Identification of an integrase-independent pathway of retrotransposition

Abstract

Retroviruses and long terminal repeat retrotransposons rely on integrase (IN) to insert their complementary DNA (cDNA) into the genome of host cells. Nevertheless, in the absence of IN, retroelements can retain notable levels of insertion activity. We have characterized the IN-independent pathway of Tf1 and found that insertion sites had homology to the primers of reverse transcription, which form single-stranded DNAs at the termini of the cDNA. In the absence of IN activity, a similar bias was observed with HIV-1. Our studies showed that the Tf1 insertions result from single-strand annealing, a noncanonical form of homologous recombination mediated by Rad52. By expanding our analysis of insertions to include repeat sequences, we found most formed tandem elements by inserting at preexisting copies of a related transposable element. Unexpectedly, we found that wild-type Tf1 uses the IN-independent pathway as an alternative mode of insertion.

Fig. 1.. Tf1 insertion takes place in the absence of integrase.

(A) The diagram shows the strategy of monitoring Tf1 retrotransposition. A drug-resistant gene, nat, with artificial intron (nat-AI) is introduced into Tf1, and the integration of Tf1 into host chromosomes allows cells to grow on plates containing Nat. The black arrows indicate the frame shift (fs) sites of PR and IN, respectively. LTR, long terminal repeat; PR, protease; RT, reverse transcriptase; IN, integrase (B) Growth phenotypes of Tf1-WT, Tf1-INfs, and Tf1-PRfs (Tf1 with a frameshift in PR) on medium containing Nat after inducing Tf1 expression. (C) Quantitative transposition analysis Tf1-WT, Tf1-INfs, and Tf1-PRfs. The y axis indicates the transposition frequency measured by the percentage of Nat-resistant cells for each strain. Error bars represent SDs.

Fig. 2.. The distribution pattern of Tf1-INfs insertions is markedly different from the insertions of Tf1-WT.

(A) Density scatterplot and linear regression analysis are shown for Tf1-WT and Tf1-INfs within intergenic and ORF sequences. The axes are the proportion of insertions in cells per intergenic or ORF region normalized as a percentage of all insertions. The correlation coefficient (R) represents the linear relationship between two datasets. R values between 0.7 and 1.0 indicate a strong correlation. R values between 0 and 0.3 indicate a weak correlation. (B) The histograms show the distance from insertion sites to the nearest ORF. The x coordinate is the distance from the 5′ and 3′ ends of ORFs. The y coordinate shows the percentage of integration events within bins of 100 bp.

Fig. 3.. Insertion sites of Tf1-INfs have bias of ATAAC.

(A) Sequence logos are shown for Tf1-WT and Tf1-INfs insertion sites. The height of each individual base at a given position reflects the level of conservation at that position. Sequence conservation is indicated as the total height of each stack measured by bits score. The relative height of bases in a stack represents base frequencies at that position. The x axis demonstrates the nucleotide positions relative to the Tf1 target site. TSD, target site duplication; the asterisks mark the nucleotides in common with the PBS. (B) Downstream flanks of the 10 most common sites of Tf1-INfs insertions. The nucleotides that match the PBS sequence are highlighted in red.

Fig. 4.. Sequences downstream and upstream of Tf1-INfs insertions show similarity to PBS and PPT, respectively.

(A) Ligation-mediated PCR with specific LTR tags provided sequence upstream and downstream of insertions. Tf1_DT, double-tagged Tf1. (B) The histograms depict the positions of insertions relative to ORFs. (C) The DNA logos are shown for sequences downstream and upstream of insertions. The lower diagram shows the sequences of the PBS and PPT at the termini of Tf1 cDNA. The asterisks mark the nucleotides in common with the PPT and PBS.

Fig. 5.. Most of the IN-independent events occur at pre-existing Tf2s.

(A) Experimental design for whole-genome sequencing of individual Tf1-INfs insertions via the PacBio approach; (B) structure diagram of Tf1-INfs insertions generated from four independent clones; (C) transposition assay performed in strains with and without 14 preexisting Tf2 elements. Tf1 plasmids used here had single-tagged Tf1with natAI. (D) Quantitative transposition analysis of Tf1-WT and Tf1-INfs in the wild type and Tf2-null strain was based on four independent replicates. Statistical differences between WT and Tf2-null were determined by two-tailed paired t test. *P < 0.05 was considered statistically significant. ***P < 0.001. Error bars represent SDs.

Fig. 6.. Homologous recombination factor Rad52 is required for IN-independent insertion.

(A) The Mre11-Rad50-Nbs1 complex is recruited to double-strand breaks (DSB) to trigger single-strand resection. RPA is responsible for the stabilization of single-stranded DNA. The Rad51/Rad52 complex binds to single-stranded DNA and facilitates strand invasion. (B) The transposition assay was conducted with Tf1-natAI in the mre11, rad50, nbs1, rad51, and rad52 deletion strains. (C) Quantitative transposition analysis of WT and deletion strains with both Tf1-WT and Tf1-INfs based on four replicates. Statistical differences between WT and deletion strains were determined by two-tailed paired t test. *P < 0.05 was considered statistically significant. **P < 0.01 and ***P < 0.001. ns, not significant.

Fig. 7.. The rad52-R45A mutation reduces the frequency of IN-independent insertion.

(A) The rad52-R45 residue is located at the DNA binding domain of Rad52, spRad52, and S. pombe Rad52. (B) Saturated cultures of WT, rad52 knockout, and rad52-R45A strains were 10-fold diluted in water and spotted onto YES plates supplemented with 5 mM HU. HU, hydroxy urea. (C) Transposition efficiency was monitored with Tf1-natAI in WT and rad52 deletion strains and in two independently generated replicates of rad52-R45A cells. (D) Immunoblot analysis of Rad52-WT and Rad52-R45A protein level in cells transformed with Tf1-WT, Tf1-INfs, and Tf1-PRfs; (E) chromatin immunoprecipitation (ChIP)–quantitative PCR to measure cDNA associated with Rad52 and Rad52-R45A. Primers are designed to amplify nat from Tf1 cDNA in which the artificial intron in the nat is spliced out. adh1 served as a negative control for Rad52 binding. *P < 0.05 was considered statistically significant. Error bars represent SDs from three independent experiments.

Fig. 8.. WT-Tf1 uses an IN-independent pathway during Tf1 expression.

(A) Tf1 expression is under the control of the nmt1 promoter, which is induced in the absence of vitamin B1. After the accumulation of insertions, expression is turned off by adding B1 to a subsequent culture (post-expression). (B) The insertions from the expression culture are plotted on the basis of their distance to the closest ORF. (C) The heatmap matrix illustrates the correlation efficiency (R) of insertions located at intergenic and ORF regions between any two datasets. Exp WT/INfs stands for Tf1-WT and Tf1-INfs datasets obtained from the expression culture; Post_WT/INfs represents Tf1-WT and T1-INfs datasets from the post-expression culture.

Fig. 9.. PBS sequence contributes to IN-independent insertion of HIV-1.

(A) Homology to PBS sequence can occur downstream of HIV-1 provirus inserted in the absence of IN activity. The table is shown for the percentage of PBS events in all of the HIV-1 IN-independent insertions identified by two groups. Reference a, (7); reference b, (9). (B) The downstream flanks of four HIV-1 insertions from virus containing the IN D64V mutant are shown in the table. The sequences are derived from both refs. a and b. The nucleotides exhibiting PBS homology are highlighted in red; the nucleotides originated from the virus that differ from the reference genome are underlined.

Fig. 10.. The models of Rad52-driven SSA.

(A) The IN-independent insertions that occurred at unique genomic sites are promoted by SSA. First, Rad52 is recruited to Tf1 cDNA that has single-stranded PBS DNA at the downstream 3′ terminus. The Tf1 cDNA searches for a PBS homology across the genome. (B) Formation of tandem elements occurs when a break forms at a replication fork near a Tf2 LTR. Tf1 cDNA anneals to a newly synthesized strand of Tf2. Repair of the paired structure results in a duplication of Tf elements separated by a single LTR.