Ribonucleoprotein particle formation is necessary but not sufficient for

Human Molecular Genetics, 2005, Vol. 14, No. 21 doi:10.1093/hmg/ddi354 Advance Access published on September 23, 2005

3237–3248

Ribonucleoprotein particle formation is necessary but not sufficient for LINE-1 retrotransposition Deanna A. Kulpa1 and John V. Moran1,2,* 1

Department of Human Genetics and 2Department of Internal Medicine, University of Michigan Medical School, Ann Arbor, MI 48109-0618, USA

Received June 16, 2005; Revised August 17, 2005; Accepted September 15, 2005

INTRODUCTION Long interspersed element-1 (LINE-1 or L1) comprises 17% of human DNA, and its mobility (i.e. retrotransposition) has had a tremendous impact on genome evolution (1,2). The vast majority of L1s are retrotransposition-defective; however, the average human genome, as represented by the Human Genome Working Draft (HGWD), is estimated to contain 80 –100 retrotransposition-competent L1s (3,4). Human L1 retrotransposition continues to result in genetic diseases and the generation of novel polymorphisms (reviewed in 5). Retrotransposition-competent human L1s are 6 kb in length and contain a 50 untranslated region (UTR) harboring an internal promoter (6,7), two non-overlapping open reading frames (ORF1 and ORF2) and a short 30 UTR that ends in a poly (A) tail (8,9). ORF1 encodes a 40 kDa nucleic acid binding protein (p40 or ORF1p; 10 –12), whereas ORF2 encodes a protein (ORF2p) with demonstrated endonuclease (EN) and reverse transcriptase (RT) activities (13 – 15). Genetic experiments indicate that both ORF1p and ORF2p demonstrate a profound cis-preference and preferentially associate with their encoding transcript to form a cytoplasmic ribonucleoprotein particle (RNP), which is a proposed

retrotransposition intermediate (11,16 – 18). After RNP formation, the L1 RNA gains access to the nucleus where it is reverse transcribed into a cDNA copy most likely via target site primed reverse transcription (TPRT) (14,19,20). Integration of the resultant L1 cDNA and the completion of retrotransposition can occur by a variety of mechanisms, and has been associated with genetic instability (21 – 25). Although ORF1p is required for retrotransposition, its exact function(s) remains unknown. In vitro experiments have shown that the amino terminus of human and mouse ORF1p contain an alpha helical coiled-coil that is required for ORF1p – ORF1p protein interactions (11,26), whereas the carboxyl terminal domain of ORF1p has demonstrated nucleic acid binding and nucleic acid chaperone activities (12,27–29). Whether ORF1p binds to specific sequences in L1 RNA remains controversial; however, it is possible that ORF1p binding to L1 RNA may protect the transcript from degradation. It is also possible that ORF1p interacts with host proteins that facilitate retrotransposition and/or is important for initial steps in TPRT. Previous studies examining the in vivo function of ORF1p have relied on its production in cell lines that naturally express high levels of endogenous L1s. Human ORF1p and L1 RNA were detected in cytoplasmic RNPs of 2102Ep and

*To whom correspondence should be addressed at: 1241 E. Catherine, 4909 Buhl, University of Michigan, Ann Arbor, MI 48109, USA. Tel: þ1 7346150456; Fax: þ1 7347633784; Email: [email protected]

# The Author 2005. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]

Downloaded from http://hmg.oxfordjournals.org/ by guest on June 9, 2013

Long interspersed elements (LINE-1s or L1s) are abundant non-LTR retrotransposons that mobilize through an RNA intermediate by target site primed reverse transcription. The L1-encoded proteins (ORF1p and ORF2p) preferentially associate with their encoding transcript to form a ribonucleoprotein particle (RNP), which is a proposed retrotransposition intermediate. Here, we have used epitope tagging to discriminate the proteins encoded by engineered L1s from those encoded by endogenously expressed L1s. We demonstrate that an L1 containing an epitope tag at the carboxyl terminus of ORF1p remains retrotranspositioncompetent and that tagged ORF1p and its encoding RNA localize to cytoplasmic RNPs. We also identified two classes of ORF1p mutants, one that severely decreased RNP formation and blocked retrotransposition, and another that allows RNP formation but reduces retrotransposition by 100-fold. Thus, these data indicate that RNP formation is important but not sufficient for L1 retrotransposition and suggest that ORF1p also may function at downstream steps in the L1 retrotransposition pathway.

3238

Human Molecular Genetics, 2005, Vol. 14, No. 21

RESULTS EtORF1p is compatible with L1 retrotransposition We engineered an 11 amino acid T7 gene10 epitope tag onto the carboxyl terminus of L1.3 ORF1p and assayed the resultant construct, pDK101, for retrotransposition in HeLa cells (Fig. 1A; 30). The 30 UTR of the L1 contains a retrotransposition indicator cassette (mneoI ), consisting of a backward copy of the neomycin phosphotransferase reporter gene (Neo ) that is interrupted with an intron, which is in the same transcriptional orientation as the L1 (Fig. 1A; 15,31). This arrangement ensures that G418-resistant foci will arise only when L1 undergoes retrotransposition (Fig. 1B; 15). We demonstrated that the resultant L1 could retrotranspose at efficiencies comparable to the positive control, which lacks the epitope tag (pJM101/L1.3; Fig. 1C).

EtORF1p detection is delayed from full-length L1 constructs We next examined the etORF1p expression profile in HeLa cells transfected with either pDK101 or pDK500. pDK500 contains a deletion of all ORF2 coding sequences, but otherwise is identical to pDK101 (Fig. 2A, see Materials and Methods). Transient transfection assays conducted with pDK500 revealed etORF1p expression 3 days post-transfection (Fig. 2A). By comparison, transient transfection assays conducted with pDK101 revealed a delay in etORF1p detection, which

Figure 1. Diagram of pDK101 and the cultured cell retrotransposition assay. (A) Diagram of pDK101. pJM101/L1.3 was modified to contain the T7 gene 10 eleven amino acid epitope tag at the C-terminus of ORF1p (black flag). The L1 contains its own 50 UTR, as well as the heterologous cytomegalovirus (CMV) immediate early promoter on the pCEP4 episomal vector (labeled boxes). The L1 30 UTR is tagged with the mneoI indicator cassette (labeled black box). The backward Neo gene is interrupted by IVS2 sequence from the g-globin gene in the forward orientation (diagonal slashed box). Splice donor (SD) and splice acceptor (SA) sites are indicated. The SV40 promoter (P0 , horizontal slashed rectangle) and the thymidine kinase polyadenylation signal (cross-hatched rectangle) drive the expression of the Neo gene. The construct also has a gene to confer resistance to hygromycin B (Hygro) as well as origins of replication for prokaryotic (pUC ori) and mammalian (EBNA1A/OriP) cells. (B) Diagram of the cultured cell retrotransposition assay. Transcripts originating from the SV40 promoter P0 will not initially confer G418 resistance to cells due to the presence of the engineered intron. After retrotransposition of a spliced L1 transcript, transcripts arising from P0 will be able to express a functional Neo gene product, which will confer G418 resistance to the host cell (15). (C) pDK101 is retrotransposition competent. pDK101 retrotransposes at similar efficiencies when compared with the untagged parent construct (pJM101/L1.3). The ORF2 mutant is an RT-active site mutant (D702A; Fig. 7A). The ORF1 mutant is a putative nucleic acid binding domain mutant (RR261–262AA; Fig. 7A). The number of cells transfected per well are indicated.

peaked at 7 – 9 days post-transfection (Fig. 2B; see Discussion for further interpretation). Due to the fluctuations in signal intensity and optimal harvest day observed in the transient assays, we sought to maximize our ability to reproducibly detect etORF1p by exploiting the hygromycin phosphotransferase gene present


NTera2D1 cells (11), and various isoforms of mouse ORF1p and L1 RNA were detected in cytoplasmic RNPs of mouse F9 teratocarcinoma cells (16), leading to the proposal that RNPs are likely retrotransposition intermediates. Although informative, these studies probably examined a heterogeneous population of ORF1p molecules generated from both retrotransposition-competent and retrotranspositiondefective L1s. Thus, the characteristics of RNPs derived from a retrotransposition-competent L1 remain unknown. The development of a cultured cell retrotransposition assay has been instrumental in both assessing the retrotransposition capacity of different L1s and elucidating individual steps in the L1 retrotransposition pathway (15, reviewed in 2). Here, we have engineered an epitope tag onto the carboxyl terminus of ORF1p (etORF1p), which allowed us to follow the fate of wild-type and mutant L1 proteins from transfected L1 constructs. We show that the resultant epitope-tagged wild-type L1 remains retrotransposition-competent, that etORF1p and its encoding RNA co-localize in cytoplasmic RNPs under various experimental conditions and that the localization of etORF1p to RNPs is independent of the function of ORF2p. In addition, we show that some retrotransposition-defective mutants in the putative ORF1p nucleic acid binding domain adversely affect the ability of ORF1p to localize to RNPs, whereas others retain the capacity to localize to RNPs. Together, these data indicate that RNP formation is important but not sufficient for L1 retrotransposition and suggest that ORF1p may function at multiple steps in the L1 retrotransposition pathway.


on the backbone of pDK101 (Fig. 1A; 15). After initiating hygromycin selection on Day 3 post-transfection, we observed a significant increase in etORF1p signal from Days 5 to 11 (Fig. 2C). As predicted from previous analyses (7,15), expression of etORF1p was more robust from pDK101 (CMVþ/L1 50 UTRþ) when compared with an expression construct containing only the L1 50 UTR (pDK171; Fig. 2D). EtORF1p detection after hygromycin selection was highly reproducible using this protocol and therefore was used for all subsequent experiments. EtORF1p and L1 RNA localize to a cytoplasmic RNP Previous studies showed that the majority of ORF1p is present in a large cytoplasmic RNP that sedimented at 160 000 g during differential centrifugation (11). To examine the partitioning of etORF1p, we performed differential centrifugation

Figure 3. Differential centrifugation of pDK101 transfected HeLa cells. (A) etORF1p is in the p160 000g fraction. The flow chart indicates the scheme to determine the cellular compartmentalization of etORF1p. The boxed areas of the flow chart indicate the source material analyzed on the western blot using the T7 antibody. Pellet fractions were resuspended according to the original packed cell volume (see Materials and Methods). EtORF1p was observed predominantly in the p160 000g fraction (indicated by arrow beside the western blot). A small amount was also observed in the 800g fraction. P and S stand for the pellet and supernatant fractions, respectively. Untransfected HeLa WCL was included as an antibody negative control. (B) Polyclonal ORF1p antibody detects etORF1p. A rabbit polyclonal antibody to human ORF1p was used to examine the 160 000g pellet and supernatant fractions. pDK101 p160 000g fractions show a specific signal identical in size to that seen on T7 western blots as indicated by the arrow. No ORF1p was detected in the 160 000g supernatant fraction. Untransfected HeLa 160 000g pellet and supernatant fractions were included as a negative control.

on extracts of HeLa cells that were transfected with pDK101. EtORF1p is detected predominantly in the 160 000 g pellet fraction (p160 000 g; Fig. 3A), and we confirmed that a protein of identical size was also detected with an ORF1 polyclonal antibody (Fig. 3B; 32). Notably, we also observed a faint band in the 800 g pellet fraction, which is enriched for nuclei and unlysed cells (Fig. 3A). However, experiments using marker antibodies revealed that the etORF1p detected in the 800 g fraction likely is derived from unlysed cells or cytoplasmic contamination (Supplementary Material, Fig. S1). To determine if L1 RNA co-localized with etORF1p in the p160 000 g fraction, we isolated total RNA from the 160 000 g pellet and supernatant fractions derived from either pDK101-, pDK171- or JM101/L1.3-transfected HeLa cells. RNA dot blot hybridization using a riboprobe specific to the sense strand of the Neo gene revealed an enrichment of L1 RNA in the p160 000 g fraction when compared with an equal amount of RNA harvested from the supernatant fraction (Fig. 4A and B). To further characterize the products identified in the RNA dot blot analysis, we performed 30 RACE on total RNA derived from the p160 000 g fraction of pDK101-transfected HeLa cells (Fig. 4A). We detected


Figure 2. EtORF1p expression in HeLa cells. (A) Diagram of the etORF1p expression construct pDK500. All ORF2 sequences were deleted from pDK101 to create pDK500 (see Materials and Methods). pDK101 and pDK500 whole cell lysates (WCLs) were harvested on Day 3 after transfection, and protein production was monitored by T7 western blotting. Samples were normalized for total protein. The western blot shows that the 40 kDa etORF1p (arrow) is detectable 3 days after transfection when expressed from pDK500, but not from the full-length construct pDK101. The bands seen migrating at 50 and 30 kDa are commonly observed cross-reacting products present also in the untransfected HeLa WCLs. (B and C) Time course of etORF1p detection from pDK101. pDK101-transfected HeLa cells were harvested at each time point from Day 2 through Day 11 after transfection, and WCLs were analyzed by western blotting with an anti-T7 antibody. The arrow indicates the position of etORF1p. (B) Ten T-1752 flasks of HeLa cells were transfected with pDK101. A single flask was harvested for each time point through Day 11 and WCLs were prepared and analyzed as described (see Materials and Methods). (C) Ten T-1752 flasks of HeLa cells were transfected with pDK101 and were selected with 200 mg/ml hygromycin B starting on Day 3 after transfection. A single flask was harvested for each time point until Day 11. WCLs were prepared and analyzed as for non-hygromycin selected cells. (D) Reduction of etORF1p from pDK171. The expression of etORF1p from pDK101 was compared with that from pDK171, a construct which lacks the pCEP4 CMV promoter. The arrow indicates the presence of etORF1p on the T7 western blot. Both samples were normalized for total protein, and both constructs are in a full-length L1 context.

3239

3240


Figure 4. L1 RNA in RNP complexes. (A) Location of primers and probes for L1 RNA detection. For the RNA dot blot, gray solid arrows above the Neo gene show the orientation of the sense and antisense riboprobes. The probes flank the engineered intron of the Neo gene and do not include intron sequences (dashed line). The Neo sense riboprobe detects the L1 promoterdriven transcript. The Neo437S and 30 RACE primers were used for 30 RACE analyses and are indicated on the expanded sequence below the Neo gene in the diagram. The relative locations of both the cryptic and the engineered SDs and SAs are indicated in the diagram. Spliced regions are indicated by the ‘V’ shaped dashed and solid black lines. The Neo437S primer sequence is detailed in Materials and Methods. The Ambion 30 RACE primer anneals to the poly(A) tail of the L1 RNA transcript. (B) RNA dot blot analysis. RNA was extracted from 160 000g pellet and supernatant material from pDK101, pDK171 or JM101/L1.3 transfected HeLa cells. The Neo sense riboprobe indicates the presence of the L1 construct RNA in the pDK101, pDK171 and JM101/L1.3 p160 000g fractions, but not in the corresponding supernatant fractions, or the untransfected HeLa background control. (C) 30 RACE analysis. 30 RACE was performed on RNA isolated from pDK101 and untransfected HeLa p160 000g material. The Neo437S and 30 RACE primer pair were used for PCR of the cDNA, and arrows indicate the presence of the expected intronless 1168 bp product, along with the presence of the cryptic splice product of 1056 bp. An unspliced product of 2070 bp is not observed. No product is observed in untransfected HeLa and water control lanes. As a control for RNA sample integrity, RT– PCR was performed using a primer to GAPDH (198 bp product). All 30 RACE products were cloned and sequenced to confirm identity.

an 1168 bp product, confirming the presence of spliced L1 RNA in the p160 000 g fraction (Fig. 4C). We also observed a 1056 bp product, which is due to the existence of a cryptic splice site in the Neo gene (Fig. 4A and C; 22). No unspliced product of 2070 bp was observed. We next tested the ability of etORF1p to remain associated with the p160 000 g complex after enzymatic or chemical treatments (Fig. 5A). After treatment with 5 U of RNase A or 500 mM sodium chloride, etORF1p was released from the resuspended 160 000 g pellet (R) and was detected in the supernatant (Su) fraction (Fig. 5B). In contrast, the localization of etORF1p was not affected by mock incubation, treatment with DNase I or treatment with low salt (Fig. 5B). Together, these data are in agreement with previous findings

(11,33) and show that etORF1p is predominantly present in a cytoplasmic RNP. Sedimentation properties of etORF1p and L1 RNA in sucrose gradients To examine the sedimentation profile of etORF1p and L1 RNA, pDK101 whole cell lysates were analyzed by centrifugation in 7 –47% continuous sucrose gradients in the presence of either Mg2þ or EDTA. In the presence of Mg2þ, etORF1p and L1 RNA were detected throughout the gradient, although the majority of complexes tracked with assembled ribosomes and polyribosomes (Fig. 6A). By comparison, whole cell lysates layered over a gradient containing 30 mM EDTA revealed a shift in both etORF1p and, to a lesser extent, L1 RNA to the upper portion of the gradient that is enriched for dissociated polyribosomes (Fig. 6B). Thus, these data indicate that the majority of etORF1p co-migrates with L1 RNA during velocity sedimentation and that these RNP complexes are stabilized by Mg2þ (see Discussion for further interpretation). ORF2p mutants do not affect etORF1p RNP localization Having established a system to specifically track the fate of etORF1p in HeLa cells, we next assessed whether mutants in ORF2p would affect etORF1p RNP formation (Fig. 7A). We hypothesized that catalytic mutants in the EN and RT domains of ORF2p would not affect etORF1p RNP formation, as these activities are thought to act exclusively during TPRT. Consistently, catalytic mutants in either the EN (H230A) or RT (D702A) domains that severely reduce retrotransposition in the cultured cell assay had no affect on the ability of etORF1p to localize to a cytoplasmic RNP (Fig. 7A and B; 14,15,18). We next tested whether ORF2p synthesis was required for etORF1p RNP formation by assaying a retrotranspositiondefective mutant containing a frameshift-induced stop mutation


Figure 5. EtORF1p biophysical characterization. (A) Flow chart of the biophysical characterization. The p160 000g material from pDK101-transfected cells was subjected to enzymatic or chemical treatments. The flow chart depicts the scheme of the treatments. Boxed areas indicate the material used in the western blot analysis. All blots used the T7 antibody to detect etORF1p. (B) EtORF1p is in RNP complex. Western blots show that RNase A and 500 mM NaCl treatments release etORF1p into the soluble fractions (Su). Treatment with DNase I or lower salt concentrations do not affect etORF1p localization to the 160 000g RNP fraction (R). The horizontal arrows indicate the presence of the 40 kDa etORF1p. Mock treated material was used as an incubation control.


3241

near the amino terminus of ORF2 that apparently abolishes ORF2p synthesis (Fig. 7A; 34). EtORF1p can be detected in an RNP complex fraction when expressed from the resultant construct, suggesting that ORF2p synthesis is dispensable for RNP formation (pDK138; Fig. 7B).

EtORF1p mutants affect RNP formation and L1 retrotransposition Previously, three highly conserved amino acid motifs in the carboxyl domain of ORF1p were shown to be important for retrotransposition (REKG235-238, RR261-262 and YPAKLS282-287; Fig. 7A; 15). Thus, we tested whether these retrotransposition-defective mutants would affect the ability of etORF1p to localize to an RNP. None of the etORF1p mutants could be detected in the p160 000 g fraction when the whole cell lysates were prepared in the absence of Mg2þ (Fig. 7C). However, in the presence of Mg2þ, both the REKG235-238AAAA and YPAKLS282-287AAAALA etORF1p proteins could be detected at low levels in the p160 000 g fraction, although the signals were never as

robust as wild-type L1 (Fig. 7C). No mutant etORF1p was ever observed in the 160 000 g supernatant fraction (Supplementary Material, Fig. S2). By comparison, the RR261262AA mutant etORF1p was not detected in whole cell lysates, the 160 000 g pellet or 160 000 g supernatant fractions, although we cannot exclude the possibility that very low amounts of the protein are present in some of these fractions (Fig. 7C; Supplementary Material, Fig. S2; data not shown). Notably, we could detect RR261-262AA etORF1p at a low level in whole cell lysates when expressed from a construct lacking ORF2 sequences (see pDK500; Fig. 1), thereby assuring that the protein itself was capable of being produced (Supplementary Material, Fig. S3). Together, these data suggest that mutations in the carboxyl terminus of ORF1p alter ORF1p-L1 RNA interactions and RNP formation, resulting in a severe reduction in retrotransposition.

RNP formation is not sufficient for retrotransposition We next examined the consequences of mutations in the di-arginine motif of ORF1p by making a series of constructs


Figure 6. Localization of etORF1p in sucrose gradient fractionation. (A) EtORF1p sedimentation in Mg2þ sucrose gradient. pDK101-transfected HeLa whole cell lysates were layered over 7 –47% sucrose containing Mg2þ. The lowest number fraction corresponds to the top of the gradient. OD254 values were plotted against time for the peak profile. The T7 and S6 antibodies were used on western blots to detect etORF1p and S6 ribosomal marker proteins, respectively. 30 RACE was used to detect the L1 construct specific RNA in gradient fractions. EtORF1p co-localizes predominantly with polyribosomes in the bottom portion of the gradient. There is a significant overlap between the etORF1p and L1 RNA signals in the gradient, although L1 RNA co-migrates with all ribosome-containing fractions. Arrows indicate the expected sizes of etORF1p, S6 and L1 RNA, respectively (B) EtORF1p sedimentation in EDTA sucrose gradient. pDK101-transfected HeLa whole cell lysates were layered over 7– 47% sucrose containing 30 mM EDTA. ORF1p shifts to the upper portion of the gradient and predominantly co-localizes with the dissociated polyribosome complexes. The L1 RNA also shifts, with several denser fractions containing lower amounts of the RNA.

3242


Figure 7. The effect of ORF1 and ORF2 mutants on etORF1p localization. (A) Diagram of ORF1 and ORF2 mutations in L1 sequence and associated retrotransposition efficiencies. The relative positions of the mutated amino acids along with their corresponding pDK construct names are indicated below the figure. Below each is a boxed number corresponding to the retrotransposition efficiency of the mutant, given as a percent of the retrotransposition efficiency of pDK101 in the cultured cell assay. Diagrammed below the RR261–262AA mutant are the single and double lysine substitution mutants. ORF2 mutants at the N-terminus or ORF2, the EN domain, or the RT domain also are indicated. (B) ORF2 mutants do not affect etORF1p localization. T7 western blots show that etORF1p is present in the p160 000g fraction from ORF2p EN-(H230A), RT-(D702A) and ORF2 null expression mutants (ORF2 Stop). All samples were collected in the presence of magnesium. Arrows indicate etORF1p. (C) Some etORF1p mutants do not form RNPs. ORF1p in non-magnesium containing p160 000g fractions are shown in T7 western blots. ORF1p is not detectable from the REKG235–238AAAA, RR261–262AA or YPAKLS282–287AAAALA mutants in the absence of Mg2þ. Magnesium containing p160 000g fractions show a reduction in etORF1p signal from REKG235–238AAAA and a faint signal from YPAKLS282– 287AAAALA on T7 western blots. The etORF1p from mutant RR261–262AA is undetectable. Arrows indicate etORF1p. The band at 50 kDa is a crossreacting product that sometimes is observed, see also Figure 1A and Supplemental Figure 2. (D) Some etORF1p mutants form RNPs. Single and double lysine substitutions allow for the ready detection of the mutant etORF1p in an RNP. RR261– 262KK, R262K and R261K all show etORF1p in the p160 000g fraction in the absence (left panel) or presence (right panel) of Mg2þ on T7 western blots.

containing arginine to lysine substitutions (i.e., RR261-262KK, R261K and R262K; Fig. 7A). When both arginines were replaced by lysines (RR261-262KK), retrotransposition was reduced to less than 1% of wild-type levels (Fig. 7A). A similar result was observed for the R261K mutation (Fig. 7A). By comparison, the R262K mutant had little effect on retrotransposition. Significantly, each mutant etORF1p localized to the p160 000 g fraction when lysates were prepared in either the presence or absence of Mg2þ (Fig. 7D). We note that the R262K mutant protein was detected reproducibly at higher levels as a broader migrating band when compared with either the wild-type, RR261-262KK or R261K mutant proteins. Together, these data indicate a possible separation of function between RNP formation and L1 retrotransposition, and suggest that ORF1p may function at steps downstream of RNP formation in the L1 retrotransposition pathway (Fig. 8).

DISCUSSION Previous attempts to study the biochemical activities associated with ORF1p have either relied on in vitro expression systems or exploited cultured cell lines that naturally express high levels of ORF1p and L1 RNA. The first approach has been useful in identifying amino acids important for ORF1p multimerization, L1 RNA binding and nucleic acid chaperone activity (10,11,26–29,35). Although valuable, these assays remove ORF1p from its native expression context, making it difficult to assess definitively whether the discovered activities are required for retrotransposition per se. The second approach has been useful in identifying potential RNP retrotransposition intermediates that contain both ORF1p and L1 RNA (11,12,36). However, those RNP particles likely are comprised ORF1p and L1 RNA molecules derived from both retrotranspositioncompetent and/or retrotransposition-defective L1s. Thus, it is


Figure 8. Model for the role of ORF1p in L1 retrotransposition. ORF1p binding to L1 RNA in cis results in RNP formation. Mutation of conserved amino acid residues in the basic C-domain of ORF1 suggests several ORF1p functions. Class I mutants (RR261–262AA, REKG235–238AAAA and YPAKLS282–287AAAALA) altering and/or reducing the efficiency of ORF1p to associate with its template severely reduce L1 retrotransposition, indicating the importance of ORF1p-mediated RNP formation. Class II mutants (RR261–262KK and R261K) enable RNP formation, but severely reduce retrotransposition, indicating that ORF1p also may act at downstream steps in TPRT.


to contaminating cytoplasmic components that co-fractionate with nuclei (Supplementary Material, Fig. S1). Moreover, we have found that mutations altering residues in the catalytic domains of ORF2p or that abolish ORF2p translation have no detectable affect on etORF1p RNP formation (40). Thus, we conclude that ORF1p binding to its template does not appear to be coordinated through ORF2p and that etORF1p is likely present at low steady-state levels in the nucleus. Interestingly, difficulty in detecting a nuclear pool of ORF1p has also been reported for mouse L1s, the Drosophila melanogaster I-factor and Xenopus laevis Tx1 elements (41 – 43). Therefore, the conundrum of ORF1p localization seems common to a variety of distantly related non-LTR retrotransposons. The ability to follow etORF1p encoded from a single L1 has allowed us to parse out two classes of retrotranspositiondefective ORF1p mutants: those that form altered RNPs containing less etORF1p (class I) and those that form RNP complexes that apparently are indistinguishable from wild-type (class II). For the class I mutants (REKG235-238AAAA, RR261-262AA and YPAKLS282-287AAAALA), etORF1p was not detected in RNPs when they were prepared in the absence of Mg2þ. Two of the mutants (REKG235-238AAAA and YPAKLS282-287AAAALA) were observed at low levels when RNPs were prepared in the presence of Mg2þ. These mutants may have a reduced RNA binding affinity, which consequently may lower the amount of etORF1p in an RNP, adversely affecting the L1 RNP structure, and resulting in a loss of retrotransposition potential (Fig. 8). Interestingly, recent biochemical studies performed on mouse ORF1p suggest that a mutation analogous to the human RR261-262AA mutation (mouse RR297-298AA) exhibits a 25-fold decrease in RNA binding affinity, which is consistent with the notion that ORF1p RNA binding contributes to its stability in an RNP (29). Together, these experiments lead us to conclude that etORF1p localization to an RNP is a necessary step in L1 retrotransposition, because blocking its localization and/or altering the amount of etORFp in the RNP is correlated strongly with an inability to retrotranspose. Class II mutants (RR261-262KK and R261K) allowed for the detection of etORF1p in the cytoplasmic RNP fraction. However, both mutations severely reduced retrotransposition. As noted above, a similar biochemical ‘phenotype’ also was observed for catalytic mutations in the EN or RT domains of ORF2p, which are hypothesized to act solely during TPRT. Thus, we hypothesize that class II mutants affect downstream steps in L1 retrotransposition, perhaps by interfering with RNP transport or initial steps of TPRT (Fig. 8). Consistent with the second possibility, Martin et al. (29) have shown that analogous mutations in mouse ORF1p reduce its DNA melting and strand exchange activities in vitro. However, it cannot be ruled out that the RR261-262KK and R261K mutations subtly perturb RNP structure or inhibit post-translational modification of ORF1p, which, in turn, inhibits RNP transport, nuclear import or other steps of retrotransposition downstream of RNP formation. In either case, these data support the hypothesis that RNP formation is necessary, but not sufficient for retrotransposition and that ORF1p may be required both for RNP formation and for downstream steps in L1 retrotransposition. In conclusion, we have developed a system that allows us to study etORF1p in cytoplasmic RNPs. Although we have


difficult to assess whether or not they represent bona fide retrotransposition intermediates or instead are ‘dead end’ complexes produced by a cohort of retrotransposition-defective L1s. Here, we have examined the fate of etORF1p molecules encoded by either retrotransposition-competent or retrotransposition-defective L1s in cultured HeLa cells against a background of endogenously expressed L1s. The majority of wild-type etORF1p co-sediments with L1 RNA in a macromolecular RNP. EtORF1p was readily dissociated from the RNP complex by treating it with either 500 mM NaCl or RNase A, but was not affected by treatment with DNase I. Despite extensive efforts, we did not detect a soluble pool of etORF1p when it was expressed from a full-length L1. Consistent with the cis-preference model of L1 retrotransposition (9,17,18), these data suggest that the majority if not all wild-type etORF1p is bound to L1 RNA and is not freely available to form complexes on other cellular RNAs. The velocity sedimentation analyses indicate the presence of a variety of etORF1p/L1 complexes. For example, some fractions toward the top of the gradient contained L1 RNA, but little or no etORF1p, which may reflect nascent ORF1p chains that still lack the carboxyl terminal epitope tag. Similarly, some fractions toward the bottom of the gradient contain relatively little L1 RNA and more etORF1p, which may represent fully assembled L1 RNPs that contain more than one ORF1p per L1 RNA. Thus, future studies are needed to determine whether all the complexes or only a subset represent bona fide L1 retrotransposition intermediates. As expected, etORF1p accumulates as cells undergo hygromycin selection when the L1 was expressed from either its native 50 UTR or a vector containing both the CMV promoter and 50 UTR. However, although both the above vectors exhibit similar retrotransposition efficiencies in HeLa cells, etORF1p RNPs were reproducibly observed at higher quantities when the L1 was expressed from the CMVþ50 UTRþ context. These data suggest that retrotransposition is not strictly correlated with the number of etORF1p RNP complexes in a cell and that greater RNP production does not necessarily equate to more L1 retrotransposition events. Thus, steps downstream of RNP formation may be rate limiting for L1 retrotransposition. The time course experiment in Figure 2 suggests a delay in etORF1p detection that is manifested when etORF1p is expressed from the context of a full-length L1. It has been demonstrated previously that full-length L1 transcript production is inefficient, owing to overall ‘A’ richness of L1 RNA (37,38). Thus, the detection of etORF1p signal from pDK500 at the Day 3 time point relative to its full-length counterpart pDK101 could be due to an increase in efficient transcript production from pDK500. However, it is also possible that the delay in etORF1p detection from a full-length L1 reflects co-ordinate translational control of ORF1p and ORF2p. Clearly, similar studies conducted with a synthetic mouse L1 could help resolve this issue (39). Although etORF1p was found predominantly in the p160 000 g fraction, differential centrifugation experiments also revealed a small amount in the 800 g pellet fraction, which is enriched for nuclei and associated cytoskeletal components. However, co-localization experiments with marker antibodies revealed that this pool of etORF1p is likely due

3243

3244


demonstrated that the presence of ORF2p is not required for ORF1p RNP formation, ORF2p association with its RNA template is a requirement for L1 retrotransposition. The system we have described will enable future studies to link L1 RNA, ORF1p and ORF2p to the cytoplasmic RNP, and should allow us to elucidate their stoichiometric requirements for L1 retrotransposition.

MATERIALS AND METHODS Cell lines

Plasmid constructs All plasmids are based on the previously described pJM101/ L1.3 construct in the pCEP4 (Invitrogen) backbone unless otherwise indicated (30). The RR261-262AA and ORF2 active site mutants were subcloned from previously described plasmid constructs (18). Site-directed mutations for other described ORF1 mutants were engineered using PCR mutagenesis, and nucleotide numbers refer to L1.3 accession number L19088 (44). pDK101 is pJM101/L1.3 modified by PCR mutagenesis to contain the T7 gene 10 eleven amino acid (MetAlaSerMetThrGlyGlyGlnGlnMetGly) epitope tag. pDK500 contains a deletion of half the inter-genic region, all of ORF2 and the 50 end of the 30 UTR from pDK101. It was created by engineering an Spe I restriction site at nucleotide 1982 in the inter-ORF region of a pBluescript KSþ version of pDK101 by PCR mutagenesis. The resultant plasmid was digested with Spe I, the 50 overhang was filled in with DNA Polymerase I, Large (Klenow) Fragment (NEB), and that DNA was digested with Not I. The liberated 1983 bp fragment was ligated to a 12.6 kb Bst Z17I – Not I digested pDK101 vector backbone, which subsequently deleted all ORF2 sequences, but retained a portion of the 30 UTR and the mNeoI cassette. pDK171 was created by sub-cloning the AgeI (nucleotide 1898 near the 30 end of ORF1) to Bst Z17I (at nucleotide 5972 of L1.3) fragment from pDK101, containing the ORF1 T7 epitope tag, into pJM101DCMV (15), creating the CMV 2 /50 UTRþ ORF1 epitope-tagged construct pDK171. pDK136 contains the sequence GCT starting at nucleotide 2676 in ORF2 to substitute Ala for His (H230A, EN domain active site mutant; 18). pDK135 contains the sequence GCT starting at nucleotide 4093 in ORF2 to substitute Ala for Asp (D702A, RT domain active site mutant; 18). pDK138 contains an AG insertion starting at nucleotide 2019 in ORF2 to create a frameshift and substitute a double stop for LeuThr (34).

Transient cultured cell retrotransposition assay The transient cultured cell retrotransposition assay was previously described (45). Briefly, HeLa cells were plated at 2 105, 2 104 and 2 103 in six-well tissue culture dishes. Twenty-four hours after plating, the cells were transfected with 1 mg of Qiagen Plasmid Midi column prepared DNA per well using FuGENE-6 from Roche Molecular Biochemicals as a vehicle. Three days post-transfection, the medium was exchanged for DMEM-complete supplemented with 400 mg/ml G418 (Gibco), and exchanged daily until selection was complete, approximately 12 days post-transfection. On Day 12, the medium was aspirated and cells were gently washed with cold 1 phosphate-buffered saline (PBS), then fixed with 2% formaldehyde, 0.2% glutaraldehyde in 1 PBS overnight at 48C. The fixed cells were thoroughly washed with water and then stained with 0.4% Bromophenol Blue overnight at room temperature. Three-day transient transfection with pDK500 and pDK101 HeLa cells were plated at 6 106 cells per T-1752 tissue culture flask, a time point referred to as Day 0. Twenty-four hours post-plating, the cells were transfected with 30 mg Qiagen Plasmid Midi column prepared DNA per flask using FuGENE-6 (45). An untransfected HeLa cell flask was kept as a background control. On Day 2 post-transfection, the medium was exchanged on each flask with DMEM-complete. On Day 3 post-transfection, the cells were harvested and whole cell lysates prepared as described in ‘Preparation of whole cell lysates’. Time course with pDK101-transfected HeLa cells Twenty flasks of HeLa cells were plated in parallel at 6 106 cells per T-1752 tissue culture flask. At 24 h, the cells were transfected with 30 mg Qiagen Plasmid Midi column prepared DNA per flask using FuGENE-6 as a vehicle. For the 10 flasks


HeLa cells were maintained as a monolayer in Dulbecco’s high glucose modified Eagle medium (DMEM) (Gibco) supplemented with 10% fetal bovine calf serum, 0.29 mg/ml L -glutamine and 100 U/ml penicillin –streptomycin (Gibco; DMEM-complete). Cells were incubated at 378C at 100% humidity in 7% CO2 and passaged using standard cell culture techniques.

pDK102 contains the sequence GCAGCGGCAGCT starting at nucleotide 1620 in ORF1 of pDK101 to substitute amino acids ArgGluLysGly to four alanines (REKG235 – 238AAAA mutant; 15). pDK105 contains the sequence GCCGCG starting at nucleotide 1698 in ORF1 to substitute two alanines for ArgArg (RR261– 262AA mutant, 18). pDK116 contains the sequence GCTGCAGCCGCACTAGCC starting at nucleotide 1761 in ORF1 to substitute AlaAlaAlaAlaLeuAla for TyrProAlaLysLeuSer (YPAKLS282 –287AAAALA mutant; 15). pDK106 contains the sequence AAGAAG starting at nucleotide 1698 in ORF1 to substitute two lysines for ArgArg (RR261– 262KK mutant). pDK107 contains the sequence AAG starting at nucleotide 1698 in ORF1 to substitute one lysine for Arg (R261K mutant). pDK108 contains the sequence AAG starting at nucleotide 1701 in ORF1 to substitute one lysine for Arg (R262K mutant).


of non-hygromycin-selected cells, the media were exchanged daily with DMEM-complete for the 11 day time course. For the 10 flasks of pDK101-transfected cells undergoing hygromycin selection, 3 days post-transfection the media were exchanged for DMEM-complete supplemented with 200 mg/ml hygromycin B (Invitrogen), and exchanged daily until 11 days post-transfection. Untransfected HeLa cells, included as a negative control, were plated in a T-1752 flask 3 days before the final harvest day, fed daily with DMEM-complete and harvested at 90% confluency. The time course experiment involved harvesting one flask per time point from both the non-hygromycin- and hygromycinselected groups and preparing a whole cell lysate fraction on the day of harvest. This lysate was flash frozen in a dry ice/ethanol bath and stored at 2808C until the time course was complete. Whole cell lysates were prepared as described below.

On harvest day, transfected HeLa cells were washed in the flask 3 with cold 1 PBS. The cells were then scraped into 10 ml of cold 1 PBS, transferred to a 15-ml conical tube and spun at 3000 g for 5 min at 48C. The PBS was aspirated and the cells were lysed in the tube by the addition of 1 ml 1.5 mM KCl, 2.5 mM MgCl2, 5 mM Tris –HCl, pH 7.4, 1% deoxycholic acid, 1% Triton X-100, 1 Complete Mini EDTA-free Protease Inhibitor Cocktail (Roche Molecular Biochemicals) per 0.5 ml packed cell volume (PCV). The cells were resuspended in the lysis buffer by gentle pipetting and incubated on ice for 5 min. After incubation, cell debris and nuclei were removed by centrifuging at 3000 g for 5 min at 48C. The supernatant was transferred to a new tube and saved as a cleared whole cell lysate fraction for western blot analysis. On average, 1/40th of a whole cell lysate fraction would be analyzed per sample on each western blot. Untransfected control HeLa cells flasks were harvested in parallel with each experiment using the same protocol. Differential centrifugation HeLa cells were plated and transfected as described above for T-1752 flasks with pDK101. Transfected cells were fed daily with DMEM-complete supplemented with hygromycin B until selection was complete at 10 days post-transfection. On harvest day, the cells were then washed in the flask by rinsing 3 with cold 1 PBS, then scraped into 10 ml cold 1 PBS and transferred to a 15-ml conical tube. Cells were spun at 3000 g for 5 min at 48C to pellet. To dounce homogenize, the PBS was aspirated off and cells were resuspended in 6.7 ml homogenization buffer (10 mM Tris – HCl, pH 8.0, 1 mM EDTA, 10 mM NaCl, 10 mM MgCl2, 1 Complete Mini EDTA-free Protease Inhibitor Cocktail) per 1 ml PCV and transferred to a 7 ml dounce homogenizer (Kontes Glass Company). The cells were then allowed to swell on ice for 10 min. Cells were disrupted using 30 strokes of an A pestle while still on ice, and then transferred to a fresh 15 ml conical tube for centrifugation at 800 g for 10 min at 48C. After centrifugation, the supernatant was carefully transferred to a new tube. The 800 g pellet was resuspended in

resuspension buffer (10 mM Tris– HCl, pH 8.0, 1 mM EDTA, 150 mM NaCl, 10 mM MgCl2, 1 Complete Mini EDTAfree Protease Inhibitor Cocktail) at one-fifth volume of PCV. The 800 g supernatant was spun at 12 000 g for 10 min 48C. After centrifugation, the 12 000 g pellet was resuspended in resuspension buffer as for the 800 g pellet at one-fifth of PCV. The 12 000 g supernatant was transferred to a new tube and spun at 160 000 g for 2 h at 48C. The 160 000 g supernatant was transferred to a separate tube and the 160 000 g pellet was resuspended in the same conditions as for the other pellet fractions. Samples were flash frozen in a dry ice/ethanol bath and stored at 2808C. Twenty-five microliters of each fraction was used for western blot analysis.

RNP ultracentrifugation Magnesium-containing preparations. Cells were plated, transfected and selected on hygromycin as described earlier at 6 106 cells per T-1752 tissue culture flask. Transfection was performed 24 h post-plating with 30 mg Qiagen Plasmid Midi column prepared DNA per flask using FuGENE-6 as a vehicle. On Day 10 post-transfection, cells were washed, lysed and the whole cell lysate fraction was prepared as described earlier in 1 ml Mg2þ lysis buffer (1.5 mM KCl, 2.5 mM MgCl2, 5 mM Tris –HCl, pH 7.4, 1% deoxycholic acid, 1% Triton X-100, 1 Complete Mini EDTA-free Protease Inhibitor Cocktail) per 0.5 ml PCV. The whole cell lysate fraction was ultracentrifuged in a Beckman Coulter Optima MAX-E ultracentrifuge at 160 000 g for 1.5 h to concentrate the RNP fraction. After ultracentrifugation, the supernatant was transferred to a new tube and the pellet fraction was resuspended in a small volume (50 – 100 ml) of dH2O þ 1 Complete Mini EDTA-free Protease Inhibitor cocktail. All samples were either used fresh or flash frozen in a dry ice/ethanol bath and stored at 2808C until used for western blot analysis. Non-magnesium protein preparations. HeLa cells were transfected with each construct in T-1752 flasks in conditions identical to the magnesium-containing preparations. On harvest day, the cells were washed 3 with cold 1 PBS. The cells were then scraped into 10 ml cold 1 PBS and transferred to a 15 ml conical tube. Cells were spun into a pellet at 3000 g for 5 min at 48C. The PBS was aspirated off the cell pellet and the cells were lysed by the addition of 1 ml non-Mg2þ lysis buffer (300 mM NaCl, 50 mM Tris – HCl, pH 7.4, 0.5% Triton X-100, 1.0 mM NEM, 1 Complete Mini EDTA-free Protease Inhibitor Cocktail) per 0.5 ml PCV. Cells were gently resuspended by pipetting and allowed to sit on ice for 1 h. After lysis, cell debris and nuclei were removed by centrifuging at 3000 g for 5 min at 48C. The supernatant (non-magnesium whole cell lysate fraction) was transferred to a new tube and ultracentrifuged at 160 000 g for 1.5 h to concentrate RNP fraction. After ultracentrifugation, the supernatant was removed to a new tube and the pellet was brought up in a small volume (50 – 100 ml) dH2O þ 1 Complete Mini EDTA-free Protease Inhibitor Cocktail. All samples were either used fresh or flash frozen


Preparation of whole cell lysates

3245

3246


in a dry ice ethanol bath and stored at 2808C until used for western blot analysis.

antisense radiolabeled riboprobes were created from this clone (46). Membranes were prehybridized and hybridized using standard protocols.

Western blot

RNA preparation To obtain the 160 000 g RNP pellet and supernatant starting material, samples were prepared using the magnesiumcontaining protocol described in ‘RNP Ultracentrifugation’. This material was used to isolate RNA by employing a TRIzol extraction protocol following the manufacturer’s instructions. Approximately one-third of the sample collected from an individual T-1752 flask was used as starting material for the extraction (30 ml of 100 ml of the 160 000 g pellet material, 300 ml of 1000 m l the 160 000 g supernatant material). The isolated RNA was resuspended in 10 –20 ml RNA Storage Solution from Ambion and then quantified by spectrophotometry. RNA dot blot RNA dot blot analysis of the 160 000 g pellet and supernatant-extracted RNA was performed on nitrocellulose membranes using standard techniques as detailed in Molecular Cloning (second edition) with a Schleicher and Schuell 96well minifold (46). Two nitrocellulose membranes were set up in parallel for the sense and antisense probes using 10 mg RNA. The probe was a 495-bp region of the Neo gene. The probe was created first by the making of a PCR product using primers Neo 437S (50 -GAGCCCCT GATGCTCTTCGTCC-30 ) and 1808As (50 -CGCACGT TAGGTAGAACAAGTTAC-30 ) from an intronless Neo template. The 495-bp intronless Neo PCR product was subcloned into pBluescript KS2 at the Sma I site. Sense and

30 RACE 30 RACE was performed on 0.25 mg RNA using Ambion’s 30 RACE RT-PCR kit and M-MLV reverse transcriptase (Promega). The RNA was reverse-transcribed according to the manufacturer’s protocol, and used as a substrate for PCR with the Ambion 30 RACE primer and Neo437S primer of the neomycin gene in the L1 construct (Neo437S primer: 50 -GAGCCCCTGATGCTCTTCGTCC-30 ). Control RT-PCR reactions for GAPDH were performed using the Ambion 30 RACE primer and a primer specific to the 30 end of GAPDH RNA (GAPDH primer: 50 -GACCCTCACTGCTG GGGAGTCC-30 ). 30 RACE products were cloned into pGEM T-Easy (Promega) and sequenced to confirm their identity. Biophysical characterization pDK101-transfected HeLa cell whole cell lysates were prepared as described for magnesium-containing samples. The whole cell lysate then was ultracentrifuged at 160 000 g for 1.5 h. The supernatant was aspirated off and the pellet material resuspended in dH2O, 1 Complete Mini EDTAfree Protease Inhibitor cocktail. Pellet material from several flasks was pooled and then divided into equal 100 ml portions for individual treatments. The control material was incubated at 378C for 60 min. For RNase A treatment, 5 U of RNase A (Roche) was added to 100 ml and incubated for 30 min at 378C. Five units of DNase I (Roche) plus 10 DNase I buffer were added to 100 ml of pellet material and incubated at 378C for 60 min. For NaCl treatments, 5 M NaCl was added to each 100 ml to give final concentrations of 100, 250 and 500 mM . All samples were incubated on ice for 60 min. Following the treatments, 0.5 M EDTA to a final concentration of 30 mM EDTA was added to all tubes. Then they were placed in an ethanol/dry ice bath to quench the reaction. The material was quickly thawed and spun at 160 000 g for 1.5 h, supernatant fraction retained and pellet material brought up in 20 ml dH2O, 1 Complete Mini EDTA-free Protease Inhibitor cocktail, all of which were loaded onto the protein gel for western blot analysis using the T7 antibody as described earlier. Additionally, 1 mg of RNA and DNA ladder was incubated with RNase A or DNase I for 1 h, and then visualized after agarose gel electrophoresis to control for enzyme activity (data not shown). Sucrose gradients Mg2þ 7 –47% sucrose gradients were prepared by layering 47% sucrose in 80 mM NaCl, 5 mM MgCl2, 20 mM Tris – HCl, pH 7.4, 1 mM DTT, 1 Complete Mini EDTA-free Protease Inhibitor Cocktail under 7% sucrose in the same buffer. For EDTA gradients, the same buffer and sucrose concentrations were used, with the substitution of 30 mM EDTA for 5 mM MgCl2. The gradient was formed using a BioComp Gradient Master using the manufacturer’s


Protein concentrations of each sample were determined prior to loading using Bio-Rad Bradford assay reagent following the manufacturer’s instructions, and were normalized for total protein. Protein samples were prepared with Laemmli sample buffer at a final concentration of 1 and run on prepoured denaturing 10% polyacrylamide gels from Bio-Rad at 200 V for 30 min. Gels were transferred to nitrocellulose membrane using Towbin transfer buffer (25 mM Tris and 192 mM glycine) in a Bio-Rad semi-dry transfer apparatus at 15 V for 45 min following the manufacturer’s instructions. After the transfer, membranes were blocked for 1 h at room temperature in 5% non-fat dry milk, 0.1% Tween and 1 TBS. Membranes were rinsed 3 with 0.1% Tween and 1 TBS. For T7 blots, membranes were incubated in 5% nonfat dry milk, 0.1% Tween, 1 TBS with T7-HRP antibody at 1:50 000 for 1 h at room temperature. For ORF1p blots, the rabbit polyclonal anti-ORF1 antibody was a generous gift from Thomas Fanning, and used at 1:5000 in 0.1%Tween, 1 TBS, followed by a goat anti-rabbit HRP conjugate secondary antibody. For S6 blots, the S6 ribosomal protein antibody from Cell Signaling Technology was incubated in 5% BSA, 1 TBS, 0.1% Tween at 1:1000 overnight at 48C. All blots were developed using Novagen SuperSignal substrate following the manufacturer’s instructions.


instructions. Whole cell lysates from pDK101-transfected HeLa cells (prepared as described earlier) were layered over the gradient and spun for 2 h at 39 000 rpm (178 000 g ) at 48C in an SW-41 rotor. Gradient fractions were harvested using a BioComp Piston Gradient Fractionator and Pharmacia Biotech UV-MII at UV254 collecting 30-s fractions. Data collection was performed using ADInstruments PowerChrom recorder and software. To concentrate the material, each fraction was ultracentrifuged at 160 000 g for 2 h, at 48C, and fraction pellets were brought up in 75 ml dH2O þ 1 Complete Mini EDTA-free Protease Inhibitor Cocktail. Samples were divided into thirds, one-third for RNA extraction using TRIzol and two-thirds for western blot analysis using T7 and S6 antibodies.

SUPPLEMENTARY MATERIAL

ACKNOWLEDGEMENTS We thank members of the University of Michigan Sequencing Core for help with sequencing and Dr Jose´ Luis Garcia Perez and Ms Amy Hulme for comments and for critically reading the manuscript. We thank Thomas Fanning and Benjamin Margolis for generously providing antibodies. The work was supported in part by a grant from the National Institutes of Health (GM60518). D.A.K. was supported in part by an NIH training grant (GM07544). The University of Michigan Cancer Center helped defray some of the DNA sequencing costs. Conflict of Interest statement. None declared.

REFERENCES 1. International Human Genome Sequencing Consortium (2001) Initial sequencing and analysis of the human genome. Nature, 409, 860 –921. 2. Moran, J.V. and Gilbert, N. (2002) Mammalian LINE-1 retrotransposons and related elements. In Craig, N.L. et al. (eds), Mobile DNA II. ASM Press, Washington DC, pp. 836– 869. 3. Sassaman, D.M., Dombroski, B.A., Moran, J.V., Kimberland, M.L., Naas, T.P., DeBerardinis, R.J., Gabriel, A., Swergold, G.D. and Kazazian, H.H. (1997) Many human L1 elements are capable of retrotransposition. Nat. Genet., 16, 37 –43. 4. Brouha, B., Schustak, J., Badge, R.M., Lutz-Prigge, S., Farley, A.H., Moran, J.V. and Kazazian, H.H. (2003) Hot L1s account for the bulk of retrotransposition in the human population. Proc. Natl Acad. Sci., 100, 5280–5285. 5. Hulme, A., Kulpa, D.A., Garcia Perez, J.L. and Moran, J.V. (2005) The impact of LINE-1 retrotransposition on the human genome. In Stankiewicz, P. and Lupski, J. (eds), Genomic Disorders: The Genomic Basis of Disease, in press. 6. Swergold, G.D. (1990) Identification, characterization, and cell specificity of a human LINE-1 promoter. Mol. Cell. Biol., 10, 6718–6729. 7. Athankiar, J.N., Badge, R.M. and Moran, J.V. (2004) A YY1 binding site is required for accurate human LINE-1 transcription initiation. Nucleic Acids Res., 32, 3846–3855. 8. Scott, A.F., Schmeckpeper, B.J., Abdelrazik, M., Comey, C.T., O’Hara, B., Rossiter, J.P., Cooley, T., Heath, P., Smith, K.D. and Margolet, L. (1987) Origin of the human L1 elements: proposed progenitor genes deduced from a consensus DNA sequence. Genomics, 1, 113 –125.

9. Dombroski, B.A., Mathias, S.L., Nanthakumar, E., Scott, A.F. and Kazazian, H.H. (1991) Isolation of an active human transposable element. Science, 254, 1805–1808. 10. Holmes, S.E., Singer, M.F. and Swergold, G.D. (1992) Studies on p40, the leucine zipper motif-containing protein encoded by the first open reading frame of an active human LINE-1 transposable element. J. Biol. Chem., 267, 19765–19768. 11. Hohjoh, H. and Singer, M.F. (1996) Cytoplasmic ribonucleoprotein complexes containing human LINE-1 protein and RNA. EMBO J., 15, 630–639. 12. Hohjoh, H. and Singer, M.F. (1997) Sequence-specific single-strand RNA binding protein encoded by the human LINE-1 retrotransposon. EMBO J., 16, 6034–6043. 13. Mathias, S.L., Scott, A.F., Kazazian, H.H., Boeke, J.D. and Gabriel, A. (1991) Reverse transcriptase encoded by a human transposable element. Science, 254, 1808–1810. 14. Feng, Q., Moran, J.V., Kazazian, H.H. and Boeke, J.D. (1996) Human L1 retrotransposon encodes a conserved endonuclease required for retrotransposition. Cell, 87, 905–916. 15. Moran, J.V., Holmes, S.E., Naas, T.P., DeBerardinis R.J., Boeke, J.D. and Kazazian, H.H. (1996) High frequency retrotransposition in cultured mammalian cells. Cell, 87, 917–927. 16. Martin, S.L. (1991) Ribonucleoprotein particles with LINE-1 RNA in mouse embryonal carcinoma cells. Mol. Cell. Biol., 11, 4804–4807. 17. Esnault, C., Maestre, J. and Heidmann, T. (2000) Human LINE retrotransposons generate processed pseudogenes. Nat. Genet., 24, 363–367. 18. Wei, W., Gilbert, N., Ooi, S.L., Lawler, J.F., Ostertag, E.M., Kazazian, H.H., Boeke, J.D. and Moran, J.V. (2001) Human L1 retrotransposition: cis preference versus trans complementation. Mol. Cell. Biol., 21, 1429–1439. 19. Luan, D.D., Korman, M.H., Jakubczak, J.L. and Eickbush, T.H. (1993) Reverse transcription of R2Bm is primed by a nick at the chromosomal target site: a mechanism for non-LTR retrotransposition. Cell, 72, 595–605. 20. Cost, G.J., Feng, Q., Jacquier, A. and Boeke, J.D. (2002) Human L1 element target-primed reverse transcription in vitro. EMBO J., 21, 5899–5910. 21. Morrish, T.A., Gilbert, N., Myers, J.S., Vincent, B.J., Stamato, T.D., Taccioli, T.E., Batzer, M.A. and Moran, J.V. (2002) DNA repair mediated by endonuclease-independent LINE-1 retrotransposition. Nat. Genet., 31, 159–165. 22. Gilbert, N., Lutz-Prigge, S. and Moran, J.V. (2002) Genomic deletions created upon LINE-1 retrotransposition. Cell, 110, 315–325. 23. Symer, D.E., Connelly, C., Szak, S.T., Caputo, E.M., Cost, G.J., Parmigiani, G. and Boeke, J.D. (2002). Human L1 retrotransposition is associated with genetic instability in vivo. Cell, 110, 327 –338. 24. Callinan, P.A., Wang, J., Herke, S.W., Garber, R.K., Liang, P. and Batzer, M.A. (2005) Alu retrotransposition-mediated deletion. J. Mol. Biol., 348, 791–800. 25. Gilbert, N., Lutz, S., Morrish, T.A. and Moran, J.V. (2005) Multiple fates of L1 retrotransposition intermediates in cultured human cells. Mol. Cell. Biol., 25, 7780–7795. 26. Martin, S.L., Li, J. and Weisz, J.A. (2000) Deletion analysis defines distinct functional domains for protein-protein and nucleic acid interactions in the ORF1 protein of mouse LINE-1. J. Mol. Biol., 304, 11–20. 27. Martin, S.L. and Bushman, F.D. (2001) Nucleic acid chaperone activity of the ORF 1 protein from the mouse LINE-1 retrotransposon. Mol. Cell. Biol., 21, 467–475. 28. Kolosha, V.O. and Martin, S.L. (2003) High-affinity, non-sequencespecific RNA binding by the open reading frame 1 (ORF1) protein from long interspersed nuclear element 1 (LINE-1). J. Biol. Chem., 278, 8112–8117. 29. Martin, S.L., Cruceanu, M., Branciforte, D., Li, P.W., Kwok, S.C., Hodges, R.S. and Williams, M.C. (2005) LINE-1 retrotransposition requires the nucleic acid chaperone activity of the ORF1 protein. J. Mol. Biol., 348, 549–561. 30. Moran, J.V., DeBerardinis, R.J. and Kazazian, H.H. (1999) Exon shuffling by L1 retrotransposition. Science, 283, 1530–1534. 31. Freeman, J.D., Goodchild, N.L. and Mager, D.L. (1994) A modified indicator gene for selection of retrotransposition events in mammalian cells. BioTechniques, 17, 47–52.


Supplementary Material is available at HMG Online.

3247

3248


40. Seleme, M.C., Busseau, I., Malinsky, S., Bucheton, A. and Teninges, D. (1999) High-frequency retrotransposition of a marked I factor in Drosophila melanogaster correlates with a dynamic expression pattern of the ORF1 protein in the cytoplasm of oocytes. Genetics, 151, 761–771. 41. Martin, S.L. and Branciforte, D. (1993) Synchronous expression of LINE-1 RNA and protein in mouse embryonal carcinoma cells. Mol. Cell. Biol., 13, 5383–5392. 42. Pont-Kingdon, G., Chi, E., Christensen. S. and Carroll, D. (1997) Ribonucleoprotein formation by the ORF1 protein of the non-LTR retrotransposon Tx1L in Xenopus oocytes. Nucleic Acids Res., 25, 3088–3094. 43. Seleme, M.C., Disson, O., Robin, S, Brun, C., Teninges, D. and Bucheton A. (2005) In vivo RNA localization of I factor, a non-LTR retrotransposon, requires a cis-acting signal in ORF2 and ORF1 protein. Nucleic Acids Res., 33: 776–85. 44. Dombroski, B.A., Scott, A.F. and Kazazian, H.H. (1993) Two additional potential retrotransposons isolated from a human L1 subfamily that contains an active retrotransposable element. Proc. Natl Acad. Sci., 90, 6513–6517. 45. Wei, W., Morrish T.A., Alisch, R.S. and Moran, J.V. (2000). A transient assay reveals that cultured human cells can accommodate multiple LINE-1 retrotransposition events. Anal. Biochem., 284, 435 –438. 46. Maniatis, T., Fritsch, E.F. and Sambrook, J. (1982) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.


32. Leibold, D.M., Swergold, G.D., Singer, M.F., Thayer, R.E., Dombroski, B.A. and Fanning, T.G. (1990) Translation of LINE-1 DNA elements in vitro and in human cells. Proc. Natl Acad. Sci., 87, 6990–6994. 33. Hohjoh, H. and Singer, M.F. (1997) Ribonuclease and high salt sensitivity of the ribonucleoprotein complex formed by the human LINE-1 retrotransposon. J. Mol. Biol., 271, 7–12. 34. Alisch, R.S. (2004) Unconventional translation of human LINE-1 ORF2. Ph.D. Dissertation, The University of Michigan, Ann Arbor. 35. Kolosha, V.O. and Martin, S.L. (1997) In vitro properties of the first ORF protein from mouse LINE-1 support its role in ribonucleoprotein particle formation during retrotransposition. Proc. Natl Acad. Sci., 94, 10155–10169. 36. Kolosha, V.O. and Martin, S.L. (1995) Polymorphic sequences encoding the first open reading frame protein from LINE-1 ribonucleoprotein particles. J. Biol. Chem., 270, 2868–2873. 37. Perepelitsa-Belancio, V. and Deininger, P. (2003). RNA truncation by premature polyadenylation attenuates human mobile element activity. Nat. Genet., 35, 363– 366. 38. Han, J., Szak, S. and Boeke, J.D. (2004). Transcriptional disruption by the L1 retrotransposon and implications for mammalian transcripts. Nature 429, 268–274. 39. Han, J. and Boeke, J.D. (2004) A highly active synthetic mouse retrotransposon. Nature, 429, 314 –318.