Recombination between Ribosomal Rna Genes

Previous workers have shown that intergeneric crosses between Salmonella typhimurium and Escherichia coli produce a high proportion of merodiploid recombinants among the viable progeny. We have examined the unequal crossover event that was responsible for a number of intergeneric merodiploids. The merodiploids that we studied were all heterozygous for the metB-argH interval and were the products of intergeneric conjugal crosses. We found that when the S. typhimurium donor had its transfer origin closely linked to metB and argH, all recombinants examined were merodiploid, and they generally arose as F-prime factors. Many of these F-prime factors had been created by recombination between flanking rrn genes in the donor. When the S. typhimurium Hfr transfer origin was more distant from the selected markers, quite different results were obtained. (1) Depending on the donor, 19-47% of the recombinants that acquired the donor argH+ or metB+ genes were merodiploid for these loci, but none of the recombinants were F-prime. (2) A majority of the merodiploids had a novel (nonparental) r m gene, indicating that unequal recombination between nonidentical rrn genes was a prevalent mechanism for establishing the merodiploidy. (3) Both tandem and nontandem duplications were found. (4) Some of the merodiploids duplicated E. coli genes in addition to acquiring S. iyphimurium genes. (5) Some merodiploids contained the oriC region from each parent. Of a total of 11 8 intergeneric merodiploids characterized from all donors, 48 different genotypes were observed, and 38 of the 48 had one or more nonparental rrn operons. ERODIPLOIDS are common among the progeny of Salmonella typhimuM rium-Escherichia coli intergeneric crosses (BARON et a l . 1968; SANDERSON 1976). The merodiploidy is maintained through F-prime factors in some of these recombinants, but in others the extra DNA is inserted into the recipient chromosome (JOHNSON et al. 1973, 1975). Despite the potential importance of intergeneric merodiploids in bacterial evolution, little attention has been paid to the illegitimate recombinational events responsible for the insertional type. Considerable information is available, however, on insertion merodiploids arising within E. coli or within S. typhimurium. In E. coli, for example, homologous ' Current address: Departments of Medicine and Endocrinology/BD-13 1, Medical College of Georgia, Augusta, Georgia 309 12. Genetics 1 1 0 365-380 July, 1985 366 A. F. LEHNER AND C. W. HILL recombination between different ribosomal RNA genes (rrn) can generate both tandem (HILL et al. 197’7) and transposed duplications (HILL and HARNISH 1982). In a survey of duplications occurring in S. typhimurium, it was found that loci lying between directly repeated rrn genes were more likely to become duplicated than loci elsewhere on the chromosome (ANDERSON and ROTH 1978). The r m operons seemed ideal targets for the type of unequal crossover required to establish intergeneric merodiploids, considering the degree of intergeneric conservation of both the nucleotide sequence (KOHNE 1968) and the map position of the seven r m operons (ANDERSON and ROTH 1978; LEHNER and HILL 1980; LEHNER, HARVEY and HILL 1984), and this paper addresses their contribution. The map positions and orientations of the rrn operons is shown in Figure 1. Given that there are seven of these operons, two of which are oriented in opposition to the other five, there are a large number of potential interactions that could lead to merodiploidy. Seven distinct mechanistic types that will be of importance in our results are depicted in Figure 2. In each mechanism shown, the novel joint(s) is created by recombination between different rrn operons. The merodiploids produced can be tandem duplications (A, B and C), transposed duplications (D and E) or more complex types (F and G). Of course, inserted merodiploids could be created by illegitimate recombination between sites other than rrn operons, and each mechanism in Figure 2 should have a counterpart that does not use rrn operons for unequal crossover. A common feature of all of the mechanisms of Figure 2 is that at least one of the partners in the creation of each novel joint is donor DNA. Mechanisms can also be imagined where both partners in forming the novel joint are recipient DNA. Three such mechanisms are shown in Figure 3. We will describe a set of merodiploid recombinants whose properties fit all of the types proposed in Figures 2 and 3 except type H. T o obtain intergeneric merodiploids, we conducted a series of crosses using an E. coli Frecipient and a variety of S. typhimurium Hfr donors. The donor markers selected were metB+ or argH+ which occur within the rrA-rmB interval. The progeny were classified as to whether they were haploid or diploid for these selected markers and analyzed as to the number and identity of the rrn operons present. This analysis took advantage of the fact that each of the seven E . coli and each of the seven S , typhimurium rrn operons occur within a distinct BamHIIPstI fragment (LEHNER, HARVEY and HILL 1984). Since the positions of the flanking restriction sites for all operons, except S. typhimurium r m H , have been mapped (BOROS, KISS and VENETIANER 1979; LEHNER, HARVEY and HILL 1984) the sizes of recombinant r m operons can be predicted and used in their identification. Once the electrophoretic profile of r m restriction fragments for a particular recombinant was obtained, we deduced its probable genotype by the following steps: (1) if r m fragments corresponded in size to parental fragments, they were assumed to be parental; (2) the parental rrn complement was considered along with the nonparental rrn band(s) to determine the most likely unequal crossover(s) responsible for generating the recombinant; (3) candidate chroINTERGENERIC MERODIPLOIDS 367 mosomal structures were considered only if they contained at least one representation, either donor or recipient, of all parts of the chromosome as well as a metB-argH duplication; (4 ) nonparental rrn fragment identifications were confirmed by comparing observed sizes with those predicted from the published restriction map of E. coli and S. typhimurium genes (LEHNER, HARVEY and HILL 1984); (5) in addition, the merodiploids were tested for the presence of F-prime factors. MATERIALS AND METHODS Bacterial strains: E. coli mutants used were K-12 derivatives. CH1392 was FsbcB15 endA hsdR4 h s d W metBl zij-116::TnIO argHl thi gal (LEHNER, HARVEY and HILL 1984). CHI396 was derived by cotransducing recA56 with srl-300::TnIO into a trpA23 Pur+ derivative of AB336 (TAYLOR and ADELBERC 1960). S. typhimurium Hfr’s were obtained from the Salmonella Genetic Stock Center, University of Calgary, Calgary, Alberta, Canada, and have been described by SANDERSON et al. (1972). Their origins of transfer are shown in Figure 1. Microbiological techniques: Media and techniques for noninterrupted mating, purification of recombinants and storage of strains have been described (LEHNER, HARVEY and HILL 1984). Sensitivity of a strain to R17 phage was determined by placing a loopful of RI7 on the surface of a soft agar layer containing the bacterial mutant. Segregation frequencies were determined by growing a culture in synthetic medium containing tetracycline but no methionine or arginine, spreading on glucose synthetic plates containing methionine and arginine to yield 150-350 colonies and replica plating to detect Met-, Argor Tet’ segregants. Restriction enzyme digestion and hydridization: Methods for isolation of genomic DNA, restriction endonuclease digestion, gel electrophoresis, transfer to diazotized paper, hybridization to E. coli [SsP]ribosomal RNA and autoradiography have been described previously (LEHNER, HARVEY and HILL 1984). F+ plasmid was prepared from an F+ segregant of KL96 (gift of M. CAPAGE) by CsCIethidium bromide centrifugation following lysozyme-RNase A-Sarkosyl treatment of logarithmically grown cells (HILL et a l . 1977). F plasmid preparations were radiolabeled to 107-108 dpmlpg with [a-”P]dCTP by the procedure of RICBY et a l . (1977).