INTRODUCTION
Animals taste receptor genes evolve in response to species-specific diets (Li et al., 2005; Zhao et al., 2010; Jiang et al., 2012; Sato et al., 2012; Liu et al., 2016; Shan et al., 2018). Food-derived compounds stimulate taste receptors on specialized cells within taste buds, causing pleasant or unpleasant sensations and, thus, influencing the dietary preferences of animals (Chandrashekar et al., 2006; Breslin, 2013; Fujikura, 2015). Among the possible taste sensations, bitter taste, mediated by bitter taste 2 receptors (TAS2Rs), is viewed as a defense system of animals against toxic compounds in food, as toxins usually taste bitter and trigger aversive reactions. In nature, there are a wide variety of bitter substances such as plant alkaloids, most of which may be toxic and harmful (Lossow et al., 2016). Given the ubiquitous bitterness that exists in the environment that species occupy, animals harbor a family of receptors that are responsive to bitter compounds (Bufe et al., 2002), and each TAS2R is commonly responsive to several bitter compounds (Meyerhof et al., 2010). Studies have demonstrated that the numbers of TAS2R vary greatly among species from a few to tens of these receptors, and they have been lost and gained frequently in the course of animal evolution (Li et al., 2013; Hayakawa et al. , 2014; Liu et al., 2016). A correlation has been revealed between the TAS2R number in a species and the fraction of plants in its diet, because plant tissues contain more toxic compounds than animal tissues, indicating that dietary toxins are a major selective force shaping the evolution of TAS2Rs (Li et al., 2013). In addition, it has been reported that the nonsynonymous K172N site in the human bitter receptor TAS2R16 gene was associated with an increased sensitivity to salicin, arbutin, and five different cyanogenic glycosides (Soranzo et al., 2005); sequence variants in the TAS2R38 gene are correlated with variable taste sensitivity to the bitter compound phenylthiocarbamide (PTC) in humans, chimpanzees, and macaques (Wooding et al., 2006; Suzuki et al., 2010). These reports suggested that receptor variants could result in different sensitivities to the same bitter compounds, leading to species- and/or population-level differences in dietary preferences, reflecting the potential roles of these variants in the transformation of animal feeding and dietary adaptation (Bufe et al., 2005; Soranzo et al., 2005; Wooding et al., 2006; Nei et al., 2008).
The giant panda (Ailuropoda melanoleuca ) is a specialized herbivore in the order Carnivora that feeds almost exclusively on highly fibrous bamboo (Wei et al., 2015; Nie et al., 2015a; Nie et al., 2019). However, paleontological and molecular evidence suggests that ancient pandas were carnivorous or omnivorous and that they switched to a plant diet at least ∼7.0 million years ago (Jin et al., 2007; Zhao et al., 2010). In addition, a recent report based on stable isotope analyses suggests that ancient pandas may have had more complex diets than modern pandas (Han et al., 2019). These multiple lines of evidences indicate that plants have been components of pandas’ diet for several million years, and this type of diet poses the challenge of tolerating large amounts of bitter compounds. Our previous work showed that there are more putatively functional TAS2Rs in the giant panda than in other carnivores, which might be expected because the abundant bitter substances encountered by pandas could lead to a requirement for more functional TAS2Rs for bitter taste perception (Shan et al., 2018). The purifying selection pressure on three TAS2R genes (TAS2R1 ,9 , and 38 ) is markedly strengthened in the species, suggesting that these three receptor gene sequences are specifically highly conserved, probably because of the presence of some sites that are functionally more important for the detection of certain bitter compounds in the panda diet. Additionally, signatures of positive selection were detected for TAS2R42 and TAS2R49 in pandas.TAS2R49 is now designated as TAS2R20 according to the last Gene Nomenclature Committee of the Human Genome Organization (http://www.genenames.org/, last accessed April 30, 2016), and this gene has been directionally selected at two nonsynonymous sites A52V and Q296H in the panda population from the Qinling Mountains (Qinling pandas) (Zhao et al., 2013; Shan et al., 2018). Consistent with this finding, field observations showed that Qinling pandas consume more bamboo leaves than pandas in other areas (Schaller et al., 1985; Pan et al., 2001); population genetic data indicated their divergence from other pandas ~0.3 million years ago and showed genetic adaptation to their environments (Zhao et al., 2013; Wei et al., 2014). These findings collectively raise the question of whether the two nonsynonymous sites in TAS2R20 are the causative base-pair changes resulting in the preference of Qinling pandas for the consumption of more bamboo leaves than the pandas from other areas. We hypothesized that the two nonsynonymous sites in TAS2R20 encode receptor variants that may decrease Qinling pandas’ taste sensitivity to bitter compounds, causing bamboo leaves to taste less bitter to the pandas.
To address this hypothesis, we first challenged pTAS2R20 with several common bitter substances (caffeine, sesquiterpene lactone, denatonium benzoate, chloroquine, picrotoxinin, cycloheximide, and nicotine), and some known bamboo-derived bitter chemicals (quercitrin, tannin, salicin, aloin, coumarin, amygdalin and galangin) in a heterologous expression system. Among these bitter compounds, pTAS2R20 was specifically activated by quercitrin, a flavonoid monomer found in various plants including bamboos. Then, we used high-performance liquid chromatography (HPLC) to quantify the quercitrin contents of the leaves ofBashania fargesii and Fargesia qinlingensis for which Qinling pandas show the strongest preference, and compared the results with the quercitrin contents of the leaves of Fargesia denudataand Bshania faberi , which other pandas prefer. Finally, four pTAS2R20 variants were generated and challenged with quercitrin. The two significantly selected nonsynonymous sites in TAS2R20 occur at amino acid position 52, where either an alanine or a valine is encoded, and position 296, where either a glutamine or a histidine is encoded, giving rise to AQ, VQ, AH, and VH receptor variants. In nature, only two haplotypes VH and AQ are found in Qinling pandas and pandas from other areas, respectively, whereas VQ and AH are mutated variants used for examining the effect of each of the two nonsynonymous sites on the function of pTAS2R20 in response to its agonist. By combining these strategies, we expected to verify the hypothesis, and reveal how polymorphisms in pTAS2R20 influence quercitrin perception in giant pandas, providing an example of the functional effects of directional selection in local population adaptation.