Radical remodeling of the Y chromosome in a recent radiation of malaria mosquitoes

Abstract

Y chromosomes control essential male functions in many species, including sex determination and fertility. However, because of obstacles posed by repeat-rich heterochromatin, knowledge of Y chromosome sequences is limited to a handful of model organisms, constraining our understanding of Y biology across the tree of life. Here, we leverage long single-molecule sequencing to determine the content and structure of the nonrecombining Y chromosome of the primary African malaria mosquito, Anopheles gambiae We find that the An. gambiae Y consists almost entirely of a few massively amplified, tandemly arrayed repeats, some of which can recombine with similar repeats on the X chromosome. Sex-specific genome resequencing in a recent species radiation, the An. gambiae complex, revealed rapid sequence turnover within An. gambiae and among species. Exploiting 52 sex-specific An. gambiae RNA-Seq datasets representing all developmental stages, we identified a small repertoire of Y-linked genes that lack X gametologs and are not Y-linked in any other species except An. gambiae, with the notable exception of YG2, a candidate male-determining gene. YG2 is the only gene conserved and exclusive to the Y in all species examined, yet sequence similarity to YG2 is not detectable in the genome of a more distant mosquito relative, suggesting rapid evolution of Y chromosome genes in this highly dynamic genus of malaria vectors. The extensive characterization of the An. gambiae Y provides a long-awaited foundation for studying male mosquito biology, and will inform novel mosquito control strategies based on the manipulation of Y chromosomes.

Keywords: Anopheles gambiae; PacBio; RNA-Seq; Y-chromosome; tandem repetitive DNA.

Fig. 1.

Summary of major Y chromosome loci, showing rapid turnover of the Y chromosome content and expression patterns in the An. gambiae species complex. (Top) Black boxes indicate Y-linkage, white boxes indicate either total absence from the species or absence from its Y chromosome, and gray boxes indicate unknown status with regard to Y-linkage. Typically, sequences indicated by gray showed CQ or RCQ values of ∼1, suggesting that they are either on both sex chromosomes or on autosomes. Details are provided in SI Appendix, Tables S6, S7, and S15. At right, the species branching order provides an evolutionary context of the changes in Y chromosome content within the past 2 My. Only YG2 is conserved and exclusively on the Y chromosome in all four species of the An. gambiae complex. (Middle) Sex-specific transcription in An. gambiae was assessed at different developmental stages and tissues, except for embryos (Em). (Bottom) The organization of the Y chromosome loci in An. gambiae, if known. ara, An. arabiensis; Car, adult carcass; Em, embryo; gam, An. gambiae; L1–L3, first- to third-instar larvae; mer, An. merus; RT, adult reproductive tissues; stDNA, satellite DNA; qua, An. quadriannulatus.

Fig. 2.

The NRY of An. gambiae mainly consists of massively amplified tandem arrays of a small number of satellites and TEs. (A) Two major regions of the An. gambiae Y, the ZAR and SAR, represent 92.3% of the sequences in Ydb (vertical bar plot). Ydb reflects the content of NRY in An. gambiae. Percentages were calculated by masking Ydb using annotated Y chromosome loci. The Left circos plot, created by homology mapping of TEs on three Y chromosome BAC clones, shows the organization of the ZAR in the three BACs. As seen in these BACs, and as independently confirmed in PacBio reads, the ZAR consists of head-to-tail tandem arrays of zanzibar, which sometimes have other transposons inserted. The arrays of zanzibar units inside each BAC are shown schematically directly inside the BAC ideograms (blue semicircular lines) enclosing the circus plot. The dark green arrows of each zanzibar unit (shown enlarged in B) represent the single LTR; lines breaking zanzibar units indicate insertions of other TEs. A few small insertions (∼200 bp) into zanzibar are too small to be visible in this plot. The asterisk in BAC10L19 corresponds to an atypical zanzibar unit that could result from recombination or misassembly. The circos plot at Right, constructed by homology mapping of satellite monomers on PacBio reads from Ydb, shows the organization of the SAR. Shown are representative examples of the occurrence of homo-monomeric tandem arrays (Y73915, Y46532, Y55593), junctions between homo-monomeric tandem arrays (Y44175), and recombinant arrays (Y2830). The recombinant arrays are interspersed with recombinant and nonrecombinant versions of AgY477 and AgY373 satellites (SI Appendix, Fig. S5). (B) Schematic of a single zanzibar unit, consisting of a gag/pol domain and a single LTR; each unit is organized in a head-to-tail tandem array (see Left circus plot in A). Shown by colored triangles are the canonical insertion sites of three other transposons (mtanga, pemba, mafia) into different zanzibar units. Percentages indicate the fraction of Ydb PacBio reads observed to carry TE insertions into zanzibar units at the precise insertion site illustrated (coordinates shown above the gag/pol domain). For example, we observed 243 of 256 (95%) PacBio reads in which pemba was inserted into zanzibar at position 2931. This phenomenon was independently confirmed in whole-genome sequencing Illumina reads. (C) Co-occurrence matrix of Y chromosome loci in PacBio reads from Ydb. These results show that satellite sequences co-occur (in the SAR), as do TEs (in the ZAR), but that the ZAR and SAR regions are largely independent.

Fig. 3.

Satellites AgY53D and AgY280 show extensive structural dynamism in males from a natural population of An. gambiae. Shown are violin plots of the log₁₀ numbers of normalized read alignments from Illumina genomic sequence derived from 40 individual male (blue) and 45 individual female (pink) mosquitoes from Cameroon, mapped to satellite monomers of AgY53D and AgY280. For comparison are similar plots of reads mapping to the presumptive male-determining gene, YG2, and to the single-copy X-linked gene, white. Numbers of read alignments to the satellite monomers varies drastically between individuals, in contrast to YG2 and white, suggesting large within-population differences in satellite abundance on the Y. Mapping reads were normalized to library size and locus length.

Fig. 4.

Physical mapping supports structural dynamism of Y chromosome sequences in the An. gambiae complex. FISH of retrotransposon zanzibar, satellite AgY53B, and gene YG5 (green signals) was performed on chromosomes of male An. gambiae Kisumu (zanzibar, YG5), An. gambiae Asembo (AgY53B), An. quadriannnulatus SANGWE, An. arabiensis Dongola, and An. merus MAF. Chromosomes were obtained from imaginal discs except for An. merus chromosomes hybridized to AgY53B, which were obtained from testes. The 18S rDNA probe (red signal) was used in all experiments except with zanzibar. Chromosomes were counterstained with DAPI (blue). (Scale bar, 2 μm; applies to all images.)

Fig. 5.

The An. gambiae X and Y chromosomes are not genetically isolated. (A) Painting of prometaphase (Left) and metaphase (Right) chromosomes from male larval imaginal discs of the An. gambiae Pimperena strain with a probe generated from microdissected Y chromosomes (labeled blue by the WGA3 kit with dNTP-Cy3). Chromosomes are counterstained with YOYO-1 (green). (Scale bars, 2 μm.) (B) Violin plots showing the log₁₀ number of normalized read alignments from 40 individual An. gambiae males (blue) and 45 females (pink) from the Cameroon population, to satellite AgY477, AgY373, AgY53A, and AgY53B monomers, compared with numbers of reads from these sources aligning to the white gene (single-copy and X-linked) and zanzibar (heavily Y-biased). Mapping reads were normalized to library size and locus length. (C) Two examples of single PacBio reads (pacbio_7224704_1 and pacbio_5551309_1) where predominantly X-linked (AgY367; shown in red) and predominantly Y-linked (AgY373, shown in blue; AgY477, shown in yellow) satellites occur in the same PacBio read. Black arrows indicate the junction between the predominantly X-linked and predominantly Y-linked satellites. The purple and orange boxes indicate inferred recombination, between AgX367-AgY373 and AgX367-AgY477, respectively. Green boxes indicate recombination between AgY477-AgY373.

Fig. 6.

Phylogeny inferred from a candidate male-determining gene on the Y chromosome, YG2, differs from the species branching order. Species topology (32) of five members of the An. gambiae complex (red branches) compared with a maximum-likelihood phylogeny inferred from a Y chromosome-specific region of the YG2 gene (green branches) sequenced from male Anopheles coluzzii (ellipse), An. gambiae (filled circles), An. arabiensis (triangles), An. quadriannulatus (open squares), and An. merus (filled squares). Samples were drawn from colonies and natural populations (see text). The YG2 tree was rooted at the midpoint; all nodes are supported by ≥95% bootstrap replicates. The topological disagreement involves An. arabiensis; in the species topology An. arabiensis is sister to An. quadriannulatus, whereas the YG2 topology indicates a sister group relationship of An. arabiensis and An. gambiae + An. coluzzii.