Identification of Two Distinct Hybridization Groups in the Genus Hafnia by 16S rRNA Gene Sequencing and Phenotypic Methods
http://www.100md.com
微生物临床杂志 2005年第7期
Microbial Diseases Laboratory, California Department of Health Services, Richmond, California 94804
ABSTRACT
A collection of 52 strains belonging to the Hafnia alvei complex were subjected to molecular (16S rRNA gene sequencing) and biochemical analysis. Based upon 16S rRNA gene sequencing results, two genetic groups were identified which correspond with previously recognized DNA hybridization group 1 (ATCC 13337T and ATCC 29926; n = 23) and DNA hybridization group 2 (ATCC 29927; n = 29). Of 46 biochemical tests used to characterize hafniae, 19 reactions (41%) yielded variable results. Of these 19 tests, 6 were determined to have discriminatory value in the separation of DNA groups 1 and 2, with malonate utilization found to be the most differential test. Test results of malonate utilization alone correctly assigned 90% of Hafnia isolates to their correct DNA group.
INTRODUCTION
The genus Hafnia presently consists of two or more distinct species (genomospecies) that are currently identified as a single taxon, H. alvei. DNA relatedness studies conducted by Steigerwalt et al. (12) in 1976 involving the genera Enterobacter and Serratia identified several strains of H. alvei ("Enterobacter hafniae") that exhibited widely divergent DNA binding ratios and degrees of mismatch (divergence) when hybridized with DNA from reference strain E. hafniae 4360-76. These results, although limited in scope by the number of strains studied, provided the first direct evidence for species diversity within the genus. Although definitive taxonomic studies regarding this species were not subsequently published, Brenner and colleagues at the Centers for Disease Control and Prevention (CDC) confirmed the existence of at least two distinct relatedness (hybridization) groups (HGs) within H. alvei (2, 3). Reference strains for these HGs include ATCC 13337T and ATCC 29926 (CDC 5632-72) for HG 1 (H. alvei sensu stricto) and ATCC 29927 (CDC 4510-73) for HG 2 (unnamed) (4).
Over the past several years, there has been renewed interest in the taxonomy of hafniae as causative agents of extraintestinal disease (6, 11) and their reputed association with bacterial gastroenteritis (7). One study found that 16S rRNA gene sequencing could be successfully used to distinguish Hafnia genomospecies (8), while a subsequent investigation achieved similar results using multilocus enzyme electrophoresis (10). This latter Australian survey also concluded that the two genomospecies differed in their genetic structure and host distribution based upon an analysis of strains recovered from a variety of mammals, reptiles, fish, amphibians, insects, and aquatic sources (10).
HG 1 contains the type strain, ATCC 13337T (Stuart's type 32011), and by definition constitutes H. alvei sensu stricto. HG 2, also currently identified as H. alvei, remains unnamed, since previous studies have failed to identify phenotypic characteristics useful in distinguishing these two species from each other (2, 3, 4). It is also unclear what the distribution and frequency of these two hybridization groups are in clinical specimens. Greipsson and Priest (5) published the last major numerical taxonomy study on Hafnia alvei in 1983 that predates recent advances in phylogenetic analysis. The goals of the present investigation were threefold: (i) to biochemically and genetically characterize a large collection of hafniae to determine their relative frequency and distribution in clinical samples, (ii) to determine whether or not phenotypic traits could be found that clearly separate each genomospecies, and finally (iii) to recharacterize the biochemical properties that presently define the genus.
MATERIALS AND METHODS
Bacterial strains. Fifty-two strains of Hafnia alvei were studied in the present investigation. These strains were initially isolated between 1970 and 2002 and were characterized by previously described biochemical tests using conventional methods (8). All 52 strains met the general definition for inclusion in the genus Hafnia, being lysine- and ornithine decarboxylase positive, arginine dihydrolase negative, and Voges Proskauer positive, as well as having growth in KCN broth and the inability to produce acid from the fermentation of D-sorbitol, raffinose, melibiose, D-adonitol, and m-inositol. A few strains deviated from one of the idealized phenotypes (e.g., raffinose positive) that is discussed within the text. Reference strains for 16S rRNA gene sequencing studies included ATCC 13337T (HG 1), ATCC 29926 (HG 1), and ATCC 29927 (HG 2). Other reference strains included in the present study were CCUG 429 (clinical isolate, source unknown), D46-NF (feces), and D67-NF (feces), the later two courtesy of A. Ismaili (University of Toronto). Of the remaining 46 strains, 25 were of fecal origin, 16 were from unknown sources, 2 were from the respiratory tract, 1 from urine, and 2 from animals (mole, wallaby).
16S rRNA gene sequencing. Partial 16S rRNA gene sequencing of all 52 H. alvei strains was performed using an Applied Biosystems ABI 377 sequencer, and sequence assembly and analysis were accomplished with MicroSeq Analysis software (8). Sequence results were expressed as percent divergence from either ATCC 13337T (H. alvei sensu stricto, DNA HG1) or ATCC 29927 (DNA HG 2). Individual H. alvei isolates were assigned to either DNA group if there was <0.5% 16S rDNA gene sequence divergence between the test isolate and the reference strain.
Biochemical characterization. All H. alvei strains were characterized in a conventional test format for 46 selected phenotypic or biochemical traits according to previously published procedures (1, 9). Motility and ONPG (o-nitrophenyl--D-galactopyranoside) tests were read at 24 h only, while most other conventional tests, including carbohydrate fermentation reactions, were read at 48 h. Decarboxylase and dihydrolase reactions, enzymatic degradation tests (DNase, lipase), and pigmentation reactions were read at 96 h. A selected subset of representative Hafnia strains (n = 20) from both DNA groups were additionally screened for the following activities: elaboration of glutamate decarboxylase, production of siderophores and alkylsulfatase, and enzymatic degradation of chitin, mucin, lecithin, elastin, tyrosine, RNA, casein, and staphylococcal cell wall. The significance of test results was determined from contingency tables using chi squared.
RESULTS
The 16S rRNA gene sequence results for 52 H. alvei strains can be seen in Fig. 1. Each data point in Fig. 1 represents one or more strains with identical sequence divergence values between ATCC 13337T and ATCC 29927. Overall, 29 strains clustered within DNA group 2; all 29 of these strains demonstrated 0.19% sequence divergence from the reference strain, ATCC 29927. Similarly, DNA group 1 contained 23 strains. With the exception of a single DNA group 1 isolate, all strains exhibited 0.09% sequence divergence to the type strain, ATCC 13337T. The singular exception, strain 6109-4-72, recovered from the stool of a 7-month-old female suspected of having salmonellosis in 1972, exhibited 0.47% and 1.04% 16S rRNA gene sequence divergence to reference strains for DNA groups 1 and 2, respectively. Twelve DNA group 2 strains (41%) and 5 DNA group 1 strains (22%) had identical 16S rRNA signatures to ATCC 29927 and ATCC 13337T, respectively.
All 52 Hafnia strains were catalase and Voges Proskauer positive, produced nitrate reductase, were lysine and ornithine decarboxylase positive, and produced acid from the fermentation of maltose, D-mannitol, and D-xylose. Conversely, all hafniae were nonpigmented and oxidase negative and failed to produce indole, phenylpyruvic acid (phenylalanine deaminase-negative), lipase or DNase, or to degrade gelatin, mucin or polypectate, or acid from the fermentation of adonitol, amygdalin, D-arabitol, dulcitol, m-inositol, lactose, melibiose, D-sorbitol, sucrose, and -methyl-D-glucoside. The remaining 19 biochemical characteristics that were variable for the genus are presented in Table 1. All Hafnia strains tested produced siderophores when assayed on Chrome Azurol S agar (9). Alkysulfatase, stapholysin, tyrosinase, protease, elastase, RNase, lecithinase, mucinase, and chitinase activities were not detected.
We then determined whether any of the traits producing variable test results listed in Table 1 were useful in distinguishing between Hafnia DNA group 1 and group 2 (Table 2). Six tests were found to have at least some discriminatory value in distinguishing between these two taxa. For five of these tests the differences in positivity rates between groups 1 and 2 were statistically significant. Two of these tests (malonate utilization, fermentation of D-arabinose) were of major potential interest in displaying high positivity rates for one DNA group and low positivity rates for the other. Malonate utilization was the single best differential test, with all but one DNA group 1 strain being positive and only 14% of DNA group 2 strains expressing this phenotype. For fermentation of D-arabinose, most group 1 strains were negative at 48 h; however, prolonged incubation of isolates to 96 h resulted in an additional eight isolates being positive (delayed) that raised the percentage to 43%.
We then looked at five strains that could not be assigned to their correct DNA groups based upon malonate utilization alone (Table 3). The single malonate-negative group 1 isolate was also D-arabinose negative. However, this strain was also citrate, esculin, and salicin positive, the later two characteristics only associated with DNA group 1 strains (Table 2). Similarly, Group 2 strains 97A-3582 and HA-6 had typical profiles for DNA group 2 (other than malonate). Strain 84A-1383 was both citrate and D-arabinose positive but was also raffinose positive, a marker found only in DNA group 2 strains, although this association was not statistically significant (P < 0.10). Strain 2199-8-71 would be difficult to place in the correct DNA group, being malonate, citrate, and D-arabinose positive.
Several studies have reported that hafniae can be biotyped based upon production of a -glucosidase active against aglycone compounds, such as salicin, arbutin, and esculin, and fermentation of D-arabinose (2, 3, 8). Since no aglycone-utilizing strains were found among DNA group 2 isolates, no biogroups were detected. However, DNA group 1 strains could be broken down into four distinct biogroups (Table 4). Overall, 87% of DNA group 1 strains fell into one of two biogroups (biogroups 1 and 2); these two groups differing only in their ability to attack aglycone compounds (-glucosidase). The remaining two groups (13%) were rarely seen and were composed of two strains attacking all three compounds (biogroup 3) and one strain (ATCC 29926) that produced acid from D-arabinose fermentation but was -glucosidase negative (biogroup 4).
DISCUSSION
The results of the present investigation support previous preliminary findings from our laboratory (8) that two DNA groups within the H. alvei complex can be distinguished by 16S rRNA gene sequencing (Fig. 1). This is supported by the separation of DNA hybridization group 1 reference strains ATCC 13337T and ATCC 29926 into one 16S cluster and DNA hybridization group 2 containing strain ATCC 29927 into another. As can be seen from Fig. 1, there is clear separation of the 52 strains tested in this study into separate clusters based upon 16S rRNA gene sequence analysis. This association is further supported by a recent phylogenetic study employing multilocus enzyme electrophoresis by Okada and Gordon (10). In that study, 161 H. alvei strains clustered into two genetically distinct groups, one DNA group containing ATCC 13337T and ATCC 29926 while the other contained ATCC 29927.
Based upon 16S rRNA gene sequencing results, both DNA group 1 (H. alvei sensu stricto) and DNA group 2 (unnamed) are commonly occurring species in clinical samples with almost an identical distribution as originally suggested by Brenner (3). DNA group 2 (unnamed) appears to be slightly more prevalent in clinical material, but this point awaits further clarification using a large collection of randomly chosen isolates obtained from many different medical centers. Since a majority of the clinical samples in the present investigation of known origin were from the gastrointestinal tract, we cannot at present determine whether invasive isolates (blood) are equally or preferentially associated with one or the other DNA group. Since extraintestinal H. alvei infections are relatively uncommon, the importance of each DNA group in systemic infections must await the collection of invasive isolates over a protracted period of time. Preliminary evidence from Australia suggests that there may be important ecologic and host differences in the two DNA groups among birds and mammals there (10).
Biochemical characterization of both DNA groups yielded several potentially useful differential tests (Table 2). The single best test was malonate, which correctly assigned 90% of all hafniae tested to their correct DNA group as determined by 16S rRNA gene sequencing. Malonate was not one of the potentially discriminatory tests proposed by Brenner (3) but was highlighted as a differential test in a previous study from our laboratory (8). Okada and Gordon (10) have since confirmed the value of malonate utilization as a highly discriminatory test for their two genetic groups. Of the remaining 10% of strains (n = 5), four could be assigned to their correct taxa by reviewing the overall pattern of reactions for six discriminatory tests listed in Table 3. This would then bring the number of strains that could be correctly assigned to one of two DNA groups to 98%.
The cumulative results from the present study and several previous investigations (2, 3, 8, 10) suggest that it is possible to develop a biochemical scheme to phenotypically identify Hafnia to one of two known DNA groups within the complex. DNA hybridization studies should be undertaken, however, to confirm the results of 16S rRNA gene sequencing. If confirmed, then infrequent atypical strains that cannot be unambiguously assigned to a given taxon by biochemical analysis can be determined to species level using 16S rRNA gene sequencing. Further studies with a larger collection of strains are needed to help define the frequency, clinical distribution, and disease spectrum of both Hafnia species.
REFERENCES
Abbott, S. L., W. K. W. Cheung, and J. M. Janda. 2003. The genus Aeromonas: biochemical characteristics, atypical reactions, and phenotypic identification schemes. J. Clin. Microbiol. 41:2348-2357.
Brenner, D. J. 1981. Introduction to the family Enterobacteriaceae, p. 1105-1127. In M. P. Starr, H. Stolp, H. G. Trüper, A. Balows, and H. G. Schlegel (ed.), The prokaryotes: a handbook on habitats, isolation, and identification of bacteria, vol. II. Springer-Verlag, New York, N.Y.
Brenner, D. J. 11 May 1999, revision date. Introduction to the family Enterobacteriaceae. In A. Balows, H. G. Trüper, M. Dworkin, W. Harder, and K.-H. Schleifer (ed.), The prokaryotes: a handbook on the biology of bacteria ecophysiology, isolation, identification, applications, 2nd ed., release 3.0. [Online.] Springer-Verlag, New York, N.Y. http://141.150.157.117:8080/prokPUB/index.htm.
Gelev, I., E. Gelev, A. G. Steigerwalt, G. P. Carter, and D. J. Brenner. 1990. Identification of the bacterium associated with haemorrhagic septicaemia in rainbow trout as Hafnia alvei. Res. Microbiol. 141:573-576.
Greipsson, S., and F. G. Priest. 1983. Numerical taxonomy of Hafnia alvei. Int. J. Syst. Bacteriol. 33:470-475.
Günthard, H., and A. Pennekamp. 1996. Clinical significance of extraintestinal Hafnia alvei isolates from 61 patients and review of the literature. Clin. Infect. Dis. 22:1040-1045.
Janda, J. M., S. L. Abbott, and M. J. Albert. 1999. Prototypal diarrheagenic strains of Hafnia alvei are actually members of the genus Escherichia. J. Clin. Microbiol. 37:2399-2401.
Janda, J. M., S. L. Abbott, S. Khashe, and W. Probert. 2002. Phenotypic and genotypic properties of the genus Hafnia. J. Med. Microbiol. 51:575-580.
Janda, J. M., S. L. Abbott, S. Khashe, and T. Robin. 1996. Biochemical investigations of biogroups and subspecies of Morganella morganii. J. Clin. Microbiol. 34:108-113.
Okada, S., and D. M. Gordon. 2003. Genetic and ecological structure of Hafnia alvei in Australia. Syst. Appl. Microbiol. 26:585-594.
Ramos, A., and D. Dámaso. 2000. Extraintestinal infection due to Hafnia alvei. Eur. J. Clin. Microbiol. Infect. Dis. 19:708-710.
Steigerwalt, A. G., G. R. Fanning, M. A. Fife-Asbury, and D. J. Brenner. 1976. DNA relatedness among species of Enterobacter and Serratia. Can. J. Microbiol. 22:121-137.(J. Michael Janda, Sharon )
ABSTRACT
A collection of 52 strains belonging to the Hafnia alvei complex were subjected to molecular (16S rRNA gene sequencing) and biochemical analysis. Based upon 16S rRNA gene sequencing results, two genetic groups were identified which correspond with previously recognized DNA hybridization group 1 (ATCC 13337T and ATCC 29926; n = 23) and DNA hybridization group 2 (ATCC 29927; n = 29). Of 46 biochemical tests used to characterize hafniae, 19 reactions (41%) yielded variable results. Of these 19 tests, 6 were determined to have discriminatory value in the separation of DNA groups 1 and 2, with malonate utilization found to be the most differential test. Test results of malonate utilization alone correctly assigned 90% of Hafnia isolates to their correct DNA group.
INTRODUCTION
The genus Hafnia presently consists of two or more distinct species (genomospecies) that are currently identified as a single taxon, H. alvei. DNA relatedness studies conducted by Steigerwalt et al. (12) in 1976 involving the genera Enterobacter and Serratia identified several strains of H. alvei ("Enterobacter hafniae") that exhibited widely divergent DNA binding ratios and degrees of mismatch (divergence) when hybridized with DNA from reference strain E. hafniae 4360-76. These results, although limited in scope by the number of strains studied, provided the first direct evidence for species diversity within the genus. Although definitive taxonomic studies regarding this species were not subsequently published, Brenner and colleagues at the Centers for Disease Control and Prevention (CDC) confirmed the existence of at least two distinct relatedness (hybridization) groups (HGs) within H. alvei (2, 3). Reference strains for these HGs include ATCC 13337T and ATCC 29926 (CDC 5632-72) for HG 1 (H. alvei sensu stricto) and ATCC 29927 (CDC 4510-73) for HG 2 (unnamed) (4).
Over the past several years, there has been renewed interest in the taxonomy of hafniae as causative agents of extraintestinal disease (6, 11) and their reputed association with bacterial gastroenteritis (7). One study found that 16S rRNA gene sequencing could be successfully used to distinguish Hafnia genomospecies (8), while a subsequent investigation achieved similar results using multilocus enzyme electrophoresis (10). This latter Australian survey also concluded that the two genomospecies differed in their genetic structure and host distribution based upon an analysis of strains recovered from a variety of mammals, reptiles, fish, amphibians, insects, and aquatic sources (10).
HG 1 contains the type strain, ATCC 13337T (Stuart's type 32011), and by definition constitutes H. alvei sensu stricto. HG 2, also currently identified as H. alvei, remains unnamed, since previous studies have failed to identify phenotypic characteristics useful in distinguishing these two species from each other (2, 3, 4). It is also unclear what the distribution and frequency of these two hybridization groups are in clinical specimens. Greipsson and Priest (5) published the last major numerical taxonomy study on Hafnia alvei in 1983 that predates recent advances in phylogenetic analysis. The goals of the present investigation were threefold: (i) to biochemically and genetically characterize a large collection of hafniae to determine their relative frequency and distribution in clinical samples, (ii) to determine whether or not phenotypic traits could be found that clearly separate each genomospecies, and finally (iii) to recharacterize the biochemical properties that presently define the genus.
MATERIALS AND METHODS
Bacterial strains. Fifty-two strains of Hafnia alvei were studied in the present investigation. These strains were initially isolated between 1970 and 2002 and were characterized by previously described biochemical tests using conventional methods (8). All 52 strains met the general definition for inclusion in the genus Hafnia, being lysine- and ornithine decarboxylase positive, arginine dihydrolase negative, and Voges Proskauer positive, as well as having growth in KCN broth and the inability to produce acid from the fermentation of D-sorbitol, raffinose, melibiose, D-adonitol, and m-inositol. A few strains deviated from one of the idealized phenotypes (e.g., raffinose positive) that is discussed within the text. Reference strains for 16S rRNA gene sequencing studies included ATCC 13337T (HG 1), ATCC 29926 (HG 1), and ATCC 29927 (HG 2). Other reference strains included in the present study were CCUG 429 (clinical isolate, source unknown), D46-NF (feces), and D67-NF (feces), the later two courtesy of A. Ismaili (University of Toronto). Of the remaining 46 strains, 25 were of fecal origin, 16 were from unknown sources, 2 were from the respiratory tract, 1 from urine, and 2 from animals (mole, wallaby).
16S rRNA gene sequencing. Partial 16S rRNA gene sequencing of all 52 H. alvei strains was performed using an Applied Biosystems ABI 377 sequencer, and sequence assembly and analysis were accomplished with MicroSeq Analysis software (8). Sequence results were expressed as percent divergence from either ATCC 13337T (H. alvei sensu stricto, DNA HG1) or ATCC 29927 (DNA HG 2). Individual H. alvei isolates were assigned to either DNA group if there was <0.5% 16S rDNA gene sequence divergence between the test isolate and the reference strain.
Biochemical characterization. All H. alvei strains were characterized in a conventional test format for 46 selected phenotypic or biochemical traits according to previously published procedures (1, 9). Motility and ONPG (o-nitrophenyl--D-galactopyranoside) tests were read at 24 h only, while most other conventional tests, including carbohydrate fermentation reactions, were read at 48 h. Decarboxylase and dihydrolase reactions, enzymatic degradation tests (DNase, lipase), and pigmentation reactions were read at 96 h. A selected subset of representative Hafnia strains (n = 20) from both DNA groups were additionally screened for the following activities: elaboration of glutamate decarboxylase, production of siderophores and alkylsulfatase, and enzymatic degradation of chitin, mucin, lecithin, elastin, tyrosine, RNA, casein, and staphylococcal cell wall. The significance of test results was determined from contingency tables using chi squared.
RESULTS
The 16S rRNA gene sequence results for 52 H. alvei strains can be seen in Fig. 1. Each data point in Fig. 1 represents one or more strains with identical sequence divergence values between ATCC 13337T and ATCC 29927. Overall, 29 strains clustered within DNA group 2; all 29 of these strains demonstrated 0.19% sequence divergence from the reference strain, ATCC 29927. Similarly, DNA group 1 contained 23 strains. With the exception of a single DNA group 1 isolate, all strains exhibited 0.09% sequence divergence to the type strain, ATCC 13337T. The singular exception, strain 6109-4-72, recovered from the stool of a 7-month-old female suspected of having salmonellosis in 1972, exhibited 0.47% and 1.04% 16S rRNA gene sequence divergence to reference strains for DNA groups 1 and 2, respectively. Twelve DNA group 2 strains (41%) and 5 DNA group 1 strains (22%) had identical 16S rRNA signatures to ATCC 29927 and ATCC 13337T, respectively.
All 52 Hafnia strains were catalase and Voges Proskauer positive, produced nitrate reductase, were lysine and ornithine decarboxylase positive, and produced acid from the fermentation of maltose, D-mannitol, and D-xylose. Conversely, all hafniae were nonpigmented and oxidase negative and failed to produce indole, phenylpyruvic acid (phenylalanine deaminase-negative), lipase or DNase, or to degrade gelatin, mucin or polypectate, or acid from the fermentation of adonitol, amygdalin, D-arabitol, dulcitol, m-inositol, lactose, melibiose, D-sorbitol, sucrose, and -methyl-D-glucoside. The remaining 19 biochemical characteristics that were variable for the genus are presented in Table 1. All Hafnia strains tested produced siderophores when assayed on Chrome Azurol S agar (9). Alkysulfatase, stapholysin, tyrosinase, protease, elastase, RNase, lecithinase, mucinase, and chitinase activities were not detected.
We then determined whether any of the traits producing variable test results listed in Table 1 were useful in distinguishing between Hafnia DNA group 1 and group 2 (Table 2). Six tests were found to have at least some discriminatory value in distinguishing between these two taxa. For five of these tests the differences in positivity rates between groups 1 and 2 were statistically significant. Two of these tests (malonate utilization, fermentation of D-arabinose) were of major potential interest in displaying high positivity rates for one DNA group and low positivity rates for the other. Malonate utilization was the single best differential test, with all but one DNA group 1 strain being positive and only 14% of DNA group 2 strains expressing this phenotype. For fermentation of D-arabinose, most group 1 strains were negative at 48 h; however, prolonged incubation of isolates to 96 h resulted in an additional eight isolates being positive (delayed) that raised the percentage to 43%.
We then looked at five strains that could not be assigned to their correct DNA groups based upon malonate utilization alone (Table 3). The single malonate-negative group 1 isolate was also D-arabinose negative. However, this strain was also citrate, esculin, and salicin positive, the later two characteristics only associated with DNA group 1 strains (Table 2). Similarly, Group 2 strains 97A-3582 and HA-6 had typical profiles for DNA group 2 (other than malonate). Strain 84A-1383 was both citrate and D-arabinose positive but was also raffinose positive, a marker found only in DNA group 2 strains, although this association was not statistically significant (P < 0.10). Strain 2199-8-71 would be difficult to place in the correct DNA group, being malonate, citrate, and D-arabinose positive.
Several studies have reported that hafniae can be biotyped based upon production of a -glucosidase active against aglycone compounds, such as salicin, arbutin, and esculin, and fermentation of D-arabinose (2, 3, 8). Since no aglycone-utilizing strains were found among DNA group 2 isolates, no biogroups were detected. However, DNA group 1 strains could be broken down into four distinct biogroups (Table 4). Overall, 87% of DNA group 1 strains fell into one of two biogroups (biogroups 1 and 2); these two groups differing only in their ability to attack aglycone compounds (-glucosidase). The remaining two groups (13%) were rarely seen and were composed of two strains attacking all three compounds (biogroup 3) and one strain (ATCC 29926) that produced acid from D-arabinose fermentation but was -glucosidase negative (biogroup 4).
DISCUSSION
The results of the present investigation support previous preliminary findings from our laboratory (8) that two DNA groups within the H. alvei complex can be distinguished by 16S rRNA gene sequencing (Fig. 1). This is supported by the separation of DNA hybridization group 1 reference strains ATCC 13337T and ATCC 29926 into one 16S cluster and DNA hybridization group 2 containing strain ATCC 29927 into another. As can be seen from Fig. 1, there is clear separation of the 52 strains tested in this study into separate clusters based upon 16S rRNA gene sequence analysis. This association is further supported by a recent phylogenetic study employing multilocus enzyme electrophoresis by Okada and Gordon (10). In that study, 161 H. alvei strains clustered into two genetically distinct groups, one DNA group containing ATCC 13337T and ATCC 29926 while the other contained ATCC 29927.
Based upon 16S rRNA gene sequencing results, both DNA group 1 (H. alvei sensu stricto) and DNA group 2 (unnamed) are commonly occurring species in clinical samples with almost an identical distribution as originally suggested by Brenner (3). DNA group 2 (unnamed) appears to be slightly more prevalent in clinical material, but this point awaits further clarification using a large collection of randomly chosen isolates obtained from many different medical centers. Since a majority of the clinical samples in the present investigation of known origin were from the gastrointestinal tract, we cannot at present determine whether invasive isolates (blood) are equally or preferentially associated with one or the other DNA group. Since extraintestinal H. alvei infections are relatively uncommon, the importance of each DNA group in systemic infections must await the collection of invasive isolates over a protracted period of time. Preliminary evidence from Australia suggests that there may be important ecologic and host differences in the two DNA groups among birds and mammals there (10).
Biochemical characterization of both DNA groups yielded several potentially useful differential tests (Table 2). The single best test was malonate, which correctly assigned 90% of all hafniae tested to their correct DNA group as determined by 16S rRNA gene sequencing. Malonate was not one of the potentially discriminatory tests proposed by Brenner (3) but was highlighted as a differential test in a previous study from our laboratory (8). Okada and Gordon (10) have since confirmed the value of malonate utilization as a highly discriminatory test for their two genetic groups. Of the remaining 10% of strains (n = 5), four could be assigned to their correct taxa by reviewing the overall pattern of reactions for six discriminatory tests listed in Table 3. This would then bring the number of strains that could be correctly assigned to one of two DNA groups to 98%.
The cumulative results from the present study and several previous investigations (2, 3, 8, 10) suggest that it is possible to develop a biochemical scheme to phenotypically identify Hafnia to one of two known DNA groups within the complex. DNA hybridization studies should be undertaken, however, to confirm the results of 16S rRNA gene sequencing. If confirmed, then infrequent atypical strains that cannot be unambiguously assigned to a given taxon by biochemical analysis can be determined to species level using 16S rRNA gene sequencing. Further studies with a larger collection of strains are needed to help define the frequency, clinical distribution, and disease spectrum of both Hafnia species.
REFERENCES
Abbott, S. L., W. K. W. Cheung, and J. M. Janda. 2003. The genus Aeromonas: biochemical characteristics, atypical reactions, and phenotypic identification schemes. J. Clin. Microbiol. 41:2348-2357.
Brenner, D. J. 1981. Introduction to the family Enterobacteriaceae, p. 1105-1127. In M. P. Starr, H. Stolp, H. G. Trüper, A. Balows, and H. G. Schlegel (ed.), The prokaryotes: a handbook on habitats, isolation, and identification of bacteria, vol. II. Springer-Verlag, New York, N.Y.
Brenner, D. J. 11 May 1999, revision date. Introduction to the family Enterobacteriaceae. In A. Balows, H. G. Trüper, M. Dworkin, W. Harder, and K.-H. Schleifer (ed.), The prokaryotes: a handbook on the biology of bacteria ecophysiology, isolation, identification, applications, 2nd ed., release 3.0. [Online.] Springer-Verlag, New York, N.Y. http://141.150.157.117:8080/prokPUB/index.htm.
Gelev, I., E. Gelev, A. G. Steigerwalt, G. P. Carter, and D. J. Brenner. 1990. Identification of the bacterium associated with haemorrhagic septicaemia in rainbow trout as Hafnia alvei. Res. Microbiol. 141:573-576.
Greipsson, S., and F. G. Priest. 1983. Numerical taxonomy of Hafnia alvei. Int. J. Syst. Bacteriol. 33:470-475.
Günthard, H., and A. Pennekamp. 1996. Clinical significance of extraintestinal Hafnia alvei isolates from 61 patients and review of the literature. Clin. Infect. Dis. 22:1040-1045.
Janda, J. M., S. L. Abbott, and M. J. Albert. 1999. Prototypal diarrheagenic strains of Hafnia alvei are actually members of the genus Escherichia. J. Clin. Microbiol. 37:2399-2401.
Janda, J. M., S. L. Abbott, S. Khashe, and W. Probert. 2002. Phenotypic and genotypic properties of the genus Hafnia. J. Med. Microbiol. 51:575-580.
Janda, J. M., S. L. Abbott, S. Khashe, and T. Robin. 1996. Biochemical investigations of biogroups and subspecies of Morganella morganii. J. Clin. Microbiol. 34:108-113.
Okada, S., and D. M. Gordon. 2003. Genetic and ecological structure of Hafnia alvei in Australia. Syst. Appl. Microbiol. 26:585-594.
Ramos, A., and D. Dámaso. 2000. Extraintestinal infection due to Hafnia alvei. Eur. J. Clin. Microbiol. Infect. Dis. 19:708-710.
Steigerwalt, A. G., G. R. Fanning, M. A. Fife-Asbury, and D. J. Brenner. 1976. DNA relatedness among species of Enterobacter and Serratia. Can. J. Microbiol. 22:121-137.(J. Michael Janda, Sharon )