Genomewide screening reveals high levels of insertional polymorphism in the human endogenous retrovirus family HERV-K(HML2): implications for present-day activity

Abstract

The published human genome sequence contains many thousands of endogenous retroviruses (HERVs) but all are defective, containing nonsense mutations or major deletions. Only the HERV-K(HML2) family has been active since the divergence of humans and chimpanzees; it contains many members that are human specific, as well as several that are insertionally polymorphic (an inserted element present only in some human individuals). Here we perform a genomewide survey of insertional polymorphism levels in this family by using the published human genome sequence and a diverse sample of 19 humans. We find that there are 113 human-specific HERV-K(HML2) elements in the human genome sequence, 8 of which are insertionally polymorphic (11 if we extrapolate to those within regions of the genome that were not suitable for amplification). The average rate of accumulation since the divergence with chimpanzees is thus approximately 3.8 x 10(-4) per haploid genome per generation. Furthermore, we find that the number of polymorphic elements is not significantly different from that predicted by a standard population genetic model that assumes constant activity of the family until the present. This suggests to us that the HERV-K(HML2) family may be active in present-day humans. Active (replication-competent) elements are likely to have inserted very recently and to be present at low allele frequencies, and they may be causing disease in the individuals carrying them. This view of the family from a population perspective rather than a genome perspective will inform the current debate about a possible role of HERV-K(HML2) in human disease.

FIG. 1.

Maximum likelihood phylogeny of HERV-K(HML2) LTRs. Filled and open circles indicate LTRs from full-length and segmentally duplicated elements, respectively. Black boxes represent taxa present in both the chimpanzee and human genome sequences, whereas red boxes represent human-specific elements. Intermingling of the two classes is probably due to a variety of factors, such as gene conversion or ancestral polymorphism. Open boxes represent taxa whose distribution could not be determined directly (the chimpanzee genome project is incomplete), and their probable distribution was estimated from their position in the phylogeny. A dashed line indicates the placement of K113, which is absent from the published human genome sequence. The large boxed region (which excludes most non-human-specific elements) in shown in more detail in Fig. 2. Scale bar shows mean number of substitutions per site.

FIG. 2.

Elements screened for insertional polymorphism. Taxon names are followed by their genomic location in parentheses. Black boxes indicate elements homozygous for the insertion in all 19 individuals surveyed, whereas those in red display insertional polymorphism, with the filled region in each box being proportional to the frequency of the inserted element in the samples. The other (nonboxed) elements gave inconclusive results, usually because of their location in regions of highly repetitive DNA. Data from all full-length elements were taken from previous reports (21, 35). A nucleotide alignment of the surveyed solo LTRs, together with their flanking sequences, is shown in Fig. S1. Scale bar shows mean number of substitutions per site.

FIG. 3.

Detection of HERV-K(HML2) preinsertion sites. Amplification of solo LTR loci within the published human genome sequence showed a preinsertion site (PRE), a solo LTR (sLTR), or a full-length element (FULL) when tested against a panel of 19 individuals. The provenance of each individual is shown in Table 1, and allele frequencies are shown in Table 2. The identity of each band was confirmed by DNA sequencing. We rescreened the previously identified polymorphism 165c5, as the originally estimated frequencies were based largely on individuals from Russia (25).

FIG. 4.

Proposed model of HERV-K(HML2) family evolution within humans. (a) At each time point, there is a large unfixed population of elements, a proportion of which are replication competent and infectious, whereas others are defective. Some of the subset of defective elements, but none of the replication-competent elements, eventually drift to fixation. The population of unfixed elements is continuously replenished by new insertions resulting from the replication of intact and unfixed elements. (b) Over time, the fixed and defective elements (i.e., A, B, and C) accumulate so that in any one genome all, or almost all, of the elements are defective, the intact and infectious elements being present only in a very small proportion of individuals.