Introduction
Supergenes are regions of the genome that contain multiple genes held in strong linkage disequilibrium due to severely restricted recombination (Schwander, Libbrecht, & Keller, 2014). They are often formed by chromosomal rearrangements, such as inversions, which inhibit crossing over during meiosis, thereby facilitating the preservation of favorable allelic combinations across numerous loci (Faria, Johannesson, Butlin, & Westram, 2019; Schwander et al., 2014). Chromosomal rearrangements are fundamental to sex chromosome evolution (Hoffmann & Rieseberg, 2008; White, Kitano, & Peichel, 2015) and are now recognized as taxonomically widespread, evolutionarily significant mediators of complex trait variation (Faria et al., 2019; Schwander et al., 2014; Wellenreuther & Bernatchez, 2018). For example, supergenes have been discovered that determine variation in social behavior, sexual dimorphism, and dispersal strategies in animals as diverse as insects (Purcell, Brelsford, Wurm, Perrin, & Chapuisat, 2014; Wang et al., 2013), birds (Küpper et al., 2016; Lamichhaney et al., 2015; Tuttle et al., 2016), and fishes (Kirubakaran et al., 2016; Pearse et al., 2019).
Despite their acknowledged role in facilitating the evolution of complex trait polymorphisms, supergenes are not without evolutionary costs. Diminished recombination and resulting linkage disequilibrium within supergenes are expected to impede the ability of selection to act on individual mutations, with the result that spread of beneficial mutations and elimination of deleterious ones are constrained by the genetic background on which they arise (Hill-Robertson interference) (Hill & Robertson, 1966; Wang et al., 2013). The subsequent reduction in the efficacy of natural selection typically culminates in some level of degeneration of the non-recombining region, usually manifested as accumulation of fixed deleterious mutations, loss of genetic diversity, suppression of gene expression, gene deletion, and/or proliferation of transposable elements (Bachtrog et al., 2011; Stolle et al., 2019; Tuttle et al., 2016). Thus, the complex makeup of supergenes as mosaics of tightly linked adaptive, neutral, and even deleterious genetic variants makes dissection and elucidation of their genotype to phenotype map especially challenging.
The fire ant Solenopsis invicta displays naturally occurring variation in fundamental, ecologically important traits comprising two social syndromes (monogyny and polygyny) that collectively are under the control of a supergene (Wang et al., 2013). The eponymous difference in the syndromes is the number of reproductive (egg-laying) queens in a colony, with colonies of the monogyne form containing only a single reproductive queen and colonies of the polygyne form containing few to many queens (Gotzek and Ross 2007; Tschinkel 2006). The difference in queen number corresponds with differences in numerous individual-level attributes including: the extent of energy reserves, types of dispersal behaviors, and reproductive ontogenies of young queens (C. DeHeer, Goodisman, & Ross, 1999; C. J. DeHeer, 2002); the fecundity of mature reproductive queens (DeHeer 2002; Tschinkel 2006); and the form of worker discrimination directed toward queens attempting to join a colony (Keller & Ross, 1998). The S. invicta supergene regulating social form arose with the appearance of multiple adjacent chromosomal inversions that span more than 13Mb of DNA on the “social chromosome” and include over 400 mapped protein-coding genes (Y. Huang, Dang, Chang, Wang, & Wang, 2018; Wang et al., 2013; Yan et al., 2020). Recombination is dramatically reduced between the inverted region of the social chromosome (denoted as the Sb haplotype) and its homologs that show synteny with the presumed ancestral haplotype (denoted SB ) (Stolle et al., 2019; Wang et al., 2013; Yan et al., 2020). Moreover,Sb/Sb homokaryotypes (homozygotes) in the invasive U.S. range effectively do not survive to reproduce (Hallar, Krieger, & Ross, 2007); thus, like the mammalian Y-chromosome, recombination is strongly reduced between variants of the Sb haplotype as well as between the Sb and SB haplotypes. The entirety of the fire ant supergene can be characterized as an enormous, complex genomic module that collectively regulates colony social form and allied traits while propagating via simple Mendelian inheritance (Keller & Ross, 1998; Wang et al., 2013).
Despite a recent surge of interest in the importance of supergenes in regulating complex trait variation (Schwander et al., 2014; Wellenreuther & Bernatchez, 2018), progress in disentangling the functional mechanisms by which they exert their phenotypic effects has been slow. As suggested above, supergenes may harbor relatively few genes with direct functional effects on adaptive traits that reside in a sea of elements with neutral or even deleterious fitness consequences, and identification of such loci is complicated by the strong linkage disequilibrium observed across the supergene region. Moreover, supergene-associated traits may be regulated by a combination of protein structural variants and variant DNA regulatory elements, necessitating careful investigation and evaluation of the role of each type, as well as possible joint effects (e.g., allele-specific expression of protein-coding genes). In S. invicta , recent studies have defined the specific inversion breakpoints in the Sb haplotype (Y. Huang et al., 2018; Yan et al., 2020) and have begun to characterize protein sequence divergence between the Sb and SB haplotypes in both the invasive U.S. range (Pracana, Priyam, Levantis, Nichols, & Wurm, 2017) and native South American range (Yan et al., 2020). However, fundamental questions remain as to how differences between the supergene haplotypes, ranging in scope from single nucleotide polymorphisms, to copy number variants, to major structural rearrangements, influence gene regulation and induce alternate complex social phenotypes (Fontana et al., 2019; Gotzek & Ross, 2007; Y. C. Huang & Wang, 2014).
To explore in greater detail the genotype to phenotype relationship between the Sb supergene haplotype and the fire ant polygyne syndrome, we generated and analyzed RNA-seq data from brains and ovaries of individual, sexually mature, similarly-aged unmated queens (gynes) of both social forms of S. invicta . Organ-specific sampling was utilized to minimize the effects of variable tissue composition, with brains and ovaries targeted because they are likely the primary organs mediating the behavioral and reproductive differences between the social forms. Samples of gynes with all three social chromosome genotypes (SB/SB, SB/Sb, and Sb/Sb ) were included in our experiment (Figure 1), which is noteworthy by virtue of inclusion of the rare gynes from polygyne nests bearing either homozygous genotype. We focus on gene expression effects of the two most conspicuous factors relevant to fire ant social organization: an individual’s social chromosome (supergene) genotype and its natal colony social form (developmental environment).