Genomic regression of claw keratin, taste receptor and light-associated genes inform biology and evolutionary origins of snakes

Regressive evolution of anatomical traits corresponds with the regression of genomic loci underlying such characters. As such, studying patterns of gene loss can be instrumental in addressing questions of gene function, resolving conflicting results from anatomical studies, and understanding the evolutionary history of clades. The origin of snakes coincided with the regression of a number of anatomical traits, including limbs, taste buds and the visual system. By studying the genomes of snakes, I was able to test three hypotheses associated with the regression of these features. The first concerns two keratins that are putatively specific to claws. Both genes that encode these keratins were pseudogenized/deleted in snake genomes, providing additional evidence of claw- specificity. The second hypothesis is whether snakes lack taste buds, an issue complicated by unequivocal, conflicting results in the literature. I found evidence that different snakes have lost one or more taste receptors, but all snakes examined retained at least some capacity for taste. The final hypothesis I addressed is that the earliest snakes were adapted to a dim light niche. I found evidence of deleted and pseudogenized genes with light- associated functions in snakes, demonstrating a pattern of gene loss similar to other historically nocturnal clades. Together these data also provide some bearing on the ecological origins of snakes, including molecular dating estimates that suggest dim light adaptation preceded the loss of limbs.


34
The evolutionary origin and diversification of lineages is frequently attributed to 35 key innovations of traits, but regressive evolution (Fong et al., 1995), the loss of formerly 36 adaptive traits through evolutionary time, can also be a powerful force in the shaping of 37 clades (Albalat and Cañestro, 2016). As phenotypic traits regress in a species, their serpent eye anatomy is strongly suggestive of an early period of dim light (scotopic) 64 adaptation in snake history (Walls, 1942). Various genes encoding visual and non-visual resulting probe encompassed the coding sequence, introns and flanking sequences of the date estimates from four recent studies and show that the ages of Serpentes (113-131.1 inactivated in a crown snake (Zheng and Wiens, 2016) approximately 113 Ma (point 187 estimate; range: 108-118.5 Ma; Fig. 1; ESM , Tables S4 and S5).
deleted from their genomes. Python bivittatus has intact copies of TAS1R2 and TAS1R3,

217
whereas the colubroids in this study lack functional copies of both genes: TAS1R2 is a 218 pseudogene and TAS1R3 returned negative BLAST results in all seven species. The genes 219 that flank TAS1R3 in amniotes (CPTP, DVL1) are located on the same contig in Vipera berus 220 and Ophiophagus hannah, providing strong evidence of whole gene deletion in these taxa.

371
Whereas taste buds appear to be widespread in squamates (Schwenk, 1985), snakes 372 are reported to completely lack them (Schwenk, 1985;Young, 1997). However, the absence 373 of taste buds within snakes is not without controversy (reviewed in Young, 1997), with 374 contradictory results reported even within the same species (Kroll, 1973;Young, 1997; genes in the serpent genomes and compared the distribution of retained genes to several 378 reptilian outgroups.

379
In contrast to crocodylians, testudines, and two non-serpent squamates, snakes   , 1997;Rage and Escuillié, 2000;Tchernov et al., 2000), and phenetic and 466 parsimony analyses that found serpent eyes to be similar to those of fishes (Caprette et al., 467 2004). Although an explanation for the adaptive basis of limb loss is rarely offered, 468 presumably it would have reduced friction for early swimming serpents. One piece of 469 evidence for the marine origins hypothesis that appears to have gone unnoticed is that to similar phenotypes, and unsurprisingly both have been invoked as possible selective 494 pressures to snakes' regressed visual system (e.g., Walls, 1942;Caprette et al., 2004).
evidence for this hypothesis, the only other squamates with tongues that are as strongly 525 forked as snakes are varanids, which also appear to completely lack taste buds (Schwenk, 526 1985;Young, 1997). Perhaps as the tongues of these clades became more strongly 527 associated with signaling gradients of molecules to the vomeronasal system, there was a 528 trade-off that led to a reduction in gustatory function.

529
The loss of light-associated genes in snakes would seem to have some bearing on the 530 question of serpent origins. The pattern of gene loss is highly convergent with Gekko 531 japonicus, crocodylians and mammals (Fig. 2) and OPNPT may be more broadly associated with adaptations to thermal stability (see 547 discussion in Emerling, 2016). The remaining light-associated genes are broadly retained 548 in vertebrates (Tomonari et al., 2008;Gerkema et al., 2013;Osborn et al., 2015;Emerling, 549 2016). Notably, they are typically found intact in the genomes of fishes and turtles, 550 including the marine-adapted Chelonia mydas, indicating that inhabiting a marine habitat 551 is unlikely to have led to the loss of these genes. Importantly, this neither implicates nor 552 rules out fossoriality as the basis for light-associated gene loss in snakes, but given that studies would benefit from investigating the patterns of non-visual opsin and UV-556 protection gene loss in fossorial species and contrasting it with nocturnal taxa.  (Walls, 1942), with the droplets being completely lost in some gecko 588 lineages (Röll, 2000 (1) and (3)