Fig 1.

Structures of a canonical SVA element and an ancestral SVA2 element. (A) A full-length SVA is defined by four domains (65, 72, 74, 81) in order from its 5′ end: (i) a CT-rich repeat, with CCCTCT being the predominant repeat unit, (ii) a domain derived from two antisense Alu-like fragments, (iii) a variable number of GC-rich tandem repeats (VNTR) with a unit size of 35 to 50 bp (65), and (iv) sequence derived from the env gene and right LTR from an extinct HERV-K10 (SINE-R) (63). (B) SVAs are derived from a sequence, referred to as SVA2 (17, 34, 44), which lacks all of the canonical SVA domains except the VNTR and contains a 3′ sequence not found in canonical SVAs (3′-U). Genomic SVA insertions terminate in a 3′ poly(A) tail and are flanked by target site duplications of various lengths (black horizontal arrows).