A Cultured Cell Retrotransposition Assay
In 1996, an assay was developed to monitor L1 retrotransposition in cultured human cells (25). In this assay, the 3' UTR of a candidate L1 was tagged with an indicator cassette designed to detect retrotransposition events (Fig. 1B). The initial retrotransposition indicator cassette (mneol) consisted of a selectable marker (NEO) containing its own promoter and poly-adenylation signal. The mneol cassette was introduced into the L1 so that the L1 and reporter gene were in opposite transcriptional orientations. The mneol gene also was interrupted by an intron in the same transcriptional orientation as the L1 (26). This arrangement ensured that G418-resistant foci would arise only when a transcript initiated from the promoter driving L1 expression underwent retrotransposition.
Using this assay, it was demonstrated that various L1s could retrotranspose at a high efficiency when expressed from an extrachromosomal expression vector in HeLa cells (25,27). Subsequent characterization of the resultant retrotransposition events revealed L1 structural hallmarks, indicating that retrotransposition was faithfully reconstituted in HeLa cells (25,28,29). Alternative retrotransposition indicator cassettes (30) and high throughput versions of the cultured cell retrotransposition assay (27,31) have been developed in recent years and these assays have been used to uncover several aspects of L1 biology:
1. Characterize functional domains in the L1-encoded proteins (22,25,32).
2. Identify retrotransposition-competent L1s in the human and mouse genomes (33-36).
Fig. 1. (A) The structure of a retrotransposition competent L1 (not to scale). A retrotransposition competent L1 contains a 5' UTR (yellow), two open reading frames—ORF1 (blue) and ORF2 (green), a 3' UTR (red), and a poly (A) tail. The direction of transcription from both the L1 promoter and the antisense promoter in the 5' UTR (L1 ASP) are denoted with arrows. Transcription factor binding sites (YY1, RUNX3, and SOX) are indicated. ORF1 contains a leucine zipper domain near its amino terminus (wavy lined box) and several conserved basic amino acid motifs near its carboxyl terminus (dotted lines). A 63-bp intergenic region (black line) separates ORF1 and ORF2. ORF2 contains an endonuclease domain (EN), a reverse transcriptase domain (RT), and a cysteine-rich domain (C). A putative NXF1(TAP) binding site is located in the region spanning the end ORF2 and the beginning of the 3' UTR (underlined and labeled). L1 usually is flanked by variable length target site duplications in the genome (bold black arrows). (B) The cultured cell retrotransposition assay. A retrotransposition-competent L1 is tagged with a retrotransposition indicator cassette (mneol; orange, labeled backwards neo) that has its own promoter and poly (A) signal (orange slashed). The cassette contains an intron in the same transcriptional orientation as the L1 (denoted by SD for splice donor and SA for splice acceptor). When this construct is transfected into HeLa cells, G418 resistant colonies will only be present if there is transcription from the L1 promoter, splicing, reverse transcription, and integration of the L1 transcript. Cells then are fixed and stained to view colonies. Representative results of this assay are shown in which HeLa cells were transfected with a wild type L1 construct (WT L1) or an L1 construct containing a mutation in the reverse transcriptase domain (RT- L1).
3. Demonstrate that the Ll-encoded proteins exhibit a cis-preference for their encoding RNA (7,9).
4. Show that the Ll-encoded proteins can act in trans to mobilize nonautonomous retrotrans-posons and cellular mRNAs (7-9).
5. Demonstrate that Lls can retrotranspose into genes and can mobilize sequences derived from their 5' and 3' flanks to new genomic locations (1,29,37,38).
6. Uncover an endonuclease-independent (but RT-dependent) L1 retrotransposition pathway (30).
7. Discover that L1 retrotransposition is associated with various forms of genetic instability in transformed cultured cells (28-30).
In addition to characterizing disease-producing L1 insertions and employing the cultured cell retrotransposition assay, there are three principal ways to study L1 biology (11). The first approach is to use biochemical and molecular biological techniques to characterize the Ll-encoded proteins and Ll RNA. These experiments have been conducted primarily on active human and mouse Lls and have been instrumental in characterizing enzymatic activities encoded by the L1 proteins (see Functional Analyses).
The second approach is to utilize phylogenetic and evolutionary analyses to examine Lls present in whole genome sequences. These methodologies have allowed the classification of human Ll subfamilies (39) and have been instrumental in identifying candidate retrotrans-position-competent Lls in the human, mouse, and rat genomes (1,40,41). More recently, comparative genomics and molecular biological approaches have been useful in examining intra- and interorganism differences in Ll content that exist between available genome sequences (42-47).
The third approach is to study L1 expression and/or retrotransposition in animal models. The use of animal models remains in its infancy, but their application has shown that full-length sense strand Ll RNA is expressed in both male and female germ cells and that human Lls can retrotranspose in the male germline prior to the onset of meiosis II (48-51). A combination of the previously described approaches has led to a greater understanding of the mechanism of Ll retrotransposition and its effects on the genome.
FUNCTIONAL ANALYSES The 5' UTR
Human Lls have a 910-bp 5' UTR that contains an internal RNA polymerase II promoter (16). To remain autonomously mobile, the Ll promoter must be able to direct transcription such that an entire Ll transcription unit can be faithfully retrotransposed to a new genomic location. Transcription has been proposed to be a limiting step in retrotransposition and most promoter activity is attributed to the first 600 bases of the 5' UTR.
A YYl-binding site, located at +l2 to +2l on the non-coding strand, is required for proper initiation of transcription at or near the first nucleotide of Ll (Fig. lA) (38,52,53). Lls lacking a functional YYl binding site can initiate transcription and retrotranspose in cultured cells, but the resultant progeny likely will not regenerate the complete 5' UTR. Thus, it is predicted that over successive rounds of retrotransposition, Lls lacking a functional YYl-binding site will become progressively shorter and ultimately retrotransposition-defective.
The Ll 5' UTR also contains two SOX binding sites at +472 to +477 and +572 to +577, as well as a RUNX3 binding site at +83 to+l0l (Fig. lA) (54,55). These sequences can bind SOX
and RUNX3 proteins in vitro, respectively. Moreover, mutations in the RUNX3 and SOX sites results in a decrease in L1 retrotransposition in cultured cells, whereas overexpression of RUNX3 protein in HeLa cells results in an increase in L1 transcription and a nearly twofold increase in retrotransposition (; Athanikar and Moran, unpublished data). However, the mechanism by which the RUNX3 and SOX proteins affect L1 transcription and retrotrans-position merits further study.
Finally, the 5' UTR contains a potent antisense promoter (L1 ASP) that can influence the expression of cellular genes (Fig. 1A) (56,57). Transcripts of cellular genes originating from the L1 ASP have been identified in both cDNA libraries and expressed sequence tag databases. Most of these transcripts were spliced correctly, but the majority only contained part of the coding region of a gene. Thus, some L1 ASP-derived transcripts have the potential to encode alternative, perhaps functional, forms of a native cellular protein. It also is intriguing to speculate that some L1 ASP-derived transcripts may act to regulate the expression of their resident gene.
Human ORF1 encodes a 338 amino acid RNA binding protein (p40 or ORF1p) that co-localizes with L1 RNA in cytoplasmic ribonucleoprotein particles (RNPs), which are proposed retrotransposition intermediates (18,19). Elegant biochemical studies conducted on mouse L1 ORF1p indicate that it possesses RNA binding and nucleic acid chaperone activities and can form a homotrimer in vitro (58-61). The amino terminus of ORF1p is predicted to fold into a coiled-coil domain that is important for ORF1p-ORF1p interactions (19,62). Human ORF1p contains a leucine zipper motif that is required for retrotransposition, perhaps by stabilizing L1 RNPs, by interacting with host factors important for retrotransposition, or by functioning at downstream steps in the retrotransposition process (; Hulme and Moran, unpublished data). Interestingly, the amino terminus of ORF1p is evolving rapidly yet mouse, rabbit, and rat L1s lack the leucine zipper domain. This rapid evolution may reflect adaptive changes that have occurred in response to host repression mechanisms over the course of L1 evolution (63).
The carboxyl terminus of ORF1p contains a number of conserved motifs that are rich in basic amino acids. Mutations in these amino acids severely reduce L1 retrotransposition in cultured cells and may affect RNA binding and nucleic acid chaperone activity of mouse ORF1p in vitro (25,58,62). However, the carboxyl terminal domain of ORF1p lacks overt sequence similarity to any known RNA-binding proteins. Thus, more biochemical studies are required to elucidate the structural and functional role of ORF1p in L1 retrotransposition.
ORF2 has the potential to encode a 150-kDa protein that has been difficult to detect in vivo (13,20). Sequence comparisons and crystallization studies indicated that the 5' terminus of L1 ORF2, as well as analogous sequences from related non-LTR retrotransposons, share homology with apurinic/apyrimidinic (AP) endonucleases (22,64,65). Subsequent biochemical analyses revealed that the amino terminus of L1 ORF2p has endonuclease activity in vitro. L1 EN purified from Escherichia coli neither shows a preference for cleaving abasic substrates in vitro nor possesses 3'-5' exonuclease or RNase H activities that are common to other AP endonucleases (66). Instead, it makes a site-specific single-stranded endonucleolytic nick at the degenerate consensus sequence (5'TTTT/A; where the "/" indicates the scissile phosphate), exposing a 3' hydroxyl residue and 5' monophosphate (66,67). Mutations in the putative L1 EN
active site abolish endonucleolytic cleavage activity in vitro and reduce L1 retrotransposition in HeLa cells by two to three orders of magnitude (22,25).
The central region of ORF2p shares homology to the reverse transcriptase domains encoded by other non-LTR retrotransposons (68,69). Biochemical studies demonstrated that ORF2p has RT activity that can extend homopolymer/oligonucleotide primer template complexes in vitro and requires divalent cations (Mg+2 is favored over Mn+2) (21,70). Mutations in the putative L1 RT active site abolish RT activity in vitro and reduce L1 retrotransposition in HeLa cells by approximately three orders of magnitude (21,25).
The carboxyl terminus of ORF2p contains a conserved cysteine-rich motif (CX3CX7HX4C) of unknown function (23). Site-directed point mutations in conserved cysteine and histidine residues within the C-domain reduce human L1 retrotransposition in HeLa cells by two orders of magnitude, but do not affect L1 RT activity as measured in a yeast expression assay (25,71). Thus, the C-domain provides a function required for retrotransposition that apparently is distinct from L1 EN or L1 RT.
The 3' UTR
The 3' UTR of mammalian L1s has a relatively weak poly (A) signal as well as a conserved guanosine-rich polypurine tract that, based on studies of rat L1s, is predicted to form an intrastrand tetraplex (25,37,72). Sequence spanning the 3' end of ORF2 and the beginning of the 3' UTR contains at least two cis-acting repeats (5'-CACA [N5] GGGA-3') that can bind the nuclear export factor NXF1(TAP) and this sequence functions as a nuclear RNA export signal in vitro (Fig. 1A) (73). Paradoxically, despite the presence of these conserved and/or functional sequences, the L1 3' UTR is dispensable for retrotransposition in HeLa cells (25). Thus, the function of the L1 3' UTR in L1 retrotransposition remains enigmatic.
Was this article helpful?