Intron Homing With Limited Exon Homology: Illegitimate Double-Strand-Break Repair in Intron Acquisition by Phage T4

Corresponding author: Marlene Belfort, Wadsworth Center, New York State Department of Health, P.O. Box 22002, Albany, NY 12201-2002., marlene.belfort{at}wadsworth.org (E-mail)

Communicating editor: G. R. SMITH

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The td intron of bacteriophage T4 encodes a DNA endonuclease that initiates intron homing to cognate intronless alleles by a double-strand-break (DSB) repair process. A genetic assay was developed to analyze the relationship between exon homology and homing efficiency. Because models predict exonucleolytic processing of the cleaved recipient leading to homologous strand invasion of the donor allele, the assay was performed in wild-type and exonuclease-deficient (rnh or dexA) phage. Efficient homing was supported by exon lengths of 50 bp or greater, whereas more limited exon lengths led to a precipitous decline in homing levels. However, extensive homology in one exon still supported elevated homing levels when the other exon was completely absent. Analysis of these "one-sided" events revealed recombination junctions at ectopic sites of microhomology and implicated nucleolytic degradation in illegitimate DSB repair in T4. Interestingly, homing efficiency with extremely limiting exon homology was greatly elevated in phage deficient in the 3'-5' exonuclease, DexA, suggesting that the length of 3' tails is a major determinant of the efficiency of DSB repair. Together, these results suggest that illegitimate DSB repair may provide a means by which introns can invade ectopic sites.

SEVERAL group I introns undergo a process termed homing, whereby the intron is efficiently inserted into intronless cognates of the intron-containing allele (DUJON 1989 ). Homing occurs by a gene conversion process that depends on expression of a site-specific endonuclease encoded within the intron (reviewed in LAMBOWITZ and BELFORT 1993 ; MUELLER et al. 1993 ; BELFORT and PERLMAN 1995 ). The intron endonuclease makes a double-strand break (DSB) in the intronless allele at or near the site of intron insertion. Intron sequences are copied from the intron-containing allele during repair of the DSB, resulting in inheritance of the intron.

Homing of the intron in the td (thymidylate synthase) gene of bacteriophage T4 is initiated by I-TevI, the td intron endonuclease, which makes a DSB in an intronless td allele 23 and 25 nucleotides (nt) upstream of the intron insertion site (BELL-PEDERSEN et al. 1990 ; Figure 1A). Repair of the DSB results in precise acquisition of the intron at the exon I-exon II junction. At least two pathways of DSB repair have been implicated in td intron homing (MUELLER et al. 1996A ), classic double-strand-break repair (DSBR; Figure 1) and synthesis-dependent strand annealing (SDSA; Figure 1). The initial steps of DSB formation, exonucleolytic processing of the ends and homologous strand invasion, are common to both models, but subsequent repair synthesis is mechanistically different. Each pathway implicated in T4 intron homing predicts a requirement for homology between donor and recipient molecules on both sides of the break. The efficiency of homing depends on exon length, suggesting that homing may require a minimal length of homology (QUIRK et al. 1989A ), as is the case in a phage -td intron model system (PARKER et al. 1996 ).

View larger version (24K): In this window In a new window Download PPT slide   Figure 1. Proposed pathways of td intron homing. The td homing site (A) comprises the intron insertion site (IS) at the junction of exon I and exon II, and the I-TevI cleavage site (CS) located 23 and 25 nt upstream. The recipient allele receives a DSB (B), which is processed by exonucleases (C). A 3' single-stranded end invades a donor allele and synapsis with homologous sequences occurs (D). In DSBR, repair synthesis leads to formation of a D-loop (E) and, after ligation, double Holliday junctions can isomerize independently (F). Cleavage by resolvase results in noncrossover or crossover products. Only the crossover products are shown (G). In SDSA, a replication bubble proceeds along the donor template DNA (E'), and the newly synthesized strand anneals with complementary sequences and serves as template for repair synthesis (F'). Release of the invading strand and ligation leads to gene conversion without crossing over of flanking markers (G'). Thick lines, homologous DNA in recipient and donor alleles; thin lines, flanking DNA of recipient; wavy lines, flanking DNA of donor; shaded lines, intron DNA; half arrow, 3' end of DNA; arrowheads, resolvase cleavage. Alternatives to these pathways have also been proposed (LUDER and MOSIG 1982 ; GEORGE and KREUZER 1996 ).

Homing of the td intron occurs via the recombination-dependent mode of DNA replication of T4 (MUELLER et al. 1996A ). Although the requisite UvsX protein, an analog of the Escherichia coli strand transferase, RecA, has been well characterized (KREUZER and MORRICAL 1994 ), several exonuclease activities that appear to be involved in replication and recombination in T4 have poorly defined roles. Models of intron homing predict a requirement for a 5'-3' nucleolytic activity to generate invasive 3' ssDNA tails, as well as a 3'-5' exonuclease activity to degrade the "resection segment" produced by the eccentric cleavage of I-TevI (Figure 1A; CLYMAN and BELFORT 1992 ). The gp46/47 complex, a required function for T4 recombination and homing, has been implicated as a 5'-3' exonuclease (MICKELSON and WIBERG 1981 ; KREUZER et al. 1995 ; MUELLER et al. 1996A ). Although a specific role for this complex has remained elusive, it likely makes a significant contribution, albeit possibly indirect, to DNA end processing during recombination (MOSIG 1998 ). T4 RNase H, which possesses 5'-3' DNA exonuclease activity (HOLLINGSWORTH and NOSSAL 1991 ), may also play a role in intron homing (MUELLER et al. 1996A ; HUANG et al. 1999 ). Although DexA, a 3'-5' exonuclease (GRUBER et al. 1988 ), and 43Exo, the 3'-5' exonuclease activity of T4 DNA polymerase, are not required for homing, in vivo and in vitro analyses suggest a role for these functions in 3' end processing (MUELLER et al. 1996A ; HUANG et al. 1999 ).

A sensitive genetic assay was developed to characterize events in which homology was limited, thereby addressing the mechanism of illegitimate DSB repair in T4. Additionally, the results are relevant to the evolution of mobile introns, for they suggest a means for intron transfer to ectopic sites.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Bacterial and phage strains:
E. coli strains CR63 (supD), the nonsense-suppressing suppressor-plus (Sup⁺) host, and MH1 (araD139 lacX74 galU^- galK^- hsr^- hsm⁺ rpsL sup⁺), the nonsuppressing (Sup°) host, were provided by K. Kreuzer (Duke University, Durham, NC). CR63cI857kan^R carrying a temperature-sensitive phage repressor was made by transduction of strain CR63 with a P1 lysate containing galK::cI857kan^R. E. coli strain OK305 (E. coli B cd^- sup⁺), a Sup° host lacking cytidine and deoxycytidine deaminase activities (HALL et al. 1967 ), was provided by D. Hall (Georgia Institute of Technology, Atlanta).

T4 38 (amB262) and T4 51 (amS29) were used to introduce 38^am and 51^am mutations into our laboratory stock of wild-type strain T4D, and T4D 38^am51^am was confirmed by cross-streak complementation tests. The td intron was deleted from this strain through a two-step procedure (BELL-PEDERSEN et al. 1989 ) by first crossing T4 38^am51^am with a plasmid carrying an intronless td variant, tdOP8, which is resistant to I-TevI cleavage. T4 38^am51^amtdOP8 was then crossed with a plasmid carrying the intronless allele, tdIn, to generate T4 38^am51^amtdIn. T4 38^am51^amtdIndexA and T4 38^am51^amtdOP8dexA containing a short insertion in the dexA allele were constructed by marker rescue from pUC18dexA^- provided by P. Gauss (Western State College, Gunnison, CO) and confirmed by growth inhibition on an E. coli optA host (GAUSS et al. 1987 ). T4 38^am51^amtdOP8 was crossed with T4K10rnh10-777 (WOODWORTH and KREUZER 1996 ), provided by K. Kreuzer, and plated on CR63() to select against T4K10 (38^am51^am denA denB) phage. T4 38^am51^amtd-OP8rnh10-777 was confirmed by the polymerase chain reaction (PCR) and crossed with pBStdIn to generate T4 38^am-51^amtdInrnh10-777.

Plasmids:
Plasmid pMP is a derivative of pKK061, which has a T4 origin of replication (ori34) in pBR322 (SELICK et al. 1988 ). To generate pMP, the EcoRI site of pKK061 was filled in using the Klenow fragment of E. coli DNA polymerase I, and the EcoRV-to-BamHI segment of the pBSKS+ polylinker (Stratagene, La Jolla, CA) was inserted at these sites in pKK061 to introduce a new EcoRI cloning site. The tdSG2::supF allele and its exon deletion variants (Table 1) were amplified from phage T4K10 tdSG2::supF by PCR and cloned into the EcoRI and BamHI sites of pMP to generate pMPtdSG2::supF. The tdSG2::supF allele contains a td intron in which the I-TevI open reading frame has been replaced with the XbaI-to-BanII fragment of pBSPLO+ containing a supF gene under control of a T4 gene 23 promoter (SELICK et al. 1988 ). A point mutation, SC79, was introduced into the intron core to inhibit splicing and render the td gene thymidylate synthase deficient (Td^-; CHANDRY and BELFORT 1987 ). The intron is flanked by the entire td coding sequence and adjacent T4 sequence extending to 766 bp upstream and 1016 bp downstream of the intron. The I-TevI expression plasmid pACYC-P_L-ORF has the HindIII-to-BamHI fragment of pKC30 (ROSENBERG et al. 1983 ), providing phage regulatory elements, in pACYC184 (CHANG and COHEN 1978 ). The I-TevI open reading frame (ORF) and flanking intron sequences are in the SmaI site of the pKC30 fragment.

Intron homing assay:
CR63cI857kan^R cells carrying a wild-type intron donor plasmid, pMPtdSG2::supF, or an exon deletion variant (Table 1), and pACYC-P_L-ORF were grown in tryptone-yeast extract broth with 100 µg/ml ampicillin and 25 µg/ml chloramphenicol at 30° to an OD₆₅₀ of 0.2. Cultures were shifted to 42° for 30 min, then infected with T4 phage at a multiplicity of 0.1 at 37° with shaking for 2 hr. Cultures were lysed with chloroform and phage lysates were titered on MH1 (Sup°) and CR63 (Sup⁺) to determine the concentration of the suppressor-containing Su⁺ and total phage, respectively.

Analysis of homing products:
To distinguish homing events (Td^-) from plasmid integrants (Td⁺), lysates were plated on an OK305 (Sup°) host to select for nonsense-suppressed (Su⁺) phage and to screen for thymidylate synthase-negative (Td^-) phage using the halo phenotype (HALL et al. 1967 ). Su⁺Td^- phage were plaque purified and homing events were confirmed by the absence of an intact copy of the tdIn allele using PCR. Recombination junction sequences were PCR amplified using intron variant SG2::supF-specific oligonucleotides W165, 5'-GCAGCTGGATATAATTCCGGGGTA-3', and W248, 5'-TGGTGGTGGGGGAAGGATTCGA-3', and T4-specific oligonucleotides W731, 5'-CCTGAACTTAGTATCACAAGCG-3', and W657, 5'-ATTCCATATCCCGTTCGTGC-3'. Oligonucleotides W731, W248, W165, and W657 are represented as P1, P2, P3, and P4, respectively, in Figure 4. PCR products were purified by centrifugation through a Centricon 100 column (Amicon, Beverly, MA) and sequenced.

View larger version (53K): In this window In a new window Download PPT slide   Figure 2. Genetic selection scheme to monitor homing. The intron recipient, T4 38am51amtdIn, was used to infect CR63cI857 carrying an I-TevI expression plasmid and an intron donor plasmid, pMPtdSG2::supF. The intron donor has a supF nonsense-suppressing gene in the intron, a mutation in the intron core () rendering an intron-containing td gene Td-, and varying lengths of exon I and/or exon II () (Table 1). Induction of I-TevI was followed by T4 infection. Lysates were assayed on a Sup° host to select for Su+ phage that had inherited the intron and a Sup+ host to quantitate total phage. Open bars, td exons; shaded bar, intron variant tdSG2::supF; Apr, ampicillin resistance; Cmr, chloramphenicol resistance. Shaded background highlights homing products. Td phenotypes: , Td+; , Td-.

View larger version (17K): In this window In a new window Download PPT slide   Figure 3. Exon-homology dependence of intron homing. (A) Homing for donors with one exon truncated. The mean ratios of Su+/total phage from three independent experiments were plotted against the length of the exon for donors with deletions in either exon I (, ) or exon II (•, ). Filled symbols, infections with I-TevI-sensitive strain T4 38am 51am tdIn; open symbols, infections with I-TevI-resistant strain T4 38am 51am tdOP8. Error bars reflect the standard deviation from the mean. The total phage burst was equivalent for all infections with the I-TevI-resistant and the I-TevI-sensitive phage. (B) Homing for donors with both exons truncated. Constructs are listed in Table 1. Plot and conventions as in A.

View larger version (39K): In this window In a new window Download PPT slide   Figure 4. Homing products from one-sided heterology events. (A) The proposed formation of homing products from an intron donor lacking exon I (left) or exon II (right). Cleavage and processing of the recipient leads to extensive degradation of the exon that is absent in the donor. The donor is invaded on one side of the intron by homologous exon sequences of the recipient, and on the other side by heterologous sequences of the recipient, leading to intron inheritence. Thin lines, td exons; wavy lines, flanking phage DNA; thick lines, vector sequences; shading, intron. marks the recombination junction and the point to which degradation occurs in the recipient. Small arrows, intron-specific (P2, P3) and phage-specific (P1, P4) oligonucleotide primers used for PCR. (B) Agarose gel analysis of PCR products. PCR amplification using primers P1 and P2 (lanes 1–6) and P3 and P4 (lanes 7–12) was performed on phage from homing assays with the wild-type donor (lanes 1 and 7), exon I-minus donor pMPEI-10 (lanes 2–5), and exon II-minus donor pMPEII-5 (lanes 8–11). Lanes 6 and 12 are control PCR reactions without template DNA. M is a 1-kb ladder with DNA lengths in kilobases indicated on the left. The product names shown below each lane correspond to those shown in C and D. (C) Deletions upstream of cleavage site. The size of the deletion accompanying intron inheritance from a donor allele lacking exon I was determined by sequencing across recombination junctions of 14 independent homing products. Independent events that had identical junctions are indicated by brackets. (D) Deletions downstream of cleavage site. The size of the deletion was determined for 21 independent homing products resulting from a donor allele lacking exon II.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Genetic assay for intron homing in T4:
An in vivo genetic assay was developed to quantitatively analyze exon homology effects on intron homing during T4 infection (Figure 2). I-TevI was supplied from a plasmid (pACYC-P_L-ORF) under control of the P_L promoter and a cI857 repressor gene carried on the host chromosome. The intron donor allele was on a compatible plasmid, pMPtdSG2::supF. In this allele the td intron ORF has been replaced with a supF gene, encoding an amber-suppressing tRNA^tyr under control of the T4 gene 23 late promoter (SELICK et al. 1988 ). The donor allele is Td^- due to a mutation in the intron's catalytic core, rendering the intron splicing defective. The intron recipient was phage T4 38^am51^amtdIn, which contains a Td⁺ intronless td allele (tdIn). Additionally, the recipient phage harbored amber mutations in genes 38 and 51, encoding essential packaging functions. These mutations prevent plaque formation on a nonsuppressing (Sup°) host (SELICK et al. 1988 ), resulting in a nonsuppressed Su^-Td⁺ phenotype (Figure 2). Plasmid integrants would contain the suppressing supF allele and two copies of the td gene (the donor allele and an uninterrupted recipient allele) and would display a Su⁺Td⁺ phenotype. In contrast, true homing events are distinguishable by their Su⁺Td^- phenotype, because inheritance of the splicing-deficient intron converts the recipient tdIn gene to Td^- (Figure 2). Td^- plaques can be readily distinguished from their Td⁺ counterparts by white halos on OK305 cells (HALL et al. 1967 ). In our assay, the ratio of Su⁺/total phage was used to quantitate levels of intron inheritance. The Td phenotype was used only when homing approached background recombination and to isolate homing products for physical analysis.

In our assay tdSG2::supF, containing full-length td exons plus flanking DNA totaling >700 bp of homology to the recipient phage in each exon, served as "wild type" (Table 1). Intron homing with the wild-type donor was efficient, with an average ratio of Su⁺/total phage equaling 6.8 x 10^-1 (Figure 3A and Figure B). Background recombination, which was determined with an isogenic recipient phage, T4 38^am51^amtdOP8, containing an I-TevI-resistant, intronless td allele (tdOP8), resulted in a mean ratio of 7.2 x 10^-3 Su⁺/total phage (Figure 3A and Figure B). Thus, intron homing levels were two orders of magnitude above background, I-TevI-independent recombination levels when homology in both exons was nonlimiting.

Analysis of homology effects on intron homing in T4:
To determine the relationship between homing and length of homology, intron donor plasmids were constructed with truncated exons (Table 1). The length of homology to recipient exons ranged from 0 to 1016 bp on one or both sides of the intron. Due to the eccentric position of the I-TevI cleavage site in the td homing site (Figure 1A), homology with exon I of the cleaved recipient occurs upstream of the DSB (Figure 1; Table 1). When homology was limiting in one exon, intron homing levels remained elevated above background recombination levels at all exon lengths studied (Figure 3A). When the limiting exon was at least 50 bp in length, homing levels were >100-fold above background. However, when either exon was reduced below 50 bp, homing levels underwent a sharp decrease, but remained elevated at least 10-fold above background levels, even when one exon was completely deleted from the donor.

When both exons were limiting, an exon length of 50 bp was again found to be critical for intron homing (Figure 3B). Although at 50 bp of homology on both sides of the break the level of homing was at least 10-fold lower than when a single exon was reduced to this length, it was still more than three orders of magnitude above background recombination (Figure 3A and Figure B). However, below 50 bp of homology on both sides a dramatic decrease in homing levels occurred (Figure 3B), whereas at 25 bp or less homing occurred at or below the level of background recombination, in contrast with comparable donor variants in which only one exon was limiting in homology.

Analysis of one-sided heterology events:
The apparent homing in the complete absence of either exon I or exon II (Figure 3A) prompted investigation of the nature of these events. Independent homing assays for donor plasmids lacking exon I or exon II were performed (Figure 4). Su⁺ phage were screened for Td^- plaques to distinguish true homing events from plasmid integrants, events that are detected increasingly when homology on one side of the break is eliminated (Figure 2). Intron homing was confirmed to have occurred by PCR analysis of plaque-purified phage (data not shown). A single homing product from each of 14 independent infections with the EI donor pMPEI-10 and 21 infections with the EII donor pMPEII-5 was selected for further analysis.

Phage DNA from independent homing events was amplified by PCR using intron-specific and phage-specific primers flanking the exon that was deleted from the donor (Figure 4A). In each event resulting from the exon I-minus or exon II-minus donor, the PCR product differed in size from the product generated by the wild-type donor (Figure 4B and data not shown). Sequence analysis of recombination junctions revealed that intron homing from the exon-minus donors resulted in deletions, representing loss in the recipient phage of the exon that was absent from the donor (Figure 4C and Figure D). In addition, various amounts of vector DNA from the donor plasmid were coinherited with the intron, explaining why in some cases a PCR product larger than that of the wild-type event was generated despite the occurrence of a deletion (Figure 4A and Figure B).

When exon I was absent from the donor allele, intron homing was associated with deletions originating at the cleavage site (CS) and extending upstream for 584–2119 bp (Figure 4C). The absence of exon II from the donor allele resulted in deletions extending downstream from the CS, but overall these were markedly smaller compared to those in sequences upstream of the CS, ranging from 18 to 776 bp (Figure 4C and Figure D). Because the CS is located 23 nt upstream of the 5' end of exon II (Figure 1A), the two events that had only 18 bp deleted resulted in recombinant phage that had a complete exon II.

Analysis of recombination junctions revealed that in each of the cases examined, a short region of homology existed between sequences in the recipient phage and sequences in the donor plasmid (Figure 5). Of the 14 events analyzed from the exon I donor (Figure 4C), 9 different sequences on the recipient phage and the donor plasmid were utilized in the repair event, indicating that the 10-fold elevation of homing above background was not due to any one particular hot spot of homology. Rather, the phage appeared to be quite promiscuous in selecting a site for initiation of repair. Similarly, the 21 events analyzed from the exon II donor (Figure 4D) utilized 13 different sequences in the recipient and the donor to effect repair. In each of the events for which the sequence was determined, a short region of homology between the donor and recipient occurred at the recombination junction. The majority of the junctions were comprised of 4–13 bp of perfect homology, although some junctions had interrupted homologies (Figure 5B). However, none of the junctions indicated a joining of donor and recipient DNA in the complete absence of homology.

View larger version (83K): In this window In a new window Download PPT slide   Figure 5. Recombination junctions resulting from one-sided heterology. Su+Td- phage from intron donors lacking exon I (A) or exon II (B) were amplified with recipient phage-specific and donor plasmid-specific primers (see Figure 4). PCR products were sequenced and recombination junctions of homing products (P) are shown along with corresponding donor (D) and recipient sequences (R). In some cases sequencing indicated the size of the deletion, but the exact junction sequence could not be determined (Figure 4, I-a, II-b, -c, -g, -j, and -l). Donor bases are indicated by underlining. Recipient bases are indicated by shading. Boxed sequences indicate the inferred crossover site. Brackets indicate sequences with three or more identical nucleotides, separated from the boxed sequence by a single mismatch. The length of sequence deleted () and the number of times each junction was isolated (n) are indicated above the sequence.

Homology effects on homing in exonuclease-deficient phage backgrounds:
Degradation of the resection segment and processing of DNA ends is required to allow precise intron insertion (Figure 1A; MUELLER et al. 1996A , MUELLER et al. 1996B ; HUANG et al. 1999 ). Homing was therefore examined under conditions of limiting homology in rnh mutant phage deficient in the 5'-3' DNA exonuclease activity of RNase H and dexA mutant phage deficient in the 3'-5' single-strand DNA exonuclease, DexA.

Homing assays were performed using the wild-type donor plasmid and plasmids carrying deletions in both exons (Table 1) with rnh or dexA recipient phage variants (Figure 6). Homing levels in the rnh recipient were reduced two- to fourfold relative to the parental recipient for donor alleles with homology ranging from 25 to >700 bp. These levels are consistent with previous analyses of intron homing in rnh phage (MUELLER et al. 1996A ; HUANG et al. 1999 ). Below 25 bp of homology, homing levels for the rnh recipient phage were equivalent to those of the wild-type phage (Figure 6A), indicating that the absence of this exonuclease does not affect homing levels when homology is very limiting.

View larger version (21K): In this window In a new window Download PPT slide   Figure 6. Homology dependence of homing in exonuclease-deficient phage. The in vivo homing assay was performed with donors containing truncations in both exons and either the parental recipient phage or exonuclease-deficient variants rnh (A) and dexA (B) (see Figure 2, Table 1). Su+/total phage ratios were plotted against the length of homology. () T4 38am51amtdIn, () T4 38am51amtdOP8, () T4 38am51am tdInrnh, () T4 38am51amtdOP8rnh (•) T4 38am51amtdIndexA, () T4 38am51amtdOP8dexA. The phage burst in the rnh infections was consistently 10-fold lower than that of the parental infections, as previously reported (HOBBS and NOSSAL 1996 ).

In contrast to the rnh phage, there was a striking elevation in homing levels in the dexA phage when homology was limiting (Figure 6B). When homology was extensive, as in the wild-type donor, the level of homing by the dexA variant was similar to that of the parental dexA⁺ phage, but as homology was decreased to 50 bp in both exons, the ratio of Su⁺/total phage in the dexA recipient was almost 100-fold greater than that of the parental recipient. At 25 bp of exon homology, homing in the dexA recipient was 1000-fold higher than with the parental phage (Figure 6B). When homology in both exons was reduced to 10 bp, homing was still 50-fold elevated with respect to the parental phage. Homing events for donors with 25-bp and 10-bp exons were confirmed by the halo assay and PCR analysis (data not shown). However, when both exons were completely deleted from the donor allele, homing levels became indistinguishable from background recombination levels. More than 10¹⁰ total phage from several independent infections were analyzed by the halo assay and/or PCR but no I-TevI-mediated intron acquisition events in the absence of flanking homology were observed. For all donor alleles, the level of background recombination, as well as the phage burst, for the dexA and parental phage recipients were similar, indicating that the exonuclease-deficient variants do not have a general replication advantage over the parental phage (Figure 6 and data not shown).

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Homology dependence of intron homing in phage T4:
Homology dependence of intron homing in phage T4: Previous analyses of intron homing and DSB repair in prokaryotes have indicated a dependence on length of homology on both sides of the break (QUIRK et al. 1989A ; CLYMAN and BELFORT 1992 ; PARKER et al. 1996 ). Because intron homing occurs naturally in the context of the T4 lytic cycle, a systematic analysis of homology dependence within this context was important, not only for mechanistic reasons, but to establish which phage factors might influence homing when homology is limiting. A genetic system was developed, which allowed quantitative analysis of intron homing levels under various conditions of exon homology and phage genotype. The use of a selectable, splicing-deficient intron provided exquisite sensitivity, allowing for detection of rare intron acquisition events and a genetic means for distinguishing between homing and plasmid integration in recipient parental and relevant mutant phage.

In phage T4, the occurrence of a DSB stimulates the initiation of recombination-dependent replication at the break site (KREUZER et al. 1995 ; GEORGE and KREUZER 1996 ; MUELLER et al. 1996A ). Accordingly, infection of cells harboring an intron donor plasmid with full-length exons with phage carrying an intronless td allele resulted in high levels (>50%) of intron inheritance in the presence of I-TevI (Figure 3A and Figure B). Homing levels remained elevated with as little as 50 bp of homology on one or both sides of the break, after which they began to decline, although they remained at least 10-fold above background, even when one exon was completely deleted from the donor allele (Figure 3A). However, homology lengths of 25 bp and less on both sides of the break resulted in homing levels that were equal to or less than background recombination levels occurring in the absence of an I-TevI cleavage site (Figure 3B).

It appears that the critical length of homology for intron homing in wild-type T4 lies between 25 and 50 bp. This is in agreement with previous studies on spontaneous recombination in T4, which indicated that 25 to 50 bp of homology is required for efficient recombination (BAUTZ and BAUTZ 1967 ; SINGER et al. 1982 ). Likewise, 35 to 50 bp of homology is required for efficient intron homing in a phage model system (PARKER et al. 1996 ). This similarity in homology length requirements likely reflects the functional components of DSB-mediated recombination in these systems. One possibility is that the strand transferase is limited by its ability to recognize or bind to short homologies. In vitro analysis of UvsX indicated that it does not catalyze pairing between molecules with <30 bp of homology (SALINAS et al. 1995 ). However, RecA effects pairing in vitro between DNA molecules sharing 15 bp of homology (HSIEH et al. 1992 ) and stimulates recombination in vivo between molecules sharing only 13 bp of homology (KING and RICHARDSON 1986 ).

Exon deletions and microhomology mark illegitimate homing events:
Analysis of those homing events with one-sided homology revealed that in each case deletions occurred in the recipient phage extending from the CS to various positions in the exon sequence corresponding to the exon that was absent from the donor. Similarly, DSB repair in a phage T7 in vitro system was associated with deletion of nonhomologous DNA occurring at the ends (LAI and MASKER 1998 ). Sequence analysis of the recombination junctions indicated that numerous short stretches of homology between the donor plasmid and the recipient phage ranging from 4 to 13 bp were used in repair of the DSB (Figure 5). This differs from a similar analysis conducted on one-sided repair events occurring in a phage -td intron model system where two out of four junctions examined appeared to form in the complete absence of homology (PARKER et al. 1996 ). Similarly, analysis of restriction endonuclease-mediated illegitimate recombination events in E. coli revealed that only 1–4 nt of homology were present at the junctions, with only a single nucleotide of homology appearing in 5 of 10 cases (KUSANO et al. 1997 ). Indeed, illegitimate DSB repair events occurring in E. coli, yeast, and mammalian systems have been attributed to an end-joining mechanism (for example, MOORE and HABER 1996 ; KUSANO et al. 1997 ). The propensity of phage T4 to utilize short regions of homology to effect DSB repair in the absence of extensive homology indicates that the T4 recombination machinery can catalyze strand transfer between very short regions of homology. Correspondingly, spontaneous deletions in T4 have been associated with recombination between direct repeats ranging from 4 to 15 bp in length (PRIBNOW et al. 1981 ; OWEN et al. 1983 ; MOSIG et al. 1998 ). The interesting question of whether this mechanism involves UvsX is difficult to address in vivo because of the requisite role of this protein in intron homing (CLYMAN and BELFORT 1992 ; MUELLER et al. 1996A ).

A comparison of the deletions produced in several independent events indicated that the extent of degradation was markedly greater on the upstream side of the cleavage site than on the downstream side (Figure 4C and Figure D). The two most likely interpretations of this result relate to the distribution of microhomologies on one hand and to access to exonucleolytic degradation on the other. First, regions of microhomology between the phage and the donor may be distributed such that on the downstream side the more lengthy regions are located closer to the break than on the upstream side. Indeed, hot spots of spontaneous deletion formation in T4 display microhomology between direct repeats, and the frequency of events occurring at particular sites corresponds well with their length of microhomology (ALLGOOD and SILHAVY 1988 ). Because the most upstream junction, located >2 kb from the cleavage site, exhibits the longest stretch of uninterrupted homology (13 bp), and appeared in 4 out of 14 events (Figure 4C and Figure 5A, I-i), it seems likely that the length of microhomology plays a role in determining the site of repair. However, computer analysis indicated that numerous segments of microhomology exist in proximal upstream regions of the donor and recipient, with at least five sites displaying six or seven perfect matches (data not shown). Another factor must therefore also be considered to account for the observed bias. Indeed, the upstream and downstream ends produced by the cleavage reaction are not equally available for exonucleolytic degradation because I-TevI remains bound to the downstream cleavage product, providing protection (MUELLER et al. 1996B ). This bias in exonucleolytic processing is likely to account in large measure for the lengthier degradation tracts in exon I than exon II (Figure 4C and Figure D), much as it does for lengthier coconversion tracts in the upstream than the downstream exon (MUELLER et al. 1996B ).

Exonuclease activity influences homing levels with limiting homology:
The gp46/47 complex and RNase H likely contribute the majority of 5'-3' end processing occurring during recombination and replication in T4. Because the gp46/47 complex is required for phage viability and the homing reaction (EPSTEIN et al. 1963 ; WIBERG 1966 ; MUELLER et al. 1996A ), we were unable to determine its influence on homing when homology was limiting. However, the essential nature of gp46/47 makes the complex less likely than nonessential exonucleases to influence homing efficiency under suboptimal conditions such as limiting homology. We therefore examined T4 RNase H, which has 5'-3' DNA exonuclease activity and has been implicated in DNA replication, repair, and intron homing (HOLLINGSWORTH and NOSSAL 1991 ; MUELLER et al. 1996A ; WOODWORTH and KREUZER 1996 ; HUANG et al. 1999 ). Homing levels in rnh phage infections were somewhat decreased compared to the wild-type phage for the majority of homology lengths analyzed (Figure 6A). However, it remains unclear whether these subtle decreases in homing levels in rnh phage are directly related to loss of exonuclease activity or due to an indirect effect on replication or repair (WOODWORTH and KREUZER 1996 ; HUANG et al. 1999 ). If the effect of RNase H on intron homing is direct, rather than through DNA replication, the enzyme is likely to generate 3' invasive ends through its 5'-3' DNA exonuclease activity (HUANG et al. 1999 ). Therefore, an rnh deficiency may lead to shorter 3' tails available for the search for homology.

Conversely, when donor alleles with limiting homology in both exons were analyzed in phage deficient in the 3'-5' exonuclease DexA, homing levels were significantly elevated above those generated with the parental recipient phage (Figure 6B). The most likely explanation for this result is that in the absence of DexA, the 3' tail, containing homologous td sequences, is not readily degraded and consequently is utilized more efficiently during the repair process. This hypothesis is supported by genetic and biochemical evidence that DexA is involved in 3' end processing during homing (HUANG et al. 1999 ). The indication that a lengthy 3' tail with an extended half-life can more productively find limiting homologies suggests that length of 3' tails is a major determinant in the efficiency of DSB repair in T4.

Nucleases and intron translocation:
Group I introns occupy a wide variety of genomic locations within a phylogenetically diverse set of organisms (LAMBOWITZ and BELFORT 1993 ; BELFORT et al. 1995 ). Among the T-even phages, group I introns are variably distributed, indicating that they may have been acquired after the divergence of these strains from a common ancestor (SHUB et al. 1988 ; QUIRK et al. 1989B ). Several observations have led to the hypothesis that intron acquisition at new sites may be mediated by intron endonuclease-stimulated DSB repair (MUELLER et al. 1993 ). Indeed, I-TevI is highly tolerant of perturbations within its target sequence, suggesting that cleavage events occur at related sequences (BRYK et al. 1993 , BRYK et al. 1995 ). Furthermore, substantial phylogenetic and molecular evidence supports DSB-mediated horizontal transfer of a group I intron to the mitochondrial genome of numerous plant species (CHO et al. 1998 ).

Considering that the recognition sequence of I-TevI spans 37 bp (BRYK et al. 1995 ), it is reasonable to speculate that an ectopic, low-specificity recognition site is likely to share some homology with the natural recognition sequence and therefore with exon sequences flanking the intron. The proclivity of the DSB repair apparatus to seek very short homologous sequences, as in phage T4 (Figure 5), suggests that these sequences may be preferentially utilized for repair. We further propose that genetic background or fluctuations in the levels of DNA nucleases may be influential factors in such DSB-mediated intron translocation. Thus, cleavage at a low specificity site with reduced 3'-5' exonucleolytic degradation and in the presence of endonuclease-mediated protection (MUELLER et al. 1996B ) may preserve 3' tails at the break site and allow the ends to participate in illegitimate DSB repair. Significantly, repair events were observed in the absence of homology on the exon II side in which the complete td coding sequence was maintained on that side, because degradation did not extend beyond the resection segment (Figure 1A, Figure 4D, and Figure 5B, II-a). Such an event would sustain gene function by maintaining the proper coding sequence of the newly invaded gene after intron splicing.

	ACKNOWLEDGMENTS

We are grateful to Gregory Lopez, Yi-Jiun Huang, and Dilip Nag for helpful discussions and comments on the manuscript. The Molecular Genetics Core facility provided oligonucleotides and automated DNA sequencing. We thank Dorie Smith for technical assistance and Maryellen Carl for preparing the manuscript. This work was supported by National Institutes of Health grants GM-39422 and GM-44844 to M.B.

Manuscript received May 12, 1999; Accepted for publication August 30, 1999.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ALLGOOD, N. D., and T. J. SILHAVY, 1988 Illegitimate recombination in bacteria, pp. 309–330 in Genetic Recombination, edited by R. KUCHERLAPATI and G. R. SMITH. American Society for Microbiology, Washington, DC.

BAUTZ, F. A. and E. K. F. BAUTZ, 1967  Transformation in phage T4: minimal recognition length between donor and recipient DNA. Genetics 57:887-895[Free Full Text].

BELFORT, M. and P. S. PERLMAN, 1995  Mechanisms of intron mobility. J. Biol. Chem. 270:30237-30240[Free Full Text].

BELFORT, M., M. E. REABAN, T. COETZEE, and J. Z. DALGAARD, 1995  Prokaryotic introns and inteins: a panoply of form and function. J. Bacteriol. 177:3897-3903[Free Full Text].

BELL-PEDERSEN, D., S. M. QUIRK, M. AUBREY, and M. BELFORT, 1989  A site-specific endonuclease and co-conversion of flanking exons associated with the mobile td intron of phage T4. Gene 82:119-126[Medline].

BELL-PEDERSEN, D., S. QUIRK, J. CLYMAN, and M. BELFORT, 1990  Intron mobility in phage T4 is dependent upon a distinctive class of endonucleases and independent of DNA sequences encoding the intron core: mechanistic and evolutionary implications. Nucleic Acids Res. 18:3763-3770[Abstract/Free Full Text].

BRYK, M., S. M. QUIRK, J. E. MUELLER, N. LOIZOS, and C. LAWRENCE et al., 1993  The td intron endonuclease makes extensive sequence tolerant contacts across the minor groove of its DNA target. EMBO J. 12:2141-2149[Medline].

BRYK, M., M. BELISLE, J. E. MUELLER, and M. BELFORT, 1995  Selection of a remote cleavage site by I-TevI, the td intron-encoded endonuclease. J. Mol. Biol. 247:197-210[Medline].

CHANDRY, P. S. and M. BELFORT, 1987  Activation of a cryptic 5' splice site in the upstream exon of the phage T4 td transcript: exon context, missplicing, and mRNA deletion in a fidelity mutant. Genes Dev. 1:1028-1037[Abstract/Free Full Text].

CHANG, A. C. Y. and S. N. COHEN, 1978  Construction and characterization of amplifiable multicopy DNA cloning vehicles derived from the P15A cryptic miniplasmid. J. Bacteriol. 134:1141-1156[Abstract/Free Full Text].

CHO, Y., Y. QIU, P. KUHLMAN, and J. D. PALMER, 1998  Explosive invasion of plant mitochondria by a group I intron. Proc. Natl. Acad. Sci. USA 95:14244-14249[Abstract/Free Full Text].

CLYMAN, J. and M. BELFORT, 1992  Trans and cis requirements for intron mobility in a prokaryotic system. Genes Dev. 6:1269-1279[Abstract/Free Full Text].

DUJON, B., 1989  Group I introns as mobile genetic elements: facts and mechanistic speculations—a review. Gene 82:91-114[Medline].

EPSTEIN, R. H., A. BOLLE, C. M. STEINBERG, E. KELLENBERGER, and E. BOY DELA TOUR et al., 1963  Physiological studies of conditional lethal mutants of bacteriophage T4D. Cold Spring Harbor Symp. Quant. Biol. 28:375-394[Abstract/Free Full Text].

GAUSS, P., M. GAYLE, R. B. WINTER, and L. GOLD, 1987  The bacteriophage T4 dexA gene: sequence and analysis of a gene conditionally required for DNA replication. Mol. Gen. Genet. 206:24-34[Medline].

GEORGE, J. W. and K. N. KREUZER, 1996  Repair of double-strand breaks in bacteriophage T4 by a mechanism that involves extensive DNA replication. Genetics 143:1507-1520[Abstract].

GRUBER, H., G. KERN, P. GAUSS, and L. GOLD, 1988  Effect of DNA sequence and structure on nuclease activity of the DexA protein of bacteriophage T4. J. Bacteriol. 170:5830-5836[Abstract/Free Full Text].

HALL, D. H., I. TESSMAN, and O. KARLSTROM, 1967  Linkage of T4 genes controlling a series of steps in pyrimidine biosynthesis. Virology 31:442-448[Medline].

HOBBS, L. J. and N. G. NOSSAL, 1996  Either bacteriophage T4 RNase H or Escherichia coli DNA polymerase I is essential for phage replication. J. Bacteriol. 178:6772-6777[Abstract/Free Full Text].

HOLLINGSWORTH, H. C. and N. G. NOSSAL, 1991  Bacteriophage T4 encodes an RNase H which removes RNA primers made by the T4 DNA replication system in vitro.. J. Biol. Chem. 266:1888-1897[Abstract/Free Full Text].

HSIEH, P., C. S. CAMERINI-OTERO, and R. D. CAMERINI-OTERO, 1992  The synapsis event in the homologous pairing of DNAs: RecA recognizes and pairs less than one helical repeat of DNA. Proc. Natl. Acad. Sci. USA 89:6492-6496[Abstract/Free Full Text].

HUANG, Y-J., M. M. PARKER, and M. BELFORT, 1999  Role of exonucleolytic degradation in group I intron homing in phage T4. Genetics 153:1501-1512[Abstract/Free Full Text].

KING, S. R. and J. P. RICHARDSON, 1986  Role of homology and pathway specificity for recombination between plasmids and bacteriophage lambda. Mol. Gen. Genet. 204:141-147[Medline].

KREUZER, K. N., and S. W. MORRICAL, 1994 Initiation of DNA replication, pp. 28–42 in Molecular Biology of Bacteriophage T4, edited by J. D. KARAM. American Society for Microbiology Press, Washington, DC.

KREUZER, K. N., M. SAUNDERS, L. J. WEISLO, and H. W. E. KREUZER, 1995  Recombination-dependent DNA replication stimulated by double-strand breaks in bacteriophage T4. J. Bacteriol. 177:6844-6853[Abstract/Free Full Text].

KUSANO, K., K. SAKAGAMI, T. YOKOCHI, T. NAITO, and Y. TOKINAGA et al., 1997  A new type of illegitimate recombination is dependent on restriction and homologous interaction. J. Bacteriol. 179:5380-5390[Abstract/Free Full Text].

LAI, Y. and W. MASKER, 1998  In vitro repair of gaps in bacteriophage T7 DNA. J. Bacteriol. 180:6193-6202[Abstract/Free Full Text].

LAMBOWITZ, A. M. and M. BELFORT, 1993  Introns as mobile genetic elements. Annu. Rev. Biochem. 62:587-622[Medline].

LUDER, A. and G. MOSIG, 1982  Two alternative mechanisms for initiation of DNA replication forks in bacteriophage T4: priming by RNA polymerase and by recombination. Proc. Natl. Acad. Sci. USA 79:1101-1105[Abstract/Free Full Text].

MICKELSON, C. and J. S. WIBERG, 1981  Membrane-associated DNase activity controlled by genes 46 and 47 of bacteriophage T4D and elevated DNase activity associated with the T4 das mutation. J. Virol. 40:65-77[Abstract/Free Full Text].

MOORE, J. K. and J. E. HABER, 1996  Cell cycle and genetic requirements of two pathways of nonhomologous end-joining repair of double-strand breaks in Saccharomyces cerevisiae.. Mol. Cell. Biol. 16:2164-2173[Abstract].

MOSIG, G., 1998  Recombination and recombination-dependent DNA replication in bacteriophage T4. Annu. Rev. Genet. 32:379-413[Medline].

MOSIG, G., N. E. COLOWICK, and B. C. PIETZ, 1998  Several new bacteriophage T4 genes, mapped by sequencing deletion endpoints between genes 56 (dCTPase) and dda (a DNA-dependent ATPase-helicase) modulate transcription. Gene 223:143-155[Medline].

MUELLER, J. E., M. BRYK, N. LOIZOS and M. BELFORT, 1993 Homing endonucleases, pp. 111–143 in Nucleases, Ed. 2, edited by S. M. LINN, R. S. LLOYD and R. J. ROBERTS. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

MUELLER, J. E., J. CLYMAN, Y. HUANG, M. M. PARKER, and M. BELFORT, 1996a  Intron mobility in phage T4 occurs in the context of recombination-dependent DNA replication by way of multiple pathways. Genes Dev. 10:351-364[Abstract/Free Full Text].

MUELLER, J. E., D. SMITH, and M. BELFORT, 1996b  Exon coconversion biases accompanying intron homing: battle of the nucleases. Genes Dev. 10:2158-2166[Abstract/Free Full Text].

OWEN, J. E., D. W. SCHULTZ, A. TAYLOR, and G. R. SMITH, 1983  Nucleotide sequence of the lysozyme gene of bacteriophage T4. Analysis of mutations involving repeated sequences. J. Mol. Biol. 165:229-248[Medline].

PARKER, M. M., D. A. COURT, K. PREITER, and M. BELFORT, 1996  Homology requirements for double-strand break-mediated recombination in a phage lambda-td intron model system. Genetics 143:1057-1068[Abstract].

PRIBNOW, D., D. C. SIGURDSON, L. GOLD, B. S. SINGER, and C. NAPOLI, 1981  rII cistrons of bacteriophage T4. DNA sequence around the intercistronic divide and positions of genetic landmarks. J. Mol. Biol. 149:337-376[Medline].

QUIRK, S. M., D. BELL-PEDERSEN, and M. BELFORT, 1989a  Intron mobility in the T-Even phages: high frequency inheritance of group I introns promoted by intron open reading frames. Cell 56:455-465[Medline].

QUIRK, S. M., D. BELL-PEDERSEN, J. TOMASCHEWSKI, W. RUGER, and M. BELFORT, 1989b  The inconsistent distribution of introns in T-even phages indicates recent genetic exchanges. Nucleic Acids Res. 17:301-315[Abstract/Free Full Text].

ROSENBERG, M., Y. S. HO, and A. SHATZMAN, 1983  The use of pKC30 and its derivatives for controlled expression of genes. Methods Enzymol. 101:123-138[Medline].

SALINAS, F., H. JIANG, and T. KODADEK, 1995  Homology dependence of UvsX protein-catalyzed joint molecule formation. J. Biol. Chem. 270:5181-5186[Abstract/Free Full Text].

SELICK, H. E., K. N. KREUZER, and B. M. ALBERTS, 1988  The bacteriophage T4 insertion/substitution vector system. A method for introducing site-specific mutations into the virus chromosome. J. Biol. Chem. 263:11336-11347[Abstract/Free Full Text].

SHUB, D. A., J. M. GOTT, M.-Q. XU, B. F. LANG, and F. MICHEL et al., 1988  Structural conservation among three homologous introns of bacteriophage T4 and the group I introns of eukaryotes. Proc. Natl. Acad. Sci. USA 85:1151-1155[Abstract/Free Full Text].

SINGER, B. S., L. GOLD, P. GAUSS, and D. H. DOHERTY, 1982  Determination of the amount of homology required for recombination in bacteriophage T4. Cell 31:25-33[Medline].

WIBERG, J. S., 1966  Mutants of bacteriophage T4 unable to cause breakdown of host DNA. Proc. Natl. Acad. Sci. USA 55:614-621[Free Full Text].

WOODWORTH, D. L. and K. N. KREUZER, 1996  Bacteriophage T4 mutants hypersensitive to an antitumor agent that induces topoisomerase-DNA cleavage complexes. Genetics 143:1081-1090[Abstract].

This article has been cited by other articles:

				E. A. Gibb and D. R. Edgell An RNA Hairpin Sequesters the Ribosome Binding Site of the Homing Endonuclease mobE Gene J. Bacteriol., April 1, 2009; 191(7): 2409 - 2413. [Abstract] [Full Text] [PDF]

				E. A. Gibb and D. R. Edgell Multiple Controls Regulate the Expression of mobE, an HNH Homing Endonuclease Gene Embedded within a Ribonucleotide Reductase Gene of Phage Aeh1 J. Bacteriol., July 1, 2007; 189(13): 4648 - 4661. [Abstract] [Full Text] [PDF]

				N. C. Friedrich, E. Torrents, E. A. Gibb, M. Sahlin, B.-M. Sjoberg, and D. R. Edgell Insertion of a homing endonuclease creates a genes-in-pieces ribonucleotide reductase that retains function PNAS, April 10, 2007; 104(15): 6176 - 6181. [Abstract] [Full Text] [PDF]

				L. Sandegren and B.-M. Sjoberg Distribution, Sequence Homology, and Homing of Group I Introns among T-even-like Bacteriophages: EVIDENCE FOR RECENT TRANSFER OF OLD INTRONS J. Biol. Chem., May 21, 2004; 279(21): 22218 - 22227. [Abstract] [Full Text] [PDF]

				P. Haugen, V. Reeb, F. Lutzoni, and D. Bhattacharya The Evolution of Homing Endonuclease Genes and Group I Introns in Nuclear rDNA Mol. Biol. Evol., January 1, 2004; 21(1): 129 - 140. [Abstract] [Full Text] [PDF]

				A. Belle, M. Landthaler, and D. A. Shub Intronless homing: site-specific endonuclease SegF of bacteriophage T4 mediates localized marker exclusion analogous to homing endonucleases of group I introns Genes & Dev., February 1, 2002; 16(3): 351 - 362. [Abstract] [Full Text] [PDF]

				D. R. Edgell, M. Belfort, and D. A. Shub Barriers to Intron Promiscuity in Bacteria J. Bacteriol., October 1, 2000; 182(19): 5281 - 5289. [Full Text]

				Y.-J. Huang, M. M. Parker, and M. Belfort Role of Exonucleolytic Degradation in Group I Intron Homing in Phage T4 Genetics, December 1, 1999; 153(4): 1501 - 1512. [Abstract] [Full Text]