The human genome in the LINE of fire

One of the most surprising revelations of the sequencing of the human genome was that nearly half of our DNA is derived from transposable element (TE) insertions, and this is likely to be an underestimate, because many TE-derived sequences have diverged beyond recognition (1). Remarkably, the vast majority of human TE sequences result from the activity of a single class of TEs known as LINE retrotransposons. Represented by the currently active LINE-1 or L1 elements, LINEs are autonomous TEs that propagate in the genome by making RNA copies of themselves that are subsequently reverse transcribed and integrated into the genome (2⇓–4). As a result of their ongoing activity during the past 150 million years, L1 elements account for approximately one-third of the human genome, by way of self-mobilization (≈500,000 copies) and transmobilization of nonautonomous Alu (≈1,100,000 copies) and SVA (≈3,000 copies) TEs and processed pseudogenes (≈11,000 copies) (1⇓⇓⇓–5). This high density of repetitive sequences poses a considerable threat to the stability of our genome, ranging from mutations and genomic alterations on TE insertion to large-scale genomic rearrangements triggered by recombination between nonallelic homologous TE sequences (2⇓–4, 6). The availability of multiple primate genome sequences now makes it possible to quantify the overall impact TE activity has had on human genome evolution. In this issue of PNAS, Han et al. (7) report the first comprehensive characterization of the genome-wide impact of human genomic deletions resulting from recombination between L1 elements, which occurred since the human-chimpanzee divergence approximately 6 million years ago.

Although some L1-driven rearrangements have resulted in beneficial innovations during evolution (8), the vast majority of these events are either neutral or deleterious to the genome. The deleterious nature of L1-L1 recombination events and their elimination from the population by negative selection (9) may help to understand why only 3 L1 recombination-mediated deletions causing human diseases have been reported to date (10⇓–12). However, the contribution of L1-L1 recombination in generating genomic deletions at an evolutionary timescale remained to be clarified. To address this issue, Han et al. (7) first computationally compared the human and chimpanzee genome sequences to identify candidate deletions that had accumulated in the human genome since the human-chimpanzee divergence. Of note is that the authenticity of the computational predictions was confirmed by experimental procedures, thus making the final dataset highly reliable. Han et al. identified 73 deletions associated with recombination between L1 elements that altogether removed nearly half a million base pairs of human DNA within just the past few million years of human evolution. This considerable amount of deleted DNA indicates that L1-L1 recombination substantially contributed to alter our genome during recent human evolution. Actually, some deletions occurred so recently that they are polymorphic for presence or absence among humans (7). Thus, L1-L1 recombination creates structural genomic variation among humans. Broadly known as copy number variants, this class of polymorphisms is increasingly recognized as a major source of variation among humans (13).

L1-L1 recombination is a major contributor to human genome plasticity.

The results of Han et al. (7) constitute a cornerstone that, together with other studies by the Batzer laboratory (14⇓⇓⇓–18), now allows us to quantify the genome-wide extent of deletions generated by the 2 major components of LINE retrotransposition activity in the human genome: L1 and Alu. Genomic deletions can be generated at 2 distinct stages of the retrotransposon life cycle: (i) at the time of insertion of the element at a new genomic locus via either classical endonuclease-dependent (19) or nonclassical endonuclease-independent (20) retrotransposition; and (ii) at a postinsertional stage, by recombination between nonallelic homologous elements potentially inserted in the genome for a long time. The emerging picture is that nearly 1 Mb of DNA has been deleted from the human genome by the products of LINE retrotransposition within just the past few million years of human evolution (Table 1). By extrapolation, these data suggest that during primate evolution, LINE retrotransposition may have contributed to as much as ≈50,000 deletion events removing ≈40 Mb of genomic sequences (Table 1). These remarkable figures serve to illustrate the profound implication of TEs as drivers of genome evolution.

The number of human-specific L1-L1 recombination deletion events identified by Han et al. (7) is relatively modest. Yet, the L1-L1 recombination process appears to be the major contributor of LINE-mediated genomic instability in terms of amounts of deleted sequences during recent human evolution (Table 1). This is because L1-L1 recombination deletions tend to be larger (>6,000 bp on average) than other types of LINE-mediated deletions (<1,000 bp on average). Because larger deletions are more likely to disrupt functional regions of the genome than shorter deletions, the lower number of L1-L1 than Alu-Alu recombination deletions detected in the human genome may reflect, at least in part, negative selection against large, deleterious L1-L1 recombination deletions (9). Consistently, only one L1-L1 recombination event reported by Han et al. (7) is involved in exon deletions. In general, L1-L1 recombination deletions tend to be preferentially distributed in genomic regions of low gene density (7) where they are less likely to be deleterious and thus negatively selected. For example, the largest deleted sequence reported by Han et al., which is also the largest TE recombination-associated deletion reported to date with ≈64,000 bp, only encompasses pseudogene and intergenic sequences. In fact, the vast majority of genomic sequences deleted by L1-L1 recombination events are L1 sequences and Han et al. suggest that this deletion process may significantly contribute to counteract the increase in genome size caused by new L1 insertions. If so, there may be an appreciable rate of L1 sequence turn over in the genome, thus further implicating LINE retrotransposition in the evolutionary dynamics of the human genome.

Interestingly, some of the L1-L1 recombination events identified by Han et al. (7) may have been resolved by nonhomologous end joining, a mechanism involved in repair of DNA damage such as double-strand breaks (6, 21). Thus, L1-L1 recombination could occasionally contribute to maintain genome integrity by repairing DNA lesions, despite concomitant deletion of intervening genomic sequences in the process. Overall, the study by Han et al. demonstrates that by being responsible for the deletion of hundreds of thousands of nucleotides in the human genome during just the past few million years. More generally, the results of Han et al. add to the growing evidence showing that LINE retrotransposition has had, and continues to have, a considerable impact on primate genome evolution. By constantly challenging the integrity of our genome, the products of LINE retrotransposition create a tremendous amount of ongoing genomic fluidity and instability, thereby durably putting the human genome in the LINE of fire.

• © 2008 by The National Academy of Sciences of the USA