PCR-amplified 16S and 23S rDNA restriction analysis for the identification of Acinetobacter strains at the DNA group level

0

INSTITUT

Paris

Res. Microbial. 1997, 148, 777-784

PASTEUR~ELSEVIER

1997

PCR-amplified 16s and 23s rDNA restriction analysis for the identification of Acinetobacter strains at the DNA group level M.I.

Garcia-Arata

(L 2, (*), P. Gerner-Smidt

c2), F. Baquero

(l) and A. Ibrahim

c3)

(I) Department of M ic robiology, Hospital Ramdn y Cajal, Carretera Colmenar Km 9.1, 28034 Madrid, f2) De artment of Clinical Microbiology, Statens Seruminstitut, Artillerivej 5, Copenhagen DK-2300, and (P) Department of Microbiology, National Defence Research Institute, S- 901-82 Umed (Sweden)

SUMMARY

The genus Acinetobacter is phenotypically rather homogeneous, but genotypicaliy heterogeneous. In this study, a simple method based on restriction analysis of a PCRamplified large fragment (4.5 kb) of most of the ribosomal operon (16s and 23s ribosomal genes and the spacer in-between) was investigated. Sixty-seven collection strains belonging to the 20 DNA groups proposed until 1993 were studied. Using the enzyme SerBAl, 25 DNA profiles were obtained. Strains belonging to DNA groups 1,3,6, TU13 and TU15 showed two profiles each, and DNA groups 4,5 and 7 showed profiles with variants showing less intensive additional bands. The remaining 12 groups showed 12 different profiles. The profiles obtained were DNA-group-specific except for one profile which was shared between the unnamed DNA group 3 and a rarely encountered genotypieally related DNA group. These two DNA groups could be separated by using the enzyme Hinfl. Twenty-five additional clinical isolates previously characterized by standard DNA-DNA hybridization were selected in a double-blind fashion for identification at the DNA group level to check the reliability of the assay. All strains were correctly identified at the DNA group level. PC&lified 16s and 23s rDNA restriction analysis is both an accurate and rapid method for the identification of Acinefobacfer at the DNA group level.

Key-words:

ARDRA, Acinetobacter,

rDNA; Diagnosis, Restriction enzyme.

INTRODUCTION

Bacteria of the genus Acinetobacter are an increasingly important cause of nosocomial infection (Dolzani et al., 1995; Garcia-Arata, 1995; Gemer-Smidt and Tjemberg, 1993; Tjemberg and Ursing, 1989; Vaneechoutte et al.,

Submitted

April

(*) Corresponding

9, 1997, accepted August 4, 1997. author.

1995). The genus is genotypically heterogeneous but phenotypically rather homogeneous (Bouvet and Grimont, 1986; Gemer-Smidt et aZ., 1991). Currently, at least 20 DNA groups have been demonstrated in the genus (table I). Bouvet and Grimont (1986) characterized 85 clinical and reference strains using DNA-DNA hybridization.

778 Table

M.I. GARCIA-ARATA I. Current

taxonomic

status

in the genus

Acinetobacter. DNA

groups

1 2

Proposed name

A. calcoaceticus(*) A. baumannii(*) unnamed(*) A. haemolyticus A. junii unnamed A. johnsonii A. lwofsii

‘1

i 5 6 7 819 10 11 12 TU13 TU14 (=BJ13) TU15 BJ14 BJ15 BJ16 BJ17 “Close to TU13” “Between 1 and 3”

unnamed unnamed

A. radioresistens unnamed(*) unnamed

unnamed unnamed unnamed unnamed unnamed unnamed(*) unnamed(*)

DNA groups 1 to 12 (Bouvet and Grimont, 1986); DNA groups TU13 to TU15 (Tjernberg and Ursing, 1989); DNA groups BJ14 to BJ17 (Bouvet and Jeanjean, 1989); DNA groups “Close to TU13” and “Between 1 & 3” (Gemer-Smidt and Tjemberg, 1993). (*) A. culcoacericus-A. baumannii complex (Gemer-Smidt er al., 1991; Gemer-Smidt and Tjemberg, 1993).

Seventy-four strains were assigned to 12 DNA groups comprising 2 to 21 strains each. Four new species were named, A. baumannii, A. haemolyticus, .A. johnsonii and A. junii. The descriptions of the species A. calcoaceticus and A. lwoffii were emended; A. Zwofii comprised strains from the phenotypically indistinguishable DNA groups 8 and 9. Five DNA groups remained unnamed. With a partially overlapping set of strains, Tjemberg and Ursing (1989) confirmed these DNA groups, but only a single A. lwofJii DNA group was detected. These authors demon-

ARDRA BJ FCR

= =

amplified ribosomal DNA restriction DNA groups as proposed by Bouvet 1989. = polymerase chain reaction.

analysis. and Jeanjean,

ET AL.

strated that the new species A. radioresistens (Nishimura et aE., 1988) corresponded to Bouvet and Grimont’s DNA group 12, and proposed three additional DNA groups (TU13, TU14, and TU15). In the same year, Bouvet and Jeanjean (1989) reported five new proteolytic DNA groups (BJ13-BJ17); one of them, BJ13, corresponded to DNA group TU14 of Tjemberg and Ursing. In a Danish survey carried out by Gemer-Smidt and Tjemberg (1993), two additional DNA groups were described. These DNA groups were not named but genotypically belonged to the A.calcoaceticus-A.baumannii complex (Gemer-Smidt et al., 1991) which comprised the closely related DNA groups 1 (A. caEcoaceticus), 2 (A. baumannii), 3, and TU13. They were denoted “between 1 and 3” and “close to TU13”, indicating their genotypic relationship to the other DNA groups in the complex. The isolates belonging to the DNA groups 2, 3 and TU13 are frequently involved in nosocomial infections, but this is not the case for strains in DNA group 1. Since phenotypic differentiation of strains belonging to each of the 20 DNA groups is extremely difficult, a simple genotypic method for distinguishing them is necessary. Most current methods are either cumbersome or do not unequivocally identify all DNA groups. Amplified ribosomal DNA restriction analysis (ARDRA) of the 16s rDNA was earlier reported to be an efficient method for identification of Acinetobacter sp. (Vaneechoutte et aE., 1995). However, it was necessary to use three restriction enzymes in order to obtain full discrimination between DNA groups. We recently extended the size of the amplified fragment to encompass both the 16s and the 23s rDNA and the spacer in-between and showed that by using just one enzyme it was possible to differentiate between four of the DNA groups in the A.caZcoaceticus-A. baumannii complex containing the

TBE TIJ

= =

UV

=

tris boric EDTA buffer. DNA groups as proposed 1989. ultraviolet (light).

by Tjemberg

and Ursing,

ACINETOBACTER

DNA GROUP

clinically most important acinetobacters (Ibrahim et aE., 1996). The aim of the present study was thus to see if, by using this approach, it was possible to reduce the number of restriction enzymes in the ARDRA assay necessary for identification of all known Acinetobacter DNA groups.

MATERIALS Bacterial

AND METHODS

DIFFERENTIATION

BY ARDRA

779

berg and Ursing, 1989). In the second phase of the study, twenty-five additional collection strains, double-blindly selected and previously identified by ribotyping and/or DNA-DNA hybridization, were analysed to test the method’s reliability as an identification tool. These collection strains included: i) clinical isolates from Denmark: 9836, 10716, 9907, 12174a, 12398, 10790, 10508, 10074, 10086, U119, 274, R204, 4, 164, 65219-84, UlOO, 284, 286, 65109-84, 225, 266 and 70819-85 (GernerSmidt et al., 1991; Gerner-Smidt and Tjernberg, 1993), and ii) clinical isolates from references (Tjernberg and Ursing, 1989) M22, M117, and M140.

strains

A collection of 67 Acinetobacter strains representing each of the 20 DNA groups described so far were included in the first part of the study to set up the method (table II). Details on the strains have been published previously (Bouvet and Grimont, 1986; Bouvet and Jeanjean, 1989; Gerner-Smidt et al., 1991; Gemer-Smidt and Tjemberg, 1993; Tjem-

Table II. Sau3AI

and Hid

restriction

profiles

Purification

of template

DNA for amplification

A lo-u1 loopful of colonies was suspended in 300 l.tl distilled water in an Eppendorf tube and boiled for 10 min. After brief centrifugation, a 2-pl portion of the supernatant was added to 98 ~1 of PCR mix.

of 67 collection

strains belonging

to the 20 DNA

groups of

Acinetobacter. Restriction DNA

group

4 5 6 ,, 7 819 10 11 &13 T&4

(BJ13)

Strain no. R583, A344, R584 R944, M42, M59, ATCC 17902-2 ATCC 17904, ATCC 9955, ATCC 17978, 50853-82, 189 ATCC 17922, M102, M79, Ml62 ATCC 19004 ATCC 17906, ATCC 17907, ATCC 19002, M197, Ml9 ATCC 17908, M124, M27, M74b, M96 ATCC 17979 39 ATCC 17909, ATCC 17946, ATCC 17969, ATCC 17923, Ml53 NCTC 5866, ATCC 17968, ATCC 9957, ATCC 17987, ATCC 17910 M198, Ml 13:2 M174, R1050, M51, M58b M17694, M16595, M16503, M109, Ml52 ATCC 17903 MlOO, ATCC M89, 17905,53893-82, M71, Ml 53937bb 14

Sau3AI 1 2 3 4 5

profiles

Hinff 1 1 2 3 3

6a-e 7a-e 8 9 lOa-e 11 12 13 14 15 17 16

TU15 B;;4

Ml18 M151a 382

18 20 1.9

BJ15 BJ16 BJ17 “Between 1 & 3” “Close to TU13” 93 9,

79 78 942 10169, 10095 10090 5804

21 22 23 24 25 4

4 4

3 4 4

780

M.I. GARCIA-ARATA

PCR amplification

and restriction

kbp

analysis

This was done as described by Ibrahim et al. (1996) with modifications. In a total volume of 100 ~1, the following reagents were added: dNTP (200 pM each), 3.3x XL buffer (Perkin Elmer), 30 ~1; primer 1 (5’ GAG TTT GAT CCT GGC TCA 3’), 20 pmol; primer 2 (5’ CCG GTC CTC TCG TAC T 3’), 20 pmol; Mg(OAc), (25mM) 6.4 ~1; DNA, loo-150 ng; distilled water to 100 ~1. The mix was covered with 100 pl mineral oil, then denatured at 94°C for 3 min. Then, 2 pl of rTth DNA polymerase XL (Perkin Elmer) were added. Thirty-five cycles were applied as follows: 94°C. 1 min.; 52”C, 1 min.; 72”C, 4 min.; with a final extension at 72°C for 10 min. After amplification, 10 pl of the PCR amplificate of each strain were run on a gel, stained with ethidium bromide (1 mg/l), and viewed on a UV transilluminator. Based on the intensity of the bands, 3 to 10 pl of the remaining PCR amplificate were digested using one of the enzymes AZuI, HaeIII, HhaI, Hid, MboI, MspI, NciI, PstI, RsaI, Sau3A1, SspI or TaqI following the manufacturers’ directions. The restriction bands were separated on a 3% “Metaphor” agarose (FMC, Vallensbaek, Denmark) in 0.5 x TBE at 200 volts for 3 h. A 123-bp (Gibco/BRL) or a mixture 100/20-bp ladder (Advances Biotech) was used as molecular weight marker. The gels were stained in ethidium bromide, viewed on a UV transilluminator and photographed. The patterns were then visually compared.

RESULTS One fragment,

ET AL.

4.5 kb in size, was obtained by PCR with all the strains, as shown in figure 1. Eleven restriction enzymes (AluI, HaeIII, HhaI, Hi&I, MspI, NciI, PstI, RsaI, Sau3A1, S,spI, and EzqI) were first tested with 10 strains belonging to different DNA groups to establish the most discriminating enzyme(s). Only one enzyme, Suu3A1, was able to provide a different pattern for each of these strains. Subsequently, this enzyme was used to digest all initial 67 collection strains. Twenty-eight different fragments were observed, yielding 25 different profiles of 7 to 11 fragments in each, 135 to 1,230 bp in size, respectively. Four fragments of approximately 185, 245, 330, and 590 bp were present in all strains. Strains belonging to DNA groups 1, 3, 6, TU13 and TU15 displayed two profiles each (figures 2

1

2

3

4

5

6

7

8

9

10

11

23.1 6.5 4.3 2.3 2.0

Fig.

1.

Amplification of the rDNA gene from different DNA groups in Acinetobacter.

Lane l=molecular weight marker (lambda phage cut with HindHI); lanes 2-ll=rDNA amplicons from strain

number(DNA group in parenthesis)39 (6), ATCC 17923 (7), Ml98 (lo), Ml74 (ll), M58b (ll), M 16503 (12), M 152 (12), 53893-82 53937bb (TU13), respectively.

Ml6595

(12),

(TU13), and

and 3), and groups 4, 5, and 7 displayed variant profiles with less intensive additional bands (profiles 6a-e, 7a-e, lOa-e, in figure 4) . These additional bands were consistently observed when PCR amplification and digestion were run in duplicate. The remaining 12 groups had a unique profile. The different profiles obtained with Sau3AI enzyme are shown in figures 1, 2, and 3. Identical profiles were obtained by using the enzyme MboI, which is an isoschizomer of Sau3AI. With the exception of strain number 5804 (DNA group “close to TU13”), which along with Sau3A1, exhibited profile 4 also displayed by most DNA group 3 strains, all Suu3AI profiles were DNA group-specific. Strain number 5804 could be differentiated from strains in the DNA group 3 by using the Hi&I restriction enzyme (table II). To confirm the ability of this simple system for DNA grouping of clinical strains, twentyfive strains were identified by Suu3AI digestion of the PCR-amplified fragment. They were identified as belonging to DNA groups 2, 3, 4, 5, 7, 8/9, 11, 12, and TU13, as shown in table III. In all cases, there was complete agreement with the classification previously obtained by conventional ribotyping/DNA hybridization identification.

ACINETOBACTER

Ml 2 12346

1

DNA GROUP

I6

3

4 Ml 67

4 8

16 9

DIFFERENTIATION

S 10

24 11

25 Ml hl 12 13 14

In2 IS

Id 16

BY ARDRA

IL? 17

h4 IS

Ml 19

781

M 20

123

Fig. 2. Sau3AI restriction

analysis of rDNA amplicons in the A. calcoaceticus-A. baumannii complex strains. Lanes 1, 7, 13 and 19=molecular weight marker (123-bp ladder); lanes2-6= strainsno. M42 (DNA group 1), A344 (l), ATCC 17903(TU13), ATCC 9955 (2) andMl62 (3), respectively; lanes %12=M102 (DNA group 3), 53937bb(T’U13), ATCC 19004 (3), 10169 (“Between 1 & 3”) and 10090(“Close to TU13”), respectively; lanes 14-18 and 20=profiles obtainedwith Hid. Arabic numbersin the columnheadscorrespondto the profile numbersin tablesII and III.

DISCUSSION

Acinetobucter spp. are increasingly important nosocomial pathogens. In order to investigate the origin and dispersal of Acinefobacter sp. in hospitals, a simple, reliable identification system is needed. Phenotypic methods are of little use in this context (Bouvet and Grimont, 1986; GemerSmidt ef al., 1991) and standard genotypic procedures are either too labour-intensive or not fully discriminatory to be really useful for routine identification purposes. The former is the case with the “gold standard” technique, DNA-DNA hybridization, where test strains must be hybridized to DNA from a panel of reference strains. The same is true for ribotyping based on restriction fragment length polymorphism in fragments carrying 16 and 23s genes. This technique has been proved to be taxonomically useful (Gemer-

Smidt, 1992), but has not been validated for DNA groups outside the A.caZcoaceticus-A.buumannii complex. Several PCR-based methods have been described. Restriction analyses of the spacer region between 16s and 23s rRNA genes have been performed on strains from 17 of the 20 DNA groups described (Nowak et al., 1995), and particularly for strains belonging to the A.culcoaceticus-A. baumannii complex (Dolzani et al., 1995; Garcia-Arata, 1995). This method was not sufficiently discriminatory in that some profiles were common to more than one DNA group. The same problem is also present with tRNA fingerprinting (Ehrenstein et al., 1996). By restriction analysis of the recA gene using two restriction enzymes, Nowak and Kur (1995) were able to discriminate between the 17 DNA groups described by Bouvet and Grimont (1986), and Bouvet and Jeanjean (1989). However, the com-

782

M.I. GARCfA-ARATA

prom he

Ia?

11

12

14

17

18

19

Ml

ET AL.

2U

21

22

23

24

25

Ml

b

492 369

123

Fig. 3. Sau3AI restriction

analysis of rDNA

amplicons

in different DNA groups in Acinetobact.cr.

Lanes 1 and 15=molecularweight marker (100/20 bp ladder) ; lane 8 =molecular weight marker (123 bp ladder); lanes 2-7=ATCC 9957 (DNA group 8/9), Ml98 (lo), Ml98 (lo), Ml6595 (12), ATCC 17905 (TU14), Ml18 (TU15) and M151a (TU15), respectively; lanes 9-14=382 (DNA group BJ14), 79 (BJ15), 78 (BJ16), 942 (BJ17), 10095 (“Between 1 & 3”) and 10090 (“Close to TU13”), respectively. Arabic numbers in the column heads correspond to the profile numbers in tables II and III.

mon and clinically significant DNA group TU13 (Gerner-Smidt, 1992; Gemer-Smidt and Tjemberg, 1993; Tjemberg and Ursing, 1989) from the A.calcoaceticus-A. baumannii complex was not included in this study. The selective restriction fragment amplification typing method described by Janssen et al. (1994) is highly discriminatory, but is too complicated to be applied in the routine laboratory setting. Vaneechoutte et al. (1995) used restriction analysis of the 16s rRNA gene (ARDRA) to identify Acinetobacter. Three enzymes, CfoI, AZuI, and Sau3A1, were necessary to discriminate between strains in the A.calcoaceticus-A. baumannii complex, and even with these three enzymes, three pairs of other DNA groups were indistinguishable from one another. We have further developed this ARDRA method by extending the region to be analysed to the 16s and the 23s rRNA gene and the spacer inbetween (Ibrahim et al., 1996). In that study, strains from DNA groups 1 (A.caZcoaceticus),

2 (A.baumannii) and 3, and TU13, all belonging to the A.caZcoaceticus-A.baumannii complex, could be differentiated using the enzyme HaeIII. However, in the present study, we found that this enzyme was not optimal for discrimination between all the other DNA groups within the genus. Fortunately, specific patterns were detected using another restriction enzyme, Sau3AI, or its isoschizomer MboI, for most DNA groups. One strain from the DNA group “close to TU13” showed the main pattern of strains from DNA group 3. However, only two isolates of the DNA group “close to TUl3” have been described so far, whereas DNA group 3 is a common pathogen (Gerner-Smidt et al., 1991; Gemer-Smidt and Tjernberg, 1993; Tjernberg and Ursing, 1989). Thus, this correspondence of patterns is not relevant in most instances. If it is judged to be important to discriminate between these two DNA groups, the enzyme Hinfl may be used.

ACINETOBACTER

DNA GROUP

pmfkM6adb6~6d6eMZ?a7b

Fig. 4. Suu3Al

restriction

7r

analysis

of rDNA

DIFFERENTIATION

7d

7e

amplicons

M,

I@. lob

loI

in different

BY ARDRA

l&I

1oC Ml

12

783

13

DNA groups 4, 5 and 7 in

Acinetobacter.

Lanes 1, 13 and 19=molecular weight marker (123 bp ladder) ; lane 7 =molecular weight marker (100/20 bp ladder); lanes 2-6=DNA group 1; lanes 8-12=DNA group 5; lanes 14-18=DNA group 7 ; lanes 20 and 21 =DNA groups 10 and 11, respectively. Arabic numbers in the column heads correspond to the profile numbers in table II.

Table III.

ARDRA identification of 25 clinical isolates of Acinetobacter.

Strain no.

Profile

16 15 5 (Suu3AI) 3 (Hid-I) 4 (Suu3AI) 12398, 10790 3 (Hid-I) 10508, 10074 3 6 U119, 274 M22, Ml 17, M140, R204 7 4, 164,65219-84 10 10086 3 UlOO, 284,286, 65109-84 11 13 225 14 70819-85, 266

9836

10716 9907, 12174a

Identification TU13 TU13 3 3 2 4 5 7 2 8J9 11 12

In the classical paper by Bouvet and Grimont (1986), strains with the A. Zwofii phenotype were split into the two DNA groups 8 and 9. Using a different DNA hybridization procedure, Tjernberg and Ursing (1989) could not reproduce this finding, and since then it has been debated as to whether to consider DNA groups 8 and 9 as a single species or not. In the present study, we could not differentiate between strains from these DNA groups, thus, supporting the former opinion. In conclusion, ARDRA, encompassing the 16s and the 23s rRNA genes and the spacer in between, using the enzyme Suu3AI, is a simple, highly accurate and rapid method for identification of Acinetobacter strains at the DNA group level.

M.I. GARCiA-ARATA

784

Acknowledgements We Paris), Leiden, sity of of the critical

wish to thank Dr. P.J.M. Bouvet (Institut Pasteur, Dr. L. Dijkshoorn, (University Hospital Leiden, The Netherlands) and Dr. I. Tjemberg, (UniverLund, Malmo, Sweden) for the donation of some reference strains. We gratefully acknowledge the comments of Dr. Luis de Rafael.

Analyse de restriction de 1’ADNr 16s et 23s amplifiC par PCR pour l’identitication de souches de Acinetobacter au niveau du groupe gknomique Le genre Acinetobacter est plutot homogbne sur le plan du phenotype mais httdrog&ne sur le plan du genotype. Nous avons CtudiCici une methode simple basee sur l’analyse par restriction dun grand fragment (45 kb) arnplifie par PCR de la majeure partie de l’op&on ribosomique (genesribosomiques 16s et 23s et espace intergenique). Nous avons examine une collection de 67 souches appartenant aux 20 groupes ADN proposesjusqu’en 1993. En utilisant l’enzyme Sau3A1, nous avons obtenu 25 profils d’ADN. Les souches appartenant aux groupes ADN 1, 3, 6, TU13 et TU15 ont r&Cl& deux protils chacune, et les groupes ADN 4, 5 et 7 ont rev% des profils dont les variants ont des bandes supplementaires moins intenses. Les 12 groupes ADN restants r&&lent 12 profils differents. Les profils obtenus sent specifiques du groupe ADN, except& pour un profil qui est partage entre le groupe ADN 3 et un groupe proche innomt, rarement rencontre ; ces deux groupes ADN peuvent &tre differencies a l’aide cliniques

de l’enzyme suppldmentaires

Hinfl. Vingt-cinq isolats prdalablement caractBrisCs

par hybridation ADN-ADN ont et& selectionnes en double-aveugle pour identification de leur groupe genomique dans le but de tester la fiabilitt de la m&rode. Toutes les souches ont et& correctement identifiees au niveau du groupe ADN. En conclusion, pour l’identification des souches de Acinetobutter, cette m&ode d’analyse du groupe ADN est B la fois rapide et fiable. Mets-cl&: Acinetobacter, ADNr, ARDRA; Diagnostic, Enzymes de restriction.

References Bouvet, P.J.M. & Grimont, P.A.D. (1986), Taxonomy of the genus Acinetobacter baumannii sp. nav., Acinetobarter haemolyticus sp. nov., Acinetobacter johnsonii

ET AL.

sp. nov., and Acinetobacter junii sp. nov. and emended descriptions of Acinetobacter calcoaceticus and Acinetobacter lwo#?i. Int. J. Syst. Bacterial., 36, 228-240. Bouvet, P.J.M. & Jeanjean, S. (1989), Delineation of new proteolytic genomic species in the genus Acinetobatter. Res. Microbial., 140, 291-299. Dolzani, L., Tonin, E., Lagatolla, C., Prandin, L. & MontiBragadin. C. (1995), Identification of Acinetobacter isolates in the A. calcoaceticus-A. baumannii complex by restriction analysis of the 16S-23s rRNA intergenic spacer sequences. J. Clin. Microbial., 33, 1108-1113. Ehrenstein, B., Bemards, A.T., Dijkshoom, L., GemerSmidt, P., Towner, K.J., Bouvet, P.J.M., Daschner, F.D. & Gmndmann,H. (1996).Acinetobacter species identification by using tRNA spacer fingerprinting. J. Clin. Microbial., 34, 24142420. Garcia-Arata, M.I. (1995), Molecular methods applied in epidemiological studies of Acinetobacter spp. in a hospital. [Estudio epidemiolbgico de Acinetobacter sp. mediante dcnicas moleculares en un medio hospitalario]. Ph. D. Thesis.Faculty of Pharmacy, Complutense University, Madrid. Gemer-Smidt, P. (1992), Ribotyping of the Acinetobacter calcoaceticus-Acinetobacter baumannii complex. J. Clin. Microbial., 30, 2680-2685. Gemer-Smidt, P., Tjemberg, I. & Ursing, J. (1991). Reliability of phenotypic tests for identification of Acinetobacter species. J. Clin. Microbial., 29,277-282. Gemer-Smidt. P. & Tjemberg, I. (1993). Acinetobacter in Denmark: molecular studies of the Acinetobucter calcoaceticus-Acinetobacter baumannii complex. APMIS, 101, 826-832. Ibrahim, A., Gerner-Smidt, P. & Sjostedt. A. (1996). Amplification and restriction endonuclease digestion of large fragment of genes coding for rRNA as a rapid method for discrimination of closely related pathogenic bacteria. J. Clin. hficrobiol., 34, 28942898. Janssen,P., Dijkshoorn, L., Maquelin, K., Coopman, R., Bouvet, P., Tjernberg, I., Zabeau, M. & Kersters, K. (1994), High resolution fingerprinting of Acinetobacter strains belonging to different genomic species. Abstract, 3rd International Symposium on the Biology of Acinetobacter. Edinburgh. Nishimura, Y., Ino, T. & Hzuka, H. (1988), Acinetobacter radioresistens sp. nov. isolated from cotton and soil. Int. J. Syst. Bacterial., 38, 209-211. Nowak, A., Burkiewicz, A. & Kur, J. (1995), PCR differentiation of seventeen genospecies of Acinetobacter. FEMS Microbial. L&t., 126, 181-188. Nowak, A. & Kur, J. (1995), Genomic speciestyping of acinetobacters by polymerase chain reaction amplification of the recA gene. FEMS Microbial. Lett., 130, 327-332. Tjemberg, I. & Ursing, J. (1989), Clinical strainsof Acinetobacter classified by DNA-DNA hybridization. APMIS, 97, 595-605. Vaneechoutte, M., Dijkshoorn, L., Tjernberg, I., Elaichouni, A., de Vos, P., Claeys, G. & Verschraegen, G. (1995), Identification of Acinetobacter genomic species by amplified ribosomal DNA restriction analysis. J. Clin. Microbial., 33, 11-15.