Gene-rich X chromosomes implicate intragenomic conflict in the evolution of bizarre genetic systems

Abstract

Haplodiploidy and paternal genome elimination (HD/PGE) are common in invertebrates, having evolved at least two dozen times, all from male heterogamety (i.e., systems with X chromosomes). However, why X chromosomes are important for the evolution of HD/PGE remains debated. The Haploid Viability Hypothesis posits that X-linked genes promote the evolution of male haploidy by facilitating purging recessive deleterious mutations. The Intragenomic Conflict Hypothesis holds that conflict between genes drives genetic system turnover; under this model, X-linked genes could promote the evolution of male haploidy due to conflicts with autosomes over sex ratios and genetic transmission. We studied lineages where we can distinguish these hypotheses: species with germline PGE that retain an XX/X0 sex determination system (gPGE+X). Because evolving PGE in these cases involves changes in transmission without increases in male hemizygosity, a high degree of X linkage in these systems is predicted by the Intragenomic Conflict Hypothesis but not the Haploid Viability Hypothesis. To quantify the degree of X linkage, we sequenced and compared 7 gPGE+X species’ genomes with 11 related species with typical XX/XY or XX/X0 genetic systems, representing three transitions to gPGE. We find highly increased X linkage in both modern and ancestral genomes of gPGE+X species compared to non-gPGE relatives and recover a significant positive correlation between percent X linkage and the evolution of gPGE. These empirical results substantiate longstanding proposals for a role for intragenomic conflict in the evolution of genetic systems such as HD/PGE.

Keywords: genomic conflict; haplodiploidy; insects; sex chromosomes; sex determination.

Fig. 1.

Schematic of different genetic systems discussed. (A) Male production and spermatogenesis under diplodiploidy, and various forms of male genetic haploidy, are shown. Blue and red letters indicate paternal and maternally derived material respectively. A and X represent autosomes and X chromosomes, respectively. Shown are HD, where males develop from unfertilized eggs; embryonic paternal genome elimination, where males eliminate their paternally inherited genome early in development; paternal genome silencing/elimination, a form of germline PGE where males silence paternal autosomes in somatic cells (indicated by the light blue A in males) and eliminate these chromosomes during meiosis; and a (somewhat simplified) form of gPGE observed in Sciarids, Cecidomyiids and Symphypleonan springtails, wherein males are produced by somatic loss of the paternal X chromosome(s), and the paternal genome is eliminated in spermatogenesis. (B-D) representative species with gPGE. (B) The Hessian fly gall midge Mayetiola destructor (Cecidiomyiidae), image by Scott Bauer and publicly available via the US Department of Agriculture. (C) The fungus gnat Bradysia coprophila (Bradysia tilicola; Sciaridae), image by Mike Palmer, used with permission. (D) The springtail Allacma fusca (Sminthuridae), image publicly available and taken by Andy Murray.

Fig. 2.

Frequency of X-linked and autosomal genes in gPGE species and related diplodiploid species, assessed by DNA read coverage. (A) Sciaroidea and outgroups within Bibionormorpha; topology based on Ševčík et al. (51). Plots for each Muller element show log2 male read coverage normalized by putative median autosomal coverage, with assigned X linkage (blue) and autosomal linkage (red) indicated. The y axis represents gene frequency scaled to the maximum in each distribution. Red dashed vertical lines at 0 indicate the expected autosomal coverage peak, blue dashed lines at -1 indicate the expected position of the X-linked peak, at half the coverage of the autosomes. Blue and black species names and genome-wide estimates represent gPGE and diplodiploid species, respectively. Percent estimates represent percent X linkage for each Muller and across each full genome, with error represented by 2 SD. As with the two springtails, for the Cecidomyiid M. destructor, female read data were available and thus male/female coverage is shown. In M. destructor genes are only included if assignments agree with previous physical mapping placements or were previously unassigned (35). (B) Whole genome autosomal and X linkage distributions for springtails diplodiploid O. cincta and gPGE A. fusca showing relative male/female coverage.

Fig. 3.

Number of ortholog pairs in which both genes are X linked, compared to the null expectation, for pairs of gPGE species from the same family. Within-family comparisons are shown, between-family comparisons in SI Appendix, Fig. S3. Color indicates Muller element. Muller elements for which species do not share X-linked orthologs are excluded, as is the F element. Shapes indicate significance via χ². Error bars represent 95% CIs computed from 10,000 bootstrap replicates. Observed/expected value if no association between X-linked orthologs is 1.