Genomic evidence of bitter taste in snakes and phylogenetic analysis of bitter taste receptor genes in reptiles

Abstract

As nontraditional model organisms with extreme physiological and morphological phenotypes, snakes are believed to possess an inferior taste system. However, the bitter taste sensation is essential to distinguish the nutritious and poisonous food resources and the genomic evidence of bitter taste in snakes is largely scarce. To explore the genetic basis of the bitter taste of snakes and characterize the evolution of bitter taste receptor genes (Tas2rs) in reptiles, we identified Tas2r genes in 19 genomes (species) corresponding to three orders of non-avian reptiles. Our results indicated contractions of Tas2r gene repertoires in snakes, however dramatic gene expansions have occurred in lizards. Phylogenetic analysis of the Tas2rs with NJ and BI methods revealed that Tas2r genes of snake species formed two clades, whereas in lizards the Tas2r genes clustered into two monophyletic clades and four large clades. Evolutionary changes (birth and death) of intact Tas2r genes in reptiles were determined by reconciliation analysis. Additionally, the taste signaling pathway calcium homeostasis modulator 1 (Calhm1) gene of snakes was putatively functional, suggesting that snakes still possess bitter taste sensation. Furthermore, Phylogenetically Independent Contrasts (PIC) analyses reviewed a significant correlation between the number of Tas2r genes and the amount of potential toxins in reptilian diets, suggesting that insectivores such as some lizards may require more Tas2rs genes than omnivorous and carnivorous reptiles.

Introduction

Vertebrates possess five basic tastes: sweet, umami, bitter, salty, and sour (Bachmanov & Beauchamp, 2007; Lindemann, 1996). It is generally believed that bitter transduction evolved as a key defense mechanism against the ingestion of toxic substances. As bitter foods generally contain harmful or toxic compounds, the bitter taste sensation is essential for animals to evaluate the potential food resources and distinguish between nutritious and toxic food (Garcia & Hankins, 1975; Glendinning, 1994). Many bitter substances come from not only plants, but also from insects and meat tissues (Cott, 2010; Drewnowski & Gomezcarneros, 2000; Garcia & Hankins, 1975;Glendinning, 1994; Howse, 1975; Unkiewicz-Winiarczyk & Gromysz-Kałkowska, 2012). The bitter perception is mediated by a group of G-protein-coupled receptors taste receptor member 2 (Tas2r) which are encoded by the Tas2r gene family (Adler et al., 2000; Chandrashekar et al., 2000; Matsunami, Montmayeur & Buck, 2000). Different animal species have distinct diets and therefore different bitter compounds are detected by different Tas2rs (Meyerhof et al., 2010). Thus the Tas2r repertoires are expected to be species-specific. Indeed, the number of intact Tas2r genes varies among species, from 0 to 37 in mammals (Feng et al., 2014; Shi & Zhang, 2006), from 1 to 12 in birds (Wang & Zhao, 2015) and 0 to 74 in fish (Bachmanov et al., 2013; Syed & Korsching, 2014), due to the duplication or loss of Tas2r genes in what is called a “birth-and-death” process, characteristic of chemosensory genes (Nei, Gu & Sitnikova, 1997). Different Tas2r repertoires among species may reflect varying dietary pressures. Generally speaking, among animals potential foods, plant tissues may contain more toxic compounds than animal tissues (Glendinning, 1994; Wang, 2004). Insects and some amphibians like newts also secrete defensive poisonous chemicals to deter predators (Brodie & Tumbarello, 1978; Howse, 1975). Hence, more functional Tas2r genes exist in herbivorous, omnivorous and insectivorous animals than in carnivorous animals, implying that distinct diets may shape the variation of Tas2r repertoire among species (Li & Zhang, 2014; Wang & Zhao, 2015). Factors other than diets like foraging pattern may also affect Tas2r numbers. For example, cetaceans swallow their food whole, and 9 out of 10 amplified Tas2r genes from 12 whales were pseudogenized (Feng et al., 2014).

Reptiles (10,450 species) live in a wide range of terrestrial, marine and freshwater environments. Among them, snakes (3,619 species) and lizards (6,263 species) comprise of 87 families, according to The Reptile Database (http://reptile-database.org/, last accessed July 18, 2017). Studies of taste genetics in vertebrate species have been mainly focused on mammals and birds (Behrens, Korsching & Meyerhof, 2014; Behrens & Meyerhof, 2013; Davis et al., 2010; Dong, Jones & Zhang, 2009; Go, 2006; Hayakawa et al., 2012; Hong & Zhao, 2014; Jiang et al., 2012; Li & Zhang, 2014; Liu et al., 2016; Nei, Gu & Sitnikova, 1997; Feng et al., 2014; Shang et al., 2017; Shi & Zhang, 2006; Shi et al., 2003; Wang & Zhao, 2015; Wang, 2004; Wu, Chen & Rozengurt, 2005; Zhou, Dong & Zhao, 2009), however, similar studies of Tas2r genes of snakes have not been performed. The strictly carnivorous diet of snakes comprises few toxins, although it may contain defensive toxic secretions released by insects or highly toxic newts (Brodie, Ridenhour & Iii, 2002; Mehrtens, 1987; Parker & Gans, 1978). Besides, snakes cannot masticate and they have to swallow the preys whole without chewing. Therefore, we hypothesized that Tas2rs would still exist but would be smaller in number in snakes according to their unique dietary and foraging pattern, compared with the number of Tas2r genes in lizards, the other suborder in Squamates, many of which are primarily insectivorous and need to to avoid intake of poisonous secretions released by insect prey. Apart from taste receptors, the taste pathway gene calcium homeostasis modulator 1 (Calhm1) that is responsible for taste signalling is also indispensable for taste function (Taruno et al., 2013). Thus, Calhm1 might be intact and functional if bitter taste is retained in snakes. To test our hypothesis, we identified Tas2r genes from eight snake species for the first time. Additionally, we identified the Calhm1 gene to explore the function of the taste signaling pathway in Squamates. We found that snakes have a considerable contraction of the Tas2rs repertoire, which was in accordance with their foraging and dietary patterns. In contrast, the Tas2r family in the two lizard species (the green anole and the Japanese gecko) we examined expanded dramatically, which was consistent with their insectivorous diet. Furthermore, Calhm1 genes were under strong purifying selection, implying the retention of bitter taste functionality in Squamates.

Materials & Methods

Genome data

The Squamates comprise two suborders: Serpentes (snake) and Lacertilia (lizard). A total of eight Serpentes genome sequences were downloaded from GenBank (http://www.ncbi.nlm.nih.gov/). They are corn snake (Pantherophis guttatus) and common garter snake (Thamnophis sirtalis) in Family Colubridae; speckled rattlesnake (Crotaus mitchellii), timber rattlesnake (Crotalus horridus), adder (Vipera berus) and brown spotted pit viper (Protobothrops mucrosquamatus) in Family Viperidae; king cobra (Ophiophagus hannah) in family Elapidae; the Burmese python (Python bivittatus) in family Pythonidae. In suborder Lacertilia, genome sequences of the green anole (Anolis carolinensis) and the Japanese gecko (Gekko japonicas) were retrieved. We also retrieved nine additional genome sequences of other reptiles to conduct further phylogenetic analysis. They were Spiny Softshell Turtle (Apalone spinifera), Green Sea Turtle (Chelonia mydas), Painted Turtle (Chrysemys picta), Diamondback Terrapin (Malaclemys terrapin) and Chinese Softshell Turtle (Pelodiscus sinensis) in Testudines; Saltwater Crocodile (Crocodylus porosus), American Alligator (Alligator mississippiensis), Chinese Alligator (Alligator sinensis) and Gharial (Gavialis gangeticus) in Crocodylia. Hence, a total of 19 genomes of reptiles were retrieved from GenBank (Table S1). The sequencing depths of the assemblies were all above 15 × except in green anole (7.1 ×) and corn snake (13 ×). Scaffold N50 sizes ranged from 2.4 to 437.3 Mb, suggesting high-coverage genomes.

Gene identification

Tas2r genes are about 900 bp without introns. First, we collected previously determined intact Tas2r protein sequences from human, mouse, lizard, chicken, frog and coelacanth genomes as queries (Li & Zhang, 2014; Syed & Korsching, 2014). We used TBLASN to search against the genomes with the E-value of 1e–10. Second, non-overlapping sequences were extracted and the blast hits shorter than 100 bp were discarded. We extracted the remaining blast hits and extended them in both 5′ and 3′ directions. They were regarded as putative Tas2r regions. Third, the deduced sequences were divided into three categories. Sequences more than 270 amino acids with a presumed start and stop codon were considered intact genes. The TMHMM method was used to make sure the seven transmembrane domains existed in intact genes (Krogh et al., 2001). Sequences longer than 100 bp with premature stop codons or frame-shifts were considered pseudogenes. Sequences longer than 100 bp containing a start codon (or a stop codon) were regarded as partial genes. Partial Tas2rs may be intact genes, but they were not sequenced completely or assembled completely (Hayakawa et al., 2014). To ensure the reliability of the partial genes, we assessed whether the partial genes are from independent loci or not (Table S2). We also performed syntenic analysis to determine whether partial genes are unique (Table S3). Lastly, we used BLASTP to search against GenBank to verify all the candidate genes belonging to the Tas2r family. The nomenclature of the Tas2r genes followed ‘Tas2r X’, in which ‘X’ represents arabic numbers one by one. See Data S1 and S2 for the nucleotide and amino acid sequences.

Phylogenetic analyses

To clarify the evolutionary history and relationships of the Tas2r genes in reptiles, phylogenetic analysis was conducted. Non-avian reptiles contain four general groups: Squamata (snakes and lizards), Crocodylia (alligators and crocodiles), Testudines (turtles) and Sphenodontia (tuatara). In the present study, tuataras were not included because no genome data was available in this group. The protein sequences of intact Tas2r genes were aligned by MUSCLE (Thompson, Higgins & Gibson, 1994) implemented in MEGA5 (Tamura et al., 2011) with manual adjustments. The western clawed frog (Xenopus tropicalis) V1R gene was used as outgroup, because among GPCRs, V1R genes are relatively close to Tas2r genes. Partial Tas2r genes were not included because most of them were too short to make a good alignment. Pseudogene sequences were also removed from the phylogenetic analysis. The best-fitting substitution model Jones-Taylor-Thornton (JTT) +G +F was determined by ModelGenerator 0.85 (Keane et al., 2006). After gaps were removed from the alignment with the pairwise deletion option, the neighbor-joining method (Saitou & Nei, 1987) implemented in MEGA5 was used to construct a phylogeny. Statistical confidence was evaluated with 1,000 replicates by the bootstrap method (Felsenstein, 1985). The BI (Bayesian Inference) (Yang & Rannala, 1997) tree was rebuilt by MrBayes 3.2.6 (Ronquist & Huelsenbeck, 2003) under the same model of substitution with 10 million generations.

Evolutionary analysis of Tas2rs repertoire

To recover evolutionary changes of Tas2r gene numbers in reptiles, we conducted a reconciliation analysis by NOTUNG 2.6 (Chen et al., 2004). This method is based on algorithms for tree rearrangement and for reconciliation with non-binary trees. It supports predicting gene gain and loss events with a parsimony-based optimization criterion. We used our BI tree as the gene tree (Fig. S1). The phylogeny and divergence times of these species were drawn from TimeTree (http://www.timetree.org/).

Phylogenetically Independent Contrasts (PIC) analyses

To explore the potential impact of diets on the evolution of the Tas2rs repertoires of reptiles, we carried out a regression analysis of Tas2r gene numbers against the diet type, which we summarized as a diet code. As insect and plant tissues contain more potential bitter compounds (toxins) than non-insect animal tissues, we coded the species as 0 (carnivore) and 1 (insectivore and herbivore) according to the amount of plant materials, insect tissues or non-insect animal tissues in their diets (Table S1). The omnivorous species painted turtle was coded as 1 because its diet appeared to contain more plant materials than animal tissues. We categorized the Tas2r genes into two sets: The first comprised putatively functional genes (intact and partial genes), and the second consisted of all genes (intact, partial and pseudogenes) because not only functional genes, but also pseudogenes may reflect the physiological demands (Rouquier, Blancher & Giorgi, 2000). Then we conducted a phylogenetically independent contrast (PIC) analysis with the R package Analyses of Phylogenetics and Evolution (APE) (Paradis, Claude & Strimmer, 2004). The correlation was assessed with nonparametric Spearman’s rank correlation coefficient (ρ).

Results

Identification of Tas2r genes

By TBLASN, we identified Tas2r genes from 19 genome sequences of reptiles. The result revealed that speckled rattlesnake and brown spotted pit viper followed the same pattern with one intact gene and one pseudogene. In the timber rattlesnake, adder and the Burmese python, the numbers were one intact gene and two pseudogenes. The corn snake and king cobra both possess two intact genes (Fig. 1). All the eight species showed absence of partial genes. Strikingly, no Tas2r gene was detected in common garter snake. The average number of intact genes of the seven snakes (except garter snake) was 1.3 and the average total number of Tas2rs was 2.4. Overall, the repertoires of Tas2rs in snake species were generally small than that in other vertebrates (Li & Zhang, 2014; Liu et al., 2016; Wang & Zhao, 2015). The nucleotide length of intact Tas2r genes of the 7 snake species was from 885 to 966 bps with the average of 904 bp and the amino acid sequence identities were ranging from 32.73% to 94.22%. By contrast, the two lizard species had a relatively larger Tas2rs repertoire. We identified 36 intact genes, 0 partial gene and 14 pseudogenes in the green anole (Anolis carolinensis). In the Japanese gecko (Gekko japonicas), the numbers were 50, 2 and 18 respectively. The average number of intact genes of the two lizards was 33 times of snakes’, with the exception of the common garter snake. The length of intact genes ranged from 870 to 1,077 with the average of 937 bp and the amino acid sequence identities were ranging from 17.24% to 95.42%. Another obvious difference between lizards and snakes was the number of pseudogenes: we discovered 14 in the green anole, and 18 in the Japanese gecko respectively (Fig. 1). In the five Testudines genomic assembles, we detected 2–11 intact Tas2r genes and 1–14 pseudogenes; the total number of Tas2r genes was 7–18 (Fig. 1). In order Crocodylia, the numbers of intact Tas2r genes, pseudogenes and total genes were 5–9, 0–5 and 5–12 respectively.

Figure 1. Numbers of intact, partial and pseudo Tas2r genes and dietary preferences (C: Carnivore; I: Insectivore; O: Omnivore; H: Herbivore) of the 19 reptile species.

Species tree and divergence times were taken from the literatures and TimeTree (http://www.timetree.org/, see Table S4). “*”indicates no Tas2r gene was as detected in the common garter snake.

Phylogenetic analysis

To investigate the phylogenetic relationship of Tas2r genes within reptiles, we conducted a phylogenetic analysis and constructed phylogenies with both the NJ and BI methods. As we can see from the BI tree (Fig. 2), the Tas2r genes of seven snake species (common garter snake was excluded) formed two monophyletic clades. In comparison, the Tas2r genes of lizards mainly clustered into two single-gene clades and four large clades, two of which were consisted of large numbers of Tas2r genes. After testing gene conversion among paralogous genes with GENECONV (Sawyer, 1989), we detected no possible events of gene conversion, suggesting that gene conversion may not have played a role in the evolution of lizards’Tas2r repertoire. We infer that the large and separate Tas2r repertoire in lizards was due to lineage-specific gene duplications, as demonstrated by our evolutionary analysis(see results below). Turtles (Testudines) had four separate ancestral nodes, two clades of which mingled with lizards and the other two with alligators and crocodiles (Crocodylia). The alligators and crocodiles had six segregated clades (Fig. 2).

Figure 2. Phylogenetic relationship of all 154 intact Tas2r genes from 19 reptiles.

The western clawed frog (Xenopus tropicalis) V1R1 gene was used as an outgroup. The tree was reconstructed using the BI method with the best-fitting model of JTT+G+F. Branch lengths were drawn to the scale. Posterior probabilities from Bayesian inference analyses are presented in percent, in which the value of 100 was not shown. The detailed information about species, gene names and posterior probabilities was shown in Fig. S1. The NJ tree showing a similar topology to this tree was provided in Fig. S2.

Evolutionary analysis of Tas2rs repertoire

To determine the evolution of Tas2r gene repertoires among non-avian reptile species, we predicted the intact Tas2rs numbers in the common ancestors of different clades. Then we performed a reconciliation analysis and inferred the birth and death of intact Tas2r genes. The BI tree was used (Fig. S1) to predict the changes of gene numbers, because the Bayesian posterior probability of each branch is above 50% and may better support the analysis than the NJ tree (Fig. S2). We found that the common ancestor of Squamata showed a gain (n = 4) when they were diverging with the common ancestor of Testudines and Crocodylia. In Squamata, the number of intact Tas2r genes in the common ancestor of snakes was only 2 (Fig. 3), indicating a reduction when snakes diverged with lizards. In contrast, green anole (36 intact genes) and the Japanese gecko (50 intact genes) both showed large numbers of species-specific duplication (28 and 40 respectively) (Figs. 1 and 2). Besides, a further gain (n = 1) happened in the common ancestor of Testudines and Crocodylia. In Testudines, gain and loss events did not happen in their common ancestor, but in each branch. There were two losses, one gain and two gains in saltwater crocodile, gharial and the American alligator, respectively. The evolutionary changes of the Tas2r gene number in the three species were consistent with the results from a recent study (Wang & Zhao, 2015). In Crocodylia, substantial gains (n = 5) were found in the Chinese softshell turtle.

Figure 3. Evolutionary changes of intact Tas2r gene numbers in 19 reptiles.

The estimated Tas2r gene numbers for ancestral lineages were indicated with black, whereas the numbers of gene gains and losses were shown with red and green, respectively. The common garter snake was not included because no Tas2rs gene was detected in the genome assembly.

Correlation between dietary preferences and Tas2rs repertoire

To test the potential impact of the dietary preferences on the evolution of Tas2r gene repertoires, we divided the 19 reptiles into carnivores, insectivores, herbivores and omnivores (Table S1). Then we coded the diet as 0 (carnivore) and 1 (insectivore, herbivore and omnivore). After performing a regression analysis of Tas2r gene numbers against the diet codes, we detected a significant positive correlation between the number of functional Tas2r genes and dietary preference (r = 0.87  P < 0.01; Fig. 4A). The same correlation was also observed between the number of total Tas2r genes and dietary preference (r = 0.8712, P < 0.01; Fig. 4B). Our findings unambiguously indicated a significant positive correlation between the repertoire of Tas2r genes in reptiles and the abundance of potential toxins in their diets.

Figure 4. Dietary preferences impact the reptilian Tas2r gene repertoires.

(A) PICs in putatively functional Tas2r gene number is positively correlated with PICs in diet content code; (B) PICs in total Tas2r gene number is also positively correlated with PICs in diet content code. The association was evaluated by the Spearman’s rank correlation coefficient (r).

Discussion

Snakes have evolved to be a nontraditional model organism with extreme physiological and morphological phenotypes, such as enhanced chemoreception (Castoe et al., 2013). Though studies on behavior and anatomy of Squamates have proposed them as alternative models for taste studies (Young, 1997), the genetic basis of taste sensation of snakes has not yet been assessed. This is partly owing to the belief that snakes rely more on vision, olfaction and even infrared sensory than on taste for feeding (Gamow & Harris, 1973; Grace et al., 1999; Halpern & Kubie, 1983; Walls, 1944). Anatomical studies revealed that snakes are lack of taste buds on the mucosa of jaw and tongue (Auen & Langebartel, 1977; Uchida, 1981; reviewed in Young, 1997). Consequently, it has been suggested that the sense of taste may be reduced (Auen & Langebartel, 1977; Uchida, 1981). In the present study, we identified Tas2r genes from the genomes of 8 snake species representing five families in Serpentes for the first time. Our results revealed that lineage-specific gene expansions and contractions leading to different numbers of Tas2rs in different lineages also happened in Squamates. The extensive gene contraction in this family found in birds and cetaceans (Dong, Jones & Zhang, 2009; Hillier et al., 2004) was here also identified in snakes. Reconciliation analysis indicated that many Tas2r genes have been lost in the common ancestor in the snake lineage (Fig. 3). By contrast, genes encoding olfactory and vomeronasal receptors all show considerable expansions in snakes (Castoe et al., 2013), implying that they may rely more on other sensory systems such as olfaction than on bitter taste sensation in detecting toxic compounds.

No Tas2r genes were detected in the common garter snake. It is likely because the genome of the common garter snake contains so much ambiguous sequencing (poly N) that might prevent identification of the true number of Tas2r genes. Studies regarding feeding habits showed that the common garter snake would exclude some toxic frogs after making oral contact with them, suggesting they could sense the bitterness of poisonous food (Brodie & Tumbarello, 1978; Dumbacher, Menon & Daly, 2009). Consequently, the numbers of Tas2r genes in the common garter snake must be regarded as tentative and needs to be validated further.

Interestingly, both of the 2 lizard species examined here displayed a noticeable expansion of the Tas2r repertoire, which was also detected in the western clawed-frog (Xenopus tropicalis) and coelacanth (Latimeria chalumnae) (Behrens, Korsching & Meyerhof, 2014; Wang & Zhao, 2015). In contrast to snakes, the tip of tongue in most lizards has a high concentration of taste papillae (Schwenk, 1985). In concordance with this anatomical evidence, responses to bitter-compound-coated preys has also been widely demonstrated in many lizards (Benes, 1969; Dumbacher, Menon & Daly, 2009; Paradis & Cabanac, 2004; Price-Rees, Webb & Shine, 2011; Pricerees, Brown & Shine, 2012; Shanbhag, Ammanna & Saidapur, 2010; Stangerhall et al., 2011). Additionally, the number of pseudogenes was considerably higher than that of snakes and other reptiles. To determine whether tandem duplication had taken place in reptilian Tas2rs as it has in mammals and birds (Wang & Zhao, 2015), we examined the genomic location for the Tas2r genes. The result indicated that some Tas2rs are aligned in arrays (Table S5). This is consistent with the birth-and-death model of multigene family evolution in which the more gene duplications occur, the more pseudogenes are found (Nei, Gu & Sitnikova, 1997).

The dramatic changes of Tas2r repertoire size may reflect different selection pressures and different evolutionary processes faced by each organism in accordance with constantly changing environments (Go, 2006). On one hand, as a driving force of evolution, diets including different sets of toxins have been suggested to have shaped the variation of Tas2r repertoire across species (Li & Zhang, 2014; Wang & Zhao, 2015). In the present study, we extended this hypothesis to a different group of Squamates. It has been demonstrated that animals with extremely narrow diets may have small functional Tas2r repertoires. For example, vampire bats which feed exclusively on blood have a considerably smaller percentage of intact Tas2rs than other bats (Hong & Zhao, 2014). Snakes are strictly carnivores, feeding on animals including small mammals, birds, eggs, frogs, lizards, snails, fish, insects or other snakes (Mehrtens, 1987; Parker & Gans, 1978). This food composition contains few toxic chemicals, except for some defensive toxic secretions released by insects or newts (Brodie, Ridenhour & Iii, 2002). On the other hand, the foraging pattern may also affect the Tas2r repertoire. For example, swallowing preys items whole without chewing results in less contact with poisonous foods. In minke whales, eight out of nine Tas2r genes identified from 12 whale genomes are pseudogenes (Feng et al., 2014): this perhaps also relates to foraging patterns. Considering the dietary and swallowing foraging pattern of snakes, a reduced Tas2r repertoire makes sense. However, the small Tas2r repertoires in snakes do not necessarily imply a reduced importance of bitter taste, as smaller repertoire size could be compensated for by a large tuning breadth to bitter substances, as has been shown in human, chicken and turkey Tas2rs (Behrens, Korsching & Meyerhof, 2014; Meyerhof et al., 2010). In contrast, the green anole and the Japanese gecko are lizards with diets that mainly consist of small insects like grasshoppers, crickets, flies spiders, and other arthropods. Since most of the prey can release defensive toxic secretions (Braswell et al., 1980; Jensen et al., 2008). Therefore, lizards may encounter more bitter substances than snakes and more Tas2r genes may be required to help avoiding certain harmful compounds in insects. Our PICs analysis clearly showed a significant positive correlation between the repertoire of Tas2rs in reptiles and the amount of potential toxins in their diets. However, this analysis will be strengthened when more Tas2r data of reptiles are available. Furthermore, our results do not necessarily mean that there are no other factors influencing the Tas2r repertoire. Extra-oral functions in other systems outside the oral cavity (e.g., in central nervous system, respiratory system or cardiovascular system) may also drive the evolution of Tas2rs genes (Behrens, Prandi & Meyerhof, 2014). Some studies have suggested neutral (rather than diet-related) reasons that are responsible for variations in chemosensory repertoire size. For example, Go found that the contraction of avian Tas2r repertoire was concomitantly with reduction of the genome size (Go, 2006). In order to determine the potential relationship between the Tas2r repertoire and genome size in reptiles, we calculated the average genome size of reptiles using data from Animal Genome Size Database (http://www.genomesize.com/), which provide genome size as haploid DNA contents (C-values, in pg). As the Tas2r repertoire of snakes was smaller than birds whereas the average genome size of snakes (2.14 pg) was larger than birds (1.43 pg), we inferred that genome size reduction was not the necessary cause for the contraction of the Tas2r repertoire in snakes (Table S6). Meanwhile, the average genome size of snakes (2.14 pg) and lizards (2.15 pg) were approximately the same, suggesting that the contraction and expansion of Tas2r genes were not related to genome size in the Squamates clade.

Apart from taste receptors, signaling pathways downstream of taste receptors are indispensable for taste function. For instance, calcium homeostasis modulator 1 (Calhm1) mediates purinergic neurotransmission of bitter taste stimulant. Calhm1 knock-out mice showed severely damaged reactions to bitter stimuli (Taruno et al., 2013). Actually, pseudogenization of Calhm1 gene occurred in some species. For example, Feng found that Tas2r genes were pseudogenized in 11 whale species. Among these species, the Beluga Whale (Delphinapterus leucas) and the Omura’s whale (Balaenoptera omurai) showed indels in Calhm1 gene, resulting in pseudogenization and functionless. However, the Calhm1 gene of 9 other whales is intact and putatively functional (Feng et al., 2014). By TBLASTN, we successfully identified intact Calhm1 genes in each genome of the eight snake species (Data S3 and S4). Furthermore, to explore the functional implications of Calhm1 in snakes, we estimated the ω ratio for the Calhm1 genes with a likelihood approach by PAML. The approach was based on the ratio of non-synonymous to synonymous substitution rates (dN/dS, ω), with ω being <1, equal to 1 and >1 indicating negative or purifying selection, neutral evolution and positive selection, respectively. The ω ratio for the Calhm1 genes is significantly smaller than 1 (ω = 0.12, Table S7), suggesting that Calhm1 genes are under strong purifying selection and thus putatively functional in snakes. Combined with the genomic data of intact and deduced functional Tas2r genes in snakes, our result explicitly indicated that snakes still retain the genetic basis for bitter taste perception.

We systematically identified the Tas2r genes in 19 reptiles, and present the first description of the Tas2r repertoires of snakes. Our analysis of genomic data demonstrates that snakes may indeed possess bitter taste perception. These results point to a reduction of Tas2rs in snakes-at least in the nine species we analyzed-and provide novel insights into the evolutionary biology of taste perception and food selection of snakes and other reptiles. Future expressional and behavioral studies regarding bitter taste of reptiles will be interesting to pursue.

Supplemental Information

Table S6. Genome Size in reptiles.

Genome size (pg) is the amount of DNA in one copy of a species’ chromosomes or the haploid nuclear DNA content.

Table S7. Likelihood ratio tests of selective pressures on Calhm1 genes of snakes.

Branch model in PAML was used to compute the ω value.

Funding Statement

This study was supported by the Special Fund for Forest Scientific Research in the Public Welfare (No. 201404420) and the National Natural Science Fund of China (No. 31372220, No. 31172119). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.