Across the animal kingdom, pigmentation delivers information intra- and interspecifically and thus is often under natural and sexual selection (Cuthill et al., 2017; Hubbard et al., 2010; Orteu & Jiggins, 2020; Protas & Patel, 2008). Pigmentation and patterning are critical for many key biological functions or interactions–including courtship, mate choice, thermoregulation, microbial resistance, crypsis, toxicity warning, and mimicry (Cuthill et al., 2017; Protas & Patel, 2008). These signals are often generated through a combination of pigment cells and reflective structures, and both color production pathways are under strong genetic control (Jablonski & Chaplin, 2017; Lopes et al., 2016; Luo et al., 2021; Nachman et al., 2003). As such, animal pigmentation genetics has emerged as a robust model system to study the phenotype-genotype relationship. The interest in this trait has led to the identification of hundreds of loci and pathways involved in vertebrate pigmentation (Elkin et al., 2022; Hidalgo et al., 2022; Kelsh, 2004; Lamoreux et al., 2010). However, studies on the genetic basis of pigmentation are often biased toward certain taxonomic groups (e.g., mice, humans), populations, and/or candidate loci (e.g., mc1r, agouti) (Elkin et al., 2022; Tapanes et al. 2022)). For example, ~70% of known pigmentation candidate genes emerged from mammalian studies, but only 5% are associated with fish (Elkin et al., 2022). Additional taxa and loci must be studied and identified to gain a more comprehensive and unbiased understanding of the genetic architecture of pigmentation.
Through characterization of the genetic architecture of pigmentation we gain insight into a vast number of evolutionary phenomena, including the predictability of phenotypic and genotypic evolution as well as the origins of adaptive genetic variation (Cuthill et al., 2017; Elkin et al., 2022; Martin & Orgogozo, 2013). So far, work suggests that often the same loci are co-opted across vast evolutionary scales to produce convergent pigmentation phenotypes (Crawford et al., 2017; Lamason et al., 2005; Miller et al., 2007; Saenko et al., 2015)). For example, oca2 underlies melanin deficiency in humans and corn snakes. Frequent identification of major effect genes (e.g. mc1r) in studies of specific populations (i.e., mice, European people) led to the assumption that pigmentation has a simple genetic architecture, with mutations of large effect generating key phenotypes (Hubbard et al., 2010; Protas & Patel, 2008; Quillen et al., 2019). However, recent evidence suggests the underlying genetic architecture of this trait may frequently be complex (highly polygenic) (Anderson et al., 2009; Jones et al., 2018). Pigmentation can exhibit rapid phenotypic and genetic change, quickly evolving within a handful of generations, adapting to new environments (Barrett et al., 2019; Jones et al., 2018). To fully understand pigmentation in light of evolutionary change, we must be able to robustly characterize the genetic architecture in more than a handful of organisms.
The threespine stickleback (Gasterosteus aculeautus; hereafter referred to as ‘stickleback’) offers an opportunity to study key evolutionary processes and patterns, such as—evolutionary predictability and adaptation and with ample genetic resources it is possible to characterize genetic architecture of adaptive traits. Marine stickleback repeatedly and rapidly colonized newly formed freshwater habitats at the end of the Pleistocene (≈12,000 years ago) (Bell & Foster, 1994). Within lakes and streams, stickleback independently adapted to the local ecological conditions–often diverging along a benthic-limnetic axis. Within a handful of lakes there has been the evolution of sympatric benthic and limnetic ecotypes which utilize the littoral and pelagic regions of the lakes, respectively (Schluter & McPhail, 1992). These sympatric ecotypes have diverged both genetically and phenotypically in repones to their divergent niches (Jones et al., 2012; Peichel et al., 2001; Schluter & McPhail, 1992). Notably, phenotypic divergence involves suites of trophic (Schluter & McPhail, 1992) and defensive traits (Vamosi & Schluter, 2004). However, the ecotypes have also diverged in several pigmentation traits (Clarke & Schluter, 2011; Greenwood et al., 2011; Gygax et al., 2018; Miller et al., 2007).
Limnetic fish exhibit greater ventral pigmentation (Miller et al., 2007) and more lateral barring than benthic fish (Greenwood et al., 2011). Increased brightness has been associated with increased use of limnetic resources (French et al., 2018; Bolnick & Ballare, 2020; Lavin & McPhail, 1986). Additionally, green pigmentation in the dorsal region is more prevalent in benthic fish (Clarke & Schluter, 2011; Gygax et al., 2018). Males of each ecotype also differ in their nuptial coloration—limnetic males exhibit redder throat patches relative to benthic stickleback, and often have an intensely blue iris (Boughman, 2001). Differences in pigmentation are predicted to be adaptive as there is covariance with the spectral qualities of each ecotype’s primary habitat (littoral vs. pelagic) (Clarke & Schluter, 2011; Rennison et al., 2016); and the preferred nest sites of the two ecotypes also differ in spectral quality (Boughman, 2001). The visual sensitivities of the two ecotypes exhibit divergence (Boughman, 2001; Rennison et al., 2016), suggesting differential perception of intra- and inter-specific pigment signals could contribute to pigmentation divergence (Boughman, 2001). Further, there are distinct predation regimes between the habitats (Vamosi & Schluter, 2004) and differential exposure to a vertebrate predator has been found to be associated to divergence of pigmentation (Gygax et al., 2018), suggesting selection due to crypsis may also drive the evolution of pigment differences.
Quantitative trait mapping (QTL) studies for some stickleback pigmentation traits have successfully identified candidate genes or genomic regions. So far, work using marine-freshwater pairs has characterized candidate regions for two pigmentation traits: lateral barring and ventral melanism (Greenwood et al., 2011; Greenwood et al., 2012). Candidate regions have also been found for both male and female nuptial coloration, specifically male red throat chroma was mapped in a benthic-limnetic pair (Malek et al., 2012) and red throat and pelvic spine pigmentation in females from allopatric stickleback populations (Yong et al., 2016). Yet, in general we know little about the genetic architecture of pigmentation traits of stickleback or how the genetic architecture varies across populations. More than 1,000 QTL have been identified for various stickleback phenotypes (behavioral, morphological or life history), but only 20 (1.7%) are associated with pigmentation traits. Furthermore, of the 27 threespine stickleback QTL studies included in a 2017 meta-analysis of stickleback QTL, only four studies mapped pigment traits (Peichel & Marques, 2017).
Here, we conducted a QTL mapping study of two melanin-based pigmentation phenotypes—melanophore density and lateral barring using threespine stickleback benthic- limnetic F2 crosses. We focused on these traits as there is experimental evidence that melanism and lateral barring are adaptive phenotypes, diverging in response to differential predation pressures (Gygax et al., 2018), which aids in vertebrate predator avoidance. Once candidate regions were identified, functional enrichment analyses were used to further characterize the resulting loci.