Evolution of the primate trypanolytic factor APOL1

Abstract

ApolipoproteinL1 (APOL1) protects humans and some primates against several African trypanosomes. APOL1 genetic variants strongly associated with kidney disease in African Americans have additional trypanolytic activity against Trypanosoma brucei rhodesiense, the cause of acute African sleeping sickness. We combined genetic, physiological, and biochemical studies to explore coevolution between the APOL1 gene and trypanosomes. We analyzed the APOL1 sequence in modern and archaic humans and baboons along with geographic distribution in present day Africa to understand how the kidney risk variants evolved. Then, we tested Old World monkey, human, and engineered APOL1 variants for their ability to kill human infective trypanosomes in vivo to identify the molecular mechanism whereby human trypanolytic APOL1 variants evade T. brucei rhodesiense virulence factor serum resistance-associated protein (SRA). For one APOL1 kidney risk variant, a two-residue deletion of amino acids 388 and 389 causes a shift in a single lysine residue that mimics the Old World monkey sequence, which augments trypanolytic activity by preventing SRA binding. A second human APOL1 kidney risk allele, with an amino acid substitution that also restores sequence alignment with Old World monkeys, protected against T. brucei rhodesiense due in part to reduced SRA binding. Both APOL1 risk variants induced tissue injury in murine livers, the site of transgenic gene expression. Our study shows that both genetic variants of human APOL1 that protect against T. brucei rhodesiense have recapitulated molecular signatures found in Old World monkeys and raises the possibility that APOL1 variants have broader innate immune activity that extends beyond trypanosomes.

Fig. 1.

Distribution of the G1 and G2 APOL1 variants across Africa. Allele frequencies of the G1 and G2 variants are indicated as blue and green wedges, respectively. Circle size reflects the number of individuals genotyped: small, <10 individuals/20 chromosomes; medium, 10–100 individuals/20–200 chromosomes; large, >100 individuals/200 chromosomes. Countries are shaded according to the subspecies of Trypanosoma brucei that cause African sleeping sickness. Darker green, gambiense types 1 and 2; light green, gambiense type 1; pink, both rhodesiense and gambiense type 1; purple, rhodesiense.

Fig. 2.

APOL1 variation in modern and archaic humans. Each leaf of the polymorphism tree (Left) and the corresponding horizontal row of the haplotype table (Right) represent a different haplotype for the 8-kbp APOL1 region between rs136150 and rs9610474 observed among 1,092 samples analyzed in phase I of the 1000 Genomes Project. Numbers to the right of the polymorphism tree reflect the numbers of times that haplotype was observed (only haplotypes at a frequency of ≥1% are shown). In the polymorphism tree, the x axis measures the average distance between haplotypes in the number of base pair differences. Haplotypes observed in out-of-Africa samples are displayed in cyan, whereas all other haplotypes are displayed in red. In the haplotype table, horizontal rows represent haplotypes across the 8-kbp region; at each polymorphic locus, ancestral alleles are green, derived alleles are red, and blue indicates that the allele could not be ascertained. Each column of the haplotype table corresponds to a different biallelic marker (i.e., base pair); rs markers are color coded: nonsynonymous coding variants are in red, synonymous coding variants are in green, and noncoding variants are in black. The base pairs defining the G1 and G2 variants are boxed in blue. Introns (thin lines) and exons (wide bars) in this 8-kbp window are illustrated above the haplotype table. Importantly, the diversity observed among the out-of-Africa haplotypes is much lower than the diversity observed in the remaining haplotypes, consistent with the hypothesis that these out-of-Africa haplotypes all originated from a recent ancestor, which would be the case for introgression from archaic humans. D, deletion; I, insertion; Ref, human genome reference.

Fig. 3.

Coding variants in APOL1. The haplotype tree (Left) is constructed from coding variants in APOL1 identified at least three times in phase 1 of the 1000 Genomes Project. The tree is read from left to right, with each branch point representing an amino acid change that creates a new haplotype (i.e., set of particular variants on a single chromosome). The 11 haplotypes created by these 11 amino acid changes are illustrated in the chart (Right). In the chart, each row represents a haplotype, with the red allele marking the haplotype-defining residues. Columns show the amino acid positions where the haplotypes may differ. p.K255R defines the haplotype shared by Neanderthal and ∼22% of Europeans. This haplotype is the reference (Ref; yellow) sequence in the current draft of the human genome (hg19). Nearly all individuals with the G1 allele (green) have both p.S342G and p.I384M, but a few individuals have p.S342G without p.I384M (no individuals have yet been identified with p.I384M but without p.S342G). p.NYK388K (blue), a deletion of asparagine and tyrosine at positions 388 and 389, defines the G2 allele. Frequencies for each allele are reported in Table S1.

Fig. 4.

C-terminal lysine residues determine APOL1/SRA binding interactions and protection against SRA-expressing trypanosomes in vivo. (A) The C-terminal domains of primate APOL1 variants. Residue identities are shown as dashes if they are identical in all sequences; otherwise, residue identities are indicated (bold, nonpolar residues; green, acidic residues; red, basic residues). Underlined hydrophobic sequences are putative membrane-spanning regions. Residues specific to G1 are circled in blue, and those lysines that are specific to baboon or G2 are circled in red. Coils denote putative α-helices; the C-terminal helix contains a heptad repeat sequence (a–g), wherein nonpolar residues occupy a and d positions. Asterisks denote the positions of the heptad-repeat leucine residues at the d position, the last of which is missing in G2 and replaced by a nonpolar alanine. Four lysine residues that are found in OWM but not human WT APOL1 are denoted K1–K4. The 2-aa deletion in the APOL1 G2 variant shifts a lysine residue into register with baboon K4. (B) Transgenic mouse sera containing comparable amounts of the indicated APOL1 variant were incubated at pH 5.0 with NHS-agarose–coupled SRA. APOL1 was detected in the unwashed suspension [load (L); diluted 1:40] and the SRA-bound fraction eluted with 0.1 M Tris [pH 8.0; bound (B); diluted 1:40 and 1:80] by immunoblotting. SG, S342G variant alone; IM, I384M variant alone. (C) Human APOL1 gene variants encoding baboon-specific lysine substitutions K3, K4, or both K3 and K4 genes were expressed in mice by HGD. (D and E) WT, SG, IM, G1, and G2 human APOL1 genes were expressed in mice by HGD. The amount of plasmid injected was 50 μg (C and D), unless otherwise indicated (E). Control mice received an empty plasmid (vector) or a plasmid containing human APOL1 (WT). On day 3 postinjection, mice were infected with 5,000 human-infective T. brucei-SRA, and their blood was monitored for parasites. When parasitemia reached 10⁹ parasites/mL, the mouse was killed; mice that survived to 21 d never developed parasitemia. Mouse plasma protein levels of each APOL1 construct were determined by immunoblot on the day of infection using a rabbit polyclonal antibody (C; plasma diluted 1:40) or a mouse monoclonal antibody (D and E). Results in D are pooled from multiple experiments. Each experiment had at least five mice per group. Survival was significantly different from WT controls. **P < 0.01; ***P < 0.001; log-rank test.

Fig. 5.

The C-terminus domain contributes to tissue pathology. (A) Sequences of APOL1 C-terminal variants and constructs screened in the livers of HGD mice. The key is as outlined in Fig. 4A. (B) Murine plasma protein levels were determined of each APOL1 construct by immunoblot 3 d post-HGD using a rabbit polyclonal antibody (plasma diluted 1:40). S, normal mouse serum; Tr, truncated APOL1 constructs; WT, full-length APOL-I. (C) The livers of HGD mice injected with saline vehicle or plasmids containing WT, baboon, K4, IM, SG, G1, G2, Tr1, Tr2, Tr2-SG, signal peptide deleted (Sig. Del.) WT, Sig. Del. G1, or Sig. Del. G2 APOL1. Livers were removed on day 5 postinjection and processed for H&E staining. Representative examples are shown of at least three mice per group. Arrowheads represent leukocyte infiltrates. Arrows represent necrosis.