Report

Male sex in houseflies is determined by Mdmd, a paralog of the generic splice factor gene CWC22

See allHide authors and affiliations

Science  12 May 2017:
Vol. 356, Issue 6338, pp. 642-645
DOI: 10.1126/science.aam5498

Disrupting housefly gene reverses sex

Sex comes in many forms, even when considered at the molecular level. In different animals, the chromosomes and specific genes that function in sex determination vary widely. As a case in point, the familiar housefly displays a highly variable sex determination system. In this animal, the male determiner (M-factor) instructs male development when it is active, but female development results when it is inactive. Sharma et al. now identify the housefly M-factor, which arose via the co-option of existing genes, gene duplication, and neofunctionalization. The findings elucidate the remarkable diversity in sex-determining pathways and the forces that drive this diversity.

Science, this issue p. 642

Abstract

Across species, animals have diverse sex determination pathways, each consisting of a hierarchical cascade of genes and its associated regulatory mechanism. Houseflies have a distinctive polymorphic sex determination system in which a dominant male determiner, the M-factor, can reside on any of the chromosomes. We identified a gene, Musca domestica male determiner (Mdmd), as the M-factor. Mdmd originated from a duplication of the spliceosomal factor gene CWC22 (nucampholin). Targeted Mdmd disruption results in complete sex reversal to fertile females because of a shift from male to female expression of the downstream genes transformer and doublesex. The presence of Mdmd on different chromosomes indicates that Mdmd translocated to different genomic sites. Thus, an instructive signal in sex determination can arise by duplication and neofunctionalization of an essential splicing regulator.

Genetic mechanisms for sex determination are not conserved among organismal groups. Illustrating this diversity, in insects, systems vary among species at the chromosomal, gene, and gene-regulation level (13). For example, in insects with male heterogamety (XX-XY system), sex can be determined by a dominant Y-linked gene or by X-chromosome dosage (4). The transformer (tra) and doublesex (dsx) genes are conserved elements of the insect sex determination pathway, but the upstream instructive signals vary (57). The polymorphic sex determination system of the housefly, Musca domestica, reflects this diversity in regulation and genes (811). Males can carry a dominant male determiner (M-factor) on the X or Y chromosome or any of the five autosomes (10, 1214).

The M-factor acts as the instructive signal for male development in the housefly. It regulates transformer (Md-tra), a binary switch that directs female differentiation when active and male differentiation when inactive. Md-tra is regulated at the splicing level. The active state of Md-tra is initially established by maternally provided Md-tra. Once activated, zygotic Md-tra will perpetuate its female-promoting function by a positive splicing feedback loop throughout development. Paternally inherited M-factor prevents this maternal activation of the zygotic Md-tra self-regulatory loop. The early embryonic presence of male-specific splice products of Md-tra indicates that this regulation has already started at the cellular blastoderm stage (11).

We hypothesized that the M-factor gene encodes a product present only in early male embryos to prevent establishment of Md-tra function. Exploiting Musca genetics, we isolated and sequenced RNA from unisexual embryos (fig. S1). Among the top 14 male-specifically expressed sequences that were absent in the female M. domestica genome assembly (15), we identified five orphan contigs of the same transcription unit (Fig. 1A and table S1), which we termed Mdmd (for Musca domestica male determiner). Subsequent analysis revealed that these sequences are present in males that carry an M-factor on chromosome Y, II, III, or V (Fig. 1B). Reverse transcription polymerase chain reaction (RT-PCR) amplification confirmed the exclusive presence of Mdmd transcripts in male embryos (Fig. 1C). Zygotic Mdmd transcripts first appear in 2- to 3-hour-old embryos (cellularized blastoderm stage), coinciding with the first zygotic expression of Md-tra (11). Mdmd expression is then maintained throughout male development until adulthood (Fig. 1D). Mdmd encodes a protein with high homology to CWC22 (complexed with Cef-1), also known as NCM (nucampholin), a spliceosome-associated protein that is required for precursor mRNA splicing and exon junction complex (EJC) assembly (16) (Fig. 1E). A BLAST survey of Mdmd over female genome scaffolds (15) identified a paralog (LOC101896466) of Mdmd that is structurally closely related to ncm genes of other insect species. In contrast to Mdmd, ncm is present and expressed in both sexes (Fig. 1, B to D). On the basis of its high sequence identity to the ncm gene in Drosophila and its conserved synteny evidenced by linkage to bicoid stability factor, we refer to this autosomal gene as the ortholog of ncm and name it Md-ncm. Mdmd shares a high degree of amino acid identity with Md-NCM in the MIF4G (85%) and MA3 (79%) domains and flanking sequences, but it displays a substantial level of divergence in the N-terminal and C-terminal regions (Fig. 1A and fig. S2). Sequence alignments reveal that Md-ncm groups with prototype ncm genes of other insect species. However, the Mdmd sequences from different M. domestica strains form a distinct outgroup, suggesting that after the duplication event, Mdmd rapidly diverged from Md-ncm (Fig. 1E and fig. S3).

Fig. 1 Mdmd is a male-specific paralog of the M. domestica CWC22 ortholog, Md-ncm.

(A) Comparison of the two paralogs, Mdmd and Md-ncm. Mdmd was initially identified by extending five male-specific RNA contigs (lines above Mdmd). Each exon contains highly conserved MIF4G and MA3 domains. Nucleotide identity is indicated in percentages. (B) Genomic amplifications with paralog-specific end primers show that Mdmd sequences are present only in males of XY, MII, MIII, and MV strains, whereas Md-ncm is present in both males and females (+/+) of each strain. (C) RT-PCR confirming the presence of Mdmd transcripts in 1- to 5-hour-old male embryos (MIII/+) but not in female embryos (+/+). (D) Developmental expression profiles of Mdmd and Md-ncm based on RT-PCR with intron-spanning primers. The upper bands in both profiles correspond to unspliced RNA and/or genomic DNA contamination, whereas the lower bands represent spliced transcripts. (E) Neighbor-joining phylogenetic tree (branch label, percent consensus support) for Mdmd and NCM/CWC22 proteins (see also the MrBayes tree in fig. S3B). Full species names and sequences are listed in the supplementary materials. L, larval stage; bp, base pairs.

Multiple nonfunctional copies of Mdmd were found next to the Mdmd gene in the genome of the MIII (M-factor located on chromosome III) strain (fig. S4). These copies may have arisen from local amplification to preserve Mdmd functionality in a nonrecombining region (fig. S4). Because of its long open reading frame (ORF), Mdmd is particularly vulnerable to the accumulation of deleterious mutations. We identified a similar arrangement of multiple Mdmd copies in MII, MV, and MY males (fig. S4). This suggests that the various M loci originated from a common ancestral Mdmd sequence, which first locally multiplied and then translocated as a cluster to different sites in the genome (fig. S5).

On silencing of Mdmd by injecting double-stranded RNA (dsRNA) into syncytial embryos of different M strains, all of the surviving M-factor–carrying individuals developed externally as males, but 56 to 88% contained fully differentiated ovaries instead of testes, with the notable exception of MI males (Fig. 2, A to C, and fig. S6). From this result, we infer that Mdmd is essential for specifying the male gonadal and germline fate, which is consistent with genetic findings that M-factor and its target Md-tra govern the sexual identity of both soma and germ line (11). Incomplete feminization may be explained by the transient nature of embryonic RNA interference. A 70% reduction of Mdmd transcript levels was observed in MIII/+ embryos 10 hours after dsRNA injection, whereas after 20 hours, levels were comparable to those in control individuals, suggesting a recovery of Mdmd expression (Fig. 2, D and E). Given that substantial levels of Mdmd transcripts were also detected in nongonadal tissues of male adults with ovaries, restored activity of Mdmd at late stages apparently prevented systemic female differentiation (Fig. 2E). To conclusively test whether Mdmd is required for overall male differentiation, loss-of-function alleles were generated in Mdmd coding sequences by nonhomologous end joining–mediated disruption with Cas9. On targeting Mdmd in the MIII strain, we recovered 59 fertile males, of which at least 10 sired female progeny carrying dominant markers tightly linked to the MIII locus, indicating loss of its male-determining function (fig. S7). These M-factor–containing individuals are phenotypically normal fertile females (Fig. 3A). Sequence analysis confirmed that these females carry structural aberrations in the Mdmd cluster (Fig. 3B). Lines M32 and M36 are most informative because the lesions specifically disrupt the ORF of Mdmd (Fig. 3C) and only abolish the protein-coding function of this Mdmd copy. We conclude that the Mdmd gene is indispensable for normal male development and may be the only gene in the cluster providing male function. Consistent with the role of M-factor as an upstream repressor of Md-tra, individuals that have Mdmd abolished by CRISPR-Cas9 exclusively express the female splice variants of Md-tra and Md-dsx (Fig. 3D).

Fig. 2 Embryonic silencing of Mdmd is transient and leads to ovarian differentiation in males.

In (A) to (C), MIII/+ individuals were injected with dsRNA against Mdmd. Scale bars, 1 mm. (A) Adult abdomen with male genital structures [claspers (cl) indicated by arrows] and, inside, fully differentiated ovaries (ov, arrowhead). (B) Dissected ovaries from the same male. (C) DAPI (4′,6-diamidino-2-phenylindole)–stained ovaries containing normal cysts composed of nurse cells (nc) and egg chambers (ec). (D) Relative levels of Mdmd mRNA 10 and 20 hours (h) after injections with dsRNA against Mdmd and dsRNA against M112 control in MIII/+ male embryos. (E) RT-PCR analysis of Mdmd transcripts and female transcripts of Md-tra (Md-traF) in normal (+/+) ovaries and (MIII/+) testes (tes) and in Mdmd dsRNA-injected (MIII/+) gonadectomized bodies (gb), testes, and ovaries.

Fig. 3 CRISPR-Cas9–induced disruption of Mdmd causes complete male-to-female transformation.

(A) F1 female of line M32 with pw+, bwb+ phenotype (left); male sibling with pw+, bwb+ phenotype (middle), and female sibling with pw, bwb phenotype (right). The phenotypic markers are described in the materials and methods and in fig. S7A. (B) CRISPR-Cas9–targeted sites sgF3 and sgFA in Mdmd (red stripes, top) and genomic amplifications of Mdmd and Md-ncm in F1 females of lines M6, M29, M31, M32, and M36 (bottom). In the upper blot, F1-R4 primers were used to amplify the ORF of Mdmd on chromosome III; in the middle blot, 1s-1as primers were used to amplify the 5′ region of different Mdmd copies; and in the lower blot, a primer pair was used that specifically amplifies Md-ncm. Absence of F1-R4 amplicons in M6, M29, and M31 indicates large deletions. (C) In M32 females, a deletion of 14 bp uncovers the sgF3 target site upstream of the MIF4G domain, causing a frame shift. In M36, a deletion of 146 bp removes the same target site and extends into the MIF4G domain. The target sequence is labeled in green in the wild-type reference sequence (pink letters, manually aligned for best fit; blue shading, sequence of mutant allele). (D) Expression of Md-tra and Md-dsx in sex-reverted females of lines M6, M29, M30, M31, and M32. Female splice variants are absent in control males (MIII/+) but present in control (+/+) and sex-reverted females. The male splice variant of Md-dsx, Md-dsxM, is only detected in control males (MIII/+). Expression of cytochrome P450 (CyP) was used as an internal standard.

On the basis of sequence similarity, we infer that Mdmd is a paralog of Md-ncm (CWC22), which, as noted, encodes a spliceosome-associate protein that is indispensable for the assembly of the EJC (16, 17). The essential functions of CWC22 are likely to be provided by Md-ncm, given that embryonic silencing of this gene leads to early lethality in both males and females (fig. S8). However, the effect of EJC on splicing is limited to certain genes (18). Changes in expression levels of EJC components also affect the splice site selection of alternatively spliced genes (19). Considering that tra is one of the targets on which EJCs preferentially assemble in Drosophila (18), it is conceivable that Md-ncm plays a crucial role in the splicing regulation of Md-tra. Because the target of M-factor, Md-tra, is alternatively spliced, this posttranscriptional regulatory function makes Mdmd an excellent candidate M-factor. Mdmd may act as a direct regulator of Md-tra by selectively promoting the male or preventing the female splicing mode. Alternatively, the high level of sequence similarity to its paralog opens the possibility that Mdmd behaves as a dominant-negative, interfering with the functions of Md-ncm in promoting female splicing of Md-tra. Further study needs to elucidate the precise role of this gene in Md-tra splicing and can contribute to a better understanding of alternative splicing regulation.

There likely exists a large source of primary signal genes, which contribute to the high diversity of sex-determining mechanisms in insects. Recently, two male determiners were characterized in mosquitoes, Nix in Aedes aegypti (20) and Yob in Anopheles gambiae (21). These genes show sequence homology neither to each other nor to Mdmd, further pointing toward the species-specific acquisition of novel male determiners in insects. Moreover, Mdmd appears to be absent in the M. domestica strain that has an M-factor mapped to chromosome I (fig. S4), suggesting that even intraspecific variation exists at the level of the primary signal. Because insect sex determination is based on alternative splicing, its role in splicing regulation may have preequipped ncm for attaining a sex determination function. The recruitment of a CWC22 duplicate for male function may be unique to the housefly, given that ncm paralogs have thus far not been found in other higher dipterans. Our study thus demonstrates that novel genes originating from duplication and neofunctionalization can adopt critical roles in essential developmental processes.

Supplementary Materials

www.sciencemag.org/content/356/6338/642/suppl/DC1

Materials and Methods

Figs. S1 to S8

Tables S1 and S2

Protein Sequences

References (2239)

References and Notes

  1. Acknowledgments: We thank E. Geuverink and A. H. Rensink for technical advice, A. Meccariello and G. Saccone for providing purified Cas9 protein, and M. Hediger for helpful advice on housefly genetics. This work was supported by the Netherlands Organisation for Scientific Research (grant no. ALW 822.02.009) and an Ubbo Emmius grant to L.W.B.; the U4 Network sponsored by the DAAD (German Academic Exchange Service) “Strategic Partnerships” program and the German Federal Ministry of Education and Research to Georg-August-University Göttingen; and the Swiss National Science Foundation (grant no. 31003A_143883) to M.D.R. We thank M. Schenkel, S. Visser, G. Saccone, and the Groningen Evolutionary Genetics group members for fruitful discussion. The nucleotide sequences of Mdmd genes from different M strains have been deposited in the DNA Data Bank of Japan/European Molecular Biology Laboratory/GenBank database (accession numbers KY020047, KY020048, KY020049, and KY020050). Deep sequencing data are available from ArrayExpress (accession number E-MTAB-5080).
View Abstract

Stay Connected to Science

Navigate This Article