Search

Each difference between the genome sequences of Escherichia coli B strains REL606 and BL21(DE3) can be interpreted in light of known laboratory manipulations plus a gene conversion between ribosomal RNA operons. Two treatments with 1-methyl-3-nitro-1-nitrosoguanidine in the REL606 lineage produced at least 93 single-base-pair mutations ( approximately 90% GC-to-AT transitions) and 3 single-base-pair GC deletions. Two UV treatments in the BL21(DE3) lineage produced only 4 single-base-pair mutations but 16 large deletions. P1 transductions from K-12 into the two B lineages produced 317 single-base-pair differences and 9 insertions or deletions, reflecting differences between B DNA in BL21(DE3) and integrated restriction fragments of K-12 DNA inherited by REL606. Two sites showed selective enrichment of spontaneous mutations. No unselected spontaneous single-base-pair mutations were evident. The genome sequences revealed that a progenitor of REL606 had been misidentified, explaining initially perplexing differences. Limited sequencing of other B strains defined characteristic properties of B and allowed assembly of the inferred genome of the ancestral B of Delbrück and Luria. Comparison of the B and K-12 genomes shows that more than half of the 3793 proteins of their basic genomes are predicted to be identical, although approximately 310 appear to be functional in either B or K-12 but not in both. The ancestral basic genome appears to have had approximately 4039 coding sequences occupying approximately 4.0 Mbp. Repeated horizontal transfer from diverged Escherichia coli genomes and homologous recombination may explain the observed variable distribution of single-base-pair differences. Fifteen sites are occupied by phage-related elements, but only six by comparable elements at the same site. More than 50 sites are occupied by IS elements in both B and K, 16 in common, and likely founding IS elements are identified. A signature of widespread cryptic phage P4-type mobile elements was identified. Complex deletions (dense clusters of small deletions and substitutions) apparently removed nonessential genes from approximately 30 sites in the basic genomes.

Antecedents of Escherichia coli B have been traced through publications, inferences, and personal communication to a strain from the Institut Pasteur in Paris used by d'Herelle in his studies of bacteriophages as early as 1918 (a strain not in the current collection). This strain appears to have passed from d'Herelle to Bordet in 1920, and from Bordet to at least three other laboratories by 1925. The strain that Gratia received from Bordet was apparently passed to Bronfenbrenner by 1924 and from him to Luria around 1941. Delbrück and Luria published the first paper calling this strain B in 1942. Its choice as the common host for phages T1-T7 by the phage group that developed around Delbrück, Luria, and Hershey in the 1940s led to widespread use of B along with E. coli K-12, chosen about the same time for biochemical and genetic studies by Tatum and Lederberg. Not all currently available strains related to B are descended from the B of Delbrück and Luria; at least three strains with somewhat different characteristics were derived independently by Hershey directly from the Bronfenbrenner strain, and a strain that appears to have passed from Bordet to Wollman is in the current Collection of the Institut Pasteur. The succession of manipulations and strains that led from the B of Delbrück and Luria to REL606 and BL21(DE3) is given, established in part through evidence from their recently determined complete genome sequences.

The effect of the E. coli termination factor rho on middle mode and anti-late T4 transcription was examined. RNA from CAM-treated rho ts-infected cells contains only about 25% of the normal T4 gene 43 RNA levels, despite the fact that the DE content of this RNA is qualitatively normal. Gene 43 RNA is not found during the early period of infection of rho+ bacteria by mot− bacteriophage, but about 20–25% normal levels are found when mot− phage infect rho ts hosts. Thus gene 43 is transcribed in two modes: a major middle mode which uses mot as an initiation cofactor and a minor early mode which seems to need mot expression to help overcome rho action. Apparently, the middle mode RNA from gene 43 is translated much more efficiently than is the early RNA transcribed from this gene. CAM RNA from rho ts infections contains nearly as much anti-late RNA as does normal RNA (without antibiotic) from these cells. Thus anti-late synthesis, which is ordinarily mot dependent, becomes mot independent if there is no rho activity in the hosts cells. These data along with experiments on an in vitro system for middle mode expression ( C. Thermes, P. Daegelen, V. de Franciscis, and E. Brody (1976) Proc. Nat. Acad. Sci. USA 73, 2569–2573) are consistent with a model in which the mot product binds to DNA where it acts either as a melting protein or as a gyrase. This DNA change leads to changed interactions both with rho and RNA polymerase.