- Email: [email protected]
Comparison of ARDRA and recA-RFLP analysis for genomic species identification of Acinetobacter spp.
FEMS Microbiology Letters 165 (1998) 357^362
Comparison of ARDRA and recA-RFLP analysis for genomic species identi¢cation of Acinetobacter spp. A. Jawad, A.M. Snelling *, J. Heritage, P.M. Hawkey Department of Microbiology, University of Leeds, Leeds LS2 9JT, UK Received 3 June 1998 ; accepted 30 June 1998
Abstract The genus Acinetobacter is subdivided into genospecies on the basis of DNA relatedness of strains. Phenotypic tests are insufficient to identify the genospecies to which an isolate belongs. The effectiveness of two previously described PCR-based methods for genospeciating Acinetobacter spp. was compared using a group of 32 well-characterised strains representing six genospecies. Amplified ribosomal DNA restriction analysis (ARDRA) correctly identified all 32 strains. Using restriction fragment length polymorphism (RFLP) of recA PCR amplimers, only six of the 32 strains were correctly identified. Heterogeneity in the recA gene sequence was demonstrated within five of the genospecies. ARDRA proved to be a reliable method whereas analysis of recA RFLP profiles did not enable the genospecies of most of the isolates of Acinetobacter spp. to be determined. z 1998 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. Keywords : Acinetobacter; recA; Restriction fragment length polymorphism ; Ampli¢ed ribosomal DNA restriction analysis; Genospecies
1. Introduction Members of the genus Acinetobacter are Gramnegative bacteria that are widely distributed in nature and can be isolated from soil, water, and human skin. Some of the genospecies are commonly found in the hospital environment and are capable of causing a wide range of nosocomial opportunistic infections, whilst others have not been linked to human disease [1^3]. The taxonomy of the genus Acinetobacter has changed extensively during the past decade and at present comprises at least 19 genospecies, seven of which have been given formal species names * Corresponding author. Tel.: +44 (113) 2335594; Fax: +44 (113) 2335649. E-mail: [email protected]
whilst the others are designated by numbers [3,4]. Six of the genospecies (A. calcoaceticus, A. baumannii, genospecies 3, `between 1 and 3', 13, and `close to 13') are very similar phenotypically, and are known collectively as the Acinetobacter calcoaceticus-Acinetobacter baumannii (Acb) complex, even though they are genetically distinct. Identi¢cation of Acinetobacter spp. to the genospecies level using phenotypic techniques is not reliable, especially when trying to identify strains belonging to the Acb complex. Given the medical importance of some of the genospecies, there is a need for alternative methods that can be applied to large numbers of strains for epidemiological studies. The delineation of the genospecies is based on DNA-DNA hybridisation grouping but this procedure is laborious
0378-1097 / 98 / $19.00 ß 1998 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 1 0 9 7 ( 9 8 ) 0 0 3 0 2 - 4
FEMSLE 8308 12-8-98
358
A. Jawad et al. / FEMS Microbiology Letters 165 (1998) 357^362
and unsuitable for use in routine microbiology laboratories [5,6]. Restriction analysis of bacterial 16S rRNA genes has proved useful in taxonomic studies to distinguish between species belonging to the same genus [7]. This technique has been applied to the genus Acinetobacter and is termed ampli¢ed ribosomal DNA restriction analysis (ARDRA) [8^10]. A 1500-bp fragment of the 16S rRNA gene is ampli¢ed using PCR and digested with di¡erent restriction enzymes to give patterns that are speci¢c for each genospecies. Nowak and Kur [11] suggested that PCR ampli¢cation of an internal fragment of the Acinetobacter recA gene, with subsequent restriction analysis with two enzymes, could also distinguish between the di¡erent genospecies. However, only a single strain from each of 17 di¡erent genospecies was tested [11]. The present study was undertaken to evaluate the recA RFLP genotyping method further, using multiple strains from di¡erent genospecies, and to compare its ability to discriminate between genospecies with that of ARDRA.
2. Materials and methods 2.1. Bacterial strains and media A total of 32 strains representing six di¡erent genospecies were used in this study. These isolates have been well characterised previously using DNA-DNA hybridisation [6,12]. The collection included the type (T ) strains of A. calcoaceticus, A. baumannii, and A. radioresistens (also known as Gp12). Strains were grown on Iso-Sensitest agar (Unipath, Basingstoke, UK) at 30³C. 2.2. Preparation of template DNA for PCR Crude template DNA was prepared by suspending 2^3 colonies from an overnight agar plate in 50 Wl of ultrapure water and heating at 95³C for 5 min to lyse the cells. This was done immediately prior to use in the PCR reactions. 2.3. Ampli¢ed ribosomal DNA restriction analysis The procedure was performed as described by Vaneechoutte et al. [8]. Brie£y, a ca. 1500-bp fragment
of the 16S rRNA gene was ampli¢ed using the universal primers 5P-TGGCTCAGATTGAACGCTGGCGGC and 5P-TACCTTGTTACGACTTCACCCCA. Reactions (50 Wl ¢nal volume) consisted of 20 pM of each primer, 200 WM (each) dATP, dCTP, dGTP, and dTTP (Pharmacia LKB Biotechnology, Inc., Piscataway, NJ), 1UPCR bu¡er (Promega, Madison, WI), 0.5 U of SuperTaq DNA polymerase (HT Biotechnology, Cambridge, UK), 1.5 mM MgCl2 , and 5 Wl crude template. Ampli¢cation reactions were performed using an Omnigene thermal cycler (Hybaid Limited, Teddington, Middlesex, UK). After an initial denaturation of 95³C for 5 min, 35 cycles of 45 s at 95³C, 45 s at 50³C and 1 min at 72³C were performed. This was followed by a ¢nal extension at 72³C for 7 min. The restriction enzymes AluI, CfoI, MboI, RsaI and MspI (Gibco BRL, Paisley, UK), were used in separate reactions to digest the amplimers, according to the manufacturer's instructions. Restriction fragments were resolved in 3% NuSieve agarose (FMC Bioproducts, Rockland, ME, USA). Pro¢les were compared with those reported by Vaneechoutte et al. [8] for reference strains. 2.4. PCR ampli¢cation of the recA gene and RFLP analysis PCR ampli¢cation of the recA gene was performed as described by Nowak and Kur [11] using forward primer rA1 (5P-CCTGAATCTTCTGGTAAAAC-3P) and reverse primer rA2 (5P-GTTTCTGGGCTGCCAAACATTAC-3P). These primers target highly conserved regions of the recA gene sequence, as derived from an alignment of the recA sequences of Acinetobacter calcoaceticus (strain BD413/ADP1, GenBank accession number L26100) and Neisseria gonorrhoeae (strain FA19, GenBank accession number X64842). PCR reactions were set up as described for ARDRA, using 20 pM of each primer, except that 1.0 mM MgCl2 was found to be optimal. The reaction mixture was run through 35 cycles of denaturation at 94³C for 1 min, annealing at 45³C for 1 min, and extension at 72³C for 1 min. Finally, a 5-min extension period at 72³C was carried out to complete partial polymerisations. The ampli¢ed PCR products were resolved using 2% w/v agarose gels. Ampli¢ed DNA was used as such for re-
FEMSLE 8308 12-8-98
A. Jawad et al. / FEMS Microbiology Letters 165 (1998) 357^362
359
striction analysis with HinfI and MboI [11]. In addition, polymorphisms in the location of AluI sites were also studied. Restriction fragment patterns were analysed by gel electrophoresis of 10 Wl of each restriction mixture using 3% NuSieve agarose. Polyacrylamide gels (6%) stained with ethidium bromide were used whenever there was a problem with the resolution of smaller fragments using the latter. Fragments sizes were estimated with reference to a 100-bp or 123-bp DNA ladder with a tolerance of þ 5%. 2.5. Southern transfer and hybridisation of recA PCR amplimers PCR amplimers were transferred from the agarose gels to a nylon membrane (Boehringer Mannheim, Lewes, Sussex, UK) using an alkaline transfer procedure according to manufacturer's instructions. As a nucleotide probe, the recA amplimer from reference strain BD413 was labelled using a PCR DigDNA labelling kit (Boehringer Mannheim). Hybridisation was performed overnight at 68³C and positive hybridisation signals were detected using the DigDNA detection kit (Boehringer Mannheim).
3. Results
Fig. 1. Restriction endonuclease digestion of 16S rRNA gene PCR amplimers with AluI (lanes 2^4), CfoI (lanes 5^7) and MboI (lanes 8^10) for three strains of A. baumannii (lanes 2, 5, 8 ATCC 19606T ; lanes 3, 6, 9 16/48 and lanes 4, 7, 10 16/49). Lan- 1, 123 bp DNA ladder (Gibco-BRL).
413 produced fragments of the expected size after restriction. The DNA probe prepared from this strain gave positive hybridisation signals with the amplimers from all of the strains tested. Twenty-
3.1. PCR ampli¢cation of the rRNA gene and RFLP analysis Using the ARDRA technique with ¢ve restriction enzymes, the genospecies designation of the 32 Acinetobacter strains was con¢rmed. Fig. 1 illustrates the RFLP patterns obtained using AluI, CfoI and MboI restriction endonucleases for three strains of A. baumannii. It can be seen that all three strains give the same pro¢les as each other. For the AluI digests, it is assumed that with each of the two largest bands, there are actually two fragments co-migrating, so that the cumulative size of the fragments matches the original amplimer. 3.2. PCR ampli¢cation of the recA gene and RFLP analysis The recA PCR amplimer from reference strain BD
Fig. 2. Restriction endonuclease digestion of the recA PCR amplimers with HinfI for six strains of A. radioresistens. Lane 1, 100-bp DNA ladder (Gibco-BRL). Lanes 2^7, A. radioresistens 266, 271, 70819-85, M109, M152 and M17694T .
FEMSLE 8308 12-8-98
360
A. Jawad et al. / FEMS Microbiology Letters 165 (1998) 357^362
with HinfI. Two di¡erent banding patterns can clearly be seen. Table 1 summarises the approximate sizes of the HinfI and MboI RFLP fragments obtained after digestion of the recA PCR products obtained from the 32 strains. Four of the six genospecies gave a unique HinfI RFLP pro¢le but in the case of A. haemolyticus and A. radioresistens two distinct pro¢les were observed amongst strains belonging to the same genospecies. Three of the six genospecies gave a unique MboI RFLP pro¢le. Strains belonging to A. calcoaceticus, A. radioresistens and Acinetobacter genospecies 3 generated two pro¢les per genospecies. Considering the HinfI and MboI pro¢les together, there were 10 distinct pro¢le combinations amongst the six genospecies. Only six of the 32 strains (the A. baumannii) could be correctly identi¢ed to the genospecies level using the pro¢les reported by Nowak and Kur [11] for reference strains. To investigate the heterogeneity amongst the recA PCR amplimers further, AluI digests were also performed (Table 1). Three di¡erent banding patterns were observed amongst the six strains of A. baumannii (Fig. 3). With two of the strains, the 420-bp recA amplimer remained intact indicating that it lacked an AluI restriction site. Similarly, the amplimers obtained from the six strains of A. haemolyticus also lacked AluI sites (Table 1). In some cases it was deduced from the cumulative length of the restriction fragments that two fragments of the same apparent
Fig. 3. Restriction endonuclease digestion of the recA PCR amplimers with AluI for six strains of A. baumannii. Lane 1, 100-bp DNA ladder. Lanes 2^7, A. baumannii 16/48, 16/49, ATCC 17904, R0211019, ATCC 9955 and ATCC 19606T .
nine of the 32 strains gave a 420-bp amplimer which hybridised with the recA probe. A. radioresistens 266 and 271 consistently gave two PCR amplimers, of 480 and 230 bp. However, the yield of the 230-bp amplimer was small and although it could not be eliminated by attempts to optimise further the PCR conditions, it did not hybridise with the recA probe nor did it make any contribution to the RFLP pattern. Fig. 2 shows the RFLP patterns of six strains of A. radioresistens obtained by digesting the amplimer Table 1 RFLP analysis of recA PCR amplimers for strains of Acinetobacter Genospecies
Strain
A. calcoaceticus
BD 413 ATCC 23055T , R584 16/48, R0211019 ATCC 17904, ATCC 9955 ATCC 19606T , 16/49 ATCC 17922 10088, 16/58 ATCC 19004, 1163, 10084 12112, R2284, R2624 R2376, R412 M188, M59, 9908 ATCC 17907, ATCC 19002 1082 266, 271 70819-85, M109, M152, M17694T
A. baumannii
Acinetobacter GP 3
Acinetobacter GP 13 A. haemolyticus
A. radioresistens
RFLP pattern produced HinfI
MboI
AluI
230, 230, 230, 230, 230, 130, 130, 130, 240, 240, 230, 230, 230, 300, 230,
210, 290, 420 420 420 210, 210, 180, 140, 140, 130, 130, 130, 360, 230,
250, 125, 240, 180, 420 250, 190, 190, 190, 190, 420 420 420 330, 290,
70, 50, 40, 30 70, 50, 40, 30 105, 50, 35 105, 50, 35 105, 50, 35 110, 70, 50, 40, 30 110, 70, 50, 40, 30 110, 70, 50, 40, 30 100, 50, 30 100, 50, 30 110, 50, 30 110, 50, 30 50, 50, 30, 30, 20 150, 30 105, 50, 30
FEMSLE 8308 12-8-98
160, 55 100, 30
80, 60, 40, 30 80, 60, 40, 30 80, 60, 40, 30, 30 140, 100, 40 140, 100, 40 60, 60, 50, 20 60, 60, 50, 20 60, 60, 50, 20 80, 30 190
120, 50 120, 115, 60 180 120, 120 120, 120, 120, 120, 120,
50 50, 50 50, 50 120 120
100, 40 120
A. Jawad et al. / FEMS Microbiology Letters 165 (1998) 357^362
size must have been generated. This is illustrated in Fig. 3, lanes 4 and 6 where ATCC 17904 and ATCC 9955 give two ca. 120-bp fragments. In total, nine AluI restriction pro¢les were observed amongst the six genospecies. Considering the RFLP pro¢les for the three enzymes together, 13 distinct pro¢les were observed amongst the six genospecies. In Table 1, isolates belonging to the same genospecies and giving the same combined RFLP pro¢le are grouped together.
4. Discussion Nowak and Kur [11] ampli¢ed an internal fragment of the recA gene from one strain of each of the 17 genospecies of Acinetobacter. Combining the HinfI and MboI RFLP patterns, they concluded that each genospecies has a unique pattern, which can be used as a means of determining the genospecies of new isolates. Using the same PCR primers, we have extended this investigation using multiple strains from each of six genospecies. The results shown in Table 1 and Figs. 2 and 3 indicate that although there are di¡erences in the recA RFLP patterns of the six di¡erent genospecies tested, there can also be signi¢cant di¡erences between strains belonging to the same genospecies, with 10 di¡erent composite HinfI/MboI pro¢les observed overall (Table 1). Only in the case of A. baumannii did a composite pro¢le appear to be genospecies speci¢c and isolates could be identi¢ed by reference to the RFLP pro¢les published previously. In contrast, the ARDRA technique correctly identi¢ed the genospecies designation of all 32 strains of Acinetobacter spp. with the use of ¢ve di¡erent restriction enzymes. ARDRA does have limitations in that in order to identify all of the recognised genospecies of Acinetobacter correctly, Vaneechoutte et al. [8] recommend the additional use of a few simple phenotypic tests. This is because the method clusters strains from DNA groups 5, 7, 10, 11, 17 and A. haemolyticus pairwise into three groups. The use of the enzyme AluI showed that the heterogeneity in the sequences of the recA amplimers is not restricted to the HinfI and MboI sites. By combining the RFLP pro¢les obtained with the three enzymes together, the 32 strains could be divided
361
into 13 distinct groups. Further sequence analysis is warranted to investigate the phylogenetic signi¢cance of these polymorphisms. The use of additional restriction enzymes might allow a higher degree of discrimination between isolates and be of use in epidemiological studies or tracing the cause of nosocomial outbreaks. It is not yet known how the di¡erences in the RFLP patterns, and hence gene sequences, relate to functional di¡erences of RecA in acinetobacter. It is plausible that the di¡erent versions may di¡er with respect to the e¤ciency with which they repair DNA damage and interact with other components of the cellular SOS response. Recently, a technique based on tRNA spacer ¢ngerprinting has been shown to be reliable for genospeciating Acinetobacter spp. and was validated using 128 strains [4]. Analysis of PCR products of the 16S-23S spacer regions [13] and sequence variations in gyrB genes [14] have also been proposed as methods that can discern the genotype of strains but have yet to be validated with large numbers of strains. Both the ARDRA and recA methods require restriction digestion of PCR amplimers and thus take similar amounts of time to perform. However, the fragments obtained with the recA-RFLP method tend to be smaller than those obtained with ARDRA and more often require the use of a polyacrylamide gel to resolve and visualise them. Our results show that ARDRA gives a rapid, economical de¢nitive result when simply attempting to determine the genospecies to which isolates of Acinetobacter belong, but further analysis of polymorphisms amongst the recA genes might yield phylogenetic groupings of greater use in studying the epidemiology and evolution of members of this genus.
Acknowledgments We would like to thank Dr. P. Gerner-Smidt, Department of Clinical Microbiology, Statens Seruminstitut, Copenhagen S, Denmark and Dr. K.J. Towner, Department of Microbiology and Public Health Laboratory Service, University Hospital, Nottingham, UK, for donating some of the strains used in this study. We thank the Ministry of Education of the Government of Pakistan for funding this work.
FEMSLE 8308 12-8-98
362
A. Jawad et al. / FEMS Microbiology Letters 165 (1998) 357^362
References [1] Bergogne-Beèreèzin, E. and Towner, K.J. (1996) Acinetobacter spp. as nosocomial pathogens: microbiological, clinical and epidemiological features. Clin. Microbiol. Rev. 9, 148^165. [2] Iqbal Hossain, M., Iqbal Kabir, A.K.M., Khan, W.A. and Fuchs, G.J. (1998) Acinetobacter bacteraemia in patients with diarrhoeal disease. Epidemiol. Infect. 120, 139^142. [3] Towner, K.J. (1997) Clinical importance and antibiotic resistance of Acinetobacter spp. J. Med. Microbiol. 46, 721^746. [4] Ehrenstein, B., Bernards, A.T., Dijkshoorn, L., Gerner-Smidt, P., Towner, K.J., Bouvet, P.J.M., Daschner, F.D. and Grundmann, H. (1996) Acinetobacter species identi¢cation by using tRNA spacer ¢ngerprinting. J. Clin. Microbiol. 34, 2414^ 2420. [5] Bouvet, P.J.M. and Grimont, P.A.D. (1986) Taxonomy of the genus Acinetobacter with the recognition of Acinetobacter baumannii sp. nov., Acinetobacter haemolyticus sp. nov., Acinetobacter johnsonii sp. nov., and Acinetobacter junii sp. nov. and emended descriptions of Acinetobacter calcoaceticus and Acinetobacter lwo¤i. Int. J. Syst. Bacteriol. 36, 228^240. [6] Tjernberg, I. and Ursing, J. (1989) Clinical strains of Acinetobacter classi¢ed by DNA-DNA hybridization. APMIS 97, 595^606. [7] Jayarao, B.M., Doreè Jr., J.J.E. and Oliver, S.P. 1992. Restriction fragment length polymorphism analysis of 16S ribosomal DNA of Streptococcus and Enterococcus species of bovine origin. J. Clin. Microbiol. 30, 2235^2240.
[8] Vaneechoutte, M., Dijkshoorn, L., Tjernberg, I., Elaichouni, A., Vos, P.D., Claeys, G. and Verschraegen, G. (1995) Identi¢cation of Acinetobacter genomic species by ampli¢ed ribosomal DNA restriction analysis. J. Clin. Microbiol. 33, 11^15. [9] Vila, J., Marcos, M.A. and Jimenez De Anta, M.T. (1996) A comparative study of PCR-based DNA ¢ngerprinting techniques for typing of the Acinetobacter calcoaceticus-Acinetobacter baumannii complex. J. Med. Microbiol. 44, 482^489. [10] Bernard, A.T., de Beaufort, A.J., Dijkshoorn, L. and van Boven, C.P.A. (1997) Outbreaks of septicaemia in neonates caused by Acinetobacter junii investigated by ampli¢ed ribosomal DNA restriction analysis (ARDRA) and four typing methods. J. Hosp. Infect. 35, 129^140. [11] Nowak, A. and Kur, J. (1995) Genomic species typing of acinetobacters by polymerase chain reaction ampli¢cation of the recA gene. FEMS Microbiol. Lett. 130, 327^332. [12] Gerner-Smidt, P., Tjernberg, I. and Ursing, J. (1991) Reliability of phenotypic tests for identi¢cation of Acinetobacter species. J. Clin. Microbiol. 29,277^282. [13] Nowak, A., Burkiewicz, A. and Kur, J. (1995) PCR di¡erentiation of seventeen genospecies of Acinetobacter. FEMS Microbiol. Lett. 126, 181^187. [14] Yamamoto, S. and Harayama, S. (1996) Phylogenetic analysis of Acinetobacter strains based on the nucleotide sequences of gyrB genes and on the amino acid sequences of their products. Int. J. Syst. Bacteriol. 46, 506^511.
FEMSLE 8308 12-8-98
Comparison of ARDRA and recA-RFLP analysis for genomic species identi¢cation of Acinetobacter spp. A. Jawad, A.M. Snelling *, J. Heritage, P.M. Hawkey Department of Microbiology, University of Leeds, Leeds LS2 9JT, UK Received 3 June 1998 ; accepted 30 June 1998
Abstract The genus Acinetobacter is subdivided into genospecies on the basis of DNA relatedness of strains. Phenotypic tests are insufficient to identify the genospecies to which an isolate belongs. The effectiveness of two previously described PCR-based methods for genospeciating Acinetobacter spp. was compared using a group of 32 well-characterised strains representing six genospecies. Amplified ribosomal DNA restriction analysis (ARDRA) correctly identified all 32 strains. Using restriction fragment length polymorphism (RFLP) of recA PCR amplimers, only six of the 32 strains were correctly identified. Heterogeneity in the recA gene sequence was demonstrated within five of the genospecies. ARDRA proved to be a reliable method whereas analysis of recA RFLP profiles did not enable the genospecies of most of the isolates of Acinetobacter spp. to be determined. z 1998 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. Keywords : Acinetobacter; recA; Restriction fragment length polymorphism ; Ampli¢ed ribosomal DNA restriction analysis; Genospecies
1. Introduction Members of the genus Acinetobacter are Gramnegative bacteria that are widely distributed in nature and can be isolated from soil, water, and human skin. Some of the genospecies are commonly found in the hospital environment and are capable of causing a wide range of nosocomial opportunistic infections, whilst others have not been linked to human disease [1^3]. The taxonomy of the genus Acinetobacter has changed extensively during the past decade and at present comprises at least 19 genospecies, seven of which have been given formal species names * Corresponding author. Tel.: +44 (113) 2335594; Fax: +44 (113) 2335649. E-mail: [email protected]
whilst the others are designated by numbers [3,4]. Six of the genospecies (A. calcoaceticus, A. baumannii, genospecies 3, `between 1 and 3', 13, and `close to 13') are very similar phenotypically, and are known collectively as the Acinetobacter calcoaceticus-Acinetobacter baumannii (Acb) complex, even though they are genetically distinct. Identi¢cation of Acinetobacter spp. to the genospecies level using phenotypic techniques is not reliable, especially when trying to identify strains belonging to the Acb complex. Given the medical importance of some of the genospecies, there is a need for alternative methods that can be applied to large numbers of strains for epidemiological studies. The delineation of the genospecies is based on DNA-DNA hybridisation grouping but this procedure is laborious
0378-1097 / 98 / $19.00 ß 1998 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 1 0 9 7 ( 9 8 ) 0 0 3 0 2 - 4
FEMSLE 8308 12-8-98
358
A. Jawad et al. / FEMS Microbiology Letters 165 (1998) 357^362
and unsuitable for use in routine microbiology laboratories [5,6]. Restriction analysis of bacterial 16S rRNA genes has proved useful in taxonomic studies to distinguish between species belonging to the same genus [7]. This technique has been applied to the genus Acinetobacter and is termed ampli¢ed ribosomal DNA restriction analysis (ARDRA) [8^10]. A 1500-bp fragment of the 16S rRNA gene is ampli¢ed using PCR and digested with di¡erent restriction enzymes to give patterns that are speci¢c for each genospecies. Nowak and Kur [11] suggested that PCR ampli¢cation of an internal fragment of the Acinetobacter recA gene, with subsequent restriction analysis with two enzymes, could also distinguish between the di¡erent genospecies. However, only a single strain from each of 17 di¡erent genospecies was tested [11]. The present study was undertaken to evaluate the recA RFLP genotyping method further, using multiple strains from di¡erent genospecies, and to compare its ability to discriminate between genospecies with that of ARDRA.
2. Materials and methods 2.1. Bacterial strains and media A total of 32 strains representing six di¡erent genospecies were used in this study. These isolates have been well characterised previously using DNA-DNA hybridisation [6,12]. The collection included the type (T ) strains of A. calcoaceticus, A. baumannii, and A. radioresistens (also known as Gp12). Strains were grown on Iso-Sensitest agar (Unipath, Basingstoke, UK) at 30³C. 2.2. Preparation of template DNA for PCR Crude template DNA was prepared by suspending 2^3 colonies from an overnight agar plate in 50 Wl of ultrapure water and heating at 95³C for 5 min to lyse the cells. This was done immediately prior to use in the PCR reactions. 2.3. Ampli¢ed ribosomal DNA restriction analysis The procedure was performed as described by Vaneechoutte et al. [8]. Brie£y, a ca. 1500-bp fragment
of the 16S rRNA gene was ampli¢ed using the universal primers 5P-TGGCTCAGATTGAACGCTGGCGGC and 5P-TACCTTGTTACGACTTCACCCCA. Reactions (50 Wl ¢nal volume) consisted of 20 pM of each primer, 200 WM (each) dATP, dCTP, dGTP, and dTTP (Pharmacia LKB Biotechnology, Inc., Piscataway, NJ), 1UPCR bu¡er (Promega, Madison, WI), 0.5 U of SuperTaq DNA polymerase (HT Biotechnology, Cambridge, UK), 1.5 mM MgCl2 , and 5 Wl crude template. Ampli¢cation reactions were performed using an Omnigene thermal cycler (Hybaid Limited, Teddington, Middlesex, UK). After an initial denaturation of 95³C for 5 min, 35 cycles of 45 s at 95³C, 45 s at 50³C and 1 min at 72³C were performed. This was followed by a ¢nal extension at 72³C for 7 min. The restriction enzymes AluI, CfoI, MboI, RsaI and MspI (Gibco BRL, Paisley, UK), were used in separate reactions to digest the amplimers, according to the manufacturer's instructions. Restriction fragments were resolved in 3% NuSieve agarose (FMC Bioproducts, Rockland, ME, USA). Pro¢les were compared with those reported by Vaneechoutte et al. [8] for reference strains. 2.4. PCR ampli¢cation of the recA gene and RFLP analysis PCR ampli¢cation of the recA gene was performed as described by Nowak and Kur [11] using forward primer rA1 (5P-CCTGAATCTTCTGGTAAAAC-3P) and reverse primer rA2 (5P-GTTTCTGGGCTGCCAAACATTAC-3P). These primers target highly conserved regions of the recA gene sequence, as derived from an alignment of the recA sequences of Acinetobacter calcoaceticus (strain BD413/ADP1, GenBank accession number L26100) and Neisseria gonorrhoeae (strain FA19, GenBank accession number X64842). PCR reactions were set up as described for ARDRA, using 20 pM of each primer, except that 1.0 mM MgCl2 was found to be optimal. The reaction mixture was run through 35 cycles of denaturation at 94³C for 1 min, annealing at 45³C for 1 min, and extension at 72³C for 1 min. Finally, a 5-min extension period at 72³C was carried out to complete partial polymerisations. The ampli¢ed PCR products were resolved using 2% w/v agarose gels. Ampli¢ed DNA was used as such for re-
FEMSLE 8308 12-8-98
A. Jawad et al. / FEMS Microbiology Letters 165 (1998) 357^362
359
striction analysis with HinfI and MboI [11]. In addition, polymorphisms in the location of AluI sites were also studied. Restriction fragment patterns were analysed by gel electrophoresis of 10 Wl of each restriction mixture using 3% NuSieve agarose. Polyacrylamide gels (6%) stained with ethidium bromide were used whenever there was a problem with the resolution of smaller fragments using the latter. Fragments sizes were estimated with reference to a 100-bp or 123-bp DNA ladder with a tolerance of þ 5%. 2.5. Southern transfer and hybridisation of recA PCR amplimers PCR amplimers were transferred from the agarose gels to a nylon membrane (Boehringer Mannheim, Lewes, Sussex, UK) using an alkaline transfer procedure according to manufacturer's instructions. As a nucleotide probe, the recA amplimer from reference strain BD413 was labelled using a PCR DigDNA labelling kit (Boehringer Mannheim). Hybridisation was performed overnight at 68³C and positive hybridisation signals were detected using the DigDNA detection kit (Boehringer Mannheim).
3. Results
Fig. 1. Restriction endonuclease digestion of 16S rRNA gene PCR amplimers with AluI (lanes 2^4), CfoI (lanes 5^7) and MboI (lanes 8^10) for three strains of A. baumannii (lanes 2, 5, 8 ATCC 19606T ; lanes 3, 6, 9 16/48 and lanes 4, 7, 10 16/49). Lan- 1, 123 bp DNA ladder (Gibco-BRL).
413 produced fragments of the expected size after restriction. The DNA probe prepared from this strain gave positive hybridisation signals with the amplimers from all of the strains tested. Twenty-
3.1. PCR ampli¢cation of the rRNA gene and RFLP analysis Using the ARDRA technique with ¢ve restriction enzymes, the genospecies designation of the 32 Acinetobacter strains was con¢rmed. Fig. 1 illustrates the RFLP patterns obtained using AluI, CfoI and MboI restriction endonucleases for three strains of A. baumannii. It can be seen that all three strains give the same pro¢les as each other. For the AluI digests, it is assumed that with each of the two largest bands, there are actually two fragments co-migrating, so that the cumulative size of the fragments matches the original amplimer. 3.2. PCR ampli¢cation of the recA gene and RFLP analysis The recA PCR amplimer from reference strain BD
Fig. 2. Restriction endonuclease digestion of the recA PCR amplimers with HinfI for six strains of A. radioresistens. Lane 1, 100-bp DNA ladder (Gibco-BRL). Lanes 2^7, A. radioresistens 266, 271, 70819-85, M109, M152 and M17694T .
FEMSLE 8308 12-8-98
360
A. Jawad et al. / FEMS Microbiology Letters 165 (1998) 357^362
with HinfI. Two di¡erent banding patterns can clearly be seen. Table 1 summarises the approximate sizes of the HinfI and MboI RFLP fragments obtained after digestion of the recA PCR products obtained from the 32 strains. Four of the six genospecies gave a unique HinfI RFLP pro¢le but in the case of A. haemolyticus and A. radioresistens two distinct pro¢les were observed amongst strains belonging to the same genospecies. Three of the six genospecies gave a unique MboI RFLP pro¢le. Strains belonging to A. calcoaceticus, A. radioresistens and Acinetobacter genospecies 3 generated two pro¢les per genospecies. Considering the HinfI and MboI pro¢les together, there were 10 distinct pro¢le combinations amongst the six genospecies. Only six of the 32 strains (the A. baumannii) could be correctly identi¢ed to the genospecies level using the pro¢les reported by Nowak and Kur [11] for reference strains. To investigate the heterogeneity amongst the recA PCR amplimers further, AluI digests were also performed (Table 1). Three di¡erent banding patterns were observed amongst the six strains of A. baumannii (Fig. 3). With two of the strains, the 420-bp recA amplimer remained intact indicating that it lacked an AluI restriction site. Similarly, the amplimers obtained from the six strains of A. haemolyticus also lacked AluI sites (Table 1). In some cases it was deduced from the cumulative length of the restriction fragments that two fragments of the same apparent
Fig. 3. Restriction endonuclease digestion of the recA PCR amplimers with AluI for six strains of A. baumannii. Lane 1, 100-bp DNA ladder. Lanes 2^7, A. baumannii 16/48, 16/49, ATCC 17904, R0211019, ATCC 9955 and ATCC 19606T .
nine of the 32 strains gave a 420-bp amplimer which hybridised with the recA probe. A. radioresistens 266 and 271 consistently gave two PCR amplimers, of 480 and 230 bp. However, the yield of the 230-bp amplimer was small and although it could not be eliminated by attempts to optimise further the PCR conditions, it did not hybridise with the recA probe nor did it make any contribution to the RFLP pattern. Fig. 2 shows the RFLP patterns of six strains of A. radioresistens obtained by digesting the amplimer Table 1 RFLP analysis of recA PCR amplimers for strains of Acinetobacter Genospecies
Strain
A. calcoaceticus
BD 413 ATCC 23055T , R584 16/48, R0211019 ATCC 17904, ATCC 9955 ATCC 19606T , 16/49 ATCC 17922 10088, 16/58 ATCC 19004, 1163, 10084 12112, R2284, R2624 R2376, R412 M188, M59, 9908 ATCC 17907, ATCC 19002 1082 266, 271 70819-85, M109, M152, M17694T
A. baumannii
Acinetobacter GP 3
Acinetobacter GP 13 A. haemolyticus
A. radioresistens
RFLP pattern produced HinfI
MboI
AluI
230, 230, 230, 230, 230, 130, 130, 130, 240, 240, 230, 230, 230, 300, 230,
210, 290, 420 420 420 210, 210, 180, 140, 140, 130, 130, 130, 360, 230,
250, 125, 240, 180, 420 250, 190, 190, 190, 190, 420 420 420 330, 290,
70, 50, 40, 30 70, 50, 40, 30 105, 50, 35 105, 50, 35 105, 50, 35 110, 70, 50, 40, 30 110, 70, 50, 40, 30 110, 70, 50, 40, 30 100, 50, 30 100, 50, 30 110, 50, 30 110, 50, 30 50, 50, 30, 30, 20 150, 30 105, 50, 30
FEMSLE 8308 12-8-98
160, 55 100, 30
80, 60, 40, 30 80, 60, 40, 30 80, 60, 40, 30, 30 140, 100, 40 140, 100, 40 60, 60, 50, 20 60, 60, 50, 20 60, 60, 50, 20 80, 30 190
120, 50 120, 115, 60 180 120, 120 120, 120, 120, 120, 120,
50 50, 50 50, 50 120 120
100, 40 120
A. Jawad et al. / FEMS Microbiology Letters 165 (1998) 357^362
size must have been generated. This is illustrated in Fig. 3, lanes 4 and 6 where ATCC 17904 and ATCC 9955 give two ca. 120-bp fragments. In total, nine AluI restriction pro¢les were observed amongst the six genospecies. Considering the RFLP pro¢les for the three enzymes together, 13 distinct pro¢les were observed amongst the six genospecies. In Table 1, isolates belonging to the same genospecies and giving the same combined RFLP pro¢le are grouped together.
4. Discussion Nowak and Kur [11] ampli¢ed an internal fragment of the recA gene from one strain of each of the 17 genospecies of Acinetobacter. Combining the HinfI and MboI RFLP patterns, they concluded that each genospecies has a unique pattern, which can be used as a means of determining the genospecies of new isolates. Using the same PCR primers, we have extended this investigation using multiple strains from each of six genospecies. The results shown in Table 1 and Figs. 2 and 3 indicate that although there are di¡erences in the recA RFLP patterns of the six di¡erent genospecies tested, there can also be signi¢cant di¡erences between strains belonging to the same genospecies, with 10 di¡erent composite HinfI/MboI pro¢les observed overall (Table 1). Only in the case of A. baumannii did a composite pro¢le appear to be genospecies speci¢c and isolates could be identi¢ed by reference to the RFLP pro¢les published previously. In contrast, the ARDRA technique correctly identi¢ed the genospecies designation of all 32 strains of Acinetobacter spp. with the use of ¢ve di¡erent restriction enzymes. ARDRA does have limitations in that in order to identify all of the recognised genospecies of Acinetobacter correctly, Vaneechoutte et al. [8] recommend the additional use of a few simple phenotypic tests. This is because the method clusters strains from DNA groups 5, 7, 10, 11, 17 and A. haemolyticus pairwise into three groups. The use of the enzyme AluI showed that the heterogeneity in the sequences of the recA amplimers is not restricted to the HinfI and MboI sites. By combining the RFLP pro¢les obtained with the three enzymes together, the 32 strains could be divided
361
into 13 distinct groups. Further sequence analysis is warranted to investigate the phylogenetic signi¢cance of these polymorphisms. The use of additional restriction enzymes might allow a higher degree of discrimination between isolates and be of use in epidemiological studies or tracing the cause of nosocomial outbreaks. It is not yet known how the di¡erences in the RFLP patterns, and hence gene sequences, relate to functional di¡erences of RecA in acinetobacter. It is plausible that the di¡erent versions may di¡er with respect to the e¤ciency with which they repair DNA damage and interact with other components of the cellular SOS response. Recently, a technique based on tRNA spacer ¢ngerprinting has been shown to be reliable for genospeciating Acinetobacter spp. and was validated using 128 strains [4]. Analysis of PCR products of the 16S-23S spacer regions [13] and sequence variations in gyrB genes [14] have also been proposed as methods that can discern the genotype of strains but have yet to be validated with large numbers of strains. Both the ARDRA and recA methods require restriction digestion of PCR amplimers and thus take similar amounts of time to perform. However, the fragments obtained with the recA-RFLP method tend to be smaller than those obtained with ARDRA and more often require the use of a polyacrylamide gel to resolve and visualise them. Our results show that ARDRA gives a rapid, economical de¢nitive result when simply attempting to determine the genospecies to which isolates of Acinetobacter belong, but further analysis of polymorphisms amongst the recA genes might yield phylogenetic groupings of greater use in studying the epidemiology and evolution of members of this genus.
Acknowledgments We would like to thank Dr. P. Gerner-Smidt, Department of Clinical Microbiology, Statens Seruminstitut, Copenhagen S, Denmark and Dr. K.J. Towner, Department of Microbiology and Public Health Laboratory Service, University Hospital, Nottingham, UK, for donating some of the strains used in this study. We thank the Ministry of Education of the Government of Pakistan for funding this work.
FEMSLE 8308 12-8-98
362
A. Jawad et al. / FEMS Microbiology Letters 165 (1998) 357^362
References [1] Bergogne-Beèreèzin, E. and Towner, K.J. (1996) Acinetobacter spp. as nosocomial pathogens: microbiological, clinical and epidemiological features. Clin. Microbiol. Rev. 9, 148^165. [2] Iqbal Hossain, M., Iqbal Kabir, A.K.M., Khan, W.A. and Fuchs, G.J. (1998) Acinetobacter bacteraemia in patients with diarrhoeal disease. Epidemiol. Infect. 120, 139^142. [3] Towner, K.J. (1997) Clinical importance and antibiotic resistance of Acinetobacter spp. J. Med. Microbiol. 46, 721^746. [4] Ehrenstein, B., Bernards, A.T., Dijkshoorn, L., Gerner-Smidt, P., Towner, K.J., Bouvet, P.J.M., Daschner, F.D. and Grundmann, H. (1996) Acinetobacter species identi¢cation by using tRNA spacer ¢ngerprinting. J. Clin. Microbiol. 34, 2414^ 2420. [5] Bouvet, P.J.M. and Grimont, P.A.D. (1986) Taxonomy of the genus Acinetobacter with the recognition of Acinetobacter baumannii sp. nov., Acinetobacter haemolyticus sp. nov., Acinetobacter johnsonii sp. nov., and Acinetobacter junii sp. nov. and emended descriptions of Acinetobacter calcoaceticus and Acinetobacter lwo¤i. Int. J. Syst. Bacteriol. 36, 228^240. [6] Tjernberg, I. and Ursing, J. (1989) Clinical strains of Acinetobacter classi¢ed by DNA-DNA hybridization. APMIS 97, 595^606. [7] Jayarao, B.M., Doreè Jr., J.J.E. and Oliver, S.P. 1992. Restriction fragment length polymorphism analysis of 16S ribosomal DNA of Streptococcus and Enterococcus species of bovine origin. J. Clin. Microbiol. 30, 2235^2240.
[8] Vaneechoutte, M., Dijkshoorn, L., Tjernberg, I., Elaichouni, A., Vos, P.D., Claeys, G. and Verschraegen, G. (1995) Identi¢cation of Acinetobacter genomic species by ampli¢ed ribosomal DNA restriction analysis. J. Clin. Microbiol. 33, 11^15. [9] Vila, J., Marcos, M.A. and Jimenez De Anta, M.T. (1996) A comparative study of PCR-based DNA ¢ngerprinting techniques for typing of the Acinetobacter calcoaceticus-Acinetobacter baumannii complex. J. Med. Microbiol. 44, 482^489. [10] Bernard, A.T., de Beaufort, A.J., Dijkshoorn, L. and van Boven, C.P.A. (1997) Outbreaks of septicaemia in neonates caused by Acinetobacter junii investigated by ampli¢ed ribosomal DNA restriction analysis (ARDRA) and four typing methods. J. Hosp. Infect. 35, 129^140. [11] Nowak, A. and Kur, J. (1995) Genomic species typing of acinetobacters by polymerase chain reaction ampli¢cation of the recA gene. FEMS Microbiol. Lett. 130, 327^332. [12] Gerner-Smidt, P., Tjernberg, I. and Ursing, J. (1991) Reliability of phenotypic tests for identi¢cation of Acinetobacter species. J. Clin. Microbiol. 29,277^282. [13] Nowak, A., Burkiewicz, A. and Kur, J. (1995) PCR di¡erentiation of seventeen genospecies of Acinetobacter. FEMS Microbiol. Lett. 126, 181^187. [14] Yamamoto, S. and Harayama, S. (1996) Phylogenetic analysis of Acinetobacter strains based on the nucleotide sequences of gyrB genes and on the amino acid sequences of their products. Int. J. Syst. Bacteriol. 46, 506^511.
FEMSLE 8308 12-8-98