Molecular phylogeny of the genus Pseudomonas based on rpoB sequences and application for the identification of isolates.

Phylogenetic relationships within the genus Pseudomonas were examined by comparing partial (about 1000 nucleotides) rpoB gene sequences. A total of 186 strains belonging to 75 species of Pseudomonas sensu stricto and related species were studied. The phylogenetic resolution of the rpoB tree was approximately three times higher than that of the rrs tree. Ribogroups published earlier correlated well with rpoB sequence clusters. The rpoB sequence database generated by this study was used for identification. A total of 89 isolates (79.5%) were identified to a named species, while 16 isolates (14.3%) corresponded to unnamed species, and 7 isolates (6.2%) had uncertain affiliation. rpoB sequencing is now being used for routine identification of Pseudomonas isolates in our laboratory.

Factors affecting reproducibility of random amplified polymorphic DNA fingerprinting.

The reproducibility of random amplified polymorphic DNA (RAPD) was tested using two different thermal cyclers and three brands of Taq DNA polymerase. Three different oligonucleotides were used to obtain patterns of amplified fragments from three DNA samples (Escherichia coli, Bacillus subtilis, and Thermococcus littoralis). Experiments were repeated three to six times. Apart from the expected between-oligonucleotide and between-DNA variations, between-thermal cycler and between-DNA polymerase variations were observed. Within the DNA-oligonucleotide-DNA-polymerase-thermal cycler, reproducibility was excellent when the thermal cycler equipped with the best temperature regulation was used, but was not as good with another brand of thermal cycler.

Enterobacter cancerogenus (Urosevic, 1966) Dickey and Zumoff 1988, a senior subjective synonym of Enterobacter taylorae Farmer et al. (1985).

Strains labelled Enterobacter cancerogenus (Erwinia cancerogena) and strains labelled Enterobacter taylorae were found to constitute a single DNA-relatedness group (S1 nuclease hybridization method). Furthermore, no phenotypic test among the conventional and nutritional tests performed could differentiate Enterobacter cancerogenus from Enterobacter taylorae. Therefore, Enterobacter cancerogenus (Urosevic, 1966) Dickey and Zumoff, 1968, is a senior subjective synonym for Enterobacter taylorae Farmer et al., 1985.

The genus Methylophaga, a new line of descent within phylogenetic branch gamma of Proteobacteria.

The genus Methylophaga with two species, M. marina and M. thalassica, comprises halophilic methylotrophic bacteria. These organisms utilise C1 compounds through the ribulose monophosphate pathway and are unable to grow on methane. Nearly complete 16S rRNA sequences were obtained for both Methylophaga species by directly sequencing the amplified 16S rRNA gene. These sequences were compared with published 16S rRNA sequences of methylotrophic strains and a large number of marine bacterial strains including several members of the alpha, beta and gamma subclasses of Proteobacteria. Phylogenetic trees were inferred using both parsimony and distance matrix methods. Each topology was analysed by bootstrap. The genus Methylophaga was found to be clearly separated from other methylotrophic bacteria and formed a distinct branch within the gamma subclass of Proteobacteria.

Universal ribotyping method using a chemically labelled oligonucleotide probe mixture.

Some of the present problems in ribotyping are associated with a lack of uniform reactivity of probes when bacterial DNAs are of phylogenetically diverse origins. To overcome these problems, a set of five oligonucleotides (referred to as OligoMix5) was selected to react with conserved sequences located near both extremities of rrs (16S rRNA gene) and near both extremities and the middle of rrl (23S rRNA gene). DNA samples from 13 bacterial species selected to represent various phylogenetic branches within the Eubacteria were cleaved by a restriction endonuclease and electrophoresed in 0.8% agarose, and the fragments were vacuum-transferred to nylon membranes and hybridized with digoxigenin-labelled OligoMix5, plasmid DNA from pKK3535 (cloned rrn operon from Escherichia coli) or pBA2 (cloned rrs from Bacillus subtilis), or acetylamino-fluorene-labelled E. coli 16 + 23S rRNA. The results showed OligoMix5 to visualize patterns in DNA from phylogenetically diverse bacteria with comparable intensity. Banding patterns (not band intensity) obtained with OligoMix5 were identical with those obtained with 16 + 23S rRNA or plasmid pKK3535 for each strain studied and represented complete ribotypes. For DNA from Gram-positive bacteria, complete ribotypes were observed after prolonged enzymatic detection of bands when probes were either E. coli 16 + 23S rRNA or pKK3535. Patterns given by plasmid pBA2 were subsets of the complete ribotypes for 9/13 strains. Each oligonucleotide of the OligoMix5 set was used as a probe to determine its contribution to the complete ribotype. The five oligonucleotide probes, used individually, visualized one to four patterns per DNA sample. Use of DNA from Xenorhabdus sp. CIP 105189 cleaved by EcoRI is suggested to control the quality of the oligonucleotide probes composing OligoMix5. Probe OligoMix5 was found to be an essential tool for ribotyping phylogenetically diverse eubacteria.

Identification of Shigella serotypes by restriction of amplified O-antigen gene cluster.

Due to the scarcity of distinctive biochemical reactions for differentiation of Shigella-Escherichia coli, antigenic analysis has long been used for identification and typing of Shigella isolates. Nevertheless, several intra- and interspecific cross-reactions have been reported to disturb serotyping assays. Shigella serotyping is also occasionally affected by the transition from the smooth (S) form to the rough (R) form. Thus, there is a need for the development of novel robust and discriminating methods for Shigella identification and typing. Characteristically, all genes specifically involved in O-antigen synthesis are clustered in E. coli, Shigella, and Salmonella. Published oligonucleotide sequences complementary to JUMPstart and gene gnd, the conserved flanking sequences upstream and downstream of O-antigen gene clusters, were used to amplify the O-antigen gene cluster of representative strains of each Shigella serotype. A unique, amplified fragment was generally observed for each serotype (size ranging from 6 kbp to 17 kbp). Clearly identifiable and reproducible patterns were obtained for each serotype after MboII digestion of the products, except for S. boydii 12 which showed two distinct patterns, and S. flexneri serotypes 1 to 5 and X and Y which showed a single pattern. A database was built with the Taxotron package allowing automated identification of clinical Shigella isolates to all known serotypes.

Clonal diversity of Vibrio cholerae O1 evidenced by rRNA gene restriction patterns.

The rRNA gene restriction patterns of 89 Vibrio cholerae O1 isolates from different geographic origins were studied. The probe was Escherichia coli 16 + 23S rRNA labelled with "ECL Gene detection system". A total of 17 rRNA gene restriction patterns were observed after BglI cleavage. Four patterns (B1 to B4) were only given by biotype cholerae (14 strains studied). Thirteen patterns (B5 to B17) were only given by biotype El Tor (75 strains studied). There was no correlation between serotypes and rRNA gene restriction patterns. This study provides arguments that (1) strains of biotypes cholerae and El Tor are different clones, (2) a cholera pandemic is not a single world-wide epidemic (due to a single clone) but rather a simultaneous occurrence of several epidemics (several clones involved), and (3) epidemic waves of biotype El Tor could be due to the emergence of new clones.

Characterization of 22 Vibrio species by gas chromatography analysis of their cellular fatty acids.

The cellular fatty acid compositions of 51 Vibrio strains belonging to 22 species as well as five Aeromonas strains were determined by using capillary gas-liquid chromatography (GLC). The major fatty acids were most often hexadecenoic, hexadecanoic and octadecenoic acids. Heptadecenoic acid was present in significant amounts in V. alginolyticus, V. natriegens, V. parahaemolyticus and "Vibrio navarrensis". Twenty fatty acids including branched and hydroxy acids were detected in the genus Vibrio. Quantitative results were treated by principal component analysis to display groups of strains. The first three components (accounting for 69% of the variance) showed the type strains of V. fischeri, V. ordalii, V. damsela, V. mediterranei, V. tubiashii, V. campbellii, V. pelagius, V. gazogenes, and V. nereis to be unclustered. V. alginolyticus (4 strains) and V. parahaemolyticus (4 strains) showed some overlap and the type strain of V. natriegens was in their neighborhood. V. harveyi (4 strains) formed a cluster and V. vulnificus was in its vicinity. V. cholerae (5 strains) overlapped with V. diazotrophicus (3 strains) and was close to the type strain of V. mimicus and V. anguillarum. V. metschnikovii (3 strains) clustered with the type strain of V. cincinnatiensis. A decision tree was devised for the identification of Vibrio species based on qualitative characteristics of fatty acid patterns. However, the following three groups, V. alginolyticus-V. parahaemolyticus-V. natriegens, V. metschnikovii-V. cincinnatiensis and V. cholerae-V. mimicus could not be split into such a decision tree.

Nucleotide sequences of 16S rRNA from ten Serratia species.

Comparison of 16S ribosomal ribonucleic acid (rRNA) sequences has emerged as a powerful tool for bacterial phylogeny. However, earlier studies often only included one or a few species per genus, and it is not sure whether the rRNA sequences could discriminate closely related species. The genus Serratia is composed of ten species, some being up to 60% related by DNA hybridization. The reverse transcriptase/primer extension method was used to determine 1,492 to 1,509 nucleotides in each of ten Serratia 16S rRNA sequences. All rRNA sequences determined were unique. The phylogenetic tree obtained with the neighbour-joining method showed a cluster of Serratia species distinct from both Escherichia coli and Proteus vulgaris. S. fonticola--whose position in the genus Serratia is questioned--was clearly included in the Serratia branch and grouped within the psychrophilic Serratia species. Variable regions in the Serratia rRNA molecules were identified and could serve as the basis for a specific probe design.

Computer identification of Escherichia coli rRNA gene restriction patterns.

A total of 191 strains of Escherichia coli comprising 164 serovar reference strains and 28 clinical strains were characterized by rRNA gene restriction patterns (ribotypes) generated after cleavage of total DNA with MluI, ClaI or HindIII restriction endonucleases and hybridization of fragments with acetylaminofluorene-labelled 16 + 23S rRNA. A wide diversity of ribotypes was observed with endonucleases MluI (104 patterns), ClaI (90 patterns) and HindIII (98 patterns). When MluI was used, 85% of patterns (11 to 15 fragments) shared five fragments 17.09, 3.94, 3.06, 2.23 and 1.76 kb in size. When these fragments were used as internal standards, the percent errors in fragment length determination was half of that obtained with an external standard. Two fragment size databases of MluI and ClaI ribotypes were built. Automatic identification was obtained after setting the percent fragment size variation tolerance (error) at 5%. MluI ribotyping is recommended as a primary epidemiological marker. Strains with similar MluI ribotype should then be submitted to ClaI ribotyping. Ribotyping with HindIII can only be the third choice, since the patterns were often uncertain due to the frequent occurrence of faint bands. Most of the studied serovars gave discrete patterns and these data provide the basis for a molecular typing system for E. coli which could possibly substitute for serotyping when the latter is not available.

Identification by in situ hybridization of segmented filamentous bacteria in the intestine of diarrheic rainbow trout (Oncorhynchus mykiss).

Nonculturable segmented filamentous bacteria (SFB) have been described in the gut of rats, mice and chickens, and 16S rRNA sequences for these organisms are available. These organisms, peripherically related to Clostridium phylogenetic group I, have been provisionally named 'Candidatus Arthromitus'. This work reports the observation of similar bacteria in the intestinal content of the distal intestine, preferentially, in the adult rainbow trout (Oncorhynchus mykiss) that exhibited episodic acute diarrhea, usually during the summer. Abdominal distension, intestinal fluid-mucus content and epithelium detachment were observed in trout. The demonstration that the observed microorganisms are bacteria and belong in the 'Candidatus Arthromitus' group was achieved by in situ hybridization with, respectively, a eubacterial probe and an oligonucleotide probe designed to react specifically with SFB 16S rRNA (encoded by the rrs gene) sequences. The sequenced rrs gene was compared with published sequences and found to be closely related to (although distinct from) other SFB sequences. Implication of these bacteria in trout diarrheic illness remains hypothetical.

Identification of Escherichia coli flagellar types by restriction of the amplified fliC gene.

A total of 182 strains of Escherichia coli (133 reference strains, 22 clinical strains, nine nonmotile strains and 18 strains derived from K-12) were characterized by HhaI restriction of the amplified flagellin gene (fliC). The amplified fliC product was a single band between 0.9 and 2.6 kbp. With the collection of reference strains which represented 48 flagellar types (H-types), a total of 62 patterns (F-types) were observed after HhaI restriction. A single F-type was associated with each of 39 H-types and more than one F-type was associated with the other nine H-types. Antigenically related H-types 12 and 45 gave a single F-type. The determination of HhaI-fliC F-types could allow deduction of all H-types and subdivision of some of these. Application of this identification system to 22 E. coli clinical isolates yielded nine F-patterns and the deduced H-types were confirmed by serotyping in all cases. Nine nonmotile strains were studied and their F-types were also identified. The proposed determination of fliC restriction patterns should be helpful for epidemiological studies.

Azospirillum irakense sp. nov., a nitrogen-fixing bacterium associated with rice roots and rhizosphere soil.

A new species of nitrogen-fixing bacteria, Azospirillum irakense, was found associated with roots and the rhizosphere of rice in the region of Diwaniyah (Qadisya), Iraq. The seven isolates, on which the species description is based, have vibrioid to S-shaped cells with one polar flagellum in liquid medium. Additional lateral flagella are seen on cells grown on nutrient agar. Poly-beta-hydroxybutyrate granules are present in cells. Nitrogen fixation occurs in microaerobic conditions. The phenotypic characters were found to be very close to those of A. amazonense with the following differences: growth occurred in the presence of 3% NaCl, and at pH 5.5 and 8.5, myo-inositol was not utilized as sole source of carbon and energy and pectin was slowly (6 to 9 days) hydrolysed. The seven studied strains formed a DNA-relatedness group distinct from other Azospirillum and Herbaspirillum species. The G + C content of the DNA was 64 to 67 mol %. The type strain is KBC1 (CIP 103311).

DNA fingerprinting of Chlamydia trachomatis by use of ribosomal RNA, oligonucleotide and randomly cloned DNA probes.

DNA fingerprinting of 15 reference strains and 24 clinical isolates of Chlamydia trachomatis, 2 strains of C. psittaci and one strain of C. pneumoniae was studied by use of universal 16 + 23S RNA from Escherichia coli, 16S rDNA-directed oligonucleotide and randomly cloned chlamydial DNA probes. The rRNA-gene restriction patterns (ribotypes) enabled the differentiation of chlamydial species. Following DNA cleavage by restriction endonuclease PvuII, lymphogranuloma venereum and trachoma biovars of C. trachomatis could be differentiated. An oligonucleotide, designed to hybridize the C. trachomatis 16S rDNA, also allowed for both species-specific identification and biovar typing of C. trachomatis human strains. Molecular typing system using 3 lambda clones containing C. trachomatis serotype E random DNA inserts, combined to ribotyping, revealed 12 groups of variable banding patterns within C. trachomatis, and could provide an alternative epidemiological tool.

Phenotypic diversity of anaerobic glycerol dissimilation shown by seven enterobacterial species.

The anaerobic glycerol pathway was studied in seven enterobacterial species selected as representative of different behaviours in terms of anaerobic glycerol dissimilation. The presence of oxidative and reductive pathways of the dha regulon in Klebsiella pneumoniae enabled the cells to grow fermentatively on glycerol. The first two enzymes of the dha regulon (glycerol dehydrogenase type I and dihydroxyacetone kinase) represent the oxidative branch, while the latter two (glycerol dehydratase and 1,3-propanediol dehydrogenase) represent the reductive branch of glycerol fermentation. The slower utilization of glycerol by K. oxytoca was attributed to low production of 1,3-propanediol. K. oxytoca lacked glycerol dehydratase and demonstrated low 1,3-propanediol dehydrogenase activity. K. planticola and K. ozaenae differed from K. pneumoniae and K. oxytoca in lacking the ability to grow on glycerol. K. planticola lacked both enzymes of the reductive branch of glycerol fermentation, and K. ozaenae possessed glycerol dehydrogenase only. K. rhinoscleromatis and Hafnia alvei, like Escherichia coli, did not possess a dha regulon. The glycerol dehydrogenase type II of H. alvei was distinct from that of E. coli. The phenotypic diversity of anaerobic glycerol dissimilation may have taxonomic applications.

Taxonomic diversity of anaerobic glycerol dissimilation in the Enterobacteriaceae.

A total of 1,123 strains representing 128 taxa in the Enterobacteriaceae (named species or subspecies and genomic species) were screened for the presence of glycerol dehydrogenases and 1,3-propanediol dehydrogenase. Only eight taxa, Citrobacter freundii sensu stricto, C. youngae, C. braakii, C. werkmanii, Citrobacter genomospecies 10 and 11, Enterobacter gergoviae and Klebsiella pneumoniae subsp. pneumoniae could grow fermentatively on glycerol and possessed both glycerol dehydrogenase type I (induced by glycerol and dihydroxyacetone) and 1,3-propanediol dehydrogenase which are typical enzymes of the anaerobic glycerol dissimilation pathway. Six other species, C. koseri, E. aerogenes, E. intermedium, K. oxytoca, K. planticola and K. terrigena could not grow fermentatively on glycerol and possessed a glycerol dehydrogenase type I but no 1,3-propanediol dehydrogenase. Other glycerol dehydrogenases types were found: type II (induced by glycerol and hydroxyacetone), type III (induced by glycerol only) and type IV (induced by hydroxyacetone only). They were widely distributed among the Enterobacteriaceae. Classification and identification may take advantage of tests exploring the dissimilation of glycerol.

rRNA gene restriction patterns of Legionella species: a molecular identification system.

A total of 28 species of Legionella could be differentiated by rRNA gene restriction patterns generated after cleavage of total DNA with either EcoRV or HindIII restriction endonucleases, and hybridization of fragments with 32P-labelled Escherichia coli 16 + 23S rRNA. Different species gave different fragment patterns. When several isolates of a species were tested, the patterns obtained were often identical. However, more than one pattern was often observed when more than one serotype was considered. The method should be useful for the identification of all species of Legionella including those exhibiting immunological cross-reactions.

Molecular typing of Salmonella enterica subsp. enterica serovar Wien by rRNA gene restriction patterns.

Analysis of digested DNA from 40 Salmonella enterica subsp. enterica serovar Wien isolates revealed five different rRNA gene restriction patterns for HindIII (H1 to H5) and five for PstI (P1 to P5). The isolates had been selected on the basis of the year of isolation, the geographic origin and the antibiotic resistance pattern and were distinguished as follows: a) 13 pre-epidemic isolates from different French towns in 1958-69 and from Senegal in 1968; b) 7 epidemic isolates from Algiers in 1969; c) 20 post-epidemic isolates from different French and Italian towns in 1970-90. Three different rRNA patterns (H1P1, H3P3 and H4P4) were observed among the pre-epidemic isolates. Conversely, the epidemic isolates, which were characterized by the previously described large antibiotic resistance patterns and by the presence of 1.3 and 80-109 MDa plasmids, belonged to the same H1P1 ribotype. All but two post-epidemic isolates were of the H1P1 ribotype. The determination of rRNA gene restriction patterns together with the plasmid content proved to be useful for a better characterization of the serovar Wien endemic and epidemic isolates.

Computer identification of Shigella species by rRNA gene restriction patterns.

We describe a MluI ribotyping scheme for Shigella which approaches correlation with serotyping. One hundred and seventeen reference strains and previously serotyped clinical isolates representing the 57 Shigella serotypes and biotypes were included in this study. A total of 51 distinct ribotypes were obtained and a database was built with them. The number of bands composing each ribotype varied from 9 to 15. The fragments ranged in size from 1.6 to 18.8 kbp. One hundred and eleven clinical isolates were successfully identified in a double blind study with standard biochemical/serologic methods, by automatic comparison of their ribotypes with our database using the software Taxotron.

Reclassification of "Propionibacterium rubrum" as P. jensenii.

The taxonomic relationship of strains previously designated as "Propionibacterium rubrum" to P. thoenii and P. jensenii was investigated by use of 16S ribosomal RNA sequence comparison, biochemical characteristics and DNA hybridization. A total of 46 strains representing the species P. jensenii and P. thoenii and the former species "P. rubrum" and also including 21 reference strains and 25 strains isolated from dairy sources were studied. The 16S rRNA sequence of strain "P. rubrum" CNRZ 85 (= ATCC 4871) was found to be almost identical to that of the type strain of P. jensenii. DNA hybridization data indicated that "P. rubrum" should belong to the species P. jensenii rather than P. thoenii, as formerly proposed. The "P. rubrum" strains should then be reclassified as a beta-haemolytic biovar of P. jensenii. The genomic species P. jensenii and P. thoenii could be differentiated by biochemical characteristics such as the production of acid from myo-inositol and starch.