Conjugational Hyperrecombination Achieved by Derepressing the LexA Regulon, Altering the Properties of RecA Protein and Inactivating Mismatch Repair in Escherichia coli K-12

Corresponding author: Alvin J. Clark, University of Arizona, Tucson, AZ 85721-0106., ajclark{at}email.arizona.edu (E-mail)

Communicating editor: G. R. SMITH

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The frequency of recombinational exchanges (FRE) that disrupt co-inheritance of transferred donor markers in Escherichia coli Hfr by F^- crosses differs by up to a factor of two depending on physiological factors and culture conditions. Under standard conditions we found FRE to be 5.01 ± 0.43 exchanges per 100-min units of DNA length for wild-type strains of the AB1157 line. Using these conditions we showed a cumulative effect of various mutations on FRE. Constitutive SOS expression by lexA gene inactivation (lexA71::Tn5) and recA gene mutation (recA730) showed, respectively, 4- and 7-fold increases of FRE. The double lexA71 recA730 combination gave an 17-fold increase in FRE. Addition of mutS215::Tn10, inactivating the mismatch repair system, to the double lexA recA mutant increased FRE to 26-fold above wild-type FRE. Finally, we showed that another recA mutation produced as much SOS expression as recA730 but increased FRE only 3-fold. We conclude that three factors contribute to normally low FRE under standard conditions: repression of the LexA regulon, the properties of wild-type RecA protein, and a functioning MutSHL mismatch repair system. We discuss mechanisms by which the lexA, recA, and mutS mutations may elevate FRE cumulatively to obtain hyperrecombination.

DURING bacterial conjugation, donor Hfr strains transfer a single strand of DNA with a leading 5'-end into recipients where this strand is converted into a double-stranded (ds) form by discontinuous complementary strand synthesis via Okazaki fragments (for references, see PANSEGRAU and LANKA 1996 ). In transconjugants, the donor DNA can be degraded in recA mutant cells or integrated into the recipient chromosome by recA-dependent homologous recombination. As a rule, integration proceeds at the level of ds donor DNA (OPPENHEIM and RILEY 1967 ). However, when the conversion of donor DNA from single-stranded (ss) to ds form is inhibited, only ssDNA appears to take part in recombination. This unusual process was called ss conjugation (BRESLER et al. 1980 ).

At least three quantitative parameters have been used to characterize conjugational recombination. These are the yield of recombinants, the linkage or co-inheritance of two transferred donor markers, and the heterogeneity of progeny resulting from multiple recombination events proceeding in one transconjugant (LEDERBERG 1957 ; BRESLER et al. 1973 ; LLOYD and BUCKMAN 1995 ). This article concerns the second of these parameters. A review of this subject is available (LANZOV 2002 ).

Linkage is disrupted by recombination between selected and unselected markers. The frequency of this disruptive recombination can be stated as either the average distance between two recombination exchanges, symbolized by , or the frequency of recombination exchanges per unit of distance (FRE). Several attempts have been made to estimate the value of FRE in conjugational recombination. We recalculated values of FRE, from 4.4/100 min up to 10.9/100 min, using a modified Haldane formula and linkage values measured in different crosses (Table 1). (See MATERIALS AND METHODS for the rationale behind our choice of this unit and a description of the formula.)

The reasons for this twofold discrepancy remain obscure. One possible reason is that the strains used by the various investigators carry different mutations affecting exchange frequency. Another reason is that different experimental conditions may affect FRE. Consistent with the former explanation, two genes controlling SOS induction (LITTLE and MOUNT 1982 ) affect FRE. One, recA441 (earlier called tif-1), leads to some derepression of the LexA regulon at 37° and even more at 42°; it increased exchange frequencies (BRESLER and LANZOV 1978 ; LLOYD 1978 ). Another, the noncleavable repressor allele lexA3, significantly reversed the effects of recA441 at 37° (LLOYD 1978 ).

In this study we have sought to answer four questions. First, are there experimental conditions that can provide constant and reproducible findings of FRE values? In other words, we tested various physiological factors to see how much they might influence FRE. Second, does derepression of the LexA regulon reproducibly alter FRE or are other functions of recA and lexA involved? Third, what is the maximal increase of FRE that can be achieved by the use of different mutations that could be involved in recombination? Fourth, are alterations in FRE connected with alterations in the mechanism of recombination?

In the search for the answers to these questions, we show that various physiological factors affect the FRE value by no more than 2-fold. Keeping physiological factors constant we show that on average about five exchanges per 100-min interval occur during conjugational recombination. Furthermore, we show that the number of exchanges can be increased by 17-fold by mutations affecting recA and the state of derepression of the LexA regulon. Finally, we show that a 26-fold increase can be achieved by additional inactivation of mismatch repair. We discuss the mechanisms possibly involved in this hyperrecombination.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Bacterial strains:
The strains used are listed in Table 2. JC18401, JC18411, and JC18414 were made for this study by allowing P1 to multiply on JCD1060 and by using the lysate to transduce, respectively, KL227, EST1450, and EST1799 to Cam^r. malB::Tn9 in JCD1060 contains the gene for Cam^r and is 75% cotransducible with lexA3 (BACHMANN 1990 ). The lexA3 genotype of JC18401, JC18411, and JC18414 was checked by P1 transduction of JC7623 to Cam^R. lexA3 confers recombination deficiency (Rec^-) and UV^s to JC7623 (LOVETT and CLARK 1983 ). The presence of recA1211 in strain JC18414 was confirmed by P1 transduction of JC18410 to Tet^r. This phenotype is conferred by srl-300::Tn10 that is present in JC18414 and is >90% cotransducible with recA (CSONKA and CLARK 1980 ). Tet^r transductants were screened for the coprotease constitutive phenotype associated with recA1211 by looking for clear plaques formed by wild-type phage, indicating that the CI repressor was constitutively degraded (CASTELLAZZI et al. 1972 ).

AB1157S was made by a P1vir transduction of the mutS215::Tn10 allele from JCD1424 into AB1157 strain, checking for both Tet^r and Mut⁺ phenotypes. A higher level of both direct (rifampicin-resistant) and reverse (lac⁺) mutants in the population of AB1157S cells compared to levels in the population of AB1157 cells was used to monitor the mutator phenotype (i.e., Mut⁺).

K197S was made by a consecutive P1vir transduction of, first, the srl⁺ allele from HB101 recA13 into K197 to obtain an intermediate strain K197T where the presence of recA730 and lexA71 alleles was checked by UV^r and Kan^r phenotypes and, second, the mutS215::Tn10 allele into the K197T strain, checking for both Tet^r and Mut⁺ phenotypes as mentioned above.

Conjugation and transduction:
Conjugation and P1 transduction were carried out essentially as described by MILLER 1972 . Except where noted, both Hfr and F^- strains were grown at 37° in mineral salts 56/2 medium at pH 7.5 (ADELBERG and BURNS 1960 ) supplemented with required growth factors. Mating was carried out in the same medium at 37° and selection was also carried out on the same medium, hardened by 1.5% agar and prewarmed to 37°, from which growth factors arginine or threonine were omitted. When selecting for Ara⁺ transconjugants, we substituted arabinose for glucose as the carbon source. Incubation of the agar plates was at 37°. Since some strains of the AB1157 line are isoleucine and valine deficient at temperatures >37° (TESSMAN and PETERSON 1985 ), both growth and selective media contained these amino acids in all crosses. If the Hfr donors used were metB mutants, both growth and selective media contained methionine. Conjugation was allowed for 60 min and the mating mixtures were gently shaken during this period. A sample of the mating mixture was diluted 1:100 with 56/2 buffer (no growth factors) and agitated for 40 sec with a vortex mixer to disrupt mating aggregates.

Some crosses were performed to compare recombination frequencies in exponential and stationary-phase cells. Cultures were prepared as follows: An overnight culture was grown with aeration in 56/2 medium in a thermostated room at 37° to reach deep stationary phase. A small portion of the culture was diluted 1:50 and aerated in a water bath shaker at 37° for 3–3.5 hr to reach middle exponential phase (OD₆₀₀ 0.5). The remaining portion was aerated in a water bath shaker at 37° for 3–3.5 hr and was diluted to the same concentration (OD₆₀₀ 0.5) as the exponential culture. Both exponential and stationary cultures were mated with Hfr KL227 exponential culture under the same condition.

Genetic analysis:
After conjugation Arg⁺, Thr⁺, or Ara⁺ transconjugants were selected; contraselection was achieved with streptomycin. The Hfr strains KL227 and KL226 (LOW 1973 ) transfer markers, respectively, as follows: origin, leu⁺, ara⁺, thr⁺, argE⁺ and origin, lacY⁺, proA⁺, leu⁺, ara⁺, thr⁺. About 200–300 unpurified colonies from each selection were inoculated as patches in a regular array on the same selective medium from which they were taken. After overnight incubation the transconjugants were replica plated onto the same selective medium, omitting the other growth factors or substituting arabinose for glucose, to score inheritance of unselected proximal markers. From two to six crosses were performed with each pair of strains and the results were averaged. Co-inheritance of leu⁺ with argE⁺ in crosses with donor KL227 was tallied directly from the selected Arg⁺ transconjugants. Co-inheritance of leu⁺ with ara⁺ and co-inheritance of leu⁺ with thr⁺ in the same crosses were calculated by including unselected with selected transconjugants. For example, some of the selected Arg⁺ transconjugants were Ara⁺. These were included with the selected Ara⁺ transconjugants to calculate the co-inheritance of leu⁺ with ara⁺. The selected Thr⁺ transconjugants that were Ara⁺ were also included. Similarly, the selected Arg⁺ transconjugants that were also Thr⁺ were included with the selected Thr⁺ transconjugants to determine co-inheritance of leu⁺ with thr⁺.

HALDANE 1919 gave a mathematical description of recombination. The Haldane formula relating genetic distance and meiotic recombination frequency is based on two assumptions: the random distribution of recombination exchanges and the conservation of genetic material during recombination. This formula relates the probability of recombination between two markers to relative distance between these markers. Co-inheritance is related exponentially to exchange frequency by the Haldane formula as modified by BRESLER and LANZOV 1978 . We used the following transformation of this formula to calculate the exchange frequency per 100 min (FRE),

where µ is the fractional co-inheritance and l is the distance in transfer minutes between the co-inherited markers. We used the following marker positions from the physical 100-min map of RUDD 1998 to calculate the distances: argE, 89.5; thr, 0.05; ara, 1.47; and leu, 1.75. The latter three are midpoint positions of the thr, ara, and leu operons. Since we did not know the kilobase coordinates of the mutations used, we chose not to use the more accurate-sounding kilobase unit for our FRE values. Also we wanted our calculated values to be in the same units as those in the previous literature on this subject. Note that it is inaccurate to equate the 100-min unit with the length of the Escherichia coli chromosome, even though the conjugational transfer time for the chromosome was measured to be 100 min (BACHMANN et al. 1976 ). We chose 100 min as our unit in order to move the decimal point two places to the right and, by so doing, create numbers that are more intuitive. If desired, our FRE values can be divided by 4.6, the number of megabases in the E. coli K-12 chromosome (BLATTNER et al. 1997 ), to convert them to a megabase unit. Uncertainty in measurement of fractional co-inheritance was determined as a deviation from the average value by making use of the program Excel-97 with formula [= 2*STDEV] and by inputting the co-inheritance values from independent repeats of the experiment (n = 2–5). Thus, the deviation gives us a confidence interval of 95.4%. Uncertainty of FRE was calculated as mean arithmetic deviation from average value of FRE measured for different selective markers.

When co-inheritance of the proximal unselected markers (leu⁺, proA⁺, or lacY⁺) with any of the distal selected markers (argE⁺, thr⁺, or ara⁺) falls below 0.50, the Haldane formula cannot be used to estimate exchange frequency. Such low co-inheritance may mean that the DNA corresponding to the unselected marker is degraded, that sequence-specific recombination occurs, or that some other unknown process affects independent inheritance of selected and unselected markers. When we encounter co-inheritance values <0.5, we say the FRE values cannot be determined.

Notes on the experimental system:
We chose to examine the argE-leu region of the E. coli chromosome because it does not span either the oriC or the ter region and hence we thought it would minimize any influence of the origin and termination of replication. Besides, the Haldane formula can be applied to the region argE-tsx to describe recombination events by the RecBCD pathway (BRESLER et al. 1978 ). We used the same minimal medium for growth, mating, and selection to prevent alteration of nutritional conditions from influencing the exchange frequency.

Measurement of ß-galactosidase synthesis:
This was carried out precisely as described by TESSMAN and PETERSON 1985 . The reproduction of their experimental protocol was necessary to compare data on the current ability of their recA Prt^c mutants to induce ß-galactosidase with data in the original article. ß-Galactosidase-specific activities reported by TESSMAN and PETERSON 1985 for recA⁺, recA1211, and recA1219 strains, respectively, were as follows: 10, 174, and 150. We found similar values: 14, 141, and 148, respectively.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Culture and physiological conditions can affect exchange frequencies twofold:
Thinking that the twofold discrepancy in published exchange frequency values might be due to differences in culture conditions, we tested the effects of temperature, pH, and the type of media. Table 3 shows that for the most part FRE values ranged around five exchanges per 100 min. Only in Table 3A did we find a statistically significant twofold difference caused by changing the temperature at which the strains were grown, mated, and plated from 37° to 44°. In Table 3C and Table 3D, there was also a significant difference between the FRE values obtained with 56/2 vs. Luria broth (LB) growth and mating media (plating pH of 6.5), but the difference was less than twofold. Nonetheless, we concluded that differences in the physiological conditions might explain the twofold discrepancy among exchange frequency values found in the literature. Consequently, we used only one set of conditions to perform the experiments described below (except where noted), namely 56/2 medium at pH 7.5 for growth, mating, and selection at 37° throughout.

Mutations activating constitutive SOS expression can affect exchange frequencies up to 17-fold:
LLOYD 1978 found that at 37° the thermosensitive coprotease allele recA441 increased the exchange frequency in the thr to leu interval approximately threefold over the frequency found with a recA⁺ recipient. Furthermore, he found that addition of the noninducible mutation lexA3 to the recA441 strain reduced the exchange frequency by twofold, almost back to wild-type levels. Table 4 shows that we basically confirmed Lloyd's findings using the constitutive coprotease allele recA1211. However, we found an average sevenfold increase of exchange frequency by looking at two intervals: a relatively long 1.7-min interval (thr–leu) or a relatively short 0.3-min interval (ara–leu). The effect of recA1211 was partially reversed by lexA3.

Additional evidence that LexA regulon derepression stimulates exchange frequencies was found by testing a strain carrying lexA71::Tn5, an inactive repressor allele. Table 5 shows that in a large 13.3-min interval (argE–leu) the effect was large but quantitatively unmeasurable; in a short 0.3-min interval (ara–leu) a 4-fold effect was measured. Interestingly, the combination recA730 lexA71 resulted in a 17-fold increase of FRE. Since recA730 is identical to recA1211 (ROCA and COX 1990 ), we can use the results in Table 4 and Table 5 to assess the additivity of effects of the recA and lexA mutations (see DISCUSSION).

Mutations of RecA protein alter exchange frequency without appreciably affecting expression of the LexA regulon:
To this point we have been concerned with the state of expression of the LexA regulon and its effect on FRE. However, the data in Table 4 and Table 5 show two indications that modifications of RecA protein can affect FRE independently of their effects on expression of the LexA regulon. First, lexA3 does not completely suppress the effects of recA1211 on FRE (Table 4). Since lexA3 is expected to prevent any induced expression of the LexA regulon, a higher FRE in the double mutant compared to wild type implies that RecA1211 protein can intrinsically increase recombination frequencies. Second, recA730 when added to the lexA71 strain increases the FRE value substantially (Table 5). Here we expect that the null lexA71 allele has already completely derepressed the LexA regulon (TESSMAN and PETERSON 1985 ) so that the increase in FRE by adding recA730 implies that RecA730 protein can intrinsically increase recombination frequencies. It bears repeating that recA1211 and recA730 are identical, although independently isolated, mutations (ROCA and COX 1990 ). Therefore, we expect their proteins to be identical. Thus, the two implications are consistent and reinforce each other.

Table 6 shows another indication that the nature of RecA protein can influence recombination frequencies independently of its effect on LexA regulon expression. In this table we look at the effects on FRE of two recA mutants isolated by TESSMAN and PETERSON 1985 as coprotease constitutive mutants derepressing the LexA regulon. These mutants show almost the same degree of regulon derepression (see MATERIALS AND METHODS), yet recA1219 increases FRE only about threefold above the wild-type value while recA1211 increases it about sevenfold. Thus, RecA1219 protein seems intrinsically less competent than RecA1211 protein for increasing recombination frequencies. We interpret these data to mean that both LexA regulon derepression and a recA effect independent of LexA play a role in altering exchange frequency. RecA1211 (RecA730) protein has the primary structure most conducive to high exchange frequencies of any mutant that we tested.

Absence of mismatch repair increases exchange frequency:
Previous experiments showed that mutations of the methyl-directed mismatch correction (MutSHL) system increased intragenic recombination (ZIEG and KUSHNER 1977 ; FEINSTEIN and LOW 1986 ). We wanted to test the possibility that a mutS mutation might also increase intergenic recombination in our system. If it did increase intergenic recombination, then we wanted to see whether that increase would augment the effects produced by lexA and recA alleles.

The data presented in Table 7 show that mutS215 increases the FRE value in the recA⁺lexA⁺ background about sixfold and increases the already high FRE value in the recA730 lexA71::Tn5 background 50%. Thus, the MutSHL system inhibits intergenic recombination at both the low level found in the recA⁺lexA⁺ strain and the elevated level found in the recA730 lexA71::Tn5 strain. Although revertants were more frequent in the populations of the mutS215 strains (data not shown), they did not exceed 0.01% of the selected recombinants (Table 7) and hence made no significant contribution to the co-inheritance frequencies measured.

FENG et al. 1996 have shown that mismatch repair capacity drops as cells enter stationary phase because of depletion of MutS and other proteins of the methyl-directed mismatch repair system. Consistent with our results with the mutS215 mutation we found that stationary-phase cells also have an elevated exchange frequency (Table 8). In the short intervals for which FRE could be calculated the increase was about twofold.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Validity of the Haldane formula for analyzing conjugational recombination:
The Haldane formula was invented to describe recombination between the linear autonomous chromosomes of eukaryotes. In the case of conjugational recombination in E. coli only one of the two interacting DNA elements is autonomous, the circular chromosome, and only one is linear, the nonautonomous exogenote. Research has shown that there is a high frequency of recombination near each of the two ends of the exogenote (LOW 1965 ; PITTARD and WALKER 1967 ; reviewed by SMITH 1991 ). The Haldane formula cannot be used to evaluate the frequency of these events because they are not randomly distributed along the chromosome. WALMSLEY 1969 published a complete theory of conjugational recombination functions and showed that the Haldane formula is a special case of his general formula.

We use the Haldane formula to describe events that occur in the midregion of the exogenote, presumably independent of the end-associated events. We show in this article (see Table 2 Table 3 Table 4 Table 5 Table 6) that, for wild type, the Haldane formula describes a relatively uniform recombination frequency for three different size intervals in the 12-min region of E. coli chromosome between argE and leu (RUDD 1998 ). We interpret the relative uniformity to validate our use of the Haldane formula. Uniformity at a higher FRE is also seen for a recA1219 mutant (Table 6), a recA1211 mutant (Table 4 and Table 6), and a lexA3 mutant (Table 4).

We note that a uniform recombination frequency in the midregion of conjugational exogenotes contrasts with the view of SMITH 1991 , which he based on earlier data from several other authors. These data (Table 9) showed an 80% co-inheritance of proximal markers located from 5 to 21 min from the selected marker. Smith's resulting view was that 80% of exogenotes show no recombination in the midregion and the 20% that do recombine do so in a small region close to the selected marker. We think that this particular view describes a situation in which midregion breaks are rare and consequently FRE is low. This thought is supported by an analysis of the data referenced by SMITH 1991 . Table 9 shows that all but 3 of the 17 FRE values calculated from these data are significantly lower than the FRE values we obtained or referenced in Table 1. This implies significant differences in procedures, conditions, or strains in the experiments referenced by SMITH 1991 compared to those that we used. Such differences could have resulted in the almost ninefold variation in FRE, a generally low frequency of recombination, and an apparent rarity of midregion breaks.

View this table: In this window In a new window   Table 9. Calculation of FRE, the frequency of conjugational recombination events per 100 min, from the data quoted by SMITH 1991 arranged according to increasing values of FRE

If midregion breaks are more frequent and randomly distributed, however, one would expect to obtain a uniform midregion recombination frequency such as we observe. We expect that the frequency of such midregion double-strand breaks and single-strand gaps will vary with experimental conditions, which is the reason we used a standard set of conditions, and with the genotype of parental strains. Therefore, our view is that recombination can occur at any position of a donor exogenote where a double-strand end occurs. Such ends would certainly exist near the origin and distal terminus of the exogenote, but they might also occur in between. In addition single-strand gaps occurring between the origin and distal terminus might initiate a recombinational exchange either directly or by being converted to double-strand ends (KUZMINOV 2001 ).

Consistent with this view BRESLER et al. 1978 found the same values of co-inheritance and, hence, of FRE, when conjugational recombination occurs by either the RecBCD or the RecFOR pathway in wild type and a recBC sbcBC multiple mutant, respectively. Nonetheless, in a strain wild type for recB, recC, and recD LANZOV et al. 1991 found that several combinations of mutations affecting the RecFOR pathway increased co-inheritance of donor markers and hence decreased FRE. This implies that the RecFOR pathway plays a large role in the midregion recombination events.

Summary of results:
One contribution of this work is to show that culture conditions can affect the frequency of recombination between thr and leu when thr is the selected marker and all zygotes that have received the donor's thr⁺ allele have also received the donor's leu⁺ allele. Although the range of effects is small (twofold), we have standardized culture conditions to allow us to isolate the effects of various mutations. Using standard conditions and ara (or thr) as the selected marker and leu as the unselected marker, we have shown that recA, lexA, and mutS alleles increase FRE. We have also shown that the combination of two of the alleles, recA730 and lexA71, leads to a higher recombination frequency than does either of the single alleles. Finally, we have shown that adding mutS215 to the combination of the recA and lexA alleles leads to a still higher recombination frequency. Contributions of each single mutation and combinations of mutations are shown in Table 10.

Ways of achieving an increase in FRE:
There are several ways to explain the effects of recA, lexA, and mutS alleles. First, the mutations might lead to selective enrichment of Leu^- recombinants at the expense of the Leu⁺ recombinants. We think this is unlikely because the recA, lexA, and mutS genes are not known to affect the relative survival of Leu⁺ and Leu^- recombinants while their products are known to play a role in the process of recombination and the related process of replication. Second, the mutations might lead to an increased frequency of integrating short regions of donor DNA. This might result from more breaks in the exogenote DNA as it is being used as a template for lagging-strand DNA synthesis. Finally, the mutations might alter the extent of heteroduplex regions of DNA formed by recombination. In this case either shortening or lengthening the heteroduplex regions could explain an increase in FRE. In the discussion that follows we explore which of these possibilities might explain our results with recA, lexA, and mutS alleles.

Derepression of the LexA regulon increases FRE:
We observed that a LexA defective mutation, lexA71, by itself increased FRE. This could be interpreted as either a direct effect on the RecA-DNA complex produced by the absence of LexA protein or an indirect effect due to the induction of one or more LexA regulon proteins. The direct effect is plausible because wild-type LexA protein can interfere with the binding of double-stranded DNA to the complex of RecA protein and single-stranded DNA (REHRAUER et al. 1996 ). Thus, the absence of LexA protein caused by the Tn5 insertion allele lexA71 may relieve the inhibition and may stimulate more frequent initiation of recombination. This could lead to shorter integrated regions of donor DNA. The indirect effect is also plausible, however, because induction of DinI protein occurs in the absence of LexA (KENYON and WALKER 1980 ). DinI protein disrupts the filament formed by RecA protein bound to single-stranded DNA by reducing the affinity of RecA protein for DNA (YASUDA et al. 1998 ). In the wild-type background this interaction allows reestablishment of LexA regulon repression (VOLOSHIN et al. 2001 ). When the LexA regulon is completely and permanently derepressed by the absence of LexA protein, the continuous presence of DinI will alter the characteristics of the complex between RecA protein and single-stranded DNA. A less-stable complex may reduce the length of heteroduplex DNA formed during recombination, leading to more exchanges between ara and leu, and thus increasing FRE. Alternatively, increased amounts of RecA protein produced in the absence of LexA might increase recombination as for interspecies recombination (MATIC et al. 1995 ) and elevated amounts of RecA protein may assist RuvAB in extending the length of heteroduplex regions produced during recombination (LLOYD and SHARPLES 1993 ).

RecA protein structure affects FRE:
Derepression of the LexA regulon is not alone in increasing FRE because the coprotease-constitutive recA730 mutation increased the FRE values in a strain already fully derepressed because of its repressor-inactivating lexA71 allele (TESSMAN and PETERSON 1985 ). This observation suggests that two mechanisms are at work in the double recA730 lexA71 mutant. One mechanism, due to lexA71, is caused by the absence of LexA protein, the continuous presence of DinI protein, or derepressed amounts of RecA730 protein. The second mechanism is due to the properties of RecA730 protein. This protein is more capable of displacing SSB protein from DNA than is wild-type RecA protein (LAVERY and KOWALCZYKOWSKI 1992 ). Thus, we hypothesize that RecA730 protein interferes with lagging-strand synthesis on the donor DNA template, thereby creating more single-stranded gaps. This also might lead to shorter regions of donor DNA being integrated, although by a mechanism different from that due to the absence of LexA protein.

The effect of RecA730 protein structure is possibly the explanation for the elevated FRE levels in the recA1211 lexA3 strain (Table 3) because recA730 and recA1211 are due to the same mutation (ROCA and COX 1990 ). In this strain substitution of lexA3 for lexA71 has prevented derepression of the LexA regulon because LexA3 protein is poorly cleavable; the rate is 1% that of wild-type protein (LIN and LITTLE 1989 ). However, an alternative explanation is that the presence of poorly cleavable LexA3 protein may inhibit the competition of RecA1211 protein with SSB, thus preventing interference with the replication of the donor DNA. Such inhibition has been shown in vitro using wild-type RecA protein (REHRAUER et al. 1996 ).

Another phenomenon in need of explanation is the phenotype of the recA1219 lexA⁺ strain. The recA1219 mutation derepressed the LexA regulon as much as recA1211 (see MATERIALS AND METHODS) but affected FRE only about one-third as much (Table 6). One possibility is that RecA1219 protein derepresses the regulon by a mechanism different from that employed by RecA1211 (and RecA730) protein. In other words RecA1219 protein may not compete more successfully than RecA⁺ protein with SSB protein for single-stranded DNA. Perhaps the complex made by this protein and single-stranded DNA has a higher affinity for LexA protein or perhaps it is a more active coprotease. Alternatively, the recA1219 mutation may enhance the unwinding activity of RecA protein (CUNNINGHAM et al. 1979 ). At present we know of no work on RecA1219 protein that would help us evaluate these speculations. Nonetheless, it seems likely that the RecA1219 protein is resistant to the effects of DinI protein or that the dinI gene is also mutant in this strain to allow permanent full derepression of the LexA regulon. Thus, there would be no effect of DinI on the RecA1219-DNA complex that might elevate FRE values.

Mismatch correction reduces FRE:
Methyl-directed mismatch repair has a well-established antirecombinational effect in bacterial experiments (RADMAN 1989 ; SHEN and HUANG 1989 ; MODRICH 1991 ). This was described for homeologous recombination proceeding between highly diverged DNA (MATIC et al. 1995 ). In this case the MutSHL system presumably prevents the formation of heteroduplex DNA. In two other cases the MutSHL system also inhibits recombination (ZIEG and KUSHNER 1977 ; FEINSTEIN and LOW 1986 ). These cases involve homologous intragenic recombination, i.e., between two very closely linked mutations in the same gene. In these cases it is unknown if the MutSHL system inhibits the formation of heteroduplex DNA or acts by some other mechanism. In fact, the inhibitory action of MutSHL in intragenic recombination is counterintuitive because our expectation is that the MutSHL system would promote intragenic recombination, not inhibit it. Our reasoning is that two intragenic mismatches are too few to be detected as a homeologous situation so that heteroduplex DNA would be formed covering both mutations. Then the MutSHL system would correct one or the other mismatched base pair to wild type and an apparent recombinant would be formed. Recognizing this, FEINSTEIN and LOW 1986 speculated that the inhibitory effect of the MutSHL system is to cocorrect both mismatched base pairs. Furthermore, they speculated that the absence of the MutSHL system prevents cocorrection, thus preserving the heteroduplex DNA to "undergo more combinations of branch migration, and/or repairs of single mismatches" by an unspecified correction system (FEINSTEIN and LOW 1986 , p. 30). We think it more reasonable to hypothesize that the MutSHL system can inhibit heteroduplex formation with as few as two mismatches and shorten the heteroduplex regions. In our experiments the ara and leu mismatches are much farther apart (perhaps 15 kb) than in the intragenic recombination experiments (<3 kb). Nonetheless, we hypothesize that the MutSHL system interferes with heteroduplex formation in this situation.

We note that there is another possibility. A very short patch (VSP) type of mismatch correction uses MutS and MutL together with Dcm to correct G:T mismatches to G:C in a small set of specific sequences (LIEB 1987 ; LIEB and REHMAT 1995 ). If it happened that either the ara or the leu mutations in our strains were susceptible to this VSP system, then recombinants might be destroyed by correction. For example, heteroduplex DNA covering the ara gene might be corrected to an ara mutant homoduplex, thus converting an incipient Ara⁺ Leu^- recombinant into an undetectable Ara^- Leu^- strain. VSP correction might also convert an Ara⁺ Leu^- recombinant into an Ara⁺ Leu⁺ recombinant. We think this possibility is extremely remote, however, because it is so sequence specific and our results are consistent with a wide range of studies showing the antirecombinogenic effects of the MutSHL system.

Cumulative effect of recA, lexA, and mutS mutations:
The most striking feature of our data is the cumulative effect of the recA730, lexA71, and mutS215 mutations. Together these mutations increase FRE to the point that there appears to be the equivalent of one crossover every 0.75 min or 30 kb approximately. Table 10 shows that adding the FRE values determined for each mutation separately yields a FRE value less than the value we observe for the multiple mutants. This implies that the mechanisms by which each mutation increases FRE interact in some way, but we know too little about the situation to draw a firm conclusion to this effect. In fact, we want to express a caution against overinterpreting the quantitative aspects of our data. As discussed above, each of the mutations we studied may have complex and partially offsetting effects. The mathematical model that we used to calculate the frequencies cannot resolve such complexities. Such a model can calculate only the net frequency of recombination; it cannot distinguish how many recombinants have been lost and how many have been gained by mutational changes affecting the mechanisms of recombination and replication. If the mismatch repair system, derepression of the LexA regulon, and specific alterations in RecA protein structure are involved in preventing recombination by one mechanism, stimulating recombination by a different mechanism and interfering with each other, the FRE values will give us an incomplete picture of the molecular events occurring.

	ACKNOWLEDGMENTS

We are very grateful to the late B. Bachmann and H. Echols, as well as to M. Berlin, A. Blinkova, K. Smith, I. Tessman, and M. Volkert for supplying bacterial strains and to L. Yurchenko for help in strain construction. V.A.L. acknowledges with gratitude all members of A. J. Clark's former research group for hospitality and providing assistance in the work. This work was supported in part by National Institutes of Health grant AI05371. The work carried out in Russia was supported by a Fogarty International Research Collaboration Award (grant 1 R03 TWO1319-01A1) and the Russian Foundation for Basic Research (grant 02-04-48332).

Manuscript received August 14, 2002; Accepted for publication January 3, 2003.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ADELBERG, E. A. and S. M. BURNS, 1960  Genetic variation in the sex factor of Escherichia coli.. J. Bacteriol. 79:321-330.[Free Full Text]

BACHMANN, B. J., 1972  Pedigrees of some mutant strains of Escherichia coli K-12. Bacteriol. Rev. 36:525-557.[Free Full Text]

BACHMANN, B. J., 1990  Linkage map of Escherichia coli K-12. Microbiol. Rev. 54:130-197.[Abstract/Free Full Text]

BACHMANN, B. J., K. B. LOW, and A. L. TAYLOR, 1976  Recalibrated linkage map of Escherichia coli K-12. Bacteriol. Rev. 40:116-167.[Free Full Text]

BIRGE, E. A., 1983  Site-specific recombination following conjugation in Escherichia coli K-12. Mol. Gen. Genet. 192:366-372.[Medline]

BLATTNER, F. R., G. PLUNKETT, III, C. A. BLOCH, N. T. PERNA, and V. BURLAND et al., 1997  The complete genome sequence of Escherichia coli K-12. Science 277:1453-1474.[Abstract/Free Full Text]

BRESLER, S. E. and V. A. LANZOV, 1978  Inducible recombination in Escherichia coli K-12: recombinogenic effect of tif1 mutation. Rep. Acad. Sci. USSR 238:715-717.

BRESLER, S. E., V. A. LANZOV, and L. R. MANUKIAN, 1973  Mechanism of genetic recombination during bacterial conjugation of Escherichia coli K-12. IV. Heterogeneity of progeny of exconjugants. Mol. Gen. Genet. 123:347-353.[Medline]

BRESLER, S. E., S. V. KRIVONOGOV, and V. A. LANZOV, 1978  Scale of the genetic map and genetic control of recombination after conjugation in Escherichia coli K-12. Hot spots of recombination. Mol. Gen. Genet. 166:337-346.[Medline]

BRESLER, S. E., I. Y. GORYSHIN, and V. A. LANZOV, 1980  Single-stranded conjugation in Escherichia coli K-12. Mol. Gen. Genet. 177:519-525.[Medline]

CASTELLAZZI, M., J. GEORGE, and G. BUTTIN, 1972  Prophage induction and cell division in E. coli. I. Further characterization of the thermosensitive mutation tif1 whose expression mimics the effect of UV irradiation. Mol. Gen. Genet. 119:139-152.[Medline]

CSONKA, L. N. and A. J. CLARK, 1980  Construction of an Hfr strain useful for transferring recA mutations between E. coli strains. J. Bacteriol. 143:529-530.[Abstract/Free Full Text]

CUNNINGHAM, R. P., T. SHIBATA, C. DASGUPTA, and C. M. RADDING, 1979  Single strands induce recA protein to unwind duplex DNA for homologous pairing. Nature 281:191-195.[Medline]

DE HAAN, P. G. and C. VERHOEF, 1966  Genetic recombination in Escherichia coli. II. Calculation of incorporation frequency and relative map distance by recombinant analysis. Mutat. Res. 3:111-117.[Medline]

DIMPFL, J. and H. ECHOLS, 1989  Duplication mutation as an SOS response in Escherichia coli: enhanced duplication formation by a constitutively activated RecA. Genetics 123:255-260.[Abstract/Free Full Text]

FEINSTEIN, S. I. and B. K. LOW, 1986  Hyper-recombining recipient strains in bacterial conjugation. Genetics 113:13-33.[Abstract/Free Full Text]

FENG, G., H. C. TSUI, and M. E. WINKLER, 1996  Depletion of the cellular amounts of the MutS and MutH methyl-directed mismatch repair proteins in stationary-phase Escherichia coli K-12 cells. J. Bacteriol. 178:2388-2396.[Abstract/Free Full Text]

HALDANE, J. B. S., 1919  The combination of linkage values and the calculation of distance between the loci of linked factors. J. Genet. 8:299-309.

JACOB, F., and E. L. WOLLMAN, 1961 Sexuality and the Genetics of Bacteria. Academic Press, New York.

KENYON, C. J. and G. C. WALKER, 1980  DNA-damaging agents stimulate gene expression at specific loci in Escherichia coli. Proc. Natl. Acad. Sci. USA 77:2819-2823.[Abstract/Free Full Text]

KUZMINOV, A., 2001  DNA replication meets genetic exchange: chromosomal damage and its repair by homologous recombination. Proc. Natl. Acad. Sci. USA 98:8461-8468.[Abstract/Free Full Text]

LANZOV, V. A., 2002 Hyper-recombination in Escherichia coli with and without SOS response, pp. 21–38 in Recent Research Development in DNA Repair and Mutagenesis, edited by M. RUIZ-RUBIO, E. ALEJANDRE-DURAN and T. ROLDAN-ARJONA. Research Signpost, Kerala, India.

LANZOV, V., I. STEPANOVA, and G. VINOGRADSKAJA, 1991  Genetic control of recombination exchange frequency in Escherichia coli K-12. Biochimie 73:305-312.[Medline]

LAVERY, P. E. and S. C. KOWALCZYKOWSKI, 1992  Biochemical basis of the constitutive repressor cleavage activity of RecA730 protein. A comparison to RecA441 and RecA803 proteins. J. Biol. Chem. 267:20648-20658.[Abstract/Free Full Text]

LEDERBERG, J., 1957  Sibling recombinants in zygote pedigrees of Escherichia coli. Proc. Natl. Acad. Sci. USA 43:1060-1064.[Free Full Text]

LIEB, M., 1987  Bacterial genes mutL, mutS, and dcm participate in repair of mismatches at 5-methylcytosine sites. J. Bacteriol. 169:5241-5246.[Abstract/Free Full Text]

LIEB, M. and S. REHMAT, 1995  Very short patch repair of T:G mismatches in vivo: importance of context and accessory proteins. J. Bacteriol. 177:660-666.[Abstract/Free Full Text]

LIN, L. L. and J. W. LITTLE, 1989  Autodigestion and RecA^- dependent cleavage of Ind^- mutant LexA proteins. J. Mol. Biol. 210:439-452.[Medline]

LITTLE, J. W. and D. W. MOUNT, 1982  The SOS-regulatory system of Escherichia coli.. Cell 29:11-22.[Medline]

LLOYD, R. G., 1978  Hyper-recombination in Escherichia coli K-12 mutants constitutive for protein X synthesis. J. Bacteriol. 134:929-935.[Abstract/Free Full Text]

LLOYD, R. G. and C. BUCKMAN, 1995  Conjugational recombination in Escherichia coli: genetic analysis of recombinant formation in Hfr x F^- crosses. Genetics 139:1123-1148.[Abstract]

LLOYD, R. G. and G. J. SHARPLES, 1993  Processing of recombination intermediates by the RecG and RuvAB proteins of Escherichia coli. Nucleic Acids Res. 21:1719-1725.[Abstract/Free Full Text]

LOVETT, S. T. and A. J. CLARK, 1983  Genetic analysis of regulation of the RecF pathway of recombination in Escherichia coli K-12. J. Bacteriol. 153:1471-1478.[Abstract/Free Full Text]

LOW, B., 1965  Low recombination frequency for markers very near the origin in conjugation in E. coli.. Genet. Res. 6:469-473.[Medline]

LOW, K. B., 1973  Rapid mapping of conditional and auxotrophic mutants of Escherichia coli K-12. J. Bacteriol. 113:798-812.[Abstract/Free Full Text]

MATIC, I., C. RAYSSIGUIER, and M. RADMAN, 1995  Interspecies gene exchange in bacteria: the role of SOS and mismatch repair systems in evolution of species. Cell 80:507-515.[Medline]

MILLER, J. H., 1972 Experiments in Molecular Genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

MODRICH, P., 1991  Mechanisms and biological effects of mismatch repair. Annu. Rev. Genet. 25:229-253.[Medline]

OPPENHEIM, A. B. and M. RILEY, 1967  Covalent union of parental DNAs following conjugation in Escherichia coli.. J. Mol. Biol. 28:503-511.[Medline]

PANSEGRAU, W. and E. LANKA, 1996  Enzymology of DNA transfer by conjugative mechanism. Prog. Nucleic Acid Res. Mol. Biol. 54:197-251.[Medline]

PITTARD, J. and E. M. WALKER, 1967  Conjugation in Escherichia coli: recombination events in terminal regions of transferred deoxyribonucleic acid. J. Bacteriol. 94:1656-1663.[Abstract/Free Full Text]

RADMAN, M., 1989 Mismatch repair and genetic recombination, pp. 169–192 in Genetic Recombination, edited by R. KUCHERLAPATI and G. SMITH. American Society for Microbiology, Washington, DC.

REHRAUER, W. M., P. E. LAVERY, E. L. PALMER, R. N. SINGH, and S. C. KOWALCZYKOWSKI, 1996  Interaction of Escherichia coli RecA protein with LexA repressor. I. LexA repressor cleavage is competitive with binding of a secondary DNA molecule. J. Biol. Chem. 271:23865-23873.[Abstract/Free Full Text]

ROCA, A. I. and M. M. COX, 1990  The RecA protein: structure and function. Crit. Rev. Biochem. Mol. Biol. 25:415-467.[Medline]

RUDD, K. E., 1998  Linkage map of Escherichia coli K-12, edition 10: the physical map. Microbiol. Mol. Biol. Rev. 62:985-1019.[Abstract/Free Full Text]

SHEN, P. and H. V. HUANG, 1989  Effect of base pair mismatches on recombination via the RecBCD pathway. Mol. Gen. Genet. 218:358-360.[Medline]

SMITH, G. R., 1991  Conjugational recombination in E. coli: myths and mechanisms. Cell 64:19-27.[Medline]

TAYLOR, A. L. and E. A. ADELBERG, 1960  Linkage analysis with very high frequency males of Escherichia coli. Genetics 45:1233-1243.[Free Full Text]

TEMPLIN, A., S. R. KUSHNER, and A. J. CLARK, 1972  Genetic analysis of mutations indirectly suppressing recB and recC mutations. Genetics 72:105-115.[Abstract/Free Full Text]

TESSMAN, E. S. and P. PETERSON, 1985  Plaque color method for rapid isolation of novel recA mutants of Escherichia coli K-12: new classes of protease-constitutive recA mutants. J. Bacteriol. 163:677-687.[Abstract/Free Full Text]

VOLOSHIN, O. N., B. E. RAMIREZ, A. BAX, and R. D. CAMERINI-OTERO, 2001  A model for the abrogation of the SOS response by an SOS protein: a negatively charged helix in DinI mimics DNA in its interaction with RecA. Genes Dev. 15:415-427.[Abstract/Free Full Text]

WALMSLEY, R. H., 1969  The general theory of mapping functions for random genetic recombination. Biophys. J. 9:421-431.

YASUDA, T., K. MORIMATSU, T. HORII, T. NAGATA, and H. OHMORI, 1998  Inhibition of Escherichia coli RecA coprotease activities by DinI. EMBO J. 17:3207-3216.[Medline]

ZIEG, J. and S. R. KUSHNER, 1977  Analysis of genetic recombination between two partially deleted lactose operons of Escherichia coli K-12. J. Bacteriol. 131:123-132.[Abstract/Free Full Text]

This article has been cited by other articles:

				D. M. Baitin, I. V. Bakhlanova, D. V. Chervyakova, Y. V. Kil, V. A. Lanzov, and M. M. Cox Two RecA Protein Types That Mediate Different Modes of Hyperrecombination J. Bacteriol., April 15, 2008; 190(8): 3036 - 3045. [Abstract] [Full Text] [PDF]

				D. M. Baitin, I. V. Bakhlanova, Y. V. Kil, M. M. Cox, and V. A. Lanzov Distinguishing Characteristics of Hyperrecombinogenic RecA Protein from Pseudomonas aeruginosa Acting in Escherichia coli. J. Bacteriol., August 1, 2006; 188(16): 5812 - 5820. [Abstract] [Full Text] [PDF]