Abstract

   Color molts from summer brown to winter white coats have evolved in
   several species to maintain camouflage year‐round in environments with
   seasonal snow. Despite the eco‐evolutionary relevance of this key
   phenological adaptation, its molecular regulation has only recently
   begun to be addressed. Here, we analyze skin transcription changes
   during the autumn molt of the mountain hare (Lepus timidus) and
   integrate the results with an established model of gene regulation
   across the spring molt of the closely related snowshoe hare
   (L. americanus). We quantified differences in gene expression among
   three stages of molt progression—“brown” (early molt), “intermediate,”
   and “white” (late molt). We found 632 differentially expressed genes,
   with a major pulse of expression early in the molt, followed by a
   milder one in late molt. The functional makeup of differentially
   expressed genes anchored the sampled molt stages to the developmental
   timeline of the hair growth cycle, associating anagen to early molt and
   the transition to catagen to late molt. The progression of color change
   was characterized by differential expression of genes involved in
   pigmentation, circadian, and behavioral regulation. We found
   significant overlap between differentially expressed genes across the
   seasonal molts of mountain and snowshoe hares, particularly at molt
   onset, suggesting conservatism of gene regulation across species and
   seasons. However, some discrepancies suggest seasonal differences in
   melanocyte differentiation and the integration of nutritional cues. Our
   established regulatory model of seasonal coat color molt provides an
   important mechanistic context to study the functional architecture and
   evolution of this crucial seasonal adaptation.

   Keywords: developmental timeline, gene expression, molt cycle, seasonal
   coat color change, transcriptomics
     __________________________________________________________________

   We analyze skin transcription changes during the autumn color molt of
   the mountain hare and integrate the results with an established model
   of gene regulation across the spring molt of the closely related
   snowshoe hare. Our work provides an important mechanistic context to
   study the functional architecture and evolution of this crucial
   seasonal adaptation.

   graphic file with name ECE3-10-1180-g004.jpg

1. INTRODUCTION

   Seasonal environments impose numerous challenges to organismal
   survival. To track seasonality, individuals cycle important biological
   processes, known as phenologies (e.g., reproduction, migration,
   hibernation, or molt), allowing a better phenotypic match with seasonal
   selective pressures (Helm et al., [42]2013; Visser, Caro, Oers,
   Schaper, & Helm, [43]2010). Accurate timing of physiological and
   behavioral changes is important for fitness. Synchronizing phenotypes
   with the environment requires a precise perception of predictable cues
   (e.g., variation in day length) that anticipate seasonal change and
   trigger regulatory cascades and shifts in gene expression, with
   consequent phenotypic change (Bradshaw & Holzapfel, [44]2007).
   Quantifying gene expression changes underlying phenological traits is
   therefore crucial to understand how organisms synchronize their
   phenotypes with the surrounding environment (Ferreira et al., [45]2017;
   Giska et al., [46]2019; Jones et al., [47]2018; Schwartz & Andrews,
   [48]2014).

   Discrete seasonal molts have evolved as energy‐efficient phenological
   adaptations to regenerate hair layers and produce coats with insulating
   properties adequate for each season (Geyfman, Plikus, Treffeisen,
   Andersen, & Paus, [49]2015; Hart, [50]1956; Ling, [51]1972). In at
   least 20 vertebrate species, seasonal molts are accompanied by the
   change from summer brown to winter white coats, allowing camouflage in
   snow‐covered environments (Mills et al., [52]2013, [53]2018). As in
   other phenological traits, seasonal molting is mostly triggered by
   annual changes in day length (photoperiod) (Bissonnette & Bailey,
   [54]1944; Hoffmann, [55]1978; Lincoln, Clarke, Hut, & Hazlerigg,
   [56]2006; Zimova et al., [57]2018). Other environmental factors, such
   as temperature or snow presence, have been suggested to influence the
   rate and completeness of the molt in different mammals (Zimova et al.,
   [58]2018 and references therein). However, molt onset seems to have