The partial nucleotide sequences of 18S ribosomal RNA (positions 1451 to 1618 in Saccharomyces cerevisiae) of 48 strains representing 46 species of ballistosporous yeasts and a related genus were determined by dideoxy sequencing with reverse transcriptase. The sequence data were analyzed by neighbor-joining method and several unrooted phylogenetic trees were drawn. Ballistosporous yeast species were separated into two groups: a group of xylose-containing species and a group of xylose-lacking species. All of the species of Bullera and Kockovaella are included in the former group, and all of the species of Ballistosporomyces, Bensingtonia, Sporobolomyces, and Sporidiobolus are included in the latter group. Species of the genus Bullera are scattered on several branches of the phylogenetic trees and species of Cryptococcus are located at several positions close to those of species in several groups of Bullera. Species of "piricola group" of Bullera, Bullera megalospora, Bullera punicea and Bullera piricola, are considered to be closely related and are far removed from where the remaining species of the genus are located. These three species are characterized by the production of extremely large asymmetrical (bilaterally symmetrical) ballistospores and are considered to be included in a different genus from Bullera. A close relationship of Kockovaella to Fellomyces is suggested. Species of Sporobolomyces are scattered on several branches of the phylogenetic trees. The same phenomenon is observed in the case of Bensingtonia and Ballistosporomyces. These observations clearly suggest that the mode of vegetative reproduction such as the productivity of ballistospores, stalked conidia and budding yeast cells, does not reflect the phylogenetic relationship among basidiomycetous yeasts and is considered to be a less significant taxonomic criterion than hitherto believed.
A numerical taxonomy of 604 strains of the genera Arthrobacter, Aureobacterium, Brevibacterium, Cellulomonas, Clavibacter, Corynebacterium, Curtobacterium, Microbacterium, Nocardia, Nocardioides, Rhodococcus, Terrabacter and Tsukamurella was undertaken based on 280 physiological characters with the aid of miniaturized tests. Clustering was by the unweighted pair group method (UPGMA) with 12 different measures of similarity. Test error, overlap between the phena and cophenetic correlation coefficients for the classifications obtained with the Jaccard coefficient (SJ), the Pearson coefficient (SP), the simple-matching coefficient (SSM), and the Dice coefficient (SD) as measures of similarity were within acceptable limits. Clusters were defined at the 55.0 to 70.5% levels (SJ). The compositions of clusters corresponded largely the delineations of 81.1 to 93.5% (SSM), 29.1 to 55.0% (SP), and 55.3 to 81.1% (SD). A total of 31 major clusters (containing five or more strains), 41 minor clusters and subclusters (containing less than five strains), and 54 single-member clusters were obtained in the UPGMA/SJ study. The following conclusions were reached: (i) A high degree of similarity between the genera Aureobacterium, Cellulomonas, Clavibacter, Curtobacterium and Microbacterium found in phylogenetic-based studies could be shown also phenetically. Strains belonging to these genera were found in SJ clusters 1 to 45, often representing closely related, or single species. (ii) Several strains of the plant pathogenic coryneform bacteria assigned to the genus Clavibacter and Curtobacterium flaccumfaciens were found to be within one SJ cluster, indicating the high similarity between these genera. The current classification of species within the genera Curtobacterium and Clavibacter is unsatisfactory; a close relationship to Microbacterium suggests a reclassification into a redefined genus. Subspecies of Clavibacter and pathovars of Curtobacterium flaccumfaciens should only be retained for practical purposes. (iii) Bacteria belonging to the genus Corynebacterium, including Corynebacterium glutamicum, C. ammoniagenes, C. minutissimum, C. striatum, C. variabilis, C. kutscheri, C. diphtheriae, C. pseudotuberculosis, and ‘C. ulcerans’ formed a separate complex of the distinct, adjacent clusters 47 to 63. The physiologically inactive species C. mycetoides, C. pseudodiphtheriticum, C. xerosis, C. renale and C. pilosum were found in clusters 116 to 126. (iv) Differences between the Arthrobacter globiformis group and the Arthrobacter nicotianae group were reflected in the structure of the phenogram, species differentiation being based on only a few characters. (v) Strains assigned to the four species of the genus Brevibacterium were grouped into two clusters; the taxonomic implications are discussed. (vi) The results of the study are largely in line with a previously published numerical survey and with chemotaxonomic and genetic data. Suggestions for an improved classification for some species is given in addition to an extensive data-base on physiological reactions for differentiation purposes.
A frequency matrix of positive results was constructed for major phena defined in a previous phenetic classification of several species of coryneform and related taxa. A total of 280 physiological characters from 441 strains were taken as basis for this identification matrix. Minimum number of diagnostic characters for the matrix was selected using computer programs for calculation of different separation indices (CHARSEP) and selection of group diagnostic properties (DIACHAR). The resulting matrix consisted of 31 phena versus 58 characters. Phena overlap within the matrix was found to be relatively small (OVERMAT program). For each phenon, the identification scores for the most typical hypothetical organism was satisfactory (MOSTTYP program). The matrix was evaluated theoretically and practically (MATIDEN program). The overall identification rate (442 strains) of the theoretical evaluation (Willcox probability>0.99) was 92.0%. In the practical evaluation (40 strains) a total of 33 strains (82.5%) were identified with a Willcox probability>0.9. For minor clusters and single strains belonging to genetically and chemotaxonomically defined species, additional identification tables are provided. When used in combination with chemotaxonomic methods, the identification matrix can improve the identification of coryneform bacteria.
The plasmid pKYM isolated from the Gram-negative bacterium Shigellasonnei has been shown to replicate via a rolling-circle mechanism (14). We have confirmed the plus origin of pKYM in a region of 173 base-pairs by cloning the fragment in pUC18 plasmid of Escherichia coli (E. coli). A derivative of pUC18 which contains this fragment multiplies normally in an E. coil polA when the Rep protein of pKYM is supplied in trans. The 173 base-pairs of pKYM contains the following characteristic structures indispensable for replication in double-stranded form: 1) the consensus sequence, 5′CTTAaggGATAaaT, of the plus origins of pUB110 plasmid family, 2) three directly repeated sequences of AATGPuC/AG downstream of the consensus sequence and 3) forty-five nucleotides upstream of the consensus sequence. The minus origin was found on a fragment containing nucleotides 281 base-pairs, that was localized upstream of the plus origin. Both the plus and minus origins are necessary for the stable multiplication of pKYM in E. coli. Although the nucleotide sequence of plus origin of pKYM has a strong homology to plus origins of the pUB110 plasmid family (14), the minus origin shares no homology to those of pUB110 plasmid family.