Molecular evolution of the primate α-/θ-defensin multigene family

Abstract

The primate α-/θ-defensin multigene family encodes versatile endogenous cationic and amphipathic peptides that have broad-spectrum antibacterial, antifungal and antiviral activity. Although previous studies have reported that α-/θ-defensin (DEFA/DEFT) genes are under birth-and-death evolution with frequent duplication and rapid evolution, the phylogenetic relationships of the primate DEFA/DEFT genes; the genetic bases for the existence of similar antimicrobial spectra among closely related species; and the evolutionary processes involved in the emergence of cyclic θ-defensins in Old World monkeys and their subsequent loss of function in humans, chimpanzees and gorillas require further investigation. In this study, the DEFA/DEFT gene repertoires from primate and treeshrew were collected, followed by detailed phylogenetic, sequence and structure, selection pressure and comparative genomics analyses. All treeshrew, prosimian and simian DEFA/DEFT genes are grouped into two major clades, which are tissue-specific for enteric and myeloid defensins in simians. The simian enteric and myeloid α-defensins are classified into six functional gene clusters with diverged sequences, variable structures, altered functional constraints and different selection pressures, which likely reflect the antimicrobial spectra among closely related species. Species-specific duplication or pseudogenization within each simian cluster implies that the antimicrobial spectrum is ever-shifting, most likely challenged by the ever-changing pathogen environment. The DEFT evolved from the myeloid DEFA8. The prosegment of θ-defensin is detected with adaptive changes coevolving with the new protein fold of mature peptide, coincident with the importance of the prosegment for the correct folding of the mature peptide. Lastly, a less-is-hitchhiking hypothesis was proposed as a possible explanation for the expansion of pseudogene DEFTP and the loss of functional DEFT, where the gain or loss of the hitchhiker is determined by its adjacent driver gene during the birth-and-death evolutionary process.

Figure 1. Phylogenetic trees of the α-/θ-defensin (DEFA/DEFT) genes in primates and treeshrews.

A: Phylogenetic tree of primate and treeshrew DEFA/DEFT genes based only on the signal-prosegment region. The BI tree is selected as the background tree. The major clades or clusters having similar topologies from all three tree-building methods (BI, NJ and ML) are combined and labeled with the BI posterior probabilities and the bootstrap support values from the NJ and ML analyses. Primate DEFA/DEFT genes are clustered into prosimian and simian clades. The simian clades are classified and named following the nomenclature used for the human DEFA/DEFT genes. The “P” in node labels denotes a pseudogene, and the “U” indicates a gene with an ambiguous locus from the species in the synteny map in Figure 7. B: Phylogenetic tree of primate and treeshrew DEFA/DEFT genes based the entire coding region. The BI tree is selected as the background tree. The treeshrew DEFA genes are the outgroups of the two separate clades.

Figure 2. Type I and type II functional divergences among simian enteric or myeloid α-defensin clusters.

A: Statistical analyses of the type I and type II functional divergences. Clusters within both simian enteric α-defensins and simian myeloid α-defensins have diverged. Pairwise coefficients (θ_ij±SE) and likelihood ratio statistics (LRT) are analyzed using the bootstrapped Gu99 method. Type I functional distance (d _F) between clusters is calculated as d _F = −ln(1−θ_ij). The effective number of functional divergence related sites (n_e) is estimated following the Gu2013 method. B: Critical amino acid sites responsible for type I and type II functional divergences from simian myeloid α-defensin clusters. These sites are identified based on the n_e cutoff after ranking posterior probabilities. C: Critical amino acid sites responsible for type I and type II functional divergences from simian enteric α-defensin clusters. D: The diagram of the sites responsible for type I and type II functional divergences in different simian clusters. Closed circles represent sites responsible for type I functional divergence, and opened circles denote sites responsible for type II functional divergence.

Figure 3. Structural flexibility and surface electrostatic potential for the simian myeloid and enteric α-defensin dimers.

A: Superimposed α-defensin dimers for the myeloid DEFA1, DEFA4 and DEFA8 proteins as well as the enteric DEFA5, DEFA6 and DEFA9 proteins. B: Diagram illustrating the intermolecular hydrogen bonds, the two distances (A22C^α-B22C^α and A11C^α-B11C^α) and the dihedral angle (A11C^α-A22C^α-B22C^α-B11C^α). C: Plots of the A22C^α-B22C^α and A11C^α-B11C^α distances during the 16 nanoseconds of molecular dynamics simulations (upper panel) and plots of the average A22C^α-B22C^α and A11C^α-B11C^α distances (lower panel). Similar A22C^α-B22C^α distances indicate that the dimer structures are well maintained during simulation processes, whereas diverse A11C^α-B11C^α distances reflect the flexibility of the dimers. Error bars represent standard deviations. D: Plots of the dihedral angle A11C^α-A22C^α-B22C^α-B11C^α. The variant dihedral angles among different clusters of myeloid and enteric α-defensins indicate different dimer topologies. E: Surface electrostatic potential generated using the smoothed trajectory from the last five frames of molecular dynamics simulations. The electrostatic potential (±10 kT/e) is colored red (−) or blue (+). The left and right views of each structure are the same as in panel A.

Figure 4. Sequence motifs and electrostatic features of primate cyclic θ-defensins.

A: The θ-defensin octadecapeptide is formed by the head-to-tail ligation of two nonapeptides. The sequence logo for the nonapeptide is generated using all the primate θ-defensins. Sites 3I and 5R are under positive selection and are marked with asterisks. B: Surface electrostatic potential of the θ-defensins without or with the positively selected sites (3I and 5R).

Figure 5. Sliding window analyses of the Ka/Ks for all DEFA, DEFT and separate DEFA clusters.

A: Sliding window analysis of the amino acid identity and Ka/Ks for all DEFA. The sequence conservation decreases from the signal peptide and prosegment to the mature peptide. The selection pressures acting on different windows are different as indicated by the Ka/Ks ratios. Fragment 26–35 in the prosegment region, fragment 56–65 in the cleavage region and most fragments in the mature peptide are under positive selection (Ka/Ks>1). The significance of Ka/Ks>1 is tested by a bootstrap method and is indicated by * (P<0.05) or ** (P<0.01). B: Sliding window analysis of the amino acid identity and Ka/Ks for DEFT. The sequence conservation also decreases from the signal peptide and prosegment to the mature peptide. The selection pressure on different regions is also variable. Compared to the same region of all DEFA and that of DEFA8, fragment 41–50 in the prosegment of DEFT is under strong positive selection. C: Sliding window analysis of Ka/Ks for each simian DEFA cluster. Only the DEFA5 and DEFA8 clusters contain fragments that are under positive selection. D: Sliding window analysis of Ka/Ks for the two prosimian clades. The fragment 56–65 in the cleavage region is also under positive selection for both prosimian clades, consistent with that of all DEFA.

Figure 6. Coevolution of the prosegment with the new protein fold of the mature peptide.

The ancestral protein fold refers to the α-defensin mature peptide, and the new protein fold refers to the θ-defensin mature peptide. Because the prosegment is important for the correct folding of the mature peptide, the prosegment coevolves with the mature peptide through adaptive changes.

Figure 7. Hitchhiking of DEFT/DEFTP during the birth-and-death evolution of the primate DEFA/DEFT multigene family.

A: Comparative synteny map of the DEFA/DEFT gene loci in humans, chimpanzees, orangutans, macaques and marmosets. The boxes highlight the two genes that are linked and duplicate together, including a driver from the DEFA1 cluster and a hitchhiker from the DEFT/DEFA10 cluster. Arrowheads indicate transcriptional orientation. Pseudogenes are in white, and functional genes are in black. The dashed-line box includes the pseudogene DEFTP containing the nonsense mutation at site 17. B: The NJ tree based on the introns of the boxed DEFT/DEFTP from humans (hspi), chimpanzees (ptro), orangutans (pabe) and macaques (mmul). The nonsense mutations (*) at codon site 17 of the pseudogenes hspi_DEFT1P, hspi_DEFT2P ptro_DEFT1P, ptro_DEFT2P and pabe_DEFT4P are underlined with dashed lines, suggesting that these DEFTP pseudogenes are derived from a common ancestor.

Figure 8. The less-is-hitchhiking hypothesis during the birth-and-death evolution of multigene families.

With the expansion of the gene (driver) under positive selection, the adjacent linked gene (hitchhiker) is duplicated in the process of segmental duplication. Pseudogenization of the hitchhiker occurs in an ancestral species following the duplication process. During the process of birth-and-death evolution, the hitchhiker that gains more copies can be either the pseudogene or the functional gene with the other one being lost, and the gain or loss of the hitchhiker is determined by the fitness of its adjacent driver. The expansion of the pseudogene and loss of the functional gene scenario is defined as the less-is-hitchhiking hypothesis.