Active Alu retrotransposons in the human genome

Abstract

Alu retrotransposons evolved from 7SL RNA approximately 65 million years ago and underwent several rounds of massive expansion in primate genomes. Consequently, the human genome currently harbors 1.1 million Alu copies. Some of these copies remain actively mobile and continue to produce both genetic variation and diseases by "jumping" to new genomic locations. However, it is unclear how many active Alu copies exist in the human genome and which Alu subfamilies harbor such copies. Here, we present a comprehensive functional analysis of Alu copies across the human genome. We cloned Alu copies from a variety of genomic locations and tested these copies in a plasmid-based mobilization assay. We show that functionally intact core Alu elements are highly abundant and far outnumber all other active transposons in humans. A range of Alu lineages were found to harbor such copies, including all modern AluY subfamilies and most AluS subfamilies. We also identified two major determinants of Alu activity: (1) The primary sequence of a given Alu copy, and (2) the ability of the encoded RNA to interact with SRP9/14 to form RNA/protein (RNP) complexes. We conclude that Alu elements pose the largest transposon-based mutagenic threat to the human genome. On the basis of our data, we have begun to identify Alu copies that are likely to produce genetic variation and diseases in humans.

Figure 1.

A genome-wide view of human Alu activity. A total of 850,044 full-length (>268 bp) genomic Alus were identified in hg18 of the reference human genome sequence and assigned to known Alu subfamilies. Alu elements frequently have sequence changes relative to consensus sequences. The number of changes for each full-length copy is indicated on the x-axis; the copy number for a given level of sequence variation is indicated on the left y-axis. Pink data points mark the mobilization activities of the 89 Alu copies that were examined in this study (labeled on the right y-axis). In sum, 8 AluJ, 27 AluS, and 54 AluY copies were tested, spanning a range of subfamilies and variation levels. Note that elements with fewer changes relative to consensus sequences (zero changes) generally had the highest levels of activity; no elements below 10% variation (28 changes) were active. Please see Supplemental Table 1 for additional details and error measurements.

Figure 2.

Alu mobilization assays. (A) Alu retrotransposition assay (Dewannieux et al. 2003). (1) Alus were cloned into a test plasmid containing the 7SL polIII enhancer and a neo retrotransposition selection cassette. The cassette contains a neo G418 resistance gene that is interrupted by the self-splicing tetrahymena intron. (2) Upon polIII transcription, the tetrahymena intron is spliced out. (3) When cotransfected with L1 ORF2p, Alu RNAs are reverse transcribed along with the neo gene, and (4) integrated into the genome, conferring G418 resistance. (5) After a 2-wk treatment with G418, resistant colonies are stained, photographed, and counted. (B) Assay results for a sample of genomic AluS elements. Activities are given relative to AluYa5 activity (100%) within each assay. Each horizontal bar indicates the mean of multiple independent assays (dots). Each dot represents the average of a single (triplicate) experiment. Gray vertical bars represent 95% confidence intervals. The percent consensus identity is indicated below each element. (C) AluY subfamily results. (D) Mobilization activities of known polymorphic AluY elements and a polymorphic AluSx. The dbSNP ss numbers are listed for each. (E) Resurrected AluJo and Sx elements. The mobilization results for artificially constructed 100% consensus AluSx and AluJo elements are compared with a highly conserved (but inactive) genomic AluJo (Jo_h10.1).

Figure 3.

How many potentially active Alu core elements in the human genome? A model was developed for estimating the number of potentially active Alu core elements in the reference human genome. A set of 33 unbiased Alu copies (see Methods) were placed into four bins according to their level of sequence variation (96.8%–100%, 93.4%–96.7%, 90%–93.3%, and <89.9%). The percentage of “active copies” was calculated for each bin, where “active” was defined as >5% of AluYa5 activity level in the mobilization assay. The percentages of active copies within each bin were then used to estimate how many genomic copies with the same levels of variation are present in the human genome (the results are depicted as numbers). The levels of activity were broken down further into “hot” (red; 100%–66.6% of AluYa5 activity level), “moderate” (yellow; 66.5%–40% of AluYa5), and “cool” (blue; 39.9%–5% of AluYa5). No elements below 90% conservation were active in the mobilization assay. This method provides a liberal estimate of the number of Alu core elements that would be active if cloned and tested in our mobilization assay. The actual number of elements expressed and mobilized from their natural genomic locations is likely to be lower than the numbers presented due to the impact of flanking genomic regions on Alu expression (see Discussion).

Figure 4.

SRP9/14 host proteins are necessary for efficient Alu retrotransposition. (A) Secondary structure representation of an AluYa5 RNA. The 124 positions (Supplemental Fig. 1) that are conserved among 70 consensus sequences and 45 experimentally tested, active Alu elements are highlighted in blue (Alu 5′ domain) and cyan (Alu 3′ domain). Positions (G25 and G159) that were mutated to prevent SRP9/14 binding are in magenta. Hash marks indicate major (magenta) and minor (gray) SRP9/14 contact sites. The inset shows an alternative base-pairing that is possible since the emergence of the AluS family and that may be responsible for the drop in SRP9/14 affinity at the transition from the AluJ to S families. Thin curves indicate U-turns, the thick, curved bar indicates a stacking interaction, and the double arrow a flexible linkage. Circles indicate tertiary base pairs between the loops. The dotted circles symbolize an alternative base pair of the respective nucleotides with G14 (left monomer) and G148 (right monomer). Additional symbols: (|) Watson-Crick base pairs; (.) wobble base pairs; (x) other (potential) base pairs. (B) Three-dimensional representation of Alu RNA and SRP9/14 binding. Positions highlighted in A have been mapped onto the crystal structure of the SRP Alu RNP (Weichenrieder et al. 2000) using the same colors. SRP9 is in red and SRP14 is in green. The 5′ and 3′ RNA ends are indicated by a large and small sphere, respectively. (C) Retrotransposition activity of consensus AluYa5 and SRP9/14 binding mutants. The mobilization activities relative to AluYa5 are shown along with representative assay plates below. (D) Relative affinities of Alu RNA mutants for SRP9/14. Left and right mutant Alu RNA monomers competed against the wild-type AluYa5 left monomer RNA (Ya5 (L.), white bars) or AluYa5 right monomer RNA (Ya5 (R.), gray bars) as labeled references in an in vitro assay based on nitrocellulose filter binding of SRP9/14. Positive values of ΔΔG reflect loss of affinity with respect to wild-type RNA. The full data set is presented in Supplemental Table 2. A representative binding experiment is shown in Supplemental Figure 2. (E) Relative affinities of Alu consensus sequences for SRP9/14. In contrast to D, truncated SRP RNA (SRP(t.)) was used as reference. These binding results agree with previous gel-shift assays examining SRP9/14 binding to AluS and Y monomers (Sarrowa et al. 1997).

Figure 5.

Model for Alu retrotransposition. Alu RNA competes with 7SL RNA for SRP9/14 binding and RNP formation. It appears that at least one SRP9/14 heterdimer is necessary for Alu mobilization, although the binding of two heterodimers provides more efficient mobilization. Alu RNPs, once formed, can dock on ribosomes. As L1 mRNA is translated, the poly(A) tail of an SRP9/14-bound Alu competes for nascent L1 ORF2 reverse transcriptase (Sinnett et al. 1991; Boeke 1997; Dewannieux et al. 2003; Mills et al. 2007). Finally, a new Alu sequence is inserted into the genome by target-primed reverse transcription (Luan et al. 1993). Modern Alu RNAs have evolved weaker SRP9/14 binding affinities, perhaps to disengage from SRP9/14 more readily during reverse transcription.