Genetic diversity and temporal variation in the E. coli population of a human host.

Electrophoretic techniques were employed to study variation in chromosomal genes encoding enzymes and in the distribution of cryptic plasmids in the E. coli population of a human host over an 11-month period. Thirteen of the 15 enzymes studied were polymorphic, and mean genetic diversity per locus was 0.39. Among 550 clones isolated from fecal samples, protein electrophoresis revealed 53 distinct electrophoretic types (ETs). Most ETs appeared on only one or a few days and were considered transients, but two (ET-12 and ET-13) were observed many times over extended periods and represented residents. Complete turnover in the transient ETs in the population occurred in periods of from two weeks to a month. ETs appearing in one month showed no particular genetic similarity to those of the previous month. - All but 4 of the 53 ETs carried one or more "cryptic" plasmids with molecular weights ranging from 1 to 80 megadaltons. With few exceptions, the plasmid composition of each ET was unique. In the course of the 11-month sampling period, there were changes in the plasmid profiles of the resident strains ET-12 and ET-13, and also in the profile of a recurrent strain, ET-2, which was isolated on four days. Modification of the plasmid profile of ET-12 involved the sequential addition of relatively high molecular weight bands. For ET-2 and ET-13, the changes in the plasmid profiles were radical, suggesting invasions of new cell types rather than merely the addition and deletion of plasmids. - The results of this study provide three lines of evidence that recombination plays a minor role in the generation of genetic diversity in the E. coli population of a single host. (1) Several pairs of loci were in strong linkage disequilibrium; compared to a randomly generated array of genotypes, the sample of ETs contained an excess of pairs differing at one or two loci and too many pairs with highly distinctive combinations of electromorphs. (2) In most cases where pairs of ETs differed at a single locus and, therefore, could reasonably have been generated by phage- or plasmid-mobilized gene transfer, the plasmid profiles of the pair members were radically different and/or the potentially transmitted alleles were not present in other ETs in the population. (3) Although ET-12 was abundant, being represented by 252 of the 550 clones sampled, the electrophoretic type most similar to ET-12 different from it at six loci, and ET-12 carried two unique alleles. We conclude that most of the genetic diversity observed in this human host is a consequence of successive invasions of E. coli genotypes.