We examined the requirement of recombination functions for marker rescue of cryptic prophage genes within the Escherichia coli chromosome. We infected lysogenic host cells with imm434 phages and selected for recombinant imm phages that had exchanged the imm434 region of the infecting phage for the heterologous 2.6-kb imm region from the prophage. Phage-encoded activity, provided by either Red or NinR functions, was required for the substitution. Red^– phages with NinR, internal NinR deletions of rap-ninH, or orf-ninC were 117-, 12-, and 5-fold reduced for imm rescue in a Rec⁺ host, suggesting the participation of several NinR activities. RecA was essential for NinR-dependent imm rescue, but had slight influence on Red-dependent rescue. The host recombination activities RecBCD, RecJ, and RecQ participated in NinR-dependent recombination while they served to inhibit Red-mediated imm rescue. The opposite effects of several host functions toward NinR- and Red-dependent imm rescue explains why the independent pathways were not additive in a Rec⁺ host and why the NinR-dependent pathway appeared dominant. We measured the influence of the host recombination functions and DnaB on the appearance of ori-dependent replication initiation and whether ori replication initiation was required for imm marker rescue.

---

MARKER rescue recombination to produce gene substitutions involves exchanges within regions of homology straddling a marker of interest. Strong modern evidence for the shuffling of phage gene modules in nature is provided by the stx phages and prophages of Escherichia coli, which share the genome organization of bacteriophage (BRüSSOW et al. 2004). Early workers identified phage-prophage marker rescue, where an infecting was capable of rescuing a gene present on a homologous cryptic prophage in a lysogenic cell. SIGNER and WEIL (1968) used a spot test involving the rescue of an h (unspecified host range) marker from rec⁺ cells with a cryptic prophage (deleted for a large portion of prophage, including the imm region) that was infected by h, and h recombinants were selected on host cells that were resistant to infection by h but sensitive to h. Using this assay, SIGNER and WEIL (1968) were able to screen hydroxylamine-treated infecting phage for deficiency in marker rescue. Several mutants with reduced ability to rescue prophage markers were subsequently mapped as recombination-defective red mutants. ECHOLS and GINGERY (1968) recovered sus⁺ recombinants that were formed by marker rescue between an infecting sus phage and a defective prophage in a lysogen. Both studies concluded that Red functions of were required for phage-prophage marker rescue in E. coli hosts defective for the host recA function. The Red-dependent recombination activity (reviewed by STAHL 1998; KUZMINOV 1999; COURT et al. 2002) depends upon the expression of genes exo and bet (or Red, Redß; combined, Red) along with gam. Red-dependent recombination is initiated by double-strand breaks, and when marked Red⁺ phages infect cells blocked for DNA replication, the x exchanges are focused near the cos ends, the only site of an initiating double-strand break (TARKOWSKI et al. 2002). Murphy and co-workers (MURPHY 1998; MURPHY et al. 2000) constructed an E. coli strain in which the cellular recBCD genes (SMITH 2001) were replaced with exo-bet and placed under lac promoter control. They found that the activities supported recombination between the cellular chromosome and linear DNA fragments at an elevated level. Recombination in these recBCD cells, lacking Gam, depended upon Exo and Beta, was greatly reduced in recA mutants, and required host recombination genes recQ, recO, recR, recF, and ruvC, but not recJ or recG (MURPHY 1998; POTEETE et al. 1999; MURPHY et al. 2000; POTEETE and FENTON 2000). The Red functions can facilitate chromosomal engineering, i.e., substituting or disrupting genes in an E. coli chromosome or plasmid (YU et al. 2000; COURT et al. 2002). In the absence of E. coli methyl-directed mismatch repair activity, the inheritance of markers from single-strand DNA oligonucleotides required only the Bet function (CONSTANTINO and COURT 2003).

The co-integration of a plasmid into a phage, each sharing limited DNA homology, is shown as a single reciprocal crossover addition reaction (Figure 1A). KING and RICHARDSON (1986) and SHEN and HUANG (1986) found that packaged plasmid-phage co-integrates were formed by homologous recombination proceeding predominantly via the E. coli RecBCD pathway (reviewed by KOWALCZYKOWSKI et al. 1994; STAHL 1998; KUZMINOV 1999) when cells were infected with int^– red^– or int-gam phages sharing limited DNA sequence homology with a small plasmid carried within the cell. HOLLIFIELD et al. (1987) found that the formation of phage-plasmid co-integrates by int-gam phage Charon 4A (BLATTNER et al. 1977) depended on an encoded phage function mapping between genes P and Q within the NinR (KRöGER and HOBOM 1982) interval, defined by deletion nin5 and including the genes ren (TOOTHMAN and HERSKOWITZ 1980) and nine open reading frames designated ninA–ninI (DANIELS et al. 1983). They identified the rap function (recombination adept with plasmid) that mapped to orf204 (ninG gene) as being required for RecBCD pathway-mediated formation of phage-plasmid co-integrates. Even though Charon 4A was nin5, it encoded a similar determinant to rap that mapped within phage 80-derived sequences included in the cloning vector. The rap product (HOLLIFIELD et al. 1987; STAHL et al. 1995; TARKOWSKI et al. 2002) represents a branch-specific endonuclease that targets recombination intermediates generated in vivo and in vitro (SHARPLES et al. 1998, 1999). It can function as a genuine Holliday junction resolvase when presented with DNA substrates containing sufficient homology at the crossover (SHARPLES et al. 2004) and can substitute for the E. coli RuvC Holliday junction resolvase in Red recombination (POTEETE et al. 2002). Yet another recombination function removed by nin5 was localized by SAWITZKE and STAHL (1992) to orf146 (ninB gene) and termed orf because it could complement single mutations in the host genes recO, recR, and recF in recBC sbcB sbcC cells and thus function in RecF-dependent pathway crossovers between two co-infecting phages. Orf, with a monomer molecular mass of 16.6 kDa, was found to have pleiotropic effects on recombination, replication, and repair in E. coli. Orf suppresses the mutant phenotype not only of recF, recO, and recR, but also of ruvAB and ruvC (POTEETE and FENTON 2000; POTEETE et al. 2002; POTEETE 2004).

View larger version (25K): In this window In a new window Download PPT slide   FIGURE 1.— (A) Homologous recombination between common sequence (rectangle) in phage and plasmid forming a phage-plasmid co-integrate structure (see text). SHEN and HUANG (1986) and HOLLIFIELD et al. (1987) used infecting phages that carried amber mutations in both genes A and B and were deleted for recombination functions int-xis-exo-bet-gam and the genes within the NinR region, which were substituted by another set of genes from Nin80. (B) Phage-prophage marker rescue of the imm genes from a cryptic prophage (lacking a cos site) within the bacterial chromosome by an infecting imm434 phage. The open rectangles are DNA strands of a circular E. coli chromosome; solid lines are DNA, as infecting phage or as cryptic prophage within E. coli chromosome; the solid rectangle is imm434; the stretched diamond structure is imm; the stippled rectangle is bio275 substitution of E. coli DNA for the prophage genes int-cIII in the Y836 chromosome (and in infecting phages in C); the solid triangle on map of Y836 represents the 431 deletion removing rightward of prophage genes and adjacent chromosomal genes. The actual bases that were deleted or substituted are described in MATERIALS AND METHODS. The drawing is to scale, except that the sizes of the imm434 and imm intervals were made equal when imm434 was smaller by 1.2 kb. (C) Maps of infecting phages used in Tables 2–5. The deletion nin5 on infecting phage encroached on the interval of base homology to the right of the imm434 marker (B) between the cryptic prophage and the infecting phage, reducing it from 2519 (or 2565) bp (with 431 endpoint) to 2257 bp. The bio275 substitution on the infecting phage increased base homology to the left of the imm434 marker: the 2281-bp homology was unchanged, but homology was increased by the size of the bio275 substitution (see HAYES et al. 1990) carried on both prophage and infecting int-red-gam phage.

 

View this table: In this window In a new window   TABLE 2 Marker rescue of imm from cryptic prophage

 

View this table: In this window In a new window   TABLE 5 CI+ limits imm rescue in Y836 at 39° following imm434 infection

 
The intercrossing between two phage genomes is routinely drawn as a single splice reaction. MOTAMEDI et al. (1999) demonstrated efficient homologous recombination between co-infecting int-gam-defective NinR phages in rec⁺ hosts, showing that the host could provide needed recombination functions for phage-phage recombination and that the Red or NinR phage activities were not required. This observation supported a prior report by STAHL et al. (1995) that Rap function had no effect on the frequency of recombination between co-infecting phages sharing DNA homology, i.e., x exchange in an infected cell. In Red-mediated x crosses occurring when DNA replication was blocked, TARKOWSKI et al. (2002) showed that the NinR products Orf and Rap influenced end focusing at double-strand breaks, but that nin5 had little or no effect on the outcome of such crosses in a recA^– host.

This study was undertaken on the basis of observations from a rapid marker rescue test (functional immunity assay: HAYES and HAYES 1986; HAYES 1991) in which cells with a cryptic imm prophage were stabbed to a fresh overlay plate containing cells lysogenized with imm434 plus added free imm434cI phage. Marker exchange between the cryptic prophage and the infecting imm434cI phage (Figure 1B) yielded imm recombinant phage that were revealed by a lysis spot (up to a 0.5-cm radius) surrounding the stabbed colony formed in the cell lawn of the imm434 lysogen. The formation of imm recombinant phages arose even when the stabbed cells (with cryptic imm prophage) carried a null mutation in the key host recombination functions recA (trace lysis), recB, recD, recG, recJ, and ruvC or were double null mutants recA recD, recD recF, and recF recJ. In these crosses, the 1.4-kb imm434 DNA interval on the infecting phage (Figure 1C) was substituted by rescue of the heterologous 2.6-kb imm DNA interval (which included the additional genes rexA-rexB) present in the cryptic prophage (K. ASAI and S. HAYES, unpublished results). We examined whether rescue of the imm region depended solely upon gene products encoded by the infecting phage. We found that 's Red or its NinR each function in the absence of the other to support imm exchange. The host requirements for the Red- or Nin-mediated imm rescue differed significantly. The two systems were partially competitive, rather than additive, when both were expressed. We monitored the influence of DnaB and host recombination functions on DNA replication from ori to assess the participation of prophage replication initiation in imm rescue.

Bacteria:

E. coli strains are shown in Table 1. Strain Y836 (infected host drawn in Figure 1B) carries a cryptic prophage, is SA500 F^– his87 relA1 strA181 (bio275 cI[Ts]857 431), and was made from Y832 (HAYES and HAYES 1986) derived from SA431 (ADHYA et al. 1968; STEVENS et al. 1971), which is a derivative of SA302 [AM3100 str^r his87 (cI[Ts]857)], itself a -lysogen of SA500 F^– his87 relA1 strA181 (HAYES 1991). The cI repressor (CI) in these strains, in addition to being temperature sensitive (Ts), carries an Ind mutation (HAYES and HAYES 1986). Y836 was made by replacing the genes int-cIII (DANIELS et al. 1983) in Y832/SA431 with bio275 DNA (Figure 1B) from the transducing phage bio275 (HAYES and HAYES 1986; HAYES et al. 1990). Both strains Y836 and Y832/SA431 are deleted for the NinR functions (KRöGER and HOBOM 1982; DANIELS et al. 1983; CHENG et al. 1995) by 431, where the left endpoint of 431 is between 40,764 and 40,810 bp DNA (HAYES 1991) within orf146 (ninB). The variants of Y836 were made by P1vir transduction of appropriate markers. The recA character of the donor alleles and the recA transductants was demonstrated by showing that Fec^– -phage, e.g., those defective for exo-bet-gam, as int-red-gam imm434 (bio275imm434), did not form plaques on these recA hosts. In contrast, the recD recA variant permitted efficient plating by Fec^– phage (AMUNDSEN et al. 1986). The grpD55 marker (SAITO and UCHIDA 1978) that was moved into Y836 from W3350 grpD55 malF::Tn10 was genetically mapped to dnaB (BULL and HAYES 1996) and shown by sequence analysis (M. HORBAY and S. HAYES, unpublished results) to be an allele of dnaB.

View this table: In this window In a new window   TABLE 1 E. coli strains

 

Plasmids:

pCH1 includes bases 34,449–41,732 (HAYES et al. 1997). pHB30^nl42-9 (abbreviated pCI⁺) was derived from pHB30 (BULL 1995), which includes pBR322 (bases 375–4286) and cI[Ts]857 (bases 34,499–34,696, 36,965–38,103, and 38,814–40,806; see Figure 2B). pCI⁺ was a cI[Ts]857 to cI⁺ ( 37,742 T to C) revertant (isolated and sequenced by M. Horbay) of pHB30. It expresses sufficiently high levels of CI repressor to prevent the plating of cI mutants and vir (data not shown). pRP42 is Amp^R ColE1 and was from M. Ptashne via SMR/HB (Table 1) and expresses the imm434 CI⁺ repressor. Cells with pRP42 infected at 39° with cI72 yielded no imm434 recombinants (frequency <1 x 10^–8) when scored on TC600() lysogens. Plasmids pTP914 and pTP915 were from A. R. Poteete. Plasmid pTP914 (POTEETE et al. 2002) is AatII-galK (N-terminal end) p_mac (TTTACA:-35; TATAAT:-10; RBS) rap SacI Kan^R from Tn903-galK (N-terminal end) BamHI pBR322:ori-bla, where the ApaI-p_mac-rap-SacI interval was cloned into pTP838 (MURPHY et al. 2000), and in pTP915 gfp replaces rap. POTEETE et al. (2002) reported that the Rap⁺ phenotype in pTP914 was unaffected by the presence or absence of the lac inducer IPTG, suggesting a basal level of expression of rap from p_mac.

View larger version (37K): In this window In a new window Download PPT slide   FIGURE 2.— Assay for prophage replication initiation from ori. (A) The nonexcisable cryptic fragment inserted (short arrow) within the E. coli chromosome remains repressed at 30° where the prophage repressor is active. Shifting cells to 38° inactivates the CI857 repressor for prophage transcription and replication initiation from ori. Multiple bidirectional replication initiation events from ori generate the onion-skin replication structure drawn at right. (B) DNA (thick solid line) fragment within the E. coli chromosome (open boxes), BstEII restriction sites within and the E. coli chromosome showing bands generated by cleavage (HAYES et al. 1990), and the region amplified to prepare a DNA probe. (C) Assay for replication initiation from ori upon shifting culture cells to 42° to induce the cryptic prophage. Cells with plasmid pCI+ are not derepressed for transcription or ori replication at 42°, permitting a comparison between any chromosomal increase for cells shifted for 1 hr at 42° with ori replication initiation arising from the derepressed cryptic prophage in Y836 cells without the plasmid that were shifted for 1 hr to 42°. (See text for discussion of survivor CFU at 42°.) (D) The influence of host recombination defects on DNA synthesis initiated from ori.

 

Phages:

nin5 (HAYES and HAYES 1986) was used for introducing the nin5 deletion into the imm434 phages. imm434 is imm434 cI as described in HAYES et al. (1998) (lysate 668a). imm434 nin5 was made by crossing nin5 x bio275 imm434 nin5. bio275 imm434 nin5 was prepared by crossing nin5 x bio275 imm434 cIBG[Ts]. bio275 imm434 cIBG[Ts] was prepared by crossing imm434 cIBG[Ts] (HAYES et al. 1998) x bio275. (Thus, all of the imm434 phages, except imm434 cI, were cIBG[Ts] and formed clear plaques at 37°.) cI[Ts]857 d2 (stock MMS1892) and cI[Ts]857 d5 (stock MMS1891) were from F. W. Stahl. imm434 d2 was prepared by crossing bio275 imm434 nin5 x cI[Ts]857 d2. imm434 d5 was prepared by crossing bio275 imm434 nin5 x cI[Ts]857 d5. bio275 imm434 d2 was prepared by crossing bio275 imm434 nin5 x imm434 d2. bio275 imm434 d5 was prepared by crossing bio275 imm434 nin5 x imm434 d5. The phages employed for marker rescue assays are diagrammed in Figure 1C. The nusA strain distinguishes phages with nin5, which are able to form plaques on it, from phages that cannot, i.e., wild type or phages with d2 or d5 (F. W. STAHL, personal communication). In the experiments reported here, the int-xis-hin-exo-bet-gam-kil gene interval (representing bases 27,731–33,303; DANIELS et al. 1983) on some imm434 infecting phages, or within the cryptic imm prophage in strain Y836, was replaced ("int-red-gam") with bio275 E. coli chromosomal DNA present on specialized transducing phage bio275. Fec^– phages with bio275 gene replacements do not form plaques on recA hosts. Phage imm434 (KAISER and JACOB 1957) carries a substitution of nonhomologous phage 434 DNA for a 2.66-kb imm DNA region (35,584–38,245 bp ) present in phage and in the cryptic prophage (DANIELS et al. 1983). Each imm region encodes the respective immunity-specific genes cI and cro and the promoters p_L and p_R, except the imm also includes genes rexA-rexB (1.28 kb not present in imm434). Phages designated NinR carry the nin5 deletion of DNA bases 40,503–43,307, including ren-ninA–ninI (DANIELS et al. 1983; HOLLIFIELD et al. 1987); d2 removes DNA bases 40,943–41,810, including the C-terminal half of orf (ninB) and most of ninC from the N-terminal end, and d5 removes DNA bases 42,925–43,183, including the C-terminal end of rap (ninG) and most of ninH from the N-terminal end (HOLLIFIELD et al. 1987).

Assay for replication initiation from induced cryptic prophage:

Single-colony isolates of strain Y836 and 13 variants were inoculated into tryptone broth (TB; 10 g Bacto tryptone, 5 g NaCl/liter) and grown overnight to saturation. Duplicate 20-ml subcultutures made from 1/100 culture dilutions were prepared and grown in a shaking water bath at 30° to midlog (A₅₇₅ 0.35) in fresh TB. Cell aliquots were diluted in buffer (0.01 M NaCl and 0.01 M Tris HCL, pH 7.8) and spread on TB agar plates incubated at 30° and 42° for 48 hr to determine cell titer and cell viability (frequency of survivor clones) upon induction of the cryptic prophage. One of the duplicate subcultures at midlog was induced for expression of the genes on the cryptic prophage by shaking the culture for 15 sec at 60° and then placing it in a shaking 42° water bath for an additional 60 min. The parallel 30° culture was grown for an additional 60 min at 30°. At the end of the 60-min growth period, DNA was prepared from the 30° and 42° culture cells using the QIAGEN (Chatsworth, CA) DNAeasy kit that can process 2 x 10⁹ cells. A culture volume equivalent to 2 x 10⁹ cells was processed per DNA sample per filter. Duplicate DNA preparations were prepared for each culture sample at 30° and 42°. The duplicate DNA samples were combined, an aliquot was diluted 1/10 in TE* buffer (0.01 M Tris HCl, pH 7.6 or 8.0, 0.001 M Na₂EDTA), and the DNA concentration was determined by spectrophotometer. Aliquots of the DNA estimated at 6.0 µg were ethanol precipitated and the pellets resuspended in 16 µl TE* for use in Southern blot analysis. A digoxigenin-dUTP (Dig)-labeled DNA probe was prepared by amplifying pCH1 plasmid DNA (300 ng) using PCR primers L22 ( bases 5'-38517–38534) and R24 ( bases 5'-40298–40281) and the Dig Hy Prime DNA labeling detection kit (Roche Applied Science). The DNA probe band (Figure 2B) from the PCR was purified by electrophoresis on a 0.7% agarose gel, extracted using the QIAGEN gel extraction procedure, eluted, and concentrated to 2.0 µg/32 µl. The extracted band was converted to a hybridization probe using the Dig-labeling procedure. The DNA probe concentration was estimated from spot coloration on a filter to be 25 ng/µl. Each DNA preparation (2 µg), from cultures of Y836 and variants, was digested with BstEII at 60° and run on a 0.7% agarose gel, with 1 µg of cI72 DNA digested with BstEII per gel as a control for band size. The gels were processed for blotting to GeneScreen Plus (DuPont, Wilmington, DE) filter paper. The Southern blots were hybridized using the Dig-labeled DNA probe at a concentration of 25 ng/ml of hybridization solution and bands were visualized using antidigoxigenin-AP according to the Dig Hy Prime labeling/detection protocol.

Marker rescue assays for imm recombinants:

The recombination event involved the rescue of a 2.6-kb nonhomologous region of prophage DNA that is flanked by homologous regions of 2 kb shared by both the prophage and the infecting phage. The infections were slightly varied from the procedure in HAYES et al. (1998). Phage of 5 x 10⁸ were mixed with 1 x 10⁸ cells, placed in an air incubator at 39° for 15 min, diluted 0.01 into TB, and shaken 90 min at the indicated infection temperature. The lysate was clarified and plated onto sensitive cells to assay total phage and onto imm434 cells to assay for imm recombinants. All cultures were made from a fresh colony on a TB agar plate (TB + 11 g Bacto agar per liter) or plate containing the antibiotic(s) corresponding to the strain's resistance marker(s). The averaged results and standard errors are for multiple independent assays. The practical cutoff for marker rescue of imm in the crosses was set to 1 x 10^–5, considered a background value even when no imm recombinant PFU were obtained. Recombinant imm phages were detected by plaque formation on W3350 or TC600 host cells lysogenized by imm434-T or transformed by plasmid pRP42. The imm rescue frequency was determined by dividing the titer of the recombinants by the total phage titer for the clarified 90-min lysate obtained on sensitive TC600 cells.

Requirement of phage genes for imm marker rescue:

We examined the phage requirement for rescuing imm from the cryptic prophage in strain Y836 (Figure 1B) using imm434 infecting phages (Figure 1C). Phage int-red-gam NinR (Table 2, line Y836, column 5) essentially failed to support imm marker rescue at 30° or 39°. Thus, phage-encoded activity is needed for gene swapping, and the E. coli recombination functions provided in the absence of the phage activities do not suffice. We term the phage-dependent acquisition of heterologous genes the kleptomania (KM) phenotype, since functions are acquired or replaced without selective necessity. The Int-Red-Gam⁺ functions, provided by imm434 nin5 phage, or the NinR⁺ functions, provided by bio275 imm434, independently supported rescue of imm from the cryptic prophage (Table 2). This shows that encodes two distinct genetic mechanisms for the KM phenotype. Marker rescue of imm was favored by the NinR⁺ functions acting alone and was 117-fold higher at 39° than the basal level observed with int-red-gam NinR phage infection.

NinR activities in Rec⁺ hosts:

The possibility that multiple NinR functions participate in NinR⁺-dependent imm rescue was tested by infections with int-red-gam phages deleted within NinR for orf-ninC or for rap-ninH. Deletion of the NinR region by nin5 reduced imm rescue by 117-fold (Table 2), whereas deletions of orf-ninC or of rap-ninH yielded 5-fold and 12-fold reductions (Table 3), suggesting the participation of the deleted functions, as well as that of unaccounted NinR activities, in imm rescue. The very inhibitory effect of nin5 on marker rescue is probably not explained by its encroachment on the rightward recombination interval, i.e., shortening it 10% (Figure 1, B and C).

View this table: In this window In a new window   TABLE 3 Influence of NinR region mutations d2 and d5 on imm marker rescue

 
Y836 cells containing a pBR322-derived multicopy plasmid expressing rap (pTP914), or a gfp control (pTP915), were infected to determine if Rap made by the plasmid could complement for the functions removed by nin5, d5, d2, and even int-red-gam (Table 4). Adding the Gfp plasmid pTP915 to Y836 cells increased the background level for imm rescue by phage int-red-gam NinR by 4-fold and the Rap plasmid pTP914 increased the background >10-fold [i.e., background 3 in Y836, 12 in Y836(pTP915), and 32 in Y836(pTP914) cells]. The presence of Rap provided by pTP914 did not substitute for all of the functions removed by NinR [i.e., comparing Y836(pTR914) infections by int-red-gam NinR⁺ and int-red-gam NinR phages]; however, imm rescue was complemented 8-fold [i.e., comparing int-red-gam rap-ninH infections of Y836(pTP914) and Y836(pTP915)]. Clearly, more than one NinR activity must be removed to incapacitate NinR-dependent imm rescue, and Rap expression alone is insufficient to complement fully nin5 phages for the loss of NinR⁺-dependent imm marker rescue. These findings further support a proposal that the NinR region encodes several functions that aid imm rescue in an otherwise Rec⁺ host.

View this table: In this window In a new window   TABLE 4 Rap participation and imm marker rescue

 

Host and phage functions influencing phage-prophage marker rescue at 39°:

The influence of host recombination functions on the KM phenotype was examined. Do the mechanisms for imm rescue involve the same key host recombination activities? We show that, in the absence of Red activity, the NinR-dependent imm rescue activity: (i) was reduced by >300-fold and 88-fold, respectively, by mutations in the host genes recA and recJ; (ii) was reduced 7- to 10-fold by inactivation of the host genes recB or recD and by the modification (mutation grpD55) of dnaB; (iii) was reduced 2- to 3-fold by inactivation of recF and recQ. We also show that (iv) a recF defect can partially suppress the loss of recJ; (v) a recF defect suppresses and stimulates NinR-dependent recombination in the absence of recD, whereas each separate mutation was inhibitory; and (vi) as noted above, imm rescue moderately, or partially, required NinR activities encoded within rap-ninH and orf-ninC. In contrast, in absence of NinR activity, the Red⁺ system activity: (i) appeared independent of defects in the host genes recA and recF; (ii) was stimulated 4-fold by defects in the host genes recB, recD, recJ, and by recF recJ double mutations; (iii) was stimulated 7-fold by defects in recQ and by recD recF double mutations; and (iv) was reduced (2-fold) by the grpD55 allele of dnaB.

The distinct requirements for the host functions strongly suggest that individual Red-dependent or NinR-dependent mechanisms support imm rescue. The RecBCD complex activity was inhibitory to the Red-dependent activity, since imm rescue was significantly stimulated in Red⁺NinR phage infections of hosts defective for recB or recD (Table 2). In contrast, it was stimulatory to NinR recombination (recB^– and recD^– mutations were inhibitory). Similarly, RecQ activity stimulates NinR-dependent rescue (recQ^– is inhibitory) and inhibits Red-dependent rescue (recQ^– is stimulatory). RecJ powerfully stimulated NinR-dependent imm rescue (recJ^– was strongly inhibitory) but inhibited Red-dependent rescue (recJ^– is stimulatory). RecF appears to have a slight stimulatory effect on NinR-dependent imm rescue, but no effect on Red-dependent imm rescue. The inhibitory effects of RecBCD and RecJ on Red-dependent imm rescue appear to be independent since the increased level of imm rescue was the same in recJ, recJ recF, recD, and recB mutants, also suggesting that the inhibitory effect of RecJ was independent of Recf. The inhibitory effect of RecBCD on Red-mediated imm rescue increased in the absence of Recf.

NinR activities in Rec^– hosts:

The influence of host mutations on imm marker rescue by int-red-gam phages deleted for part of the NinR region was examined (Table 3). We observed two unanticipated results: (1) The deletion of orf-ninC or of rap-ninH suppressed the requirement of recJ for NinR⁺-dependent imm rescue by 27- and 9-fold, respectively (i.e., compared to int-red-gam NinR⁺ infection of Y836 recJ, Table 2), and (2) there was a dramatic loss in imm rescue in a recD recF host for infections by int-red-gam d2 or d5 compared to a int-red-gam NinR⁺ infection (Table 2), and the levels were even lower than those for the int-red-gam NinR infection. We noted earlier that RecF was not required for Red activity, but in the absence of RecD, proved somewhat inhibitory to both Red- and NinR-dependent imm rescue. Was the level of RecF inhibition magnified by removal of the orf-ninC or rap-ninH genes within NinR? (d2 and d5 do not reduce the size of the recombining interval as both lie to the right of the interval between imm and 431; see Figure 1.)

Prophage replication initiation from ori stimulates imm rescue by infecting imm434 phage:

prophage replication initiation and gene expression are repressed by the prophage CI857 repressor in strain Y836 and its variants for cells grown at 30°. However, upon shifting the cells to 38°, the repressor is thermally denatured and transcription initiates from the promoters p_L and p_R, the early genes N and cro-cII-O-P are expressed, and bidirectional replication is initiated from ori within gene O (Figure 2A). HAYES (1979) observed an increase of 5–6 DNA copies by 20 min after shifting cells with a nonexcising cryptic prophage cI857Q-Jb from 30° to 42°. Subsequently, C. HAYES and S. HAYES (unpublished results) found that ori-dependent replication forks induced in Y836 escaped bidirectionally into the contiguous E. coli chromosome, e.g., replicating the adjacent gal operon 15–20 kb left of attL, but stalled <200 kb from ori.

The initiation of ori replication in strain Y836 from the int-red-gam-kil prophage is shown by comparing DNA amplification of the 3675-bp BstEII fragment (Figure 2C), including ori, from thermally induced and noninduced cultures. Replication from ori escapes rightward as seen by amplification of the 4250-bp BstEII fragment (HAYES et al. 1990), which is mainly E. coli DNA. The addition of a multicopy pCI⁺ plasmid (7298-bp band, Figure 2C), expressing wild-type CI⁺ repressor, prevented ori replication initiation, i.e., amplification of the 3675-bp prophage fragment.

The initiation of divergent replication forks from the derepressed cryptic prophage in Y836 was accompanied by cell killing (HAYES et al. 1990), with the frequency of mutant survivor cells appearing at <10^–5. Figure 2, C and D (separate assays and gels), shows survivor frequencies of Y836 CFU at 1 or 2 x 10^–6, respectively. The addition of the pCI⁺ plasmid to Y836 cells prevented the loss in cell viability seen when the cells were shifted to 39° (Table 5) or 42° (Figure 2C) due to CI⁺ inhibition of prophage induction.

The imm rescue from Y836 cells infected at 39° by imm434 Red⁺ or NinR⁺ phages was seven- to ninefold higher than that for the parallel infections carried out at 30°, where the prophage genes are repressed (Table 2, 30° infection results in footnote "c"). We sought to determine if the increased frequency for imm rescue was temperature specific or if it was influenced by transcription and ori replication initiation from the induced cryptic prophage. In Table 5 we show that (i) the plasmid loss from Y836[pCI⁺] cells grown at 39° was <0.5% (i.e., <1/220 colonies) per experiment, greatly limiting the possibility for imm rescue from cells with induced prophages that had lost pCI⁺; (ii) the frequency of imm recombinants formed at 39°, where both the Red⁺ and NinR⁺ recombination functions are expressed from imm434 (which is insensitive to the CI repressor), was 212-fold below that for the infection of Y836 cells (without plasmid) by imm434 at 39°; and (iii) Y836[pCI⁺] cells were 7-fold more inhibitory for imm rescue at 39° than wereY836 cells infected at 30° (Table 5 and footnote c in Table 2). We conclude that ori-dependent replication initiation, or simply the derepression of gene expression, stimulates imm rescue. Relevant to these observations, (i) the formation of imm434 progeny phage particles released from the infected Y836[pCI⁺] cells was not reduced by pCI⁺, and (ii) viable imm recombinants were not formed upon infecting W3350[pCI⁺] cells (data not shown). (Note that while pCI⁺ includes a small portion of DNA within sieB left of imm and the stretch O[C-terminal half]-P-ren right of imm, it is deleted for the gene sequences sieB-N-p_L-rexB-rexA and cro-cII-O (Figure 2B) so that any potential imm recombinants rescued by imm434 infection would be defective for growth.)

Replication initiation at ori requires an interaction between the P gene product and the host DnaB helicase (MALLORY et al. 1990). BULL and HAYES (1996) transduced the mutation grpD55 from the original isolate (SAITO and UCHIDA 1978) into W3350 and mapped it to dnaB. The grpD55 allele in the original isolate and in the allele moved into W3350 were sequenced and found to have identical dual missense mutations within dnaB (M. HORBAY and S. HAYES, unpublished results). forms plaques (at high efficiency) on E. coli W3350 dnaBgrpD55 cells at 30° but not at the restrictive temperature of 39° (BULL and HAYES 1996), suggesting that the P:DnaB interaction required for ori replication breaks down at 39°. However, the dnaBgrpD55 allele does not significantly interfere with E. coli replication, since the cells can form a lawn on plates incubated at 39°–42°. The influence of dnaBgrpD55 on ori replication initiation in Y836 was examined. The dnaB allele strongly prevented ori replication initiation from the 3675-bp prophage fragment and suppressed cell killing by the induced prophage (Figure 2C).

The formation of imm recombinants from infections of Y836 dnaBgrpD55 cells at 39° was reduced by nine-, eight-, and twofold in Red⁺ NinR⁺, int-red-gam NinR⁺, and Red⁺ NinR phage infections (Table 2). We conclude that NinR-dependent imm marker rescue is notably stimulated by replication initiation from the cryptic prophage fragment, whereas Red-dependent imm rescue is much less dependent upon this event, suggesting again that encodes two mechanistically distinct recombination activities.

Studies being reported elsewhere (M. HORBAY, C. HAYES and S. HAYES, unpublished results) are relevant to these observations. We asked if the absence of plaque formation on W3350 dnaBgrpD55 cells at 42° reflected the absence of a phage burst. We quantified plaque formation, free phage, and phage bursts from cells infected at MOI's of 0.01 and 5 at 30° and 42° with different ori phages and with the oriP22 hybrid phage cI857(18,12)P22, which is insensitive to the influence of the dnaBgrpD55 allele. Recombination-driven phage replication was able to bypass the formal ori replication initiation step required for generating phage progeny if two homologous phage genomes entered a cell (M. HORBAY, C. HAYES and S. HAYES, unpublished results), whereas the low MOI infections yielded no burst at the restrictive temperature, in agreement with the plaque assay results, but strong burst at 30°.

Formation of imm recombinants:

We observed a >1000-fold difference in the frequencies of imm recombinant phages formed upon varying the availability of phage and host functions (Table 2). At the low end, int-red-gam NinR⁺ and NinR phages produced no imm recombinants upon infection of the recA host. Was this because imm exchange occurred but the formation of a mature imm particle required phage replication, because exchange between the infecting phage and prophage could not occur, or because there was an interdependence of replication and recombination? (We will discuss phage maturation without replication, i.e., the "free-loader" concept, below.) HAYES (1979)(see FURTH and WICKNER 1983) surveyed the requirement of host replication proteins for ori replication initiation. We are unaware of a similar assay for the influence of recombination proteins on replication initiation from ori. Since imm marker rescue was diminished in the absence of ori replication initiation from prophage (i.e., by CI blocking prophage induction or the conditional defect of the dnaBgrpD55 allele), it was important to learn if host recombination functions influenced the appearance of replication initiation from ori.

We assayed ori replication from variants of Y836 with null mutations in host recombination genes (Figure 2D). In comparison to results seen for the Rec⁺ parent (Figure 2, C and D), the appearance of the onion-skin complex, representing multiple bidirectional replication initiation events (Figure 2A) arising from ori, was reduced in all of the variants, especially those with null alleles in recA, recB, recD, recQ, and recJ. Addressing the above question, ori replication initiation, albeit reduced, proceeded in the recA host. Thus, int-red-gam phages should undergo ori replication initiation in this host, and thus the absence of imm recombinants likely involves a requirement for RecA in phage-prophage exchange or maturation.

The most significant defect in DNA amplification from ori was seen in the Y836 recJ variant. (The strongest of three DNA amplification assays is shown in Figure 2D.) The recJ effect was suppressed partially by an additional null mutation in recF. Another experiment with Y836 recJ, which was identical to that for Y836 dnaBGrpD55 (Figure 2C, i.e., one culture set included pCI⁺ to visualize the ori prophage bands from 30° and 42° cultures with blocked prophage induction), revealed little if any amplification of the ori band (data not shown) from the induced Y836 recJ culture cells. These results suggest that recJ participates in the appearance of the onion-skin DNA complex initiated from ori (Figure 2, A and C). However, as described below, will form a plaque on a recJ host.

NinR and recJ influence plaque size:

The plaque size of imm434 variants on E. coli hosts was measured in parallel at 30°. Plaques (>15/phage/host; 10% standard error) averaged 0.35 mm for int-red-gam NinR⁺ and int-red-gam NinR phages on TC600 Rec⁺, and, respectively, 0.35 and 0.18 mm on Y836 and 0.22 and 0.16 mm on Y836 recJ. Clearly, the ori phages somehow replicate and mature on a host defective for recJ, but the extent of phage progeny released was compromised if we roughly assume that plaque size has some relationship to phage burst. The NinR⁺ activity doubled the plaque size of int-red-gam phage on Y836 and somewhat suppressed for the loss of recJ. Even in the presence of Red activity, the NinR region strongly contributed to plaque size, since in the same assay imm434cI formed 0.90 and 0.53 mm plaques on TC600 and Y836 cells, whereas imm434NinR plaques were 0.44 and 0.18 mm, respectively.

We summarize in Figure 3 the influence of phage and host recombination functions on the rescue of a cryptic prophage imm region by an infecting heteroimmune imm434 phage. Major differences in the imm marker rescue requirements were seen for infecting phages deleted for int-red-gam or for NinR. This suggests the occurrence of two distinct phage-encoded recombination pathways, the Red-dependent path and the NinR-dependent path. The deletion of both paths eliminated imm marker rescue, even in the full presence of wild-type host recombination functions.

View larger version (47K): In this window In a new window Download PPT slide   FIGURE 3.— Summary of recombination factors influencing NinR-dependent and Red-dependent phage-prophage marker rescue. The open rectangles are DNA strands of circular E. coli chromosome; thin lines are DNA, as infecting phage or as cryptic prophage within the E. coli chromosome; the solid rectangle is imm434; the stretched diamond is imm; the feathered tails of arrows and the solid arrowheads respectively represent the 5'- and 3'-ends of phage DNA strands. The potential imm recombination intermediates include: 1, replication D-loop formed within cryptic prophage after replication initiation from ori when the cells are placed at or above 39°, generating intermediate A; 2, exonucleolytic activity at 5'-ends of linear infecting imm434 phage, resulting in 3'-single-stranded DNA overhangs generating intermediate B; 3, a nick is introduced into the circularized monomeric phage genome [or cos end(s) remains unligated], which is converted to a gap by the joint action of a DNA helicase and a single-stranded 5'–3' exonuclease producing intermediate C.

 
To make sense of the many observations, we consider the following: Some mutants have unexpected phenotypes, for which we cannot rule out strain or allele-specific effects. However, in summary, the host recombination activities RecBCD, RecF, RecJ, and RecQ stimulated (participated in) NinR-dependent recombination while they served to reduce (inhibit) Red-mediated imm marker rescue (Figure 3). RecA was essential for NinR, but had slight influence on Red-dependent recombination. RecF was not required for Red activity, but in absence of RecD, proved somewhat inhibitory to both Red- and NinR-dependent imm rescue. The Red-dependent imm rescue was mainly unaffected by host mutations in recA, recF, or a double mutant in recA recD, but was stimulated by host mutations in recB, recD, recJ, and recQ and double mutations in recF recJ and recD recF, whose products may compete for recombination intermediates generated by Red pathway functions. Thus, several of the host functions had opposite effects toward NinR- and Red-dependent imm rescue, e.g., as noted, RecBCD stimulated NinR but inhibited the Red-dependent recombination. Because of these opposite effects, the combination of both Red⁺ and NinR⁺ recombination activities expressed by a imm434 infecting phage (i.e., the sum of independent Red⁺- and NinR⁺-dependent frequencies) was not additive for imm rescue in the Rec⁺ Y836 host. Due to the inhibitory effect of several of the host functions on Red and their stimulatory effect on NinR-dependent recombination, we suggest that most of the imm rescue seen with the Red⁺ NinR⁺ infections of Rec⁺ Y836 cells is NinR⁺ dependent; i.e., NinR⁺ is the dominant pathway in a Rec⁺ host.

It was suggested that the Red functions Beta and Exo (double-stranded 5'-exonuclease) share functionality with RecA and RecJ (CORRETTE-BENNETT and LOVETT 1995). While RecA mediates recombination via single-strand invasion, Exo and Beta generate recombinants by a process called single-strand annealing (see COURT et al. 2002), possibly analogous to double-strand break repair in yeast. Our results showing a fourfold stimulation of Red-dependent marker rescue (by NinR phages) in the absence of recJ, which encodes a single-stranded 5' exonuclease (LOVETT and KOLODNER 1989), agree with observations by MURPHY (1998) and with the finding that RecJ participates in Red-mediated x recombination in the absence of Rap activity (TARKOWSKI et al. 2002).

The requirements for phage-prophage marker rescue have gone mainly unexplored since the studies by ECHOLS and GINGERY (1968) and SIGNER and WEIL (1968), who suggested that 's Int or Red functions support marker rescue recombination in a RecA^– host. The results reported here agree and show that the Red-dependent pathway is capable of eliciting substantial imm rescue in the absence of RecA activity. However, the investigators in the 1960s suggested that the host Rec⁺ functions could provide for marker rescue recombination in the absence of Red functions. This was unsupported by our results showing virtually no imm phage-prophage marker rescue in the absence of the Red or the NinR functions encoded by phage . We explain the latter discrepancy by assuming that the marker rescue infection assays carried out in the absence of Red functions actually provided unrecognized NinR⁺ functionality within the infecting phages employed, which was attributed to the host Rec⁺ functions.

HOLLIFIELD et al. (1987) identified the NinR function rap as an essential requirement for phage-plasmid co-integration at a site of shared DNA homology (Figure 1A) using infecting phages deleted for int-gam. The Rap protein was identified as a Ser/Thr phosphatase (VOEGTLI et al. 2000) and as a Holliday junction resolvase (SHARPLES et al. 2004). We show that the nin5 deletion of NinR functions on a int-red-gam infecting phage was complemented partially by expressing rap from a plasmid within the infected cells. The deletion of rap-ninH on a int-red-gam infecting phage inhibited imm recombination 9-fold, while similarly providing Rap from a plasmid within the infected cells increased imm rescue. The partial complementation of both nin5 and rap-ninH deletions by Rap expressed from a plasmid indirectly implies that Rap is important for imm rescue, but it is unlikely to be the only participating NinR function. An analogous infection with a orf-ninC phage similarly decreased imm rescue (5-fold). Thus, the orf and/or ninC product contributes, along with Rap (and possibly NinH), to NinR-pathway-dependent recombination. Of relevance to this study, STAHL et al. (1995) reported a very modest (2.5-fold) effect of Rap on the frequency of apparent patch (substitution) recombinant formation between a plasmid and a phage whereas the production of co-integrate (apparent single splice or addition) products was stimulated 17-fold.

The formation of NinR-dependent imm recombinants was far lower at 30° than at 39°. Two procedures that inhibited replication initiation from the cryptic prophage, i.e., the addition to cells of a CI⁺ plasmid or of the grpD55 allele of dnaB (both shown to prevent ori replication initiation), both reduced imm marker rescue by 212- and 9-fold, respectively, at 39°. Thus, imm marker rescue is enhanced by ori replication initiation from the cryptic prophage. The products of genes O-P participate in loading the host dnaB product at ori (LEARN et al. 1997), contributing eventually to the open DNA complex represented by intermediate A in Figure 3, which likely is subjected to replicative inhibition (see HAYES and HAYES 1986) by CI due to a requirement for transcriptional activation of ori (DOVE et al. 1969; FURTH and WICKNER 1983). An enhancement of imm recombination due to the formation of intermediate A could involve DnaB driving branch migration (KAPLAN and O'DONNELL 2002). The greater inhibition of imm rescue in Y836[pCI⁺] than in Y836 dnaBgrpD55 cells at 39° or in Y836 cells (with CI857 repressor) at 30° may result from excess CI⁺ severely inhibiting transcriptional derepression.

The amplification level of onion-skin replication was not a critical requirement for NinR-dependent imm marker rescue, although its formation clearly proved of equal importance to the requirements for recB and recD. The accumulation of the onion-skin product of ori replication initiation was greatly reduced in the recJ host and was eliminated in the dnaBgrpD55 strain. Yet the recJ mutation was significantly more inhibitory to NinR-dependent imm marker rescue than was the altered dnaB allele. Moderate ori replication accumulated as an onion-skin intermediate in strains defective for recA, recB, and recD, yet there was an absolute requirement of RecA and an intermediate requirement of RecB and RecD for NinR-dependent imm rescue. recQ encodes a helicase that translocates in the 3' 5' direction when bound to single-stranded DNA (UMEZU et al. 1990). Amplification of the onion-skin product of ori replication initiation was minimal in recQ mutants, but there was less than a threefold reduction (observed in 30 separate assays) in the level of NinR-dependent imm rescue.

The induction of ori-dependent replication from a cryptic prophage kills the host cell, probably by several mechanisms (DATTA et al. 2005). Rare survivor cells include mutants with large deletions removing part of or the entire fragment (HAYES et al. 1990; HAYES 1991). The endpoints in 435 such deletions arising from cells with an induced cI857431 prophage were mapped and none of the deletion endpoints was at or near attL (HAYES 1991). Thus, the attractive possibility that the int product efficiently promotes unequal crossing over between att and secondary sites within the prophage or bacterial chromosome to generate deletions was unsupported. The int-cIII prophage segment in the cI857431 prophage was substituted in strain Y836 to reduce further the possibility for concerted illegitimate deletion of the prophage fragment. The appearance of imm recombinants from imm434 nin5 infections was ascribed to Red-dependent recombination activities, but also includes the potential for int-dependent attB x attP site-specific recombination driving marker rescue; however, we provide no further evidence for or against this possibility. We note that ENQUIST and SKALKA (1973) found no evidence for Int-mediated recombination between circular monomers to produce packagable concatemers; i.e., no more phages were produced on infection of recA^– cells with int⁺red^– gam^– point mutants than with bio11 int^– red^– gam^– phage. We consider unlikely the possibility that a significant fraction of imm marker rescue is explained by illegitimate excision of the replicating cryptic prophage out of the chromosome, i.e., enabling the infecting to recombine with a prophage circle rather than with the chromosome. The excision of a fragment (circular or not) from a replicating cryptic prophage imbedded within the chromosome was not observed here by gel blotting studies that identified the single-copy 3675-bp ori fragment excised by BstEII from a noninduced prophage.

Whether promoted by Red or by NinR, mechanistically, the detection of imm marker rescue from a chromosomal cryptic prophage, lacking cos, requires gene substitution and replacement of a functionally equivalent (although nonhomologous) DNA module in the imm434 infecting phage plus packaging of the recombinant genome to form a mature phage particle. Models for DNA substitution involve two independent splice events flanking imm or a single invasion event extended by branch migration, leaving a large patch of unpaired DNA that is ultimately resolved into distinct daughter copies by replication. The quite different requirements for Red- vs. NinR-promoted imm rescue suggest independent mechanisms.

The formation of mature imm recombinant particles "should" also depend upon sufficient replication of the recombinant molecule to produce a molecular intermediate able to serve as a substrate for DNA packaging, i.e., one with two cos sites. Packaged phage genomes are occasionally formed from substrates with less than two cos sites (LITTLE and GOTTESMAN 1971; YARMOLINSKY 1971; ENQUIST and SKALKA 1973). Similarly, imm lysogens grown in heavy medium and then infected with imm434 in light medium yielded some viable phage particles containing DNA (fully heavy repressed prophage) predominantly made before infection (PTASHNE 1965); the thermal induction of a prophage in a lysogenic cell unable to synthesize DNA at the inducing temperature also yielded some infective centers, requiring the formation of a mature virus particle (FANGMAN and FEISS 1969). Several of the experiments described here included an infection of cells on which plaque formation was virtually abolished: (1) int-red-gam phage plated onto recA-defective hosts and (2) ori phage plated onto dnaB grpD55 host cells at restrictive temperature. In (1), the absence of plaque formation is explained by the "classical" Red-Rec bypass model (see SMITH 1983; AMUNDSEN et al. 1986). Linear multimers (concatemers) formed via late stage rolling-circle replication are destroyed by the ExoV activity of the RecBCD complex in the absence of Gam, and a dimer is not formed by recombination between two monomeric circles in the absence of RecA, resulting in the absence of a molecular substrate with two cos sites for genome packaging. recD null mutants (of recA⁺ cells) have a hyperrecombination phenotype (KUZMINOV 1999). The introduction of a recD^– mutation into a recA^– host restores the ability of Red^– phage to form a plaque on a recA host, somehow suppressing the destruction of linear concatemers by ExoV (AMUNDSEN et al. 1986). Linear multimers are produced by plasmids in recBC sbcBC (ExoV- and ExoI-defective) cells and are formed by the rolling-circle type of plasmid DNA replication dependent upon the RecF pathway genes recA (BIEK and COHEN 1986), recF, recJ (COHEN and CLARK 1986; SILBERSTEIN and COHEN 1987), recO, and recQ (KUSANO et al. 1989). Indeed, the link between accelerated concatemer formation and plasmid instability led to the identification of recD mutants (BIEK and COHEN 1986).

STAHL et al. (1997) suggested that DNA replicated in recA mutant cells is a good substrate for annealing, providing DNA ends, presumably as tips of rolling circles. We showed that defects in host recombination genes modulated the accumulation of the onion-skin replication product, suggesting that the recombination proteins are required for either ori replication or the stability of the accumulated DNA, a topic requiring further analysis. (For example, the product from the double recD recA or recD recF mutants is greater than that from the three single mutations.) We suggest that the onion-skin product is a target for host nucleases and that some of the DNA copies of this region are broken and degraded. Presumably, the broken ends arising from (ori) replication could stimulate recombination by serving as a substrate for strand invasion or by invading the replicon of the infecting phage. This may help explain why imm marker rescue increased when the cryptic prophage was able to initiate ori replication at 39° and reduced when ori replication was blocked. No imm recombinants were detected from int-red-gam imm434 NinR⁺ infections of recA or recA recD hosts, even though ori replication should occur from prophage and infecting phage in both hosts, and in the recA recD host, linear concatemers should be stabilized. The absence of imm recombinants for the int-red-gam imm434 NinR⁺ infection of a recA recD host provides strong support for the requirement of RecA for NinR-dependent marker rescue, since neither replication nor the potential to form stable linear concatemers for genome packaging was limiting.

In plating situation (2) above, the absence of plaque formation by an ori phage on dnaBgrpD55 host cells at restrictive temperature is explained by an inability of the infecting phage to initiate replication, as observed. Nevertheless, when the ori infecting phages were NinR⁺ or Red⁺, mature imm ori recombinants arose in the lysates of the infected dnaBgrpD55 cells. We offer two observations that demonstrate that these results are not without precedent:

1. Biophysical studies: MCMILIN and RUSSO (1972)(p. 55) reported, "under conditions which block DNA duplication, unduplicated DNA can mature, including molecules which have recombined in the host." STAHL et al. (1972) extended this observation and coined the term "free-loader" phage to describe phage produced under replication-blocked conditions, whose synthesis depended upon bacterial and phage recombination systems.
   
2. Infections at MOI >1: SCLAFANI and WECHSLER (1981) showed that at low MOI (0.1) no phage particles were produced in cells lacking a functional dnaB product, yet at high MOI a significant proportion of the cells can produce phage. M. HORBAY, C. HAYES and S. HAYES (unpublished results) found that ori replication initiation is needed for a phage burst when the MOI of the infecting phage is <1 (e.g., the situation in a plaque assay with many-fold excess cells to phage). But infections introducing two phage genomes per cell yield bursts of up to 30 phage/infected cell following infections into cells where ori replication is blocked.
   

From each of these four separate observations, we assume that recombination mechanisms can bypass the need for ori replication to produce a mature phage. (MOTAMEDI et al. 1999 and KUZMINOV 1999 discuss postulates for recombination intermediates initiating replication.) The imm bursts, shown here to depend upon Red or NinR phage functions, suggest that phage-prophage recombination will permit the maturation of a phage particle in a host cell where both the prophage and the infecting phage are blocked for ori replication initiation. Similarly, plasmids of diverse origin sometimes generate open circular by-products with single-stranded tails, i.e., linear multimers, in cells defective for ExoV and ExoI (COHEN and CLARK 1986; SILBERSTEIN and COHEN 1987). The dependence of plasmid linear multimer formation on the functional recA, recF, and recJ genes was interpreted to reflect integrative recombination between tails and plasmid circles operating as an additional mechanism for tail elongation (see KUSANO et al. 1989). SILBERSTEIN et al. (1990) examined -mediated synthesis of plasmid linear multimers and proposed that (a) concatemer formation involved RecE, RecF, or Red pathway-dependent recombination between DNA double-stranded ends and (b) recombination-dependent priming of DNA synthesis at a 3' OH end of an external, single-stranded DNA that invades a homologous sequence on a circular template. Recombination-independent models for switching from the early -mode to the late -mode of linear multimer replication include DNA synthesis at a 3' OH end of a nicked strand and displacement of the 5'-end or the conversion of ori-dependent bidirectional -replication to unidirectional replication accompanied by a break in the lagging strand downstream from the replication fork remaining active.

CLARK et al. (2001) described seven short boundary sequence intervals, each highly conserved per interval, which divided the metabolic regions of many lambdoid phages into six modules (CAMPBELL and BOTSTEIN 1982). They proposed that the boundary sequences were "foci for genetic recombination" in lambdoid phages and served to assort the modules and to increase genetic mosaicism. BLASTP search results (ALTSHUL et al. 1997) for the NinR region reading frames reveal widely shared homologies with other phages and bacterial genomes. We propose that these NinR clusters (KM modules) have a role in stimulating exchanges between the boundary sequences of infecting phages and cellular prophages and thus participate in driving phage evolutionary diversity (HENDRIX et al. 1999; KAMEYAMA et al. 1999; WEISBERG et al. 1999; JUHALA et al. 2000; JOHANSEN et al. 2001; BRüSSOW et al. 2004).

We thank H. J. Bull, A. R. Poteete, S. M. Rosenberg, F. W. Stahl, and M. Horbay for phage, plasmids, and bacterial strains. We are grateful for the participation of J. S. Booth, N. Pastershank, and H. Phillips during preliminary stages of this study; M. Horbay for sharing unpublished results; H. Bull for discussions on recombination models; and R. Slavcev for a BLASTP search of phage NinR genes. The Natural Science and Engineering Research Council of Canada provided grant support.

ADHYA, S., P. CLEARY and A. CAMPBELL, 1968 A deletion analysis of prophage lambda and adjacent regions. Proc. Natl. Acad. Sci. USA 61: 956–962.[Free Full Text]

ALTSCHUL, S. F., T. L. MADDEN, A. A. SCHäFFER, J. ZHANG, Z. ZHANG et al., 1997 Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25: 3389–3402.[Abstract/Free Full Text]

AMUNDSEN, S. K., A. F. TAYLOR, A. M. CHAUDHURY and G. R. SMITH, 1986 recD, the gene for an essential third subunit of exonuclease V. Proc. Natl. Acad. Sci. USA 83: 5558–5562.[Abstract/Free Full Text]

BACKMANN, B. J., 1987 Derivatives and genotypes of some mutant derivatives of Escherichia coli K-12, pp. 1192–1219 in Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology, Vol. 2, edited by F. C. NEIDHARDT, J. I. INGRAHAM, K. B. LOW, B. MAGASANIK, M. SCHAECHTER et al. American Society for Microbiology, Washington, DC.

BIEK, D. P., and S. N. COHEN, 1986 Identification and characterization of recD, a gene affecting plasmid maintenance and recombination in Escherichia coli. J. Bacteriol. 167: 594–603.[Abstract/Free Full Text]

BLATTNER, F. R., B. G. WILLIAMS, A. E. BLECHEL, K. DENNISTON-THOMPSON, H. E. FABER et al., 1977 Charon phages: safer derivatives of bacteriophage lambda for DNA cloning. Science 196: 161–169.[Abstract/Free Full Text]

BRüSSOW, H., C. CANCHAYA and W.-D. HARDT, 2004 Phages and the evolution of bacterial pathogens: from genomic rearrangements to lysogenic conversion. Microbiol. Mol. Biol. Rev. 68: 560–602.[Abstract/Free Full Text]

BULL, H. J., 1995 Bacteriophage lambda replication-coupled processes: genetic elements and regulatory choices. Ph.D. Thesis, University of Saskatchewan, Saskatoon, SK, Canada.

BULL, H. J., and S. HAYES, 1996 The grpD55 locus appears to be an allele of dnaB. Mol. Gen. Genet. 252: 755–760.[Medline]

CAMPBELL, A., and D. BOTSTEIN, 1982 Evolution of the lambdoid phages, pp. 365–380 in Lambda II, edited by R. W. HENDRIX, J. W. ROBERTS, F. W. STAHL and R. A. WEISBERG. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

CHENG, S. C., D. L. COURT and D. I. FRIEDMAN, 1995 Transcription termination signals in the nin region of bacteriophage lambda: identification of rho-dependent termination regions. Genetics 140: 875–887.[Abstract]

CLARK, A. J., W. INWOOD, T. CLOUTIER and T. S. DHILLON, 2001 Nucleotide sequence of coliphage HK620 and the evolution of lambdoid phages. J. Mol. Biol. 311: 657–679.[CrossRef][Medline]

COHEN, A., and A. J. CLARK, 1986 Synthesis of linear plasmid multimers in Escherichia coli K-12. J. Bacteriol. 167: 327–335.[Abstract/Free Full Text]

CONSTANTINO, N., and D. L. COURT, 2003 Enhanced levels of lambda Red-mediated recombinants in mismatch repair mutants. Proc. Natl. Acad. Sci. USA 100: 15748–15753.[Abstract/Free Full Text]

CORRETTE-BENNETT, S., and S. T. LOVETT, 1995 Enhancement of RecA strand-transfer activity by the RecJ exonuclease of Escherichia coli. J. Biol. Chem. 270: 6881–6885.[Abstract/Free Full Text]

COURT, D. L., J. A. SAWITZKE and L. C. THOMASON, 2002 Genetic engineering using homologous recombination. Annu. Rev. Genet. 36: 361–388.[CrossRef][Medline]

DANIELS, D. L., J. L. SCHRODER, W. SZYBALSKI, F. SANGER, A. R. COULSON et al., 1983 Complete annotated lambda sequence, pp. 519–676 in Lambda II, edited by R. W. HENDRIX, J. W. ROBERTS, F. W. STAHL and R. A. WEISBERG. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

DATTA, I., S. BANIK-MAITI, L. ADHIKARI, S. SAU, N. DAS et al., 2005 The mutation that makes Escherichia coli resistant to P gene-mediated host lethality is located within the DNA initiator gene dnaA of the bacterium. J. Biochem. Mol. Biol. 38: 89–96.[Medline]

DOVE, W. F., E. HARGROVE, M. OHASHI, F. HAUGLI and A. GUPTA, 1969 Replicator activation in lambda. Jpn. J. Genet. 44(Suppl.): 11–22.

ECHOLS, H., and R. GINGERY, 1968 Mutants of bacteriophage defective in vegetative genetic recombination. J. Mol. Biol. 34: 239–249.[CrossRef][Medline]

ENQUIST, L. W., and A. SKALKA, 1973 Replication of bacteriophage DNA dependent on the function of host and viral genes. I. Interaction of red, gam, and rec. J. Mol. Biol. 75: 185–212.[CrossRef][Medline]

FANGMAN, W. L., and M. FEISS, 1969 Fate of DNA in a bacterial host defective in DNA synthesis. J. Mol. Biol. 44: 103–116.[CrossRef][Medline]

FURTH, M. E., and S. H. WICKNER, 1983 Lambda DNA replication, pp. 145–173 in Lambda II, edited by R. W. HENDRIX, J. W. ROBERTS, F. W. STAHL and R. A. WEISBERG. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

HARRIS, R. S., S. LONGERICH and S. M. ROSENBERG, 1994 Recombination in adaptive mutation. Science 264: 258–260.[Abstract/Free Full Text]

HAYES, S., 1979 Initiation of coliphage lambda replication, lit, oop RNA synthesis, and effect of gene dosage on transcription from promoters P_L, P_R, and P_R'. Virology 97: 415–438.[CrossRef][Medline]

HAYES, S., 1991 Mapping ethanol induced deletions. Mol. Gen. Genet. 231: 139–149.[CrossRef][Medline]

HAYES, S., and C. HAYES, 1986 Spontaneous mutations suppress inhibition of bacteriophage growth by nonimmune exclusion phenotype of defective prophage. J. Virol. 58: 835–842.[Abstract/Free Full Text]

HAYES, S., D. DUNCAN and C. HAYES, 1990 Alcohol treatment of defective lambda lysogens is deletionogenic. Mol. Gen. Genet. 222: 17–24.[Medline]

HAYES, S., H. J. BULL and J. TULLOCH, 1997 The Rex phenotype of altruistic cell death following infection of a lysogen by T4rII mutants is suppressed by plasmids expressing OOP RNA. Gene 189: 35–42.[CrossRef][Medline]

HAYES, S., C. HAYES, H. J. BULL, L. A. PELCHER and R. A. SLAVCEV, 1998 Acquired mutations in phage genes O or P that enable constitutive expression of a cryptic N⁺cI[Ts]cro^– prophage in E. coli cells shifted from 30°C to 42°C, accompanied by loss of imm and Rex⁺ phenotypes and emergence of a non-immune exclusion state. Gene 223: 115–128.[CrossRef][Medline]

HENDRIX, R. W., M. C. M. SMITH, R. N. BURNS, M. E. FORD and G. F. HATFULL, 1999 Evolutionary relationships among diverse bacteriopages and prophages: all the world's a phage. Proc. Natl. Acad. Sci. USA 96: 2192–2197.[Abstract/Free Full Text]

HOLLIFIELD, W. C., E. N. KAPLAN and H. V. HUANG, 1987 Efficient RecABC-dependent, homologous recombination between coliphage lambda and plasmids requires a phage Nin region gene. Mol. Gen. Genet. 210: 248–255.[CrossRef][Medline]

JOHANSEN, B. K., Y. WASTESON, P. E. GRANUM and S. BRYNESTAD, 2001 Mosaic structure of Shiga- toxin-2-encoding phages isolated from Escherichia coli O157:H7 indicates frequent gene exchange between lambdoid phage genomes. Microbiology 147: 1929–1936.[Abstract/Free Full Text]

JUHALA, R. J., M. E. FORD, R. L. DUDA, A. YOULTON, G. F. HATFULL et al., 2000 Genomic sequences of bacteriophages HK97 and HK022: pervasive genetic mosaicism in the lambdoid bacteriophages. J. Mol. Biol. 299: 27–51.[CrossRef][Medline]

KAISER, A. D., and F. JACOB, 1957 Recombination between related temperate bacteriophages and the genetic control of immunity and prophage localization. Virology 4: 509–521.[CrossRef][Medline]

KAMEYAMA, L., L. FERNANDEZ, J. CALDERON, A. ORTIZ-ROJAS and T. A. PATTERSON, 1999 Characterization of wild lambdoid bacteriopages: detection of a wide distribution of phage immunity groups and identification of a nus-dependent, nonlambdoid phage group. Virology 263: 100–111.[CrossRef][Medline]

KAPLAN, D. L., and M. O'DONNELL, 2002 DnaB drives DNA branch migration and dislodges proteins while encircling two DNA strands. Mol. Cell 10: 647–657.[CrossRef][Medline]

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

KOWALCZYKOWSKI, S. C., D. A. DIXON, A. K. EGGLESTON, S. D. LAUDER and W. M. REHRAUER, 1994 Biochemistry of homologous recombination in Escherichia coli. Microbiol. Rev. 58: 401–465.[Abstract/Free Full Text]

KRöGER, M., and G. HOBOM, 1982 A chain of interlinked genes in the Nin region of bacteriophage lambda. Gene 20: 25–38.[CrossRef][Medline]

KUSANO, K., K. NAKAYAMA and H. NAKAYAMA, 1989 Plasmid-mediated lethality and plasmid multimer formation in an Escherichia coli recBC sbcBC mutant. Involvement of RecF recombination pathway genes. J. Mol. Biol. 209: 623–634.[Medline]

KUZMINOV, A., 1999 Recombinational repair of DNA damage in Escherichia coli and bacteriophage . Microbiol. Mol. Biol. Rev. 63: 751–813.[Abstract/Free Full Text]

LEARN, B. A., S.-J. UM, L. HUANG and R. MCMACKEN, 1997 Cryptic single-stranded-DNA binding activities of the phage P and Escherichia coli DnaC replication initiation proteins facilitate the transfer of E. coli DnaB helicase onto DNA. Proc. Natl. Acad. Sci. USA 94: 1154–1159.[Abstract/Free Full Text]

LITTLE, J. W., and M. GOTTESMAN, 1971 Defective lambda particles whose DNA carries only a single cohesive end, pp. 371–394 in The Bacteriophage Lambda, edited by A. D. HERSHEY. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

LLOYD, R. G., and C. BUCKMAN, 1991 Genetic analysis of the recG locus of Escherichia coli K-12 and of its role in recombination and DNA repair. J. Bacteriol. 173: 1004–1011.[Abstract/Free Full Text]

LOVETT, S. T., and R. D. KOLODNER, 1989 Identification and purification of a single-stranded-DNA-specific exonuclease encoded by the recJ gene of Escherichia coli. Proc. Natl. Acad. Sci. USA 86: 2627–2631.[Abstract/Free Full Text]

MADIRAJU, M. V., and A. J. CLARK, 1991 Effect of RecF protein on reactions catalyzed by RecA protein. Nucleic Acids Res. 19: 6295–6300.[Abstract/Free Full Text]

MALLORY, J. B., C. ALFANO and R. MCMACKEN, 1990 Host virus interactions in the initiation of bacteriophage DNA replication. J. Biol. Chem. 265: 13297–13307.[Abstract/Free Full Text]

MCMILIN, K. D., and V. E. A. RUSSO, 1972 Maturation and recombination of bacteriophage lambda DNA molecules in the absence of DNA duplication. J. Mol. Biol. 68: 49–55.[CrossRef][Medline]

MOTAMEDI, M. R., S. K. SZIGETY and S. M. ROSENBERG, 1999 Double-strand-break repair recombination in Escherichia coli: physical evidence for a DNA replication mechanism in vivo. Genes Dev. 13: 2889–2903.[Abstract/Free Full Text]

MURPHY, K. C., 1998 Use of bacteriophage recombination functions to promote gene replacement in Escherichia coli. J. Bacteriol. 180: 2063–2071.[Abstract/Free Full Text]

MURPHY, K. C., K. G. CAMPELLONE and A. R. POTEETE, 2000 PCR-mediated gene replacement in Escherichia coli. Gene 246: 321–330.[CrossRef][Medline]

POTEETE, A. R., 2004 Modulation of DNA repair and recombination by the bacteriophage Orf function in Escherichia coli K-12. J. Bacteriol. 186: 2699–2707.[Abstract/Free Full Text]

POTEETE, A. R., and A. C. FENTON, 2000 Genetic requirements of phage Red-mediated gene replacement in Escherichia coli K-12. J. Bacteriol. 182: 2336–2340.[Abstract/Free Full Text]

POTEETE, A. R., A. C. FENTON and K. C. MURPHY, 1999 Roles of RuvC and RecG in phage Red-mediated recombination. J. Bacteriol. 181: 5402–5408.[Abstract/Free Full Text]

POTEETE, A. R., A. C. FENTON and H. R. WANG, 2002 Recombination-promoting activity of the bacteriophage Rap protein in Escherichia coli K-12. J. Bacteriol. 184: 4626–4629.[Abstract/Free Full Text]

PTASHNE, M., 1965 The detachment and maturation of conserved lambda prophage DNA. J. Mol. Biol. 11: 90–96.[Medline]

SAITO, H., and H. UCHIDA, 1978 Initiation of the DNA replication of bacteriophage lambda in Escherichia coli K12. J. Mol. Biol. 113: 1–25.

SAWITZKE, J. A., and F. W. STAHL, 1992 Phage has an analog of Escherichia coli recO, recR, and recF genes. Genetics 130: 7–16.[Abstract]

SCLAFANI, R. A., and J. A. WECHSLER, 1981 Growth of phages and X174 under P1ban protein control in the absence of host dnaB function. Virology 113: 314–322.[CrossRef][Medline]

SHARPLES, G. J., F. E. BENSON, G. T. ILLING and R. G. LLOYD, 1990 Molecular and functional analysis of the ruv region of Escherichia coli K12 reveals three genes involved in DNA repair and recombination. Mol. Gen. Genet. 221: 219–226.[Medline]

SHARPLES, G. J., L. M. CORBETT and I. R. GRAHAM, 1998 Rap protein is a structure-specific endonuclease involved in phage recombination. Proc. Natl. Acad. Sci. USA 95: 13507–13512.[Abstract/Free Full Text]

SHARPLES, G. J., L. M. CORBETT and P. MCGLYNN, 1999 DNA structure specificity of Rap endonuclease. Nucleic Acids Res. 27: 4121–4127.[Abstract/Free Full Text]

SHARPLES, G. J., F. A. CURTIS, P. MCGLYNN and E. L. BOLT, 2004 Holliday junction binding and resolution by the Rap structure-specific endonuclease of phage . J. Mol. Biol. 340: 739–751.[CrossRef][Medline]

SHEN, P., and H. V. HUANG, 1986 Homologous recombination in Escherichia coli dependence on substrate length and homology. Genetics 112: 441–457.[Abstract/Free Full Text]

SIGNER, E. R., and J. WEIL, 1968 Recombination in bacteriophage . I. Mutants deficient in general recombination. J. Mol. Biol. 34: 261–271.[CrossRef][Medline]

SILBERSTEIN, S., and A. COHEN, 1987 Synthesis of linear multimers of oriC and pBR322 derivatives in Escherichia coli K-12: role of recombination and replication functions. J. Bacteriol. 169: 3131–3137.[Abstract/Free Full Text]

SILBERSTEIN, Z., S. MAOR, I. BERGER and A. COHEN, 1990 Lambda Red-mediated synthesis of plasmid linear multimers in Escherichia coli K12. Mol. Gen. Genet. 223: 496–507.[Medline]

SMITH, G., 1983 General recombination, pp. 175–210 in Lambda II, edited by R. W. HENDRIX, J. W. ROBERTS, F. W. STAHL and R. A. WEISBERG. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

SMITH, G. R., 2001 Homologous recombination near and far from DNA breaks: alternative roles and contrasting views. Annu. Rev. Genet. 35: 243–274.[CrossRef][Medline]

STAHL, F. W., 1998 Recombination in phage : one geneticist's historical perspective. Gene 223: 95–102.[CrossRef][Medline]

STAHL, F. W., K. D. MCMILIN, M. M. STAHL, R. E. MALONE, Y. NOZU et al., 1972 A role for recombination in the production of "free-loader" lambda bacteriophage particles. J. Mol. Biol. 68: 57–67.[CrossRef][Medline]

STAHL, F. W., M. M. STAHL, R. E. MALONE and J. M. CRASEMANN, 1980 Directionality and nonreciprocality of chi-stimulated recombination in phage . Genetics 94: 235–248.[Abstract/Free Full Text]

STAHL, F. W., C. E. SHURVINTON, L. C. THOMASON, S. HILL and M. M. STAHL, 1995 On the clustered exchanges of the RecBCD pathway operating on phage . Genetics 139: 1107–1121.[Abstract]

STAHL, M. M., L. THOMASON, S. R. POTEETE, T. TARKOWSKI, A. KUZMINOV et al., 1997 Annealing vs. invasion in phage recombination. Genetics 147: 961–977.[Abstract]

STEVENS, W., S. ADHYA and W. SZYBALSKI, 1971 Origin and bidirectional orientation of DNA replication in coliphage lambda, pp. 515–533 in The Bacteriophage Lambda, edited by A. D. HERSHEY. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

TARKOWSKI, T. A., D. MOONEY, L. C. THOMASON and F. W. STAHL, 2002 Gene products encoded in the ninR region of phage participate in Red-mediated recombination. Genes Cells 7: 351–363.[Abstract]

TOOTHMAN, P., and I. HERSKOWITZ, 1980 Rex-dependent exclusion of lambdoid phages II. Determinants of sensitivity to exclusion. Virology 102: 147–160.[CrossRef][Medline]

UMEZU, K., K. NAKAYAMA and H. NAKAYAMA, 1990 Escherichia coli RecQ protein is a DNA helicase. Proc. Natl. Acad. Sci. USA 87: 5363–5367.[Abstract/Free Full Text]

VOEGTLI, W. C., D. J. WHITE, N. J. REITER, F. RUSNAK and A. C. ROSENZWEIG, 2000 Structure of the bacteriophage Ser/Thr protein phosphatase with sulfate ion bound in two coordination modes. Biochemistry 39: 15365–15374.[CrossRef][Medline]

WEISBERG, R. A., M. E. GOTTESMAN, R. W. HENDRIX and J. W. LITTLE, 1999 Family values in the age of genomics: comparative analysis of temperate bacteriophage HK022. Annu. Rev. Genet. 33: 565–602.[CrossRef][Medline]

YARMOLINSKY, M. B., 1971 Making and joining DNA ends, pp. 97–111 in The Bacteriophage Lambda, edited by A. D. HERSHEY. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

YU, D., H. M. ELLIS, E-C. LEE, N. A. JENKINS, N. G. COPELAND et al., 2000 An efficient recombination system for chromosome engineering in Escherichia coli. Proc. Natl. Acad. Sci. USA 97: 5978–5983.[Abstract/Free Full Text]

Communicating editor: S. LOVETT

International Access Link

Online ISSN: 1943-2361  Print ISSN: 0016-6731
Copyright 2009 by the Genetics Society of America
phone: 412-268-1812   fax: 412-268-1813   email: genetics-gsa{at}andrew.cmu.edu