DISCUSSION
In this study, we showed that two directionally selected nonsynonymous sites, A52V and Q296H, are associated with decreased sensitivity to quercitrin in Qinling pandas, which confers upon the pandas a dietary preference for high-quercitrin bamboo leaves. In agreement with the hypothesis that polymorphisms in sensory receptor genes may alter perception by encoding functionally distinct receptor types (Bufe et al., 2005; Wooding et al., 2006), we add new evidence that genetic variations in a bitter receptor gene (pTAS2R20 ) are correlated with differences in bitterness (quercitrin) recognition. Bitter taste perception provides an important means for animals to detect bitter compounds in the environments that they occupy. As animals explore new environments and diets, they may encounter distinct bitterness, which imposes selective pressure on their bitter taste receptor genes, leading to individual differences in bitter taste sensitivity (Shi et al., 2003; Wooding et al., 2006; Li et al., 2013; Shan et al., 2018). The TAS2R genes of the giant panda are expected to have experienced selective pressure during the course of the dietary shift of pandas from carnivorous to omnivorous to herbivorous and, finally, their specialization for bamboo consumption. Our previous study demonstrated that the giant and red pandas harbor more putative functional bitter taste receptors than other carnivores, and several of their TAS2R genes seem to have experienced selective pressure, as these species are challenged with many more bitter compounds in diets (Shan et al., 2018). Although natural selective signatures on several TAS2R genes in the pandas has been revealed, the functional effects of the selective pressure on bitter taste perception have not been examined. The present study has provided a fine example of the effects of directional selection on giant panda dietary preferences, providing direct evidence of a close correlation between taste function and the pandas’ feeding ecology.
pTAS2R20 was shown to be specifically activated by quercitrin, a flavonoid monomer found in various plants including bamboos that has previously been identified as the agonist of a bitterness receptor of the herbivorous insect Papilio hospiton (Sollai et al., 2015). Studies in humans have suggested that bitter taste receptors are usually broadly tuned to recognize numerous compounds possessing common chemical groups responsible for mediating receptor-agonist interactions (Bufe et al., 2005; Meyerhof et al., 2010). Given that a limited number of bitter compounds were assayed here and that bamboo produces flavone compounds, the actual number of dietary agonists for TAS2R20 may be greater. However, the two directionally selected sites, A52V and Q296H, indeed conferred TAS2R20 with a significantly decreased sensitivity to quercitrin, which may reduce aversion to quercitrin in Qinling pandas. Consequently, although higher quercitrin was contained in the bamboo leaves from B. fargesii and F. qinlingensis in Qinling Mountains, Qinling pandas taste the leaves better than other pandas taste the bamboo leaves from F. denudate and B. faberi in other areas. Therefore, the two directionally selected nonsynonymous sites may explain why Qinling pandas prefer to consume more bamboo leaves than other pandas do.
Recent studies on chemical substances in the giant panda staple food bamboo have focused on nutrition and minerals (Christian et al., 2015; Nie et al., 2015b), but there are few systematic reports of bitter substances. Thus, we screened agonists for pTAS2R20 from a limited number of common chemicals found in bamboos and other plants. According to previous studies indicating that giant pandas are inclined to ingest bamboos that contain less tannin (Zhao et al., 2001; Hansen et al., 2010), we listed quercitrin and tannin as important potential agonists of the receptor. Here, functional expression experiments confirmed that quercitrin is an agonist of pTAS2R20 that may play a significant role in feeding preference. Although tannin is the correct agonist, as we previously conjectured (Zhao et al., 2013), this result is not surprising, since giant pandas possess numerous bitter receptors, and some of them may specifically recognize tannic acid.
The giant panda belongs to the order Carnivora (Wei et al., 2014); nevertheless, it has evolved certain characteristics of herbivores including a skull, jaw musculature and dentition that are suitable for a fibrous diet and a pseudothumb that is specifically used for handling bamboo (Zhang et al., 2007; Hu et al., 2017). Mutations in DUOX2led to a decrease in thyroxine levels, which helped to reduce energy consumption for adaptation to a low-energy bamboo diet (Nie et al., 2015a). These studies have emphasized behavioral, physiological, morphological, and genetic adaptations. However, it is rare to reveal the molecular basis of feeding adaptation in giant pandas from the perspective of bitterness receptors function, with the exception of similar reports on the pseudogenization of the umami taste receptor gene TAS1R1, which is associated with eating meat (Hu et al., 2017). This research fills a gap in explaining the evolution of diet adaptation from the perspective of the molecular function of bitterness receptors, especially in non-model animals. This study is also a typical example for linking genome-scan candidate gene to ecological adaptation at functional verification level.