The retrohoming of linear group II intron RNAs in Drosophila melanogaster occurs by both DNA ligase 4-dependent and -independent mechanisms

Abstract

Mobile group II introns are bacterial retrotransposons that are thought to have invaded early eukaryotes and evolved into introns and retroelements in higher organisms. In bacteria, group II introns typically retrohome via full reverse splicing of an excised intron lariat RNA into a DNA site, where it is reverse transcribed by the intron-encoded protein. Recently, we showed that linear group II intron RNAs, which can result from hydrolytic splicing or debranching of lariat RNAs, can retrohome in eukaryotes by performing only the first step of reverse splicing, ligating their 3' end to the downstream DNA exon. Reverse transcription then yields an intron cDNA, whose free end is linked to the upstream DNA exon by an error-prone process that yields junctions similar to those formed by non-homologous end joining (NHEJ). Here, by using Drosophila melanogaster NHEJ mutants, we show that linear intron RNA retrohoming occurs by major Lig4-dependent and minor Lig4-independent mechanisms, which appear to be related to classical and alternate NHEJ, respectively. The DNA repair polymerase θ plays a crucial role in both pathways. Surprisingly, however, mutations in Ku70, which functions in capping chromosome ends during NHEJ, have only moderate, possibly indirect effects, suggesting that both Lig4 and the alternate end-joining ligase act in some retrohoming events independently of Ku. Another potential Lig4-independent mechanism, reverse transcriptase template switching from the intron RNA to the upstream exon DNA, occurs in vitro, but gives junctions differing from the majority in vivo. Our results show that group II introns can utilize cellular NHEJ enzymes for retromobility in higher organisms, possibly exploiting mechanisms that contribute to retrotransposition and mitigate DNA damage by resident retrotransposons. Additionally, our results reveal novel activities of group II intron reverse transcriptases, with implications for retrohoming mechanisms and potential biotechnological applications.

Figure 1. Models for retrohoming of Ll.LtrB group II intron lariat and linear RNAs.

(A) Retrohoming of lariat RNA. RNPs containing lariat RNA recognize the DNA target site (ligated E1–E2 sequence) and carry out both steps of reverse splicing, resulting in insertion of the intron RNA between E1 and E2. The IEP uses its En domain to cleave the bottom strand between positions +9 and +10 of E2, and then uses the 3′ end of the cleaved DNA strand as a primer for reverse transcription of the inserted intron RNA. The resulting full-length intron cDNA is extended directly into E1 by continued DNA synthesis. Retrohoming is completed by a process that includes removal of the 5′ overhang on the bottom strand, degradation or displacement of the intron RNA template strand, top-strand DNA synthesis by a host DNA polymerase, and sealing of nicks by a host DNA ligase . (B) Retrohoming of linear RNA. RNPs containing linear intron RNA recognize the DNA target site and carry out the first step of reverse splicing, resulting in ligation of the 3′ end of the intron RNA to the 5′ end of E2. The IEP then uses its En domain to cleave the bottom strand between positions +9 and +10, generating a primer for reverse transcription of the intron RNA, as in lariat RNA retrohoming. However, because the 5′ end of the linear intron RNA is unattached, the resulting cDNA cannot be extended directly into E1 and is instead linked to the 5′ exon DNA by an error-prone process that sometimes leads to precise insertion of the intron RNA, but often gives imprecise 5′ junctions due to deletion of E1 sequences, 5′-intron truncations, and/or insertion of extra nucleotide residues at the ligation junction. As for lariat RNA, retrohoming of the linear intron RNA is completed by degradation or displacement of the intron RNA template strand, top-strand DNA synthesis, and sealing of nicks by host enzymes. E1 and E2, 5′ and 3′ exon, respectively; CS, bottom-strand cleavage site; IS, intron-insertion site.

Figure 2. Retrohoming efficiencies of linear and lariat group II intron RNAs in wild-type and mutant Drosophila.

(A) Microinjection assay for retrohoming of lariat and linear group II intron RNAs. D. melanogaster precellular blastoderm are microinjected with the Amp^R-recipient plasmid pBRR3-ltrB, which contains an Ll.LtrB target site (ligated E1–E2 sequence) cloned upstream of a promoterless tet ^R gene, followed by separate microinjection of Ll.LtrB RNPs containing linear or lariat intron RNAs with a phage T7 promoter sequence inserted near their 3′ end. The embryos are incubated at 30°C for 1 h, during which the intron integrates into the target site in the recipient plasmid, placing the T7 promoter upstream of the promoterless tet ^R gene. Nucleic acids are then extracted and transformed into E. coli HMS174(DE3) for plating assays, and retrohoming efficiencies are calculated as the ratio of (Tet^R+Amp^R)/Amp^R colonies. T1 and T2, E. coli rrnB transcription terminators; Tφ, phage T7 transcription terminator. (B) Retrohoming efficiencies of lariat and linear Ll.LtrB RNPs in D. melanogaster wild-type (w¹¹¹⁸ and Or-R) and mutant embryos were determined, as described in panel A and Materials and Methods. The bar graphs show retrohoming efficiency in the indicated mutant embryos relative to that of wild-type embryos assayed in parallel in ten independent experiments with different combinations of strains (Table S1). The values are the mean for at least three independent determinations for each mutant, with the error bars indicating the standard error. The retrohoming efficiency of ≤0.5% wild type for linear RNPs in the polQ mutant is an upper limit, as only a single Tet^R+Amp^R colony was recovered in three separate experiments, although 5′-intron integration junctions could be detected by using a more sensitive PCR assay in all experiments (see Figure 3).

Figure 3. PCR analysis of integration junctions from lariat and linear intron RNA retrohoming in wild-type and mutant strains.

Retrohoming assays with lariat and linear RNPs were done as described in Figure 2A and Materials and Methods, and DNA was extracted from 80 pooled embryos for each strain. 5′- and 3′-integration junctions were amplified by PCR, using primers that flank the junction (5′junction, forward primer P1 and reverse primer LtrB933a; 3′ junction, forward primer P3 and reverse primer P4; see Materials and Methods). The PCR products were analyzed in a 1% agarose gel, which was stained with ethidium bromide. Precise 5′ and 3′ junctions for lariat intron RNA retrohoming and precise 3′ junctions for linear intron RNA retrohoming were confirmed by sequencing junctions from at least 10 randomly selected Tet^R+Amp^R colonies or PCR products from pooled embryos for all strains (not shown).

Figure 4. Sequences of 5′-integration junctions from linear intron RNA retrohoming in wild-type and mutant strains.

5′-integration junctions of DNA extracted from 80 pooled embryos for each strain were amplified by PCR, as described in Figure 3, then TOPA-TA cloned, amplified by colony PCR, and sequenced, as described in Materials and Methods. (A) wild-type w¹¹¹⁸; (B) lig4 ⁻; (C) ku70 ⁻; (D) lig4⁻; P{lig4⁺}; (E) wild-type Or-R; (F) polQ ⁻. Inserted or mutant nucleotide residues are shown in lower case letters; microhomologies between intron and exon end sequences prior to ligation are shown in parentheses; and inserted sequences that match or are complementary to nearby 5′-exon or intron sequences are underlined. Freq., frequency of occurrence.

Figure 5. Characteristics of 5′-integration junctions resulting from linear intron RNA retrohoming in wild-type and mutant strains.

The bar graphs show the percentage of 5′-integration junctions with (A) exon 1 deletions, (B) 5′-intron truncations, (C) extra nucleotide additions, and (D) microhomologies in the indicated strains. For the polQ ⁻ embryos, where the number of unique junction sequences recovered was smaller than for the other strains, the percentage of junctions having the indicated characteristics was calculated both as a percentage of total junctions (left bar) and a percentage of unique junctions (right bar, asterisk).

Figure 6. Template switching of LtrA from the 5′ end of the Ll.LtrB intron RNA to exon 1 DNA or RNA.

The Ll.LtrB intron RT (LtrA protein; 40 nM) was incubated with artificial substrates corresponding to the 5′ end of Ll.LtrB intron (Ll.LtrB RNA; 40 nM) with an annealed 5′-³²P-labeled DNA primer c (Pri c; 44 nM) in presence of exon 1 (E1) DNA or RNA (40 nM; black and red, respectively), as diagrammed in schematics to the left of the gel. The substrates were incubated with dNTPs (200 µM) in reaction medium containing 450 mM NaCl, 5 mM MgCl₂, 20 mM Tris-HCl, pH 7.5, and 1 mM DTT for 30 min at 30°C. After terminating the reaction by extraction with phenol-CIA, the products were analyzed in a denaturing 15% polyacrylamide gel. Lanes (1) and (2) ³²P-labeled Pri c incubated without and with LtrA, respectively; (3) and (4) LtrA incubated with ³²P-labeled Pri c and E1 DNA or RNA, respectively; (5) and (6) LtrA incubated with Ll.LtrB RNA with annealed ³²P-labeled Pri c and E1 DNA or RNA, respectively; (7–9) LtrA incubated with Ll.LtrB RNA with annealed ³²P-labeled Pri c and E1 DNA or RNA with annealed complementary DNA oligonucleotides to leave a blunt end (exon 1 AS) or a 5′-bottom-strand overhang (exon 1 AS+9). Bands excised for sequencing (Figure 7) are indicated in the gel. In the schematics, DNA and RNA oligonucleotides are shown in black and red, respectively; LtrA is shown as a gray oval; and the direction of cDNA synthesis is indicated by a green arrow. The numbers to the right of the gel indicate the positions of 5′-end labeled size markers (10-bp DNA ladder, Invitrogen).

Figure 7. DNA sequences resulting from template switching of LtrA from Ll.LtrB RNA to exon 1 DNA or RNA.

(A) and (B) Sequences of DNA products resulting from template switching of the Ll.LtrB RT (LtrA protein) from the 5′ end of the Ll.LtrB intron RNA template/Pri c DNA substrate to exon 1 DNA (lane 5) or RNA (lane 6), respectively. (C) Sequences of DNA products resulting from template switching of the Ll.LtrB RT from the 5′ end of Ll.LtrB RNA/primer c substrate to double-stranded exon 1 DNA with a 5′ bottom-strand overhang (AS+9; lane 9). Bands were excised from the gel, cloned, and sequenced, as described in Materials and Methods. The substrate and expected cDNA or DNA product sequences are shown boxed above each set of experimentally determined DNA product sequences. Extra or mutant nucleotide residues are shown in lower case letters. Freq., frequency of occurrence; *, ³²P-label at 5′ end of primer c.

Figure 8. Model for ligation of the intron cDNA to exon 1 during linear group II intron retrohoming.

Partial reverse splicing of the linear intron RNA into the DNA target site is followed by bottom-strand cleavage between E2 positions +9 and +10 and synthesis of a cDNA of the attached linear intron RNA. After cDNA synthesis reaches the 5′ end of the intron RNA template, extra nucleotide residues are added to the 3′ end of the cDNA generating microhomologies that enable annealing of the cDNA strand to the top strand of E1. In Drosophila, the repair DNA polymerase θ is the major enzyme responsible for this extra nucleotide addition, but some extra nucleotide addition may also be done by other host DNA polymerases or by the Ll.LtrB RT. The annealing of the cDNA requires unwinding and/or resection of the bottom strand, leading to loss of the bottom-strand 5′ overhang resulting from the initial double-strand break by the group II intron RNP. In the final step, the annealed cDNA is ligated to the bottom strand of E1 by DNA ligase 4 or an alternate ligase. The retrohoming of the linear intron RNA is completed by degradation or displacement of the intron RNA template strand, second-strand DNA synthesis, and sealing of nicks by host enzymes. E1 and E2, 5′ and 3′ exons, respectively.