e-dokumen.id

Salvemini et al. BMC Evolutionary Biology 2011, esults and DiscussionIsolation and molecular characterization of the AeadsxgeneThe molecular cloning of the Aedes aegypti dsx gene(Aeadsx) was performed by combining a classical PCRbased approach with the available bioinformatic andgenomic tools. The later used the non-sex-specificregion of the Anopheles gambiae DSX protein (236 aa)as a virtual probe for a BLASTP search of the EnsemblAe. aegypti AaegL1 genomic database (http://www.ensembl.org/) (see Methods for further strategy details).A putative male-specific cDNA sequence was thenobtained by EST analysis, using An. gambiae dsx malespecific nucleotide sequences (kindly provided by Pannuti A. and Lucchesi J., Emory University - USA, priorpublication). Using specific primer pairs for the putativemale-specific EST and common genomic sequences, weperformed an RT-PCR analysis on RNA samplesextracted from adult sexed Ae. aegypti mosquitoes. Withthis approach, we successfully amplified sex-specific products of the Aeadsx gene, with two female-specific products of 1.5 kb and 2.0 kb and a male-specific productof 1.0 kb (Figure 1B). cDNA products were cloned andsequenced, and their conceptual translation of incomplete ORFs and comparison with DSX isoforms revealedstriking conservation of the DSX DM domain and thefemale-specific carboxy domain encoded by the 2.0 kbproduct. Through 5’-3’ RACE PCR, we next obtainedlonger cDNAs containing fully open reading frames.The 3 cDNA clones were named as AeadsxF1 (2846 bp),Aeadsx F2 (2384 bp) and Aeadsx M1 (1918 bp), whichencoded three Aedes aegypti DSX isoforms, DSXF1 (278aa), DSXF2 (267 aa) and DSXM1 (548 aa), respectively.The alignment of the Aeadsx cDNA sequences with theAedes corresponding genomic sequence was used todefine the exon/intron organization and the alternativesplicing events leading to distinct sex-specific mRNAs(Figure 1A).These data confirmed the identification of the Ae.aegypti dsx gene, which produced sex-specific transcripts by alternative splicing.Aeadsx genomic organization and evolutionThe Aeadsx gene spans a very large 450 kb long genomic region, located in supercontig 1.370, and consists ofat least eight exons with seven introns that vary markedly in length from 208 to 274,879 bp.The first four Aeadsx exons (2-3a-3b-4) are nonsex-specific and encode the 248 amino acid commonN-terminus region of AeaDSX proteins. The four exonsare followed by two alternatively spliced female-specificexons (5a-5b), encoding 30 and 19 amino acid sequences,respectively, and two male-specific exons (6-7), the firstone encoding the male-specific protein domain of 300 aaPage 3 of 19and the second one constituting the 3’ UTR sequence(Figure 1A).The first four Aeadsx exons correspond by homologyto the second, third and fourth exons of the Angdsx,suggesting that we missed the real first Aeadsx exon(Figure 2). Interestingly, some of the Aedes dsx cDNAsexhibit in this second exon the presence of a nonsex-specific alternative splicing event by intron retentionof a short coding 63-nt region. The correspondingencoded 21-amino acid tract is highly conserved inAngDSX isoforms (Figure 1 and 3A) but lacks any conservation with DSX isoforms from other dipteran species (data not shown). This finding suggests that thisshort region is used in vivo and should play a functionalrole that seems to be under positive natural selection.Because the N-termini of the encoded Anopheles andAedes DSX proteins are highly conserved, starting fromthe first putative methionine, this selective constraintsuggests that the dsx translation start site is conservedin both species. As in the 5’ UTR of the isolated cDNAthere is lack of sequence corresponding to the Angdsxexon 1, and additional upstream genomic sequences ofAeadsx are likely still to be identified, including the promoter region.The Aedes exon 3a and 3b exhibit a low level ofamino acid conservation. In particular, exon 3a seems tocorrespond to the unique Anopheles exon 3 (see alignment in Figure 3). The short exon 3b encodes 16 aminoacids with no homology to DSX or other known proteins and it constitute an event of exon gain. Exon 5 inAedes is separated into two exons (5a and 5b). Theseexons (461 bp and 465 bp, respectively) maintain thefemale-specific alternative splicing regulation (Figure 2).Interestingly, it seems that dsx underwent to eitherintron gain in the Aedes or intron loss in the Anopheleslineages.Finally, Aedes exons 6 and 7 (1.1 kb and 0.5 kb) correspond to the Anopheles male-specific exons 6 and 7.The exon 6 encodes in both species the male-specificDSX portion and includes part of the 3’ UTR. The exon7 corresponds to the remaining 3’ UTR.The Aedes dsx exons show variable sequence similarity,conserved exon-intron junctions, an exon gain (3b exon)and an intron gain (intron between 5a and 5b exons)events and hence partial structural correspondence to theAnopheles dsx exons. In particular, with respect to theAnopheles homologue, Aeadsx common exons 2 and 4are largely conserved in size and content, while sexspecifically regulated exons as well as exons 3a and 3bshare a lower level of sequence identity. On the contrary,an extensive divergence in intron structure (intron position and length) was observed that reflects overall genomicdifferences between the species (Figure 2). The dsx gene inAn. gambiae is contained within an 85-kb genomic region;

Salvemini et al. BMC Evolutionary Biology 2011, age 4 of 19Figure 1 Aeadsx gene. (A) Genomic organization, splicing variants and protein isoforms of dsx in Aedes aegypti. Male-specific and femalespecific exons/protein regions are marked in blue and red, respectively. Exons and introns are not shown to scale. Translational start and stopsites and the poly(A) addition sites are marked. Rectangular striped box within exon 2 represents a 63-bp intronic sequence alternativelyremoved in Aeadsx transcripts of both sexes (see Figure 5B.3 for further details). All transcripts shown in this picture retain the 63-bp intronicsequence, which encodes an in-frame non-conserved 21-aa sequence. Transcripts without the 63-bp are not shown in this picture and insubsequent paper figures. (B) RT-PCR amplification of Aeadsx sex-specific transcripts. Primers used in this amplification are indicated as short redarrows in Figure 1A.therefore, the corresponding Ae. aegypti genomicsequence is approximately 5.3-fold larger. This differenceis due to the presence of very large introns in theAe. aegypti homologue, with an average intron size of 64kb in contrast to the observed average intron size forAngdsx (15 kb). This is undoubtedly reflective of the overall differences in genome organization of the two speciesbecause the An. gambiae genome size is about 243 Mb,while Ae. aegypti is about five-fold larger at about 1.31 Gb.Most of this difference is due to the high frequency ofrepetitive sequences in the Ae. aegypti genome [26].An analysis of Aeadsx introns with CENSOR software(http://www.girinst.org/censor/index.php) [30] revealedthe presence of multiple copies of repetitive elements,the most abundant of which are the NON-LTR/JockeyLINE-1 AA elements [31], detected in 34 copies, and theNON-LTR/SINE Feilai element [32], detected in 30copies (Additional file 1, Table S1). Interestingly, thecomparative analysis of CENSOR outputs of Aeadsxintrons revealed that two out of three sex-specificallyregulated introns, the short intron 5 (208 bp long) andintron 6 (10392 bp long), significantly deviate in thenumber of repetitive elements per kb (indicated asNoRE/kb) and in the percentage of nucleotides of repetitive elements relative to intron nucleotides (indicated asREbp). Aeadsx intron 5 contains no repetitive element at

Salvemini et al. BMC Evolutionary Biology 2011, age 5 of 19Figure 2 Comparative genomic structure of the D. melanogaster, Ae. aegypti and An. gambiae dsx genes. Comparative genomic structureof the D. melanogaster, Ae. aegypti and An. gambiae dsx genes. Green boxes represent the OD1 and OD2 domain-encoding exons. Black boxesrepresent exons encoding protein regions conserved in mosquitoes but not in fruit flies. Alternative male-specific and female-specific exons arerepresented as blue boxes and pink boxes, respectively. Green dots represent canonical acceptor/donor splicing sites. Red dots represent weakacceptor/donor splicing sites. White and green rectangles represent, respectively, TRA/TRA-2 binding sites and Nasonia dsxRE. In Drosophila, theDmdsx gene is located in a 45-kb region on chromosome 3R and is organized into six exons and five introns, with three common exonsfollowed by a female-specific and two male-specific exons. DmdsxF translation initiates at the AUG within exon 2 and terminates within thefemale-specific exon 4, while in the case of DmdsxM, translation begins at the same AUG and terminates within the first male-specific exon 5.all, while Aeadsx intron 6 presents a NoRE/kb value of0.6 and a REbp value of 7% with respect to the meanNoRE/kb and REbp values of non-sex-specific introns,which were 1.13 and 18%, respectively (Additional file 2,Figure S1). This finding suggests that there may havebeen positive selective pressure on these two intronicregions against repetitive elements to preserve sex-specific alternative splicing regulation.We used the entire supercontig 1.370 dsx-containingregion of Ae. aegypti to analyze the presence and natureof genomic microsynteny between Ae. aegypti and An.gambiae. We compared the amino acid sequences of allputative genes in Aedes supercontig 1.370 with putativegenes of the syntenic region in An. gambiae obtainedfrom the Ensembl precomputed tBLAT DNA-DNAcomparison for the two insect species.Gene density in these regions was apparently higher(twice) in An. gambiae relative to Ae. aegypti. The relative synteny quality, expressed as a percentage and calculated by dividing the number of conserved genes inboth syntenic regions by the total number of genes inboth regions [33], was 62%. Out of a total of 16 genesin these regions, 10 homologues were found includingdsx. Interestingly, microsynteny of the prospero genewith the dsx gene, previously described also for Anopheles gambiae, Apis mellifera and Tribolium castaneum[23], was not conserved in the Aedes aegypti dsxcontaining region (Additional file 3, Figure S2). Thisfinding suggests that a chromosomal rearrangementmay have occurred after the split between the two species, bringing the Aedes prospero gene into a differentgenomic position outside of supercontig 1.370 that isnot yet mapped on any Aedes chromosome. A putativeAedes prospero gene (AAEL002769) is located in supercontig 1.67, and a mapped chromosome position is notyet available for this supercontig.In Aedes aegypti, a putative gene (AAEL009111)encoding a phosphodiesterase is located in Aeadsxintron 2 (position 918233-980385 of supercontig 1.370)in the opposite direction of transcription with respect toAeadsx. The Anopheles homologue of this gene(AGAP004054) is located downstream of the prosperogene in the Angdsx-containing microsyntenic region, inthe opposite direction of transcription with respect tothe Aedes counterpart. In this case a rearrangement,occurred during evolution, moved the gene from oneposition nearby the dsx gene into the gene itself or viceversa.Comparison of AeaDSX isoformsThe alignment of AeaDSX isoforms with the DSX isoforms of D. melanogaster and An. gambiae is presented

Salvemini et al. BMC Evolutionary Biology 2011, age 6 of 19Figure 3 Multiple sequence alignment of DSX homologues. Protein sequence alignment of DSX isoforms in Drosophila melanogaster,Anopheles gambiae and Aedes aegypti. The sequences are divided into a region that is common to males and females (A), a first female-specificregion (B), a second female-specific region (C) and a male-specific region (D). The amino-terminal DNA binding (OD1) and oligomerizationdomains (OD2) are boxed in grey. The asterisk (*) indicates six amino acids whose replacements has been shown to abolish DNA-binding activityin D. melanogaster; (**) double asterisks indicate the three amino acids specific for the DSX DM domain. Intron positions are indicated by solidtriangles. The amino acid stretch marked in rectangular box corresponds to the 63-bp sequence removed in some but not all Aeadsx transcripts.This event leads to the in-frame deletion of the indicated 21-amino acid tract (see Figure 5B.3 for further details). Also the conserved removedamino acid stretch of An. gambiae is marked in rectangular box. Bold letters indicate amino acid identity in the homologous proteins. Gaps wereintroduced in the alignments to maximize similarity. The comparison of protein sequences was performed using Clustal-W (1.82).

Salvemini et al. BMC Evolutionary Biology 2011, n Figure 3. Drosophila DSX proteins essentially containtwo domains, OD1 and OD2, which serve as interfacesfor protein and DNA interactions [4,34]. The nonsex-specific OD1 is composed of an atypical zinc-fingerdomain (DM), which directly binds to target sequenceson the DNA. OD2 is an oligomerization domain thathas a common region and a female-specific portion.A sequence alignment of Aeadsx isoforms shows a highdegree of sequence conservation in the N-terminus upto the unique OD1 domain and the common part of theOD2 domain. Furthermore, within the OD1 domain, wehave found full conservation of the six residues (C, H,H, C, C, R) essential for DNA-binding activity in D. melanogaster [35] and the three residues (E, T, Q) recentlyidentified to be specific to the insect DSX DM domain[23]. These nine residues were also conserved inAn. gambiae DSXs [19].The region that links OD1 and OD2 is less conservedin Aedes and Anopheles and lacks the low complexityregion from Drosophila that contains a large number ofhistidine, glycine and alanine residues (Figure 3A).The C-terminal region of DSXF proteins shows a very highdegree of sequence conservation among different insects,and this region in Aedes aegypti, encoded by the AeadsxF1transcript, is also very highly conserved (Figure 3B).The predicted protein encoded by the AeadsxF2 transcript differs from the AeaDSXF1 protein due to a short19-aa alternative female-specific C-terminus with noobvious similarity to other DSX isoforms. However, thecorresponding nucleotide sequence encoding the shortAeaDSXF2 C-terminal region is highly conserved in theuntranslated region of the Anopheles gambiae femalespecific dsx exon 5 (Figure 3C). This strong sequenceconservation of an untranslated region of the An.gambiae dsx gene could be related to either the use ofthis region for alternative splicing events (still to beidentified in Anopheles) leading to a coding frame inAnopheles, as in Aedes, or to the involvement of thisregion in the splicing mechanism underlying sex-specificregulation of this region (see below).The predicted male-specific protein encoded byAeadsx M1 is remarkably poorly conserved, displayingonly very short stretches of similarity, as observed forDrosophila (Figure 3D) and others dipteran species (datanot shown).Finally, paralog search in the Aedes aegypti genomevia the BLASTP matching function, using the threeAeaDSX isoforms as virtual probes, revealed a singleputative paralog gene (AAEL004696). However, thisgene encodes a protein which shares with AeaDSXisoforms only a very well-conserved DM domain. ThisDM domain exhibits conservation of only one out ofthree DSX-DM domain-specific amino acids identifiedby Oliveira et al., [23].Page 7 of 19Phylogenetic relationship and molecular evolution of theAedes aegypti dsx geneAs remarked above, the dipteran DSX protein is essentially characterized by the two OD1 and OD2 domains,which constitute the most conserved portion of the dsxgene across species. The high degree of conservation ofthese two domains is expected according with their regulative role in protein-protein and protein-DNAinteractions.To determine the phylogenetic position of AeaDSXF1,we defined a combined Aedes aegypti dsx nucleotidesequence, containing regions encoding respectively thenon-sex-specific OD1 domain and the complete OD2domain (non-sex-specific portion and the female-specificportion; nucleotides 94 to 288 joined with nucleotides 607to 837 of GenBank DQ440532 CDS), and we aligned itwith the corresponding homologous regions of dsxsequences from fourteen other dipteran species: Drosophila melanogaster, Ceratitis capitata, Bactrocera tryoni,B. correcta, B. dorsalis, B. oleae, Anastrepha bistrigata,A. serpentina, A. amita, A. fraterculus, Anopheles gambiae,Musca domestica, Lucilia cuprina and Megaselia scalaris.Figure 4A shows the maximum parsimony and neighborjoining trees for this alignment. We used the lepidopteranBombyx mori and Danaus plexippus dsx sequences to rootboth trees. The resulting topologies agreed well with thetaxonomy of the order Diptera and showed high confidence levels in the groups defined. As expected, based onphylogenetic relationships, both dendrograms for the combined dsx nucleotide sequence grouped Aedes aegypti withAnopheles gambiae, and all the remaining dipteran speciesformed another group, while the lepidopteran Bombyxand Danaus representatives clustered in a basal clade.Our analysis, performed with combined OD1/OD2sequences, confirms the topologies obtained for insectsdsx genes by using the whole DSX F or DSX commonportions encoding sequence [36,37] the combined OD1/non-sex-specific OD2 portion sequences [23].We utilized the same dipteran dsx OD1-OD2 combined nucleotide sequences to perform an analysis ofthe nucleotide variation across corresponding codingregions to examine whether the sequences evolve underpurifying constraint or positive selection for amino acidchanges. These analyses were conducted on the OD1and OD2 coding sequences separately and on the OD2coding sequence partitioned into the common regionand the female-specific region (Figure 4B).Even though diffuse purifying selection was detectedin both OD1 and OD2 domains, the OD2 domainshowed a relaxation of selective constraints compared tothe OD1 domain. The difference between the dN/dSvalues was due to the non-synonymous substitutionrate, in particular in the female-specific region wheredN was significantly higher than the dN values of the

Salvemini et al. BMC Evolutionary Biology 2011, age 8 of 19Figure 4 Phylogenetic and molecular evolutionary analyses. (A) A phylogenetic tree based on the combined dsx nucleotide sequencesencoding OD1 and OD2 domains in six dipteran families and two lepidopteran species. The consensus of six equally parsimonious trees (treelength 757 and parsimony-informative characters 206) and the neighbor-joining tree (417 total characters) obtained using the combinednucleotide sequences are shown with bootstrap support above branches (shown only when greater than 50%). Taxonomic relationships areindicated in the right margin of the trees. The topology was rooted with the dsx corresponding sequences from the lepidopteran B. mori andDanaus plexippus. (B) Comparison of dipteran female-specific dsx coding sequences and localization of OD1 and OD2 domains. Pairwisesynonymous (dS) and non-synonymous (dN) substitution rates and the mean pairwise ratio (dN/dS) values are placed above the correspondingcoding sequence.OD1 domain and the common part of OD2 domain.These findings suggest that OD1-OD2 domains of theDSXF protein have stringent structure/function relationships leading to a constrained evolution. However, thefemale-specific region of the OD2 domain exhibits alower level of constraint. Amino acid changes in thefemale-specific portion of the DSX protein that affectdimerization dynamics might induce changes in thetranscriptional regulation of dsx target genes (possiblyincluding heterochronic, heterotopic, heterometric orheterotypic variations, as well as recruitment/loss of specific target genes).Hence, it is conceivable that even though DSX belongsto a highly conserved transcription factor protein family,

Salvemini et al. BMC Evolutionary Biology

dae), a dsx ortholog was previously isolated, and it maintains sex-specific regulation by alternative splicing and putative TRA/TRA-2 binding sites in the female-specific exon [18,19]. However, despite the availability of a genome sequence, it is still unclear whether dsx is also under the control of the TRA-related and TRA-2 ortho-