Clonal analysis of the Actinobacillus pleuropneumoniae population in a geographically restricted area by multilocus enzyme electrophoresis.

The genetic diversity among 250 isolates of Actinobacillus pleuropneumoniae from lungs of pigs with pleuropneumonia and from tonsils of apparently healthy pigs at slaughter was estimated by multilocus enzyme electrophoresis. The Danish strains were derived from both specific-pathogen-free and conventional herds. Sixty-six percent of the isolates belonged to three electrophoretic types (ETs) of a total of 37 ETs detected. While five biotype 2 isolates constituted a separate ET closely related to biotype 1 isolates, the type strain of the species (Shope 4074) belonged to its own ET, with a genetic distance of 0.30 from its nearest neighbor. Isolates of serotypes traditionally considered to have less pathogenic potential (serotypes 6, 10, and 12) from herds with acute outbreaks of pleuropneumonia belonged to the same ETs as isolates from apparently healthy pigs, suggesting that factors such as cross immunity and management may lead to divergent clinical results. Isolates from four herds harboring more than one serotype showed distinct profiles between the serotypes, indicating no or only limited chromosomal recombination among clones. Isolates from tonsils belonged to the same ET as isolates from lungs. The same ET was isolated from widely different parts of the world. Evidence from this study indicates that multilocus enzyme electrophoresis may be a valuable tool for the epidemiological analysis of A. pleuropneumoniae.     

Characterization of isolates of Mycobacterium avium serotypes 4 and 8 from patients with AIDS by multilocus enzyme electrophoresis.

Isolates of Mycobacterium avium serotypes 4 and 8 originating from patients with AIDS in New York City, Los Angeles, or San Francisco were further characterized by multilocus enzyme electrophoresis. Reference strains used to produce typing antisera were also examined. Thirty-one electrophoretic types (ETs) were found among 58 isolates of serotype 4, while 10 ETs were identified among 21 isolates of serotype 8. One major ET was found within each serotype, and these two ETs were closely related, separated by a genetic distance of only 0.05. Six ETs were found in more than one city. In four cases, isolates of serotypes 4 and 8 shared the same ET. Multilocus enzyme electrophoresis in combination with serotyping should be helpful in locating the specific infection sources of these commonly isolated opportunistic pathogens.     

Genetic diversity in relation to serotype in Escherichia coli.

The extent of chromosomal-gene diversity among 261 isolates of Escherichia coli sharing single O, K, or H antigens and various combinations thereof was estimated by multilocus enzyme electrophoresis, which detects allelic variation in structural genes. The results of this study indicate that the genetic diversity among isolates sharing single antigenic determinants can approach or equal that observed among randomly chosen strains; that the magnitude of the diversity varies among antigens; and that the genetic diversity is reduced, but not eliminated, among isolates sharing two antigenic determinants. With one exception, isolates of the same O:K:H serotype were of identical or closely related electrophoretic types (ETs). Isolates of the same ET generally shared the same combination of antigenic determinants, but some ETs included isolates of different serotypes. The implications of these findings for epidemiological research and the clone hypothesis of population structure are discussed, and possible evolutionary mechanisms causing antigenic divergence and convergence are considered.     

Genetic relationships and clonal population structure of serotype 2 strains of Neisseria meningitidis.

Two hundred and thirty-four strains of Neisseria meningitidis, including 94 serotype 2a, 111 serotype 2b, and 19 serotype 2c isolates, together with 10 isolates that were serotyped as 2 with polyvalent antiserum but did not react with monoclonal antibodies, were characterized by the electrophoretic mobilities of 15 metabolic enzymes. Of these enzymes, 14 were polymorphic, and 56 distinctive combinations of alleles at the enzyme loci (electrophoretic types) were identified, among which the mean genetic diversity per locus was 0.413, or about 75% of that recorded for the species N. meningitidis as a whole. Mean genetic diversity among electrophoretic types of the same serotype (2a, 2b, or 2c) was, however, on average, less than half the total species diversity, and no multilocus genotypes were shared between isolates of the different serotypes, which belong to distinctive clonal lineages. Recent temporal changes in the frequencies of recovery of pathogenic strains of serotypes 2a and 2b in South Africa and North America resulted from clone replacement in these populations rather than evolutionary modification of the serotype protein of the initially dominant clones.     

Population structure of Australian isolates of Streptococcus suis.

The genetic diversity of 109 isolates of Streptococcus suis, which were recovered mainly from Australian pigs, was examined by multilocus enzyme electrophoresis. The collection was genetically diverse. Sixty-five electrophoretic types (ETs) were recognized, with a mean genetic diversity per enzyme locus of 0.512, or 0.431 when the number of isolates in each ET was considered. Serotype diversity varied, being greatest for isolates of capsular serotype 15 (0.364), and then diminishing in the order of serotypes 9, 1, 4, 1/2, 2, 7, and 3 (0.120). On average, isolates from these eight serotypes represented 4.13 separate clonal groups per serotype. This diversity indicated that serotyping of S. suis for subspecific differentiation is not a reliable technique for identifying specific strains and is not a good predictor of the genetic background of a given isolate. No tendency for isolates recovered from healthy pigs to be genetically distinct from those from diseases animals was found, nor were there consistent differences between isolates recovered from animals with different disease syndromes (meningitis, pneumonia, and septicemia). Danish reference strains of serotypes 1, 2, and 7 each belonged to one of the same clonal groupings of these types found in Australia, but Danish strains of serotypes 3, 4, 6, and 8 and a strain of serotype 1 from the United Kingdom were each genetically distinct from the Australian isolates. Generally, isolates in the same ET belonged to the same serotype, but one ET contained isolates of types 6 and 6/16, and three were made up of isolates of types 2 and 1/2. One isolate of serotype 2, which was recovered from a human with meningitis, belonged to the same ET as two isolates of serotype 2 that were recovered from pigs. The human infection was therefore likely to have been zoonotic.     

Genetic diversity of penicillin G-resistant Neisseria meningitidis from Spain.

Genotypic and phenotypic diversity among 16 penicillin G-resistant (Penr) isolates of Neisseria meningitidis recovered from human blood or cerebrospinal fluid in Spain was compared with that among 12 penicillin-susceptible (Pens) isolates by the use of multilocus enzyme electrophoresis, serotyping, auxotroph testing in chemically defined media, and sodium dodecyl sulfate-polyacrylamide gel electrophoresis of penicillin-binding proteins (PBPs). Thirteen distinctive multilocus enzyme genotypes (electrophoretic types [ETs] ) were identified among the 28 isolates. There was slightly less genetic diversity among the eight ETs of Penr isolates (H = 0.385) than among the eight ETs of Pens isolates (H = 0.431). Cluster analysis demonstrated two distinctive complexes of ETs and one ET that was not closely related to either complex. The possibility of a singular clonal origin of penicillin G-resistant isolates was excluded by the observations that resistance occurred in isolates of each of the two distantly related complexes of ETs, that three of the four ETs represented by multiple isolates included both susceptible and resistant strains, and that serotypes and growth requirements were not associated with the resistance phenotype. The 28 isolates showed a relatively homogeneous pattern of four PBPs, with apparently reduced penicillin G binding by PBP 3 of the Penr isolates.     

Genetic relationships of serologically nontypable and serotype b strains of Haemophilus influenzae.

A collection of 242 strains of Haemophilus influenzae, including 65 nontypable (unencapsulated) isolates and 177 encapsulated serotype b isolates recovered largely from children with invasive and noninvasive diseases in the United States, was characterized by the electrophoretic mobilities of 15 metabolic enzymes presumably encoded by chromosomal genes. All enzymes were polymorphic for three to seven electromorphs, and 94 distinctive multilocus genotypes (electrophoretic types [ETs]) were distinguished, among which mean genetic (allelic) diversity was 0.500. Isolates recovered from cases of invasive or noninvasive diseases did not differ significantly in level of genetic variation. The observation that 29 ETs were represented exclusively by serotype b isolates and that each of the 65 nontypable isolates was of a unique ET strongly confirmed the hypothesis that unencapsulated clinical isolates are not merely phenotypic variants of the common serotype b cell lines. Rather, the two types of isolates are distinctive subsets of the multilocus chromosomal genotypes of the species as a whole. Serotype b capsule occurred in three groups of isolates that are distantly related in multilocus enzyme genotype. Isolates of four closely related nontypable biotype IV ETs associated with obstetrical infections or neonatal bacteremia were highly divergent from all others examined and may be specifically distinct. A phylogenetic scenario was proposed in which the ancestor of H. influenzae was encapsulated and the nontypable clones arose by convergent evolutionary loss of the ability to synthesize or extracellularly express a polysaccharide capsule.     

Genetic relationships of penicillin-susceptible and -resistant Streptococcus pneumoniae strains isolated on different continents.

Sixty-six strains of Streptococcus pneumoniae isolated in different parts of the world, 46 resistant and 22 susceptible to penicillin, were subdivided by multilocus enzyme electrophoresis into 28 distinct electrophoretic types (ETs). The ETs to which penicillin-susceptible strains were assigned differed from those containing resistant isolates of the same serotype. Five common clones could be recognized among the penicillin-resistant bacteria by combining the ETs, the antigenic properties of penicillin-binding proteins PBP 1a and 2b, and the tetracycline and chloramphenicol resistance profiles. Two clones were found in Finland and were associated with capsular serotypes 6B and 23F, respectively. Two clones were from Spain (type 6B and 9V, respectively). The fifth clone was isolated in South Africa and in Spain and contained both serotype 23F isolates and one type 19F strain. The other resistant strains were represented by rare isolates distributed among 12 other ETs, confirming that resistance to penicillin has evolved by multiple branches. Because capsular type was mixed in several ETs, the results also demonstrate that it may vary among very closely related pneumococci.     

Clonal nature of Salmonella typhi and its genetic relatedness to other salmonellae as shown by multilocus enzyme electrophoresis, and proposal of Salmonella bongori comb. nov

Crude cell extracts of 26 isolates of Salmonella serotype typhi (S. typhi) and 48 other Salmonella isolates representing 28 serotypes and seven DNA hybridization subgroups were analyzed for electrophoretic variants of 24 metabolic enzymes by starch gel electrophoresis. All strains of S. typhi had identical isoenzyme patterns, indicating that they were a single clone. All of the enzymes detected in the remaining strains were polymorphic, and the degree of genetic variation was quite high. The average number of alleles per enzyme locus was 4.7, and the mean genetic diversity per locus was 0.556. Thirty-two distinct allele profiles, or electrophoretic types (ETs), were found in these 48 strains of Salmonella serotypes other than S. typhi. Analysis of the genetic relationships of the ETs to each other showed that, with one exception, the ETs formed subgroups that were consistent with the subgroupings based on DNA hybridization studies. ET profiles were not always linked to specific serologic patterns. These data show that multilocus enzyme electrophoresis has a potential application in epidemiologic and taxonomic studies of salmonellae, although it is not differential for S. typhi. We also propose a new species, Salmonella bongori comb. nov., a new combination base on the elevation of Salmonella choleraesuis subsp. bongori to the level of species.     

Characterization of Streptococcus agalactiae strains by multilocus enzyme genotype and serotype: identification of multiple virulent clone families that cause invasive neonatal disease.

The chromosomal genotypes of 277 isolates of 16 serotypes of Streptococcus agalactiae were characterized by analysis of electrophoretically demonstrable allele profiles at 12 metabolic enzyme loci. The collection comprised the type strain and 276 strains recovered from French symptomatic and asymptomatic subjects. Sixty-one distinctive electrophoretic types (ETs), representing multilocus clonal genotypes, were identified. Cluster analysis of the ETs revealed two primary phylogenetic divisions separated by a genetic distance of 0.62, Division I contained 67 isolates which could be assigned to 13 ETs. Twenty-seven of these isolates were from samples of cerebrospinal fluid (CSF) from neonatal meningitis patients. Two ETs, separated by a genetic distance of 0.217, contained 26 of these 27 isolates. Division II contained 210 isolates, of which 27 were isolated from CSF. This division was more polymorphic and included 48 ETs. Spanning a genetic distance of 0.3, three clusters and one ET were identified within this group. Twenty-four of 27 strains isolated from CSF belonged to one cluster, and 19 of them belonged to two adjacent ETs with a genetic distance of 0.083. Fifty-five of the 68 serotype Ia strains and 24 of the 26 serotype Ib strains were each confined to one of the evolutionary lineages, and 85 of the 86 strains which carried protein antigen c belonged to phylogenetic division II. Most of the type III organisms were assigned to two clone families. The characteristics of this French population argue for the existence of particular groups of strains responsible for neonatal meningitis and demonstrate that serotyping can supply information about the genetic distribution of strains.     

Difference in structure between type b and nontypable Haemophilus influenzae populations

The extent of chromosomal genetic variability and the genetic structure of Haemophilus influenzae populations was analyzed. A total of 119 isolates from humans in G?teborg, Sweden, and Birmingham, Ala., and 16 strains from a type culture collection were characterized for capsular type, biotype, outer membrane protein profile, and enzyme electrophoretic type (ET). The results of this study indicate that the bacteria identified as H. influenzae are a genetically extremely variable array of organisms. For the six enzymes studied, the estimated mean genetic diversity was 0.57 (approximately 20% higher than the corresponding estimate for Escherichia coli). Two lines of evidence indicate that despite its ability to recombine by transformation, H. influenzae maintains a largely clonal population structure. Although there is considerable potential for generating different genotypes, there were only 88 distinct ETs among the 135 strains, and isolates of the same ET and biotype were recovered at frequencies greater than would be anticipated at random. This evidence for a clonal population structure holds for uncapsulated as well as capsulated strains. However, these data also suggest that the stability of H. influenzae clones (clone persistence time) may be less than that of the nontransforming species E. coli. The ET data indicate that there is somewhat less variability among H. influenzae strains that express the same capsular antigens, biotype, and outer membrane proteins than among randomly chosen isolates. Nevertheless, there is substantial genetic variation among isolates within each of these classes and combinations thereof. There is also variation in these typing characteristics among strains of the same ET. These observations and those on genetic variability and population structures have implications for the characterization of H. influenzae isolates in clinical and epidemiological studies.     

Genetic structure of population of Bacillus cereus and B. thuringiensis isolates associated with periodontitis and other human infections

The genetic diversity and relationships among 35 Bacillus cereus and Bacillus thuringiensis isolates recovered from marginal and apical periodontitis in humans and from various other human infections were investigated using multilocus enzyme electrophoresis. The strains were isolated in Norway, except for three strains isolated from periodontitis patients in Brazil. The genetic diversity of these strains was compared to that of 30 isolates from dairies in Norway and Finland. Allelic variation in 13 structural gene loci encoding metabolic enzymes was analyzed. Twelve of the 13 loci were polymorphic, and 48 unique electrophoretic types (ETs) were identified, representing multilocus genotypes. The mean genetic diversity among the 48 genotypes was 0.508. The genetic diversity of each source group of isolates varied from 0.241 (periodontal infection) to 0.534 (dairy). Cluster analysis revealed two major groups separated at a genetic distance of greater than 0.6. One cluster, ETs 1 to 13, included solely isolates from dairies, while the other cluster, ETs 14 to 49, included all of the human isolates as well as isolates from dairies in Norway and Finland. The isolates were serotyped using antiflagellar antiserum. A total of 14 distinct serotypes were observed. However, little association between serotyping and genotyping was seen. Most of the strains were also analyzed with pulsed-field gel electrophoresis, showing the presence of extrachromosomal DNA in the size range of 15 to 600 kb. Our results indicate a high degree of heterogeneity among dairy strains. In contrast, strains isolated from humans had their genotypes in one cluster. Most strains from patients with periodontitis belonged to a single lineage, suggesting that specific clones of B. cereus and B. thuringiensis are associated with oral infections.     

Clonal relationships and variation in virulence among Escherichia coli strains of avian origin

Average genetic relatedness among 44 Escherichia coli strains of serotypes O1, O2, and O78 isolated mainly from birds with colibacillosis or swollen-head syndrome from France or Saudi Arabia was estimated based on allelic variation detected by multilocus enzyme electrophoresis. For 20 enzyme-encoding loci, we resolved 2.8 alleles per locus and distinguished 17 electrophoretic types (ETs) that were used to mark naturally occurring cell lineages or clones. On average, ETs differed at 37% of their loci. Forty-eight percent of the isolates represent three ETs, two of which belong to previously defined complexes of clones identified in avian disease in North America and Europe. Virulence of strains, assessed in experimental infections of day-old chicks, showed little variation among isolates of a clone, but was significantly variable among isolates of different clone complexes. These findings add support to the evidence that a majority of avian isolates that cause colibacillosis belong to a few cosmopolitan pathogenic clones and indicate a substantial between-clone component of pathogenicity.     

Comparison of multilocus enzyme electrophoresis and random amplified polymorphic DNA analysis for molecular subtyping of Cryptococcus neoformans. The Cryplococcal Disease Active Surveillance Group.

We evaluated multilocus enzyme electrophoresis (MEE) and random amplified polymorphic DNA (RAPD) for their usefulness in subtyping 344 Cryptococcus neoformans clinical isolates obtained from four U.S. metropolitan areas in 1992 to 1994. MEE and RAPD with five primers both discriminated between the two varieties of C. neofromans. MEE divided C. neoformans var. neoformans isolates into 15 enzyme electrophoretic subtypes (ETs) arranged in three complexes. The predominant ET 1 complex contained 10 ETs, with isolates from 70% of patients in 1 ET. RAPD with five primers further sorted this predominant ET into 19 subtypes, with 60% of isolates sorting into three RAPD types. The ET 8 MEE complex, containing three ETs, could not be divided further by RAPD. The ET 7 complex (two ETs) included isolates from all serotype AD patients. Although both MEE and RAPD identified isolates of C. neoformans var. gattii, neither distinguished between serotypes B and C. These results showed that the two C. neoformans varieties could be identified by MEE or RAPD profile as well as by biochemical methods. RAPD improved the discriminatory power of MEE for isolates within the ET 1 complex but with other ETs offered little additional sensitivity over MEE and was less sensitive than MEE with isolates of C. neoformans var. gattii. This information will be useful in identifying particular environmental sources of disease-causing exposures, in seeking clusters of cases, and in determining whether an infecting strain changes over time.     

Clonal diversity of Neisseria meningitidis from a population of asymptomatic carriers.

Genetic diversity and relationships among 109 isolates of Neisseria meningitidis obtained from throat cultures of healthy individuals in Norway in 1984 were assessed by analyzing electrophoretically demonstrable allelic variation at 15 enzyme-encoding chromosomal genes. Seventy-eight distinctive electrophoretic types (ETs), representing multilocus genotypes, were identified. The mean genetic diversity per locus among the 78 ETs (0.538) was equivalent to that among 19 ETs represented by 66 isolates collected from patients with meningococcal disease in Norway in the first 5 months of 1984. The clonal composition of the collection of carrier strains was, however, quite different from that of strains from patients. The two groups of clones, the ET-5 complex and the ET-37 complex, that were responsible for 91% of the cases of systemic disease in Norway in 1984 were identified in only 7 and 9%, respectively, of the throat cultures from healthy individuals, and their frequencies in the human population sampled were only 0.7% for clones of the ET-5 complex and 0.9% for those of the ET-37 complex. The complex of clones that was most frequently represented by isolates from carriers (19%) has never been recovered from patients with meningococcal disease in Norway or elsewhere, which suggests that these clones have a low virulence potential. Children attending the same day care center or school seldom harbored the same clone in their throats.     

Antigenic and genetic variation in Legionella pneumophila serogroup 6.

Legionella pneumophila subsp. pneumophila serogroup 6 is second in importance only to L. pneumophila serogroup 1 as a cause of legionellosis. Monoclonal antibody (MAb) reactivity and multilocus enzyme electrophoretic analyses were used to subtype serogroup 6 isolates as a potential aid for epidemiologic and virulence studies. Forty-eight serogroup 6 isolates submitted to the Centers for Disease Control from 1980 to 1985 were examined by these methods. The isolates were divided into two groups based on differential reactivity with two MAbs. Thirty-two of the isolates were of a single electrophoretic type (ET) and were reactive with both MAbs. The remaining 16 isolates were distributed among 10 ETs and were reactive with one or both MAbs. The mean genetic diversity for serogroup 6, as determined from the degree of variability at 20 enzyme loci, was found to be essentially the same as that for L. pneumophila subsp. pneumophila as a whole. The ETs of serogroup 6 isolates were unique but closely related genetically to the ETs of L. pneumophila subsp. pneumophila serogroups 1 to 5, 7, and 8. The range of serogroup 6 subtypes distinguished by MAbs and enzyme electrophoresis suggests that the combination of these two methods can be useful as a typing system.     

Comparison of ribotyping and multilocus enzyme electrophoresis for subtyping of Listeria monocytogenes isolates.

Ribotyping was compared with multilocus enzyme electrophoresis (MEE) for subtyping 305 Listeria monocytogenes isolates from clinical and nonclinical sources. For ribotyping, EcoRI-restricted genomic DNA fragments of L. monocytogenes strains were separated by agarose gel electrophoresis, and Southern blots were probed with a cloned Escherichia coli rrnB operon (plasmid pKK3535) labeled with digoxigenin. The L. monocytogenes isolates were divided into 28 distinct ribotypes, while MEE analysis divided the same isolates into 78 electrophoretic types (ETs). On the basis of their ribotype profiles, the strains were divided into two subgroups. The ribotype alpha (RT alpha) subgroup contained serotypes 1/2a, 1/2c, and 3a, and the ribotype beta (RT beta) subgroup contained serotypes 1/2b, 3b, 4b, and 4ab. This division is in complete agreement with MEE analysis, which divides the species into two subgroups (ET groups A and B), with the same serotype distribution in each subgroup. Overall, MEE was more discriminating than ribotyping. However, in several instances ribotyping discriminated between isolates within the same ET. Ribotyping was more discriminating for serotypes 1/2a, 1/2c, and 3a (Simpson's Index for Diversity [DI] = 0.81) than for serotypes 1/2b and 4b (DI = 0.76). A substantial proportion (69%) of serotype 1/2b and 4b strains clustered in five ETs and five ribotypes. These data suggest that ribotyping and MEE do not provide adequate discrimination between strains of serotypes 1/2b and 4b. Methods such as pulsed-field gel electrophoresis and random amplified polymorphic DNA analysis should be explored for further discrimination of strains of these serotypes.     

Genotypic heterogeneity of strains of Citrobacter diversus expressing a 32-kilodalton outer membrane protein associated with neonatal meningitis.

Genetic diversity and relationships among 42 strains of Citrobacter diversus recovered from the cerebrospinal fluid of human infants with meningitis and from other clinical sources in the United States were estimated on the basis of electrophoretically detectable allelic variation in 20 genes encoding metabolic enzymes. Sixteen distinctive multilocus enzyme genotypes were identified, among which the mean genetic diversity per locus was 0.244. The recovery of isolates of the same genotype in several regions of the United States and over periods as long as 20 years indicates that the population structure of C. diversus is clonal. There was little association between multilocus enzyme genotype and biotype, piliation, or presence of a 32-kilodalton outer membrane protein. The observation that the 32-kilodalton outer membrane protein, which is expressed predominantly by strains recovered from infants with meningitis, occurs in a variety of genotypically diverse clones belonging to several phylogenetic lineages supports the hypothesis that this protein confers virulence.     

Evidence of recombination and an antigenically diverse immunoglobulin A1 protease among strains of Streptococcus pneumoniae.

The genetic relationships among 114 isolates of Streptococcus pneumoniae representing mainly nine serotypes that frequently cause severe childhood disease in Northern Europe were examined by use of multilocus enzyme electrophoresis. A comparison was made of the corresponding antigenic variations of excreted immunoglobulin A1 (IgA1) proteases detected by enzyme neutralization assays. Allelic variation at 13 gene loci among 70 electrophoretic types disclosed a comparatively low mean genetic diversity per locus (H = 0.319). In contrast, IgA1 proteases showed extensive antigenic diversity as 17 different inhibition types were distinguished. A lack of overall clonality was apparent from the linkage equilibrium of alleles harbored by 28 isolates chosen to represent the genetic diversity of the study population. However, certain clones, such as those marked by identical electrophoretic type, serotype, and IgA1 protease type, persisted for a sufficiently long time to enable clonal spread between distant geographic areas. Among clonally related isolates, examples illustrating a shift of capsular serotype or IgA1 protease type supported the view that recombination occurs in vivo in corresponding genes. In conclusion, over time, horizontal genetic exchange appears to be sufficiently frequent to disrupt the clonal structure otherwise generated by binary fission in natural populations of S. pneumoniae. The clonal instability combined with considerable antigenic heterogeneity renders the pneumococcal IgA1 protease less attractive as a potential component of future vaccines.     

Analysis of relationships among isolates of Citrobacter diversus by using DNA fingerprints generated by repetitive sequence-based primers in the polymerase chain reaction.

Oligonucleotide probes which match consensus sequences of the repetitive extragenic palindromic (REP) element hybridize to genomic DNA of diverse bacterial species. Primers based on the REP sequence generate complex band patterns with genomic DNA in the polymerase chain reaction (PCR), a technique named REP-PCR. We used REP-PCR with genomic DNA to fingerprint 47 isolates of Citrobacter diversus. Previously, 37 were assigned electrophoretic types (ETs) by multilocus enzyme electrophoresis and 35 were evaluated by using outer membrane protein profiles. Fingerprints were compared by visual inspection and by similarity coefficients (SimCs) based on the number of common bands versus total bands between two given isolates. DNA fingerprints were highly similar visually for patient pairs and outbreak-related sets. SimCs for these were > or = 0.952. Fingerprints of isolates with different ETs generally were distinctive. Among 21 unrelated isolates representing 15 ETs, only 6 of 210 comparisons had SimCs of > or = 0.952. REP-PCR rapidly generated DNA fingerprints which were highly similar for epidemiologically linked isolates of C. diversus and distinct for previously characterized strains within this species. The ability of this method to discriminate between C. diversus isolates with the same biotype was similar to that of multilocus enzyme electrophoresis and outer membrane protein profiles. REP-PCR may be useful in evaluation of apparent outbreaks of this or other bacterial species which possess these extragenic, repetitive elements.