SVA elements are nonautonomous retrotransposons that cause disease in humans

Abstract

L1 elements are the only active autonomous retrotransposons in the human genome. The nonautonomous Alu elements, as well as processed pseudogenes, are retrotransposed by the L1 retrotransposition proteins working in trans. Here, we describe another repetitive sequence in the human genome, the SVA element. Our analysis reveals that SVA elements are currently active in the human genome. SVA elements, like Alus and L1s, occasionally insert into genes and cause disease. Furthermore, SVA elements are probably mobilized in trans by active L1 elements.

Figure 1

The sequence of SVA retrotransposition events. A full-length SVA element of unknown origin retrotransposed into the empty-site sequence on chromosome 4, represented on contigs AC037439 and AC068352. During subsequent transcription of the new SVA at this site, the SVA poly A was bypassed in favor of a downstream poly A, producing the first transduction event. The full-length SVA and transduction 1 inserted into a new genomic location at target site 2 on chromosome 3 to produce the sequence found in contig AC016142. During subsequent transcription of the new SVA on chromosome 3, both the SVA poly A and the poly A from transduction 1 were bypassed in favor of a poly A farther downstream, producing the second transduction event. The transcript containing a full-length SVA element and both transduction events inserted into target site 3 on chromosome 1, which is in exon 5 of the α-spectrin gene. However, the integration process resulted in a structure that was 5′ truncated and inverted compared with the precursor, a common process in L1-mediated retrotransposition. RNA is represented by a dashed line; DNA is represented by a solid line.

Figure 2

Structure of a full-length SVA element. Most SVA elements in the genome are flanked by L1-like TSDs. The 5′ end of a full-length SVA element consists of a variable number of (CCCTCT) hexameric repeats, followed by sequence with homology to antisense Alu sequence (arrows represent approximate areas of homology), a VNTR region, a SINE-R region, and an AATAAA poly A signal, followed immediately by a poly A tail, that is A(n). SINE-R sequence is related to the 3′ LTR sequence of HERV-K10 (regions of identity and percent identity indicated).

Figure 3

Analysis of the SVA VNTRs. ClustalW alignment of the 35 SVA_SPTA1 VNTRs (A) and associated cladogram (B), which shows evidence of VNTR expansion by unequal homologous recombination. Notice that the individual VNTRs from a block of VNTRs (18–23) are most closely related to the corresponding individual VNTRs from block 24–29; that is, VNTR18 is most similar to VNTR24, VNTR19 is most similar to VNTR25, VNTR20 is most similar to VNTR26, etc.