Identification of Campylobacter coli isolates from animals and humans by bacterial restriction endonuclease DNA analysis

Ninety-nine Campylobacter coli isolates were examined by bacterial restriction endonuclease DNA analysis (BRENDA) with HindIII. Isolates from poultry from the same environment had identical patterns, patterns of isolates carried by suckling piglets were generally the same as those of isolates recovered from their dams, and one human patient yielded the same BRENDA type when sampled 6 weeks later. The 14 human isolates examined produced 11 distinct BRENDA types. Forty-three C. coli isolates from pigs were represented by 20 BRENDA types. Ten C. coli isolates from the feces of gulls yielded five different BRENDA types. Thirty-two C. coli isolates from live chickens and processed chicken yielded five different BRENDA types. Three human isolates had identical DNA patterns; two were from brothers living in the same house, and the third was from a human with no apparent relationship to the brothers. Another human isolate was identical to a poultry isolate. None of the pig strains had DNA patterns resembling those of human strains, nor were the DNA patterns like those of any strains recovered from poultry or gulls. Four C. coli isolates were subcultured onto agar 23 times over a period of 45 days, and their BRENDA patterns were preserved. BRENDA shows great promise for use in epidemiological studies of C. coli.     

Strain-specific differentiation of environmental Escherichia coli isolates via denaturing gradient gel electrophoresis (DGGE) analysis of the 16S-23S intergenic spacer region

Denaturing gradient gel electrophoresis (DGGE) was applied to the 16S-23S rRNA intergenic spacer region (ISR) as a means to evaluate strain level differences in Escherichia coli. The ISRs of 81 environmental E. coli isolates obtained from bovine, poultry, and human sources yielded a total of 41 unique DGGE banding patterns, with identical patterns and common bands within each source and no overlapping patterns among sources. An additional 51 isolates from two nearby streams yielded 45 unique banding patterns with no overlap between sites. However, two of the isolates from the streams had identical banding patterns to those from two of the source isolates, resulting in a total of 84 unique DGGE banding patterns out of 132 isolates identified in this study. These results revealed high diversity among environmental E. coli isolates, which made it difficult to unambiguously ascribe strains found in water samples to specific host organisms.     

Characterization of Listeria monocytogenes isolates from food animal clinical cases: PFGE pattern similarity to strains from human listeriosis cases

Twenty-one isolates of Listeria monocytogenes from food animal clinical cases that involved meningitis or meningoencephalitis, encephalitis, mastitis and abortion were characterized by serotyping and pulsed-field gel electrophoresis (PFGE) in order to improve our understanding of the genetic links between individual strains and strains recovered from human listeriosis cases. Results showed that five of the isolates were serotype 1/2a, six were 1/2b, nine were 4b, and one was untypeable. A caprine, two bovine and an ovine brain isolate shared identical PFGE patterns indicating that strains of L. monocytogenes are not host specific. Other isolates exhibited distinct patterns that were not shared, indicating a genetic diversity. Dendrogram analysis revealed that PFGE patterns of the isolates clustered primarily according to serotype. We compared the PFGE types obtained for these isolates with PFGE types for human clinical isolates present in the CDC national PulseNet database. Six (29%) of the twenty-one strains had patterns that were indistinguishable from pathogenic human isolates in the database. Our observations offer preliminary evidence that food animals could be significant reservoirs of L. monocytogenes that lead to human infections and support the inclusion of PFGE patterns of veterinary clinical isolates in the national PulseNet database for increased surveillance.     

Speciating Campylobacter jejuni and Campylobacter coli isolates from poultry and humans using six PCR-based assays

Six previously published polymerase chain reaction (PCR) assays each targeting different genes were used to speciate 116 isolates previously identified as Campylobacter jejuni using routine microbiological techniques. Of the 116 isolates, 84 were of poultry origin and 32 of human origin. The six PCR assays confirmed the species identities of 31 of 32 (97%) human isolates and 56 of 84 (67%) poultry isolates as C. jejuni. Twenty eight of 84 (33%) poultry isolates were identified as Campylobacter coli and the remaining human isolate was tentatively identified as Campylobacter upsaliensis based on the degree of similarity of 16S rRNA gene sequences. Four of six published PCR assays showed 100% concordance in their ability to speciate 113 of the 116 (97.4%) isolates; two assays failed to generate a PCR product with four to 10 isolates. A C. coli-specific PCR identified all 28 hippuricase gene (hipO)-negative poultry isolates as C. coli although three isolates confirmed to be C. jejuni by the remaining five assays were also positive in this assay. A PCR-restriction fragment length polymorphism assay based on the 16S rRNA gene was developed, which contrary to the results of the six PCR-based assays, identified 28 of 29 hipO-negative isolates as C. jejuni. DNA sequence analysis of 16S rRNA genes from four hipO-negative poultry isolates showed they were almost identical to the C. jejuni type strain 16S rRNA sequences ATCC43431 and ATCC33560 indicating that assays reliant on 16S rRNA sequence may not be suitable for the differentiation of these two species.     

DNA fingerprinting of Cryptosporidium parvum isolates using amplified fragment length polymorphism (AFLP)

The genetic variability of 10 Cryptosporidium parvum isolates of human and animal origin was investigated using amplified fragment length polymorphism (AFLP). Analysis of fluorescent dye-labeled amplified products was carried out using an ABI PRISMS 377 DNA sequencer and ABI PRISMS GeneScan software. One-hundred and twelve primer combinations were evaluated using a single C. parvum isolate. The patterns generated were highly reproducible. For subsequent study, a subset of 9 primer pairs that yielded 30-90 DNA fragments after the polymerase chain reaction, within the size range of 50-500 bp, was used to screen the 10 C. parvum isolates, including 7 bovine, 1 equine, and 2 of human origin. The animal isolates produced identical fingerprint patterns with every primer combination tested. Of the 2 human isolates tested, 1 of the isolates, passaged in calves, generated the same AFLP DNA banding patterns as the animal isolates, whereas the other isolate, obtained directly from human feces, produced unique patterns. Polymorphism, detected by comparison of the fingerprint patterns of the latter human isolate with the common pattern shared by all other isolates, ranged from 17 to 35% for the 9 primer pairs. The results show that AFLP is a useful method for differentiating C. parvum isolates into 2 distinct genotypes.     

Restriction fragment length polymorphism analysis of Cryptosporidium parvum isolates of bovine and human origin.

Cryptosporidium parvum oocysts isolated from different hosts and geographical areas were compared by restriction endonuclease analysis of repetitive DNA: Iowa (bovine), Florida (bovine), New York (bovine), Peru (human), Brazil (human), and Mexico (human). Southern blot hybridization analysis was performed using the restriction endonuclease enzyme Eco RI and the DNA probe pV47-2. The probe hybridized with 18 bands present in all the isolates. The Brazilian, Mexican, and Peruvian human isolates had an additional common band of 4.3 kbp that was absent in the bovine isolates. Two extra bands of 14 and 12 kbp were present in the Brazilian isolate whereas the Mexican isolate had an extra band of 14 kbp. When the Iowa and Peru C. parvum isolates were passed twice through calves, oocysts recovered from both passages showed identical banding patterns, suggesting that recombination of the repetitive sequences was not altered during sexual reproduction. The DNA digested with other restriction endonucleases were tested confirming differences between isolates. A genomic DNA library is currently being produced to better define isolate variation in C. parvum.     

Genetic analyses of Acanthamoeba isolates from contact lens storage cases of students in Seoul, Korea.

We conducted both the small subunit ribosomal DNA (SSU rDNA) polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) and mitochondrial (mt) DNA RFLP analyses for a genetic characterization of Acanthamoeba isolates from contact lens storage cases of students in Seoul, Korea. Twenty-three strains of Acanthamoeba from the American Type Culture Collection and twelve clinical isolates from Korean patients were used as reference strains. Thirty-nine isolates from contact lens storage cases were classified into seven types (KA/LS1, KA/LS2, KA/LS4, KA/LS5, KA/LS7, KA/LS18, KA/LS31). Four types (KA/LS1, KA/LS2, KA/LS5, KA/LS18) including 33 isolates were regarded as A. castellanii complex by riboprints. KA/LS1 type was the most predominant (51.3%) in the present survey area, followed by KA/LS2 (20.9%), and KA/LS5 (7.7%) types. Amoebae of KA/LS1 type had the same mtDNA RFLP and riboprint patterns as KA/E2 and KA/E12 strains, clinical isolates from Korean keratitis patients. Amoebae of KA/LS2 type had the identical mtDNA RFLP patterns with A. castellanii Ma strain, a corneal isolate from an American patient as amoebae of KA/LS5 type, with KA/E3 and KA/E8 strains from other Korean keratitis patients. Amoebae of KA/LS18 type had identical patterns with JAC/E1, an ocular isolate from a Japanese patient. Three types, which remain unidentified at species level, were not corresponded with any clinical isolate in their mtDNA RFLP and riboprint patterns. Out of 39 isolates analyzed in this study, mtDNA RFLP and riboprint patterns of 33 isolates (84.6%) were identical to already known clinical isolates, and therefore, they may be regarded as potentially keratopathogenic. These results suggest that contact lens wearers in Seoul should pay more attention to hygienic maintenance of contact lens storage cases for the prevention of Acanthamoeba keratitis.     

Genotypic analysis of Escherichia coli recovered from product and equipment at a beef-packing plant

AIMS: To identify sources of Escherichia coli on beef by characterizing strains of the organism on animals, equipment and product at beef-packing plant. METHODS AND RESULTS: Generic E. coli were recovered from hides, carcasses, beef trimmings, conveyers and ground beef during the summer of 2001 (750 isolates) and winter of 2002 (500 isolates). The isolates were characterized by Random Amplification of Polymorphic DNA (RAPD). The numbers of E. coli recovered from dressed carcasses were less than the numbers recovered from hides. The numbers recovered from chilled carcasses were too few for meaningful analysis of the strains present on them but the numbers recovered from trimmings and ground beef were larger. The RAPD patterns showed that the majority of isolates from hides, carcasses, beef trimmings, conveyers and ground beef were of similar RAPD types, but a few unique RAPD types were recovered from only one of those sources. The E. coli populations present on the hides of incoming animals and in the beef-processing environment were highly diverse. Randomly selected E. coli isolates from each of the five sources were further characterized by pulsed-field gel electrophoresis (PFGE). Most genotypes of E. coli defined by PFGE corresponded to the E. coli types defined by RAPD. CONCLUSIONS: The hides of the incoming animals appeared to be only one of the sources of the E. coli on trimmings and in ground beef, as additional sources were apparently present in equipment used for carcass breaking. SIGNIFICANCE AND IMPACT OF THE STUDY: This study indicates that hazardous microbiological contamination of meat may occur after the dressing of carcasses at commercial beef-packing plants, which suggests that attention should be given to the control of the contamination of meat during carcass breaking as well as during the dressing of carcasses.     

Genetic diversity of Escherichia coli isolated from urban rivers and beach water

Repetitive element anchored PCR was used to evaluate the genetic profiles of Escherichia coli isolated from surface water contaminated with urban stormwater, sanitary sewage, and gull feces to determine if strains found in environmental samples reflect the strain composition of E. coli obtained from host sources. Overall, there was less diversity in isolates collected from river and beach sites than with isolates obtained from human and nonhuman sources. Unique strain types comprised 28.8, 29.2, and 15.0% of the isolate data sets recovered from stormwater, river water, and beach water, respectively. In contrast, 50.4% of gull isolates and 41.2% of sewage isolates were unique strain types. River water, which is expected to contain E. coli strains from many diffuse sources of nonpoint source pollution, contained strains most closely associated with other river water isolates that were collected at different sites or on different days. However, river sites impacted by sewage discharge had approximately 20% more strains similar to sewage isolates than did sites impacted by stormwater alone. Beach sites with known gull fecal contamination contained E. coli most similar to other beach isolates rather than gull isolates collected at these same sites, indicating underrepresentation of possible gull strains. These results suggest large numbers of strains are needed to represent contributing host sources within a geographical location. Additionally, environmental survival may influence the composition of strains that can be recovered from contaminated waters. Understanding the ecology of indicator bacteria is important when interpreting fecal pollution assessments and developing source detection methodology.     

Characterization of Escherichia coli populations from gulls, landfill trash, and wastewater using ribotyping

Due to their opportunistic and gregarious nature, gulls may be important reservoirs and vectors for anthropogenically derived fecal pathogens in coastal areas. We used ribotyping, a genotypic bacterial source tracking method, to compare populations of Escherichia coli among herring gulls Larus argentatus, great black-backed gulls L. marinus, wastewater, and landfill trash in New Hampshire and Maine, USA. Concentrations of E. coli in gull feces varied widely among individuals, but were generally high (6.0 x 10(1) to 2.5 x 10(9) g(-1) wet weight). Of 39 E. coli isolates from L. argentatus, 67% had banding patterns that were > or = 90% similar to those from wastewater and trash, whereas only 39% of 36 L. marinus isolates exhibited > or = 90% similarity to these sources. Strains of E. coli from gulls matched (> or = 90% similarity) more strains from wastewater (39% matching) than from trash (15% matching). E. coli isolates from L. marinus feces exhibited a greater diversity of banding patterns than did isolates from L. argentatus. There were more unique E. coli banding patterns in trash samples than in wastewater, and higher diversity indices in the former compared to the latter. These findings suggest that both species of gulls, especially L. argentatus, obtain fecal bacteria from wastewater and landfill trash, which they may transport to recreational beaches and waters. Our results also indicate that E. coli populations may vary widely between gull species, and between the anthropogenic habitats that they frequent, i.e. landfills and wastewater treatment facilities.     

Molecular epidemiological analysis of bloodstream isolates of Candida albicans from a university hospital over a five-year period

We assessed the genetic relations and epidemiological links among bloodstream isolates of Candida albicans, which were obtained from a university hospital over a period of five years. The 54 bloodstream isolates from the 38 patients yielded 14 different karyotypes, 29 different patterns after digestion with SfiI (REAG-S), and 31 different patterns after digestion with BssHII (REAG-B) when analyzed using three different pulsed-field gel electrophoresis (PFGE) typing methods. In 11 patients with serial bloodstream isolates, all strains from each patient had the same PFGE pattern. The dendrograms for all of the strains revealed that the distribution of similarity values ranged from 0.70 to 1.0 in the REAG-S patterns, and from 0.35 to 1.0 in the REAG-B patterns. Overall, the combination of the three different PFGE methods identified 31 distinct types, reflecting the results obtained using the REAG-B alone different. different Five PFGE types were shared among 22 isolates from 12 patients. These types of strains were more frequently associated with central venous catheter-related fungemia than the other 26 type strains (92% versus 31%; P < 0.005). Of five PFGE types, four isolates were determined to be epidemiologically related: each of these types was primarily from two or three patients who had been hospitalized concurrently within the same intensive care unit. Our results suggest that the REAG-B constitutes perhaps the most useful PFGE method for investigating C. albicans candidemia and also shows that a relatively high proportion of C. albicans candidemia may be associated with exogenous acquisition of clonal strains.     

Molecular fingerprinting of isolates of the genus Peptostreptococcus using rRNA genes from Escherichia coli and P. anaerobius.

Restriction fragment length polymorphisms (RFLPs) of rRNA genes were evaluated as a tool for intra- and interspecies differentiation of Peptostreptococcus isolates. RFLPs from a collection of 20 clinical isolates and five ATCC strains representing five Peptostreptococcus spp. (P. anaerobius, P. asaccharolyticus, P. magnus, P. micros and P. prevotii) were obtained by hybridization of Southern blots of HindIII- or EcoRI-digested genomic DNA with three probes: probe A, a 0.98 kb HindIII fragment with a partial 16S rRNA gene sequence from P. anaerobius ATCC 27337; probe B, cloned Escherichia coli rrnB operon in plasmid pKK3535; and probe C, E. coli 16S and 23S rRNA. The hybridization patterns varied, but all yielded RFLPs useful for both intra- and inter-species differentiation. RFLPs of P. asaccharolyticus clinical isolates were closely related to each other and differed significantly from those of the ATCC type strains. The profiles of P. prevotii differed from those of the other four species studied, and based on the HindIII- and EcoRI-generated RFLPs, the strains in this species are more heterogeneous than the other four species studied.     

Genome type analysis of adenovirus type 4 isolates, recently obtained from clinically different syndromes in some areas in Japan

DNA cleavage analyses with EcoRI, HindIII and BamHI were carried out to investigate genome types of recent Ad 4 isolates obtained from acute respiratory disease (8 strains), and ocular disease (11 strains) in Japan. DNA cleavage patterns of all 19 isolates studied were identical regardless of whether they were recovered from respiratory tract or conjunctiva, but were distinct from that of the prototype strain.     

Typings of Campylobacter jejuni isolated from patients of 2 outbreak cases by genotypic and phenotypic methods

We compared Campylobacter jejuni strains isolated from the patient stools associated with two food-borne diarrheal outbreak cases by the serotypic methods (Lior and Penner systems) and the genotypic methods (restriction fragment length polymorphism (RFLP) of flaA gene and pulsed-field gel electrophoresis (PFGE)). Fla-RFLP was based on the digestion of 410 bp DNA fragment by MboI restriction enzyme amplified from a 5' portion of C. jejuni flaA gene. Six distinctive fla-RFLP patterns were identified by examining 29 serotype reference strains and 58 strains isolated from the patients infected with C. jejuni independently. In the first outbreak case, 4 isolates were shown to be the same patterns each other by the fla-RFLP and PFGE, and by the Lior serotyping, except the Penner system that serotyped into 2 distinct types. On the other hand, in the second case, out of 10 isolates, 5 isolates were identical by the both genotypic and the both serotypic methods, and 4 isolates were not differentiated by the fla-RFLP and Penner system, but were separated into 4 types by PFGE in a little difference. The rest isolate was completely different from the other isolates by the all of methods used now. The findings suggest that the second case occurred by the infection of at least 3 different strains of C. jejuni.     

Effect of incubation temperature on isolation of Campylobacter jejuni genotypes from foodstuffs enriched in Preston broth

Preston broth and agar incubated at either 37 or 42 degrees C have been widely used to isolate campylobacters from foodstuffs. The consequences of using either incubation temperature were investigated. Retail packs of raw chicken (n = 24) and raw lamb liver (n = 30) were purchased. Samples were incubated in Preston broth at 37 and 42 degrees C and then streaked onto Preston agar and incubated as before. Two Campylobacter isolates per treatment were characterized. Poultry isolates were genotyped by random amplification of polymorphic DNA (RAPD), pulsed-field gel electrophoresis (PFGE), and flagellin PCR-restriction fragment length polymorphism, and lamb isolates were genotyped by RAPD only. In total, 96% of the poultry and 73% of the lamb samples yielded campylobacters. The lamb isolates were all Campylobacter jejuni, as were 96% of the poultry isolates, with the remainder being Campylobacter lari. The incubation temperature had no significant effect on the number of positive samples or on the species isolated. However, genotyping of the C. jejuni isolates revealed profound differences in the types obtained. Overall (from poultry and lamb), the use of a single incubation temperature, 37 degrees C, gave 56% of the total number of RAPD C. jejuni genotypes, and hence, 44% remained undetected. The effect was especially marked in the poultry samples, where incubation at 37 degrees C gave 47% of the PFGE genotypes but 53% were exclusively recovered after incubation at 42 degrees C. Thus, the incubation temperature of Preston media selects for certain genotypes of C. jejuni, and to detect the widest range, samples should be incubated at both 37 and 42 degrees C. Conversely, genotyping results arising from the use of a single incubation temperature should be interpreted with caution.     

Silver-resistant Enterobacteriaceae from hospital patients

The inclusion of agar medium containing 0.5 mM AgNO3 in the hospital laboratory replicating system for routine antibiotic-susceptibility determinations resulted in identification of species of Enterobacteriaceae (Escherichia coli, Enterobacter cloacae, Klebsiella pneumoniae, Proteus mirabilis, and Citrobacter freundii) with silver resistance. Since the study began in October, 1975, 11 in-hospital patients receiving silver sulfadiazine for burn wound prophylaxis have yielded silver-resistant bacteria from their infected burns. During this treatment routine burn-site cultures from these patients yielded 230 isolates of Enterobacteriaceae, including 211 which were sulfonamide-resistant, 97 of which were also silver-resistant, and 38 of which were untested for silver resistance. Seven silver-resistant but sulfonamide-sensitive isolates were incidentally recovered from respiratory specimens from four nonburn patients with silver tracheostomy tubes, one silver-resistant sulfonamide-sensitive isolate was recovered from a small infected burn on the foot of an Emergency Room patient. Previous treatment of this burn was unknown. Representative AgNO3-resistant E. coli isolates from four patients were serologically untypable. Serotyping of representative isolates of K. pneumoniae showed a diversity of types except from two patients who had been in the same ward at the same time.     

Genetic diversity and antibiotic resistance patterns in a campylobacter population isolated from poultry farms in Switzerland

The diversity and genetic interrelation of Campylobacter jejuni and C. coli isolated from Swiss poultry were assessed by three independent typing methods. Samples were derived prior to slaughter from 100 randomly selected flocks (five birds per flock) raised on three different farm types. The observed flock prevalence was 54% in total, with 50% for conventional and 69% for free-range farms. Birds held on farms with a confined roaming area had the lowest prevalence of 37%. Campylobacter isolates were characterized by amplified fragment length polymorphism (AFLP), restriction fragment length polymorphism of flaA PCR fragments (flaA-RFLP), and disk diffusion testing for eight antimicrobial agents that are commonly used in veterinary or human medicine in Switzerland. Analysis of the genotypic results indicates that the Campylobacter population in Swiss poultry is genetically highly diverse. Nevertheless, occasionally, isolates with identical or nearly identical characteristics were isolated from different farms or farm types in different locations. Genetic typing by AFLP and flaA-RFLP was found to be complementary. The majority of isolates (67%) were susceptible to all tested antibiotics; however, single, double, and triple resistances were observed in 7%, 23%, and 2% of the strains, respectively. There was no correlation between genotype and antibiotic resistance. Surprisingly, sulfonamide resistance was frequently found together with streptomycin resistance. Our findings illustrate the results of common genetic exchange in the studied bacterial population.     

Campylobacter jejuni isolated from retail poultry meat, bovine feces and bile, and human diarrheal samples in Japan: comparison of serotypes and genotypes

To determine the significance of poultry and bovine as infectious sources of Campylobacter jejuni in Japan, the serotype distribution and pulsed-field gel electrophoresis (PFGE) patterns of poultry and bovine isolates were compared with those of isolates from patients with diarrhea in Akita (Japan). Serotypes O:2 and O:4-complex were common in human, poultry, and bovine isolates, and serotype O:23,36,53 was common in human and bovine isolates. SmaI PFGE patterns of isolates belonging to these serotypes were generated. Eight PFGE patterns were shared by poultry and human isolates and three patterns were shared by human and bovine isolates. Further analysis of the isolates having the same SmaI PFGE pattern by KpnI PFGE confirmed that four patterns and two patterns were still shared by poultry and human isolates, and bovine and human isolates, respectively. Thus, serotypic and genotypic data indicated a possible link between sporadic human campylobacteriosis and C. jejuni from retail poultry and bovine bile and feces, suggesting that bovine serves as an infectious source of C. jejuni in Japan, as is observed in other countries.     

Campylobacter jejuni infection within a laboratory animal production unit

A conventional laboratory animal production unit in which rats, mice, guineapigs and rabbits were bred in one building and cats maintained in a separate, but adjacent area was examined for the presence of intestinal thermophilic Campylobacter spp. Campylobacter jejuni was recovered from 18.84% of 552 animals. The infection rate was highest amongst the cats (51.7%), with rats being the second most commonly infected (23.2%), whereas only 7.7% of guineapigs and a single rabbit (1%) were positive. Campylobacter-like organisms were cultured from 10% of the mice, but these bacteria failed to grow on subsequent subculturing. By using bacterial restriction endonuclease DNA analysis (BRENDA), a single type of C. jejuni was identified from all isolates recovered from the rats, guineapigs and a rabbit, suggesting a common source of infection. In contrast, there were 5 different BRENDA patterns derived from cat isolates. No isolates of C. jejuni were obtained from humans working within the unit or from animal bedding or the immediate environment, although it was suggested that the organism may have entered and spread within the unit from sawdust.     

Estimation of ribosomal RNA operon (rrn) copy number in Acinetobacter isolates and potential of patterns of rrn operon-containing fragments for typing strains of members of this genus

The copy number of the rrn operon in 70 strains of Acinetobacter including the type strains of almost all the genomic species with validated names was estimated after digestion of their genomic DNA by the restriction enzymes BglII and PstI, and Southern blotting. Copy number estimates varied between and among species, with between 3 and 7 rrn operon copies detected. Copy number estimates obtained from the same strain with the two enzymes sometimes varied. BglII generated RFLP patterns of the rrn containing fragments obtained from Southern blots after agarose gel electrophoresis were examined for their value in identifying Acinetobacter isolates. This method was very reproducible with the same fragment pattern always generated from the same isolate on repeated analysis. Often multiple strains of the same genomic species gave identical or very similar patterns (e.g. Acinetobacter baylyi), clustering closest together on the dendrogram generated after numerical analysis of these patterns. However, with some, like BG5 and BG8, the patterns derived from the different strains, some of which had been placed in this genomic species from DNA:DNA hybridization data, varied considerably to each other and to the type strain. Little similarity was seen when relationships between these strains based on these patterns were compared to those using DNA:DNA hybridization data. Often these patterns could be used to question earlier identification of strains using phenotypic characters. Thus, strain AB82 thought to belong to genomic species 5 gave an identical pattern to A. bouvetii(T) (DSM 14964). In some cases this pattern analysis suggested that novel species of Acinetobacter might exist among the strains examined.