Evolution of Acidic Mammalian Chitinase Genes (CHIA) Is Related to Body Mass and Insectivory in Primates

Evolution of Acidic Mammalian Chitinase Genes (CHIA) Is Related to Body Mass and Insectivory in... Abstract Insects are an important food resource for many primates, but the chitinous exoskeletons of arthropods have long been considered to be indigestible by the digestive enzymes of most mammals. However, recently mice and insectivorous bats were found to produce the enzyme acidic mammalian chitinase (AMCase) to digest insect exoskeletons. Here, we report on the gene CHIA and its paralogs, which encode AMCase, in a comparative sample of nonhuman primates. Our results show that early primates likely had three CHIA genes, suggesting that insects were an important component of the ancestral primate diet. With some exceptions, most extant primate species retain only one functional CHIA paralog. The exceptions include two colobine species, in which all CHIA genes have premature stop codons, and several New World monkey species that retain two functional genes. The most insectivorous species in our sample also have the largest number of functional CHIA genes. Tupaia chinensis and Otolemur garnettii retain three functional CHIA paralogs, whereas Tarsius syrichta has a total of five, two of which may be duplications specific to the tarsier lineage. Selection analyses indicate that CHIA genes are under more intense selection in species with higher insect consumption, as well as in smaller-bodied species (<500 g), providing molecular support for Kay’s Threshold, a well-established component of primatological theory which proposes that only small primates can be primarily insectivorous. These findings suggest that primates, like mice and insectivorous bats, may use the enzyme AMCase to digest the chitin in insect exoskeletons, providing potentially significant nutritional benefits. digestive enzymes, acidic mammalian chitinase, insectivory, dietary adaptations Introduction All primates include some insects in their diet, whether through accidental consumption or through active insectivory (Raubenheimer and Rothman 2013). Degree of insectivory in living primates ranges from nearly exclusive (e.g., Tarsius spp.) to complementary (e.g., the Callitrichidae) to supplemental, such as in the great apes (McGrew 2001). Although insects are a significant source of energy and protein for many living primates, this is especially true for small-bodied primates (Kay 1984). Due to their sparse distribution in most environments, insects are usually energetically costly to find and catch, making it difficult for a large-bodied primate to fill their nutritional demands solely with insects (Raubenheimer and Rothman 2013). Small-bodied primates have relatively higher metabolic requirements per unit body mass, but are small enough that the insects they catch suffice to meet their nutritional needs (Fleagle 2013). Kay (1984) calculated that a primate above 500 g will simply not be able to catch enough insects in a day to satisfy its daily energy requirements (Kay 1984; Kay and Covert 1984). Following from this is a classic concept in primatology, Kay’s Threshold (Kay 1984), which predicts that only species below 500 g will be insectivores, whereas only species above this weight will be folivorous (fig. 1). Among frugivorous primates, those that are smaller (≤1 kg) will typically rely on insects as their source of protein, whereas those that are larger will use leaves (Gingerich 1980; Kay 1984; Fleagle 2013) (fig. 1). Social insects, including ants and termites, represent an exception to this rule because they occur as clumped resources across time and space and can be efficiently preyed upon by larger primates (Isbell 1998). In the case of termites, extractive foraging tools are often used (Goodall 1986; McGrew 1992; van Schaik 2003; Souto et al. 2011). Although the nutrient composition of insects varies widely, geometric analyses show that insects eaten by nonhuman primates tend to have high protein-to-fat ratios and are important sources of minerals that may not otherwise be included in the diet (Raubenheimer and Rothman 2013), making them a valuable resource for extant primates and a possible driving force in primate evolution. Fig. 1. View largeDownload slide Correlation between primate diets and body size. Figure concept by Kay (1984) and modified in Fleagle (2013), reproduced with permission. Fig. 1. View largeDownload slide Correlation between primate diets and body size. Figure concept by Kay (1984) and modified in Fleagle (2013), reproduced with permission. The Visual Predation Hypothesis posits that insect predation was the adaptive pressure leading to the evolution of the primate visual system and other morphological features (Cartmill 1972, 1992, 2012). A suite of adaptations is associated with insectivorous primates, including molars with crests that are used to masticate insect exoskeletons (Kay 1975), simple guts with low stomach-to-small intestine ratios (Chivers and Hladik 1980), relatively larger home ranges (Clutton Brock and Harvey 1980) and small body size (Kay 1984). The primate visual system and grasping hands have also been suggested as adaptations for preying on insects (Bishop 1964; Cartmill 1972, 1992, 2012). Despite the high nutritional value of insects, there are drawbacks to consuming them. One such drawback is that they are often protected by exoskeletons, which are made up of the structural carbohydrate chitin (Finke 2007). Chitin makes up between 2 and 20% of an insect’s drymatter and is considered to be indigestible by most primates, unless their digestive systems contain chitinolytic enzymes (Rothman et al. 2014). Given the digestive challenges posed by chitinous exoskeletons (Strait and Vincent 1998; Rothman et al. 2014), paired with potentially significant energy and amino acid returns if they are digested (Finke 2007; Rothman et al. 2014), endogenously producing such a chitinolytic enzyme could have important adaptive benefits for insectivorous primates, complementing the dental, behavioral, and morphological adaptations discussed above. Indeed, mice and insectivorous bats have been shown to digest chitin using an enzyme produced in the stomach called acidic mammalian chitinase (AMCase) (Whitaker et al. 2004; Strobel et al. 2013; Ohno et al. 2016). This chitinolytic digestive enzyme is produced in the gastric chief cells, where other digestive enzymes are also secreted (Strobel et al. 2013; Ohno et al. 2016). Studies in mice further showed that AMCase is resistant to degradation by the proteases found in the stomach, such as pepsin C, trypsin and chymotrypsin, and breaks down chitin in the presence of these enzymes, as well as at an acidic pH (Ohno et al. 2016). Even though chitinolytic activity has also been observed in the gastric juices of two primates (Perodicticus potto and Cebus capucinus) (Cornelius et al. 1976; Jeuniaux and Cornelius 1978), it was long believed that primates (and most other mammals) did not produce such an enzyme and could not digest chitin (Cork and Kenagy 1989; Oftedal et al. 1991; Simunek and Bartonova 2005; Strobel et al. 2013; Ohno et al. 2016). Instead, it was thought that insect-eating primates had fast gut-transit times to quickly pass indigestible exoskeletons (Gaulin 1979; Milton 1984). More recently, one study found that some primates harbor chitin-digesting microbes (Macdonald et al. 2014) and another study identified an acidic mammalian chitinase in macaques (Macaca fascicularis) that is expressed in the stomach and effectively digests chitin at an acidic pH (Krykbaev et al. 2010). However, because macaques are not very insectivorous and in humans AMCase has been associated with type-2 immune response, such as asthma, allergies, eye diseases, and parasite defense (Zhu et al. 2004; Reese et al. 2007; Musumeci et al. 2009; Bucolo et al. 2011; Muzzarelli et al. 2012; Vannella et al. 2016), any potential benefit of AMCase for insectivorous primates remains unresolved. AMCase is encoded by the gene CHIA or one of its paralogs. In primates, two functional CHIA genes have been identified (Krykbaev et al. 2010): hCHIA and mCHIA. Although hCHIA remains functional in humans, mCHIA has a premature stop codon. In macaques, this state is reversed and mCHIA is functional, whereas hCHIA has a premature stop codon (Krykbaev et al. 2010). Here, we present data on the acidic mammalian chitinase gene (CHIA) in a large comparative sample of nonhuman primates (n = 34) and one treeshrew (Tupaia chinensis). Primates are an order of mammals with over 230 species that have a range of different dietary ecologies (Groves 2001). Across primates, insect consumption varies from practically 0% (e.g., colobine monkeys, which are limited to ingesting insects incidentally when eating leaves or fruit) to almost 100% (Tarsius spp. are completely faunivorous, with most of their diet consisting of arthropods), making them an ideal group for a comparative study of dietary adaptations associated with insectivory. We investigate the potential of AMCase as a digestive enzyme adaptation for insectivory by comparing the paralogous CHIA gene sequences across primates with different levels of insectivory. We also test the strength of selection on CHIA as a function of insect consumption. Finally, we further explicate the relationship between insectivory and body size, as proposed by Kay’s threshold (Kay 1984). Results We successfully sequenced two CHIA genes in 12 primate species for which whole-genome sequences are not available. The sequences we generated for Callithrix jacchus contained numerous differences to the reference genome sequence. Since our sequences were more parsimonious in the comparative context of our study, both compared with sequences generated by us and the whole-genome sequences of closely related species, we believe that our sequences are most likely accurate. All sequences have been deposited in GenBank (accession numbers in supplementary table 1, Supplementary Material online). We further found both homologous sequences (mCHIA and hCHIA) in the 23 genomes we surveyed, with the exception of Microcebus murinus in which we could only positively identify one complete CHIA sequence (hCHIA). In the tarsier (Tarsius syrichta), galago (Otolemur garnettii), and Chinese treeshrew (Tupaia chinensis) genomes we identified >2 CHIA sequences. Both the galago and the Chinese treeshrew genomes had a third CHIA gene sequence (CHIA3), whereas the tarsier genome included three additional CHIA sequences (CHIA3, CHIA4, CHIA5) for a total of five homologous genes. Phylogenetic trees generated from an alignment of the coding sequences (fig. 2) show a deep split between the CHIA genes, indicating that mCHIA and hCHIA (and likely CHIA3) arose in a duplication event that was ancestral to primates and treeshrews and are not independently duplicated genes. Although we did not find complete CHIA3 sequences in any primates other than the tarsier and galago, partial sequences were identified in some genomes. Fig. 2. View largeDownload slide Evolutionary relationships of the CHIA genes in primates. The tree was inferred with PhyML (Guindon et al. 2010) using the HKY85 nucleotide substitution model. Node labels indicate percent bootstrap support (1000 replicates) and branches are scaled by number of substitutions per site. Tree is rooted at the midpoint. Because this tree is based only on the CHIA loci, not all relationships are resolved in a way that is consistent with primate phylogeny. Notably, the placements of Tupaia chinensis mCHIA and CHIA3, and Old World monkey hCHIA do not reflect the likely organismal relationships among some taxa. Fig. 2. View largeDownload slide Evolutionary relationships of the CHIA genes in primates. The tree was inferred with PhyML (Guindon et al. 2010) using the HKY85 nucleotide substitution model. Node labels indicate percent bootstrap support (1000 replicates) and branches are scaled by number of substitutions per site. Tree is rooted at the midpoint. Because this tree is based only on the CHIA loci, not all relationships are resolved in a way that is consistent with primate phylogeny. Notably, the placements of Tupaia chinensis mCHIA and CHIA3, and Old World monkey hCHIA do not reflect the likely organismal relationships among some taxa. CHIA Pseudogenizations Although all species except the mouse lemur (Microcebus murinus) had two complete CHIA sequences, one of these sequences often contained frameshift-causing indels or nonsense mutations leading to premature stop codons and likely rendering the gene nonfunctional. As a result, most primates only retain one full-length, and likely functional CHIA paralog (fig. 3). With some exceptions, hCHIA remains functional only in apes, whereas only mCHIA is functional in most monkeys. Two species in our data set did not retain any functional CHIA genes. In Rhinopithecus bieti and Nasalis larvatus both the mCHIA and hCHIA sequences contained premature stop codons (figs. 3 and 4a). Several New World primates, on the other hand, retained full-length sequences of both mCHIA and hCHIA, including Callithrix jacchus, Cebus capucinus, Saimiri boliviensis, S. sciureus, and Saguinus fuscicollis (figs. 3 and 4b). All three CHIA sequences in the galago and treeshrew, and all five sequences in the tarsier were free from any indels or premature stop codons (figs. 3 and 4c and d). Fig. 3. View largeDownload slide Evolutionary relationships as inferred from CHIA sequences, including timing of CHIA pseudogenization events. *The pseudogenizing mutation found in the Miopithecus ogouensis hCHIA sequence is the same as the one found in the Papionini (Macaca spp., Papio spp, etc.), but is not found in the other Cercopithecini. It is unclear what accounts for this unexpected pattern, but possible explanations include ancestral polymorphism or ancient hybridization. Fig. 3. View largeDownload slide Evolutionary relationships as inferred from CHIA sequences, including timing of CHIA pseudogenization events. *The pseudogenizing mutation found in the Miopithecus ogouensis hCHIA sequence is the same as the one found in the Papionini (Macaca spp., Papio spp, etc.), but is not found in the other Cercopithecini. It is unclear what accounts for this unexpected pattern, but possible explanations include ancestral polymorphism or ancient hybridization. Fig. 4. View largeDownload slide Most primate species have one full-length CHIA sequence, with some exceptions. (a) In the colobine monkeys Rhinopithecus bieti and Nasalis larvatus both mCHIA and hCHIA sequences have a premature stop codon. (b) In some New World monkeys (in order: Cebus capucinus, Callithrix jacchus, Saguinus fuscicollis, Saimiri sciureus, S. boliviensis) both mCHIA and hCHIA are full-length. (c) The treeshrew (Tupaia chinensis) and the Northern greater galago (Otolemur garnettii), two insectivores, have three full-length CHIA sequences, whereas (d) the tarsier (Tarsius syrichta), the most insectivorous of all primates, has a total of five CHIA genes. Photos by (in order) Israel Didham (with pers. permission), Charles J. Sharp, Steven G. Johnson, Manfred Werner, Marie de Carne, Dave Pape, Julie Langford, JJ Harrison, Milan Kořínek (with pers. permission), and Pierre Fidenci; reproduced with permission via Wikimedia Commons unless otherwise noted. Fig. 4. View largeDownload slide Most primate species have one full-length CHIA sequence, with some exceptions. (a) In the colobine monkeys Rhinopithecus bieti and Nasalis larvatus both mCHIA and hCHIA sequences have a premature stop codon. (b) In some New World monkeys (in order: Cebus capucinus, Callithrix jacchus, Saguinus fuscicollis, Saimiri sciureus, S. boliviensis) both mCHIA and hCHIA are full-length. (c) The treeshrew (Tupaia chinensis) and the Northern greater galago (Otolemur garnettii), two insectivores, have three full-length CHIA sequences, whereas (d) the tarsier (Tarsius syrichta), the most insectivorous of all primates, has a total of five CHIA genes. Photos by (in order) Israel Didham (with pers. permission), Charles J. Sharp, Steven G. Johnson, Manfred Werner, Marie de Carne, Dave Pape, Julie Langford, JJ Harrison, Milan Kořínek (with pers. permission), and Pierre Fidenci; reproduced with permission via Wikimedia Commons unless otherwise noted. Interestingly, across our sample of primate species (n = 34), premature stop codons were independently introduced into the hCHIA or mCHIA sequences numerous times, through frameshift or nonsense mutations at various sites along the sequence. The mCHIA gene lost function independently at least six times in primates: three times in the apes and three times in the colobine monkeys (fig. 3). Premature stop codons in the hCHIA gene arose at least seven times: three times in New World monkeys, twice in colobine monkeys, and two (possibly three) times in cercopithecine monkeys (fig. 3), a subfamily that includes the tribes Cercopithecini and Papionini (table 1). The Papionini (Macaca spp, Mandrillus leucophaeus, Cercocebus atys, and Papio anubis) share a deletion in exon 8 that causes a frameshift and premature stop codon, while most of the Cercopithecini (Cercopithecus mitis, Allenopithecus nigroviridis, Allochrocebus lhoesti, Erythrocebus patas, and Chlorocebus spp.) share a frameshifting deletion and premature stop codon at the beginning of exon 11. Even though Miopithecus ogouensis is considered to be part of the tribe Cercopithecini (Tosi et al. 2002, 2005), this species shares the exon 8 deletion with the Papionini and lacks the exon 11 deletion characteristic of its tribe (fig. 3). Exons 8 and 11 were sequenced again in another lab, using a different Miopithecus sample, confirming these results. At this time, it is unclear what accounts for this pattern. Possible explanations include polymorphism in the common ancestor of the Cercopithecini and Papionini, or ancient hybridization. Table 1. Species Included in This Study with Annual Average Insect Consumption and Average Body Weight of Adult Females in Grams.a Species  Common Name  Average Insect Consumption (%)  Average Body Weight (Adult Female, in g)  Aotus nancymaae  Nancy Ma’s night/owl monkey  <15  780  Callicebus moloch  Red-bellied titi monkey  12  956  Callithrix jacchus  Common marmoset  7.2  381  Saguinus fuscicollis  Saddleback tamarin  28.3  358  Cebus capucinus  White-faced capuchin  31.4  2540  Saimiri boliviensis  Black-capped squirrel monkey  no wild data  711  Saimiri sciureus  Common squirrel monkey  53.4  662  Sapajus apella  Tufted capuchin  32.6  2520  Allenopithecus nigroviridis  Allen’s swamp monkey  9  3180  Allochrocebus lhoesti  L’Hoest’s monkey  8.8  3450  Cercocebus atys  Sooty mangabey  26  6200  Cercopithecus mitis  Blue monkey  17.5  4250  Chlorocebus aethiops  Grivet  15.4  2980  Chlorocebus sabaeus  Green monkey  15.4  3300  Erythrocebus patas  Patas monkey  23.5  6500  Macaca fascicularis  Long-tailed macaque  4.1  3590  Macaca mulatta  Rhesus macaque  0  5370  Macaca nemestrina  Pig-tailed macaque  12.2  6500  Mandrillus leucophaeus  Drill  no wild data  12500  Miopithecus talapoin  Talapoin  35  1120  Papio anubis  Olive baboon  2  13300  Colobus angolensis  Tanzanian black&white colobus  0  6935  Colobus guereza  Guereza  0  8550  Nasalis larvatus  Proboscis monkey  0  9820  Rhinopithecus bieti  Black snub-nosed monkey  0  9960  Rhinopithecus roxellana  Golden snub-nosed monkey  0  11600  Nomascus leucogenys  Northern white-cheeked gibbon  4  7320  Gorilla gorilla gorilla  Western lowland gorilla  7.7  71500  Pan paniscus  Bonobo  2  33200  Pan troglodytes  Common chimpanzee  6.4  41600  Pongo abelii  Sumatran orangutan  11.1  35600  Tarsius syrichta  Philippine tarsier  90  117  Microcebus murinus  Gray mouse lemur  8  63  Otolemur garnettii  Northern greater galago  50  734  Tupaia chinensis  Northern treeshrew  >50  200  Species  Common Name  Average Insect Consumption (%)  Average Body Weight (Adult Female, in g)  Aotus nancymaae  Nancy Ma’s night/owl monkey  <15  780  Callicebus moloch  Red-bellied titi monkey  12  956  Callithrix jacchus  Common marmoset  7.2  381  Saguinus fuscicollis  Saddleback tamarin  28.3  358  Cebus capucinus  White-faced capuchin  31.4  2540  Saimiri boliviensis  Black-capped squirrel monkey  no wild data  711  Saimiri sciureus  Common squirrel monkey  53.4  662  Sapajus apella  Tufted capuchin  32.6  2520  Allenopithecus nigroviridis  Allen’s swamp monkey  9  3180  Allochrocebus lhoesti  L’Hoest’s monkey  8.8  3450  Cercocebus atys  Sooty mangabey  26  6200  Cercopithecus mitis  Blue monkey  17.5  4250  Chlorocebus aethiops  Grivet  15.4  2980  Chlorocebus sabaeus  Green monkey  15.4  3300  Erythrocebus patas  Patas monkey  23.5  6500  Macaca fascicularis  Long-tailed macaque  4.1  3590  Macaca mulatta  Rhesus macaque  0  5370  Macaca nemestrina  Pig-tailed macaque  12.2  6500  Mandrillus leucophaeus  Drill  no wild data  12500  Miopithecus talapoin  Talapoin  35  1120  Papio anubis  Olive baboon  2  13300  Colobus angolensis  Tanzanian black&white colobus  0  6935  Colobus guereza  Guereza  0  8550  Nasalis larvatus  Proboscis monkey  0  9820  Rhinopithecus bieti  Black snub-nosed monkey  0  9960  Rhinopithecus roxellana  Golden snub-nosed monkey  0  11600  Nomascus leucogenys  Northern white-cheeked gibbon  4  7320  Gorilla gorilla gorilla  Western lowland gorilla  7.7  71500  Pan paniscus  Bonobo  2  33200  Pan troglodytes  Common chimpanzee  6.4  41600  Pongo abelii  Sumatran orangutan  11.1  35600  Tarsius syrichta  Philippine tarsier  90  117  Microcebus murinus  Gray mouse lemur  8  63  Otolemur garnettii  Northern greater galago  50  734  Tupaia chinensis  Northern treeshrew  >50  200  a Primate body weight data from (Smith and Jungers 1997); only data from wild primates were used. Tupaia chinensis data from PanTHERIA (Jones et al. 2009). Detailed dietary data and references can be found in supplementary table 3, Supplementary Material online. Signatures of Catalytically Active Chitinases In all of the full-length CHIA amino acid sequences we found that the signatures of catalytically active chitinases were conserved (fig. 5): these have a conserved glutamate and the consensus sequence DXXDXDXE at the active site (Synstad et al. 2004; Krykbaev et al. 2010). In addition, catalytically active chitinases further have a chitin-binding domain at the C-terminus, containing six cysteines, which are essential for attaching the enzyme to the chitin (Tjoelker et al. 2000). All of the full-length CHIA sequences in our study had conserved chitin-binding domains that contained all six cysteines (fig. 5). Fig. 5. View largeDownload slide Partial amino acid sequence alignment of functional CHIA genes. The conserved motif for the chitinase catalytic site (DXXDXDXE) from residue 134 to 141 and the chitin-binding domain from residue 440 to 490 are shown. The cysteines in the chitin-binding domain are highlighted. Fig. 5. View largeDownload slide Partial amino acid sequence alignment of functional CHIA genes. The conserved motif for the chitinase catalytic site (DXXDXDXE) from residue 134 to 141 and the chitin-binding domain from residue 440 to 490 are shown. The cysteines in the chitin-binding domain are highlighted. CODEML Analyses Within the CODEML program in the PAML package (Yang 2007), we used three pairs of site models to test whether any of the sites in mCHIA or hCHIA are subject to positive selection. Results of the CODEML analysis for mCHIA suggested that all sites in this gene are under purifying or neutral selection (table 2). Comparison of the models M0 and M3 falsified the null hypothesis that the same dN/dS ratio (ω) applies to all sites of the mCHIA gene (χ2 = 79.54; df = 4; P < 0.00001); however, this result is expected for most functional proteins. Interestingly, CODEML estimated all three discrete ω groups below 1.00, and it placed 75.5% of sites into categories that have a dN/dS ratio of 0.09. Comparison of models M1a and M2a (χ2 = 0, df = 2, P = 1) and models M7 and M8 (χ2 = 2.85, df = 2, P = 0.24) both failed to support any hypothesis of positive selection. Overall, these models indicate that most mCHIA codons appear to be under purifying selection with a smaller number of sites under neutral selection (16–19%, table 2). Table 2. CodeML Results.a Gene  Model  ln(L)  Parameter Estimates  Test  LR  P Value  Positively Selected Sites  mCHIA  M0  −5377.772  ω = 0.251, k = 3.835          M1a  −5338.362  k = 3.907;          ω0 = 0.114 (81.12%); ω1 = 1.00 (18.87%)  M2a  −5338.362  k = 3.907;  M1–M2  0      ω0 = 0.114 (81.12%); ω1 = 1.00 (13.69%); ω2 = 1.00 (5.19%)  M3  −5338.001  k = 3.861;  M0–M3  79.542  < 0.001    ω0 = 0.1 (77.85%); ω1 = 0.354 (0.0002%); ω2 = 0.866 (22.14%)  M7  −5339.595  k = 3.830, α = 0.348; β = 0.965          M8  −5338.170  k = 3.872, α = 1.341; β = 8.405;  M7–M8  2.851  0.240    p0 = 0.841; ωs = 1.00 (15.89%)  hCHIA  M0  −6138.723  ω = 0.320, k =  4.132          M1a  −6079.211  k = 4.177;          ω0 = 0.140 (77.26%); ω1 = 1.00 (22.74%)  M2a  −6074.863  k = 4.283;  M1–M2  8.696  0.013    ω0 = 0.146 (77.36%); ω1 = 1.00 (21.68%); ω2 = 4.327 (0.96%)  M3  −6074.675  k = 4.219;  M0–M3  128.097  < 0.001    ω0 = 0.00 (31.07%); ω1 = 0.345 (60.07%); ω2 = 1.639 (8.86%)  M7  −6081.171  k = 4.160, α = 0.369; β = 0.758          M8  −6073.789  k = 4.235, α = 0.457; β = 1.025;  M7–M8  14.763  0.001  36P(0.953*), 62Q(0.971*), 164R(0.915), 280H(0.941)  p0 = 0.985; ω = 3.503 (1.52%)  Gene  Model  ln(L)  Parameter Estimates  Test  LR  P Value  Positively Selected Sites  mCHIA  M0  −5377.772  ω = 0.251, k = 3.835          M1a  −5338.362  k = 3.907;          ω0 = 0.114 (81.12%); ω1 = 1.00 (18.87%)  M2a  −5338.362  k = 3.907;  M1–M2  0      ω0 = 0.114 (81.12%); ω1 = 1.00 (13.69%); ω2 = 1.00 (5.19%)  M3  −5338.001  k = 3.861;  M0–M3  79.542  < 0.001    ω0 = 0.1 (77.85%); ω1 = 0.354 (0.0002%); ω2 = 0.866 (22.14%)  M7  −5339.595  k = 3.830, α = 0.348; β = 0.965          M8  −5338.170  k = 3.872, α = 1.341; β = 8.405;  M7–M8  2.851  0.240    p0 = 0.841; ωs = 1.00 (15.89%)  hCHIA  M0  −6138.723  ω = 0.320, k =  4.132          M1a  −6079.211  k = 4.177;          ω0 = 0.140 (77.26%); ω1 = 1.00 (22.74%)  M2a  −6074.863  k = 4.283;  M1–M2  8.696  0.013    ω0 = 0.146 (77.36%); ω1 = 1.00 (21.68%); ω2 = 4.327 (0.96%)  M3  −6074.675  k = 4.219;  M0–M3  128.097  < 0.001    ω0 = 0.00 (31.07%); ω1 = 0.345 (60.07%); ω2 = 1.639 (8.86%)  M7  −6081.171  k = 4.160, α = 0.369; β = 0.758          M8  −6073.789  k = 4.235, α = 0.457; β = 1.025;  M7–M8  14.763  0.001  36P(0.953*), 62Q(0.971*), 164R(0.915), 280H(0.941)  p0 = 0.985; ω = 3.503 (1.52%)  a According to our hypothesis-testing framework, no mCHIA sites were found to be under positive selection; the vast majority (>75%) appear to be under purifying selection. Regarding hCHIA, models assuming positive selection outperformed null models, with 1.0–1.5% of sites found to be under positive selection. Results for the same analyses for the hCHIA gene indicated that a small number of sites (1.0–1.5%) may be subject to positive selection (table 2). Both the comparison of models M1a and M2a (χ2 = 8.70, df = 2, P = 0.013) and of models M7 and M8 (χ2 = 14.76, df = 2, P = 0.001) favored the hypothesis that sites in hCHIA are under positive selection over the null hypothesis. Bayes Empirical Bayes analysis indicated two sites that had a significant probability of being under positive selection, 36 P and 62Q (table 2). As with the mCHIA gene, the majority of sites in hCHIA appear to be under purifying selection. RELAX Analyses Although CODEML results suggested that most sites in mCHIA and hCHIA are under purifying selection, we also tested whether the strength of purifying selection acting on these sites varies across different branches of our phylogeny using the program RELAX. Given two categories of branches within a phylogeny, a set of test branches and a set of “background” or “reference” branches, RELAX tests whether the strength of selection was relaxed or intensified in one of these sets compared with the other (Wertheim et al. 2015). For hCHIA, RELAX results supported the hypothesis that selection was relaxed in species in which the gene has become pseudogenized (k = 0.02, P = 0, LR = 74.26, table 3). Compared with branches in which hCHIA remains functional, the ω values of pseudogene branches were shifted towards neutrality (ω = 1) indicating relaxed selection in these species (fig. 6a). Branch specific inferences of the selection intensity parameter (k) under the General Descriptive model in RELAX showed that the branches under more intense selection are ones with functional hCHIA genes, such as Saimiri sciureus (k = 1.99), Otolemur garnettii (k = 2.07), Tupaia chinensis (k = 2.47), and Cebus capucinus (k = 4.74) (fig. 7a). These species also have some of the highest insect intakes in our sample (table 1). Results of a RELAX test including only functional hCHIA sequences supported the hypothesis that hCHIA is under more intense selection in species with higher insect consumption than in species with lower insect consumption (k = 0.20, P < 0.001, LR = 26.6, table 3). The ω values of species with lower insect consumption were shifted toward neutrality compared with those of species with higher insect intake (fig. 6b), but the majority of sites remained below ω = 1 (0.165, 78%). We found similar results for our test of Kay’s threshold (fig. 6c). Selection on hCHIA was relaxed in species with body weights above this 500 g threshold (k = 0.67, LR = 7.68, P = 0.006, table 3). Table 3. RELAX Results.a Gene  Test Branches  Reference Branches  Model  log L  AICc  LR  P Value  mCHIA  Pseudogenes  Functional genes  Null  −5892.80  11956.49      Alternative  −5876.86  11926.63  31.88  <0.001  Partitioned Exploratory  −5866.18  11913.36  53.24  <0.001  Below average insect consumption  Above average insect consumption  Null  −4778.17  9691.10      Alternative  −4769.42  9675.61  17.50  <0.001  Partitioned Exploratory  −4764.26  9673.38  27.82  <0.001  Above Kay’s threshold  Below Kay’s threshold  Null  −4778.17  9691.10      Alternative  −4767.75  9672.29  20.84  <0.001  Partitioned Exploratory  −4765.17  9675.22  26.00  <0.001  hCHIA  Pseudogenes  Functional genes  Null  −6824.57  13820.01      Alternative  −6787.44  13747.78  74.26  <0.001  Partitioned Exploratory  −6787.46  13755.90  74.22  <0.001  Below average insect consumption  Above average insect consumption  Null  −4795.67  9685.95      Alternative  −4782.37  9661.39  26.60  <0.001  Partitioned Exploratory  −4781.84  9668.44  27.66  <0.001  Above Kay’s threshold  Below Kay’s threshold  Null  −4795.68  9685.98      Alternative  −4791.84  9680.33  7.68  0.006  Partitioned Exploratory  −4788.79  9682.33  13.78  0.017  Gene  Test Branches  Reference Branches  Model  log L  AICc  LR  P Value  mCHIA  Pseudogenes  Functional genes  Null  −5892.80  11956.49      Alternative  −5876.86  11926.63  31.88  <0.001  Partitioned Exploratory  −5866.18  11913.36  53.24  <0.001  Below average insect consumption  Above average insect consumption  Null  −4778.17  9691.10      Alternative  −4769.42  9675.61  17.50  <0.001  Partitioned Exploratory  −4764.26  9673.38  27.82  <0.001  Above Kay’s threshold  Below Kay’s threshold  Null  −4778.17  9691.10      Alternative  −4767.75  9672.29  20.84  <0.001  Partitioned Exploratory  −4765.17  9675.22  26.00  <0.001  hCHIA  Pseudogenes  Functional genes  Null  −6824.57  13820.01      Alternative  −6787.44  13747.78  74.26  <0.001  Partitioned Exploratory  −6787.46  13755.90  74.22  <0.001  Below average insect consumption  Above average insect consumption  Null  −4795.67  9685.95      Alternative  −4782.37  9661.39  26.60  <0.001  Partitioned Exploratory  −4781.84  9668.44  27.66  <0.001  Above Kay’s threshold  Below Kay’s threshold  Null  −4795.68  9685.98      Alternative  −4791.84  9680.33  7.68  0.006  Partitioned Exploratory  −4788.79  9682.33  13.78  0.017  a Model-fits of null, alternative, and partitioned exploratory models inferred by the program RELAX (Wertheim et al. 2015). For each CHIA paralog, selective patterns were compared between species with 1) pseudogenes and functional genes, 2) below and above average insect consumption, and 3) body weights above and below Kay’s threshold (500 g). Fig. 6. View largeDownload slide Patterns of natural selection across mCHIA and hCHIA. The best fitting model (as determined by AICc) for each RELAX analysis is shown. Three ω parameters and the percentage of sites they represent are plotted for test (blue/dark) and reference (red/light) branches. The vertical gray and dashed line at ω = 1 indicates neutral evolution. Asterisks indicate significant differences (P < 0.05) between test and reference branches. Fig. 6. View largeDownload slide Patterns of natural selection across mCHIA and hCHIA. The best fitting model (as determined by AICc) for each RELAX analysis is shown. Three ω parameters and the percentage of sites they represent are plotted for test (blue/dark) and reference (red/light) branches. The vertical gray and dashed line at ω = 1 indicates neutral evolution. Asterisks indicate significant differences (P < 0.05) between test and reference branches. Fig. 7. View largeDownload slide Branch specific relaxation parameters inferred for (a) hCHIA and (b) mCHIA genes under the General Descriptive model in RELAX (Wertheim et al. 2015). Branches are colored based on the selection intensity parameter k. A higher k value (light/red) indicates intensified selection, whereas a lower k value (dark/blue) indicates relaxed selection. Scale bars indicate number of substitutions per site. Very long branches were truncated (indicated by breaks) to avoid obscuring the variation present in the remaining branches. Branches with extremely high k values are highlighted in gray. Fig. 7. View largeDownload slide Branch specific relaxation parameters inferred for (a) hCHIA and (b) mCHIA genes under the General Descriptive model in RELAX (Wertheim et al. 2015). Branches are colored based on the selection intensity parameter k. A higher k value (light/red) indicates intensified selection, whereas a lower k value (dark/blue) indicates relaxed selection. Scale bars indicate number of substitutions per site. Very long branches were truncated (indicated by breaks) to avoid obscuring the variation present in the remaining branches. Branches with extremely high k values are highlighted in gray. The initial RELAX results for mCHIA suggested a more complex pattern. The RELAX test comparing pseudogene branches to branches with functional mCHIA genes was significant for selection intensification (k = 1.88, P <0.001) acting on pseudogene branches. However, the ω distributions of the best-fitting model, Partitioned Exploratory (LR = 53.24, P = <0.001, table 3), suggested that most sites (99.89%) of mCHIA pseudogenes are shifted toward neutrality (0.88–0.981) compared with most sites (97%) of functional genes (ω = 0.23–0.04). Only a very small number of hCHIA pseudogene sites (0.11%) were pushed far above neutrality (ω = 325) (fig. 6d). Similarly, the RELAX test comparing mCHIA between species with high and low insect consumption indicated intensified selection on branches with lower insect intake (k = 2.06, LR = 17.50, P < 0.001, table 3), but the best-fitting model, Partitioned Exploratory (fig. 6e), suggested a pattern of relaxation. When comparing species above and below Kay’s threshold, RELAX results reject the hypothesis that mCHIA is under relaxed selection in larger-bodied species, suggesting that they are instead under intensified selection (k = 3.81, LR = 20.8, P < 0.001, table 3) with ω values shifted away from neutrality (fig. 6e). Interestingly, branch specific inferences of k under the General Descriptive model show that three branches in our mCHIA sample that were part of the reference branches in all three RELAX tests have very strong selective regimes, two extremely so (fig. 7b). Otolemur garnettii (k = 3.89), Tarsius syrichta (k = 34.59), and Tupaia chinensis (k = 50) all appear to be under intense selection. It is likely that these extreme values and the long branch lengths observed in these species were biasing the RELAX test (Pond SLK, pers. comm.). We therefore ran these models again, leaving the O. garnettii, T. syrichta, and Tu. chinensis branches “unclassified” to exclude them from the hypothesis test and eliminate any bias contributed by these outliers. In support of our predictions, the RELAX test indicated that mCHIA pseudogene branches are under relaxed selection compared with branches with functional genes (k = 0.58, LR = 15.08, P < 0.001), and that selection on mCHIA is relaxed in species with lower insect consumption (k = 0.42, LR = 11.78, P < 0.001). The model comparing selection between smaller-bodied and larger-bodied species could not be retested, because only two species below 500 g remained in our sample after removing the aforementioned species. Discussion Here, we present the first comparative study of CHIA genes in primates, including species with a variety of dietary ecologies and different levels of insectivory. Our findings are consistent with the hypothesis that primates produce acidic mammalian chitinase (AMCase) as a digestive enzyme for the breakdown of insect exoskeletons and that this enzyme is under more intense selection in more insectivorous and smaller-bodied primates. We have identified a number of homologs of the CHIA gene in primates, mCHIA, hCHIA, CHIA3, CHIA4, and CHIA5. Three of these, mCHIA, hCHIA, and CHIA3 were previously identified by Krykbaev et al (2010) and are most likely ancestral to the Euarchonta, as they are also found in the treeshrew (Tupaia chinensis) (figs. 2 and 3). Because we could only identify a complete and putatively functional sequence for CHIA3 in the treeshrew, the galago (Otolemur garnettii), and the tarsier (Tarsius syrichta) (fig. 4), we conclude that CHIA3 was likely lost early in the anthropoid lineage (fig. 3). This loss is consistent with our current understanding of the ecological changes that are associated with the emergence of crown anthropoids. These changes include a shift to a diurnal lifestyle, an increase in body size, and a greater dietary reliance on fruit, rather than insects (Ross 1996; Ross 2000; Williams et al. 2010). The additional two genes are only found in the Philippine tarsier (T. syrichta, table 1), a member of the genus Tarsius, the most insectivorous primate taxon (Gursky 2011). These genes likely arose as duplications after the tarsier lineage split from the anthropoids between 60 and 70 Ma (Pozzi et al. 2014; Di Fiore et al. 2015; Kistler et al. 2015), but it is unclear which of the three ancestral CHIA genes gave rise to CHIA4 and CHIA5, as they are most similar to each other and equidistant from tarsier mCHIA, hCHIA, and CHIA3 (fig. 2). With the exception of Microcebus murinus (see Future Directions), the genes mCHIA and hCHIA are found in all species in our sample, but show a pattern of independent pseudogenization events that is consistent with our hypothesis that acidic mammalian chitinase is a digestive enzyme evolved and retained for insectivory. Most primates include at least some insects in their diet (Raubenheimer and Rothman 2013) and we find that most primates have one functional CHIA gene (fig. 3). Only two species in our sample did not retain any functional CHIA genes, the black snub-nosed monkey (Rhinopithecus bieti) and the proboscis monkey (Nasalis larvatus) (figs. 3 and 4a). Both of these monkeys are colobines, a subfamily of primates that are highly folivorous and generally do not include insects in their diet (table 1), with the exception of incidental ingestion of insects attached to leaves or found in fruit. Further, the only species in which >1 CHIA gene remained functional are species with above-average levels of insectivory. This includes some of the New World monkeys, which are generally both smaller-bodied and more reliant on insects than Old World monkeys (Gaulin 1979; Terborgh 1983). In the common marmoset (Callithrix jacchus), saddleback tamarin (Saguinus fuscicollis), the common and Bolivian squirrel monkeys (Saimiri sciureus and S. boliviensis, respectively), and the white-faced capuchin (Cebus capucinus) (fig. 4b), both mCHIA and hCHIA were free of premature stop codons and were conserved in the catalytic site and chitin-binding domain (fig. 5). Finally, the most insectivorous primate species in our sample are also the species with the greatest number of (putatively) functional CHIA genes, the galago and tarsier having three and five functional CHIA paralogs, respectively (fig. 4c and d). Tarsiers are the only primates that are entirely faunivorous (Gursky 2011), feeding mostly on arthropods and sometimes small vertebrates (Niemitz 1984; Gursky 2007). Although there are no gene expression data for the tarsier digestive system, it is noteworthy that, despite consuming arthropods almost exclusively, tarsier feces do not contain visible insect exoskeletons. Instead, their feces are described as “powder-like” (Gursky S, personal communication), suggesting that the exoskeletons ingested by the tarsier are broken down completely during gut transit. Such complete digestion is even more impressive considering that tarsiers scarcely chew their arthropod prey, despite having an extremely wide gape and powerful chewing muscles (Jablonski and Crompton 1994; Jablonski 2003). It has been argued that the sharp, pointed teeth of tarsiers are merely instruments of puncture rather than chewing (Prinz et al. 2003), a concept that elevates the functional importance of enzymatic digestion in the stomach and small intestine. In this light, the unique CHIA duplications we have documented here are likely to be important digestive adaptations for tarsiers. The primates with >1 functional CHIA gene are not just more insectivorous but are also among the smaller species in our sample (table 1), in accordance with what is predicted by Kay’s threshold (fig. 1). Because it is costly for small-bodied primates to fill their relatively smaller guts with indigestible bulk, such as leaves (Gaulin 1979; Fleagle 2013), the ability to efficiently digest insect exoskeletons using AMCase might be an especially valuable adaptation for these species. The results of our selection analyses show that CHIA genes are under intensified purifying selection in more insectivorous species (fig. 7a and b) and also partially support the hypothesis that selection on CHIA genes is relaxed in larger-bodied primates (fig. 6c). While the initial mCHIA RELAX test did not find support for this hypothesis (fig. 6e) and had inconsistent results for the other two tests (pseudogenes vs. functional genes [fig. 6c], above-average vs. below-average insectivory [fig. 6d]) it is worth noting that, in accordance with our predictions, mCHIA appears to be under extreme selection in the three most insectivorous species (Tupaia chinensis, Tarsius syrichta, and Otolemur garnettii). Further, it is likely that the extremely high k values and long branch lengths of these lineages (fig. 7b) masked other selective patterns across the alignment and biased the RELAX model (Pond, personal communication). This is supported by the results of additional analyses in which Tu. chinensis, T. syrichta, and O. garnettii were left “unclassified” and not included in the reference branches. Here, we found support for the hypotheses that mCHIA is under relaxed selection in 1) pseudogene branches and 2) less insectivorous species. Overall, these results suggest that having multiple functional CHIA paralogs is an adaptive advantage for insect-eating primates. Having multiple CHIA genes may correspond to increased AMCase production, as has been found for other digestive enzymes (Perry et al. 2007; Axelsson et al. 2013) or, alternatively, the different CHIA paralogs may have diverged in function (Zhang 2003), potentially specializing in the digestion of various types of arthropod prey. The exoskeletons of arthropods are not made up of pure chitin, but rather of a matrix of chitin and other compounds (Finke 2007), the exact composition of which varies across species (No et al. 1989; Kramer et al. 1995; Tomberlin et al. 2002; Nation 2015). Considering the diversity of arthropods consumed by primates (Raubenheimer and Rothman 2013; McGrew 2014; O'Malley and McGrew 2014), it is conceivable that having multiple CHIA paralogs does not merely increase the amount of AMCase secreted, but leads to the expression of different specialized forms of the enzyme. Our results may also shed some light on primate origins. Two commonly cited and competing hypotheses to explain the evolution of primate features, such as the visual system and grasping hands, are the Visual Predation Hypothesis (Cartmill 1972, 1992, 2012) and the Terminal Branch Feeding Hypothesis (Sussman 1991). According to the former, primate vision and hands evolved to facilitate the predation on insects, whereas the latter proposes that primate features evolved to allow for the exploitation of angiosperms at the ends of terminal branches. Although these two hypotheses do not have to be mutually exclusive, our results show that early primates likely had three CHIA genes, suggesting that insects were an important component of the ancestral primate diet, which favors the Visual Predation Hypothesis. Our results complement a recent study on primate opsin genes which found a shift to blue sensitivity in the SWS1 opsin of early primates, suggesting a lifestyle that required enhanced visual acuity (Melin et al. 2016). Such a shift is indicative of a need to resolve cryptic or fast-moving prey, rather than angiosperm reproductive tissues (Melin et al. 2016), and, like our results, supports the hypothesis that insects were an important food source for early primates. Limitations and Future Directions We were unable to confirm complete CHIA sequences in most of the currently available lemur genomes. In the mouse lemur (Microcebus murinus) genome we identified the full hCHIA sequence and found partial sequences of one, or possibly two, similar genes. However, we could not identify all exons of these additional genes with confidence, an issue that was also encountered with other lemur genomes. We therefore chose to focus on other taxa, but look forward to revisiting the evolutionary history of chitinase genes in the strepsirrhines as additional high-quality genomes become available. Similar issues were encountered with two other, nonprimate genomes, the Sunda colugo (Galeopterus variegatus) and the pen-tailed treeshrew (Ptilocercus lowii) (Mason et al. 2016). In the colugo, this is likely because the genes have diverged substantially, whereas the low-coverage of the genome may have been an issue in the case of Ptilocercus. Given that the diet of the Sunda colugo consists mainly of leaves (Dzulhelmi and Abdullah 2009), it is perhaps unsurprising that the CHIA gene sequence is not conserved in this species. One limitation of the current study is the available data on primate diets. During our review of the literature, we encountered many inconsistencies in the ways that data on food intake are collected, making it difficult to compare levels of insect consumption across species and even populations. In addition, identifying and observing insect-feeding is notoriously difficult (Frazão 1991), so it is very likely that insect consumption in many primates, and especially arboreal primates, is underestimated. Like exoskeletons, the cell walls of fungi and lichens are composed of chitin (Stevens and Hume 1995). Several primate species feed on fungi, most notably Callimico goeldii (up to 29% of feeding time) (Porter 2001; Hanson et al. 2006), and lichens are a staple for many colobines (Kirkpatrick et al. 2001; Fashing et al. 2007). Although it was outside of the scope of the present study, the relationship between AMCase and mycophagy deserves to be examined in future studies. Finally, without gene expression data for primate digestive systems, we cannot be certain whether the various CHIA genes are actually expressed in the stomachs of all primates. Krykbaev and colleagues report that mCHIA is highly expressed in the stomach of Macaca fascicularis, but expression data for humans suggests that the role of hCHIA may be more complicated. Although some studies show that hCHIA is expressed in the human stomach (Boot et al. 2005; Krykbaev et al. 2010), there is disagreement over whether the gene actually translates into a functional chitinase in the stomach. One study found that chitinolytic activity in the gastric juice was present in 80% of their participants, but absent in the other 20% (Paoletti et al. 2007), whereas another study failed to detect any evidence of chitinase in the human digestive system (Goto et al. 2003). It is plausible that this may be due to dietary difference across human populations and invites further study in groups with long histories of insect consumption. Conclusion Even though several studies from the 1970s suggested that primates have chitinolytic enzymes (Cornelius et al. 1976; Jeuniaux and Cornelius 1978; Kay and Sheine 1979), the notion that chitin was indigestible by the endogenous digestive enzymes of primates and other mammals has persisted (Cork and Kenagy 1989; Oftedal et al. 1991; Šimůnek and Bartoňová 2005; Strobel et al. 2013; Ohno et al. 2016). Here we present evidence that suggests insect-eating primates share an adaptation found in insectivorous bats (Vespertilionidae) and mice (Mus musculus) (Strobel et al. 2013; Ohno et al. 2016) and use the enzyme acidic mammalian chitinase to digest the chitin in insect exoskeletons. The efficient digestion of insect exoskeletons is likely to have important adaptive benefits for all insect-eating primates, through the potentially significant energy and amino acid returns from the digestion of the polysaccharide chitin (Finke 2007; Rothman et al. 2014) and, for small-bodied primates, by reducing the amount of indigestible bulk in their guts. Materials and Methods In this study, we collected and analyzed sequence data on the CHIA paralogs of 34 primate species from 27 genera with varying levels of insect consumption (table 1). We mined the published genomes of 22 primates and one treeshrew (Tupaia chinensis) for CHIA-like sequences using BLAST and sequenced the CHIA genes in additional primate species that do not have publicly available genomes. DNA samples for Callicebus moloch, Callithrix jacchus, Saguinus fuscicollis, and Saimiri sciureus were obtained from Coriell Biorepositories (see supplementary table 1, Supplementary Material online). Dr. George Perry provided extracted DNA from Sapajus apella. Extracted DNA from Erythrocebus patas was provided by Dr. Todd Disotell. DNA for the following samples was provided by one of us (AJT): Allenopithecus nigroviridis, Allochrocebus lhoesti, Cercopithecus mitis, Chlorocebus aethiops, Colobus guereza kikuyensis, and Miopithecus ogouensis. Genome Mining We conducted BLAST searches against the whole-genome sequences of the following taxa, using the Macaca fascicularis CHIA gene sequences (Krykbaev et al. 2010) as queries: Callithrix jacchus, Saimiri boliviensis, Aotus nancymaae, Cebus capucinus imitator, Cercocebus atys, Mandrillus leucophaeus, Papio anubis, Macaca mulatta, Macaca nemestrina, Chlorocebus sabaeus, Nasalis larvatus, Rhinopithecus roxellana, Rhinopithecus bieti, Colobus angolensis, Gorilla gorilla gorilla, Pan paniscus, Pan troglodytes, Pongo abelii, Nomascus leucogenys, Tarsius syrichta, Microcebus murinus, Otolemur garnettii, and Tupaia chinensis whole-genome sequences (GenBank accession numbers in supplementary table 1, Supplementary Material online). Amplification, Sequencing, and Assembly We designed PCR primers for the amplification of each of the 11 exons of both CHIA paralogs (mCHIA, hCHIA) in Old World and New World monkeys. We first tried to find regions that were conserved across a wide range of primates, including both platyrrhines and catarrhines, however, it was difficult to find intron regions that were suitable as PCR primer sites. Separate sets of PCR primers therefore had to be designed for catarrhine species and for platyrrhine species for most exons (supplementary table 2, Supplementary Material online). After aligning the CHIA sequences of multiple catarrhines (Macaca fascicularis, Mandrillus leucophaeus, Nasalis larvatus, Cercocebus atys, Papio anubis, Rhinopithecus roxellana, and Chlorocebus sabaeus) and platyrrhines (Saimiri boliviensis, Callithrix jacchus, and Aotus nancymaae), ambiguities in the consensus sequence were masked and primers were designed using Primer3 (http://primer3.ut.ee/). PCRs were carried out with one of two commercial kits. One kit was the Qiagen Fast Cycling PCR Kit MasterMix, which was used following manufacturer’s protocols, but with a proportionally reduced volume of 15 µl. Reactions were run at 95 °C for 5 min followed by 35 cycles of 96 °C for 15 s, 56–59 °C for 15 s, and 68 °C for 35 s, and a final step at 72 °C for 7 min. PCR products were purified with exonuclease I and shrimp alkaline phosphatase (ExoSAP) or, when gels showed the amplification of nontarget sequences (i.e., multiple bands), with the Qiagen QIAquick Gel Extraction Kit. Purified PCR products were sequenced directly on the Applied Biosystems 3500 Genetic Analyzer using the Applied Biosystems BigDye Terminator v3.1 Cycle Sequencing Kit and the same primers as used for PCR. Alternatively, PCRs were carried out in 25 µl with the Promega GoTaq G2 Master Mix and run under the following conditions: 95° for 2 min followed by 35–40 cycles of 95° for 20 s, 55°–60° for 25 s, and 72° for 30–60 s. These cycles were followed by a final extension phase at 72° for 5 min. Some of the exons of the hCHIA gene were difficult to amplify in Co. guereza, Ce. mitis, A. lhoesti, Ch. aethiops, M. ogouensis, and Al. nigroviridis. For these taxa, we used a long range PCR approach with Promega GoTaq Long PCR Master Mix to amplify the entire hCHIA gene (8–9 kb) under the following thermal cycling conditions: 95° for 2 min followed by 35 cycles of 95° for 20 s, 57° for 25 s, and 65° for 18 min. These long-range reactions were followed by a final extension phase of 72° for 10 min. These whole-gene products were then cleaned up using Millipore Microcon centrifugal filters by spinning the products at 500 rcf for 4 min, and then spinning the inverted filter to recover only dsDNA at 1,000 rcf for 3 min. We then diluted the cleaned-up DNA concentrate with 10 µl of nuclease-free water. This step was taken in order to quantitate the DNA so that we could use an appropriate amount of these products for subsequent nested PCRs, which targeted the problematic exons mentioned above. The nested products were sequenced directly at the Molecular Cloning Laboratories (San Francisco, CA) on an ABI 3730XL sequencer. Reads were assembled, mapped to the Macaca fascicularis reference sequence, concatenated into coding sequences, and translated using Geneious v. 9.1.8. Coding regions and translated amino acid sequences were aligned with MAFFT alignment server v. 7. Sequence Analyses Sequences were visually inspected for frameshift mutations causing premature stop codons; these were considered to be CHIA pseudogenes. To determine whether full-length CHIA sequences were likely coding for a functional enzyme, we inspected the translated amino acid sequences for conservation of the catalytic site motif and chitin-binding domain, signatures of functional chitinases (Tjoelker et al. 2000; Krykbaev et al. 2010). Premature stop codons were removed from pseudogene sequences by excising the insertion(s) or deletion(s) that caused the premature stop codon. We looked for evidence of positive selection acting on sites along both CHIA genes in primates with the CODEML program in the PAML package (Yang 2007). For these analyses, premature stop codons were removed from pseudogene sequences by excising the insertion(s) or deletion(s) that caused the premature stop codon. We used site-specific models (M0: null, M1a: nearly neutral selection, M2a: positive selection, M3: discrete, M7: beta, and M8: beta and ω > 1) to determine if there is variation in the ratios of nonsynonymous to synonymous nucleotide substitutions (dN/dS or ω) across sites along the mCHIA and hCHIA alignments and to test for evidence of positively selected sites (Yang 2007). Model fit was evaluated using likelihood ratio tests (LRT). The software program RELAX, which is part of the HyPhy package, tests for relaxed versus intensified selection in a codon-based phylogenetic framework (Wertheim et al. 2015). Given two sets of branches in a phylogeny (foreground and background, or test and reference branches), RELAX tests whether selection is intensified or relaxed in one set versus the other. To do this, RELAX uses a branch-site evolutionary (BS-REL) model to estimate the distribution of ω for each of the two branch sets and then compares this distribution using two models. In the null model the selection intensity parameter k is constrained to 1, causing the ω distribution to be the same on both test and reference branches, while under the alternative model k is allowed to vary. If the latter model is a significantly better fit (as determined by a LRT), this suggests that selection on the test branches is either relaxed (k < 1) or intensified (k > 1) compared with the reference branches (Wertheim et al. 2015). In addition to the null and alternative models, the partitioned exploratory model can provide a quantitative measure of selection patterns in the test and reference branches. The partitioned exploratory model is a less constrained model that allows the proportion of sites in each category of ω to vary between test and reference branches. Here we used RELAX to test whether selection was relaxed or intensified along any of the branches in our phylogeny as a function of either insect consumption or body size. Specifically, we hypothesized that selection on CHIA genes would be relaxed in lineages with 1) pseudogenizing mutations, 2) below average insect consumption, and 3) body size above Kay’s threshold (500 g). Phenotypic Data Data on insect consumption and body size were taken from the literature. Most of the data on body size came from (Smith and Jungers 1997) and only measures from wild specimens were used. To avoid skews introduced by large adult males in sexually dimorphic species we only used adult female body weight measures for all species. Average adult female body weight for each species is presented in table 1. Species-specific information on average annual insect consumption was taken from studies of wild primates, the results of many of which are collected in Campbell et al (2011). Annual averages of insect consumption used in our analyses are presented in table 1, while detailed information on all data used to calculate these averages is given in supplementary table 3, Supplementary Material online. Most of these studies present data on the percent of time spent feeding, and we did not include estimates that included foraging time. A few studies presented data based on stomach contents. Because of the inconsistent ways in which dietary data is collected and presented in the literature, we decided to group species into one of two categories for the analyses, rather than using insect intake as a continuous variable. Species were classified as relatively more or less insectivorous depending on whether their average annual insect consumption was higher or lower than the average annual insect consumption across all species included in our sample (16.58%). Furthermore, where possible we used data from 1) studies that covered at least 12 months to account for seasonal variation in insect consumption, and 2) multiple studies for each species to account for interpopulation differences. The diets of a few of the species included here have not been studied in the wild; in these cases, we used data from a closely related species where available (Saimiri boliviensis and Aotus nancymaae), or where unavailable (Mandrillus leucophaeus) left the branch “unclassified” in the RELAX analyses involving dietary data. Supplementary Material Supplementary data are available at Molecular Biology and Evolution online. Acknowledgments We thank the editors and two anonymous reviewers for their thoughtful comments and suggestions. Dr. Robert Scott provided helpful feedback throughout this study. We thank P.J. Perry, Todd Disotell, and the Coriell Biorepository for providing DNA samples. This work was supported by an Albert Fellows Dissertation Research Award from the Center of Human Evolutionary Studies at Rutgers University, the International Primatological Society, and the Kent State University College of Arts and Sciences. References Axelsson E, Ratnakumar A, Arendt M-L, Maqbool K, Webster MT, Perloski M, Liberg O, Arnemo JM, Hedhammar Å, Lindblad-Toh K. 2013. The genomic signature of dog domestication reveals adaptation to a starch-rich diet. Nature  495( 7441): 360– 364. Google Scholar CrossRef Search ADS PubMed  Bishop A. 1964. Use of the hand in lower primates. In: Buettner-Janush J, editor. Evolutionary and genetic biology of primates . 2nd ed. New York: Elsevier. p. 133– 225. Google Scholar CrossRef Search ADS   Boot RG, Bussink AP, Verhoek M, de Boer PAJ, Moorman AFM, Aerts JMFG. 2005. Marked differences in tissue-specific expression of chitinases in mouse and man. J Histochem Cytochem . 53( 10): 1283– 1292. http://dx.doi.org/10.1369/jhc.4A6547.2005 Google Scholar CrossRef Search ADS PubMed  Bucolo C, Musumeci M, Musumeci S, Drago F. 2011. Acidic mammalian chitinase and the eye: implications for ocular inflammatory diseases. Front Pharmacol.  2: 43. Google Scholar CrossRef Search ADS PubMed  Campbell CJ, Fuentes A, MacKinnon KC, Bearder SK, Stumpf RM, editors. 2011. Primates in perspective . 2nd ed. New York: Oxford University Press Cartmill M. 1972. Arboreal adaptations and the origin of the order Primates. In: Tuttle RH, editor. The functional and evolutionary biology of primates . Chicago: Aldine-Atherton. p. 97– 122. Cartmill M. 1992. New views on primate origins. Evol Anthropol . 1( 3): 105– 111. http://dx.doi.org/10.1002/evan.1360010308 Google Scholar CrossRef Search ADS   Cartmill M. 2012. Primate origins, human origins, and the end of higher taxa. Evol Anthropol . 21( 6): 208– 220. http://dx.doi.org/10.1002/evan.21324 Google Scholar CrossRef Search ADS PubMed  Chivers DJ, Hladik CM. 1980. Morphology of the gastrointestinal tract in primates: comparisons with other mammals in relation to diet. J Morphol.  166( 3): 337– 386. Google Scholar CrossRef Search ADS PubMed  Clutton Brock TH, Harvey PH. 1980. Primates, brains and ecology. J Zool.  190( 3): 309– 323. Google Scholar CrossRef Search ADS   Cork SJ, Kenagy GJ. 1989. Nutritional value of hypogeous fungus for a forest‐dwelling ground squirrel. Ecology  70( 3): 577– 586. Google Scholar CrossRef Search ADS   Cornelius C, Dandrifosse G, Jeuniaux C. 1976. Chitinolytic enzymes of the gastric mucosa of Perodicticus potto (primate prosimian): purification and enzyme specificity. Int J Biochem.  7( 9–10): 445– 448. Google Scholar CrossRef Search ADS   Di Fiore A, Chaves PB, Cornejo FM, Schmitt CA, Shanee S, Cortés-Ortiz L, Fagundes V, Roos C, Pacheco V. 2015. The rise and fall of a genus: complete mtDNA genomes shed light on the phylogenetic position of yellow-tailed woolly monkeys, Lagothrix flavicauda, and on the evolutionary history of the family Atelidae (Primates: Platyrrhini). Mol Phylogenet Evol.  82 Pt B: 495– 510. Google Scholar CrossRef Search ADS PubMed  Dzulhelmi MN, Abdullah MT. 2009. Foraging ecology of the Sunda colugo (Galopterus variegatus) in Bako National Park, Sarawak, Malaysia. Malayan Nat J.  61: 285– 294. Fashing PJ, Mulindahabi F, Gakima J-B, Masozera M, Mununura I, Plumptre AJ, Nguyen N. 2007. Activity and ranging patterns of Colobus angolensis ruwenzorii in Nyungwe Forest, Rwanda: possible costs of large group size. Int J Primatol.  28( 3): 529– 550. Google Scholar CrossRef Search ADS   Finke MD. 2007. Estimate of chitin in raw whole insects. Zoo Biol . 26( 2): 105– 115. http://dx.doi.org/10.1002/zoo.20123 Google Scholar CrossRef Search ADS PubMed  Fleagle JG. 2013. Primate adaptation and evolution . 3rd ed. Amsterdam: Academic Press. Frazão E. 1991. Insectivory in free-ranging bearded saki (Chiropotes satanas chiropotes). Primates  32( 2): 243– 245. Google Scholar CrossRef Search ADS   Gaulin SJ. 1979. A Jarman/Bell model of primate feeding niches. Hum Ecol.  7( 1): 1– 20. http://dx.doi.org/10.1007/BF00889349 Google Scholar CrossRef Search ADS   Gingerich PD. 1980. Dental and cranial adaptations in Eocene Adapidae. Z Morphol Anthropol.  71( 2): 135– 142. Google Scholar PubMed  Goodall J. 1986. The chimpanzees of Gombe: patterns of behavior . Cambridge (MA): Belknap Press of Harvard University Press. Goto M, Fujimoto W, Nio J, Iwanaga T, Kawasaki T. 2003. Immunohistochemical demonstration of acidic mammalian chitinase in the mouse salivary gland and gastric mucosa. Arch Oral Biol . 48( 10): 701– 707. http://dx.doi.org/10.1016/S0003-9969(03)00150-X Google Scholar CrossRef Search ADS PubMed  Groves CP. 2001. Primate taxonomy . Washington (DC): Smithsonian. Guindon S, Dufayard J-F, Lefort V, Anisimova M, Hordijk W, Gascuel O. 2010. New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of PhyML 3.0. Syst Biol . 59( 3): 307– 321. Google Scholar CrossRef Search ADS PubMed  Gursky SL. 2007. The Spectral Tarsier . Abingdon (UK): Routledge. Gursky SL. 2011. Tarsiiformes. In: Campbell CJ, Fuentes A, MacKinnon KC, Bearder SK, Stumpf RM, editors. Primates in perspective . New York: Oxford University Press. p. 79– 90. Hanson AM, Hall MB, Porter LM, Lintzenich B. 2006. Composition and nutritional characteristics of fungi consumed by Callimico goeldii in Pando, Bolivia. Int J Primatol.  27( 1): 323– 346. http://dx.doi.org/10.1007/s10764-005-9014-z Google Scholar CrossRef Search ADS   Isbell LA. 1998. Diet for a small primate: insectivory and gummivory in the (large) patas monkey (Erythrocebus patas pyrrhonotus). Am J Primatol . 45( 4): 381– 398. http://dx.doi.org/10.1002/(SICI)1098-2345(1998)45:4<381::AID-AJP5>3.0.CO;2-S Google Scholar CrossRef Search ADS PubMed  Jablonski NG, Crompton RH. 1994. Feeding behavior, mastication, and tooth wear in the western tarsier (Tarsius bancanus). Int J Primatol.  15( 1): 29– 59. http://dx.doi.org/10.1007/BF02735233 Google Scholar CrossRef Search ADS   Jablonski NG. 2003. The evolution of the tarsiid niche. In: Wright PC, Simons EL, Gursky SL, editors. Tarsiers past, present, and future . New Brunswick (NJ): Rutgers University Press. p. 35– 49. Jeuniaux C, Cornelius C. 1978. Distribution and activity of chitinolytic enzymes in the digestive tract of birds and mammals. In: Muzzarelli RAA, Pariser ER, editors. Proceedings of the first international conference on chitin/chitosan. Boston (MA): Massachusetts Institute of Technology. p 542–549. Jones KE, Bielby J, Cardillo M, Fritz SA, O'Dell J, Orme CDL, Safi K, Sechrest W, Boakes EH, Carbone C. 2009. PanTHERIA: a species‐level database of life history, ecology, and geography of extant and recently extinct mammals. Ecology  90( 9): 2648– 2648. Google Scholar CrossRef Search ADS   Kay RF, Covert HH. 1984. Anatomy and behaviour of extinct primates. In: Chivers DJ, Wood BA, Bilsborough A, editors. Food acquisition and processing in primates . Boston (MA): Springer US. p. 467– 508. Google Scholar CrossRef Search ADS   Kay RF, Sheine WS. 1979. On the relationship between chitin particle size and digestibility in the primate Galago senegalensis. Am J Phys Anthropol . 50( 3): 301– 308. http://dx.doi.org/10.1002/ajpa.1330500303 Google Scholar CrossRef Search ADS   Kay RF. 1975. The functional adaptations of primate molar teeth. Am J Phys Anthropol . 43( 2): 195– 216. http://dx.doi.org/10.1002/ajpa.1330430207 Google Scholar CrossRef Search ADS PubMed  Kay RF. 1984. On the use of anatomical features to infer foraging behavior in extinct primates. In: Cant J, Rodman PS, editors. Adaptations for foraging in nonhuman primates . New York: Columbia University Press. p. 21– 53. Kirkpatrick RC, Zou RJ, Dierenfeld ES, Zhou HW. 2001. Digestion of selected foods by Yunnan snub-nosed monkey Rhinopithecus bieti (Colobinae). Am J Phys Anthropol . 114( 2): 156– 162. Google Scholar CrossRef Search ADS PubMed  Kistler L, Ratan A, Godfrey LR, Crowley BE, Hughes CE, Lei R, Cui Y, Wood ML, Muldoon KM, Andriamialison H, et al.   2015. Comparative and population mitogenomic analyses of Madagascar's extinct, giant ‘subfossil’ lemurs. J Hum Evol.  79: 45– 54. Google Scholar CrossRef Search ADS PubMed  Kramer KJ, Hopkins TL, Schaefer J. 1995. Applications of solids NMR to the analysis of insect sclerotized structures. Insect Biochem Mol Biol.  25( 10): 1067– 1080. http://dx.doi.org/10.1016/0965-1748(95)00053-4 Google Scholar CrossRef Search ADS   Krykbaev R, Fitz LJ, Reddy PS, Winkler A, Xuan D, Yang X, Fleming M, Wolf SF. 2010. Evolutionary and biochemical differences between human and monkey acidic mammalian chitinases. Gene  452( 2): 63– 71. Google Scholar CrossRef Search ADS PubMed  Macdonald C, Barden S, Foley S. 2014. Isolation and characterization of chitin-degrading micro-organisms from the faeces of Goeldi’s monkey, Callimico goeldii. J Appl Microbiol.  116( 1): 52– 59. Google Scholar CrossRef Search ADS PubMed  Mason VC, Li G, Minx P, Schmitz J, Churakov G, Doronina L, Melin AD, Dominy NJ, Lim NTL, Springer MS, et al.   2016. Genomic analysis reveals hidden biodiversity within colugos, the sister group to primates. Sci Adv.  2( 8): e1600633. Google Scholar CrossRef Search ADS PubMed  McGrew WC. 1992. Chimpanzee material culture: implications for human evolution . Cambridge: Cambridge University Press. Google Scholar CrossRef Search ADS   McGrew WC. 2001. The other faunivory: primate insectivory and early human diet. In: Stanford CB, Bunn HT, editors. Meat-eating and human evolution . Oxford: Meat-eating and Human Evolution. p. 160– 178. McGrew WC. 2014. The “other faunivory” revisited: insectivory in human and non-human primates and the evolution of human diet. J Hum Evol.  71: 4– 11. Google Scholar CrossRef Search ADS PubMed  Melin AD, Wells K, Moritz GL, Kistler L, Orkin JD, Timm RM, Bernard H, Lakim MB, Perry GH, Kawamura S, et al.   2016. Euarchontan opsin variation brings new focus to primate origins. Mol Biol Evol . 33( 4): 1029– 1041. Google Scholar CrossRef Search ADS PubMed  Milton K. 1984. The role of food-processing factors in primate food choice. In: Rodman PS, Cant JGH, editors. Adaptations for foraging in nonhuman primates . New York: Columbia University Press. p. 249– 279. Musumeci M, Aragona P, Bellin M, Maugeri F, Rania L, Bucolo C, Musumeci S. 2009. Acidic mammalian chitinase in dry eye conditions. Cornea  28( 6): 667– 672. Google Scholar CrossRef Search ADS PubMed  Muzzarelli RAA, Boudrant J, Meyer D, Manno N, DeMarchis M, Paoletti MG. 2012. Current views on fungal chitin/chitosan, human chitinases, food preservation, glucans, pectins and inulin: a tribute to Henri Braconnot, precursor of the carbohydrate polymers science, on the chitin bicentennial. Carbohydr Polym.  87( 2): 995– 1012. Google Scholar CrossRef Search ADS   Nation JL. 2015. Insect physiology and biochemistry , 3rd ed. Boca Raton (FL): CRC Press. Niemitz C. 1984. Biology of tarsiers . Stuttgart: Gustav Fischer. No HK, Meyers SP, Lee KS. 1989. Isolation and characterization of chitin from crawfish shell waste. J Agric Food Chem.  37( 3): 575– 579. http://dx.doi.org/10.1021/jf00087a001 Google Scholar CrossRef Search ADS   O’Malley RC, McGrew WC. 2014. Primates, insects and insect resources. J Hum Evol.  71: 1– 3. Google Scholar CrossRef Search ADS PubMed  Oftedal OT, Whiten A, Southgate DAT, Van Soest PJ. 1991. The nutritional consequences of foraging in primates: the relationship of nutrient intakes to nutrient requirements. Philos Trans R Soc Lond B Biol Sci . 334: 161– 170. http://dx.doi.org/10.1098/rstb.1991.0105 Google Scholar CrossRef Search ADS PubMed  Ohno M, Kimura M, Miyazaki H, Okawa K, Onuki R, Nemoto C, Tabata E, Wakita S, Kashimura A, Sakaguchi M, et al.   2016. Acidic mammalian chitinase is a proteases-resistant glycosidase in mouse digestive system. Sci Rep . 6: 37756. Google Scholar CrossRef Search ADS PubMed  Paoletti MG, Norberto L, Damini R, Musumeci S. 2007. Human gastric juice contains chitinase that can degrade chitin. Ann Nutr Metab . 51( 3): 244– 251. http://dx.doi.org/10.1159/000104144 Google Scholar CrossRef Search ADS PubMed  Perry GH, Dominy NJ, Claw KG, Lee AS, Fiegler H, Redon R, Werner J, Villanea FA, Mountain JL, Misra R, et al.   2007. Diet and the evolution of human amylase gene copy number variation. Nat Genet.  39( 10): 1256– 1260. Google Scholar CrossRef Search ADS PubMed  Porter LM. 2001. Dietary differences among sympatric Callitrichinae in northern Bolivia: Callimico goeldii, Saguinus fuscicollis and S. labiatus. Int J Primatol.  22: 961– 992. http://dx.doi.org/10.1023/A:1012013621258 Google Scholar CrossRef Search ADS   Pozzi L, Hodgson JA, Burrell AS, Sterner KN, Raaum RL, Disotell TR. 2014. Primate phylogenetic relationships and divergence dates inferred from complete mitochondrial genomes. Mol Phylogenet. Evol.  75: 165– 183. Google Scholar CrossRef Search ADS PubMed  Prinz JF, Silwood CJL, Claxson AWD, Grootveld M. 2003. Simulated digestion status of intact and exoskeletally-punctured insects and insect larvae: a spectroscopic investigation. Folia Primatol.  74( 3): 126– 140. Google Scholar CrossRef Search ADS PubMed  Raubenheimer D, Rothman JM. 2013. Nutritional ecology of entomophagy in humans and other primates. Annu Rev Entomol . 58: 141– 160. http://dx.doi.org/10.1146/annurev-ento-120710-100713 Google Scholar CrossRef Search ADS PubMed  Reese TA, Liang H-E, Tager AM, Luster AD, Van Rooijen N, Voehringer D, Locksley RM. 2007. Chitin induces accumulation in tissue of innate immune cells associated with allergy. Nature  447( 7140): 92– 96. Google Scholar CrossRef Search ADS PubMed  Ross C. 1996. Adaptive explanation for the origins of the anthropoidea (primates). Am J Primatol . 40( 3): 205– 230. http://dx.doi.org/10.1002/(SICI)1098-2345(1996)40:3<205::AID-AJP1>3.0.CO;2-1 Google Scholar CrossRef Search ADS   Ross CF. 2000. Into the light: the origin of anthropoidea. Annu Rev Anthropol . 29( 1): 147– 194. http://dx.doi.org/10.1146/annurev.anthro.29.1.147 Google Scholar CrossRef Search ADS   Rothman JM, Raubenheimer D, Bryer MAH, Takahashi M, Gilbert CC. 2014. Nutritional contributions of insects to primate diets: implications for primate evolution. J Hum Evol.  71: 59– 69. Google Scholar CrossRef Search ADS PubMed  Šimůnek J, Bartoňová H. 2005. Effect of dietary chitin and chitosan on cholesterolemia of rats. Acta Veterinaria Brno.  74( 4): 491– 499. Google Scholar CrossRef Search ADS   Smith RJ, Jungers WL. 1997. Body mass in comparative primatology. J Hum Evol.  32( 6): 523– 559. http://dx.doi.org/10.1006/jhev.1996.0122 Google Scholar CrossRef Search ADS PubMed  Souto A, Bione CBC, Bastos M, Bezerra BM, Fragaszy D, Schiel N. 2011. Critically endangered blonde capuchins fish for termites and use new techniques to accomplish the task. Biol Lett.  7( 4): 532– 535. Google Scholar CrossRef Search ADS PubMed  Stevens CE, Hume ID. 1995. Comparative physiology of the vertebrate digestive system . 2nd ed. Cambridge: Cambridge University Press. Strait SG, Vincent J. 1998. Primate faunivores: physical properties of prey items. Int J Primatol.  19( 5): 867– 878. http://dx.doi.org/10.1023/A:1020397430482 Google Scholar CrossRef Search ADS   Strobel S, Roswag A, Becker NI, Trenczek TE, Encarnação JA. 2013. Insectivorous bats digest chitin in the stomach using acidic mammalian chitinase. PLoS One  8( 9): e72770. Google Scholar CrossRef Search ADS PubMed  Sussman RW. 1991. Primate origins and the evolution of angiosperms. Am. J. Primatol . 23( 4): 209– 223. http://dx.doi.org/10.1002/ajp.1350230402 Google Scholar CrossRef Search ADS   Synstad B, Gaseidnes S, van Aalten DMF, Vriend G, Nielsen JE, Eijsink VGH. 2004. Mutational and computational analysis of the role of conserved residues in the active site of a family 18 chitinase. Eur J Biochem.  271( 2): 253– 262. Google Scholar CrossRef Search ADS PubMed  Terborgh J. 1983. Five new world primates: a study in comparative ecology . Princeton (NJ): Princeton University Press. Tjoelker LW, Gosting L, Frey S, Hunter CL, Trong HL, Steiner B, Brammer H, Gray PW. 2000. Structural and functional definition of the human chitinase chitin-binding domain. J Biol Chem . 275( 1): 514– 520. Google Scholar CrossRef Search ADS PubMed  Tomberlin JK, Sheppard DC, Joyce JA. 2002. Selected life-history traits of black soldier flies (Diptera: Stratiomyidae) reared on three artificial diets. Ann Entomol Soc Am.  95: 379– 386. http://dx.doi.org/10.1603/0013-8746(2002)095[0379:SLHTOB]2.0.CO;2 Google Scholar CrossRef Search ADS   Tosi AJ, Buzzard PJ, Morales JC, Melnick DJ. 2002. Y-chromosome data and tribal affiliations of Allenopithecus and Miopithecus. Int J Primatol.  23( 6): 1287– 1299. http://dx.doi.org/10.1023/A:1021178904824 Google Scholar CrossRef Search ADS   Tosi AJ, Detwiler KM, Disotell TR. 2005. X-chromosomal window into the evolutionary history of the guenons (Primates: Cercopithecini). Mol Phylogenet Evol.  36( 1): 58– 66. http://dx.doi.org/10.1016/j.ympev.2005.01.009 Google Scholar CrossRef Search ADS PubMed  van Schaik CP, Ancrenaz M, Borgen G, Galdikas B, Knott CD, Singleton I, Suzuki A, Utami SS, Merrill M. 2003. Orangutan cultures and the evolution of material culture. Science  299( 5603): 102– 105. Google Scholar CrossRef Search ADS PubMed  Vannella KM, Ramalingam TR, Hart KM, de Queiroz Prado R, Sciurba J, Barron L, Borthwick LA, Smith AD, Mentink-Kane M, White S. 2016. Acidic chitinase primes the protective immune response to gastrointestinal nematodes. Nat Immunol.  21: 73– 79. Wertheim JO, Murrell B, Smith MD, Kosakovsky Pond SL, Scheffler K. 2015. RELAX: detecting relaxed selection in a phylogenetic framework. Mol Biol Evol . 32( 3): 820– 832. http://dx.doi.org/10.1093/molbev/msu400 Google Scholar CrossRef Search ADS PubMed  Whitaker JO, Dannelly HK, Prentice DA. 2004. Chitinase in insectivorous bats. J Mammal.  85( 1): 15– 18. http://dx.doi.org/10.1644/1545-1542(2004)085<0015:CIIB>2.0.CO;2 Google Scholar CrossRef Search ADS   Williams BA, Kay RF, Kirk EC. 2010. New perspectives on anthropoid origins. Proc Natl Acad Sci U S A . 107( 11): 4797– 4804. Google Scholar CrossRef Search ADS PubMed  Yang Z. 2007. PAML 4: phylogenetic analysis by maximum likelihood. Mol Biol Evol . 24( 8): 1586– 1591. http://dx.doi.org/10.1093/molbev/msm088 Google Scholar CrossRef Search ADS PubMed  Zhang J. 2003. Evolution by gene duplication: an update. Trends Ecol Evol.  18( 6): 292– 298. http://dx.doi.org/10.1016/S0169-5347(03)00033-8 Google Scholar CrossRef Search ADS   Zhu Z, Zheng T, Homer RJ, Kim Y-K, Chen NY, Cohn L, Hamid Q, Elias JA. 2004. Acidic mammalian chitinase in asthmatic Th2 inflammation and IL-13 pathway activation. Science  304( 5677): 1678– 1682. Google Scholar CrossRef Search ADS PubMed  © The Author 2017. Published by Oxford University Press on behalf of the Society for Molecular Biology and Evolution. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Molecular Biology and Evolution Oxford University Press

Evolution of Acidic Mammalian Chitinase Genes (CHIA) Is Related to Body Mass and Insectivory in Primates

Loading next page...
 
/lp/ou_press/evolution-of-acidic-mammalian-chitinase-genes-chia-is-related-to-body-TfRLqkbiYF
Publisher
Oxford University Press
Copyright
© The Author 2017. Published by Oxford University Press on behalf of the Society for Molecular Biology and Evolution. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com
ISSN
0737-4038
eISSN
1537-1719
D.O.I.
10.1093/molbev/msx312
Publisher site
See Article on Publisher Site

Abstract

Abstract Insects are an important food resource for many primates, but the chitinous exoskeletons of arthropods have long been considered to be indigestible by the digestive enzymes of most mammals. However, recently mice and insectivorous bats were found to produce the enzyme acidic mammalian chitinase (AMCase) to digest insect exoskeletons. Here, we report on the gene CHIA and its paralogs, which encode AMCase, in a comparative sample of nonhuman primates. Our results show that early primates likely had three CHIA genes, suggesting that insects were an important component of the ancestral primate diet. With some exceptions, most extant primate species retain only one functional CHIA paralog. The exceptions include two colobine species, in which all CHIA genes have premature stop codons, and several New World monkey species that retain two functional genes. The most insectivorous species in our sample also have the largest number of functional CHIA genes. Tupaia chinensis and Otolemur garnettii retain three functional CHIA paralogs, whereas Tarsius syrichta has a total of five, two of which may be duplications specific to the tarsier lineage. Selection analyses indicate that CHIA genes are under more intense selection in species with higher insect consumption, as well as in smaller-bodied species (<500 g), providing molecular support for Kay’s Threshold, a well-established component of primatological theory which proposes that only small primates can be primarily insectivorous. These findings suggest that primates, like mice and insectivorous bats, may use the enzyme AMCase to digest the chitin in insect exoskeletons, providing potentially significant nutritional benefits. digestive enzymes, acidic mammalian chitinase, insectivory, dietary adaptations Introduction All primates include some insects in their diet, whether through accidental consumption or through active insectivory (Raubenheimer and Rothman 2013). Degree of insectivory in living primates ranges from nearly exclusive (e.g., Tarsius spp.) to complementary (e.g., the Callitrichidae) to supplemental, such as in the great apes (McGrew 2001). Although insects are a significant source of energy and protein for many living primates, this is especially true for small-bodied primates (Kay 1984). Due to their sparse distribution in most environments, insects are usually energetically costly to find and catch, making it difficult for a large-bodied primate to fill their nutritional demands solely with insects (Raubenheimer and Rothman 2013). Small-bodied primates have relatively higher metabolic requirements per unit body mass, but are small enough that the insects they catch suffice to meet their nutritional needs (Fleagle 2013). Kay (1984) calculated that a primate above 500 g will simply not be able to catch enough insects in a day to satisfy its daily energy requirements (Kay 1984; Kay and Covert 1984). Following from this is a classic concept in primatology, Kay’s Threshold (Kay 1984), which predicts that only species below 500 g will be insectivores, whereas only species above this weight will be folivorous (fig. 1). Among frugivorous primates, those that are smaller (≤1 kg) will typically rely on insects as their source of protein, whereas those that are larger will use leaves (Gingerich 1980; Kay 1984; Fleagle 2013) (fig. 1). Social insects, including ants and termites, represent an exception to this rule because they occur as clumped resources across time and space and can be efficiently preyed upon by larger primates (Isbell 1998). In the case of termites, extractive foraging tools are often used (Goodall 1986; McGrew 1992; van Schaik 2003; Souto et al. 2011). Although the nutrient composition of insects varies widely, geometric analyses show that insects eaten by nonhuman primates tend to have high protein-to-fat ratios and are important sources of minerals that may not otherwise be included in the diet (Raubenheimer and Rothman 2013), making them a valuable resource for extant primates and a possible driving force in primate evolution. Fig. 1. View largeDownload slide Correlation between primate diets and body size. Figure concept by Kay (1984) and modified in Fleagle (2013), reproduced with permission. Fig. 1. View largeDownload slide Correlation between primate diets and body size. Figure concept by Kay (1984) and modified in Fleagle (2013), reproduced with permission. The Visual Predation Hypothesis posits that insect predation was the adaptive pressure leading to the evolution of the primate visual system and other morphological features (Cartmill 1972, 1992, 2012). A suite of adaptations is associated with insectivorous primates, including molars with crests that are used to masticate insect exoskeletons (Kay 1975), simple guts with low stomach-to-small intestine ratios (Chivers and Hladik 1980), relatively larger home ranges (Clutton Brock and Harvey 1980) and small body size (Kay 1984). The primate visual system and grasping hands have also been suggested as adaptations for preying on insects (Bishop 1964; Cartmill 1972, 1992, 2012). Despite the high nutritional value of insects, there are drawbacks to consuming them. One such drawback is that they are often protected by exoskeletons, which are made up of the structural carbohydrate chitin (Finke 2007). Chitin makes up between 2 and 20% of an insect’s drymatter and is considered to be indigestible by most primates, unless their digestive systems contain chitinolytic enzymes (Rothman et al. 2014). Given the digestive challenges posed by chitinous exoskeletons (Strait and Vincent 1998; Rothman et al. 2014), paired with potentially significant energy and amino acid returns if they are digested (Finke 2007; Rothman et al. 2014), endogenously producing such a chitinolytic enzyme could have important adaptive benefits for insectivorous primates, complementing the dental, behavioral, and morphological adaptations discussed above. Indeed, mice and insectivorous bats have been shown to digest chitin using an enzyme produced in the stomach called acidic mammalian chitinase (AMCase) (Whitaker et al. 2004; Strobel et al. 2013; Ohno et al. 2016). This chitinolytic digestive enzyme is produced in the gastric chief cells, where other digestive enzymes are also secreted (Strobel et al. 2013; Ohno et al. 2016). Studies in mice further showed that AMCase is resistant to degradation by the proteases found in the stomach, such as pepsin C, trypsin and chymotrypsin, and breaks down chitin in the presence of these enzymes, as well as at an acidic pH (Ohno et al. 2016). Even though chitinolytic activity has also been observed in the gastric juices of two primates (Perodicticus potto and Cebus capucinus) (Cornelius et al. 1976; Jeuniaux and Cornelius 1978), it was long believed that primates (and most other mammals) did not produce such an enzyme and could not digest chitin (Cork and Kenagy 1989; Oftedal et al. 1991; Simunek and Bartonova 2005; Strobel et al. 2013; Ohno et al. 2016). Instead, it was thought that insect-eating primates had fast gut-transit times to quickly pass indigestible exoskeletons (Gaulin 1979; Milton 1984). More recently, one study found that some primates harbor chitin-digesting microbes (Macdonald et al. 2014) and another study identified an acidic mammalian chitinase in macaques (Macaca fascicularis) that is expressed in the stomach and effectively digests chitin at an acidic pH (Krykbaev et al. 2010). However, because macaques are not very insectivorous and in humans AMCase has been associated with type-2 immune response, such as asthma, allergies, eye diseases, and parasite defense (Zhu et al. 2004; Reese et al. 2007; Musumeci et al. 2009; Bucolo et al. 2011; Muzzarelli et al. 2012; Vannella et al. 2016), any potential benefit of AMCase for insectivorous primates remains unresolved. AMCase is encoded by the gene CHIA or one of its paralogs. In primates, two functional CHIA genes have been identified (Krykbaev et al. 2010): hCHIA and mCHIA. Although hCHIA remains functional in humans, mCHIA has a premature stop codon. In macaques, this state is reversed and mCHIA is functional, whereas hCHIA has a premature stop codon (Krykbaev et al. 2010). Here, we present data on the acidic mammalian chitinase gene (CHIA) in a large comparative sample of nonhuman primates (n = 34) and one treeshrew (Tupaia chinensis). Primates are an order of mammals with over 230 species that have a range of different dietary ecologies (Groves 2001). Across primates, insect consumption varies from practically 0% (e.g., colobine monkeys, which are limited to ingesting insects incidentally when eating leaves or fruit) to almost 100% (Tarsius spp. are completely faunivorous, with most of their diet consisting of arthropods), making them an ideal group for a comparative study of dietary adaptations associated with insectivory. We investigate the potential of AMCase as a digestive enzyme adaptation for insectivory by comparing the paralogous CHIA gene sequences across primates with different levels of insectivory. We also test the strength of selection on CHIA as a function of insect consumption. Finally, we further explicate the relationship between insectivory and body size, as proposed by Kay’s threshold (Kay 1984). Results We successfully sequenced two CHIA genes in 12 primate species for which whole-genome sequences are not available. The sequences we generated for Callithrix jacchus contained numerous differences to the reference genome sequence. Since our sequences were more parsimonious in the comparative context of our study, both compared with sequences generated by us and the whole-genome sequences of closely related species, we believe that our sequences are most likely accurate. All sequences have been deposited in GenBank (accession numbers in supplementary table 1, Supplementary Material online). We further found both homologous sequences (mCHIA and hCHIA) in the 23 genomes we surveyed, with the exception of Microcebus murinus in which we could only positively identify one complete CHIA sequence (hCHIA). In the tarsier (Tarsius syrichta), galago (Otolemur garnettii), and Chinese treeshrew (Tupaia chinensis) genomes we identified >2 CHIA sequences. Both the galago and the Chinese treeshrew genomes had a third CHIA gene sequence (CHIA3), whereas the tarsier genome included three additional CHIA sequences (CHIA3, CHIA4, CHIA5) for a total of five homologous genes. Phylogenetic trees generated from an alignment of the coding sequences (fig. 2) show a deep split between the CHIA genes, indicating that mCHIA and hCHIA (and likely CHIA3) arose in a duplication event that was ancestral to primates and treeshrews and are not independently duplicated genes. Although we did not find complete CHIA3 sequences in any primates other than the tarsier and galago, partial sequences were identified in some genomes. Fig. 2. View largeDownload slide Evolutionary relationships of the CHIA genes in primates. The tree was inferred with PhyML (Guindon et al. 2010) using the HKY85 nucleotide substitution model. Node labels indicate percent bootstrap support (1000 replicates) and branches are scaled by number of substitutions per site. Tree is rooted at the midpoint. Because this tree is based only on the CHIA loci, not all relationships are resolved in a way that is consistent with primate phylogeny. Notably, the placements of Tupaia chinensis mCHIA and CHIA3, and Old World monkey hCHIA do not reflect the likely organismal relationships among some taxa. Fig. 2. View largeDownload slide Evolutionary relationships of the CHIA genes in primates. The tree was inferred with PhyML (Guindon et al. 2010) using the HKY85 nucleotide substitution model. Node labels indicate percent bootstrap support (1000 replicates) and branches are scaled by number of substitutions per site. Tree is rooted at the midpoint. Because this tree is based only on the CHIA loci, not all relationships are resolved in a way that is consistent with primate phylogeny. Notably, the placements of Tupaia chinensis mCHIA and CHIA3, and Old World monkey hCHIA do not reflect the likely organismal relationships among some taxa. CHIA Pseudogenizations Although all species except the mouse lemur (Microcebus murinus) had two complete CHIA sequences, one of these sequences often contained frameshift-causing indels or nonsense mutations leading to premature stop codons and likely rendering the gene nonfunctional. As a result, most primates only retain one full-length, and likely functional CHIA paralog (fig. 3). With some exceptions, hCHIA remains functional only in apes, whereas only mCHIA is functional in most monkeys. Two species in our data set did not retain any functional CHIA genes. In Rhinopithecus bieti and Nasalis larvatus both the mCHIA and hCHIA sequences contained premature stop codons (figs. 3 and 4a). Several New World primates, on the other hand, retained full-length sequences of both mCHIA and hCHIA, including Callithrix jacchus, Cebus capucinus, Saimiri boliviensis, S. sciureus, and Saguinus fuscicollis (figs. 3 and 4b). All three CHIA sequences in the galago and treeshrew, and all five sequences in the tarsier were free from any indels or premature stop codons (figs. 3 and 4c and d). Fig. 3. View largeDownload slide Evolutionary relationships as inferred from CHIA sequences, including timing of CHIA pseudogenization events. *The pseudogenizing mutation found in the Miopithecus ogouensis hCHIA sequence is the same as the one found in the Papionini (Macaca spp., Papio spp, etc.), but is not found in the other Cercopithecini. It is unclear what accounts for this unexpected pattern, but possible explanations include ancestral polymorphism or ancient hybridization. Fig. 3. View largeDownload slide Evolutionary relationships as inferred from CHIA sequences, including timing of CHIA pseudogenization events. *The pseudogenizing mutation found in the Miopithecus ogouensis hCHIA sequence is the same as the one found in the Papionini (Macaca spp., Papio spp, etc.), but is not found in the other Cercopithecini. It is unclear what accounts for this unexpected pattern, but possible explanations include ancestral polymorphism or ancient hybridization. Fig. 4. View largeDownload slide Most primate species have one full-length CHIA sequence, with some exceptions. (a) In the colobine monkeys Rhinopithecus bieti and Nasalis larvatus both mCHIA and hCHIA sequences have a premature stop codon. (b) In some New World monkeys (in order: Cebus capucinus, Callithrix jacchus, Saguinus fuscicollis, Saimiri sciureus, S. boliviensis) both mCHIA and hCHIA are full-length. (c) The treeshrew (Tupaia chinensis) and the Northern greater galago (Otolemur garnettii), two insectivores, have three full-length CHIA sequences, whereas (d) the tarsier (Tarsius syrichta), the most insectivorous of all primates, has a total of five CHIA genes. Photos by (in order) Israel Didham (with pers. permission), Charles J. Sharp, Steven G. Johnson, Manfred Werner, Marie de Carne, Dave Pape, Julie Langford, JJ Harrison, Milan Kořínek (with pers. permission), and Pierre Fidenci; reproduced with permission via Wikimedia Commons unless otherwise noted. Fig. 4. View largeDownload slide Most primate species have one full-length CHIA sequence, with some exceptions. (a) In the colobine monkeys Rhinopithecus bieti and Nasalis larvatus both mCHIA and hCHIA sequences have a premature stop codon. (b) In some New World monkeys (in order: Cebus capucinus, Callithrix jacchus, Saguinus fuscicollis, Saimiri sciureus, S. boliviensis) both mCHIA and hCHIA are full-length. (c) The treeshrew (Tupaia chinensis) and the Northern greater galago (Otolemur garnettii), two insectivores, have three full-length CHIA sequences, whereas (d) the tarsier (Tarsius syrichta), the most insectivorous of all primates, has a total of five CHIA genes. Photos by (in order) Israel Didham (with pers. permission), Charles J. Sharp, Steven G. Johnson, Manfred Werner, Marie de Carne, Dave Pape, Julie Langford, JJ Harrison, Milan Kořínek (with pers. permission), and Pierre Fidenci; reproduced with permission via Wikimedia Commons unless otherwise noted. Interestingly, across our sample of primate species (n = 34), premature stop codons were independently introduced into the hCHIA or mCHIA sequences numerous times, through frameshift or nonsense mutations at various sites along the sequence. The mCHIA gene lost function independently at least six times in primates: three times in the apes and three times in the colobine monkeys (fig. 3). Premature stop codons in the hCHIA gene arose at least seven times: three times in New World monkeys, twice in colobine monkeys, and two (possibly three) times in cercopithecine monkeys (fig. 3), a subfamily that includes the tribes Cercopithecini and Papionini (table 1). The Papionini (Macaca spp, Mandrillus leucophaeus, Cercocebus atys, and Papio anubis) share a deletion in exon 8 that causes a frameshift and premature stop codon, while most of the Cercopithecini (Cercopithecus mitis, Allenopithecus nigroviridis, Allochrocebus lhoesti, Erythrocebus patas, and Chlorocebus spp.) share a frameshifting deletion and premature stop codon at the beginning of exon 11. Even though Miopithecus ogouensis is considered to be part of the tribe Cercopithecini (Tosi et al. 2002, 2005), this species shares the exon 8 deletion with the Papionini and lacks the exon 11 deletion characteristic of its tribe (fig. 3). Exons 8 and 11 were sequenced again in another lab, using a different Miopithecus sample, confirming these results. At this time, it is unclear what accounts for this pattern. Possible explanations include polymorphism in the common ancestor of the Cercopithecini and Papionini, or ancient hybridization. Table 1. Species Included in This Study with Annual Average Insect Consumption and Average Body Weight of Adult Females in Grams.a Species  Common Name  Average Insect Consumption (%)  Average Body Weight (Adult Female, in g)  Aotus nancymaae  Nancy Ma’s night/owl monkey  <15  780  Callicebus moloch  Red-bellied titi monkey  12  956  Callithrix jacchus  Common marmoset  7.2  381  Saguinus fuscicollis  Saddleback tamarin  28.3  358  Cebus capucinus  White-faced capuchin  31.4  2540  Saimiri boliviensis  Black-capped squirrel monkey  no wild data  711  Saimiri sciureus  Common squirrel monkey  53.4  662  Sapajus apella  Tufted capuchin  32.6  2520  Allenopithecus nigroviridis  Allen’s swamp monkey  9  3180  Allochrocebus lhoesti  L’Hoest’s monkey  8.8  3450  Cercocebus atys  Sooty mangabey  26  6200  Cercopithecus mitis  Blue monkey  17.5  4250  Chlorocebus aethiops  Grivet  15.4  2980  Chlorocebus sabaeus  Green monkey  15.4  3300  Erythrocebus patas  Patas monkey  23.5  6500  Macaca fascicularis  Long-tailed macaque  4.1  3590  Macaca mulatta  Rhesus macaque  0  5370  Macaca nemestrina  Pig-tailed macaque  12.2  6500  Mandrillus leucophaeus  Drill  no wild data  12500  Miopithecus talapoin  Talapoin  35  1120  Papio anubis  Olive baboon  2  13300  Colobus angolensis  Tanzanian black&white colobus  0  6935  Colobus guereza  Guereza  0  8550  Nasalis larvatus  Proboscis monkey  0  9820  Rhinopithecus bieti  Black snub-nosed monkey  0  9960  Rhinopithecus roxellana  Golden snub-nosed monkey  0  11600  Nomascus leucogenys  Northern white-cheeked gibbon  4  7320  Gorilla gorilla gorilla  Western lowland gorilla  7.7  71500  Pan paniscus  Bonobo  2  33200  Pan troglodytes  Common chimpanzee  6.4  41600  Pongo abelii  Sumatran orangutan  11.1  35600  Tarsius syrichta  Philippine tarsier  90  117  Microcebus murinus  Gray mouse lemur  8  63  Otolemur garnettii  Northern greater galago  50  734  Tupaia chinensis  Northern treeshrew  >50  200  Species  Common Name  Average Insect Consumption (%)  Average Body Weight (Adult Female, in g)  Aotus nancymaae  Nancy Ma’s night/owl monkey  <15  780  Callicebus moloch  Red-bellied titi monkey  12  956  Callithrix jacchus  Common marmoset  7.2  381  Saguinus fuscicollis  Saddleback tamarin  28.3  358  Cebus capucinus  White-faced capuchin  31.4  2540  Saimiri boliviensis  Black-capped squirrel monkey  no wild data  711  Saimiri sciureus  Common squirrel monkey  53.4  662  Sapajus apella  Tufted capuchin  32.6  2520  Allenopithecus nigroviridis  Allen’s swamp monkey  9  3180  Allochrocebus lhoesti  L’Hoest’s monkey  8.8  3450  Cercocebus atys  Sooty mangabey  26  6200  Cercopithecus mitis  Blue monkey  17.5  4250  Chlorocebus aethiops  Grivet  15.4  2980  Chlorocebus sabaeus  Green monkey  15.4  3300  Erythrocebus patas  Patas monkey  23.5  6500  Macaca fascicularis  Long-tailed macaque  4.1  3590  Macaca mulatta  Rhesus macaque  0  5370  Macaca nemestrina  Pig-tailed macaque  12.2  6500  Mandrillus leucophaeus  Drill  no wild data  12500  Miopithecus talapoin  Talapoin  35  1120  Papio anubis  Olive baboon  2  13300  Colobus angolensis  Tanzanian black&white colobus  0  6935  Colobus guereza  Guereza  0  8550  Nasalis larvatus  Proboscis monkey  0  9820  Rhinopithecus bieti  Black snub-nosed monkey  0  9960  Rhinopithecus roxellana  Golden snub-nosed monkey  0  11600  Nomascus leucogenys  Northern white-cheeked gibbon  4  7320  Gorilla gorilla gorilla  Western lowland gorilla  7.7  71500  Pan paniscus  Bonobo  2  33200  Pan troglodytes  Common chimpanzee  6.4  41600  Pongo abelii  Sumatran orangutan  11.1  35600  Tarsius syrichta  Philippine tarsier  90  117  Microcebus murinus  Gray mouse lemur  8  63  Otolemur garnettii  Northern greater galago  50  734  Tupaia chinensis  Northern treeshrew  >50  200  a Primate body weight data from (Smith and Jungers 1997); only data from wild primates were used. Tupaia chinensis data from PanTHERIA (Jones et al. 2009). Detailed dietary data and references can be found in supplementary table 3, Supplementary Material online. Signatures of Catalytically Active Chitinases In all of the full-length CHIA amino acid sequences we found that the signatures of catalytically active chitinases were conserved (fig. 5): these have a conserved glutamate and the consensus sequence DXXDXDXE at the active site (Synstad et al. 2004; Krykbaev et al. 2010). In addition, catalytically active chitinases further have a chitin-binding domain at the C-terminus, containing six cysteines, which are essential for attaching the enzyme to the chitin (Tjoelker et al. 2000). All of the full-length CHIA sequences in our study had conserved chitin-binding domains that contained all six cysteines (fig. 5). Fig. 5. View largeDownload slide Partial amino acid sequence alignment of functional CHIA genes. The conserved motif for the chitinase catalytic site (DXXDXDXE) from residue 134 to 141 and the chitin-binding domain from residue 440 to 490 are shown. The cysteines in the chitin-binding domain are highlighted. Fig. 5. View largeDownload slide Partial amino acid sequence alignment of functional CHIA genes. The conserved motif for the chitinase catalytic site (DXXDXDXE) from residue 134 to 141 and the chitin-binding domain from residue 440 to 490 are shown. The cysteines in the chitin-binding domain are highlighted. CODEML Analyses Within the CODEML program in the PAML package (Yang 2007), we used three pairs of site models to test whether any of the sites in mCHIA or hCHIA are subject to positive selection. Results of the CODEML analysis for mCHIA suggested that all sites in this gene are under purifying or neutral selection (table 2). Comparison of the models M0 and M3 falsified the null hypothesis that the same dN/dS ratio (ω) applies to all sites of the mCHIA gene (χ2 = 79.54; df = 4; P < 0.00001); however, this result is expected for most functional proteins. Interestingly, CODEML estimated all three discrete ω groups below 1.00, and it placed 75.5% of sites into categories that have a dN/dS ratio of 0.09. Comparison of models M1a and M2a (χ2 = 0, df = 2, P = 1) and models M7 and M8 (χ2 = 2.85, df = 2, P = 0.24) both failed to support any hypothesis of positive selection. Overall, these models indicate that most mCHIA codons appear to be under purifying selection with a smaller number of sites under neutral selection (16–19%, table 2). Table 2. CodeML Results.a Gene  Model  ln(L)  Parameter Estimates  Test  LR  P Value  Positively Selected Sites  mCHIA  M0  −5377.772  ω = 0.251, k = 3.835          M1a  −5338.362  k = 3.907;          ω0 = 0.114 (81.12%); ω1 = 1.00 (18.87%)  M2a  −5338.362  k = 3.907;  M1–M2  0      ω0 = 0.114 (81.12%); ω1 = 1.00 (13.69%); ω2 = 1.00 (5.19%)  M3  −5338.001  k = 3.861;  M0–M3  79.542  < 0.001    ω0 = 0.1 (77.85%); ω1 = 0.354 (0.0002%); ω2 = 0.866 (22.14%)  M7  −5339.595  k = 3.830, α = 0.348; β = 0.965          M8  −5338.170  k = 3.872, α = 1.341; β = 8.405;  M7–M8  2.851  0.240    p0 = 0.841; ωs = 1.00 (15.89%)  hCHIA  M0  −6138.723  ω = 0.320, k =  4.132          M1a  −6079.211  k = 4.177;          ω0 = 0.140 (77.26%); ω1 = 1.00 (22.74%)  M2a  −6074.863  k = 4.283;  M1–M2  8.696  0.013    ω0 = 0.146 (77.36%); ω1 = 1.00 (21.68%); ω2 = 4.327 (0.96%)  M3  −6074.675  k = 4.219;  M0–M3  128.097  < 0.001    ω0 = 0.00 (31.07%); ω1 = 0.345 (60.07%); ω2 = 1.639 (8.86%)  M7  −6081.171  k = 4.160, α = 0.369; β = 0.758          M8  −6073.789  k = 4.235, α = 0.457; β = 1.025;  M7–M8  14.763  0.001  36P(0.953*), 62Q(0.971*), 164R(0.915), 280H(0.941)  p0 = 0.985; ω = 3.503 (1.52%)  Gene  Model  ln(L)  Parameter Estimates  Test  LR  P Value  Positively Selected Sites  mCHIA  M0  −5377.772  ω = 0.251, k = 3.835          M1a  −5338.362  k = 3.907;          ω0 = 0.114 (81.12%); ω1 = 1.00 (18.87%)  M2a  −5338.362  k = 3.907;  M1–M2  0      ω0 = 0.114 (81.12%); ω1 = 1.00 (13.69%); ω2 = 1.00 (5.19%)  M3  −5338.001  k = 3.861;  M0–M3  79.542  < 0.001    ω0 = 0.1 (77.85%); ω1 = 0.354 (0.0002%); ω2 = 0.866 (22.14%)  M7  −5339.595  k = 3.830, α = 0.348; β = 0.965          M8  −5338.170  k = 3.872, α = 1.341; β = 8.405;  M7–M8  2.851  0.240    p0 = 0.841; ωs = 1.00 (15.89%)  hCHIA  M0  −6138.723  ω = 0.320, k =  4.132          M1a  −6079.211  k = 4.177;          ω0 = 0.140 (77.26%); ω1 = 1.00 (22.74%)  M2a  −6074.863  k = 4.283;  M1–M2  8.696  0.013    ω0 = 0.146 (77.36%); ω1 = 1.00 (21.68%); ω2 = 4.327 (0.96%)  M3  −6074.675  k = 4.219;  M0–M3  128.097  < 0.001    ω0 = 0.00 (31.07%); ω1 = 0.345 (60.07%); ω2 = 1.639 (8.86%)  M7  −6081.171  k = 4.160, α = 0.369; β = 0.758          M8  −6073.789  k = 4.235, α = 0.457; β = 1.025;  M7–M8  14.763  0.001  36P(0.953*), 62Q(0.971*), 164R(0.915), 280H(0.941)  p0 = 0.985; ω = 3.503 (1.52%)  a According to our hypothesis-testing framework, no mCHIA sites were found to be under positive selection; the vast majority (>75%) appear to be under purifying selection. Regarding hCHIA, models assuming positive selection outperformed null models, with 1.0–1.5% of sites found to be under positive selection. Results for the same analyses for the hCHIA gene indicated that a small number of sites (1.0–1.5%) may be subject to positive selection (table 2). Both the comparison of models M1a and M2a (χ2 = 8.70, df = 2, P = 0.013) and of models M7 and M8 (χ2 = 14.76, df = 2, P = 0.001) favored the hypothesis that sites in hCHIA are under positive selection over the null hypothesis. Bayes Empirical Bayes analysis indicated two sites that had a significant probability of being under positive selection, 36 P and 62Q (table 2). As with the mCHIA gene, the majority of sites in hCHIA appear to be under purifying selection. RELAX Analyses Although CODEML results suggested that most sites in mCHIA and hCHIA are under purifying selection, we also tested whether the strength of purifying selection acting on these sites varies across different branches of our phylogeny using the program RELAX. Given two categories of branches within a phylogeny, a set of test branches and a set of “background” or “reference” branches, RELAX tests whether the strength of selection was relaxed or intensified in one of these sets compared with the other (Wertheim et al. 2015). For hCHIA, RELAX results supported the hypothesis that selection was relaxed in species in which the gene has become pseudogenized (k = 0.02, P = 0, LR = 74.26, table 3). Compared with branches in which hCHIA remains functional, the ω values of pseudogene branches were shifted towards neutrality (ω = 1) indicating relaxed selection in these species (fig. 6a). Branch specific inferences of the selection intensity parameter (k) under the General Descriptive model in RELAX showed that the branches under more intense selection are ones with functional hCHIA genes, such as Saimiri sciureus (k = 1.99), Otolemur garnettii (k = 2.07), Tupaia chinensis (k = 2.47), and Cebus capucinus (k = 4.74) (fig. 7a). These species also have some of the highest insect intakes in our sample (table 1). Results of a RELAX test including only functional hCHIA sequences supported the hypothesis that hCHIA is under more intense selection in species with higher insect consumption than in species with lower insect consumption (k = 0.20, P < 0.001, LR = 26.6, table 3). The ω values of species with lower insect consumption were shifted toward neutrality compared with those of species with higher insect intake (fig. 6b), but the majority of sites remained below ω = 1 (0.165, 78%). We found similar results for our test of Kay’s threshold (fig. 6c). Selection on hCHIA was relaxed in species with body weights above this 500 g threshold (k = 0.67, LR = 7.68, P = 0.006, table 3). Table 3. RELAX Results.a Gene  Test Branches  Reference Branches  Model  log L  AICc  LR  P Value  mCHIA  Pseudogenes  Functional genes  Null  −5892.80  11956.49      Alternative  −5876.86  11926.63  31.88  <0.001  Partitioned Exploratory  −5866.18  11913.36  53.24  <0.001  Below average insect consumption  Above average insect consumption  Null  −4778.17  9691.10      Alternative  −4769.42  9675.61  17.50  <0.001  Partitioned Exploratory  −4764.26  9673.38  27.82  <0.001  Above Kay’s threshold  Below Kay’s threshold  Null  −4778.17  9691.10      Alternative  −4767.75  9672.29  20.84  <0.001  Partitioned Exploratory  −4765.17  9675.22  26.00  <0.001  hCHIA  Pseudogenes  Functional genes  Null  −6824.57  13820.01      Alternative  −6787.44  13747.78  74.26  <0.001  Partitioned Exploratory  −6787.46  13755.90  74.22  <0.001  Below average insect consumption  Above average insect consumption  Null  −4795.67  9685.95      Alternative  −4782.37  9661.39  26.60  <0.001  Partitioned Exploratory  −4781.84  9668.44  27.66  <0.001  Above Kay’s threshold  Below Kay’s threshold  Null  −4795.68  9685.98      Alternative  −4791.84  9680.33  7.68  0.006  Partitioned Exploratory  −4788.79  9682.33  13.78  0.017  Gene  Test Branches  Reference Branches  Model  log L  AICc  LR  P Value  mCHIA  Pseudogenes  Functional genes  Null  −5892.80  11956.49      Alternative  −5876.86  11926.63  31.88  <0.001  Partitioned Exploratory  −5866.18  11913.36  53.24  <0.001  Below average insect consumption  Above average insect consumption  Null  −4778.17  9691.10      Alternative  −4769.42  9675.61  17.50  <0.001  Partitioned Exploratory  −4764.26  9673.38  27.82  <0.001  Above Kay’s threshold  Below Kay’s threshold  Null  −4778.17  9691.10      Alternative  −4767.75  9672.29  20.84  <0.001  Partitioned Exploratory  −4765.17  9675.22  26.00  <0.001  hCHIA  Pseudogenes  Functional genes  Null  −6824.57  13820.01      Alternative  −6787.44  13747.78  74.26  <0.001  Partitioned Exploratory  −6787.46  13755.90  74.22  <0.001  Below average insect consumption  Above average insect consumption  Null  −4795.67  9685.95      Alternative  −4782.37  9661.39  26.60  <0.001  Partitioned Exploratory  −4781.84  9668.44  27.66  <0.001  Above Kay’s threshold  Below Kay’s threshold  Null  −4795.68  9685.98      Alternative  −4791.84  9680.33  7.68  0.006  Partitioned Exploratory  −4788.79  9682.33  13.78  0.017  a Model-fits of null, alternative, and partitioned exploratory models inferred by the program RELAX (Wertheim et al. 2015). For each CHIA paralog, selective patterns were compared between species with 1) pseudogenes and functional genes, 2) below and above average insect consumption, and 3) body weights above and below Kay’s threshold (500 g). Fig. 6. View largeDownload slide Patterns of natural selection across mCHIA and hCHIA. The best fitting model (as determined by AICc) for each RELAX analysis is shown. Three ω parameters and the percentage of sites they represent are plotted for test (blue/dark) and reference (red/light) branches. The vertical gray and dashed line at ω = 1 indicates neutral evolution. Asterisks indicate significant differences (P < 0.05) between test and reference branches. Fig. 6. View largeDownload slide Patterns of natural selection across mCHIA and hCHIA. The best fitting model (as determined by AICc) for each RELAX analysis is shown. Three ω parameters and the percentage of sites they represent are plotted for test (blue/dark) and reference (red/light) branches. The vertical gray and dashed line at ω = 1 indicates neutral evolution. Asterisks indicate significant differences (P < 0.05) between test and reference branches. Fig. 7. View largeDownload slide Branch specific relaxation parameters inferred for (a) hCHIA and (b) mCHIA genes under the General Descriptive model in RELAX (Wertheim et al. 2015). Branches are colored based on the selection intensity parameter k. A higher k value (light/red) indicates intensified selection, whereas a lower k value (dark/blue) indicates relaxed selection. Scale bars indicate number of substitutions per site. Very long branches were truncated (indicated by breaks) to avoid obscuring the variation present in the remaining branches. Branches with extremely high k values are highlighted in gray. Fig. 7. View largeDownload slide Branch specific relaxation parameters inferred for (a) hCHIA and (b) mCHIA genes under the General Descriptive model in RELAX (Wertheim et al. 2015). Branches are colored based on the selection intensity parameter k. A higher k value (light/red) indicates intensified selection, whereas a lower k value (dark/blue) indicates relaxed selection. Scale bars indicate number of substitutions per site. Very long branches were truncated (indicated by breaks) to avoid obscuring the variation present in the remaining branches. Branches with extremely high k values are highlighted in gray. The initial RELAX results for mCHIA suggested a more complex pattern. The RELAX test comparing pseudogene branches to branches with functional mCHIA genes was significant for selection intensification (k = 1.88, P <0.001) acting on pseudogene branches. However, the ω distributions of the best-fitting model, Partitioned Exploratory (LR = 53.24, P = <0.001, table 3), suggested that most sites (99.89%) of mCHIA pseudogenes are shifted toward neutrality (0.88–0.981) compared with most sites (97%) of functional genes (ω = 0.23–0.04). Only a very small number of hCHIA pseudogene sites (0.11%) were pushed far above neutrality (ω = 325) (fig. 6d). Similarly, the RELAX test comparing mCHIA between species with high and low insect consumption indicated intensified selection on branches with lower insect intake (k = 2.06, LR = 17.50, P < 0.001, table 3), but the best-fitting model, Partitioned Exploratory (fig. 6e), suggested a pattern of relaxation. When comparing species above and below Kay’s threshold, RELAX results reject the hypothesis that mCHIA is under relaxed selection in larger-bodied species, suggesting that they are instead under intensified selection (k = 3.81, LR = 20.8, P < 0.001, table 3) with ω values shifted away from neutrality (fig. 6e). Interestingly, branch specific inferences of k under the General Descriptive model show that three branches in our mCHIA sample that were part of the reference branches in all three RELAX tests have very strong selective regimes, two extremely so (fig. 7b). Otolemur garnettii (k = 3.89), Tarsius syrichta (k = 34.59), and Tupaia chinensis (k = 50) all appear to be under intense selection. It is likely that these extreme values and the long branch lengths observed in these species were biasing the RELAX test (Pond SLK, pers. comm.). We therefore ran these models again, leaving the O. garnettii, T. syrichta, and Tu. chinensis branches “unclassified” to exclude them from the hypothesis test and eliminate any bias contributed by these outliers. In support of our predictions, the RELAX test indicated that mCHIA pseudogene branches are under relaxed selection compared with branches with functional genes (k = 0.58, LR = 15.08, P < 0.001), and that selection on mCHIA is relaxed in species with lower insect consumption (k = 0.42, LR = 11.78, P < 0.001). The model comparing selection between smaller-bodied and larger-bodied species could not be retested, because only two species below 500 g remained in our sample after removing the aforementioned species. Discussion Here, we present the first comparative study of CHIA genes in primates, including species with a variety of dietary ecologies and different levels of insectivory. Our findings are consistent with the hypothesis that primates produce acidic mammalian chitinase (AMCase) as a digestive enzyme for the breakdown of insect exoskeletons and that this enzyme is under more intense selection in more insectivorous and smaller-bodied primates. We have identified a number of homologs of the CHIA gene in primates, mCHIA, hCHIA, CHIA3, CHIA4, and CHIA5. Three of these, mCHIA, hCHIA, and CHIA3 were previously identified by Krykbaev et al (2010) and are most likely ancestral to the Euarchonta, as they are also found in the treeshrew (Tupaia chinensis) (figs. 2 and 3). Because we could only identify a complete and putatively functional sequence for CHIA3 in the treeshrew, the galago (Otolemur garnettii), and the tarsier (Tarsius syrichta) (fig. 4), we conclude that CHIA3 was likely lost early in the anthropoid lineage (fig. 3). This loss is consistent with our current understanding of the ecological changes that are associated with the emergence of crown anthropoids. These changes include a shift to a diurnal lifestyle, an increase in body size, and a greater dietary reliance on fruit, rather than insects (Ross 1996; Ross 2000; Williams et al. 2010). The additional two genes are only found in the Philippine tarsier (T. syrichta, table 1), a member of the genus Tarsius, the most insectivorous primate taxon (Gursky 2011). These genes likely arose as duplications after the tarsier lineage split from the anthropoids between 60 and 70 Ma (Pozzi et al. 2014; Di Fiore et al. 2015; Kistler et al. 2015), but it is unclear which of the three ancestral CHIA genes gave rise to CHIA4 and CHIA5, as they are most similar to each other and equidistant from tarsier mCHIA, hCHIA, and CHIA3 (fig. 2). With the exception of Microcebus murinus (see Future Directions), the genes mCHIA and hCHIA are found in all species in our sample, but show a pattern of independent pseudogenization events that is consistent with our hypothesis that acidic mammalian chitinase is a digestive enzyme evolved and retained for insectivory. Most primates include at least some insects in their diet (Raubenheimer and Rothman 2013) and we find that most primates have one functional CHIA gene (fig. 3). Only two species in our sample did not retain any functional CHIA genes, the black snub-nosed monkey (Rhinopithecus bieti) and the proboscis monkey (Nasalis larvatus) (figs. 3 and 4a). Both of these monkeys are colobines, a subfamily of primates that are highly folivorous and generally do not include insects in their diet (table 1), with the exception of incidental ingestion of insects attached to leaves or found in fruit. Further, the only species in which >1 CHIA gene remained functional are species with above-average levels of insectivory. This includes some of the New World monkeys, which are generally both smaller-bodied and more reliant on insects than Old World monkeys (Gaulin 1979; Terborgh 1983). In the common marmoset (Callithrix jacchus), saddleback tamarin (Saguinus fuscicollis), the common and Bolivian squirrel monkeys (Saimiri sciureus and S. boliviensis, respectively), and the white-faced capuchin (Cebus capucinus) (fig. 4b), both mCHIA and hCHIA were free of premature stop codons and were conserved in the catalytic site and chitin-binding domain (fig. 5). Finally, the most insectivorous primate species in our sample are also the species with the greatest number of (putatively) functional CHIA genes, the galago and tarsier having three and five functional CHIA paralogs, respectively (fig. 4c and d). Tarsiers are the only primates that are entirely faunivorous (Gursky 2011), feeding mostly on arthropods and sometimes small vertebrates (Niemitz 1984; Gursky 2007). Although there are no gene expression data for the tarsier digestive system, it is noteworthy that, despite consuming arthropods almost exclusively, tarsier feces do not contain visible insect exoskeletons. Instead, their feces are described as “powder-like” (Gursky S, personal communication), suggesting that the exoskeletons ingested by the tarsier are broken down completely during gut transit. Such complete digestion is even more impressive considering that tarsiers scarcely chew their arthropod prey, despite having an extremely wide gape and powerful chewing muscles (Jablonski and Crompton 1994; Jablonski 2003). It has been argued that the sharp, pointed teeth of tarsiers are merely instruments of puncture rather than chewing (Prinz et al. 2003), a concept that elevates the functional importance of enzymatic digestion in the stomach and small intestine. In this light, the unique CHIA duplications we have documented here are likely to be important digestive adaptations for tarsiers. The primates with >1 functional CHIA gene are not just more insectivorous but are also among the smaller species in our sample (table 1), in accordance with what is predicted by Kay’s threshold (fig. 1). Because it is costly for small-bodied primates to fill their relatively smaller guts with indigestible bulk, such as leaves (Gaulin 1979; Fleagle 2013), the ability to efficiently digest insect exoskeletons using AMCase might be an especially valuable adaptation for these species. The results of our selection analyses show that CHIA genes are under intensified purifying selection in more insectivorous species (fig. 7a and b) and also partially support the hypothesis that selection on CHIA genes is relaxed in larger-bodied primates (fig. 6c). While the initial mCHIA RELAX test did not find support for this hypothesis (fig. 6e) and had inconsistent results for the other two tests (pseudogenes vs. functional genes [fig. 6c], above-average vs. below-average insectivory [fig. 6d]) it is worth noting that, in accordance with our predictions, mCHIA appears to be under extreme selection in the three most insectivorous species (Tupaia chinensis, Tarsius syrichta, and Otolemur garnettii). Further, it is likely that the extremely high k values and long branch lengths of these lineages (fig. 7b) masked other selective patterns across the alignment and biased the RELAX model (Pond, personal communication). This is supported by the results of additional analyses in which Tu. chinensis, T. syrichta, and O. garnettii were left “unclassified” and not included in the reference branches. Here, we found support for the hypotheses that mCHIA is under relaxed selection in 1) pseudogene branches and 2) less insectivorous species. Overall, these results suggest that having multiple functional CHIA paralogs is an adaptive advantage for insect-eating primates. Having multiple CHIA genes may correspond to increased AMCase production, as has been found for other digestive enzymes (Perry et al. 2007; Axelsson et al. 2013) or, alternatively, the different CHIA paralogs may have diverged in function (Zhang 2003), potentially specializing in the digestion of various types of arthropod prey. The exoskeletons of arthropods are not made up of pure chitin, but rather of a matrix of chitin and other compounds (Finke 2007), the exact composition of which varies across species (No et al. 1989; Kramer et al. 1995; Tomberlin et al. 2002; Nation 2015). Considering the diversity of arthropods consumed by primates (Raubenheimer and Rothman 2013; McGrew 2014; O'Malley and McGrew 2014), it is conceivable that having multiple CHIA paralogs does not merely increase the amount of AMCase secreted, but leads to the expression of different specialized forms of the enzyme. Our results may also shed some light on primate origins. Two commonly cited and competing hypotheses to explain the evolution of primate features, such as the visual system and grasping hands, are the Visual Predation Hypothesis (Cartmill 1972, 1992, 2012) and the Terminal Branch Feeding Hypothesis (Sussman 1991). According to the former, primate vision and hands evolved to facilitate the predation on insects, whereas the latter proposes that primate features evolved to allow for the exploitation of angiosperms at the ends of terminal branches. Although these two hypotheses do not have to be mutually exclusive, our results show that early primates likely had three CHIA genes, suggesting that insects were an important component of the ancestral primate diet, which favors the Visual Predation Hypothesis. Our results complement a recent study on primate opsin genes which found a shift to blue sensitivity in the SWS1 opsin of early primates, suggesting a lifestyle that required enhanced visual acuity (Melin et al. 2016). Such a shift is indicative of a need to resolve cryptic or fast-moving prey, rather than angiosperm reproductive tissues (Melin et al. 2016), and, like our results, supports the hypothesis that insects were an important food source for early primates. Limitations and Future Directions We were unable to confirm complete CHIA sequences in most of the currently available lemur genomes. In the mouse lemur (Microcebus murinus) genome we identified the full hCHIA sequence and found partial sequences of one, or possibly two, similar genes. However, we could not identify all exons of these additional genes with confidence, an issue that was also encountered with other lemur genomes. We therefore chose to focus on other taxa, but look forward to revisiting the evolutionary history of chitinase genes in the strepsirrhines as additional high-quality genomes become available. Similar issues were encountered with two other, nonprimate genomes, the Sunda colugo (Galeopterus variegatus) and the pen-tailed treeshrew (Ptilocercus lowii) (Mason et al. 2016). In the colugo, this is likely because the genes have diverged substantially, whereas the low-coverage of the genome may have been an issue in the case of Ptilocercus. Given that the diet of the Sunda colugo consists mainly of leaves (Dzulhelmi and Abdullah 2009), it is perhaps unsurprising that the CHIA gene sequence is not conserved in this species. One limitation of the current study is the available data on primate diets. During our review of the literature, we encountered many inconsistencies in the ways that data on food intake are collected, making it difficult to compare levels of insect consumption across species and even populations. In addition, identifying and observing insect-feeding is notoriously difficult (Frazão 1991), so it is very likely that insect consumption in many primates, and especially arboreal primates, is underestimated. Like exoskeletons, the cell walls of fungi and lichens are composed of chitin (Stevens and Hume 1995). Several primate species feed on fungi, most notably Callimico goeldii (up to 29% of feeding time) (Porter 2001; Hanson et al. 2006), and lichens are a staple for many colobines (Kirkpatrick et al. 2001; Fashing et al. 2007). Although it was outside of the scope of the present study, the relationship between AMCase and mycophagy deserves to be examined in future studies. Finally, without gene expression data for primate digestive systems, we cannot be certain whether the various CHIA genes are actually expressed in the stomachs of all primates. Krykbaev and colleagues report that mCHIA is highly expressed in the stomach of Macaca fascicularis, but expression data for humans suggests that the role of hCHIA may be more complicated. Although some studies show that hCHIA is expressed in the human stomach (Boot et al. 2005; Krykbaev et al. 2010), there is disagreement over whether the gene actually translates into a functional chitinase in the stomach. One study found that chitinolytic activity in the gastric juice was present in 80% of their participants, but absent in the other 20% (Paoletti et al. 2007), whereas another study failed to detect any evidence of chitinase in the human digestive system (Goto et al. 2003). It is plausible that this may be due to dietary difference across human populations and invites further study in groups with long histories of insect consumption. Conclusion Even though several studies from the 1970s suggested that primates have chitinolytic enzymes (Cornelius et al. 1976; Jeuniaux and Cornelius 1978; Kay and Sheine 1979), the notion that chitin was indigestible by the endogenous digestive enzymes of primates and other mammals has persisted (Cork and Kenagy 1989; Oftedal et al. 1991; Šimůnek and Bartoňová 2005; Strobel et al. 2013; Ohno et al. 2016). Here we present evidence that suggests insect-eating primates share an adaptation found in insectivorous bats (Vespertilionidae) and mice (Mus musculus) (Strobel et al. 2013; Ohno et al. 2016) and use the enzyme acidic mammalian chitinase to digest the chitin in insect exoskeletons. The efficient digestion of insect exoskeletons is likely to have important adaptive benefits for all insect-eating primates, through the potentially significant energy and amino acid returns from the digestion of the polysaccharide chitin (Finke 2007; Rothman et al. 2014) and, for small-bodied primates, by reducing the amount of indigestible bulk in their guts. Materials and Methods In this study, we collected and analyzed sequence data on the CHIA paralogs of 34 primate species from 27 genera with varying levels of insect consumption (table 1). We mined the published genomes of 22 primates and one treeshrew (Tupaia chinensis) for CHIA-like sequences using BLAST and sequenced the CHIA genes in additional primate species that do not have publicly available genomes. DNA samples for Callicebus moloch, Callithrix jacchus, Saguinus fuscicollis, and Saimiri sciureus were obtained from Coriell Biorepositories (see supplementary table 1, Supplementary Material online). Dr. George Perry provided extracted DNA from Sapajus apella. Extracted DNA from Erythrocebus patas was provided by Dr. Todd Disotell. DNA for the following samples was provided by one of us (AJT): Allenopithecus nigroviridis, Allochrocebus lhoesti, Cercopithecus mitis, Chlorocebus aethiops, Colobus guereza kikuyensis, and Miopithecus ogouensis. Genome Mining We conducted BLAST searches against the whole-genome sequences of the following taxa, using the Macaca fascicularis CHIA gene sequences (Krykbaev et al. 2010) as queries: Callithrix jacchus, Saimiri boliviensis, Aotus nancymaae, Cebus capucinus imitator, Cercocebus atys, Mandrillus leucophaeus, Papio anubis, Macaca mulatta, Macaca nemestrina, Chlorocebus sabaeus, Nasalis larvatus, Rhinopithecus roxellana, Rhinopithecus bieti, Colobus angolensis, Gorilla gorilla gorilla, Pan paniscus, Pan troglodytes, Pongo abelii, Nomascus leucogenys, Tarsius syrichta, Microcebus murinus, Otolemur garnettii, and Tupaia chinensis whole-genome sequences (GenBank accession numbers in supplementary table 1, Supplementary Material online). Amplification, Sequencing, and Assembly We designed PCR primers for the amplification of each of the 11 exons of both CHIA paralogs (mCHIA, hCHIA) in Old World and New World monkeys. We first tried to find regions that were conserved across a wide range of primates, including both platyrrhines and catarrhines, however, it was difficult to find intron regions that were suitable as PCR primer sites. Separate sets of PCR primers therefore had to be designed for catarrhine species and for platyrrhine species for most exons (supplementary table 2, Supplementary Material online). After aligning the CHIA sequences of multiple catarrhines (Macaca fascicularis, Mandrillus leucophaeus, Nasalis larvatus, Cercocebus atys, Papio anubis, Rhinopithecus roxellana, and Chlorocebus sabaeus) and platyrrhines (Saimiri boliviensis, Callithrix jacchus, and Aotus nancymaae), ambiguities in the consensus sequence were masked and primers were designed using Primer3 (http://primer3.ut.ee/). PCRs were carried out with one of two commercial kits. One kit was the Qiagen Fast Cycling PCR Kit MasterMix, which was used following manufacturer’s protocols, but with a proportionally reduced volume of 15 µl. Reactions were run at 95 °C for 5 min followed by 35 cycles of 96 °C for 15 s, 56–59 °C for 15 s, and 68 °C for 35 s, and a final step at 72 °C for 7 min. PCR products were purified with exonuclease I and shrimp alkaline phosphatase (ExoSAP) or, when gels showed the amplification of nontarget sequences (i.e., multiple bands), with the Qiagen QIAquick Gel Extraction Kit. Purified PCR products were sequenced directly on the Applied Biosystems 3500 Genetic Analyzer using the Applied Biosystems BigDye Terminator v3.1 Cycle Sequencing Kit and the same primers as used for PCR. Alternatively, PCRs were carried out in 25 µl with the Promega GoTaq G2 Master Mix and run under the following conditions: 95° for 2 min followed by 35–40 cycles of 95° for 20 s, 55°–60° for 25 s, and 72° for 30–60 s. These cycles were followed by a final extension phase at 72° for 5 min. Some of the exons of the hCHIA gene were difficult to amplify in Co. guereza, Ce. mitis, A. lhoesti, Ch. aethiops, M. ogouensis, and Al. nigroviridis. For these taxa, we used a long range PCR approach with Promega GoTaq Long PCR Master Mix to amplify the entire hCHIA gene (8–9 kb) under the following thermal cycling conditions: 95° for 2 min followed by 35 cycles of 95° for 20 s, 57° for 25 s, and 65° for 18 min. These long-range reactions were followed by a final extension phase of 72° for 10 min. These whole-gene products were then cleaned up using Millipore Microcon centrifugal filters by spinning the products at 500 rcf for 4 min, and then spinning the inverted filter to recover only dsDNA at 1,000 rcf for 3 min. We then diluted the cleaned-up DNA concentrate with 10 µl of nuclease-free water. This step was taken in order to quantitate the DNA so that we could use an appropriate amount of these products for subsequent nested PCRs, which targeted the problematic exons mentioned above. The nested products were sequenced directly at the Molecular Cloning Laboratories (San Francisco, CA) on an ABI 3730XL sequencer. Reads were assembled, mapped to the Macaca fascicularis reference sequence, concatenated into coding sequences, and translated using Geneious v. 9.1.8. Coding regions and translated amino acid sequences were aligned with MAFFT alignment server v. 7. Sequence Analyses Sequences were visually inspected for frameshift mutations causing premature stop codons; these were considered to be CHIA pseudogenes. To determine whether full-length CHIA sequences were likely coding for a functional enzyme, we inspected the translated amino acid sequences for conservation of the catalytic site motif and chitin-binding domain, signatures of functional chitinases (Tjoelker et al. 2000; Krykbaev et al. 2010). Premature stop codons were removed from pseudogene sequences by excising the insertion(s) or deletion(s) that caused the premature stop codon. We looked for evidence of positive selection acting on sites along both CHIA genes in primates with the CODEML program in the PAML package (Yang 2007). For these analyses, premature stop codons were removed from pseudogene sequences by excising the insertion(s) or deletion(s) that caused the premature stop codon. We used site-specific models (M0: null, M1a: nearly neutral selection, M2a: positive selection, M3: discrete, M7: beta, and M8: beta and ω > 1) to determine if there is variation in the ratios of nonsynonymous to synonymous nucleotide substitutions (dN/dS or ω) across sites along the mCHIA and hCHIA alignments and to test for evidence of positively selected sites (Yang 2007). Model fit was evaluated using likelihood ratio tests (LRT). The software program RELAX, which is part of the HyPhy package, tests for relaxed versus intensified selection in a codon-based phylogenetic framework (Wertheim et al. 2015). Given two sets of branches in a phylogeny (foreground and background, or test and reference branches), RELAX tests whether selection is intensified or relaxed in one set versus the other. To do this, RELAX uses a branch-site evolutionary (BS-REL) model to estimate the distribution of ω for each of the two branch sets and then compares this distribution using two models. In the null model the selection intensity parameter k is constrained to 1, causing the ω distribution to be the same on both test and reference branches, while under the alternative model k is allowed to vary. If the latter model is a significantly better fit (as determined by a LRT), this suggests that selection on the test branches is either relaxed (k < 1) or intensified (k > 1) compared with the reference branches (Wertheim et al. 2015). In addition to the null and alternative models, the partitioned exploratory model can provide a quantitative measure of selection patterns in the test and reference branches. The partitioned exploratory model is a less constrained model that allows the proportion of sites in each category of ω to vary between test and reference branches. Here we used RELAX to test whether selection was relaxed or intensified along any of the branches in our phylogeny as a function of either insect consumption or body size. Specifically, we hypothesized that selection on CHIA genes would be relaxed in lineages with 1) pseudogenizing mutations, 2) below average insect consumption, and 3) body size above Kay’s threshold (500 g). Phenotypic Data Data on insect consumption and body size were taken from the literature. Most of the data on body size came from (Smith and Jungers 1997) and only measures from wild specimens were used. To avoid skews introduced by large adult males in sexually dimorphic species we only used adult female body weight measures for all species. Average adult female body weight for each species is presented in table 1. Species-specific information on average annual insect consumption was taken from studies of wild primates, the results of many of which are collected in Campbell et al (2011). Annual averages of insect consumption used in our analyses are presented in table 1, while detailed information on all data used to calculate these averages is given in supplementary table 3, Supplementary Material online. Most of these studies present data on the percent of time spent feeding, and we did not include estimates that included foraging time. A few studies presented data based on stomach contents. Because of the inconsistent ways in which dietary data is collected and presented in the literature, we decided to group species into one of two categories for the analyses, rather than using insect intake as a continuous variable. Species were classified as relatively more or less insectivorous depending on whether their average annual insect consumption was higher or lower than the average annual insect consumption across all species included in our sample (16.58%). Furthermore, where possible we used data from 1) studies that covered at least 12 months to account for seasonal variation in insect consumption, and 2) multiple studies for each species to account for interpopulation differences. The diets of a few of the species included here have not been studied in the wild; in these cases, we used data from a closely related species where available (Saimiri boliviensis and Aotus nancymaae), or where unavailable (Mandrillus leucophaeus) left the branch “unclassified” in the RELAX analyses involving dietary data. Supplementary Material Supplementary data are available at Molecular Biology and Evolution online. Acknowledgments We thank the editors and two anonymous reviewers for their thoughtful comments and suggestions. Dr. Robert Scott provided helpful feedback throughout this study. We thank P.J. Perry, Todd Disotell, and the Coriell Biorepository for providing DNA samples. This work was supported by an Albert Fellows Dissertation Research Award from the Center of Human Evolutionary Studies at Rutgers University, the International Primatological Society, and the Kent State University College of Arts and Sciences. References Axelsson E, Ratnakumar A, Arendt M-L, Maqbool K, Webster MT, Perloski M, Liberg O, Arnemo JM, Hedhammar Å, Lindblad-Toh K. 2013. The genomic signature of dog domestication reveals adaptation to a starch-rich diet. Nature  495( 7441): 360– 364. Google Scholar CrossRef Search ADS PubMed  Bishop A. 1964. Use of the hand in lower primates. In: Buettner-Janush J, editor. Evolutionary and genetic biology of primates . 2nd ed. New York: Elsevier. p. 133– 225. Google Scholar CrossRef Search ADS   Boot RG, Bussink AP, Verhoek M, de Boer PAJ, Moorman AFM, Aerts JMFG. 2005. Marked differences in tissue-specific expression of chitinases in mouse and man. J Histochem Cytochem . 53( 10): 1283– 1292. http://dx.doi.org/10.1369/jhc.4A6547.2005 Google Scholar CrossRef Search ADS PubMed  Bucolo C, Musumeci M, Musumeci S, Drago F. 2011. Acidic mammalian chitinase and the eye: implications for ocular inflammatory diseases. Front Pharmacol.  2: 43. Google Scholar CrossRef Search ADS PubMed  Campbell CJ, Fuentes A, MacKinnon KC, Bearder SK, Stumpf RM, editors. 2011. Primates in perspective . 2nd ed. New York: Oxford University Press Cartmill M. 1972. Arboreal adaptations and the origin of the order Primates. In: Tuttle RH, editor. The functional and evolutionary biology of primates . Chicago: Aldine-Atherton. p. 97– 122. Cartmill M. 1992. New views on primate origins. Evol Anthropol . 1( 3): 105– 111. http://dx.doi.org/10.1002/evan.1360010308 Google Scholar CrossRef Search ADS   Cartmill M. 2012. Primate origins, human origins, and the end of higher taxa. Evol Anthropol . 21( 6): 208– 220. http://dx.doi.org/10.1002/evan.21324 Google Scholar CrossRef Search ADS PubMed  Chivers DJ, Hladik CM. 1980. Morphology of the gastrointestinal tract in primates: comparisons with other mammals in relation to diet. J Morphol.  166( 3): 337– 386. Google Scholar CrossRef Search ADS PubMed  Clutton Brock TH, Harvey PH. 1980. Primates, brains and ecology. J Zool.  190( 3): 309– 323. Google Scholar CrossRef Search ADS   Cork SJ, Kenagy GJ. 1989. Nutritional value of hypogeous fungus for a forest‐dwelling ground squirrel. Ecology  70( 3): 577– 586. Google Scholar CrossRef Search ADS   Cornelius C, Dandrifosse G, Jeuniaux C. 1976. Chitinolytic enzymes of the gastric mucosa of Perodicticus potto (primate prosimian): purification and enzyme specificity. Int J Biochem.  7( 9–10): 445– 448. Google Scholar CrossRef Search ADS   Di Fiore A, Chaves PB, Cornejo FM, Schmitt CA, Shanee S, Cortés-Ortiz L, Fagundes V, Roos C, Pacheco V. 2015. The rise and fall of a genus: complete mtDNA genomes shed light on the phylogenetic position of yellow-tailed woolly monkeys, Lagothrix flavicauda, and on the evolutionary history of the family Atelidae (Primates: Platyrrhini). Mol Phylogenet Evol.  82 Pt B: 495– 510. Google Scholar CrossRef Search ADS PubMed  Dzulhelmi MN, Abdullah MT. 2009. Foraging ecology of the Sunda colugo (Galopterus variegatus) in Bako National Park, Sarawak, Malaysia. Malayan Nat J.  61: 285– 294. Fashing PJ, Mulindahabi F, Gakima J-B, Masozera M, Mununura I, Plumptre AJ, Nguyen N. 2007. Activity and ranging patterns of Colobus angolensis ruwenzorii in Nyungwe Forest, Rwanda: possible costs of large group size. Int J Primatol.  28( 3): 529– 550. Google Scholar CrossRef Search ADS   Finke MD. 2007. Estimate of chitin in raw whole insects. Zoo Biol . 26( 2): 105– 115. http://dx.doi.org/10.1002/zoo.20123 Google Scholar CrossRef Search ADS PubMed  Fleagle JG. 2013. Primate adaptation and evolution . 3rd ed. Amsterdam: Academic Press. Frazão E. 1991. Insectivory in free-ranging bearded saki (Chiropotes satanas chiropotes). Primates  32( 2): 243– 245. Google Scholar CrossRef Search ADS   Gaulin SJ. 1979. A Jarman/Bell model of primate feeding niches. Hum Ecol.  7( 1): 1– 20. http://dx.doi.org/10.1007/BF00889349 Google Scholar CrossRef Search ADS   Gingerich PD. 1980. Dental and cranial adaptations in Eocene Adapidae. Z Morphol Anthropol.  71( 2): 135– 142. Google Scholar PubMed  Goodall J. 1986. The chimpanzees of Gombe: patterns of behavior . Cambridge (MA): Belknap Press of Harvard University Press. Goto M, Fujimoto W, Nio J, Iwanaga T, Kawasaki T. 2003. Immunohistochemical demonstration of acidic mammalian chitinase in the mouse salivary gland and gastric mucosa. Arch Oral Biol . 48( 10): 701– 707. http://dx.doi.org/10.1016/S0003-9969(03)00150-X Google Scholar CrossRef Search ADS PubMed  Groves CP. 2001. Primate taxonomy . Washington (DC): Smithsonian. Guindon S, Dufayard J-F, Lefort V, Anisimova M, Hordijk W, Gascuel O. 2010. New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of PhyML 3.0. Syst Biol . 59( 3): 307– 321. Google Scholar CrossRef Search ADS PubMed  Gursky SL. 2007. The Spectral Tarsier . Abingdon (UK): Routledge. Gursky SL. 2011. Tarsiiformes. In: Campbell CJ, Fuentes A, MacKinnon KC, Bearder SK, Stumpf RM, editors. Primates in perspective . New York: Oxford University Press. p. 79– 90. Hanson AM, Hall MB, Porter LM, Lintzenich B. 2006. Composition and nutritional characteristics of fungi consumed by Callimico goeldii in Pando, Bolivia. Int J Primatol.  27( 1): 323– 346. http://dx.doi.org/10.1007/s10764-005-9014-z Google Scholar CrossRef Search ADS   Isbell LA. 1998. Diet for a small primate: insectivory and gummivory in the (large) patas monkey (Erythrocebus patas pyrrhonotus). Am J Primatol . 45( 4): 381– 398. http://dx.doi.org/10.1002/(SICI)1098-2345(1998)45:4<381::AID-AJP5>3.0.CO;2-S Google Scholar CrossRef Search ADS PubMed  Jablonski NG, Crompton RH. 1994. Feeding behavior, mastication, and tooth wear in the western tarsier (Tarsius bancanus). Int J Primatol.  15( 1): 29– 59. http://dx.doi.org/10.1007/BF02735233 Google Scholar CrossRef Search ADS   Jablonski NG. 2003. The evolution of the tarsiid niche. In: Wright PC, Simons EL, Gursky SL, editors. Tarsiers past, present, and future . New Brunswick (NJ): Rutgers University Press. p. 35– 49. Jeuniaux C, Cornelius C. 1978. Distribution and activity of chitinolytic enzymes in the digestive tract of birds and mammals. In: Muzzarelli RAA, Pariser ER, editors. Proceedings of the first international conference on chitin/chitosan. Boston (MA): Massachusetts Institute of Technology. p 542–549. Jones KE, Bielby J, Cardillo M, Fritz SA, O'Dell J, Orme CDL, Safi K, Sechrest W, Boakes EH, Carbone C. 2009. PanTHERIA: a species‐level database of life history, ecology, and geography of extant and recently extinct mammals. Ecology  90( 9): 2648– 2648. Google Scholar CrossRef Search ADS   Kay RF, Covert HH. 1984. Anatomy and behaviour of extinct primates. In: Chivers DJ, Wood BA, Bilsborough A, editors. Food acquisition and processing in primates . Boston (MA): Springer US. p. 467– 508. Google Scholar CrossRef Search ADS   Kay RF, Sheine WS. 1979. On the relationship between chitin particle size and digestibility in the primate Galago senegalensis. Am J Phys Anthropol . 50( 3): 301– 308. http://dx.doi.org/10.1002/ajpa.1330500303 Google Scholar CrossRef Search ADS   Kay RF. 1975. The functional adaptations of primate molar teeth. Am J Phys Anthropol . 43( 2): 195– 216. http://dx.doi.org/10.1002/ajpa.1330430207 Google Scholar CrossRef Search ADS PubMed  Kay RF. 1984. On the use of anatomical features to infer foraging behavior in extinct primates. In: Cant J, Rodman PS, editors. Adaptations for foraging in nonhuman primates . New York: Columbia University Press. p. 21– 53. Kirkpatrick RC, Zou RJ, Dierenfeld ES, Zhou HW. 2001. Digestion of selected foods by Yunnan snub-nosed monkey Rhinopithecus bieti (Colobinae). Am J Phys Anthropol . 114( 2): 156– 162. Google Scholar CrossRef Search ADS PubMed  Kistler L, Ratan A, Godfrey LR, Crowley BE, Hughes CE, Lei R, Cui Y, Wood ML, Muldoon KM, Andriamialison H, et al.   2015. Comparative and population mitogenomic analyses of Madagascar's extinct, giant ‘subfossil’ lemurs. J Hum Evol.  79: 45– 54. Google Scholar CrossRef Search ADS PubMed  Kramer KJ, Hopkins TL, Schaefer J. 1995. Applications of solids NMR to the analysis of insect sclerotized structures. Insect Biochem Mol Biol.  25( 10): 1067– 1080. http://dx.doi.org/10.1016/0965-1748(95)00053-4 Google Scholar CrossRef Search ADS   Krykbaev R, Fitz LJ, Reddy PS, Winkler A, Xuan D, Yang X, Fleming M, Wolf SF. 2010. Evolutionary and biochemical differences between human and monkey acidic mammalian chitinases. Gene  452( 2): 63– 71. Google Scholar CrossRef Search ADS PubMed  Macdonald C, Barden S, Foley S. 2014. Isolation and characterization of chitin-degrading micro-organisms from the faeces of Goeldi’s monkey, Callimico goeldii. J Appl Microbiol.  116( 1): 52– 59. Google Scholar CrossRef Search ADS PubMed  Mason VC, Li G, Minx P, Schmitz J, Churakov G, Doronina L, Melin AD, Dominy NJ, Lim NTL, Springer MS, et al.   2016. Genomic analysis reveals hidden biodiversity within colugos, the sister group to primates. Sci Adv.  2( 8): e1600633. Google Scholar CrossRef Search ADS PubMed  McGrew WC. 1992. Chimpanzee material culture: implications for human evolution . Cambridge: Cambridge University Press. Google Scholar CrossRef Search ADS   McGrew WC. 2001. The other faunivory: primate insectivory and early human diet. In: Stanford CB, Bunn HT, editors. Meat-eating and human evolution . Oxford: Meat-eating and Human Evolution. p. 160– 178. McGrew WC. 2014. The “other faunivory” revisited: insectivory in human and non-human primates and the evolution of human diet. J Hum Evol.  71: 4– 11. Google Scholar CrossRef Search ADS PubMed  Melin AD, Wells K, Moritz GL, Kistler L, Orkin JD, Timm RM, Bernard H, Lakim MB, Perry GH, Kawamura S, et al.   2016. Euarchontan opsin variation brings new focus to primate origins. Mol Biol Evol . 33( 4): 1029– 1041. Google Scholar CrossRef Search ADS PubMed  Milton K. 1984. The role of food-processing factors in primate food choice. In: Rodman PS, Cant JGH, editors. Adaptations for foraging in nonhuman primates . New York: Columbia University Press. p. 249– 279. Musumeci M, Aragona P, Bellin M, Maugeri F, Rania L, Bucolo C, Musumeci S. 2009. Acidic mammalian chitinase in dry eye conditions. Cornea  28( 6): 667– 672. Google Scholar CrossRef Search ADS PubMed  Muzzarelli RAA, Boudrant J, Meyer D, Manno N, DeMarchis M, Paoletti MG. 2012. Current views on fungal chitin/chitosan, human chitinases, food preservation, glucans, pectins and inulin: a tribute to Henri Braconnot, precursor of the carbohydrate polymers science, on the chitin bicentennial. Carbohydr Polym.  87( 2): 995– 1012. Google Scholar CrossRef Search ADS   Nation JL. 2015. Insect physiology and biochemistry , 3rd ed. Boca Raton (FL): CRC Press. Niemitz C. 1984. Biology of tarsiers . Stuttgart: Gustav Fischer. No HK, Meyers SP, Lee KS. 1989. Isolation and characterization of chitin from crawfish shell waste. J Agric Food Chem.  37( 3): 575– 579. http://dx.doi.org/10.1021/jf00087a001 Google Scholar CrossRef Search ADS   O’Malley RC, McGrew WC. 2014. Primates, insects and insect resources. J Hum Evol.  71: 1– 3. Google Scholar CrossRef Search ADS PubMed  Oftedal OT, Whiten A, Southgate DAT, Van Soest PJ. 1991. The nutritional consequences of foraging in primates: the relationship of nutrient intakes to nutrient requirements. Philos Trans R Soc Lond B Biol Sci . 334: 161– 170. http://dx.doi.org/10.1098/rstb.1991.0105 Google Scholar CrossRef Search ADS PubMed  Ohno M, Kimura M, Miyazaki H, Okawa K, Onuki R, Nemoto C, Tabata E, Wakita S, Kashimura A, Sakaguchi M, et al.   2016. Acidic mammalian chitinase is a proteases-resistant glycosidase in mouse digestive system. Sci Rep . 6: 37756. Google Scholar CrossRef Search ADS PubMed  Paoletti MG, Norberto L, Damini R, Musumeci S. 2007. Human gastric juice contains chitinase that can degrade chitin. Ann Nutr Metab . 51( 3): 244– 251. http://dx.doi.org/10.1159/000104144 Google Scholar CrossRef Search ADS PubMed  Perry GH, Dominy NJ, Claw KG, Lee AS, Fiegler H, Redon R, Werner J, Villanea FA, Mountain JL, Misra R, et al.   2007. Diet and the evolution of human amylase gene copy number variation. Nat Genet.  39( 10): 1256– 1260. Google Scholar CrossRef Search ADS PubMed  Porter LM. 2001. Dietary differences among sympatric Callitrichinae in northern Bolivia: Callimico goeldii, Saguinus fuscicollis and S. labiatus. Int J Primatol.  22: 961– 992. http://dx.doi.org/10.1023/A:1012013621258 Google Scholar CrossRef Search ADS   Pozzi L, Hodgson JA, Burrell AS, Sterner KN, Raaum RL, Disotell TR. 2014. Primate phylogenetic relationships and divergence dates inferred from complete mitochondrial genomes. Mol Phylogenet. Evol.  75: 165– 183. Google Scholar CrossRef Search ADS PubMed  Prinz JF, Silwood CJL, Claxson AWD, Grootveld M. 2003. Simulated digestion status of intact and exoskeletally-punctured insects and insect larvae: a spectroscopic investigation. Folia Primatol.  74( 3): 126– 140. Google Scholar CrossRef Search ADS PubMed  Raubenheimer D, Rothman JM. 2013. Nutritional ecology of entomophagy in humans and other primates. Annu Rev Entomol . 58: 141– 160. http://dx.doi.org/10.1146/annurev-ento-120710-100713 Google Scholar CrossRef Search ADS PubMed  Reese TA, Liang H-E, Tager AM, Luster AD, Van Rooijen N, Voehringer D, Locksley RM. 2007. Chitin induces accumulation in tissue of innate immune cells associated with allergy. Nature  447( 7140): 92– 96. Google Scholar CrossRef Search ADS PubMed  Ross C. 1996. Adaptive explanation for the origins of the anthropoidea (primates). Am J Primatol . 40( 3): 205– 230. http://dx.doi.org/10.1002/(SICI)1098-2345(1996)40:3<205::AID-AJP1>3.0.CO;2-1 Google Scholar CrossRef Search ADS   Ross CF. 2000. Into the light: the origin of anthropoidea. Annu Rev Anthropol . 29( 1): 147– 194. http://dx.doi.org/10.1146/annurev.anthro.29.1.147 Google Scholar CrossRef Search ADS   Rothman JM, Raubenheimer D, Bryer MAH, Takahashi M, Gilbert CC. 2014. Nutritional contributions of insects to primate diets: implications for primate evolution. J Hum Evol.  71: 59– 69. Google Scholar CrossRef Search ADS PubMed  Šimůnek J, Bartoňová H. 2005. Effect of dietary chitin and chitosan on cholesterolemia of rats. Acta Veterinaria Brno.  74( 4): 491– 499. Google Scholar CrossRef Search ADS   Smith RJ, Jungers WL. 1997. Body mass in comparative primatology. J Hum Evol.  32( 6): 523– 559. http://dx.doi.org/10.1006/jhev.1996.0122 Google Scholar CrossRef Search ADS PubMed  Souto A, Bione CBC, Bastos M, Bezerra BM, Fragaszy D, Schiel N. 2011. Critically endangered blonde capuchins fish for termites and use new techniques to accomplish the task. Biol Lett.  7( 4): 532– 535. Google Scholar CrossRef Search ADS PubMed  Stevens CE, Hume ID. 1995. Comparative physiology of the vertebrate digestive system . 2nd ed. Cambridge: Cambridge University Press. Strait SG, Vincent J. 1998. Primate faunivores: physical properties of prey items. Int J Primatol.  19( 5): 867– 878. http://dx.doi.org/10.1023/A:1020397430482 Google Scholar CrossRef Search ADS   Strobel S, Roswag A, Becker NI, Trenczek TE, Encarnação JA. 2013. Insectivorous bats digest chitin in the stomach using acidic mammalian chitinase. PLoS One  8( 9): e72770. Google Scholar CrossRef Search ADS PubMed  Sussman RW. 1991. Primate origins and the evolution of angiosperms. Am. J. Primatol . 23( 4): 209– 223. http://dx.doi.org/10.1002/ajp.1350230402 Google Scholar CrossRef Search ADS   Synstad B, Gaseidnes S, van Aalten DMF, Vriend G, Nielsen JE, Eijsink VGH. 2004. Mutational and computational analysis of the role of conserved residues in the active site of a family 18 chitinase. Eur J Biochem.  271( 2): 253– 262. Google Scholar CrossRef Search ADS PubMed  Terborgh J. 1983. Five new world primates: a study in comparative ecology . Princeton (NJ): Princeton University Press. Tjoelker LW, Gosting L, Frey S, Hunter CL, Trong HL, Steiner B, Brammer H, Gray PW. 2000. Structural and functional definition of the human chitinase chitin-binding domain. J Biol Chem . 275( 1): 514– 520. Google Scholar CrossRef Search ADS PubMed  Tomberlin JK, Sheppard DC, Joyce JA. 2002. Selected life-history traits of black soldier flies (Diptera: Stratiomyidae) reared on three artificial diets. Ann Entomol Soc Am.  95: 379– 386. http://dx.doi.org/10.1603/0013-8746(2002)095[0379:SLHTOB]2.0.CO;2 Google Scholar CrossRef Search ADS   Tosi AJ, Buzzard PJ, Morales JC, Melnick DJ. 2002. Y-chromosome data and tribal affiliations of Allenopithecus and Miopithecus. Int J Primatol.  23( 6): 1287– 1299. http://dx.doi.org/10.1023/A:1021178904824 Google Scholar CrossRef Search ADS   Tosi AJ, Detwiler KM, Disotell TR. 2005. X-chromosomal window into the evolutionary history of the guenons (Primates: Cercopithecini). Mol Phylogenet Evol.  36( 1): 58– 66. http://dx.doi.org/10.1016/j.ympev.2005.01.009 Google Scholar CrossRef Search ADS PubMed  van Schaik CP, Ancrenaz M, Borgen G, Galdikas B, Knott CD, Singleton I, Suzuki A, Utami SS, Merrill M. 2003. Orangutan cultures and the evolution of material culture. Science  299( 5603): 102– 105. Google Scholar CrossRef Search ADS PubMed  Vannella KM, Ramalingam TR, Hart KM, de Queiroz Prado R, Sciurba J, Barron L, Borthwick LA, Smith AD, Mentink-Kane M, White S. 2016. Acidic chitinase primes the protective immune response to gastrointestinal nematodes. Nat Immunol.  21: 73– 79. Wertheim JO, Murrell B, Smith MD, Kosakovsky Pond SL, Scheffler K. 2015. RELAX: detecting relaxed selection in a phylogenetic framework. Mol Biol Evol . 32( 3): 820– 832. http://dx.doi.org/10.1093/molbev/msu400 Google Scholar CrossRef Search ADS PubMed  Whitaker JO, Dannelly HK, Prentice DA. 2004. Chitinase in insectivorous bats. J Mammal.  85( 1): 15– 18. http://dx.doi.org/10.1644/1545-1542(2004)085<0015:CIIB>2.0.CO;2 Google Scholar CrossRef Search ADS   Williams BA, Kay RF, Kirk EC. 2010. New perspectives on anthropoid origins. Proc Natl Acad Sci U S A . 107( 11): 4797– 4804. Google Scholar CrossRef Search ADS PubMed  Yang Z. 2007. PAML 4: phylogenetic analysis by maximum likelihood. Mol Biol Evol . 24( 8): 1586– 1591. http://dx.doi.org/10.1093/molbev/msm088 Google Scholar CrossRef Search ADS PubMed  Zhang J. 2003. Evolution by gene duplication: an update. Trends Ecol Evol.  18( 6): 292– 298. http://dx.doi.org/10.1016/S0169-5347(03)00033-8 Google Scholar CrossRef Search ADS   Zhu Z, Zheng T, Homer RJ, Kim Y-K, Chen NY, Cohn L, Hamid Q, Elias JA. 2004. Acidic mammalian chitinase in asthmatic Th2 inflammation and IL-13 pathway activation. Science  304( 5677): 1678– 1682. Google Scholar CrossRef Search ADS PubMed  © The Author 2017. Published by Oxford University Press on behalf of the Society for Molecular Biology and Evolution. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com

Journal

Molecular Biology and EvolutionOxford University Press

Published: Mar 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Unlimited reading

Read as many articles as you need. Full articles with original layout, charts and figures. Read online, from anywhere.

Stay up to date

Keep up with your field with Personalized Recommendations and Follow Journals to get automatic updates.

Organize your research

It’s easy to organize your research with our built-in tools.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve Freelancer

DeepDyve Pro

Price
FREE
$49/month

$360/year
Save searches from
Google Scholar,
PubMed
Create lists to
organize your research
Export lists, citations
Read DeepDyve articles
Abstract access only
Unlimited access to over
18 million full-text articles
Print
20 pages/month
PDF Discount
20% off