15:4005–4012

Park Y, Taylor MFJ, Feyereisen R (1999) Voltage-gated sodiumchannel genes hscp and hDSC1 of Heliothis virescens F.genomic organization. Insect Mol Biol 8:161–170

Peixoto AA, Smith LA, Hall JC (1997) Genomic organization andevolution of alternative exons in a Drosophila calcium channelgene. Genetics 145:1003–1013

Plummer NW, McBurney MW, Meisler MH (1997) Alternativesplicing of the sodium channel SCN8A predicts a truncatedprotein in fetal brain and non-neuronal cells. J Biol Chem272:24008–24015

Plummer NW, Galt J, Jones JM, Burgess DL, Sprunger LK,Kohrman DC, Meisler MH (1998) Exon organization, codingsequence, physical mapping, and polymorphic intragenicmarkers for the human neuronal sodium channel gene SCN8A.Genomics 54:287–296

Sarao R, Gupta SK, Auld VJ, Dunn RJ (1991) Developmentallyregulated alternative RNA splicing of rat brain sodium channelmRNAs. Nucleic Acids Res 19:5673–5679

Schwarz TL, Tempel BL, Papazian DM, Jan YN, Jan LY (1988)Multiple potassium-channel components are produced by al-ternative splicing at the Shaker locus in Drosophila. Nature331:137–142

Smith TJ, Lee SH, Ingles PJ, Knipple DC, Soderlund DM (1997)The L1014F point mutation in the house fly Vssc1 sodiumchannel confers knockdown resistance to pyrethroids. InsectBiochem Mol Biol 27:807–812

Soderlund DM, Knipple DC (1999) Knockdown resistance toDDT and pyrethroids in the house fly (Diptera: Muscidae):from genetic trait to molecular mechanism. Ann Entomol SocAm 92:909–915

Thackeray JR, Ganetzky B (1994) Developmentally regulated al-ternative splicing generates a complex array of Drosophila parasodium channel isoforms. J Neurosci 14:2569–2578

Thackeray JR, Ganetzky B (1995) Conserved alternative splicingpatterns and splicing signals in the Drosophila sodium channelgene para. Genetics 141:203–214

Warmke JW, Reenan RAG, Wang P, Qian S, Arena JP, Wang J,Wunderler D, Liu K, Kaczorowski GJ, Van der Ploeg LHT,Ganetzky B, Cohen CJ (1997) Functional expression of Dros-ophila para sodium channels: modulation by the membraneprotein tipE and toxin pharmacology. J Gen Physiol 110:119–133

Williamson MS, Martinez-Torres D, Hick CA, Devonshire AL(1996) Identification of mutations in the housefly para-typesodium channel gene associated with knockdown resistance(kdr) to pyrethroid insecticides. Mol Gen Genet 252:51–60

133