Functional evidence that a recently evolved Drosophila sperm-specific gene boosts sperm competition

Abstract

In many species, both morphological and molecular traits related to sex and reproduction evolve faster in males than in females. Ultimately, rapid male evolution relies on the acquisition of genetic variation associated with differential reproductive success. Many newly evolved genes are associated with novel functions that might enhance male fitness. However, functional evidence of the adaptive role of recently originated genes in males is still lacking. The Sperm dynein intermediate chain multigene family, which encodes a Sperm dynein intermediate chain presumably involved in sperm motility, originated from complex genetic rearrangements in the lineage that leads to Drosophila melanogaster within the last 5.4 million years since its split from Drosophila simulans. We deleted all the members of this multigene family resident on the X chromosome of D. melanogaster by chromosome engineering and found that, although the deletion does not result in a reduction of progeny number, it impairs the competence of the sperm in the presence of sperm from wild-type males. Therefore, the Sperm dynein intermediate chain multigene family contributes to the differential reproductive success among males and illustrates precisely how quickly a new gene function can be incorporated into the genetic network of a species.

The acquisition of new genes is a fundamental process in the origin of phenotypic novelty that facilitates adaptation in natural populations (1, 2). The incorporation of newly evolved genes in the genetic repertoire of a species depends on the fitness advantage that these new genes confer during their spread within species and, subsequently, on the action of long-standing selection pressures preventing their elimination from the genome (3, 4). Newly evolved genes associated with sex and reproduction, especially those eliciting sexual conflict, seem to be more prevalent than gene novelties unrelated to these biological functions (5–8). For example, many newly evolved genes have been shown to acquire male-biased expression in both Drosophila (9, 10) and primates (5, 11), possibly contributing to the variation in fertility among males. How these recently originated genes became integrated into pathways and how they conferred a male-related fitness advantage can be deciphered unambiguously only by performing molecular and functional analyses in which these new functions are impaired and the phenotypic consequences monitored. Although still scarce, this type of study has confirmed the contribution of newly evolved genes to sperm production (12), sperm function upon fertilization (13), sperm individualization and motility (14, 15), and male courtship (16). Among these case studies, only one (16) involves a species-specific gene. Species specificity is important because species-specific genes are the most likely to provide a precise portrait of the early stages of an evolving gene function that still retains many of its original features, which could be obscured by subsequent mutations (2). Therefore, phenotypic characterization of species-specific genes relevant for male-related functions remains crucial for understanding their role in male adaptation.

Sperm dynein intermediate chain (Sdic) is a chimeric gene originated within the last 5.4 million years in the lineage leading to Drosophila melanogaster (17). Sdic resulted from a segmental duplication that included two genes, Annexin X (AnnX) and short wing (sw) (also known as “Cdic”), followed by a series of deletions and the exonization and de novo evolution of regulatory sequences from intronic and exonic sequences of the parental genes (18, 19). The gene Sdic has been hypothesized to be present as a tandem array of 10 copies (18), although only four have been annotated (Fig. 1). Sequence comparison of the annotated copies suggests a scenario in which two duplications following the formation of Sdic gave rise to the current molecular organization of the whole multigene family. This event was estimated to have occurred 102–180 thousand years ago (20). Sdic presumably encodes a sperm-specific axonemal dynein intermediate chain. Sdic1, the copy adjacent to AnnX, has been shown to be testis-specific, and the protein that it encodes was found in both seminal vesicles and maturing spermatocytes, especially along the tails of mature sperm, suggesting a role in sperm motility. This putative role in reproductive functions makes the Sdic multigene family a suitable target for genetic incompatibilities during the incipient reproductive isolation stages in ancestral populations of D. melanogaster (18). Further, it has been proposed that Sdic is under sexual selection (21). The fast formation of the Sdic multigene family, which may be coupled with the occurrence of at least one selective sweep (18, 22, but see ref. 23), lends further support to the adaptive role played by Sdic. Two population-diversity signals substantiate the occurrence of a selective sweep. The first is the significant decrease of nucleotide variation around the genes Sdic and sw. The second hallmark is a frequency spectrum of DNA polymorphisms skewed toward rare variants in the region surrounding the gene Sdic in D. melanogaster but not in the orthologous region of Drosophila simulans (21, 22). However, there is no functional evidence supporting the relevance of Sdic to male fitness, nor was there a precise picture of the biological processes in which Sdic is involved.

Fig. 1.

Gene organization of the genomic region harboring the Sdic multigene family at 19C1 of the X chromosome of D. melanogaster. (Upper) Four copies of Sdic [Sdic1 (CG9580), Sdic2 (CG33497), Sdic3 (CG32823), Sdic4 (CG33499)] are reported in FlyBase release 5.42 (29). Upstream of each Sdic copy, there is a stretch of DNA containing the coding exons 2–4 of AnnX. One and two FRT-bearing TEs, downstream and upstream of the Sdic multigene family, respectively, were used to generate two deletions differing at their proximal breakpoints. Location of the TEs used: P{XP}d03903, X: 20,068,001 [+]; PBac{RB}e03601, X: 20,108,124 [+]; and PBac{WH}f02348, X: 20,108,283 [−] (filled and empty triangles). Deletions Df(1)FDD-0053243 (P{XP}d03903×PBac{RB}e03601) and Df(1)FDD-0053249 (P{XP}d03903×PBac{WH}f02348) encompass 40,122, and 40,281 nt, respectively. Cen, centromere; tel, telomere. (Lower) Span of the two deletions and tiling BACS in the region. Genes and the distances between them are not to scale.

Results and Discussion

We knocked out the Sdic multigene family at 19C1 on the X chromosome by inducing its deletion through a nonallelic homologous recombination event between two flippase recombination target (FRT)-bearing transposable elements (TEs) (24, 25). This approach allowed us to circumvent two difficulties that alternative gene silencing techniques such as RNAi or homologous recombination encounter because of the characteristics of Sdic. The first difficulty is the number of functional Sdic copies. Consistent with genome-wide expression surveys (26), RT-PCR experiments indicated that, in addition to Sdic1, Sdic3 also is expressed in 1- and 5-d-old w¹¹¹⁸ males (Fig. S1). The second difficulty arises from the high sequence similarity between Sdic and its parental gene sw (18), which is essential for development through the larval and pupal stages (this work and refs. 27 and 28). Specifically, this high sequence similarity can result in an undesired off-target effect if the RNAi technique is used. Thus, we selected strains carrying FRT-bearing TEs flanking the Sdic multigene family to delete it entirely while minimizing the effects on adjacent genes. We created two deletions of slightly different sizes (Fig. 1), both including the region from the most distal copy of the Sdic cluster (Sdic1) to the flanking gene immediately upstream of the cluster (sw). The insertion site of TE P{XP}d03903 is 448 nt upstream of the transcription start site of the gene AnnX, whereas the insertion sites of TEs PBac{RB}e03601 and PBac{WH}f02348 are respectively 2,809 and 2,650 nt upstream of the transcription start site of the gene Obstructor-A. According to release 5.42 of the D. melanogaster genome sequence, the two deletions span more than ∼40 kb (29). Individuals carrying X chromosomes without the Sdic multigene family are identified readily by eye color. Flies carrying the original FRT-bearing TEs are red-eyed because of their w⁺ markers, whereas deletion-bearing flies are white-eyed (in G3) (Fig. S2). The deletions generated are homozygous lethal because of the absence of the gene sw and therefore were maintained in heterozygous females, which carry the X-chromosome balancer FM7h.

We retrieved one strain carrying the deficiency Df(1)FDD-0053243 (strain A⁻) and two strains carrying the deficiency Df(1)FDD-0053249 (strains E⁻ and F⁻). Concurrently, siblings from the same mating events, but for which the generation of the deletion failed, were used to establish control strains: two (strains B⁺ and D⁺) for Df(1)FDD-0053243, and two (strains H⁺ and I⁺) for Df(1)FDD-0053249). Because the essential gene sw also is located in the deleted chromosomal stretch, it was reintroduced after pertinent mating with a sw transgene-containing donor strain (Fig. S3). The resulting males were genotyped as follows: (i) by testing for the presence of the hybrid TE resulting from the recombination between the FRT sequences leading to the deletion of the Sdic multigene family (Fig. 2 A, C, and D); (ii) by amplifying the distal breakpoint, which should be possible only in strains without the deletion (Fig. 2 A and B); and (iii) by confirming the absence of the transcripts for Sdic1 and Sdic3 (Fig. S1). The genotypes of all of the strains were confirmed.

Fig. 2.

Molecular verification of the presence or absence of the Sdic multigene family at 19C1 in males carrying X chromosomes derived from seven strains generated in the course of our experiments. Males assayed are those obtained upon the reintroduction of the sw transgene. (A) Outline of the expected outcome from different diagnostic PCRs depending on the arrangement of the X chromosome (St, the Sdic multigene family is not deleted: B⁺, D⁺, H⁺, and I⁺; Df, the Sdic multigene family is deleted: A⁻, E⁻, and F⁻). Up to two different amplicons can be amplified successfully in males carrying the St arrangement (primers in blue). These two amplicons span the distal breakpoint of the intended deficiency and differ in size. The resulting amplicon is ∼7 kb long if a TE (P{XP}d03903; filled triangle) is inserted in that location and otherwise is 165 nt long. In the latter case the X chromosome carries a TE (PBac{RB}e03601 or PBac{WH}f02348; open triangle) in the proximal breakpoint region, depending on the deficiency. For the males carrying the intended deficiency, we performed two complementary diagnostic PCRs. The amplicon generated by the primers in red confirms the presence of a hybrid TE (half-filled triangle) as a result of a nonallelic homologous recombination event that generates the deletion. The amplicon generated by the primers in green confirms the existence of an extra DNA fragment, which should correspond to the hybrid TE. (B) PCR results that confirm the absence of the deletion (two amplicons generated by blue primers) in males with the St arrangement. (C) PCR results that confirm the presence of a hybrid TE in males with the Df arrangement (amplicon generated by the primers in red). (D) PCR results that confirm the presence of an extra DNA fragment in males with the Df arrangement (amplicon generated by the primers in green). w¹¹¹⁸ males were used as a control. H₂O, negative control (no genomic DNA was added); L, ladder.

A cursory examination of the males carrying the deletion of the Sdic multigene family revealed no obvious morphological differences in the testis or impairment in sperm motility (Movies S1 and S2). Next, we tested whether the knockout lines exhibited discernible effects on male fertility. We first monitored the progeny number from days 1–8, which includes a period before sexual maturity (30) when sperm exhaustion in young males can result in adaptive differences (31). We mated w¹¹¹⁸ females with males that carried the deletions and with control counterparts (hereafter known as “experimental males”) to test for differences in progeny number and sex ratio (Fig. S3). Progeny were detected in experiments involving males without the Sdic multigene family, thus ruling out any potential problem related to the storage and fertilizing capability of the sperm of these males. Importantly, no statistically significant differences were found between males with and without the Sdic multigene family either at particular time points or over the whole timeframe examined (Fig. 3).

Fig. 3.

Fertility test for males with and without the Sdic multigene family. Different nonsimultaneous pairwise comparisons among males were performed: A⁻ vs. B⁺; E⁻ vs. I⁺; and F⁻ vs. H⁺. Males A⁻, E⁻, and F⁻ (blue), unlike the remaining males (red), do not carry the Sdic multigene family at 19C1. (A) Total progeny. No differences were found; the Student's t test. P value is given top of each comparison (df: A⁻ vs. B⁺, 11.18; E⁻ vs. I⁺, 5.98; F⁻ vs. H⁺, 20.95). The sex ratio (female to male) also was calculated, and no differences were found (Wilcoxon test; A⁻ vs. B⁺, P = 0.7983; E⁻ vs. I⁺, P = 0.9273; F⁻ vs. H⁺, P = 0.9783). The number of males that successfully participated in the experiment is indicated above each bar. (B) Progeny number over a period of 8 d. Error bars show 95% confidence intervals.

In D. melanogaster, females often mate with more than one male, leading to competition among the sperm of different males (32–36). This competition can result in differential retention and movement of sperm into the female's sperm storage organs, yielding differences in fertilization success among the males (37, 38). Because the loss of function of the Sdic multigene family might affect sperm motility, the ability of the sperm to compete might be impacted. We performed two sperm-displacement assays in which the females mated consecutively with two different males, an experimental male and a reference male (39) (Fig. S4). In the “offense” experiment, which evaluated the ability of the sperm from the experimental male (the second male in Fig. S4B) to displace or inactivate the sperm from the first male, we detected significantly lower sperm competence in males without the Sdic multigene family than in control males (Fig. 4A and Table S1). The median reduction in the proportion of progeny sired by the experimental male relative to the total progeny was 1.40–1.73% (set 1: A⁻, B⁺, w¹¹¹⁸; set 2: E⁻, I⁺, w¹¹¹⁸). These results were not caused by differences in number of progeny, which is associated mostly with the second mating (Kruskal–Wallis test; set 1: H = 2.33, df = 2, P = 0.31; set 2: H = 1.40, df = 2, P = 0.50), in good agreement with the results of the fertility test. In the “defense” experiment, which evaluated the ability of sperm from the experimental male (the first male in Fig. S4C) to resist the displacement or inactivation by sperm from the reference male, the same trend was detected, but differences between strains without the Sdic multigene and control strains were not statistically significant (Fig. 4B and Table S1). Further, from these sperm-displacement assays, we were able to assess the extent of refractoriness and remating rates of the females exposed to different experimental males (Materials and Methods). We found no evidence that the deletion of the Sdic multigene family at 19C1 had any effect on these two phenotypes (Table S2).

Fig. 4.

Sperm competence conferred by the Sdic multigene family at 19C1. Median proportion of female progeny sired by males with and without Sdic in offense (A) and defense (B) sperm-competition experiments (Fig. S4). The two competition indexes calculated in these experiments are referred to as “P2” and “P1,” respectively (Materials and Methods). Two sets of strains were assayed nonsimultaneously; a common control was included in both experiments (set 1: A⁻, B⁺, w¹¹¹⁸; set 2: E⁻, I⁺, w¹¹¹⁸). A⁻ and E⁻, males carrying the deletion; B⁺, I⁺, and w¹¹¹⁸, males carrying the St arrangement. Error bars indicate interquartile range. The datasets of both experiments were not distributed normally even upon the common arcsine square root transformation of the P1 and P2 indexes. Given the impossibility of determining the paternity of the male progeny, only female progeny were considered. The number of females that participated successfully in double-mating experiments is indicated above each bar. Although males without the Sdic multigene family at 19C1 exhibited less sperm competence overall than the males without the deletion, this difference was statistically significant only in the offense experiment (Table S1). The result of the pairwise comparisons using the Stell–Dwass test is shown at top of each comparison. The total number of female progeny scored was 18,915 for the offense experiment and 13,271 for the defense experiment.

We found evidence of fitness differences associated with the presence of the Sdic multigene family at 19C1. These differences are confined to the competence of the sperm in laboratory conditions and therefore are likely to be exacerbated in nature because of the opportunity for the female to mate with multiple males (40), as documented in D. melanogaster (41). Thus, our findings strongly support the adaptive relevance of Sdic in the context of male reproductive success in D. melanogaster, a species of a large population size (42). Further, no differences were found for the number of progeny and sex ratio in benign conditions (i.e., without competition) or in phenotypes elicited in the females in the sperm-displacement assays as a result of the deletion of Sdic. Several nonmutually exclusive explanations can account for the absence of an effect on number of progeny. The first possibility is that the Sdic multigene family is not functionally indispensable or relevant for this trait. For example, the role of the Sdic multigene family currently may be redundant or compensated by another gene in the genome with an identical or overlapping role. Indeed, some EST sequences of the gene sw, known to be involved in sperm differentiation (43), encode protein isoforms very similar to the putative SDIC protein, although their respective amino ends clearly differ (18, 19). The extent to which these differences influence the efficiency of SW protein to recapitulate fully the function of SDIC protein cannot be assessed with the available information.

A second possibility is that there are other copies of Sdic that are not annotated in the current genome assembly of D. melanogaster. We evaluated this possibility by searching for additional Sdic copies. Several lines of evidence suggest that not all of the Sdic copies have been annotated. First is the inherent difficulty of accurately reconstructing regions of repetitive nature (44). Second, most heterochromatin, including the Y chromosome, is poorly represented in the current assembly of D. melanogaster. We performed similarity searches for homologous sequences of Sdic in D. melanogaster genome release 5.40, including the collections of small scaffolds that are unmapped or involve conflicting information and often correspond to heterochromatic regions (the U and Uextra arms; ref. 45). We found evidence of Sdic-like sequences. To distinguish between redundant traces of currently annotated copies and putatively unannotated copies, we semiautomatically assembled all sequence traces with sw/Sdic similarity. Remarkably, we found nucleotide variants not consistent with those annotated as Sdic and sw in the reference genome. Specifically, the analysis of the nucleotide differences revealed eight well-supported haplotypes, including one corresponding to sw (Fig. S5). This result suggests that there are at least three extra copies of sw/Sdic, thus resulting in a total of eight, and conforms better to early estimates (18). In fact, the average coverage, assuming a single sw/Sdic gene copy, is about 120×, ∼13 times higher than the average sequence coverage estimated for the D. melanogaster X chromosome (9×) (46). Any of these extra copies might explain the lack of effect on phenotypes such as number of progeny if located outside the main cluster at 19C1 on the X chromosome. We hybridized an ∼2-kb probe of Sdic, which is almost identical in sequence to sw, to mitotic chromosomes of the D. melanogaster strains y¹; cn¹ bw¹ sp¹ and OR-R. We found only one strong hybridization signal close to the euchromatin–heterochromatin junction of the X chromosome (Fig. S6). Unless an extra copy of Sdic had diverged beyond the level of detection of the in situ hybridization technique, these results eliminate the possibility that such a copy resides outside 19C1 and indicate that all copies of Sdic, whether annotated or not, were deleted by our silencing approach.

Genes encoding sperm axonemal motor proteins have been proposed to influence the phenomenon of sperm competition (47). Our results indicate that the Sdic multigene family is one of the genes that affect sperm performance in a way that cannot be compensated by other genes. Interestingly, the Sdic multigene family is located on the X chromosome, which has been identified by association studies as an important reservoir for genetic factors influencing sperm competition in addition to the autosomes (48), where most Acp genes reside (49, 50). Sdic epitomizes the birth and integration of a new gene function into the genetic network underlying male-related fitness over a very short evolutionary time span. Taken together, our results uphold the notion that Darwinian selection actually operated in this genomic region (22) and also provide functional evidence for the relevance of Sdic in intrasexual selection, an evolutionary mechanism that fuels the steady and rapid evolution of genomes (6, 7).

Materials and Methods

Fly Husbandry.

Table S3 lists the strains used. Fly cultures were grown and maintained on dextrose-cornmeal-yeast medium at 25 °C in a temperature-controlled chamber. Fly manipulation, sorting, and scoring were carried out under CO₂ anesthesia unless otherwise stated. Dissections were performed using Ringer's solution (51). Examination of testes morphology and sperm motility was performed using phase contrast microscopy with an automated 90i Nikon motorized microscope.

Deletion Generation and Molecular Genotyping.

Deletion of the Sdic multigene family was performed following as described previously (25). To verify the presence/absence of the deletion of the Sdic multigene family, genomic DNA from males obtained upon the reintroduction of the sw transgene was extracted (25), and its integrity was verified by gel electrophoresis. Red Hot TaqDNA polymerase (ABgene) and TaKaRa Ex Taq DNA polymerase (TaKaRa) were used for amplifying fragments shorter and longer than 2.5 kb, respectively. Presence of a recombinant hybrid TE that has DNA stretches of the two original TEs can be present only in males carrying the deletion (25). Likewise, primers designed outside the chromosomal stretch to be deleted can amplify only in males carrying the deletion. Conversely, primers designed flanking the distal TE insertion, i.e., P{XP}d03903, can amplify only in males that do not carry the deletion and should encompass P{XP}d03903. Because only one of the two chromosomes with the original TEs can be recovered in any given female (Fig. S2, G3), two amplicons of different sizes are possible. Table S4 lists the primers used and the expected size of the amplicons.

RNA Preparation and RT-PCR Experiments.

Males obtained upon the reintroduction of the sw transgene were separated from females at the time of collection, allowed to age (1 or 5 d), snap-frozen in liquid nitrogen, and kept at −80 °C until needed. Total RNA extraction was performed with TRIzol (Invitrogen) following the manufacturer's instructions, and its integrity was verified by gel electrophoresis. To avoid genomic DNA contamination, 20 μg of total RNA of each sample was treated with Turbo DNA-free (Ambion). Poly(A)+ mRNA was reverse-transcribed with an oligo(dT) using SuperScript II reverse transcriptase (Invitrogen) in the presence of RNase Inhibitor (Roche). PCRs were done using Red Hot TaqDNA polymerase (ABgene) to detect expression of the genes Sdic1 and Sdic3 (Table S4).

Male Fertility Test.

We compared strains with and without the Sdic multigene family: A⁻ against B⁺; E⁻ against I⁺; and F⁻ against H⁺. We focused on three parameters: total progeny, number of progeny over time, and sex ratio. Individual naive males from crosses between the self-maintained deficiency stocks and sw transgene-containing donor strain were isolated within 6–8 h after eclosion and aged for 24 h before mating with three 4- to 5-d-old virgin w¹¹¹⁸ females in fresh vials. Subsequently, each male was transferred to a different vial with three different virgin females every day for 8 d. The inseminated females were left in their original vials for 3 d after mating and then were transferred to fresh vials every 3 d until they were discarded on the ninth day. The offspring eclosing from days 1–8 were collected and scored phenotypically until no more individuals emerged. For the strains interrogated, 10–15 males were tested initially, although the data collected correspond only to males that survived throughout the 8-d period. Flies were transferred by aspiration.

Sperm Competence Test.

We essentially followed the procedures in ref. 39 for the two sperm-displacement assays performed, the offense and defense experiments. For each experimental male by experiment combination, we used 60–70 females. The offense experiment was done by first mating virgin white-eyed females, w¹¹¹⁸, with red-eyed wild-type OR-R males in mass mating for 2 h on day 1. Next, each individual female was transferred into a fresh vial (vial 1), and females were allowed to oviposit for 2 d. On day three, three males were added to vial 1 and allowed to mate overnight. The next morning, females were transferred into fresh vials (vial 2), and males were discarded. The females were left in vial 2 for 3 d and then were transferred into new vials (vial 3), where they were kept in for another 3 d before being discarded. For the defense experiment, we followed the same experimental design except that the experimental male mated first, and the red-eyed wild-type male mated second. For both types of experiments, vial 1 was examined for the presence of offspring to confirm that successful mating had occurred with the first male; females with an unsuccessful first mating were excluded from downstream analysis. Females with a successful first mating had their offspring recorded and scored for eye color on days 13 and 17 after the initiation of oviposition in vials 2 and 3; these offspring actually reflect the competition between sperm from different males. Females that died or did not produce progeny from the second male were detected and also were excluded from downstream analysis. Male offspring were not considered, because those sired by the first and second males were phenotypically indistinguishable. For each female included in downstream analyses, we calculated a sperm competition index depending on the type of experiment: P1 for the defense experiment and P2 for the offense experiment. P1 gauges the proportion of progeny sired by the experimental male when it was the first to mate compared with the total progeny sired by both the first and second males; P2 is equivalent to P1 but compares the progeny sired by the experimental male when the experimental male was second rather than first to mate (52, 53). The distributions of P1 and P2 values were examined using the Shapiro–Wilk test before and after arcsine square root transformation (54, 55). Differences in the sperm competence among experimental males, as inferred from P1 and P2 values, were assessed using the Kruskal–Wallis test; Steel–Dwass tests were used in pairwise comparisons accounting for multiple testing. Flies were transferred by aspiration.

Remating Ability and Refractoriness Tests.

“Remating ability” refers to the proportion of females that remated with the experimental male when the experimental male was the second to mate in the offense displacement assay. “Refractoriness” refers to the proportion of females that did not remate after their first mating with the experimental male in the defense displacement assay. Differences in these two proxies were assessed using the χ² test followed by randomization tests of independence based on 100,000 Monte Carlo simulations.

Computational Searches for Extra Copies of Sdic.

The presence of additional copies of Sdic was evaluated by performing TBLASTN, BLASTN (56), and HMMER3 (57) similarity searches (cutoff e-value of 10⁻⁵) against the D. melanogaster reference sequence release 5.40, including the U and Uextra arms. The U arm contains scaffolds from Celera that could not be allocated into the D. melanogaster chromosomes or that have conflicting localization (45). The Uextra arm consists mostly of small scaffolds that could not be consistently joined to existing larger ones (29). To conduct the HMMER3 similarity searches, we first used MAFFT (v. 6.846) (58) to align the nucleotide and amino acid sequences of the four annotated Sdic copies. For the amino acid alignment, we selected the larger Sdic1 isoform. We then used the gene and protein alignments to build the two corresponding Hidden Markov Model profiles for nucleotide and amino acid sequences.

We inferred the minimum number of Sdic gene copies by analyzing the patterns of nucleotide differences across all D. melanogaster Sdic-like sequence traces. Given that the D. melanogaster reference sequence derives from a highly inbred individual (45), we estimated the minimum number of sw/Sdic copies as the different number of haplotypes identified among the sw/Sdic traces. To do so, we conducted sw/Sdic similarity searches against the D. melanogaster trace archive using a cutoff e-value of 10⁻²⁰⁰. We mapped all these traces (BLAST hits) to the annotated Sdic and sw genes using Burrows-Wheeler Aligner (59). To remove possible sequencing errors, which would artifactually increase the number of new haplotypes, we masked all singleton variants and calculated the number of haplotypes along the alignment (sliding window) using VariScan (60).

Chromosome Preparations and FISH Analysis.

Mitotic chromosomes from larval neuroblasts were prepared as described (61). DNA probes were labeled by nick translation using Biotin-Nick Translation Mix (Roche) or Digoxigenin-Nick Translation Mix (Roche). Hybridization signals were detected using FITC-conjugated avidin (Vector Laboratories) or with rhodaminated antidigoxigenin antibodies. Mitotic chromosome preparations were stained with DAPI. FISH procedures were performed as reported (62). Biotinylated anti-avidin antibodies were used to amplify the hybridization signal. Digital images of FISH signals and DAPI staining were recorded separately by CCD camera and were pseudocolored and merged using the Adobe Photoshop CS3 program.

Statistical analyses were performed using JMP 4.1 and SAS 9.2 (SAS Institute).