Rapid Evolution of Glycan Recognition Receptors Reveals an Axis of Host-Microbe Arms Races beyond Canonical Protein-Protein Interfaces

Abstract

Detection of microbial pathogens is a primary function of many mammalian immune proteins. This is accomplished through the recognition of diverse microbial-produced macromolecules including proteins, nucleic acids, and carbohydrates. Pathogens subvert host defenses by rapidly changing these structures to avoid detection, placing strong selective pressures on host immune proteins that repeatedly adapt to remain effective. Signatures of rapid evolution have been identified in numerous immunity proteins involved in the detection of pathogenic protein substrates, but whether similar signals can be observed in host proteins engaged in interactions with other types of pathogen-derived molecules has received less attention. This focus on protein-protein interfaces has largely obscured the study of fungi as contributors to host-pathogen conflicts, despite their importance as a formidable class of vertebrate pathogens. Here, we provide evidence that mammalian immune receptors involved in the detection of microbial glycans have been subject to recurrent positive selection. We find that rapidly evolving sites in these genes cluster in key functional domains involved in carbohydrate recognition. Further, we identify convergent patterns of substitution and evidence for balancing selection in one particular gene, MelLec, which plays a critical role in controlling invasive fungal disease. Our results also highlight the power of evolutionary analyses to reveal uncharacterized interfaces of host-pathogen conflict by identifying genes, like CLEC12A, with strong signals of positive selection across mammalian lineages. These results suggest that the realm of interfaces shaped by host-microbe conflicts extends beyond the world of host-viral protein-protein interactions and into the world of microbial glycans and fungi.

Keywords: balancing selection; evolutionary conflict; host–pathogen interactions; microbial glycans; pattern recognition receptor; rapid evolution.

Fig. 1.

Positive selection across mammalian carbohydrate recognition PRRs. (A) Positive selection analyses of 26 glycan PRRs in primates (left column), rodents (middle), and bats (right column). Colored boxes indicate whether evidence of positive selection was supported by PAML analyses only (medium blue) or by both PAML and BUSTED analyses (dark blue). Genes with no evidence for positive selection are represented by pale blue boxes. Statistical cutoffs were P < 0.05 for PAML M7 versus M8 likelihood ratio tests and for BUSTED analysis. (B) Patterns of positive selection mapped onto a phylogenetic tree of the human CTLD domains. Only genes from the gene set with CTLDs are represented. Colored circles represent evidence of positive selection in the primate (orange), rodent (purple), and/or bat (blue) lineages. Genes with black circles were not analyzed in this study because of unclear ortholog relationships across mammals but do have important roles in pathogen detection in mammals. Numbers indicate bootstrap values from phylogenetic tree construction using IQ-TREE.

Fig. 2.

Diversification of Langerin (CD207) ligand-binding interfaces in all mammalian lineages. (A) Positively selected residues (triangles) predicted by PAML (Model 8, BEB > 0.9) cluster primarily in the extracellular portion of Langerin (CD207), with many in the CTLD. A number of positively selected sites in the CTLD are common across primates (orange triangles), rodents (purple triangles), and bats (blue triangles). (B) Agreement between different algorithms for identifying site-specific positive selection in Langerin of different mammalian groups. Listed residue numbers correspond to the position in the human Langerin sequence. Single letter residues correspond to the amino acid identity in human (primates, left), house mouse (rodents, middle), or black flying fox (bats, right) sequences. Bolded residues are those predicted to be under positive selection across all mammals by one or more tests. (C) Positively selected sites mapped onto a crystal structure of the human Langerin CTLD (gray, PDB:3p5d) in complex with a mannose ligand (yellow) and Ca²⁺ ion (magenta) (Feinberg et al. 2011). Positively selected sites in all three lineages (colored in green) along with several sites from rodent (blue) and bat (purple) analyses are shown with sidechains and surround the ligand binding site.

Fig. 3.

Single nucleotide polymorphisms in primate populations converge on a single site in Melanin Lectin (CLEC1A). (A) Patterns of conservation and variation at amino acid position 26 of MelLec across primates. Most primate species carry the ancestral alanine allele (orange highlighting), whereas single nucleotide polymorphisms in both humans (glycine, green highlighting) and squirrel monkeys (valine, pink highlighting) confer missense mutations. (B) Genotypes of 19 unrelated squirrel monkey gDNA samples from three S. boliviensis subspecies. The sex and the amino acid identity at position 26 for each individual are indicated, with heterozygous individuals indicated as carrying both Ala and Val amino acids (A/V in Black-capped and Peruvian squirrel monkeys). (C) (top) Geographic distribution of the glycine 26 allele (green) at SNP rs2306894 in human populations. Allele frequencies are shown for populations from the 1KG Project and the HGDP. Individuals carrying the Ala26 allele (orange) have been previously shown to have higher risk of invasive fungal infections in stem-cell transplant patients (Stappers et al. 2018). (middle) Tajima's D values for populations from the HGDP and 1KG and (bottom) β⁽¹⁾ for populations from the 1KG project showing evidence of balancing selection at the MelLec locus. For both plots, * empirical P-value < 0.05, ** empirical P-value < 0.01. Population abbreviations are as follows: AMR, America; AFR, Africa; EUR, Europe; CSA, Central-South Asia; ME, Middle East; SAS, South Asia; EAS, East Asia; OC, Oceania.

Fig. 4.

Extensive positive selection in CLEC12A across mammals reveals a new host–pathogen battleground. (A) Diagram showing sites under positive selection in CLEC12A in primates (orange triangles), rodents (purple triangles) and bats (blue triangles). Indicated sites were predicted by PAML (Model 8, BEB > 0.9). Locations of the CTLD and transmembrane domain are indicated on the left. (B)–(D) dN/dS values for CLEC12A were calculated across the species phylogenies of primates (B), rodents (C), and bats (D) using PAML (free ratios, Model = 1 setting). Lineages with elevated dN/dS values (>1), suggestive of positive selection along that branch, are indicated with colored lines. Calculated dN/dS values are listed above each branch and for branches lacking either nonsynonymous or synonymous sites; ratios of the respective substitution numbers (N:S) are indicated.