|Home | About | Journals | Submit | Contact Us | Français|
Genotypic and phenotypic analyses were carried out to clarify the taxonomic position of the naturally transformable Acinetobacter sp. strain ADP1. Transfer tDNA-PCR fingerprinting, 16S rRNA gene sequence analysis, and selective restriction fragment amplification (amplified fragment length polymorphism analysis) indicate that strain ADP1 and a second transformable strain, designated 93A2, are members of the newly described species Acinetobacter baylyi. Transformation assays demonstrate that the A. baylyi type strain B2T and two other originally identified members of the species (C5 and A7) also have the ability to undergo natural transformation at high frequencies, confirming that these five strains belong to a separate species of the genus Acinetobacter, characterized by the high transformability of its strains that have been cultured thus far.
Acinetobacter sp. strain ADP1's metabolic versatility (17, 26) combined with its ability to undergo natural genetic transformation at high frequencies have made it the most widely studied strain of the genus. However, due in large part to its lack of clinical significance, strain ADP1 has thus far not been included in phylogenetic studies and has yet to be given a species designation. The recent publication of the genomic sequence of ADP1 will no doubt bring increased attention to this important strain (2, 17). In light of this, and because of strain ADP1's importance as a model organism, it is important to establish its taxonomic position and to eliminate any confusion regarding the relatedness of this benign strain to the other members of the Acinetobacter genus, some of which are important nosocomial pathogens, including A. baumannii and the closely related unnamed species 3 and 13TU (4).
Due to its long history and widespread distribution, any attempt to understand the taxonomy of strain ADP1 first requires a review of its pedigree. Strain ADP1 is a derivative of the soil isolate Acinetobacter sp. strain BD4 (ATCC 33304) (12, 15). Strain BD4 forms mucoid colonies that make its use in the laboratory difficult (12). As a result, most laboratories have chosen to use a microencapsulated mutant of BD4 that was originally designated “strain BD413” (ATCC 33305) (14) but which has also become known as “strain ADP1” (22). Today, the designation “ADP1” is encountered most frequently, but “BD413” is also still in use. It should be noted that strain BD4 and all of its derivatives, including strain ADP1, were for many years labeled Acinetobacter calcoaceticus, a name given to all members of the genus prior to subdivision of the genus based on DNA-DNA hybridization data (6).
Although natural transformation has been reported for a small number of Acinetobacter strains other than ADP1, the transformation frequencies reported for those strains were at least 100- to 1,000-fold lower than those obtained with ADP1 (14, 15, 21). However, a second isolate, strain 93A2, was recently shown to possess a level of competence equal to that of ADP1 (27). Strain 93A2 was isolated from a natural stream, using kynurenine as its sole carbon source, and was included in the study by Baumann and colleagues which first brought taxonomic cohesiveness to the genus (3). 93A2's high level of genetic competence plus the level of nucleotide sequence similarity it was shown to share (98% to 100% identity) with strain ADP1 by sequence analysis of three specific genes (trpE, mutS, and the 16S rRNA gene) (27) prompted its inclusion in this study.
This report presents genotypic and phenotypic data demonstrating that strains ADP1 and 93A2 are members of the recently described species Acinetobacter baylyi (7). In addition, we show that the three original members of the A. baylyi species (7), including the type strain (strain B2T), are competent for natural transformation at frequencies equaling those of strain ADP1. These data demonstrate that members of the A. baylyi species share the ability to undergo natural transformation at high frequencies, a trait that sets them apart from all other described genomic species of the genus Acinetobacter.
Previously, some of us presented DNA-DNA hybridization data showing that BD413, a synonym of ADP1, did not belong to the genomic species 1 to 15 or to “A. venetianus” (9).
Table Table11 summarizes the data on the origins of the A. baylyi strains studied and the tests carried out for each strain. DNA-based genotyping methods have proven useful for identifying the taxonomic positions of unclassified Acinetobacter isolates (8-10, 24, 25). We used three of these techniques, transfer DNA (tDNA)-PCR fingerprinting, 16S rRNA gene sequence analysis, and amplified fragment length polymorphism (AFLP) analysis, in our initial attempts to determine whether strains ADP1 and 93A2 grouped within any of the described genomic species of the genus Acinetobacter or whether they were members of an as-yet undescribed species.
Using capillary electrophoresis, the length of the tDNA-intergenic spacers for strains ADP1, BD4, and 93A2 were determined as described previously (10). The patterns obtained for these three strains were identical, as would be expected for a wild-type strain and a mutant derivative, and their patterns also were identical to that of the type strain, A. baylyi B2T. The pattern is composed of amplified tDNA spacer fragments with lengths (standard deviations are in parentheses) of 90.5 (0.1), 103.9 (0.3), 117.8 (0.1), 125 (0.1), 199.5 (0.1), 209.8 (0.2), 234.6 (0.1), and 241.7 bp (0.1 bp). No other matches were observed between these strains and published tDNA-PCR data. A dendrogram (Fig. (Fig.1),1), based on the similarity between tDNA-PCR fingerprints as calculated by the differential base pairs algorithm (1), confirms the close relatedness of ADP1 and A. baylyi B2T and their separate position within the genus Acinetobacter.
The 16S rRNA gene nucleotide sequences for strains ADP1 and 93A2 were determined and analyzed using previously reported methods (23). The sequence determined for the ADP1 isolate used in our laboratories was in agreement with the ADP1 sequence that had previously been deposited in the database (GenBank accession no. AY289925). Comparison of the 16S rRNA gene sequences for ADP1 and 93A2 relative to those in the database demonstrated that they shared their highest similarity (98.2% and 98.4%, respectively) with A. baylyi B2T (accession no. AF509820). As shown in Fig. Fig.2,2, based on their 16S rRNA gene sequences, strains ADP1, 93A2, and the other members of A. baylyi form a separate cluster within the genus.
High-resolution genomic fingerprinting (AFLP) and cluster analysis of AFLP profiles were performed as described previously (18). By this method, genomic fingerprints are generated by selective amplification of restriction fragments (13). EcoRI and MseI were used as restriction enzymes, and Cy5-labeled EcoRI plus A primer and MseI with or without C primer were used for selective amplification (with A and C as selective nucleotides). After electrophoretic separation, the profiles were compared to those of a database of more than 200 strains of all described genomic species (13, 18, 19, 25) of the Leiden University Hospital database using cluster analysis. An isolate was identified to the species it grouped with at or above 50% (18). It was found that the A. baylyi strains formed a distinct cluster at 77.73% ± 2.43%, while this cluster was linked at only 26.14% ± 7.15% with strains of other described species (data not shown). A dendrogram of the A. baylyi cluster with type and reference strains of all described named and unnamed species is shown in Fig. Fig.33.
Phenotypic analysis was carried out for strains ADP1, 93A2, B2T, A7, and C5 as described previously (20), and the results further confirmed the close relationship of the strains. All five strains produced acid from d-glucose, and no strain showed hemolytic activity on sheep blood agar or produced gelatinase. When inoculated onto brain heart infusion broth (Oxoid), the strains showed clear growth after 1 day at 37°C but no growth at 44°C. In assimilation tests, modified from the method of Bouvet and Grimont (5), all strains utilized dl-lactate, 4-aminobutyrate, trans-aconitate, citrate (Simmons), l-aspartate, azelate, malonate, 4-hydroxybenzoate, ethanol, acetate, 2,3-butanediol, d-gluconate, and d-glucose with clear positivity within 2 days of incubation. No strain showed growth on β-alanine, l-histidine, histamine, l-phenylalanine, phenylacetate, levulinate, citraconate, l-tartrate, l-leucine, or l-ornithine within 10 days. These properties are distinct from those of all 32 named and unnamed species hitherto described within the genus Acinetobacter (4, 7, 11, 18, 19, 25).
Finally, transformation assays were performed to test whether the original A. baylyi strains identified by Carr et al. (7) shared the same high level of genetic competence as strains ADP1 and 93A2. Linearized plasmid DNA (pZR80) (16) containing an ADP1 lipA allele in which a kanamycin resistance cassette was inserted was used as donor DNA in transformations with all of the recipients. The assays were performed as previously reported (27), and transformation frequencies were calculated by plating transformants on Luria broth plates containing 25 μg/ml kanamycin. The results showed that the average transformation frequencies (three replicates; standard deviation, ≤0.1) for all five recipients (ADP1, 93A2, B2T, C5, and A7) were of the same order of magnitude (Table (Table1).1). These results confirm that the three previously described A. baylyi strains are highly competent for transformation at levels matching those of ADP1 and 93A2.
This paper demonstrates that the well-studied Acinetobacter sp. strain ADP1 and the independently isolated strain 93A2 are members of the species A. baylyi. We show that the ability to undergo highly efficient natural transformation, a trait that has made A. baylyi strain ADP1 an important model organism, is shared by the other members of the species as well. We propose that the transformation phenotype is the trait which best differentiates this group from all other Acinetobacter species.
We thank Janny Gruwel, Ingrid de Bruijn, Leen Van Simaey, and Catharine De Ganck for technical assistance and Elliot Juni for providing strain BD4.
Thierry De Baere is indebted to the Fund for Scientific Research Flanders (FWO) for a position as postdoctoral fellow. Work done in the laboratory of L. Nicholas Ornston was funded by NIH grant GM63628.