Genomic signatures of sensory adaptation and evolution in pangolins Authors: Jun Li, Ce Guo, Meiling Xie, Kai Wang, Xianghe Wang, Bishan Zou, Fanghui Hou, Chongyang Ran, Shiman Bi, Yanchun Xu, Yan Hua Published: 2024 ABSTRACT BackgroundPangolin is one of the most endangered mammals with many peculiar characteristics, yet the understanding of its sensory systems is still superficial. Studying the genomic basis of adaptation and evolution of pangolin’s sensory system is expected to provide further potential assistance for their conservation in the future.ResultsIn this study, we performed a comprehensive comparative genomic analysis to explore the signature of sensory adaptation and evolution in pangolins. By comparing with the aardvark, Cape golden mole, and short-beaked echidna, 124 and 152 expanded gene families were detected in the genome of the Chinese and Malayan pangolins, respectively. The enrichment analyses showed olfactory-related genomic convergence among five concerned mammals. We found 769 and 733 intact OR genes, and 704 and 475 OR pseudogenes in the Chinese and Malayan pangolin species, respectively. Compared to other mammals, far more intact members of OR6 and OR14 were identified in pangolins, particularly for four genes with large copy numbers (OR6C2, OR14A2, OR14C36, and OR14L1). On the genome-wide scale, 1,523, 1,887, 1,110, and 2,732 genes were detected under positive selection (PSGs), intensified selection (ISGs), rapid evolution (REGs), and relaxed selection (RSGs) in pangolins. GO terms associated with visual perception were enriched in PSGs, ISGs, and REGs. Those related to rhythm and sound perception were enriched in both ISGs and REGs, ear development and morphogenesis were enriched in ISGs, and mechanical stimulus and temperature adaptation were enriched in RSGs. The convergence of two vision-related PSGs (OPN4 and ATXN7), with more than one parallel substituted site, was detected among five concerned mammals. Additionally, the absence of intact genes of PKD1L3, PKD2L1, and TAS1R2 and just six single-copy TAS2Rs (TAS2R1, TAS2R4, TAS2R7, TAS2R38, TAS2R40, and TAS2R46) were found in pangolins. Interestingly, we found two large insertions in TAS1R3, distributed in the N-terminal ectodomain, just in pangolins.ConclusionsWe found new features related to the adaptation and evolution of pangolin-specific sensory characteristics across the genome. These are expected to provide valuable and useful genome-wide genetic information for the future breeding and conservation of pangolins.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12864-024-11063-x. Background As the link to the interaction between animals and the environment, the sensory systems are crucial to the basic activities of animals for survival, which has been suggested by many empirical studies [1]. The environment transmits external information as specific stimuli to animals, meanwhile, animals collect these stimuli by different senses (such as vision, audition, olfaction, and gustation) and further translate it into perception [2]. Providing insights into the sensory features and their genetic bases of animals could play a pivotal role in the conservation and management of wildlife [3]. The pangolin pertains to the order Pholidota and is characterized by overlapping keratin scales over the whole body, a specialized diet on termites and ants, a long and muscular tongue, toothless jaws, elongated rostrum, burrowing ability, and nocturnal habits [4–7]. Pangolin should have received a great number of attentions for only its unique morphological features, however, it become the globally most trafficked wild animal and one of the endangered mammals on earth [8, 9]. The sensory patterns of pangolins remain mysterious, and uncovering the sensory genomic adaption and evolution can help to understand their sensory features and answer how they interact with the environment [10]. However, the genomic evolution of their sensory system is largely unknown. A recent study provided genetic evidence of the remarkable olfaction and suspected poor vision, based on the significant expansion of the olfactory receptor (OR) gene family and the pseudogenization of certain vision-related genes (e.g. BFSP and GUCA1C) in pangolins compared with carnivorous animals [4]. An increasing number of cases support the idea that the gain or loss of receptor genes might severely affect sensation [11, 12]. Accordingly, further uncovering the turnover of the OR gene family, including the duplication and loss of genes, is indispensable to elucidate the genetic mechanisms of the evolution of the sensitive olfaction for pangolins [13]. Additionally, further evidence is still needed to support their degraded vision mentioned in the previous study [14, 15], and sound and taste perception in pangolins have not been meticulously reported. Several high-quality pangolin genomes have been published and released, which provided valuable and rich genomic resources. To find genetic evidence that can help interpret the sensory patterns in pangolins, the present study focused on the genome-wide genetic characteristics related to pivotal senses. The scope of this study was not limited to reported olfaction and vision but included gustation, audition, and other senses. Firstly, the comparative and convergent evolutionary analysis was performed among 15 high-quality genomes of two pangolins (Manis pentadactyla and Manis javanica of Pholidota), three mammals with similar nocturnal, burrowing, and ant-eating lifestyles, indicating similar sensory patterns, of different orders (Orycteropus afer in Tubulidentata, Chrysochloris asiatica in Afrosoricida, Tachyglossus aculeatus in Monotremata), and outgroups. For all senses, we focused on (i) the adaptive pressure on pangolins compared to their outgroups, and (ii) the convergent evolutionary signals among five mammals with similar lifestyles. Moreover, as for their conspicuous and preliminarily reported sense of smell, the intact genes and pseudogenes in the OR gene family (Class I and Class II) were identified and the duplication and loss of ORs in pangolins were further evaluated. Finally, the receptor genes related to different tastes (salty, sour, sweet, umami, and bitter) were identified to reveal the gustatory evolutionary characteristics in pangolins. Our results showed the significant expansion and rapid evolution of ORs, the loss and relaxed selection of taste receptor genes, the positive adaptation related to light and sound perception, and the relaxed evolutionary constraints associated with mechanical stimulus and temperature response in pangolins. These could be helpful for a comprehensive understanding of the sensory adaptation and evolution, and provide new insights into the conservation of pangolins. Materials and methods Genome collection According to the assembly quality assessed using the Benchmarking Universal Single-Copy Orthologs (BUSCO, v5.2.2) [16], we collected high-quality genomes from 15 species available in the China National GeneBank Database (CNGBdb) and the National Center for Biotechnology Information (NCBI). These species included the Chinese pangolin (M. pentadactyla), Malayan pangolin (M. javanica), aardvark (O. afer), Cape golden mole (C. asiatica), short-beaked echidna (T. aculeatus), giant panda (Ailuropoda melanoleuca), domestic cat (Felis catus), dog (Canis lupus familiaris), cattle (Bos taurus), horse (Equus caballus), house mouse (Mus musculus), human (Homo sapiens), Asian elephant (Elephas maximus), platypus (Ornithorhynchus anatinus), and chicken (Gallus gallus) as the outgroup species, which were used for comparative genomic analysis (Table S1). Except for the genomes of two mammals in Monotremata (short-beaked echidna, 94.0%; platypus, 92.6%), the BUSCO completeness values for all other genomes exceeded 95.0% (95.7–98.4%). Gene family analysis The orthologous groups across taxa were constructed using OrthoMCL (v1.4) [17]. During the data filtering process, the longest annotated transcript variants were selected, while sequences with a protein length of less than 30 amino acids were excluded. Additionally, the maximum percentage of stop codons for each sequence was limited to 20%. An all-versus-all BLASTp was performed with an e-value cutoff of ≤ 1E − 05 to identify homologous pairs of sequences. The results of BLASTp were then converted into a normalized similarity matrix, which was analyzed using a Markov Cluster algorithm (MCL) for the family clustering, with an inflation value set to 1.5. Based on the single-copy orthologs obtained with OrthoMCL, multiple sequence alignment was performed using MUSCLE (v3.8.1551) [18] with default parameters, and a maximum-likelihood (ML) phylogenetic tree of 15 species was further reconstructed using RAxML (v8.0.19) [19]. Divergence times were estimated with the MCMCTREE program in PAML (v4.7) [20], with the fossil calibration nodes listed in Table S2. The pairwise divergence times were estimated using TimeTree (v5, https://timetree.org/) [21]. Based on the phylogenetic tree of 15 species and the clustering of gene family by MCL, CAFE (v4) [22] was employed to analyze gene family expansion and contraction in each branch by default parameters. To compare the expansion across the clades of five concerned mammals (M. pentadactyla, M. javanica, O. afer, C. asiatica, and T. aculeatus), we identified the expanded gene family on each of five branches based on the tree of five species using CAFE and further filtered the intersections between these results and the gene families shared by the target clade and outgroups based on the tree of 15 species, and the final results were applied to the comparison of five species. Natural selection analysis To detect signals of sensory-related adaptive evolution across the genome, three types of analyses were performed based on the species tree, according to the published phylogenetic relationships of mammals [23, 24]. We separately designated the basal branch of pangolins, that of aardvark and Cape golden mole, and the branch of short-beaked echidna as the foreground branch, while other species within the same branches were regarded as background branches. The signal of positive selection was detected using the branch-site model in the CODEML program from PAML (v4) [20]. A likelihood ratio test (LRT) was performed (null hypothesis: model = 2, Nssites = 2, fix_omega = 1, omega = 1; alternative hypothesis: model = 2, Nssites = 2) to identify positively selective signals with a significant difference between the null and alternative hypotheses (p-value < 0.05). Additionally, the positively selected amino acid sites were predicted using the Bayes empirical Bayes (BEB) approach. Sites with probability values greater than 95% were considered positively selected and subsequently tested using the false discovery rate (FDR) method. Genes with at least one positively selected site (FDR < 0.05) and a significant difference between the two hypotheses (LRT, p-value < 0.05) were identified as positively selected genes (PSGs). To identify the rapidly evolving genes (REGs), we employed the branch model in the CODEML program. The dN/dS ratios for each branch were calculated using the one-ratio model (M0, model = 0), and the two-ratio model (M2, model = 2) was performed by designating the candidate foreground branch. According to the results of LRT (FDR < 0.05) and the dN/dS ratios for the foreground branch (dN/dS < 1), the genes under rapid evolution were screened. The tests for intensified or relaxed selection were performed by running RELAX in HYPHY (v2.5.48) [25], which could generate a selection intensity parameter k. A k value greater than 1 (k > 1) or less than 1 (k < 1), with a significant p-value (< 0.05), indicates that the selection strength is intensified or relaxed along the tested foreground branch [26]. Genes under the intensified selection (ISGs) and relaxed selection (RSGs) were identified. The visualization of amino acid substitution sites was performed using MultiDisp (http://structure.bmc.lu.se/MultiDisp/index.html). Tests for convergence According to the description by Yuan et al. (2020) [27], a site-wise test on the alignment data was conducted to verify whether the sites were subject to parallel substitutions using RAxML to test for convergence among five concerned mammals. This was achieved through the calculation of site-wise log-likelihood support (SSLS) and site-wise likelihood values (ΔSSLS = lnL (H0) - lnL (HA)), which involved comparing two hypotheses (the null model, H0, representing the species tree; the alternative model, HA, representing the constraint of monophyly of concerned species; Figure S1). A negative ΔSSLS value indicates parallel substitution, while a positive value supports the species phylogeny. The sites with parallel substitutions were ultimately selected based on the following criteria: (i) the amino acid residues of both pangolins were identical; (ii) the site was substituted with the same amino acid in more than 80% of the five concerned species; (iii) more than 80% of other species exhibited different amino acids; (iv) the ΔSSLS value was negative. Genes with the parallel substituted amino acid sites were screened to indicate the convergent evolution among two pangolins, aardvark, Cape golden mole, and short-beaked echidna (CEGs). Functional enrichment analysis To clarify the functions associated with the evolution and adaptation of pangolins, we used Metascape [28] to perform the functional annotation, and enrichment in Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases for the candidate genes. We obtained the significantly enriched GO terms and KEGG pathways for the genes exhibiting adaptive and evolutionary signatures (PSGs, REGs, CEGs, ISGs, and RSGs), as well as those within expanded gene families. Identification of intact sensory-related genes and pseudogenes For the key sensory-related genes, the protein-coding sequences of H. sapiens and M. musculus were downloaded from the NCBI database and used as query sequences. These query sequences were then aligned to each genome using BLAST (v2.2.29+) with an e-value threshold of 1E-05. Solar (v0.9.6) was used to combine those blast-hits belonging to the same query proteins into one predicted gene, and the combined hits were extracted from the genome [29]. Additionally, GeneWise (v2.4.1) [30] was utilized to predict gene structures and obtain newly annotated gene models. The functions and domains of the sequences were predicted using the NR and Pfam databases [31]. If a newly annotated gene shares the same function and domain sequences as those in the annotation file released with the genome, we defined these genes as intact genes. Finally, we selected sensory-related genes from these genes as candidate genes. To identify loss-of-function sensory-related genes, we focused on pseudogenes, which were with loss-of-function features, while incomplete genes (truncated or edge genes) were out of our consideration [32]. The genome-wide identification and classification of pseudogenes was conducted using a combination of criteria, including homology, intron-exon structure, and the presence of stop codons and frameshifts, by PseudoPipe (v2.0) for the genomes of two pangolins [33]. The following cutoffs were applied to determine the final candidate pseudogenes: (i) amino acid sequence identity greater than 40%; (ii) BLAST e-value less than 1E-10; and (iii) the pseudogene must cover more than 70% of the parent gene. The high-confidence pseudogenes were classified as PSSD (processed pseudogene or retropseudogene) and DUP (duplicated pseudogene) [34, 35]. Assessment of the duplication and loss of olfactory receptor genes (ORs) in pangolins We screened for intact OR genes and OR pseudogenes based on the methods from previous studies in 14 mammals [11, 36]. For the identification of intact ORs, it is additional to note that the candidates were restricted to those with seven alpha-helical transmembrane (TM) regions, as OR genes are members of the G-protein coupled receptor superfamily [37]. To estimate gene duplication and loss, the reconstructed tree of the ORs was compared with the species tree by lineage using Notung (v2.9) with the default parameters [38]. The ORs can be classified into Class I (four subfamilies: subfamily 51, OR51; subfamily 52, OR52; subfamily 55, OR55; subfamily 56, OR56) and Class II (nine subfamilies: subfamily 1/3/7, OR1/3/7; subfamily 2/13, OR2/13; subfamily 4, OR4; subfamily 5/8/9, OR5/8/9; subfamily 6, OR6; subfamily 10, OR10; subfamily 11, OR11; subfamily 12, OR12; subfamily 14, OR14). According to this information, we obtained the subfamilies ORs ascribed. Additionally, we calculated the proportions of intact genes and pseudogenes relative to the total number of ORs were to evaluate the distribution of ORs in two pangolins. Evolutionary analysis of candidate genes To conduct an evolutionary analysis for the pivotal candidate gene sets (certain OR subfamilies, and TAS1Rs and TAS2Rs) from multiple taxa, the maximum likelihood (ML) tree was constructed using IQ-TREE (v2.1.4) with 1,000 bootstrap replicates [39]. Multiple sequence alignment was performed using MAFFT (v7.313) [40]. The phylogenetic tree was visualized and modified with Figtree (v1.4.4) [41] and iTOL (v6, https://itol.embl.de/) [42]. To explore the effects of the mutations on the functions and protein structures of the candidates, the SWISS-MODEL workspace (https://swissmodel.expasy.org) was utilized to build protein structure models using ProMod3 (v3.3.0). The global and per-residue model quality was assessed using the QMEAN scoring function [43–45]. PCR amplification for verification To verify the pivotal insertions of TAS1R3 in pangolins, PCR amplification was performed using tissue samples from 22 individuals of two pangolin species collected at the Guangzhou Wildlife Rescue Center in Guangdong Province, China. These pangolin individuals died of natural causes. Total DNA was extracted using the EasyPure Genomic DNA Kit (Transgen Biotech, Beijing, China). After assessing the quality of the DNA with agarose gel electrophoresis (AGE), the target sequence (~ 1300 bp, encompassing the two insertions of TAS1R3) was amplified using primers (Forward, 5’-gtcagctacggcgccagc-3’; Reverse, 5’-ctgggctgtgcaggtagct-3’) designed by Primer Premier (v5.0) [46]. The polymerase chain reaction (PCR) was performed in a total volume of 25 µl, which included 12.5 µl of 2× High Fidelity PCR Master Mix (BBI, Shanghai, China), 1 µl of genomic DNA, 1 µl of each forward and reverse primer, and 9.5 µl of sterilized ddH2O. The PCR conditions were as follows: pre-denaturation at 94 °C for 5 min; 5 cycles of denaturation at 94 °C for 30 s, annealing at 60–63.2 °C for 30 s, and extension at 72 °C for 1 min; followed by 30 cycles of denaturation at 94 °C for 30 s, annealing at 52 °C for 30 s, and extension at 72 °C for 1 min; with a final extension at 72 °C for 5 min. Sanger sequencing was performed using the Applied Biosystems™ 3730XL (Thermo Fisher Scientific, USA), and the data were analyzed using DNASTAR Lasergene SeqMan Pro (v7.1.0) [47]. Results Expanded gene families related to sensory systems We first constructed a phylogenetic tree based on 7,742 single-copy orthologous genes shared among 15 species (Fig. 1A). We found that both the phylogenetic relationships and divergence times were consistent with the results of previous studies [23, 24], providing a reliable species tree for subsequent analyses. Further, we found 47 expanded gene families within the evolutionary lineage of pangolins (Manis). These gene families were significantly enriched in various sensory-related GO terms, including “olfactory receptor activity” (GO:0004984), “odorant binding” (GO:0005549), “sensory perception of smell” (GO:0007608), and “detection of chemical stimulus involved in sensory perception of smell” (GO:0050911), as well as one KEGG pathway (“Olfactory transduction”, hsa04740) (Table S3 & S4). The results of comparative analysis among five concerned mammals revealed that 124, 152, 178, 200, and 158 gene families were expanded in the genomes of the Chinese pangolin, Malayan pangolin, aardvark, Cape golden mole, and short-beaked echidna, respectively. Three terms, including “G protein-coupled receptor activity” (GO:0004930), “G protein-coupled receptor signaling pathway” (GO:0007186), and “sensory perception of smell” (GO:0007608), were enriched in five species. And three olfactory-related terms (“olfactory receptor activity”, GO:0004984; “detection of chemical stimulus involved in sensory perception of smell”, GO:0050911; “pheromone receptor activity”, GO:0016503) were enriched in all species except for the short-beaked echidna. The enrichment of terms associated with “odorant binding” (GO:0005549) and “response to pheromone” (GO:0019236) were found in both pangolins and aardvark (Fig. 1B). Furthermore, several terms related to temperature adaptation (“temperature homeostasis”, GO:0001659; “regulation of cellular response to heat”, GO:1900034; “cellular response to cold”, GO:0070417), vision (“regulation of rhodopsin mediated signaling pathway”, GO:0022400), and rhythm (“negative regulation of circadian rhythm”, GO:0042754) were exclusively enriched in the Chinese pangolin, and those associated with UV response (“response to UV-B”, GO:0010224; “response to UV-C”, GO:0010225) were uniquely enriched in the Malayan pangolin (Table S5-S9). Accordingly, KEGG enrichment results showed that the pathway associated with “Olfactory transduction” (hsa04740) was significantly enriched in all five species. The enrichment of “Phototransduction” (hsa04744) was found in both the Malayan pangolin and aardvark, and the pathways related to “Thermogenesis” (hsa04714) and “Retinol metabolism” (hsa00830) were exclusively enriched in the Chinese pangolin and Cape golden mole, respectively (Table S10-S14). Fig. 1Genome-wide comparative analysis between pangolins and other mammalian species. (A) Phylogenetic tree and divergence times of 15 species. The red and green numbers above each branch represent the numbers of significantly expanded and contracted gene families, respectively. (B) Gene Ontology (GO) enrichment results of the expanded gene families in five concerned species, including M. pentadactyla (orange), M. javanica (yellow), O. afer (blue), C. asiatica (green), and T. aculeatus (red). GO terms related to sensory systems, commonly and uniquely enriched in five species, are indicated in different colors Genome-wide comparative analysis between pangolins and other mammalian species. (A) Phylogenetic tree and divergence times of 15 species. The red and green numbers above each branch represent the numbers of significantly expanded and contracted gene families, respectively. (B) Gene Ontology (GO) enrichment results of the expanded gene families in five concerned species, including M. pentadactyla (orange), M. javanica (yellow), O. afer (blue), C. asiatica (green), and T. aculeatus (red). GO terms related to sensory systems, commonly and uniquely enriched in five species, are indicated in different colors Duplication and loss of olfactory receptor (OR) genes in pangolins In total, we detected 10,356 intact OR genes from the genomes of 14 mammals. In pangolins, we detected 769 intact OR genes in the Chinese pangolin (Class I, 133; Class II, 656) and 733 intact OR genes in the Malayan pangolin (Class I, 123; Class II, 610), distributed across 17 OR subfamilies (excluding OR55) (Table S15). The total number of intact OR genes varied among species, particularly for Class II genes (Fig. 2A). In Class II, a larger number of OR6 and OR14 members were identified in both two pangolins compared to other species that diverged from the most recent common ancestor, manifested as the large number of copies of OR6C2 in OR6 (Chinese pangolin, 104 copies; Malayan pangolin, 58 copies; giant panda, 4 copies; dog, 2 copies; domestic cat, 7 copies), and OR14A2 (Chinese pangolin, 6 copies; Malayan pangolin, 8 copies; giant panda, 2 copies; dog, no functional copy; domestic cat, single copy), OR14C36 (Chinese pangolin, 17 copies; Malayan pangolin, 16 copies; giant panda and dog, single copy; domestic cat, no functional copy), and OR14L1 (Chinese pangolin, 7 copies; Malayan pangolin, 8 copies; giant panda, dog, and domestic cat, no functional copy) in OR14 (Figure S2). Moreover, a total of 704 (Class I, 18; Class II, 686) and 475 (Class I, 31; Class II, 444) OR pseudogenes were identified across 16 OR subfamilies (excluding OR3 and OR55) in the Chinese pangolin (681 PSSD, 23 DUP) and the Malayan pangolin (458 PSSD, 17 DUP), respectively (Table S16). We did not find any pseudogenes for OR55 or OR3 in either pangolin. Notably, the highest number of pseudogenes was found in OR14 in both pangolin genomes (Fig. 2A). Accordingly, we observed a greater number of pseudogenes in the two pangolin genomes compared to those found in the aardvark and Cape golden mole, with much more pseudogenes of OR6 and OR11 detected in the Chinese pangolin than Malayan pangolin. The loss and duplication analyses of ORs revealed 292 duplications and 1,210 losses in the lineages of the common ancestor of two pangolins, additionally, more duplications and losses were found in pangolins than mammals of the order Carnivore (Figure S3). Furthermore, we assessed the proportions of intact genes and pseudogenes relative to the total number of ORs in two pangolins (Chinese pangolin, 1,473; Malayan pangolin, 1,208). In the Chinese pangolin, we detected a higher proportion of Class II ORs (91.1%, with intact genes at 44.5% and pseudogenes at 46.6%) compared to Class I ORs (8.9%, with intact genes at 7.7% and pseudogenes at 1.2%). Similarly, in the Malayan pangolin, the proportion of Class I ORs (12.8%, with intact genes at 10.2% and pseudogenes at 2.6%) was considerably lower than that of Class II ORs (87.2%, with intact genes at 50.5% and pseudogenes at 36.7%) (Fig. 2B). The distribution of intact genes across different OR families varied, with no intact genes found in OR55 for both two pangolins. The intact ORs accounted for more than 80% of two subfamilies (OR51 and OR52) in the Chinese pangolin, and three subfamilies (OR51, OR56, and OR12) in the Malayan pangolin (Table S17). The number of intact members of OR5/8/9 in the genomes of two pangolins was much higher than the other three concerned mammals, and the Chinese pangolin exhibited a greater number of intact OR6 members compared to the others (Fig. 2C). Fig. 2Evolution of the olfactory receptor (OR) gene family in 14 mammals. (A) Distribution of intact OR genes and OR pseudogenes in 14 mammals. The screened ORs include members from four subfamilies (OR51, OR52, OR55, and OR56) in Class I, and nine subfamilies (OR1/3/7, OR2/13, OR4, OR5/8/9, OR6, OR10, OR11, OR12, and OR14) in Class II. Different colors represent various OR subfamilies (Class I in blue; Class II ranging from green to brown). (B) Comparison of the proportions of intact OR genes and OR pseudogenes relative to the total number of OR genes between M. pentadactyla and M. javanica. (C) Heat map illustrating the number of intact OR genes in each subfamily in five concerned mammals (two pangolins, O. afer, C. asiatica, and T. aculeatus) Evolution of the olfactory receptor (OR) gene family in 14 mammals. (A) Distribution of intact OR genes and OR pseudogenes in 14 mammals. The screened ORs include members from four subfamilies (OR51, OR52, OR55, and OR56) in Class I, and nine subfamilies (OR1/3/7, OR2/13, OR4, OR5/8/9, OR6, OR10, OR11, OR12, and OR14) in Class II. Different colors represent various OR subfamilies (Class I in blue; Class II ranging from green to brown). (B) Comparison of the proportions of intact OR genes and OR pseudogenes relative to the total number of OR genes between M. pentadactyla and M. javanica. (C) Heat map illustrating the number of intact OR genes in each subfamily in five concerned mammals (two pangolins, O. afer, C. asiatica, and T. aculeatus) Adaptive and convergent evolution associated with sensation In total, we identified 1,523, 1,887, 2,732, and 1,110 genes exhibiting positive selection, intensified selection, relaxed selection, and rapid evolution in the pangolin lineage, respectively. Additionally, the comparative genomic analysis involving pangolins, aardvark, Cape golden mole, and short-beaked echidnas revealed a total of 697 parallel substitutions corresponding to 630 genes (CEGs). Notably, 7.9% of these genes (50 out of 630) contained more than one site with parallel substitution. Enrichment analysis showed that several GO terms associated with visual perception and cilium were significantly enriched in PSGs, ISGs, CEGs, and REGs (Fig. 3A & B and Tables S18-S21). One GO term related to “cellular response to UV” (GO:0034644) was also enriched in CEGs (Fig. 3A). The GO terms associated with rhythm (“circadian rhythm”, GO:0007623; “rhythmic process”, GO:0048511) and sound perception (“sensory perception of sound”, GO:0007605) were significantly enriched in both ISGs and REGs (Fig. 3B). Furthermore, terms related to ear development and morphogenesis (“inner ear development”, GO:0048839; “inner ear morphogenesis”, GO:0042472; “inner ear auditory receptor cell differentiation”, GO:0042491; “middle ear morphogenesis”, GO:0042474; “outer ear morphogenesis”, GO:0042473; “cartilage morphogenesis”, GO:0060536), and fear response (“behavioral fear response”, GO:0001662) were also enriched in ISGs (Table S19). And GO terms associated with mechanical stimulus response (“cellular response to mechanical stimulus”, GO:0071260), and GO terms and KEGG pathways related to temperature response and regulation (“positive regulation of cold-induced thermogenesis”, GO:0120162; “cellular response to cold”, GO:0070417; “negative regulation of cold-induced thermogenesis”, GO:0120163; “regulation of cellular response to heat”, GO:1900034; “heat shock protein binding”, GO:0031072) were enriched in RSGs (Table S22 & S23). However, we did not observe significant enrichment related to sound perception or rhythm in PSGs or REGs for the aardvark and Cape golden mole (Table S24 & S25), or short-beaked echidna (Table S26 & S27). Simultaneously, several genes with evolutionary signatures and involved in sensory perception gained our attention, including those involved in vision (5 PSGs & CEGs, 22 PSGs, and 4 CEGs), gustation (3 PSGs and 4 RSGs), olfaction (9 PSGs, 4 REGs, and 2 RSGs), audition (11 PSGs, 12 ISGs, and 2 REGs), and circadian behavior and rhythm (4 PSGs, 4 ISGs, 3 REGs, and 7 RSGs) (Fig. 3C). Among the five vision-related genes subject to both positive selection and convergent evolution, two genes (OPN4 and ATXN7) had more than one parallel substituted site, specifically two (739R and 750 N) for OPN4 and three (343 A, 874 L, and 986 S) for ATXN7 (Table S28). We visualized the amino acid sites with evolutionary signatures for OPN4, which encodes a photoreceptive opsin protein expressed within the ganglion and amacrine cell layers of the retina, across fourteen mammalian species (Fig. 3D). Fig. 3Genomic signatures of sensory evolution in pangolins. (A) Significantly enriched sensory-related Gene Ontology (GO) terms for positively selected genes (PSGs) and convergent evolutionary genes (CEGs). (B) Bubble diagram illustrating GO enrichment results related to sensory systems for genes evolved under intensified selection (ISGs) and rapid evolution (REGs). (C) Sensory-related genes with evolutionary signatures (positive, intensified, and relaxed selection, as well as rapid evolution) in pangolins, along with parallel substituted features among pangolins, aardvark, Cape golden mole, and short-beaked echidna. A Venn diagram displays the sensory-related PSGs, CEGs, and the intersection of PSGs and CEGs (PSGs & CEGs). PSGs, ISGs, RSGs, and REGs are highlighted in brown, red, green, and purple, respectively. (D) Partial distribution of amino acid sites demonstrating adaptation and parallel substitutions of OPN4, which is associated with visual perception, in 14 mammals. A total of 20 amino acid sites with signatures of positive selection (solid line frame) or parallel substitution (dotted line frame) are explicitly shown in five mammals, including the Chinese pangolin (M. pentadactyla), Malayan pangolin (M. javanica), aardvark (O. afer), Cape golden mole (C. Asiatica), and short-beaked echidna (T. aculeatus). Green and purple markers indicate the sites under positive selection in the lineage of pangolins (Manis) and the Malayan pangolin (Manis javanica), respectively Genomic signatures of sensory evolution in pangolins. (A) Significantly enriched sensory-related Gene Ontology (GO) terms for positively selected genes (PSGs) and convergent evolutionary genes (CEGs). (B) Bubble diagram illustrating GO enrichment results related to sensory systems for genes evolved under intensified selection (ISGs) and rapid evolution (REGs). (C) Sensory-related genes with evolutionary signatures (positive, intensified, and relaxed selection, as well as rapid evolution) in pangolins, along with parallel substituted features among pangolins, aardvark, Cape golden mole, and short-beaked echidna. A Venn diagram displays the sensory-related PSGs, CEGs, and the intersection of PSGs and CEGs (PSGs & CEGs). PSGs, ISGs, RSGs, and REGs are highlighted in brown, red, green, and purple, respectively. (D) Partial distribution of amino acid sites demonstrating adaptation and parallel substitutions of OPN4, which is associated with visual perception, in 14 mammals. A total of 20 amino acid sites with signatures of positive selection (solid line frame) or parallel substitution (dotted line frame) are explicitly shown in five mammals, including the Chinese pangolin (M. pentadactyla), Malayan pangolin (M. javanica), aardvark (O. afer), Cape golden mole (C. Asiatica), and short-beaked echidna (T. aculeatus). Green and purple markers indicate the sites under positive selection in the lineage of pangolins (Manis) and the Malayan pangolin (Manis javanica), respectively Taste adaptation and evolution in pangolins According to the identification of receptor genes associated with salty (SCNN1A, SCNN1B, and SCNN1C), sour (PKD1L3, PKD2L1, and OTOP1), sweet and umami (TAS1Rs), and bitter (TAS2Rs) tastes, intact genes for SCNN1A, SCNN1B, SCNN1C, OTOP1, TAS1R1, TAS1R3, and TAS2Rs were identified in all 14 mammals (Table S29). We did not find intact genes for PKD1L3, PKD2L1, or TAS1R2 in either pangolin genomes. Intact PKD1L3 was absent in the aardvark, Cape golden mole, and short-beaked echidna, while PKD2L1 was not found in the Cape golden mole. Furthermore, a total of 43 intact TAS1Rs and 276 intact TAS2Rs were obtained (Table S29 & S30). In each pangolin genome, six intact single-copy TAS2Rs, including TAS2R1, TAS2R4, TAS2R7, TAS2R38, TAS2R40, and TAS2R46, were identified (Table S30). Moreover, three pseudogenes in TAS2Rs (TAS2R7 and two copies of TAS2R40) were detected in the Chinese pangolin, and one (TAS2R40) was found in the Malayan pangolin. The phylogenetic tree of intact taste receptor genes (TAS1Rs and TAS2Rs) from 14 mammals was reconstructed. Interestingly, TAS2R4, TAS2R38, TAS2R40, and TAS2R46 in pangolins displayed a separate phylogenetic clade from those in other mammals (Fig. 4B). TAS1R3, the gene responsible for sweet and umami taste receptors, exhibited relaxed selection in pangolins. Based on the preliminary results from intact gene identification, we found two long-fragment insertions (Insertion_1, 44 amino acids, positions 429–473; Insertion_2, 47 amino acids, positions 541–587) in pangolins compared to other mammals (Table S31). The 3D-model (GMQE = 0.83) indicated that two insertions were located in the N-terminal ectodomain (Fig. 4C). This was confirmed by the partial sequence of TAS1R3 (~ 1300 bp), which encompassed the two insertions, from 22 pangolin individuals using PCR amplification (Table S32). According to the alignment results for all sequences, there were five and six amino acid substitution sites between two pangolins at Insertion_1 and Insertion_2, respectively (Figure S4). Fig. 4Evolution of taste-related genes in 14 mammals. (A) The number of intact bitter receptor genes (TAS2Rs) in 14 mammalian species. (B) The phylogenetic tree of the taste receptor gene family (TAS1Rs and TAS2Rs, including 319 genes). The light green and pink shadows represent TAS1Rs and TAS2Rs, respectively. The branches of TAS in pangolins (TAS1R1, TAS1R3, TAS2R1, TAS2R4, TAS2R7, TAS2R38, TAS2R40, and TAS2R46) are highlighted in deep pink. The arc markers indicate that the clade encompasses all members of the genes in 14 species. (C) A 3D-model of TAS1R3 featuring two long-fragment insertions (Insertion_1, 44 amino acids, Positions 429–473; Insertion_2, 47 amino acids, Positions 541–587) in pangolins, generated using SWISS-MODEL (GMQE = 0.83) Evolution of taste-related genes in 14 mammals. (A) The number of intact bitter receptor genes (TAS2Rs) in 14 mammalian species. (B) The phylogenetic tree of the taste receptor gene family (TAS1Rs and TAS2Rs, including 319 genes). The light green and pink shadows represent TAS1Rs and TAS2Rs, respectively. The branches of TAS in pangolins (TAS1R1, TAS1R3, TAS2R1, TAS2R4, TAS2R7, TAS2R38, TAS2R40, and TAS2R46) are highlighted in deep pink. The arc markers indicate that the clade encompasses all members of the genes in 14 species. (C) A 3D-model of TAS1R3 featuring two long-fragment insertions (Insertion_1, 44 amino acids, Positions 429–473; Insertion_2, 47 amino acids, Positions 541–587) in pangolins, generated using SWISS-MODEL (GMQE = 0.83) Discussion Significant expansion and rapid evolution of ORs in pangolins As one of the most ancient senses, mammals widely use olfaction to monitor the environment, allowing recognition, predation, communication, and reproduction [48]. It is commonly believed that pangolins have a superior sense of olfaction [49]. In this study, the first observation is the expansion of the OR gene family in pangolins, which is consistent with the previous conclusions and confirms that there might be more reliance on the sense of smell compared to carnivorous animals, by the genome-wide gene family analysis [5]. We found the convergent expansion associated with the perception of chemical stimulus and pheromone among pangolins and the other three concerned species (Fig. 1B). These results suggested that pangolins’ chemical communication, which relied mainly on scent, might play a crucial role in their various daily activities, as they possess large anal glands that emit odorous secretions [50]. Previous studies supported that olfaction exhibits rapid evolution among vertebrates, and the varied abilities of olfaction among mammals might be attributed to the birth and death of the OR gene family driven by niche-specific adaptation [51–53]. In previous studies, the functional and nonfunctional ORs have been screened using the scaffold-level pangolin genomes and used to look at the diversity and evolution of the vertebrate chemoreceptor gene repertoire [54, 55]. In these cases, pangolins were just one of the components of the mammals, which were in comparison with other taxa (such as rodents, amphibians, birds, and so on). In our study, to focus on the evolutionary features of pangolin genomes, we further identified the intact and pseudogenetic members of OR subfamilies in detail using the higher-quality genomes and performed the comparison among mammals. The reason for excluding incomplete genes is to ensure true positive results. Although a large number of ORs have lost their functions, there were still more intact ORs in the genomes of pangolins, especially manifested as the expansion of OR6C2 in OR6 and three genes (OR14A2, OR14C36, and OR14L1) in OR14, in the comparison with those of carnivorous animals (Fig. 2). Furthermore, both the exceptional gene turnover through duplication and loss of functional copies of receptor genes and rapid rates of molecular evolution can lay the groundwork for rapid adaptation [56, 57]. We also found the signatures of positive selection (OR5I1 and OR6K2) and rapid evolution (OR5AS1, OR6N1, and OR6N2) on certain ORs in pangolins (Fig. 3C). These above have given the adaptive basis to illustrate the perception abilities of olfaction for pangolins. Loss and relaxed selection of intact taste receptor genes in pangolins As one of the representative myrmecophagous mammals, pangolin evolved strong forelimbs with long claws to excavate nests, tubular mouth, and slender tongue to capture and swallow heterodyne, and distinctive stomach structures and decomposition abilities to digest large numbers of ants or termites with chitin exoskeletons and toxic substance, along the long-term dietary specialization [58, 59]. However, the genetic mechanisms of taste perception could be fascinating due to the eternal exposure to the stimulation of acids according to the specialized myrmecophagy for pangolins, and that remains uncovered [8, 59]. Our identification results showed the universal loss of intact receptor genes responsible for the gustatory response to acids (PKD1L3 and PKD2L1) in pangolins, and this was similar to the other three myrmecophagous species. These indicated the shutdown of the acid perception pathway for pangolins, suggesting that the ability to sustain the prolonged acid-stimulated behavior could be associated with adaptive neglect. It has previously been proposed that TAS1Rs and TAS2Rs might evolve in response to diet preferences [60]. It is worth noting that carnivorous mammals are more prone to lose TAS1R2 compared to herbivores and omnivores [55]. It has been reported that herbivores have more TAS2Rs than carnivores to detect toxic compounds in plants, and insects can acquire toxicity directly from plants through diet [61, 62]. In the study of 41 Laurasiatherian species, their results also showed a lower number and diversity of TAS2Rs in Chinese pangolin than other entomophagous taxa, such as common shrew and insectivorous bats [63]. We found the loss of TAS1R2, and an extremely low number and diversity of TAS2Rs in pangolin genomes, however, not in the other three ant-eating species. Our results were consistent with the previous findings and indicated the interesting evolution of TAS1R2 and TAS2Rs in pangolins. Nevertheless, it still needs further investigation whether the lacking in pangolins is shaped by the specialized diet. Gene TAS1R3, which is found under relaxed selection in pangolins (Fig. 4C), is the main component of the heterodimers responding to amino acids (TAS1R1/1R3) and sweet-tasting substances (TAS1R2/1R3), and providing the umami and sweet tastes [64, 65]. As a member of family C in G protein-coupled receptor superfamily, TAS1R3 is characterized by a large N-terminal ectodomain as long as ~ 560 amino acids, which can be implicated in ligand binding and dimerization [66, 67]. We detected that two long-fragment insertions of the amino acid sequence were the unique features of TAS1R3 in pangolins (Figure S4). That was estimated to probably cause the structural variation of the N-terminal domain, which could potentially affect its function [68]. Unexpectedly, we did not detect convergent evolutionary signals for either of the taste candidate genes among five myrmecophagous mammals. This indicated the diversity of taste perception among them, even though their diet phenotypes were convergent. In the future, it is necessary to conduct behavioral tests to perform the functional verification of taste perception for pangolins. Other sensations in pangolins Existing research focused mainly on the olfaction of pangolins, and other sensory modalities were largely neglected. As for the vision, in addition to visual perception, the significant enrichment signals associated with cilium were also detected in the genes under selection (PSGs, ISGs, and CEGs) in the pangolin branch (Fig. 3A&B), and photoreceptors are the modified cilium containing phototransduction component [69]. Moreover, we detected the intensified selection and rapid evolution associated with circadian behavior and rhythm, the positive selection and convergent evolution on gene OPN4, which is the candidate for photoreceptor-mediated circadian entrainment [70, 71]. Combined with the relaxed selection on gene MTNR1A, which is responsible for the melatonin effects on mammalian circadian rhythm and reproductive alterations affected by day length [72], these might indicate the adaptation related to the light-mediated but not physiological rhythm regulation in pangolins (Fig. 3C&D). It is widely known that the circadian rhythm affects the cyclical life activities of animals, such as diet, sleep, exercise, and reproduction. As a typical nocturnal and burrowing animal, the potential effects of rhythm adaptation on the life activities of pangolins still need to be further explored and interpreted. Furthermore, the intensified selection pressure associated with the sensory perception of sound, ear morphogenesis and development, and auditory receptor cell differentiation in pangolins indicated the potentially adaptive features related to auditory perception. Unlike olfaction and vision which have been followed with interest for pangolins, their auditory sensation is still mysterious, and our results provided the preliminary genetic basis for an in-depth investigation of their hearing sense. In addition to the senses mentioned above (vision, audition, olfaction, and gustation), animals also have other sensory systems, such as vestibular sense, proprioception, kinesthesia, nociception, and thermoception, providing information about balance, body position and movement, pain and defense, and temperature and humidity [73]. In the pangolin branch, we found the significant relaxation of evolutionary pressure associated with response to temperature and the intensification of selected constraints involved in behavioral fear response, based on the signatures of relaxed and intensified selection of related gene sets. These indicated an attractive clue to further explore the adaptive evolution of nociception and thermoreception, which might give dramatic contributions for the comprehension of their living habits, for pangolins. In a nutshell, this study explored the potential signatures of sensory adaptation and evolution in pangolins by comparative genomic analysis, and the correlation between the genomic characters and sensory functions might require further functional verification. Furthermore, it must be noted that the assembly quality can have a critical impact on the results of comprehensive multi-species comparative analyses [74]. In this study, except for the genome of platypus (92.6%), the BUSCO completeness values of the other 14 genomes were high (94.0-98.4%), indicating the sequencing and assembly procedure had successfully constructed a reliable representation of the full set of genes in these genomes [75]. This suggested the reliability of our results. Of course, more accurate genomic evolutionary characteristics can result from higher quality assembly resources of genomes. Conclusion According to the comparison with other mammal species, a series of new genomic features associated with the adaptation and evolution of the sensory system in pangolins were found in this study. Our results showed the significant expansion and rapid evolution of ORs in pangolins, and the convergence with the other three concerned mammals was detected, providing further evidence for understanding their developed olfaction. Furthermore, we revealed the loss of intact sour taste receptor genes (PKD1L3 and PKD2L1) and the evolutionary characteristics of TAS1Rs and TAS2Rs in pangolins, including the lack of TAS1R2, two long-fragment insertions in TAS1R3, and low quantity and diversity of TAS2Rs. There was no evidence of the taste-related convergent evolutionary signal among five myrmecophagous mammals, suggesting a unique evolution of taste in pangolins. The positive adaptation related to light and sound perception and the relaxed evolutionary constraints associated with mechanical stimulus and temperature response were also detected in pangolins. These results will help to better understand the interaction between pangolins and the environment, which may future provide the genetic basis across the genome for the conservation of pangolins. Electronic supplementary material Below is the link to the electronic supplementary material. Supplementary Material 1: Tables S1-S32 Supplementary Material 1: Tables S1-S32 Supplementary Material 2: Figures S1-S4 Supplementary Material 2: Figures S1-S4