BOX-PCR Genomic Fingerprinting Research Articles

Survival of escherichia coli in Water Microcosm Study and Rethinking its Use as Indicator

We investigated the fate of Escherichia coli in natural waters, addressing survival, viable but non-culturable (VBNC) state, changes in phenotype and genomic diversification under laboratory microcosm environment. Five E. coli isolates (three marine, one clinical and one laboratory strain) were inoculated in microcosms of sterile distilled water, as well as, water collected from river, pond, sea, and estuary. A viable count was measured every week for up to 33 weeks. Strains which undergone VBNC state in microcosms were resuscitated. The revived isolates were tested to determine changes in antibiotic susceptibility, thermotolerance and genetic fingerprinting. We found extended persistence of five E. coli strains in water microcosms at room temperature for durations ranging from four to 33 weeks. Even in distilled water, some isolates persisted up to seven weeks; proving E. coli have excellent tenacity to survive. At VBNC state in microcosms, bacteria were small coccoid in shape and formed biofilms in most cases. Resuscitated laboratory strains showed changes in antibiotic susceptibility and thermotolerance compared to the original counterpart. However, the revived marine isolates showed very little to no difference in those tests. Resuscitation media could not revived the pathogenic isolates. The REP- and BOX-PCR genomic fingerprinting also showed some genetic diversification among resuscitated strains during the extended survival and VBNC state. The overall observation supports for the E. coli to survive in natural waters and to enter into VBNC state. It ultimately proves that their presence in water does not necessarily indicate faecal contamination but still beyond acceptance.

Phylogenetic diversity of rhizobia nodulating native Mimosa gymnas grown in a South Brazilian ecotone.

Floristic surveys performed in "Campos Gerais" (Paraná, Brazil), an ecotone of Mata Atlântica and Cerrado biomes, highlights the richness and relative abundance of the family Fabaceae and point out the diversity and endemism of Mimosa spp. Our study reports the genetic diversity of rhizobia isolated from root nodules of native/endemic Mimosa gymnas Barneby in three areas of Guartelá State Park, an important conservation unit of "Campos Gerais". Soils of the sample areas were characterized as sandy, acid, poor in nutrients and organic matter. The genetic variability among the isolates was revealed by BOX-PCR genomic fingerprinting. Phylogeny based on 16S rRNA gene grouped the strains in a large cluster including Paraburkholderia nodosa and P. bannensis, while recA-gyrB phylogeny separated the strains in two groups: one including P. nodosa and the other without any described Paraburkholderia species. MLSA confirmed the separate position of this second group of strains within the genus Paraburkholderia and the nucleotide identity of the five concatened housekeeping genes was 95.9% in relation to P. nodosa BR 3437T. Phylogram based on symbiosis-essential nodC gene was in agreement with 16S rRNA analysis. Our molecular phylogenetic analysis support that Paraburkholderia are the main symbionts of native Mimosa in specific edaphic conditions found in South America and reveal the importance of endemic/native leguminous plants as reservoirs of novel rhizobial species.

Vibrio vulnificus: From Oyster Colonist to Human Pathogen.

Vibrio vulnificus is a Gram-negative halophilic bacterium commonly found in warm coastal waters. The bacterium can cause severe gastroenteritis from consumption of raw seafood as well as wound infections and necrotizing fasciitis, with mortality rates for sepsis and wound infection at 50% and 17%, respectively [1]. Although infections are rare, V. vulnificus is responsible for the most deaths caused by Vibrios [1] and has the highest per case economic impact of all food-related diseases in the United States [2]. Additionally, the geographical area impacted by V. vulnificus is expanding due to global warming and rising sea temperatures, such that the incidence of infection has risen worldwide [3]. Even more troublesome, recent studies have identified V. vulnificus in previously unaffected regions, suggesting there may be future increase of infections [3]. Risk factors for infection include advanced age, male gender, and underlying disease, particularly liver cirrhosis, immunodeficiency, diabetes, and hemochromatosis [1]. However, illness occurs from less than 1 in 10,000 raw oyster meals served to persons with cirrhotic liver, suggesting factors beyond host susceptibility contribute to productive infection [1]. One factor to be considered is whether human-pathogenic V. vulnificus are relatively rare in food and the environment. As V. vulnificus is a diverse bacterial species [4], the specific factor(s) that might define why some isolates cause disease in humans, while the majority of strains are seemingly non-pathogenic, is poorly understood. The identity of these “human-tropic” factors remains the single largest question regarding V. vulnificus pathogenesis. Discovery and characterization of these factors may then facilitate identifying potentially deadly strains. However, this requires researchers to address critical questions regarding V. vulnificus human pathogenesis in the context of its natural environment. Can Pathogenic Potential be Predicted by Genomes? For years, researchers have attempted to identify genotypic or phenotypic markers to classify this species into “pathogenic” or “non-pathogenic” strains. However, this has proven difficult due to the nature of the bacterium: V. vulnificus is naturally competent and frequently exchanges DNA via horizontal gene transfer (HGT), complicating classification systems [4]. One method has been to classify V. vulnificus into biotypes, with Biotype 1 (BT1) originally composed of human clinical and related environmental strains, Biotype 2 (BT2) containing mostly eel pathogens, and Biotype 3 (BT3) representing a recent outbreak of systemic hybrid strains geographically limited to Israel [5, 6]. BT1 strains were further classified by allelic variation in the gene vcg into Clinical (C) and Environmental (E) types [7]. This split of BT1 into two lineages was apparent also by multilocus sequence typing (MLST) [8], with most clinical isolates clustered. However, this linkage of clinical isolates to a single phylogenetic branch has recently been challenged. First, PCR and MLST-based studies showed that V. vulnificus is in linkage disequilibrium, meaning there is a nonrandom association of alleles at multiple loci [9]. Additionally, BOX-PCR genomic fingerprinting identified 52 unique genotypes, demonstrating that the association between genotype and strain source (environmental or clinical) was not significant. This reveals that genotypic profiles alone are not sufficient predictors of virulence. These authors suggest instead the existence of two different ecotypes, lineages A and B, which represent two major monophyletic groups. A separate whole-genome SNP analysis of hundreds of clinical and environmental V. vulnificus strains suggested two ecotypes, A and B [10]. Unexpectedly, BT2 strains previously categorized separately, overlapped with BT1 in the B ecotype, and BT3 strains emerged as an independent branch within the B ecotype. In both the BOX-PCR [9] and whole-genome SNP analyses [10], most of the previously characterized clinical isolates sort to the A lineage or ecotype, but now many sort to the B groupings. Thus, phylogeny does not seem to hold the predictive value for pathogenicity as the number of strains included in these studies expands. Indeed, the genome-wide SNP analysis suggests that it is HGT of chromosomal segments between strains that leads to the continuous evolution of this pathogen [10] and that pathogenesis approaches could be more informative than phylogenetic approaches to understanding V. vulnificus pathogenicity.

Genotypic Diversity and Population Structure of Vibrio vulnificus Strains Isolated in Taiwan and Korea as Determined by Multilocus Sequence Typing

The genetic diversity and population structure of Vibrio vulnificus isolates from Korea and Taiwan were investigated using PCR-based assays targeting putative virulence-related genes and multilocus sequence typing (MLST). BOX-PCR genomic fingerprinting identified 52 unique genotypes in 84 environmental and clinical V. vulnificus isolates. The majority (> 50%) of strains had pathogenic genotypes for all loci tested; moreover, many environmental strains had pathogenic genotypes. Although significant (p < 0.05) inter-relationships among the genotypes were observed, the association between genotype and strain source (environmental or clinical) was not significant, indicating that genotypic characteristics alone are not sufficient to predict the isolation source or the virulence of a given V. vulnificus strain and vice versa. MLST revealed 23–35 allelic types per locus analyzed, resulting in a total of 44 unique sequence types (STs). Two major monophyletic groups (lineages A and B) corresponding to the two known lineages of V. vulnificus were observed; lineage A had six STs that were exclusively environmental, whereas lineage B had STs from both environmental and clinical sources. Pathogenic and nonpathogenic genotypes predominated in MLST lineages B and A, respectively. In addition, V. vulnificus was shown to be in linkage disequilibrium (p < 0.05), although two different recombination tests (PHI and Sawyer’s tests) detected significant evidence of recombination. Tajima’s D test also indicated that V. vulnificus might be comprised of recently sub-divided lineages. These results suggested that the two lineages revealed by MLST correspond to two distinct ecotypes of V. vulnificus.

Population Structure and Diversity of Phenazine-1-Carboxylic Acid Producing Fluorescent Pseudomonas spp. from Dryland Cereal Fields of Central Washington State (USA)

Certain strains of the rhizosphere bacterium Pseudomonas fluorescens contain the phenazine biosynthesis operon (phzABCDEFG) and produce redox-active phenazine antibiotics that suppress a wide variety of soilborne plant pathogens. In 2007 and 2008, we isolated 412 phenazine-producing (Phz(+)) fluorescent Pseudomonas strains from roots of dryland wheat and barley grown in the low-precipitation region (<350 mm annual precipitation) of central Washington State. Based on results of BOX-PCR genomic fingerprinting analysis, these isolates, as well as the model biocontrol Phz(+) strain P. fluorescens 2-79, were assigned to 31 distinct genotypes separated into four clusters. All of the isolates exhibited high 16S rDNA sequence similarity to members of the P. fluorescens species complex including Pseudomonas orientalis, Pseudomonas gessardii, Pseudomonas libanensis, and Pseudomonas synxantha. Further recA-based sequence analyses revealed that the majority of new Phz(+) isolates (386 of 413) form a clade distinctly separated from P. fluorescens 2-79. Analysis of phzF alleles, however, revealed that the majority of those isolates (280 of 386) carried phenazine biosynthesis genes similar to those of P. fluorescens 2-79. phzF-based analyses also revealed that phenazine genes were under purifying selection and showed evidence of intracluster recombination. Phenotypic analyses using Biolog substrate utilization and observations of phenazine-1-carboxylic acid production showed considerable variability amongst members of all four clusters. Biodiversity indices indicated significant differences in diversity and evenness between the sampled sites. In summary, this study revealed a genotypically and phenotypically diverse group of phenazine producers with a population structure not seen before in indigenous rhizosphere-inhabiting Phz(+) Pseudomonas spp.

Impact of Continuous Cropping of Lettuce on the Disease Dynamics of Bottom Rot and Genotypic Diversity of Rhizoctonia solani AG 1-IB

The impact of continuous cropping of lettuce on the disease dynamics of bottom rot and genotypic diversity of the causal pathogen Rhizoctonia solani AG 1-IB was studied over 3 years with two crops per year within a field naturally infested with R. solani the pathogen. This field had not had lettuce cultivated in it for 7 years. The disease incidence (DI) and disease severity (DS) were assessed at each harvest and mapped. Surprisingly, a high DI was already observed in the first crop of year one of this field study. In addition, the pathogen was also found to be evenly distributed. Severely infected plants occurred mainly in patches, and the position varied between crops. A significant increase in DS was already observed in the second year, and both temperature conditions and continuous cropping influenced the DS on average over time. Rhizoctonia isolates were randomly collected from the first crop in 1999 and the sixth crop in 2001. The genotypic diversity within the subgroup of R. solani AG 1-IB was analysed by BOX-PCR genomic fingerprinting and the aggressiveness of isolates by bioassay. The fingerprints revealed a high level of genotypic diversity within the AG 1-IB field population. However, continuous cropping was found not to have an impact on genotypic diversity and aggressiveness.

Epidemiology of Pseudomonas cichorii, the Cause of Lettuce Midrib Rot

Bacterial midrib rot, caused by Pseudomonas cichorii, has become a serious threat to the production of greenhouse butterhead lettuce (Lactuca sativa L. var. capitata) in Belgium. Currently, there are no strategies for controlling this pathogen. Therefore, greenhouse experiments were conducted to obtain more knowledge about the epidemiology of P. cichorii on butterhead lettuce. Greenhouse butterhead lettuce becomes susceptible to lettuce midrib rot infections at head formation, and a single overhead irrigation with water containing 102 CFU/ml P. cichorii was sufficient to cause disease. The use of surface drip irrigation instead of overhead sprinkler irrigation significantly reduced midrib rot incidence in the greenhouse. P. cichorii isolates can be divided into subgroups based on BOX-PCR genomic fingerprinting, with isolates belonging to subgroup C1 and C2 being more virulent than those of (or related to) subgroup C3. P. cichorii infections with distinct symptoms comparable to midrib rot have also been observed on field-grown crisphead lettuce in California and Japan which, respectively, are referred to as ‘varnish spot’ or ‘tar’. We showed that symptom expression is strongly influenced by the lettuce cultivar group, irrespective of the P. cichorii isolate, resulting in varnish spot/tar on crisphead lettuce and midrib rot on butterhead or cutting group lettuce.

Survival and catabolic performance of introducedPseudomonasstrains during phytoremediation and bioaugmentation field experiment

A long-term field experiment was carried out to estimate the efficiency of bioaugmentation in combination with phytoremediation for oil shale chemical industry solid waste dump area remediation. Soil samples for microbiological and chemical analysis were collected during 3 years after bacterial biomass application. Microbial communities in soil samples were analysed using both culture-based and molecular methods. The survival of the introduced bacterial strains was confirmed by cultivation-based Box-PCR genomic fingerprints and denaturing gradient gel electrophoresis fingerprinting of the 16S rRNA and lmPH genes. The introduced bacterial strains as well as corresponding catabolic genes were recovered several years after biomass application, predominantly from the rhizosphere of birches. Soil samples from bioaugmented plots showed an elevated potential for degradation of phenolic compounds even 40 months after treatment. Based on our results we can conclude that the introduced Pseudomonas strains both survived, and their metabolic traits have persisted at the contaminated site over a long period of time.

Pseudomonas cichorii as the causal agent of midrib rot, an emerging disease of greenhouse-grown butterhead lettuce in Flanders

Bacterial midrib rot of greenhouse-grown butterhead lettuce ( Lactuca sativa L. var. capitata) is an emerging disease in Flanders (Belgium) and fluorescent pseudomonads are suspected to play an important role in the disease. Isolations from infected lettuces, collected from 14 commercial greenhouses in Flanders, yielded 149 isolates that were characterized polyphasically, which included morphological characteristics, pigmentation, pathogenicity tests by both injection and spraying of lettuce, LOPAT characteristics, FAME analysis, BOX-PCR fingerprinting, 16S rRNA and rpoB gene sequencing, as well as DNA–DNA hybridization. Ninety-eight isolates (66%) exhibited a fluorescent pigmentation and were associated with the genus Pseudomonas. Fifty-five of them induced an HR + (hypersensitive reaction in tobacco leaves) response. The other 43 fluorescent isolates were most probably saprophytic bacteria and about half of them were able to cause rot on potato tuber slices. BOX-PCR genomic fingerprinting was used to assess the genetic diversity of the Pseudomonas midrib rot isolates. The delineated BOX-PCR patterns matched quite well with Pseudomonas morphotypes defined on the basis of colony appearance and variation in fluorescent pigmentation. 16S rRNA and rpoB gene sequence analyses allowed most of the fluorescent isolates to be allocated to Pseudomonas, and they belonged to either the Pseudomonas fluorescens group, Pseudomonas putida group, or the Pseudomonas cichorii/ syringae group. In particular, the isolates allocated to this latter group constituted the vast majority of HR + isolates and were identified as P. cichorii by DNA–DNA hybridization. They were demonstrated by spray-inoculation tests on greenhouse-grown lettuce to induce the midrib rot disease and could be re-isolated from lesions of inoculated plants. Four HR + non-fluorescent isolates associated with one sample that showed an atypical midrib rot were identified as Dickeya sp.

Diversity of endophytic enterobacteria associated with different host plants

Fifty-three endophytic enterobacteria isolates from citrus, cocoa, eucalyptus, soybean, and sugar cane were evaluated for susceptibility to the antibiotics ampicillin and kanamycin, and cellulase production. Susceptibility was found on both tested antibiotics. However, in the case of ampicillin susceptibility changed according to the host plant, while all isolates were susceptible to kanamycin. Cellulase production also changed according to host plants. The diversity of these isolates was estimated by employing BOX-PCR genomic fingerprints and 16S rDNA sequencing. In total, twenty-three distinct operational taxonomic units (OTUs) were identified by employing a criterion of 60% fingerprint similarity as a surrogate for an OTU. The 23 OTUs belong to the Pantoea and Enterobacter genera, while their high diversity could be an indication of paraphyletic classification. Isolates representing nine different OTUs belong to Pantoea agglomerans, P. ananatis, P. stewartii, Enterobacter sp., and E. homaechei. The results of this study suggest that plant species may select endophytic bacterial genotypes. It has also become apparent that a review of the Pantoea/Enterobacter genera may be necessary.

Diversity of a Soybean Bradyrhizobial Population Adapted to an Indian Soil

Soybean (Glycine max) is an introduced crop in India. Over the years it has been regularly inoculated with indigenous rhizobia. In this study genetic diversity has been studied at a site where soybean has been regularly grown with inoculation. Rhizobia were plant trapped using soybean varieties as host, and fingerprinted using BOX-PCR. BOX-PCR genomic fingerprints of 69 isolates from the nodules of 4 soybean varieties Pusa22, Bragg, PK1041 and PK1142 showed a high level of genetic diversity. The population profiles of the 69 isolates clustered them into 10 groups. Root nodule isolates from the four varieties were Bradyrhizobium japonicum types, growing in 4–7 days with typical colonies which were found to be genetically distinct from the USDA and SEMIA strains of B. japonicum and B. elkanii. Also the genotype of the host plant seemed to be one of the factors determining the diversity. The high diversity could be attributed both to lateral transfer of genetic material between inoculant and indigenous strains and to genomic rearrangements during the adaptation to the Indian soils.

Differentiation of xanthomonads causing the bacterial leaf spot of poinsettia in China from the pathotype strain of Xanthomonas axonopodis pv. poinsettiicola

In October 2003, a new bacterial disease with symptoms similar to those caused by Xanthomonas axonopodis pv. poinsettiicola was observed on poinsettia leaves at a flower nursery in Zhejiang Province of China. Three Xanthomonas strains were isolated from infected plants and classified as X. axonopodis. They were differentiated from the pathotype strain LMG849 of X. axonopodis pv. poinsettiicola causing bacterial leaf spot of poinsettia by comparison of pathogenicity, substrate utilization and BOX-PCR genomic fingerprints.

High diversity among environmental Escherichia coli isolates from a bovine feedlot.

Approximately 280 Escherichia coli isolates were isolated from a bovine feedlot at the University of Connecticut campus via enrichment in lauryl tryptose broth and random selection from MacConkey plates. The E. coli subspecies diversity was estimated by employing whole-cell BOX-PCR genomic fingerprints. A total of 89 distinct operational taxonomic units (OTUs) were identified by employing a criterion of 85% fingerprint similarity as a surrogate for an OTU, while the Chao1 index estimated the E. coli population richness at 128 OTUs. One genotype (at a similarity level of 60%) dominated the population at 66% regardless of sampling depth or location, while no significant vertical distribution pattern was observed in terms of genotype, mobility, antibiotic resistance profile, or biofilm-forming ability. Motility, measured by a soft agar assay, had a very broad range among the E. coli population and was positively correlated with biofilm-forming ability in minimal medium (Spearman's rank correlation coefficient r = 0.619, P < 10(-4)) but not in Luria broth. Only an estimated 48% of the population possessed gene agn43, which encodes Ag43, a phase-variable outer membrane protein that has been implicated in biofilm formation in minimal medium. We observed significantly more biofilm formation in both minimal medium and Luria broth for agn43(+) strains, with a larger effect in minimal medium. This study represents an exhaustive inventory of extant E. coli population diversity at a bovine feedlot and reveals significant subspecies heterogeneity in interfacial behavior.

Identification of iron-reducing Thermus strains as Thermus scotoductus.

Thermus strain SA-01, previously isolated from a deep (3.2 km) South African gold mine, is closely related to Thermus strains NMX2 A.1 and VI-7 (previously isolated from thermal springs in New Mexico, USA, and Portugal, respectively). Thermus strains SA-01 and NMX2 A.1 have also been shown previously to grow using nitrate, Fe(III), Mn(IV) or S(O) as terminal electron acceptors and to be capable of reducing Cr(VI), U(VI), Co(III), and the quinone-containing compound anthraquinone-2,6-disulfonate. The objectives of this study were to determine the phylogenetic positions of the three known metal-reducing Thermus strains and to determine the phylogenetic significance of metal reduction within the genus Thermus. Phylogenetic analyses of 16S rDNA sequences, BOX PCR genomic fingerprinting, and DNA-DNA reassociation analyses indicated that these strains belong to the previously described genospecies T. scotoductus. The morphologies and lipid fatty acid profiles of these metal-reducing strains are consistent with their identification as T. scotoductus; however, the T. scotoductus strains tested in this study evinced a wide intraspecies variability in some other phenotypic traits, e.g., carbon substrate utilization and pigmentation. Iron reduction occurred in all strains of T. scotoductus tested except the mixotrophic, sulfur-oxidizing strain IT-7254. Thermus strains belonging to other species did not reduce Fe(III) to Fe(II) or reduced it only poorly.

Diversity of activated sludge bacteria receiving the 3-chloroaniline-degradative plasmid pC1gfp

Plasmid pC1 of Delftia acidovorans CA28, encoding the oxidative deamination of 3-chloroaniline (3-CA), was tagged with a mini-Tn 5 transposon containing the gfp gene. The labelled plasmid, designated pC1gfp, was subsequently transferred to Pseudomonas putida UWC3 and the plasmid transfer from this donor to the bacterial community in activated sludge was studied. Mating experiments were performed on both Luria–Bertani agar plates and directly in liquid activated sludge. Green fluorescent colonies appearing on mineral medium containing 3-CA as sole nitrogen source were picked up and verified to be true pC1gfp-harbouring transconjugants. REP- and BOX-PCR genomic fingerprinting revealed a large diversity in the transconjugant collection, indicating the occurrence of multiple plasmid transfer events. Remarkably, Luria–Bertani agar plate matings yielded a different set of transconjugants compared to plasmid transfers directly in liquid activated sludge. From the plate matings, mainly Aeromonas sp. was isolated, whereas D. acidovorans strains were predominant in the transconjugant collection from liquid activated sludge. The majority of the isolates showed 3-CA deamination, probably to chlorocatechol, but several pC1gfp-harbouring transconjugants performed a rapid and complete 3-CA degradation.

Genotypic and phenotypic diversity within species of purple nonsulfur bacteria isolated from aquatic sediments.

To assess the extent of genotypic and phenotypic diversity within species of purple nonsulfur bacteria found in aquatic sediments, a total of 128 strains were directly isolated from agar plates that had been inoculated with sediment samples from Haren and De Biesbosch in The Netherlands. All isolates were initially characterized by BOX-PCR genomic DNA fingerprinting, and 60 distinct genotypes were identified. Analyses of 16S rRNA gene sequences of representatives of each genotype showed that five and eight different phylotypes of purple nonsulfur bacteria were obtained from the Haren and De Biesbosch sites, respectively. At the Haren site, 80.5% of the clones were Rhodopseudomonas palustris, whereas Rhodoferax fermentans and Rhodopseudomonas palustris were numerically dominant at the De Biesbosch site and constituted 45.9 and 34.4% of the isolates obtained, respectively. BOX-PCR genomic fingerprints showed that there was a high level of genotypic diversity within each of these species. The genomic fingerprints of Rhodopseudomonas palustris isolates were significantly different for isolates from the two sampling sites, suggesting that certain strains may be endemic to each sampling site. Not all Rhodopseudomonas palustris isolates could degrade benzoate, a feature that has previously been thought to be characteristic of the species. There were differences in the BOX-PCR genomic fingerprints and restriction fragment length polymorphisms of benzoate-coenzyme A ligase genes and form I and form II ribulose 1,5-bisphosphate carboxylase/oxygenase (RubisCO) genes between benzoate-degrading and non-benzoate-degrading genotypes. The ability to distinguish these two Rhodopseudomonas palustris groups based on multiple genetic differences may reflect an incipient speciation event resulting from adaptive evolution to local environmental conditions.

Taxonomic study of aromatic-degrading bacteria from deep-terrestrial-subsurface sediments and description of Sphingomonas aromaticivorans sp. nov., Sphingomonas subterranea sp. nov., and Sphingomonas stygia sp. nov.

Phylogenetic analyses of 16S rRNA gene sequences by distance matrix and parsimony methods indicated that six strains of bacteria isolated from deep saturated Atlantic coastal plain sediments were closely related to the genus Sphingomonas. Five of the strains clustered with, but were distinct from, Sphingomonas capsulata, whereas the sixth strain was most closely related to Blastobacter natatorius. The five strains that clustered with S. capsulata, all of which could degrade aromatic compounds, were gram-negative, non-spore-forming, non-motile, rod-shaped organisms that produced small, yellow colonies on complex media. Their G + C contents ranged from 60.0 to 65.4 mol%, and the predominant isoprenoid quinone was ubiquinone Q-10. All of the strains were aerobic and catalase positive. Indole, urease, and arginine dihydrolase were not produced. Gelatin was not liquified, and glucose was not fermented. Sphingolipids were present in all strains; 2OH14:0 was the major hydroxy fatty acid, and 18:1 was a major constituent of cellular lipids. Acid was produced oxidatively from pentoses, hexoses, and disaccharides, but not from polyalcohols and indole. All of these characteristics indicate that the five aromatic-degrading strains should be placed in the genus Sphingomonas as currently defined. Phylogenetic analysis of 16S rRNA gene sequences, DNA-DNA reassociation values, BOX-PCR genomic fingerprinting, differences in cellular lipid composition, and differences in physiological traits all indicated that the five strains represent three previously undescribed Sphingomonas species. Therefore, we propose the following new species: Sphingomonas aromaticivorans (type strain, SMCC F199), Sphingomonas subterranea (type strain, SMCC B0478), and Sphingomonas stygia (type strain, SMCC B0712).