Vision-related arrestins that are expressed in photoreceptor cells regulate both dim-light (rod arrestin, also called S-antigen, encoded by SAG) and bright-light (cone arrestin, encoded by ARR3) vision (Lamb et al., 2018). Rod arrestin is expressed in both rod and cone photoreceptor cells, whereas cone arrestin is expressed at low levels in cone cells (Craft et al., 1994; Chan et al., 2007; Nikonov et al., 2008). Except for much lower expression level compared with SAG (Chan et al., 2007), there are also reports of the pseudogenization of ARR3 in multiple mammalian lineages, including some fossorial, nocturnal and aquatic species living in low light environments (Zhou et al., 2013; Emerling and Springer, 2014; Indrischek et al., 2017, 2022; Zheng et al., 2022).

After binding to visual pigments (rhodopsin for dim-light vision and cone pigments for bright-light vision), which are light-sensing molecules in photoreceptor cells, arrestins terminate their ability to initiate phototransduction (Lamb et al., 2018). Arrestins can slow the release rate of retinal during the Meta II stage, thus allowing the reduction of toxic all-trans-retinal and protect photoreceptor cells (Sommer and Farrens, 2006; Sommer et al., 2014). At the same time, by accelerating the decay of the longstanding Meta III state, arrestins can lead to faster visual pigment regeneration and rod dark adaptation (Frederiksen et al., 2016). Moreover, visual arrestins have been reported to participate in other functions, such as the regulation of dopamine receptors in the circadian cycle (Deming et al., 2015).

Several groups of amniotes, such as penguins, sea turtles, pinnipeds and cetaceans, have independently evolved a deep-diving ability, and correspondingly specialized vision (Levenson and Schusterman, 1999; Kröger and Katzir, 2008; Reuter and Peichl, 2008). Some of these species, such as beaked whale and elephant seal (Robinson et al., 2012; Berrow et al., 2018), can dive deeper than 1,000 m below sea level, into the aphotic zone (Warrant and Locket, 2004), for predation. Interestingly, the dim-light visual pigment rhodopsin from these species have been reported to possess key phenotypic substitutions that allow fast retinal release rates, which could be adaptive for rapid dim-light sensing needed during diving (Xia et al., 2021). Both adaptive and divergent evolution of SAG has been reported in birds, reptiles and mammals (Wu et al., 2016, 2018; Schott et al., 2019), including whales (Chiu, 2019; McGowen et al., 2020), however, the role of rod arrestin in dim-light adaptation in these deep-diving taxa has not been fully resolved. Here, we performed extensive sequence analyses of amniote SAG genes, as well as the cone arrestin gene ARR3 for comparison, to determine whether adaptive evolution has occurred in rod arrestin sequences of deep-diving taxa.

After obtaining SAG and ARR3 coding sequences from 257 amniotes, we identified additional ARR3 genes, that were potential pseudogenes, from 43 mammal species, some of which had previously been reported, especially those in species with degraded color vision (Supplementary Table S1). In detail, ARR3 candidate pseudogenes were identified in 6 cetaceans, 13 rodents, 3 sirenians, Loxodonta africana, Dasypus novemcinctus, Erinaceus europaeus, Sorex araneus, Condylura cristata and Manis javanica that have frame-shifting indel(s) and/or pre-mature stop codon in their coding regions. Additionally, gene sequences from Chrysochloris asiatica, Elephantulus edwardii and Sarcophilus harrisii contain missing exon(s) segments, and in a few species, including Choloepus didactylus, Echinops telfairi, Trichosurus vulpecula and Tachyglossus aculeatus, no evidence for the gene could be found (Supplementary Figure S1; Supplementary Table S2).

In cetaceans, apart from the reported losses of ARR3 in Lipotes vexillifer and Physeter catodon (Zhou et al., 2013; McGowen et al., 2020), our searches revealed that the ARR3 genes in Balaenoptera musculus, Kogia breviceps, Mesoplodon bidens and Ziphius cavirostris are likely pseudogenes. Notably, some of these species have also lost their SWS1 and M/LWS opsin genes (Meredith et al., 2013). In baleen whales, except Balaenoptera musculus, all three other species from this genus possess intact ARR3 coding sequences. In another fully aquatic group Sirenia, we found that all three species, including the extinct Hydrodamalis gigas and the reported pseudogene in Trichechus manatus (Emerling and Springer, 2014), have lost their ARR3 gene, likely in their common ancestor as they share an inactivating mutation in exon 8 (Supplementary Figure S1). Given that dichromatic color vision exists in sirenians, in contrast to cetaceans and pinnipeds (Newman and Robinson, 2006), the loss of ARR3 in their cones is potentially compensated by the functional SAG, similar to the case in some other mammals (Emerling and Springer, 2014).

Interestingly, we report that most echolocating bats also have candidate pseudogenized ARR3 genes. We investigated a total of 46 bats and found that 35 species have likely lost this gene (7 of which need further verification), with almost all of them being echolocating bats (Figure 1). As two molossid species (Tadarida brasiliensis and Molossus molossus) have intact ARR3 coding regions, the loss of ARR3 in echolocating bats likely occurred multiple times. On the other hand, Old World fruit bats that do not have laryngeal echolocation generally have intact ARR3 coding regions, with four species (especially Eonycteris spelaea) needing further verification. In addition to the evidence from opsin sequences (Zhao et al., 2009), our findings for bat ARR3 strengthens the link between visual gene loss and sensory trade-offs in bat vision and hearing.

After excluding candidate pseudogenes, we estimated the selection pressure acting on SAG and ARR3 genes across both mammals and saurians and tested for evidence of positive selection on SAG in deep-diving lineages. For the SAG gene, multiple lineages, in both Mammalia and Sauria, have ω values greater than one when the free-ratio model was applied (p < 0.001), suggesting potential positive selection. These lineages included the deep-diving penguins, pinnipeds and cetaceans, but not the sea turtles (Figures 2A,B). Varied site-wise selection pressures for the SAG sequences in the penguin, pinniped or cetacean clades suggest diversified evolution for their dim-light related arrestin in each of the groups (Figure 2C). However, in-depth lineage-specific branch-site and two-ratio model tests on the ancestral penguin, pinniped and cetacean lineages failed to obtain evidence for positively selected sites on these three ancestral lineages (Supplementary Table S3), with only the cetacean ancestral branch having significantly different selection pressure (p = 0.02) from the background lineages (Table 1). Interestingly, in addition to validated the reported divergent evolution found in cetacean SAG (Chiu, 2019), clade model C tests further revealed that the SAG sequences from two other deep-diving groups also have significantly higher ω values compared with non-deep-diving lineages (Table 1; Supplementary Table S4). The divergent evolution of cetacean SAG has been suggested to be associated with their deep diving, and a key substitution (Q69R) has been verified for having a role in increasing the formation of rhodopsin Meta II (Chiu, 2019). In addition to their specialized rhodopsins (significantly faster retinal release rates and possibly faster dark adaptation; Xia et al., 2021; Dungan and Chang, 2022), SAG could be another gene contributing to the acute dim-light vision of deep-diving species, by regulating Meta II formation and thus possibly accelerating the rod dark adaptation (Frederiksen et al., 2016). Since the key substitution found in whales (Chiu, 2019) is not shared with penguins and seals (Supplementary Figure S2), their potential fast adaptation to deep-diving visual perception via rod arrestin might be due to different molecular mechanisms.

As a comparison to the findings for SAG, the same analyses were performed for the amniote ARR3 genes, which show much lower levels of expression in cones than the SAG (Nikonov et al., 2008). No evidence for positively selected sites in the ancestral lineages for deep-diving species was found for ARR3 (Supplementary Table S3). Interestingly, ARR3 sequences from both penguins and pinnipeds, but not cetaceans, had ω values not significantly different from those from other lineages, adding further evidence that only dim-light vision, and not color vision, is the key factor for their acute visual perception during deep diving (Table 1; Supplementary Table S5). For cetaceans, since ARR3 pseudogenes were found in some species (Supplementary Figure S1), and they have experienced widespread losses of opsin and some other visual genes (Meredith et al., 2013; Springer et al., 2016; McGowen et al., 2020), the significantly higher ω value for the ARR3 sequences is unlikely due to adaptation to color vision.

Apart from phenotypic evolution of rhodopsin in deep-diving vertebrates (Xia et al., 2021), rhodopsin function has also been measured for the ancestor of Archosaur, which was shown to have a similar transducin activation rate as that of bovine rhodopsin (Chang et al., 2002). We therefore carried out evolutionary analyses of the archosaur SAG and ARR3 sequences to determine whether their visual arrestins experienced adaptive evolution. Interestingly, SAG showed signals for positive selection, but not ARR3 (Table 2), suggesting that their rod arrestin might be an important molecule involved in their rhodopsin-mediated adaptation to dim-light. The results from both deep-diving taxa and the archosaur ancestor suggest that SAG is a critical gene for dim-light adaptation, possibly aiding in responding speed to low light environments, although functional assays are essential to fully explain the adaptations.

Widespread pseudogenization events were observed for mammalian ARR3 genes, but not SAG, suggesting a stronger functional importance for rod arrestin. Thus, accelerated evolution of SAG in penguins, pinnipeds and cetaceans could be a visual adaptation for prey seeking in quickly changing dim-light environments due to deep dives. These findings add more evidence that in addition to the dim-light visual pigment rhodopsin, rod arrestin is another critical molecule underlying their unique dim-light adaptation. Furthermore, the finding of degenerated ARR3 genes in echolocating bats and positive selection on SAG in the ancestral Archosaur suggests that the evolution of arrestins underly, at least in part, visual adaptations in diverse taxa.

The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding author.

YL designed the study. XG, YC, and YL performed data analyses. XG, YC, DI, and YL wrote the paper. All authors contributed to the article and approved the submitted version.

The study was supported by the Fundamental Research Funds for the Central Universities (GK202102006 and GK202107024 to YL and 2021CSLY017 to XG) and the Natural Science Basic Research Program of Shaanxi (2021JM-197 to YL).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

The Supplementary material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fevo.2022.1069088/full#supplementary-material