Interspecies gene exchange in bacteria: the role of SOS and mismatch repair systems in evolution of species.

Analysis of interspecies matings between S. typhimurium and E. coli indicates that the genetic barrier that separates these (and perhaps many other) related species is primarily recombinational. The structural component of this barrier is genomic sequence divergence. The mismatch repair enzymes act as potent inhibitors of interspecies recombination, whereas the SOS system acts as an inducible positive regulator. Interspecies mating triggers a RecBC-dependent SOS response in female bacteria that increases recombination mainly through overproduction of the RecA protein. Mismatch repair acts to reduce the mutation rate and recombination between similar sequences, whereas SOS acts to increase both. These opposing activities allow mismatch repair and SOS systems to determine both the rate of accumulation of sequence divergence and the extent of genetic isolation, which are the key components of the speciation process.

Structure of recombinants from conjugational crosses between Escherichia coli donor and mismatch-repair deficient Salmonella typhimurium recipients.

To get more insight into the control of homologous recombination between diverged DNA by the Mut proteins of the long-patch mismatch repair system, we have studied interspecies Escherichia coli/Salmonella typhimurium recombination. Knowing that the same recombination pathway (RecABCD) is responsible for intraspecies and interspecies recombination, we have now studied the structure (replacement vs. addition-type or other rearrangement-type recombinants) of 81 interspecies recombinants obtained in conjugational crosses between E. coli donor and mutL, mutS, mutH, mutU or mut+ S. typhimurium recipients. Taking advantage of high interspecies sequence divergence, a physical analysis was performed on one third of the E. coli Hfr genome, which was expected to be transferred to S. typhimurium F- recipients during 40 min before interruption of the mating. Probes specific for each species were hybridized on dot blots of genomic DNA, or on colonies, and the composition of the rrn operons was determined from purified genomic DNA. With very few exceptions, the structure of these interspecies recombinants corresponds to replacements of one continuous block of the recipient genome by the corresponding region of the donor genome.

[Control of homologous recombination by mismatch base repair system in bacteria: implications concerning the chromosome stability and the evolution of species]. / Contrôle de la recombinaison homologue par le système de correction des mésappariements de bases chez les bactéries: implications concernant la stabilité chromosomique et l'évolution des espèces.

The generalized mismatch repair system controls, in bacteria, the homologous recombination between diverged (homologous) DNA. It thus constitute, together with the sequence divergence, a barrier to recombination between bacteria of different species as we have shown for E. coli and S. typhimurium. It is moreover, by preventing the recombination between diverged repeated sequences, a key component of the chromosome stability.

Interspecific recombination between Escherichia coli and Salmonella typhimurium occurs by the RecABCD pathway.

Interspecific recombination in conjugation between Escherichia coli and Salmonella typhimurium is several orders of magnitude lower than intraspecies recombination and is dependent on the RecA function. This low efficiency is due to a 20% divergence in the DNA sequence. The methyl-directed (mut H,L,S dependent) mismatch repair system appears to control the fidelity of homologous recombination; inactivating one of the Mut functions increases the interspecies recombination at least by 10(3)-fold. The interspecific recombination in mutS or mutL mutants is only approximately 10-fold lower than recombination in homospecific crosses as found after correction for the efficiency of mating and DNA transfer by zygotic induction experiments. The interspecific recombination is dependent on the RecABCD pathway: it was abolished in a recA mutant and decreased approximately 10(3)-fold in a recC mutant.

The barrier to recombination between Escherichia coli and Salmonella typhimurium is disrupted in mismatch-repair mutants.

The requirement for DNA sequence homology in generalized genetic recombination is greatly relaxed in bacterial mutL, mutS and mutH mutants deficient in mismatch repair. In such mutants, intergeneric recombination occurs efficiently between Escherichia coli and Salmonella typhimurium, which are approximately 20% divergent in DNA sequence. This finding has implications for speciation, for regulating recombination between diverged repeated sequences, and for hitherto difficult interspecies hybridizations.

Spontaneous and UV-induced mutations in Escherichia coli K-12 strains with altered or absent DNA polymerase I.

The induction of mutations to valine resistance and to rifampin resistance occurs after UV irradiation in bacteria carrying a deletion through the polA gene (delta polA), showing that DNA polymerase I (PolI) is not an essential enzyme for this process. The PolI deletion strain showed a 7- to 10-fold-higher spontaneous mutation frequency than the wild type. The presence in the deletion strain of the 5'----3' exonuclease fragment on an F' episome caused an additional 10-fold increase in spontaneous mutation frequency, resulting in mutation frequencies on the order of 50- to 100-fold greater than wild type. The mutator effect associated with the 5'----3' exonuclease gene fragment together with much of the effect attributable to the polA deletion was blocked in bacteria carrying a umuC mutation. The mutator activity therefore appears to reflect constitutive SOS induction. Excision-proficient polA deletion strains exhibited increased sensitivity to the lethal effect of UV light which was only partially ameliorated by the presence of polA+ on an F' episome. The UV-induced mutation rate to rifampin resistance was marginally lower in delta polA bacteria than in bacteria carrying the polA+ allele. This effect is unlikely to be caused by the existence of a PolI-dependent mutagenic pathway and is probably an indirect effect caused by an alteration in the pattern of excision repair, since it did not occur in excision-deficient (uvrA) bacteria. An excision-deficient polA deletion strain possessed UV sensitivity similar to that of an isogenic strain carrying polA+ on an F' episome, showing that none of the functions of PolI are needed for postreplication repair in the absence of excision repair. Our data provide no evidence for a pathway of UV mutagenesis dependent on PolI, although it remains an open question whether PolI is able to participate when it is present.

Purification and characterization of an inducible Escherichia coli DNA polymerase capable of insertion and bypass at abasic lesions in DNA.

We have investigated the ability of DNA polymerases from SOS-induced and uninduced Escherichia coli to incorporate nucleotides at a well-defined abasic (apurinic/apyrimidinic) DNA template site and to extend these chains from this unpaired 3' terminus. A DNA polymerase activity has been purified from E. coli, deleted for DNA polymerase I, that appears to be induced 7-fold in cells following treatment with nalidixic acid. Induction of this polymerase (designated DNA polymerase X) appears to be part of the SOS response of E. coli since it cannot be induced in strains containing a noncleavable form of the LexA repressor (Ind-). The enzyme is able to incorporate nucleotides efficiently opposite the abasic template lesion and to continue DNA synthesis. Although we observe an approximate 2-fold induction of DNA polymerase III in cells treated with nalidixic acid, several lines of evidence argue that DNA polymerase X is unrelated to DNA polymerase III (pol III). In contrast to pol X, pol III shows almost no detectable ability to incorporate at or extend beyond the abasic site; incorporation efficiency at the abasic lesion is at least 100-fold larger for pol X compared to pol III holoenzyme, pol III core, or pol III* (the polymerase III holoenzyme subassembly lacking the beta subunit). Pol X does not cross-react with polyclonal antibody directed against pol III holoenzyme complex or with monoclonal antibody prepared to the alpha subunit of pol III. Despite these structural and biochemical differences, pol X appears to interact specifically with the beta subunit of the pol III holoenzyme in the presence of single-stranded binding protein. Pol X has a molecular mass of 84 kDa. Our results indicate that this novel activity is likely to be identical to DNA polymerase II of E. coli.

Cloning of rabies virus matrix protein mRNA and determination of its amino acid sequence.

A cDNA clone of mRNA for rabies virus matrix (M) protein has been identified. The clone hybridizes to an mRNA species from rabies virus-infected cells, whose size correlates to the size of the M protein in rabies virions, and selects an mRNA that translates into a polypeptide corresponding in size to M protein. The nucleotide sequence of the cloned cDNA was determined and from this a complete amino acid sequence for M protein was deduced. The deduced sequence of 202 amino acids bears no detectable sequence homology with vesicular stomatitis virus M protein although these proteins may share functional homology.

Partial replicas of UV-irradiated bacteriophage T4 genomes and their role in multiplicity reactivation.

A physicochemical study was made of the replication and transmission of UV-irradiated T4 genomes. The data presented in this paper justify the following conclusions. (i) For both low and high multiplicity of infection there was abundant replication from UV-irradiated parental templates. It exceeded by far the efficiency predicted by the hypothesis that a single lethal hit completely prevents replication of the killed phage DNA: i.e., some dead phage particles must replicate parts of thier DNA. (ii) Replication of the UV-irradiated DNA was repetitive as shown by density reversal experiments. (iii) Newly synthesized progeny DNA originating from UV-irradiated templates appeared as significantly shorter segments of the genomes than progeny DNA produced from non-UV-irradiated templates. A good correlation existed between the number of UV hits and the number of random cuts that would be needed to reduce replication fragments to the length observed. (iv) The contribution of UV-irradiated parental DNA among progeny phage in multiplicity reactivation was disposed in shorter subunits than was the DNA from unirradiated parental phage. It is important to emphasize that it was mainly in the form of replicative hybrid. These conclusions appear to justify excluding interparental recombination as a prerequisite for multiplicity reactivation. They lead directly to some form of partial replica hypothesis for multiplicity reactivation.

Late replication and recombination in the vegetative pool of T4.

The rates and extents of replication are the same for all members of the vegetative pool, whether already residing (progeny) or newly entered (superinfecting). Thus, no member of the pool is sequestered in a replicative complex. Amber N82 infections of nonpermissive host result in extensive breakdown of phage DNA. The extent of fragmentation observed depends on the multiplicity of infection and whether phage ligase is present. Hence, parental DNA suffers single-strand nicks which can be repaired by ligase only if recombination does not interfere. The physiological role of ligase in compensating for such nicks is reemphasized. Superinfecting genomes recombine very rapidly with progeny molecules whose combined lengths are approximately six times that of the superinfecting genomic fragment. The superinfecting phage does not replicate before recombining. Therefore, the lack of replication poses no barrier to efficient recombination.