Nocturnality is a key evolutionary innovation of mammals that enables mammals to occupy relatively empty nocturnal niches. Invasion of ancestral mammals into nocturnality has long been inferred from the phylogenetic relationships of crown Mammalia, which is primarily nocturnal, and crown Reptilia, which is primarily diurnal, although molecular evidence for this is lacking. Here we used phylogenetic analyses of the vision genes involved in the phototransduction pathway to predict the diel activity patterns of ancestral mammals and reptiles. Our results demonstrated that the common ancestor of the extant Mammalia was dominated by positive selection for dim-light vision, supporting the predominate nocturnality of the ancestral mammals. Further analyses showed that the nocturnality of the ancestral mammals was probably derived from the predominate diurnality of the ancestral amniotes, which featured strong positive selection for bright-light vision. Like the ancestral amniotes, the common ancestor of the extant reptiles and various taxa in Squamata, one of the main competitors of the temporal niches of the ancestral mammals, were found to be predominate diurnality as well. Despite this relatively apparent temporal niche partitioning between ancestral mammals and the relevant reptiles, our results suggested partial overlap of their temporal niches during crepuscular periods.
Mammals are hypothesized to have experienced selection for nocturnality during their early evolution1,2,3,4 and some argue nocturnality emerged in the Late Carboniferous almost 300 mya5, well before the origin of Mammalia. The nocturnality of the ancestral mammals is primarily inferred from the enhanced hearing and olfaction based on morphologic traits of Mesozoic fossil mammals6,7 and reduced daytime vision and activity(e.g., dichromatic color vision) in most living mammals8. Phenotypic traits preserved in early mammalian fossils such as diminutive body size, a less sprawling body posture, secondary palate and turbinals, etc. are consistent with the evolution of endothermy, which would enable mammals to be active at night when temperatures are lower and are consistent with the nocturnality hypothesis9. Evolution of increasing nocturnality of ancestral mammals is hypothesized to be a result of competition and/or predation of reptiles, which are predominately diurnal due to their ectothermic dependence on external warmth to achieve optimum activity1,3,10,11. While these studies have advanced our understanding of the diel activity pattern of ancestral mammals, a robust test for the invasion of ancestral mammals into the nocturnal niche using molecular data that specifically targets visual ability has not been conducted in a phylogenetic framework.
In vision, phototransduction is a process involving a signal transduction cascade that converts light into an electrical signal in photoreceptor cells. In the vertebrate retina two types of photoreceptor cells,(i.e., rods and cones) are specialized for different aspects of vision. Rods are extremely sensitive to light and are specialized for dim-light vision(e.g., scotopic vision), while cones have high spatial resolution and are dedicated to bright-light vision(e.g., photopic vision). The functional specializations of rods and cones are a result of different genes involved in their phototransduction cascades12. However, despite using different genes, the phototransduction cascades in rods and cones are similar and both start with the absorption of light in visual pigments that activates the opsins. The opsins then activate transducin(Gt), which subsequently activates the cGMP phosphodiesterase(PDE), reducing concentration of cGMP and eventually leading to closure of a cyclic nucleotide-gated(CNG) ion channel and hyperpolarizing the cell(Fig. 1). Along with the activation of these phototransduction cascades, a timely photoresponse recovery is responsible for a high temporal resolution, which is considered to be important for motion detection13.
The differentiation of the dim-light vision genes and the bright-light vision genes in photoreceptors provides an opportunity to test the nocturnality hypothesis of ancestral mammals. We would expect that a switch from diurnality to nocturnality in ancestral mammals would be accompanied by a strong positive selection for dim-light vision and that this selection would not have occurred in other diurnal tetrapod clades. In support of this, one recent study in raptors reveals that the phototransduction genes evolution is consistent with their diel activity variations, showing that the diurnal raptors(Accipitriformes) feature a positive selection for bright-light vision, while the nocturnal owls feature a positive selection for dim-light vision, suggesting a molecular phyloecological approach to investigate different diel activity patterns through analyzing the adaptive evolution of the bright-light vision genes and the dim-light vision genes14. In this study, with phototransduction gene sequences available from a broad range of tetrapods(Supplementary Table 1), we analyzed the adaptive evolution of 34 phototransduction-related genes that are known to be involved in rod and cone phototransduction pathway12(Fig. 1) within a phylogenetic context(Fig. 2). Our results indicate increasing selection for nighttime vision within mammals, consistent with the primarily nocturnal lifestyles of extant taxa. Consistent with the recent study in birds14, we find concordant patterns between the diel activity patterns and the selection of the phototransduction genes within both mammals and reptiles. Our study highlights the power of using phototransduction genes to predict and/or reconstruct the diel activity patterns of both extant and extinct taxa using a phylogenetic context.
Materials and Methods
The coding sequences of 34 vision genes involved in the phototransduction pathway were downloaded from the public databases GenBank and Ensembl(Supplementary Table 1). To maximize phylogenetic coverage, one or two represented species(depending on their sequences availabilities) from each major taxa of amniotes were used. The sequences of each gene were aligned using software webPRANK(http://www.ebi.ac.uk/goldman-srv/webprank/)15. The algorithm implemented in this software is considered to generate a more reliable alignment than other alignment methods and is capable of reducing false positive results of positive selection inference based on the branch-site model16. During the analyses, the sequence alignments were manually inspected for quality and the species with short sequences or low sequence identities(e.g., long indels and multiple ambiguous bases Ns) were removed. After such pruning, only the high-quality alignment results were used for subsequent positive selection analyses.
Positive selection analyses
We used codon-based maximum likelihood methods implemented in the Codeml program in PAML 4.8a package17 for our positive selection analyses. The maximum likelihood methods estimate the ratio of non-synonymous to synonymous substitutions per site(dN/dS or ω), and ω < 1, ω = 1 and ω > 1 indicate purifying selection, neutral evolution and positive selection, respectively. For our analyses, the species relationships(Fig. 2) were constructed based on published studies18,19,20,21,22,23, and the branch model, branch-site model and clade model C were used to detect positive selection along the branches or clades of interest. For the analyses, each branch or clade of interest was respectively used as the foreground branch or clade while all others were treated as the background. Likelihood ratio tests(LRT) were then employed to compare the goodness of fit of the more complex models with their corresponding null models.
We first used the two-rate branch model to test the strength and nature of selection along the foreground branches of interest. The two-rate model allows ω variations between foreground branches and background branches, and its goodness of fit was compared with the one-rate model, which assumes one single ω value for every branches base on the LRT. To further examine if the ω value of the foreground branch is greater than one, the likelihood values of the two-ratio model was then compared with the two-ratio model with a constraint of ω = 1.
We also used a more powerful branch-site model to detect positively selected sites along the branches of interest. With this analyses, we used the recommended Test 2, which compares a modified model A with its corresponding null model with ω = 1 fixed. The modified model A assumes four classes of sites. The site class 0(0 < ω0 < 1) and site class 1(ω1 = 1) respectively include evolutionarily conserved and neutral codons in both background branches and foreground branches, while the site classes 2a and 2b respectively include evolutionarily conserved(0 < ω0 < 1) or neutral(ω1 = 1) codons in the background branches, but are allowed to be under positive selection(ω2 > 1) along the foreground branches. When the LRT is significant, the Bayes Empirical Bayes method was used to identify positively selected sites.
We used the clade model C to identify the sites under divergent selection pressure between the foreground clade and the background clade. Clade model C assumes three classes of sites and the site class 0(0 < ω0 < 1) and site class 1(ω1 = 1) includes evolutionarily conserved and neutral codons, respectively, while the site classes 2 allows codons under divergent selection pressures between foreground clade and background clade. The clade model C was compared with a modified null model M2a_rel for LRT since M2a_rel is believed to better account for among-site variation in selective constraint and help to reduce false positive results24.
Robustness tests of positive selection genes
We examined the effects of the phylogenetic uncertainty and the initial value variations of kappa and omega on our positive selection results. To examine the effects of phylogenetic uncertainty on our results, we used another mammalian phylogeny that differs from the molecular phylogeny we used in several aspects. The new mammalian phylogeny is based on both molecular and morphological data25 and one of its main differences from our phylogeny is the separation of Xenarthra from Afrotheria, and Xenarthra is treated as a sister taxa of all other placentals. In addition, several other inconsistences within Afrotheria and Euarchonta are also noted, for instance, Scandentia is treated as more closely related to Dermoptera than to Glires. In addition to the phylogenetic uncertainty, we also evaluated the dependences of our positive selection results on the initial value variations of kappa and omega. For this, we respectively used two different initial values of kappa(kappa = 0.5, 3.0) and omega(omega = 0.5, 2.0) for our positive selection analyses.
Ancestral sequence reconstruction
We reconstructed ancestral amino acid sequences of LWS, RH1, SWS1 and SWS2 to analyze their possible amino acid replacements along branches of interests. For this, we used the amino acid-based marginal reconstruction implemented in the empirical Bayes approach in PAML 4.8a26. With the marginal reconstruction, the character was assigned to one single interior node and the character with the highest posterior probability is used as the best reconstruction. Two different amino acid substitution models, JTT and Poisson, were used to examine the consistency of our results. The model JTT assumes different substitution rates of different amino acids while the Poisson model assumes the same substitution rate of all amino acids. The analyses based on the JTT and Poisson models generated similar results, and for convenience only the results based on the JTT model are shown.
Protein structure reconstruction
To understand the potential effects of the positively selected site and the critical amino acid replacement S159A of LWS in the ancestral mammals, we reconstructed three-dimensional structure of LWS at the internal node representing the last common ancestor of modern mammals based on the ancestral amino acid reconstruction mentioned above. For 3D structure reconstruction, we used homology modelling techniques implemented in SWISS-MODEL(http://swissmodel.expasy.org/)27. The 3D structure was reconstructed based on the template 2j4y.1.A, which is the first best hit according to the global quality estimation(GMQE).
Analyses of gene losses in reptiles
We failed to find sequences of the two dim-light vision genes PDE6A and GNGT1 from all reptiles(including birds) in nr/nt database, suggesting their losses from all reptiles studied. To further test the possible losses of these two genes in reptiles, we respectively used coding sequences of these two genes from primitive mammals(platypus and Tasmanian devil) as a query to blast against the database of whole genome-shotgun contigs(WGS) of reptiles in GenBank, and no sequences for PDE6A or GNGT1 were found in reptiles. Specifically, for PDE6A, we used both its whole coding sequences and each of its six longest exon sequences of playtupus(Ornithorhynchus anatinus) as queries to blast against WGS of reptiles using web-based blastn. Similarly, for GNGT1, we used both its whole coding sequences and each of its two exon sequences of Tasmanian devil(Sarcophilus harrisii) to blast against WGS of reptiles. The best hits(e < = 1e - 5) of PDE6A and GNGT1 were subsequently used to blast against nr/nt database using the web-based blastn and almost all the best hits showed 100% coverage and 100% sequence identity with genes PDE6B/ PDE6C and GNGT2/GNG11, respectively, instead of our target PDE6A and GNGT1, suggesting the losses of the two dim-light vision genes from all the reptiles studied.
Results and Discussion
Evolution of nocturnality in mammals
We tested for adaptive evolution of vision genes using different models implemented in PAML17. Positively selected genes(PSGs) were only identified when using the branch-site model(Table 1, Supplementary Table 2), while no PSGs were detected with the branch model and the clade model C. The variable capabilities of these models in identifying PSGs may be partly attributed to their different underlying assumptions. For example, it has been shown that positive selection is less likely to affect all sites, and thus the ω values averaged over all sites by the branch model is almost never >128. Similarly, for the clade model C, the ω values are calculated as an average over all lineages in certain clades29. The ω values averaged over all sites or all lineages may have weakened power in detecting positive selection. On the other hand, the branch-site model is built on an underlying hypothesis that positive selection tends to affect only a few sites along particular lineages28, which may explain why the branch-site model has improved power in detecting PSGs compared with other models. For the PSGs identified by the branch-site model, their positive selection signals remained robust when phylogenetic uncertainty and variations of the initial values of kappa and omega(ω) in the positive selection analyses were considered. Specifically, to test the possible occurrence of nocturnality in ancestral mammals, we analyzed the adaptive evolution of vision genes along the ancestral branch(branch a in Fig. 2) of all three extant clades of mammals(Monotremata, Marsupialia and Placentalia) and we found four dim-light vision genes(GRK1, PDE6A, PDE6B and CNGA1) under positive selection(Fig. 1, Table 1). Of the four PSGs, PDE6A, PDE6B and CNGA1 are involved in the activation of phototransduction pathway, and GRK1 is involved in the photoresponse recovery. PDE6A and PDE6B are known to encode two hydrolytic subunits α and β of phosphodiesterase and CNGA1 encodes a protein forming CNG ion channel. GRK1 plays a role in deactivation of activated rhodopsin(RH1) and contributes to rhodopsin recovery. Additionally, two other vision genes(GUCY2D and RGS9) involved in photoresponse recovery were found to be under positive selection. GUCY2D encodes guanylyl cyclase involved in the resynthesis of cGMP, and RGS9 plays a role in deactivation of transducin. Moreover, RGS9BP, which plays a role in inactivation of transducin, was found to be under positive selection along the branch(combined branches a and b due to sequence unavailability of monotremes) leading to Theria. These PSGs involved in both the phototransduction activation and the photoresponse recovery likely contribute to the evolution of increased visual sensitivity as well as enhanced motion detection capability in ancestral mammals in dim-light conditions.
We also found the long-wavelength sensitive opsin LWS and the short-wavelength sensitive opsin SWS1 were under positive selection in ancestral mammals(Fig. 1, Table 1). LWS showed strong positive selection along the common ancestor branch(branch a) of extant mammals, and SWS1, which is lost from monotremes1, showed positive selection along the branch(combined branches a and b) leading to Theria. Regarding the strong positive selection of LWS, we further used the Bayes Empirical Bayes model30 to identify the positively selected sites and only one such site was detected with a high posterior probability(0.956), which was determined to be located in the transmembrane domain(Fig. 1, Supplementary Fig. 1). We subsequently reconstructed the ancestral sequence of LWS based on an empirical Bayes approach26 and 19 amino acid replacements along the ancestral mammal branch were found(Supplementary Fig. 2). Among the 19 amino acid replacements, one critical amino acid replacement(S159A, corresponding to S164A in bovine rhodopsin) that is known to decrease the wavelength of maximal absorption(λmax) of LWS by 7 nm was identified31(Supplementary Fig. 2, Supplementary Table 3), suggesting a short-wavelength shift of LWS. Given the short-wavelength shift of LWS, we further reconstructed ancestral sequences of three other opsins SWS1, SWS2 and RH1 along the ancestral mammalian branch; we did not find any critical amino acid replacements associated with their spectral turnings(Supplementary Table 3)(data not shown), suggesting their λmax values were unchanged.
The short-wavelength shift of LWS may have helped ancestral mammals to maximize photon absorption by tuning their LWS spectral sensitivity to the relatively abundant short-wavelength light in twilight32. If this is the case, our results suggest crepuscular activity of ancestral mammals along with increasing nocturnality, consistent with a recent study that argues a ‘twilight phase’ of mammalian evolution given the possible retention of three cone opsins(LWS, SWS1 and SWS2)33. Moreover, the positive selection results found in ancestral mammals are very similar with ancestral owls that also show the positive selection of the dim-light vision genes(SLC24A1, CNGB1) and the cone opsin genes(LWS and SWS2), and particularly, the LWS of ancestral owls is found to harbor the same one critical amino acid replacement(corresponding to S164A in bovine rhodopsin) as that of ancestral mammals leading to a short-wavelength shift by reducing its λmax of 7 nm14. The similar results between ancestral mammals and ancestral owls may be a result of their adaption to similar diel activity patterns.
We further analyzed the selection on vision genes along the main branches within mammals(branches b-i in Fig. 2, Supplementary Table 2) to examine the additional adaptations to light environments. For the Monotremata branch(branch e), only RGS9 was under positive selection, suggesting that there has been few vision modification along this branch since its evolution. Within Theria, sister to Monotremata, three PSGs(GUCY2D, RGS9 and RH1) with strong positive selection signals were found along the ancestral therian branch(branch b). GUCY2D and RGS9 are involved in photoresponse recovery. RH1 is the rod-specific rhodopsin and further ancestral sequence reconstruction identified one critical amino acid replacement(N83D) leading to an increase of the λmax of RH1 by 2 nm(Supplementary Table 3) along the ancestral therian branch(Supplementary Fig. 3). This long-wavelength shift of RH1 may assist in maximizing photon absorption in nocturnal conditions when relatively long-wavelength light dominates34. The strong positive selection of these three genes trends toward further enhancement of both the dim-light vision and the photoresponse recovery in the ancestral therians subsequent to stem mammals. Moreover, testing for positive selection within Theria showed that ancestral Marsupialia(branch f) harbored two PSGs including one dim-light vision gene(RH1) and one bright-light vision gene(CNGA3). Within ancestral Placentalia(branch c or combined branches b and c) one dim-light vision gene(PDE6B) and one photoresponse recovery gene(GUCA1B) were found to be under positive selection. Our results support a trend of increased adaptation to the dim-light conditions from ancestral Mammalian through Theria and to ancestral Placentalia. A further examination of positive selection along the main branches(branches g-i) within Placentalia only found two PSGs(RH1 and SWS1) along the ancestral branch(branch h) of Laurasiatheria, while no PSGs were detected along the other two branches(branches g and i) including Euarchontoglires, Afrotheria and Xenarthra, suggesting fewer subsequent modifications in vision in these groups. Thus these three subclades(Euarchontoglires, Afrotheria and Xenarthra) might retain pleisiomorphic visual adaptation similar to ancestral placentals, perhaps in part be due to their relatively early origins prior to the K/T boundary22,35, a time when diurnal niches were occupied by other species and dinosaurs still dominated terrestrial habitats. Considering our partial taxa coverage for each of the three subclades, further studies incorporating more samples would be useful for a robust inference of their diel activity patterns.
Nocturnality as a derived trait
To further test whether the possible nocturnality of ancestral mammals evolved from a diurnal ancestor, we examined the adaptive evolution of vision genes along the common ancestral branch of amniotes(branch s in Fig. 2). Among the 34 genes analyzed, 4 genes(CNGA3, GNB3, LWS and PDE6B) which are involved in the activation of the phototransduction pathway were found to be under positive selection(Fig. 1, Table 1). Among the four PSGs, CNGA3, GNB3 and LWS are bright-light vision genes and PDE6B is a dim-light vision gene. CNGA3, which encodes a subunit of the CNG ion channel, presented the strongest positive selection signal(ω2a = 447.444, ω2b = 447.444, df = 1, P = 0.002). GNB3 encodes the β subunit of transducin and showed the second strongest positive selection. LWS was under relatively strong positive selection and one positively selected site located in the transmembrane domain was identified(Supplementary Fig. 1). The only dim-light vision gene, PDE6B, showed the weakest positive selection signal compared with the other three bright-light vision genes. The relatively strong positive selection of the bright-light vision genes in ancestral amniotes probably helped to promote their visual acuity in daytime and suggests their predominately diurnal lifestyle. The adaptation of ancestral amniotes to bright-light environments may have occurred during the habitat shift from aquatic or semi-aquatic habitats of amphibian ancestors to their terrestrial habitats, where light is more intense and less attenuated by water.
Given the diurnality of early amniotes inferred here, the nocturnality of ancestral mammals is likely a derived character. Moreover, our finding of a few positively selected bright-light vision genes along the branch leading to extant mammals suggests that origin of dim-light vision and true nocturnality of the common ancestors of extant mammals likely occurred only once, while the timing of the occurrence of the nocturnality remains unknown. It is well known that mammals, as one of the members of synapsid lineage, appear relatively late and before their appearance, synapsids had experienced a long evolutionary history, which is sufficient to allow their adaptation to different niches(e.g., temporal niches)5,7. A recent phenotypic study implies that the invasion of synapsids into nocturnality may have occurred early extending back to the Late Carboniferous5. If this is the case, one possibility is that the nocturnality of ancestral mammals might be retained from that of their synapsid progenitors. However, uncertainties of the timing exist since the fossil synapsid taxa used in the phenotypic study may not represent the common ancestor of extant mammals, and hence the timing of the nocturnal invasion of the common ancestor of extant mammals still remains unknown.
Temporal niches differentiation between ancestral mammals and reptiles
Invasion of ancestral mammals into nocturnality is hypothesized to be a result of competitive exclusion and/or predation by reptiles, which are believed to be predominately diurnal due to their reliance on solar radiation for warmth36. Among reptiles, the insectivorous lizards, like early mammals, occupy the insectivoran niche and may be one of the main competitors of ancestral mammals10,36. Here we tested the diurnality of reptiles by analyzing positive selection along different branches within reptiles(Fig. 2). The initial examination along the ancestral branch(branch j) of the extant reptiles studied showed very strong positive selection signals in five genes(Fig. 1, Table 1). Among the five PSGs, three genes GNAT2, LWS and CNGB3 are bright-light vision genes and one gene SLC24A1 is a dim-light vision gene. GNAT2, which encodes transducin α subunit, and LWS both showed the strongest positive selection signals. CNGB3 is involved in the formation of CNG ion channel. The only positively selected dim-light gene SLC24A1 showed relatively strong positive selection. SLC24A1 encodes the Na+/Ca2+-K+ exchanger(NCKX), which removes free calcium in the rod outer segment and contributes to the restoration of the cGMP concentration and hence photoresponse recovery. In addition, GUCY2F, which encodes guanylyl cyclase involved in the resynthesis of cGMP and contributes to the photoresponse recovery, was also found to be under positive selection. Our results therefore suggest that positive selection for bright-light vision dominated the ancestral reptiles, and hence support the hypothesis that ancestral reptiles were predominately diurnal. In support of this, we failed to find two dim-light vision genes PDE6A and GNGT1 from all reptiles(including birds) in both nr/nt database and the whole-genome shotgun contigs database, suggesting the loss of these genes in ancestral reptiles. Previous studies also find the absence of the PDE6A gene in birds and non-avian reptiles37,38. Additionally, no transcripts of PDE6A and GNGT1 were detected in a recent study about the retinal transcriptome sequencing of 15 bird species including both diurnal and nocturnal taxa as well14. The possible loss of the two dim-light vision genes from ancestral reptiles may be due to relaxed selection on dim-light vision associated with their predominate diurnality. In addition to the ancestral reptiles, we detected almost only bright-vision genes to be under positive selection along the branches(branches k, o and p) leading to Squamata(Table 1, Supplementary Table 2), suggesting their predominate diurnality as well. We should note here that not all reptilian taxa we examined showed predominately diurnal activities, with the various branches related to turtles, crocodiles and birds harboring considerable PSGs related to the dim-light vision(Supplementary Table 2), demonstrating the abilities of their vision in dim-light.
Here we detected the diel activity patterns of relevant taxa based on the molecular phyloecological approach. These patterns are largely consistent with traditional hypothesis that ancestral mammals are predominately nocturnal and reptiles are mainly diurnal1. This molecular phyloecological approach can be useful to reconstruct the ancestral character states of extinct taxa. However, we would like to note some limitations of this molecular phyloecological approach: i) The approach is only restricted to those taxa from which their gene sequences are available, and thus can not be applied to those extinct taxa lacking sequence availabilities or representatives of their derived living taxa. ii) Little is known about the effects of the absence of extinct taxa on the phylogenetic analyses of positive selection. Regarding our focal amniotes, there are numerous extinct species, while our study is only capable of involving living amniotes. Given that previous morphological studies show an improved ancestral trait reconstruction by incorporating fossil information relative to those based on extant taxa alone39,40,41, future integrated analyses incorporating both molecular and morphological results would lead to more robust inferences of ancestral character states. iii) Even for the taxa with numerous derived living taxa, the effects of using relatively less represented entities on the positive selection results are less clear. Though our study covers a broad range of tetrapods, the sample size is relatively small for each of certain subclades, and our study is largely restricted to those taxa from which their genomes are available. Nonetheless, one previous study shows that the sample sizes seem to have little effects on positive selection results42, and thus our positive selection results may still be informative. In general, an increased taxa sampling is considered to improve the power to detect positive selection43, so future study incorporating more taxa with intensive sampling in a supertree context may be more powerful in detecting positive selection. iv) For the large-scale phylogenetic analyses spanning a long evolutionary history, like our tetrapod phylogeny, the gene sequences among taxa would be highly divergent, which may lead to substitution saturation and possible false positive results. However, the branch-site model test has been demonstrated to be robust against highly divergent sequences and the substitution saturation is found to have nearly no effects on the false positive results44, suggesting the robustness of our positive selection results.
Ancestral mammals are inferred to be nocturnal based on multiple lines of morphological evidence in fossils7,8,45. In this study, we conducted comparative evolutionary analyses of genes involved in vision pathways and revealed positively selected genes involved in dim-light vision within ancestral Mammalia, supporting the nocturnality of early mammals. The positively selected vision genes of ancestral mammals incorporated almost all of the principle components of the phototransduction pathway involved in both phototransduction activation and photoresponse recovery(Fig. 1). These results suggest substantial modification of the visual system within subsequent branches of Mammalia toward crown Placentalia trending toward an increased visual sensitivity and an enhanced ability to detect motion. These abilities may be important for mammals to hunt flying insects in dim-light conditions7. Our results further show that the nocturnality of the common ancestors of extant mammals was clearly derived from the predominately diurnal ancestral amniotes, perhaps driven in part by competition exclusion of predominately diurnal reptiles. Despite the large differentiation of the temporal niches between ancestral mammals and reptiles, our study provides evidence of both the crepuscularity of ancestral mammals and the dim-light activities of some reptile taxa, and thus suggests their possible partial overlap during crepuscular periods. Consistent with a recent bird study14, our study shows that the phototransduction genes evolution corresponds very well with the diel activity variations in mammals and reptiles, thus further tests the effectiveness of the molecular phyloecological approach for inferring the diel activity patterns, which remains to be addressed in other systems. Finally, our study includes partial representative taxa of reptiles and mammals, future studies incorporating more taxa from both groups may provide more profound information.
How to cite this article: Wu, Y. et al. Invasion of Ancestral Mammals into Dim-light Environments Inferred from Adaptive Evolution of the Phototransduction Genes. Sci. Rep. 7, 46542; doi: 10.1038/srep46542(2017).
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Gerkema, M. P., Davies, W. I., Foster, R. G., Menaker, M. & Hut, R. A. The nocturnal bottleneck and the evolution of activity patterns in mammals. Proc Roy Soc Lond B Biol Sci 280, 20130508(2013).
Jerison, H. More on why birds and mammals have big brains. Am Nat 105, 185–189(1971).
Polyak, S. The Vertebrate Visual System. 954–956(University of Chicago Press, 1957).
Walls, G. L. The Vertebrate Eye and its Adaptive Radiation. 671–687(The Cranbrook Institute of Science, 1942).
Angielczyk, K. & Schmitz, L. Nocturnality in synapsids predates the origin of mammals by over 100 million years. Proc Roy Soc Lond B Biol Sci 281, 2014 1642(2014).
Jerison, H. J. Evolution of the Brain and Intelligence. 256–281(Academic Press, 1973).
Kemp, T. S. The Origin and Evolution of Mammals. 183–184(Oxford University Press, 2005).
Heesy, C. P. & Hall, M. I. The nocturnal bottleneck and the evolution of mammalian vision. Brain Behav Evol 75, 195–203(2010).
Crompton, A., Taylor, C. R. & Jagger, J. A. Evolution of homeothermy in mammals. Nature 272, 333–336(1978).
Kermack, D. & Kermack, K. The Evolution of Mammalian Characters. 89–99(Croom Helm, 1984).
Livesey, P. J. Learning and Emotion: a Biological Synthesis. 61–69(Lawrence Erlbaum Associates, 1986).
Lamb, T. D. Evolution of phototransduction, vertebrate photoreceptors and retina. Progr Retin Eye Res 36, 52–119(2013).
Umino, Y. et al. The relationship between slow photoresponse recovery rate and temporal resolution of vision. J Neurosci 32, 14364–14373(2012).
Wu, Y. et al. Retinal transcriptome sequencing sheds light on the adaptation to nocturnal and diurnal lifestyles in raptors. Sci Rep 6, 33578, doi: 10.1038/srep33578(2016).
Löytynoja, A. & Goldman, N. webPRANK: a phylogeny-aware multiple sequence aligner with interactive alignment browser. BMC Bioinformatics 11, 579(2010).
Fletcher, W. & Yang, Z. The effect of insertions, deletions, and alignment errors on the branch-site test of positive selection. Mol Biol Evol 27, 2257–2267(2010).
Yang, Z. PAML 4: phylogenetic analysis by maximum likelihood. Mol Biol Evol 24, 1586–1591(2007).
Cardillo, M., Bininda-Emonds, O. R., Boakes, E. & Purvis, A. A species-level phylogenetic supertree of marsupials. J Zool 264, 11–31(2004).
Crawford, N. G. et al. More than 1000 ultraconserved elements provide evidence that turtles are the sister group of archosaurs. Biol Lett 8, 783–786(2012).
Crawford, N. G. et al. A phylogenomic analysis of turtles. Mol Phylogenet Evol 83, 250–257(2015).
Guillon, J.-M., Guéry, L., Hulin, V. & Girondot, M. A large phylogeny of turtles(Testudines) using molecular data. Contrib Zool 81, 147–158(2012).
Meredith, R. W. et al. Impacts of the Cretaceous terrestrial revolution and KPg extinction on mammal diversification. Science 334, 521–524(2011).
Pyron, R. A., Burbrink, F. T. & Wiens, J. J. A phylogeny and revised classification of Squamata, including 4161 species of lizards and snakes. BMC Evol Biol 13, 93(2013).
Weadick, C. J. & Chang, B. S. An improved likelihood ratio test for detecting site-specific functional divergence among clades of protein-coding genes. Mol Biol Evol 29, 1297–1300(2012).
O’Leary, M. A. et al. The placental mammal ancestor and the post–K-Pg radiation of placentals. Science 339, 662–667(2013).
Yang, Z., Kumar, S. & Nei, M. A new method of inference of ancestral nucleotide and amino acid sequences. Genetics 141, 1641–1650(1995).
Biasini, M. et al. SWISS-MODEL: modelling protein tertiary and quaternary structure using evolutionary information. Nucleic Acids Res 42, W252–W258(2014).
Yang, Z. & Nielsen, R. Codon-substitution models for detecting molecular adaptation at individual sites along specific lineages. Mol Biol Evol 19, 908–917(2002).
Bielawski, J. P. & Yang, Z. A maximum likelihood method for detecting functional divergence at individual codon sites, with application to gene family evolution. J Mol Evol 59, 121–132(2004).
Yang, Z., Wong, W. S. & Nielsen, R. Bayes empirical Bayes inference of amino acid sites under positive selection. Mol Biol Evol 22, 1107–1118(2005).
Yokoyama, S. Molecular evolution of color vision in vertebrates. Gene 300, 69–78(2002).
Endler, J. A. The color of light in forests and its implications. Ecol Monogr 63, 1–27(1993).
Davies, W. I., Collin, S. P. & Hunt, D. M. Molecular ecology and adaptation of visual photopigments in craniates. Mol Ecol 21, 3121–3158(2012).
Veilleux, C. C. & Cummings, M. E. Nocturnal light environments and species ecology: implications for nocturnal color vision in forests. J Exp Biol 215, 4085–4096(2012).
Bininda-Emonds, O. R. et al. The delayed rise of present-day mammals. Nature 446, 507–512(2007).
Crompton, A. in Comparative Physiology: Primitive MammalsVol. 4(eds Knut Schmidt-Nielsen, Liana Bolis & Charles Richard Taylor ) 1–12(Cambridge University Press, 1980).
Huang, D., Hinds, T. R., Martinez, S. E., Doneanu, C. & Beavo, J. A. Molecular determinants of cGMP binding to chicken cone photoreceptor phosphodiesterase. J Biol Chem 279, 48143–48151(2004).
Lagman, D., Franzén, I. E., Eggert, J., Larhammar, D. & Abalo, X. M. Evolution and expression of the phosphodiesterase 6 genes unveils vertebrate novelty to control photosensitivity. BMC Evol Biol 16, 124(2016).
Slater, G. J. Phylogenetic evidence for a shift in the mode of mammalian body size evolution at the Cretaceous‐Palaeogene boundary. Methods Ecol Evol 4, 734–744(2013).
Slater, G. J. & Harmon, L. J. Unifying fossils and phylogenies for comparative analyses of diversification and trait evolution. Methods Ecol Evol 4, 699–702(2013).
Slater, G. J., Harmon, L. J. & Alfaro, M. E. Integrating fossils with molecular phylogenies improves inference of trait evolution. Evolution 66, 3931–3944(2012).
Yang, Z. Likelihood ratio tests for detecting positive selection and application to primate lysozyme evolution. Mol Biol Evol 15, 568–573(1998).
Yang, Z. Maximum likelihood estimation on large phylogenies and analysis of adaptive evolution in human influenza virus A. J Mol Evol 51, 423–432(2000).
Gharib, W. H. & Robinson-Rechavi, M. The branch-site test of positive selection is surprisingly robust but lacks power under synonymous substitution saturation and variation in GC. Mol Biol Evol 30, 1675–1686(2013).
Hall, M. I., Kamilar, J. M. & Kirk, E. C. Eye shape and the nocturnal bottleneck of mammals. Proc Roy Soc Lond B Biol Sci 279, 4962–4968(2012).
We gratefully acknowledge Katie Solari and Hannah Frank in the Hadly lab at Stanford University for their comments on our manuscript draft. This research was supported by the National Natural Science Foundation of China(Grant 31200276), the grant of China Scholarship Council, the Fundamental Research Funds for the Central Universities and assisted by a Visiting Scholarship for Y.W. to Stanford University.
The authors declare no competing financial interests.
About this article
Cite this article
Wu, Y., Wang, H. & Hadly, E. Invasion of Ancestral Mammals into Dim-light Environments Inferred from Adaptive Evolution of the Phototransduction Genes. Sci Rep 7, 46542 (2017). https://doi.org/10.1038/srep46542
The Journal of Experimental Biology (2020)
Physiological and Biochemical Zoology (2019)
Proceedings of the Royal Society B: Biological Sciences (2019)
BMC Evolutionary Biology (2019)
Molecular Biology and Evolution (2018)