Positive selection of primate genes that promote HIV-1 replication

Abstract

Evolutionary analyses have revealed that most host-encoded restriction factors against HIV have experienced virus-driven selection during primate evolution. However, HIV also depends on the function of many human proteins, called host factors, for its replication. It is not clear whether virus-driven selection shapes the evolution of host factor genes to the extent that it is known to shape restriction factor genes. We show that five out of 40 HIV host factor genes (13%) analyzed do bear strong signatures of positive selection. Some of these genes (CD4, NUP153, RANBP2/NUP358) have been characterized with respect to the HIV lifecycle, while others (ANKRD30A/NY-BR-1 and MAP4) remain relatively uncharacterized. One of these, ANKRD30A, shows the most rapid evolution within this set of genes and is induced by interferon stimulation. We discuss how evolutionary analysis can aid the study of host factors for viral replication, just as it has the study of host immunity systems.

Keywords: Arms race; Co-factors; Genetic conflict; Host factors; Paleovirology.

Fig 1. Evolutionary analysis of genes encoding HIV host factors

(A) Each circle represents a whole-genome RNA interference screen previously conducted to identify human genes important for HIV replication (Brass et al., 2008; König et al., 2008; Yeung et al., 2009; Zhou et al., 2008). Indicated within overlap regions are the numbers of human genes identified in multiple screens. (B) A histogram shows whole-gene dN/dS values previously calculated for 10,376 orthologous gene trios from the human, chimpanzee, and rhesus genomes (Rhesus Macaque Genome Sequencing and Analysis Consortium et al., 2007). On the x-axis is the average dN/dS calculated over the length of each gene for this three-species tree. On the y-axis is the number of gene trios with this average dN/dS value. On this distribution are overlayed the dN/dS values for similar trios hand-curated for known HIV restriction factors (red asterisks), and for the 39 putative host factors analyzed in this study (gray asterisks). dN/dS values were calculated using the M0 model in PAML (Yang, 1997). The restriction factors are (from left to right on the distribution): SAMHD1, Tetherin, ZAP, TRIM22, TRIM5, APOBEC3H, APOBEC3DE, APOBEC3G. (C) Example sliding window analyses for CXCR4 and CD4. The x-axis is in base pairs and represents the length of each gene. dN/dS was calculated in sliding windows moving along the length of the gene. The three pairwise species comparisons made are color-coded (blue, human vs orangutan; red, human vs rhesus; green, rhesus vs marmoset). For each comparison, the window with the highest dN/dS value was tested for statistical significance using a Monte Carlo simulation (Comeron, 1999), and two windows where dN/dS is significantly greater than 1 (p < 0.001 and p < 0.004) are indicated. (D) The maximum dN/dS value observed in each pairwise sliding window analysis is summarized. An asterisk (*) indicates instances where dN/dS is significantly greater than 1 (p < 0.05).

Fig 2. ANKRD30Aand MAP4 exhibit signatures of positive selection

A cladogram illustrates the relationship of the primate species analyzed in this study. The 20 species used in the positive selection analysis are indicated with an asterisk next to the species name. Numbers at the top refer to codon coordinates in two genes, ANKRD30A and MAP4. These four codons shown were identified as evolving under positive selection, and as expected the amino acids encoded at these sites are highly variable between species. The arbitrarily colored boxes represent different amino acids encoded. All codons under positive selection in ANKRD30A and MAP4 were sequenced from individuals representing small population sets of different primate species (indicated with a triangle at the end of the branch on the primate cladogram). For humans, SNPs were identified in human SNP databases. Non-synonymous SNPs identified are indicated by 2 squares next to each other. A hyphen indicates lack of information, because ANKDR30A could not be sequenced from these species. These four codons were chosen for illustration because they are under positive selection and also bear SNPs in human or primate populations.

Fig 3. ANKRD30A expression in humans

(A) A human tissue cDNA panel was probed for ANKDR30A expression using a primer set that spans an intron. As a positive control, CRFK cells were engineered to stably express part of ANKRD30A. Human DNA was used to show the product size obtained from genomic template (with the intron). Because ANKRD30A was originally identified as a gene associated with breast cancers, we also demonstrate its expression in a breast carcinoma line (HCC1937). (B) Expression of ANKRD30A was assessed in various laboratory cell lines. Cells were grown both with and without stimulation by 1000 U/ml interferon-β, and mRNA was harvested. Expressed and spliced transcripts where detected by RT-PCR. Primers were also designed to amplify spliced transcripts of NUP153 as a control for the amount of input material in each reaction. (C) A summary of genome-wide screens for HIV host factors (Brass et al., 2008; Jäger et al., 2012; König et al., 2008; Yeung et al., 2009; Zhou et al., 2008), along with relevant details of the methods used.