A taxonomic study of the proteus-providence group with especial reference to the role of plasmids.

(1) A numerical taxonomic survey of the Proteus-Providence area was carried out. One hundred and forty three organisms from both this group and the rest of the Enterobacteriaceae were subjected to 178 morphological, biochemical and physiological tests, and the results were analysed by computer. Proteus-Providence strains clustered into two areas which contained six well-separated groups. These groups corresponded to the six Proteus-Providence species. Area A comprised Prot.mirabilis and Prot.vulgaris, and Area B Prot.morganii, Prot.rettgeri, Prov.stuartii and Prov.alcalifaciens. It is proposed that the classification of the the Proteus-Providence group be re-organised to accommodate these and other findings of other workers. (2) A number of the strains were used to generate a total of 65 new strains by curing of selected markers, and by plasmid infection. High curing frequencies were obtained using the compounds acridine orange, ethidium bromide and chlorpromazine. Using these compounds, it was found that the ability to swarm on solid media could be readily cured from most Prot.vulgaris strains and a few Prot.mirabilis strains, and that the character gelatinase activity could also be eliminated by the use of these agents. Some strains were also cured for antibiotic resistance. (3) A number of different plasmid transfer techniques were attempted, and the poor "recipient ability" of this group noted. However, R plasmid transfer was effected using donor strains from within the study to suitable recipient strains. One strain was also re-infected with the "swarming plasmid". Out of 24 plasmids recieved from Drs. N. Datta and R.W, Hedges (originally from strain within the Proteus-Providence area), five were transferred to suitable recipient organisms to generate a further eight new strains. Transfer of the ability to utilise ethanol as sole carbon source was also demonstrated and the resulting progeny selected for further study. (4) All the new strains were tested for the 178 characters in the numerical taxonomic survey. Various attempts were made to eliminate test differences between parent and new strains due to lack of reproducibility. The test differences were examined; in general cured strains lost characters but plasmid-infected strains both lost and gained characters, indicating either elimination of an incompatibile resident plasmid by the incoming plasmid, or plasmid-mediated membrane changes affecting permeability. Many of the characters in question were the abilities to ferment sugars. (5) Attempts were made to confirm the plasmid-mediated nature of some of the characters, by repeating the tests using different methods, and curing and plasmid transfer experiments. Further evidence was amassed for the existence of some plasmids in some strains which carried genes for swarming, and gelatinase activity. The characteristic production of long forms by swarming cells was shown to be absent in nearly all the non-swarming cured derivatives of these strains. Transfer of gelatinase activity was demonstrated and there was some evidence that this character was linked to casein hydrolysis in some strains, and possibly ampicillin resistance. (6) In the numerical taxonomic survey of both the original strains and the new strainsa large number of the new strains clustered near the parental strains, because either the number of character changes involved were very small, or these characters were not important characteristic features of that particular cluster. A number of strains, however, were markedly affected by removal of, or infection with plasmids. Only one strain was excluded from the Proteus-Providence area as a result of these genetic manipulations. This strain was an atypical Prot.vulgaris and the evidence indicates this organism was originally an intermediate strain of unknown and perhaps previously undescribed identity which acquired characteristics of the Prot.vulgaris cluster by acquiring the "swarming plasmid". There was evidence to suggest that a Prot.mirabilis strain had acquired a plasmid conferring gelatinase activity and other characters which were features of Prot.vulgaris causing the strain to cluster with this species. Removal of the plasmid caused the strain to join the Prot.mirabilis cluster. The remaining strain movements involved strains derived by plasmid infection and these generally resulted in gross changes in identity, within the Proteus-Providence area, (7) The implications of these findings in bacterial taxonomy are discussed. It is concluded that in general the genotype of an organism is conserved by the exclusion of "unwanted" plasmids and the maintainance of stable populations of plasmids, which may, as this work suggests, confer stable diagnostic characters of the host's identity.