ATLAS: a system to selectively identify human-specific L1 insertions

Abstract

Retrotransposition of L1 LINEs (long interspersed elements) continues to sculpt the human genome. However, because recent insertions are dimorphic, they are not fully represented in sequence databases. Here, we have developed a system, termed "ATLAS" (amplification typing of L1 active subfamilies), that enables the selective amplification and display of DNA fragments containing the termini of human-specific L1s and their respective flanking sequences. We demonstrate that ATLAS is robust and that the resultant display patterns are highly reproducible, segregate in Centre d'Etude du Polymorphisme Humain pedigrees, and provide an individual-specific fingerprint. ATLAS also allows the identification of L1s that are absent from current genome databases, and we show that some of these L1s can retrotranspose at high frequencies in cultured human cells. Finally, we demonstrate that ATLAS also can identify single-nucleotide polymorphisms within a subset of older, primate-specific L1s. Thus, ATLAS provides a simple, high-throughput means to assess genetic variation associated with L1 retrotransposons.

Figure  1

Principle of ATLAS. a, Overview of LI structure. A schematic of a full-length Ta-1d L1 retrotransposon indicates the position of 5′ and 3′ UTRs, ORFs, and Ta-1d–specific and Ta-subfamily–specific primers. b, Principle of ATLAS, as depicted by a flow chart of the ATLAS procedure. Digestion of genomic DNA and ligation to a GC-rich oligonucleotide linker create a library of fragments flanked by defined sequences (A–C) (top). The denatured single-stranded DNA can self-anneal within the common linker sequences to form a frying pan–shaped intermediate that is inefficiently amplified (A and B). By contrast, subfamily-specific L1 primers can anneal within the frying pan (C), leading to extension products containing a single linker, which subsequently can undergo exponential amplification (bottom). Black bars indicate genomic DNA sequence, white bars indicate L1 sequence, and speckled bars indicate linker sequences. Restriction sites (R) that are compatible with the linker and the L1-specific (L1) and linker-specific (X) primers are shown.

Figure  2

Representative ATLAS array. Two independent ATLAS display patterns from the DNAs from three unrelated individuals (lanes 1–3) were generated with primers specific to either the 5′ UTR L1 Ta-1d subfamily (left) or the 3′ UTR L1 Ta subfamily (right). The libraries were labeled in duplicate, resulting in four display lanes per individual. Lane 4 reactions were conducted in the absence of restriction digestion; lane 5 reactions were conducted in the absence of genomic DNA; lane 6 reactions were conducted in the absence of T4 DNA ligase; lane 7 reactions were conducted in the absence of linker; and lane 8 reactions were conducted in the absence of primary amplification products. “A” and “C” indicate invariant bands, whereas “B” and “D” indicate variant bands. Size standards are indicated to the left of the gel.

Figure  3

ATLAS's reproducibility and robustness. a, High reproducibility. Genomic DNA samples (D1 and D2) from two unrelated individuals were digested, in duplicate, with AseI, and each digestion was ligated, in duplicate, to a compatible linker (L1–L4). Each of the purified ligation reactions (A1–A8) was amplified in duplicate, and the resultant primary amplifications (L1–L16) were radiolabeled in duplicate. The resultant ATLAS display patterns are shown for each of the 16 replicates. b, Robustness. Top, 5′ ATLAS performed, in duplicate, on 10 unrelated individuals (lanes 1–10). Controls (lanes 11–15) are the same as in figure 2. Bottom, Genomic DNAs (∼500 ng), used to generate the above 5′ ATLAS display patterns, resolved on a 0.8% agarose gel that contained ethidium bromide. Note the significant amount of degradation in the DNA sample from individual 3. MW = molecular weight markers; λ DNA = genomic DNA from bacteriophage lambda.

Figure  4

Dimorphic Ta-1d L1s, identified by 5′ ATLAS. a, Independent assay to assess the dimorphic status of Ta-1d L1s. PCR typing assays were developed to confirm the presence (filled)/absence (empty) status of Ta-1d L1s identified in the 5′ ATLAS reactions shown in panel b. Empty sites, lacking the L1 insertion, can be amplified with primers flanking the insertion site (e.g., RB907A and RB907B). Filled sites, containing the L1 insertion, can be amplified with a Ta-subfamily–specific L1 3′ UTR primer and a 3′ flanking primer (e.g., RB3PA and RB907B). b, Results of PCR typing experiments. Top, Amplification of empty site through PCR typing by use of the RB907A and RB907B primers. Bottom, Amplification of filled site through PCR typing by use of the RB3PA and RB907B primers. Note that individual DNA samples (1–10) can be homozygous for the presence of the L1 (F/F; lanes 2, 8, and 9), heterozygous (EF; lanes 1, 4, 5, and 7), or homozygous for the absence of the L1 (E/E; lanes 3, 6, and 10). The 11th reaction (Ctrl) is a control PCR with genomic DNA omitted. MW = molecular weight markers.

Figure  5

ATLAS products' segregation in families. 5′ Ta-1d–specific ATLAS was performed on a three-generation CEPH pedigree (1331). FF = paternal grandfather; FM = paternal grandmother; F = father; S/D = son/daughter; M = mother; MF = maternal grandfather; MM = maternal grandmother. “C1” denotes a control sample in which water was substituted for genomic DNA at digestion; “C2” denotes a control sample in which water was substituted for library DNA; and “C3” and “C4” denote control samples in which water was substituted for the primary PCR products in the labeling reaction. Arrows (“A”–“D”) indicate dimorphic array products. MW = molecular weight markers.

Figure  6

Use of ATLAS to identify SNPs within older L1s. a, Segregation of a dimorphic 5′ (Ta-1d) ATLAS product in CEPH family 1331. A dimorphic 5′ ATLAS display product (see “D” in fig. 5) is shown. b, Identification, through sequencing, of a cosegregating SNP at the L1 primer site. Primers placed outside the AseI restriction site and Ta-1d L1 primer-annealing site enabled PCR amplification of the product from each individual (1–16) in CEPH family 1331. Direct sequencing of the products revealed that an ATLAS display product occurred only when an individual inherited an allele, of the L1, containing an exact match to the Ta-1d primer-annealing site. Individuals 3 and 5 are heterozygous for the matched/mismatched sequence (arrows indicate the double G/C peak). In contrast, individuals 10 and 14 are homozygous for the mismatched allele (arrows indicate the single G peak).