Divergence of T2R chemosensory receptor families in humans, bonobos, and chimpanzees

Abstract

T2R (Tas2R) genes encode a family of G protein-coupled gustatory receptors, several involved in bitter taste perception. So far, few ligands for these receptors have been identified, and the specificity of most T2Rs is unclear. Differences between individual T2Rs result in altered taste perception in either specificity or sensitivity. All 33 human T2Rs are characterized by significant sequence homology. However, with a total of eight pseudogenes and >83 coding region single-nucleotide polymorphisms, the family displays broad diversity. The underlying variability of individual T2Rs might be the source for personalized taste perception. To test this hypothesis and also to identify T2Rs that possibly function beyond bitter taste, we compared all human T2R genes with those of the closely related primate species Pan paniscus (bonobo) and Pan troglodytes (chimpanzee). The differences identified range from large sequence alterations to nonsynonymous and synonymous changes of single base pairs. In contrast to olfactory receptors, no human-specific loss of the amount of functional genes was observed. Taken together, the results indicate ongoing evolutionary diversification of T2R receptors and a role for T2Rs in dietary adaptation and personalized food uptake.

Chemoreception is a species-specific process that is a prerequisite for survival in a given environment. Throughout the animal kingdom, different molecular sensory systems evolved to serve this purpose. For gustation, recent advances have shed light on the intricacies of the receptor-based machineries expressed in signal-specific cell types (1–8). Most mammalian gustatory senses include the five recognized modalities: sweet, sour, bitter, salty, and umami (monosodium glutamate and 5′ nucleotide monophosphate). Different taste cell types with distinct specificities are located within a taste bud expressing taste receptors on the apical surface. Salty and sour taste use Na⁺ and H⁺ selective ion channels, which permeabilize the membrane for these ions to facilitate direct depolarization (7). The remaining modalities, bitter, sweet, and umami taste, are mediated by G protein-coupled receptors. Nutrients such as amino acids and sugars bind to T1R receptors expressed in cells that relay the binding of a ligand to higher cortical areas (1, 3, 4, 9–12). For umami taste, it is the T1R1/T1R3 heterodimer that binds both glutamate and nucleotides (1, 3, 13). Sweet tastants are perceived at high concentrations by the T1R2 and/or T1R3 monomers. For low concentrations, T1R2 and T1R3 form heterodimers, which also have been shown to bind artificial sweeteners (14).

The largest family of taste receptors is the Tas2R or T2R family of G protein-coupled receptors, which have been linked to bitter taste (5, 6). The human T2R family is composed of 33 members with a high degree of similarity, including eight pseudogenes and at least 83 single-nucleotide polymorphisms (SNPs), indicating large variability (15). Currently, only a few ligands for these receptors have been identified, such as phenylthiocarbamide (PTC) for human T2R38 (16), strychnine for human T2R10 (17), salicin for human T2R16 (17), and denatonium benzoate for human T2R4 (5).

Differences within individual T2R amino acid sequences lead to differences in ligand specificity and taste perception. In mouse T2R5, this has been demonstrated first with the identification of the genetic basis for different cycloheximide taster strains (5). Another example is the identification of T2R38 as the receptor for phenylthiocarbamide (PTC), where taster and nontaster alleles differ by only three amino acid replacements, Pro-49 to Ala, Ala-262 to Val, and Val-269 to Ile (16).

The T2R genes, like odorant receptors and several other G protein-coupled receptors, do not contain introns, and valuable insight can be obtained by comparing genes of closely related as well as more distant species (6, 18). Regions or even single amino acid residues that are conserved among distant species can indicate the areas that are essential for the function of the protein. Regions that change among closely related species or individuals within a species are likely to be involved in fine tuning the actions of the protein, in the case of the T2Rs either altering the ligand-binding region or affecting signaling and coupling to the G protein.

Humans' closest relatives in the animal kingdom are pygmy chimpanzees or bonobos (Pan paniscus) followed by chimpanzees (Pan troglodytes). We have analyzed all T2R receptors of pygmy chimpanzees and compared them to those of humans and those available for chimpanzees to identify patterns of variability and conservation alike. Several of the great ape T2Rs have significant differences from the human homologues leading to amino acid changes that could have an impact on the function of the encoded protein. Moreover, some genes that exist in humans turn out to be pseudogenes in the animals studied and vice versa. Most differences are at the single base pair level, and some of them are identical to SNPs within the human population.

Materials and Methods

DNA Preparation and Sequence Analysis. Genomic DNA from an adult bonobo blood sample was isolated by using the Hi Speed Plasmid Midi Kit (Qiagen, Chatsworth, CA). Homologues of all known human T2R genes were amplified by PCR by using High Fidelity polymerase (Roche Diagnostics) and primers specific for each gene (Table 2, which is published as supporting information on the PNAS web site). Standard PCR conditions were used throughout, with annealing temperatures and MgCl₂ concentrations optimized for each primer pair. All PCR reactions were carried out in quadruple before being pooled to reduce the likelihood of PCR-induced errors. Amplified T2R genes were purified (QIAquick PCR Purification Kit, Qiagen), incubated with Taq polymerase (Roche Diagnostics) and dNTPs at 72°C for 10 min, purified by gel extraction (QIAquick Gel Extraction Kit, Qiagen), and ligated into pGEM-TEasy (Promega). After transformation into Escherichia coli DH5α (Invitrogen), at least three individual clones for each T2R have been sequenced by using SP6 and T7 primers (MWG Biotech, Ebersberg, Germany). The resulting sequences were analyzed by using seqman ii sequence analysis software (DNASTAR, Madison, WI). If there were differences among clones, further sequencing was carried out until a consensus could be confidently established. Calculation of nucleotide substitution rates was performed by using the Jukes–Cantor model implemented in the distmat program of the emboss suite (19).

Nomenclature of Human T2R Genes. Literature data on T2R nomenclature are not in agreement, and many genes are duplicated in nucleotide databases, because they were identified by several groups. To identify duplicates in this study, blast searches were performed against a nonredundant database (www.ncbi.nlm.nih.gov/BLAST) and the human genome (www.ncbi.nlm.nih.gov/genome/guide/human) (20) (Table 1). Names on the left are taken from the human genome and are used in our study. T2R55^$$ (21) is the same as T2R24 (6), which cannot be found in the databases. T2R55^$ (15) is not in the databases but is identical to T2R48 (19). T2R56** (15) and T2R56* (21) are different. The former is a SNP of T2R49 (17), whereas the latter is the same as T2R60 (15). T2R46 and T2R54 are the same; however, T2R46 is 30 bp (10 aa) shorter. The genomic draft has the DNA and protein sequence annotated as T2R46, but the DNA sequence corresponds to T2R54. The other listed T2Rs are either identical (62PS/ps1, 63PS/ps6, 64PS/ps2, and 65PS/ps4) or slight variations of each other probably due to SNPs (38/61, 41/59, 43/59, 44/53, 46/54, and 50/51).

Table 1. Human T2Rs.

Name	Ref.	Equivalent	Ref.
T2R38	Bufe et al. (17)	T2R61	Conte et al. (15)
T2R39	Bufe et al. (17)	T2R57	Conte et al. (15)
T2R40	Bufe et al. (17)	T2R58	Conte et al. (15)
T2R41	Bufe et al. (17)	T2R59	Conte et al. (15)
T2R43	Bufe et al. (17)	T2R52	Conte et al. (15)
T2R44	Bufe et al. (17)	T2R53	Conte et al. (15)
T2R46	Bufe et al. (17)	T2R54	Conte et al. (15)
T2R48	Bufe et al. (17)	T2R55§	Conte et al. (15)
T2R49	Bufe et al. (17)	T2R56**	Conte et al. (15)
T2R50	Bufe et al. (17)	T2R51	Conte et al. (15)
T2R55$$	Shi et al. (20)	T2R24	Adler et al. (6)
T2R60	Conte et al. (15)	T2R56*	Shi et al. (20)
T2R62PS	Conte et al. (15)	hPS1	Shi et al. (20)
T2R63PS	Conte et al. (15)	hPS6	Shi et al. (20)
T2R64PS	Conte et al. (15)	hPS2	Shi et al. (20)
T2R65PS	Conte et al. (15)	hPS4	Shi et al. (20)

Nomenclature of doubly cited T2R entries in the literature.

Results

Sequence Analysis of Human, Bonobo and Chimpanzee T2Rs. The T2R genes of P. paniscus (bonobo or pygmy chimpanzee) are distinct although clearly related to those available through the P. troglodytes (chimpanzee) sequencing consortium (www.ensembl.org/Pan_troglodytes). In many cases, T2R sequences of P. troglodytes are not available or are incomplete. Qualitatively, the differences among the three species range from large deletions/insertions among specific orthologs to conversions from pseudogenes to functional genes and vice versa to individual synonymous or nonsynonymous base-pair replacements. Most differences among the species are distributed without an apparent pattern across the sequences. Within the T2R family, no significant clustering of differences in a given structural feature such as cytoplasmic or extracellular domains or the putative transmembrane regions could be detected (Fig. 4, which is published as supporting information on the PNAS web site, for a full alignment with all differences). However, certain individual genes did display differences at the C terminus, indicating possible alterations in the downstream signal processing and the interaction with the G protein.

Conversions Among Functional Genes and Pseudogenes. The occurrence of pseudogenes and the relationship among the T2R genes of humans, bonobos, and chimpanzees are outlined in the cladogram in Fig. 1. For clarity, we will refer to human, bonobo, and chimpanzee T2Rs as hT2R, bT2R, and cT2R, respectively.

Fig. 1.

Cladogram of pseudogenes in hT2R, bT2R, and cT2R DNA. Upon divergence from the common primate ancestor, T2R65PS, PS5, PS3, PS8, PS7, and T2R63PS remained nonfunctional across all species. T2R9 is a pseudogene only in bonobos, whereas T2R62 is a pseudogene only in humans. T2R64 is functional only in bonobos. Moreover, in bonobos, a new T2R with close relation to T2R46 has been identified (not shown; see text). The graph indicates no species-specific loss of T2R function upon primate divergence and instead species-dependent specialization. The cladogram for all T2Rs analyzed is available in Fig. 5, which is published as supporting information on the PNAS web site.

Straightforward analysis of differences in the genes is sometimes hindered by polymorphisms for a given T2R with sequences different in the DNA database entry from the sequence published in the human genome database. With hT2R62, the genomic DNA sequence contains stop codons at position 624 and 793. The corresponding database entry (NG_002653) has an 18-bp deletion in the middle of the gene removing the stop codon at base 624. The two stop codons in the human genomic sequence are read through in the great apes, leading to full length bT2R62 and cT2R62 genes (Fig. 2A).

Fig. 2.

Pseudogenes of selected human, bonobo, and cT2Rs. (A) For hT2R62, an 18-bp indel at position 625 inserts an additional stop codon into the genomic entry that is absent in the database entry. A read through of these two stop codons in both bonobos and chimpanzees converts c/bT2R62 into a functional gene. (B) A read through of the stop codon in hT2R64 leads to a functional gene in bonobos. (C) In bT2R9, deletion of thymidine 431 induces a frame shift, introducing a downstream stop codon.

The hT2R64 pseudogene has a termination codon found between base pairs 757 and 759, which is absent in both bonobos and chimpanzees, allowing a read through. Therefore, in bonobos, bT2R64 becomes a full-length gene; however, in chimpanzees, cT2R64 is still a pseudogene due to a single base pair insertion at base 106 (not shown), resulting in a frame shift and several downstream termination codons (Fig. 2B).

In contrast to human pseudogenes, whose homologues are full-length genes in either bonobos or chimpanzees, several full-length genes in humans are found to be pseudogenes in bonobos, chimpanzees, or both. The bonobo T2R9 homologue (bT2R9) has a base pair deletion at position 431, leading to a frameshift and a stop codon 78 base pairs downstream. T2R9 therefore becomes a pseudogene in the bonobo; however, the deletion is not present in the chimpanzee gene, which remains a full length gene (Fig. 2C).

C-Terminal Differences Among Primate T2Rs. The human genome entry for hT2R39 has a single base-pair insertion at position 967, leading to a frame shift and a C terminus with 16 altered amino acids compared to the original sequence in the National Center for Biotechnology Information database (accession nos. AF494230 and NM_176881). This single additional base pair is also present in both bT2R39 and cT2R39 (Fig. 3A).

Fig. 3.

C-terminal alterations across species. (A) A single-nucleotide deletion of an adenosine at position 967 of the database hT2R39 leads to a frameshift affecting the subsequent 15 C-terminal amino acids. (B) A SNP at position 955 of T2R7 introduces a stop codon in the database version that is found neither in the human genome nor in the great apes. (C) The bonobo T2R45 has an extended 3′ end to the human ORF. Tryptophan replaces the stop codon found in humans, adding 27 coding base pairs.

The bonobo and chimpanzee T2R7 gene are both 21 bp longer than human T2R7, as deposited in the database. In bonobos, the GGA codon for glycine replaces the TGA stop codon of the human gene and is followed by 21 coding base pairs before an in-frame TAG termination codon resulting in seven additional amino acids at the protein C terminus. This polymorphism is also found within humans, as evidenced by different entries for hT2R7 in GenBank and the genomic database (Fig. 3B). A similar read through occurs in the bonobo T2R45 gene, where the stop codon present in humans is replaced by TGG encoding tryptophan. This is followed by 27 additional coding base pairs and a TAG stop codon, resulting in a protein that is 10 amino acids longer than the human counterpart (Fig. 3C). In all cases studied, it is single base-pair replacements and not large-scale insertions that lead to substantial rearrangements of the receptor C terminus.

Single Base-Pair Differences Among the Three T2R Families. The most frequent differences found among the T2R sequences studied are at the single base-pair level. Bonobo and chimpanzee sequences were compared to the database entries for each hT2R. As expected, in most cases, bT2R and cT2R families are more closely related to each other than they are to humans.

A comparison of the differences among the three primate T2R gene families with entries found within the human SNP database indicates that some of the differences between the great ape genes and their human counterparts can also be found as SNPs within the human population. For example, the three SNPs required for phenylthiocarbamide (PTC) tasting are all present in both bT2R38 and cT2R38, implying that the individuals from whom the DNA was extracted can taste PTC. However, in most examples, human SNPs are distinct from the differences found between the human and primate genes, suggesting that these genes have continued to evolve since the separation of humans and the great apes.

Some of the great ape T2Rs are almost identical to their human counterparts, whereas others show a large degree of difference. In general, the nucleotide substitution rates between bT2Rs and hT2Rs range from 0.11 in h/bT2R3 to 1.83 in h/bT2R63PS (for a detailed listing, see Fig. 5). At one end of the spectrum of functional genes, bT2R3 has only a single synonymous base-pair replacement compared with the human homologue, and receptor bT2R4 has not a single base pair different from hT2R4. At the other extreme, bT2R13 and bT2R46 were most different from their human counterparts, with a total of up to 13 base pair differences, 10 of which are nonsynonymous leading to alteration of amino acids (Fig. 4).

A Previously Undescribed T2R in Bonobos. After amplification of the bonobo bT2R46 ORF and analysis of the resulting sequence, a number of clones were identified that did not align completely with hT2R46. The sequence of these clones appears to be a hybrid of two hT2R genes, with the first 500 bases similar to hT2R45 and the remainder similar to hT2R46. blast searches using the entire 930 bases of this previously undescribed bT2R do not indicate a homologue in the human genome; instead, the closest matches are T2R45 and T2R46, as would be expected from the hybrid nature of the gene. Both attempts to amplify and clone a human version of this previously undescribed bT2R, and a genome database search yielded no results, indicating that the gene is not present in humans. The putative gene has been called bT2R66.

Discussion

The T2R gene repertoire of humans closest evolutionary relatives, bonobo chimpanzees (P. paniscus), was cloned, sequenced, and analyzed. Sequences were compared with their human homologues and the available sequences of the common chimpanzee (P. troglodytes) to determine alterations, which possibly affect their function.

Despite a high degree of identity among hT2R, bT2R, and cT2R families (>98% between humans and bonobos), all members display significant across-species differences. Most notably, some pseudogenes in one species appear to be functional genes in the other species or vice versa (Fig. 1). Moreover, additional divergence is seen in the development of different 3′-ends for homologous genes. Those differences in the T2R protein C termini (Fig. 3) are likely to affect interactions with the corresponding G protein, leading to an enhanced or reduced efficiency of downstream signal transduction (22, 23). Finally, differences in individual base pairs among all homologues of the three species add a third degree of diversity. Being at the root of molecular divergence, single base-pair replacements can turn a pseudogene into a functional gene, as found for T2R62. However, in a more subtle way, these differences can contribute to the fine tuning of a given receptor for its ligand.

In a recent study, a similar analysis was performed on the divergence between all human and chimpanzee olfactory receptors (24). In contrast to the T2R taste receptor family, the human olfactory receptor repertoire is much larger and includes several hundred genes, which, like the T2Rs, are not spliced (19). Interestingly, Gimelbrant et al. (24) find no evidence for positive selection on the olfactory receptor repertoire after the chimpanzee–human divergence, and Gilad et al. (25) even described human-specific loss of olfactory receptor genes. This is in contrast to the data presented here on T2R taste receptors, where the numbers of functional genes remained similar after species bifurcation. Instead, certain identified differences imply adaptation of the receptor sequence.

The discovered differences between gustatory and olfactory receptor repertoires are logical from a biological perspective. Primates do not fundamentally rely on their sense of smell, possibly since the acquisition of color vision (26), and no strong selection pressure was imposed on their olfactory receptor repertoire. This is different for gustation, which, in contrast to olfaction, serves not preingestive orientation but more imminent needs, such as control and evaluation of ingested food. In particular with bitter taste, the need for rapid adaptation of fundamental sensory mechanisms was relevant to maintain a life-saving warning mechanism.

Recognition of nutrients such as sugars (sweet) and monosodium glutamate (umami) is not necessarily subject to this kind of selection because of the ubiquitous nature of these components. Both sweet and umami taste are mediated by multiply spliced T1R receptors, with T1R3 being common to both modalities.

A fast selection of the correct gustatory repertoire is not only beneficial but also critical for survival of an evolving species. The common ancestors of humans, bonobos, and chimpanzees expanded their habitat beyond central Africa some five and a half million years ago (27). As a result, they had to survive in a changed environment with a different diet. Therefore, in an extreme view, T2R taste receptors can be seen as a mirror image of the biological environment consisting of plants with toxic metabolites. Accurate tasting using a responsive taste receptor repertoire is the quintessential prerequisite for meaningful taste memory, the subsequent hedonic evaluation, and ultimately an optimal food choice.

Finally, T2Rs have been discovered in tissues unrelated to gustation. Found in the nose (28), they might contribute to the perception of flavors. Expressed in rodent (29, 30) and human intestinal tissues (unpublished results), they imply a significance for chemical perception larger than just sensory stimulation. If gastro-intestinal taste cell systems contribute to the complex analysis of ingested food to determine its metabolic value, T2R genes might be involved with food choice and preference. In this case, additional pressure for adequate T2R selection can be expected (31).