RAPD-PCR typing of Acinetobacter isolates from activated sludge systems designed to remove phosphorus microbiologically

RAPD-PCR typing of Acinetobacter isolates from activatedsludge systems designed to remove phosphorusmicrobiologically

E. Carr1, H. Eason1, S. Feng1, A. Hoogenraad2, R. Croome2, J. Soddell1,K. Lindrea1 and R. Seviour1

1Biotechnology Research Centre, La Trobe University, Bendigo, Victoria, and 2Department of Environmental

Management and Ecology, La Trobe University, Wodonga, Victoria, Australia

514/8/00: received 18 August 2000 and accepted 31 October 2000

E. CARR, H. EASON, S. FENG, A . HOOGENRAAD, R. CROOME, J . SODDELL, K . L INDREA

AND R. SEVIOUR. 2001.

Aims: This study investigated whether there were differences in RAPD ®ngerprints between

already described genomic species of Acinetobacter and those from activated sludge systems.

Whether plant-speci®c populations of acinetobacters exist was also examined.

Methods and Results: Fifty-two isolates of Acinetobacter from four biological phosphorus

removal (EBPR) systems of different con®gurations, and the known genomic species, were

characterized using RAPD-PCR, and fragments separated on agarose gels. Patterns were

analysed using Gel Pro software and data analysed numerically. RAPD-PCR produced patterns

suggesting that many environmental isolates differ from known genomic species. In two cases,

strains from individual plants clustered closely enough together to imply that there may be

plant-speci®c populations of acinetobacters.

Conclusions: The data suggest that current understanding of the taxonomic status of

Acinetobacter may need modifying to accommodate non-clinical isolates, as many of the clusters

emerging after numerical analysis of RAPD-PCR fragments from activated sludge isolates were

quite separate from the clusters containing the already described genomic species. Some

evidence was also obtained from the clusters generated to support a view that particular

populations of Acinetobacter may occur in individual activated sludge plants.

Signi®cance and Impact of the Study: These data suggest that the current understanding

of the systematics of Acinetobacter, based as it is almost exclusively on clinical isolates, may need

drastic revision to accommodate environmental strains. They also suggest that a re-examination

of the importance and role of Acinetobacter in the activated sludge process may be appropriate.

INTRODUCTION

Members of the genus Acinetobacter are ubiquitous organ-

isms of increasing interest and importance to clinical

microbiologists, as some are serious nosocomial pathogens

and show increasing resistance to many commonly pre-

scribed antibiotics (Bergogne-Berezin and Towner 1996).

This genus has a very confused taxonomic history but is now

placed in the c-Proteobacteria on 16S rDNA data; it is

considered to be phylogenetically coherent (Rainey et al.1994; Ibrahim et al. 1997) and can be delineated unequivo-

cally using the transformation assay of Juni (1972).

The species delineation of Acinetobacter still poses prob-

lems, even after Bouvet and Grimont (1986), Bouvet and

Jeanjean (1989) and Tjernberg and Ursing (1989) recognized

19 genomic species using DNA:DNA hybridization. Only

seven of these species have been validly named. Tjernberg and

Ursing (1989) showed that their DNA group 14 corresponded

to DNA group 13 of Bouvet and Jeanjean (1989), while their

DNA group 15 was novel and quite different to the BouvetCorrespondence to: Dr R. Seviour, Biotechnology Research Centre, La Trobe

University, Bendigo, Victoria 3552, Australia

(e-mail: [email protected]).

ã 2001 The Society for Applied Microbiology

Journal of Applied Microbiology 2001, 90, 309±319

and Jeanjean (1989) group 15. Tjernberg and Ursing (1989)

also proposed that their DNA group 13 was closely related to

the Bouvet and Grimont groups 1, 2 and 3, and they and

colleagues have proposed that all four should be amalgamated

into a single species, A. calcoaceticus. Tjernberg and Ursing

(1989) also suggested their DNA groups 10 and 11 were most

closely related to each other, as were groups 4 and 6 and

Bouvet and Jeanjean (1989) groups 15 and 16. However, such

interspeci®c relationships have not always been con®rmed

when other characterization methods, both phenotypic and

molecular, have been used (e.g. Nowak and Kur 1995;

Bergogne-Berezin and Towner 1996; Vila et al. 1996; Dijk-

shoorn et al. 1998; Jawad et al. 1998; Koeleman et al. 1998).

Certainly, 16S rDNA sequence comparisons (Rainey et al.1994) gave different interspecies relationships to some of

these (e.g. Rudant et al. 1999), although Yamamoto et al.(1999) obtained a better, if not complete, agreement with the

DNA groupings using gyr B DNA sequence comparisons. As

a consequence of the inconsistencies arising from the many

different typing methods used for Acinetobacter, reliable

methods for unequivocally identifying many members of this

genus are still not available. Furthermore, it is now clear that

more clinical genomic species than those currently recognized

from the collective work of Bouvet and Grimont (1986),

Bouvet and Jeanjean (1989) and Tjernberg and Ursing (1989)

exist (Dijkshoorn et al. 1998; Yamamoto et al. 1999).

Acinetobacters are readily isolated from activated sludge

systems removing phosphorus (Beacham et al. 1990; Mino

et al. 1998). There is some evidence that genomic species 5,

7 and 8/9 are more commonly seen in these environmental

samples (Wiedmann-Al-Ahmad et al. 1994; Knight et al.1995) than DNA groups 2 and 3 and Tjernberg and Ursing

(1989) group 13, which predominate in clinical species.

Identi®cation methods developed for the clinical isolates are

unlikely always to be applicable to environmental isolates

(Soddell et al. 1993). It is also likely that many novel isolates

of Acinetobacter await description from habitats other than

clinical specimens (Vaneechoutte et al. 1999), so it is

probable that we still understand little about the ecology

and true biodiversity of this genus (Dijkshoorn et al. 1998).

The importance of Acinetobacter in biological phosphate

removal in activated sludge systems, once widely accepted

(e.g. Jenkins and Tandoi 1991), has been questioned from

studies using 16S rRNA sequences as determinants of

biodiversity (Wagner et al. 1994; Bond et al. 1995, 1999;

Kampfer et al. 1996; Christensson et al. 1998).

The current study attempted to understand better the

biodiversity and ecology of Acinetobacter in four enhanced

biological phosphate removal (EBPR) activated sludge

systems, by clarifying their taxonomic status in relation to

the currently recognized DNA groups. The characterization

method chosen was RAPD-PCR, which has been success-

fully applied to epidemiological studies with bacteria, and in

resolving interspeci®c relationships in other genera (Pacheco

et al. 1996; Moschetti et al. 1998; Vickery et al. 1998). The

antibiotic resistance patterns of individual isolates was also

determined to see if these patterns differed in populations

from the different plants.

MATERIALS AND METHODS

Cultures of Acinetobacter used in study

Four different EBPR plants were sampled at regular (usually

fortnightly) intervals from April to December in 1995.

These plants were at Bendigo, Ballarat and Wodonga in

Victoria and Albury in NSW, Australia. Their plant

operational parameters are given in Table 1, and all were

removing phosphate during the study period. The methods

for isolating Acinetobacter spp. from biomass samples within

24 h of taking them were those described by Knight et al.(1995). All the Gram-negative, oxidase-negative coccobacilli

obtained were then screened using the Juni (1972) transfor-

mation assay for members of the genus Acinetobacter. The 52

isolates used in this study were selected randomly from the

total number obtained to re¯ect different sampling times.

All were stored at )80°C. Those from Albury, Bendigo,

Ballarat and Wodonga are pre®xed with A (11 strains), B (16

strains), C (12 strains) and D (13 strains), respectively. Also

included were the 19 recognized genomic species, kindly

supplied by Prof. Bouvet, Institut Pasteur, Paris (referred

to as BG strains) and Prof. Tjernberg, Lund University,

Sweden (TU strains). Their sources and culture collection

numbers are given in Table 2. All the environmental isolates

were screened for their ability to accumulate polyphosphate

as detailed by Beacham et al. (1992).

Phenotypic identi®cation of Acinetobacterisolates to species level

The Biolog GN system was used in attempts to characterize

and identify the isolates con®rmed as Acinetobacter spp.

Table 1 Operation parameters of the four activated sludge plants

used in study

Wastewater treatment plants

Albury Ballarat Wodonga Bendigo

P mg g)1 (MLSS) 24á34 38á11 14á03 31á26

P mg l)1 117á75 87á00 133á25 126á58

Typical MLSS mg l)1 4800 2300 9500 4100

MLVSS/MLSS 0á77 0á80 0á82 0á80

Sludge age (days) 16 (est) 3±5 18 25

P = Phosphorus; MLSS = Mixed Liquor Suspended Solids; MLVSS

= Mixed Liquor Volatile Suspended Solids.

310 E. CARR ET AL .

ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 309±319

These isolates were also identi®ed using the phenotypic

identi®cation schemes described by Soddell et al. (1993).

Numerical analysis of the Biolog data was carried out

and dendrograms drawn based on SSM coef®cients and the

UPGMA algorithm using the NTSYS-pc software (Exeter

Software, New York, USA).

DNA extraction for RAPD-PCRof Acinetobacter isolates

All Acinetobacter isolates were grown in nutrient broth

(Oxoid) at 30°C for 48 h with shaking, and cells were

harvested by centrifugation. Pellets were resuspended in

TE buffer (10 mmol l)1 Tris; 1 mmol l)1 EDTA, pH 8),

and cell lysis occurred with addition of 30 ll 10% v/v

sodium dodecyl sulphate (SDS) plus 3 ll proteinase K

(20 mg ml)1, Sigma). Many of these isolates produced

substantial capsular material which seemed to interfere with

cell lysis and DNA extraction. Therefore, 100 ll 5 mol l)1

NaCl, and 80 ll 1% w/v cetyl trimethyl ammonium

bromide (CTAB) in 0á7 mol l)1 NaCl, were added, and the

mixture incubated at 65°C for 10 min before extraction with

phenol:chloroform:isoamyl alcohol (25:24:1). After centrif-

ugation, the aqueous phase was again extracted with

chloroform:isoamyl alcohol (24:1). Genomic DNA was then

precipitated with ice-cold ethanol; the DNA was collected

by centrifugation, washed in cold 70% (v/v) ethanol and air-

dried before being resuspended in TE buffer. Finally, 3 ll

RNAse (1 mg ml)1 Sigma) were added and the mixture

incubated at 37°C for 30 min. The DNA was again

precipitated with ethanol as above and the pellet air dried

and redissolved in TE buffer overnight. Purity of DNA was

checked by the A260:A280 nm ratio which, if greater than 1á8,

was taken to indicate that purity was acceptable. If below

1á8, the DNA solution was re-extracted with chloroform:iso-

amyl alcohol (24:1).

Genomic ®ngerprinting of Acinetobacterusing RAPD-PCR

Seven primers, all obtained from Bresatec (Adelaide,

Australia), were used in this study. Four were discarded as

when screened, the reproducibility of the band patterns

obtained with the known genomic species was poor between

replicate runs (data not presented). Primers 5¢GCTTGTGAAC 3¢ and 5¢ CGGCAGCGCC 3¢ gave

reproducible patterns but often, only a small numbers of

bands was generated with many of the strains (data not

presented). The one eventually chosen for this study was the

decamer 5¢ GCTTGTGAAC 3¢ because it gave the highest

number of bands with the known genomic species in

screening trials. The importance of standardizing genomic

®ngerprinting protocols has been convincingly demonstrated

by Koeleman et al. (1998), so the following method was

Table 2 Known genomic species of

Acinetobacter used in this studyAbbreviation used in study Species name Origin/Culture collection numbers

BG1 Acinetobacter calcoaceticus ATCC 23055T/CIP 81á08T

BG2 Acinetobacter baumannii ATCC 19606T/CIP 70á34T

BG3 Acinetobacter sp. 3 ATCC 19004/CIP 70á29

BG4 Acinetoabcter haemolyticus ATCC 17906T/CIP 64á3T

BG5 Acinetobacter junii ATCC 17908T/CIP 64á5T

BG6 Acinetobacter sp. 6 ATCC 17979/CIP A165

BG7 Acinetobacter johnsonii ATCC17909T/CIP 64á6T

BG8 Acinetobacter lwof®i NCTC 5866T/CIP 64á10T

BG9 Acinetobacter sp. 9 ATCC 9957/CIP 70á31

BG10 Acinetobacter sp. 10 ATCC 17924/CIP 70á12

BG11 Acinetobacter sp. 11 ATCC 11171/CIP 63á46

BG12 Acinetobacter radioresistens SEIP 12á81

BJ13 Acinetobacter sp. 13 (BJ) ATCC 17905/CIP 64á2BJ14 Acinetobacter sp. 14 (BJ) K.Irino 105/85

BJ15 Acinetobacter sp. 15 (BJ) M.M. Adam Ac606 180 : 40 va

BJ16 Acinetobacter sp. 16 (BJ) ATCC 17988/CIP 70á18

BJ17 Acinetobacter sp. 17 (BJ) SEIP Ac87á314

TU13 Acinetobacter sp. 13 (TU) ATCC 17903

TU14 Acinetobacter sp. 14 (TU) ATCC 151a1

TU15 Acinetobacter sp. 15 (TU) ATCC 71

ATCC = American Type Culture Collection, Maryland, USA; CIP = Collection de l'Institut

Pasteur, Paris; SEIP = Collection du Service des Enterobacteries, Institut Pasteur, Paris.

ACINETOBACTER FROM ACTIVATED SLUDGE 311

ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 309±319

rigorously adhered to for all the analyses in this study. In

most cases, DNA that was extracted at least twice from each

strain almost always gave identical RAPD patterns, even

when analyses were carried out up to ®ve months apart. Any

occasional differences noticed in the patterns between runs

of the same strain were limited to presence or absence of one

or two faint bands (e.g. as in lanes K and L in Fig. 2, which

are the poorest replicate patterns obtained in this study).

Some of the data obtained for replicate runs after numerical

analysis of the patterns for several strains are presented in

the dendrograms given in Figs 3 and 5.

PCR reaction mixes contained 2á5 ll of a 10´ reaction

buffer (200 mmol l)1 (NH4)2SO4; 750 mmol l)1 Tris HCl

buffer pH 9; 0á1% v/v Tween (80), 2á5 ll 2 mmol l)1

deoxynucleotide triphosphate mixture (Pharmacia), 3á5 ll

25 mmol l)1 MgCl2, 0á9 ll 20 lmol l)1 primer (Bresatec),

2 units Taq polymerase (Integrated Sciences, Melbourne)

and 100 ng DNA template from each Acinetobacter isolate.

DNA ampli®cation was achieved with a Hybaid Omnigene

thermocycler (Integrated Sciences, Melbourne, Australia)

and the following protocol. After initial denaturation at 94°C

for 3 min, followed by 34 cycles at 92°C for 30 s, 40°C for

60 s and 72°C for 3 min, a ®nal single cycle at 92°C for 30 s,

40°C for 60 s and 72°C for 10 min was performed.

Ampli®cation products were run on 1% agarose gels with

1´ TAE buffer at 90 V for 1á5 h, and bands visualized with

ethidium bromide. Gels were viewed and photographed

with a Polaroid/FotoDyne (FotoDyne, Hartland, Wiscon-

sin, USA) camera illuminator using Polaroid 665 ®lm.

Negatives of gel images were scanned into Adobe Photoshop

(Macintosh version 4á01), and the gel patterns analysed

using Gel Pro Analyser (Media Cybernetics, Silver Spring,

Maryland, USA), which could compensate for slightly

slanted gel runs. One Acinetobacter strain, BG 16, was run

on all gels as an internal marker, and all bands from other

isolates normalized against those of this isolate using GelPro

Analyser (Media Cybernetics, Silver Spring, Maryland,

USA). Numerical analyses and dendrogram constructions

were carried out as described earlier.

Antibiotic resistance patterns for activatedsludge Acinetobacter isolates

Antibiotic resistance for selected isolates was determined

using the Vitek AMS system (bioMerieux Vitek, Melbourne,

Australia) and the GNS-406 susceptibility card containing

the following antibiotics: amikacin, amoxicillin/clavulanic

acid, ampicillin, aztreonam, cefotaxime, ceftazidime, ceph-

alothin, cipro¯oxacin, gentamycin, imipenem, nitrofuran-

toin, nor¯oxacin, tricarcillin/clavulanic acid, tobramycin,

trimethaprim/sulfamethoxazole and trimethoprim. Resist-

ance or susceptibility was based on MIC values determined

for each antibiotic by Vitek (Vitek Technical Bulletin).

RESULTS

Phenotypic properties of the Acinetobacterisolates

All 52 strains used in this study were con®rmed as belonging

to the genus Acinetobacter with the Juni (1972) assay.

However, most could not be identi®ed to species level with

either the Biolog GN or the phenotypic characterization

schemes used earlier by Soddell et al. (1993). Of those

which were (<30%), most emerged as genomic species 7

(A. johnsonii), 11 and 15. Many which identi®ed as genomic

species BG15 with Biolog GN appeared as A. johnsonii when

the other schemes were used. Numerical analyses using the

95 Biolog characters for a selection of these strains and the

known genomic species produced a dendrogram (Fig. 1)

where few of the isolates from the four plants cluster closely

with the recognized genomic species of Bouvet and Grimont

(1986), Bouvet and Jeanjean (1989) or Tjernberg and Ursing

(1989). Isolates from individual plants were usually scattered

widely throughout the dendrogram, although some strains

from a single plant (e.g. C2, C5, C6 and C7 from Ballarat,

and D2, D5, D6, D7, D12 and D13 from Wodonga), and

occasionally different plants (e.g. A6, A15 and B9, and B17,

D4 and D11), were very closely related to each other on

their substrate utilization patterns. Less than 50% of these

isolates (23/52) could store polyphosphate when examined

in pure culture with the methylene blue and Neisser stains,

and these did not belong to any particular genomic species

where identi®cation was possible.

PCR ®ngerprinting of known genomic speciesof Acinetobacter

The patterns obtained with the primer 5¢ GCTTGTGAAC

3¢ after RAPD-PCR of the known genomic species are

shown in Fig. 2. Most yielded a distinctive RAPD ®nger-

print with the primer used, and the relationships between

them following numerical analysis of these patterns are

shown in the dendrogram (Fig. 3). Brie¯y, all strains cluster

at an SSM of 0á55. Several distinct clusters emerged and

among these, isolates BG1, BG2 and BG3, BG7, BG8 and

BG9, and BJ14 and BJ15, clustered most closely together.

Only one strain of each of these known genomic species was

examined here.

RAPD-PCR of activated sludge Acinetobacterisolates

Examples of the patterns obtained from the Acinetobacterisolates from the four different plants are given in Fig. 4 and

the dendrogram constructed after numerical analysis of

these, and including the known genomic species, is presen-

312 E. CARR ET AL .

ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 309±319

ted in Fig. 5. They show that generally, the environmental

strains produced patterns quite different to those of the

known genomic species and thus, did not cluster closely

with them. There were a few exceptions (see below).

In general, isolates from each particular EBPR plant

tended to cluster together, as shown with those from Ballarat

(C isolates) which form a single group except for C8 and

C12, the latter being most closely related to some Bendigo

isolates. The Albury (A) strains, like the Bendigo (B)

isolates, fall into one quite tight cluster or are distributed

throughout the dendrogram. Strains from Wodonga (D) fall

into three main clusters that are scattered throughout the

dendrogram, while some isolates e.g. D4, D11 and D13

appear to be identical to each other, even though obtained

from the plant at different times of the year. Only in a very

few cases do these activated sludge isolates form clusters at

high SSM values with the known genomic species. In some

cases, these groupings support those obtained after analysis

of the Biolog characterization data (Fig. 1), as in the case of

most of the isolates from Ballarat (C) and Wodonga (D

isolates), which clustered together in both.

Antibiotic resistance patterns of activatedsludge isolates of Acinetobacter

The antibiotic resistance pattern results determined by the

Vitek system are given in Table 3. Not all the isolates could

be tested with this system because many, although not

identi®ed as A. johnsonii with Biolog, failed to grow at 37°C,

the temperature at which this system operates. With the

possible exception of ampicillin and ceftazidine, very similar

resistance patterns were obtained for isolates from the four

plants, although the percentages of isolates resistant to each

of the individual antibiotics did vary. Thus, only isolates

Fig. 1 Dendrogram of Acinetobacter spp. after

numerical analysis with UPGMA and SSM of

data obtained from characterization of strains

using Biolog

ACINETOBACTER FROM ACTIVATED SLUDGE 313

ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 309±319

resistant to aztreonam, cephalophin, nitrofurantoin and

trimethoprim of the antibiotics were seen in all four plants.

With the possible exception of trimethoprim and ampicillin,

a higher percentage of Australian hospital Acinetobacterisolates have been recorded as having resistance to these

antibiotics too. None were resistant to imipenem, amakacin,

clavulanic acid supplemented ampicillin, cipro¯oxacin,

gentamycin, clavulanic acid supplemented tricarcillin and

trimethoprim/sulfa, although hospital isolates resistant to

all these except imipenem have been reported in Australia.

Fig. 2 RAPD-PCR patterns from currently recognized genomic species of Acinetobacter. For other details, see text. Lanes A, K, L and V: BJ16 (marker

strain); lanes B±J and M and N: BG 1±10 and BG 11 and 12 (respectively); lanes O±S: BJ13±BJ17 (respectively); lanes T and U: TU14 and 15

Fig. 3 Dendrogram of RAPD-PCR patterns

of the currently recognized genomic species of

Acinetobacter after their numerical analysis

using UPGMA and SSM. Results from ana-

lysis of duplicate strains are asterisked

314 E. CARR ET AL .

ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 309±319

None of the Acinetobacter isolates screened here were

identi®ed as members of the A. calcoaceticus±A. baumanniicomplex, i.e., those normally associated with clinical spec-

imens.

DISCUSSION

This study has investigated the biodiversity of Acinetobacterstrains from four activated sludge plants of different

con®gurations, but all of which were removing phosphorus.

It has determined, on the RAPD-PCR criteria used here,

that strains different to the described genomic species from

clinical habitats exist in activated sludge, an ecosystem long

known to contain large numbers of culturable Acinetobacterstrains (e.g. Knight et al. 1995). This supports earlier work

based on phenotypic (Soddell et al. 1993) and soluble

protein PAGE (Maszenan et al. 1997) characterization. Even

isolates obtained from clinical specimens have often fallen

outside the currently accepted taxonomic framework for this

genus (e.g. Gerner-Smidt and Frederiksen 1993).

Many different methods have been applied in attempts

to sort out the confused interspecies relationships within

Acinetobacter, but none give results which agree completely

with the relationships suggested from the DNA-DNA

hybridization data (Bouvet and Grimont 1986; Bouvet and

Jeanjean 1989; Tjernberg and Ursing 1989). The character-

ization method selected for the isolates in this study was

RAPD-PCR, which has been used successfully elsewhere to

resolve intrageneric relationships in bacteria, including

acinetobacters from activated sludge (Wiedmann-Al-Ahmad

et al. 1994). In some, if not all (e.g. Tynkkynen et al. 1999)

comparative studies, RAPD-PCR has proved to be as

discriminatory in achieving this with both prokaryotes and

eukaryotes as several of the other genomic ®ngerprinting

techniques available (Koeleman et al. 1998; Arenal et al.1999).

Results obtained here showed that PCR-RAPD could

differentiate between all the currently recognized genomic

species of Acinetobacter. Although BG8 and BG9 could not

be separated either phenotypically by Bouvet and Grimont

(1986) or using other criteria (e.g. Tjernberg and Ursing

1989; Vaneechoutte et al. 1995), they were distinguished

readily here on their RAPD patterns (Fig. 2). An ability to

separate BG8 and BG9 was reported by Weidmann-

Al-Ahmad et al. (1994) although in their RAPD study,

BG7 and BG9 gave the same RAPD patterns, as they did

also in the study by Nowak and Kur (1995) with REP-PCR.

Numerical analysis of the RAPD patterns from BG1, BG2,

BG3 and TU13 (Fig. 3) supports the close relationship

between BG1, BG2 and BG3 (but not TU13) proposed

earlier from other methods of characterization (Gerner-

Smidt 1992; Gerner-Smidt and Tjernberg 1993). Genomic

species 10 and 11, also considered very similar (Tjernberg

and Ursing 1989; Bernards et al. 1995; Ibrahim et al. 1997;

Maszenan et al. 1997; Vaneechoutte et al. 1999), gave easily

recognizable different RAPD patterns (Fig. 2). Other

genomic ®ngerprinting techniques (Wiedmann-Al-Ahmad

et al. 1994; Nowak and Kur 1995; Janssen and Dijkshoorn

1996) could also separate them.

The reasons for these variations in interspecies relation-

ships with different typing methods are not clear, but the

patterns obtained here were highly reproducible. It is

recognized that only a single primer for PCR-RAPD was

used and only one strain of each genomic species was

analysed in this study. Clearly, more strains of each of these

known genomic species should be analysed but as detailed

Fig. 4 RAPD-PCR patterns of representative strains of Acinetobacter obtained from activated sludge systems. Lanes A, F, K, L, P and U: BJ16

(marker strain); lanes B, C, D, E and G: C2, C4, C6, C8 and C10; lanes H, I and J: D4, D8 and D10; lanes M and N: A29 and A34; lanes O, Q, R, S

and T:B6, B10, B11, B17 and B19

ACINETOBACTER FROM ACTIVATED SLUDGE 315

ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 309±319

above, the literature suggests that disagreements like those

noted here are not uncommon within Acinetobacter charac-

terization studies.

Even allowing for the fact that only single isolates of each

of the known genomic species of Acinetobacter were used

here, it is thought that the main ®nding of this study still

holds. With a few exceptions (e.g. D10 and BJ15), the

activated sludge strains analysed here did not cluster closely

with them. Instead, these usually clustered quite separately,

as seen with cluster BG5, BG7, BG8, BG9, BJ13 and BJ14,

and cluster BG1, BG2, BG3, BG4, and BG12. (Fig. 5). Of

the 30% of strains which were identi®ed phenotypically

here, most emerged as BG7, BG11 and BJ15, and the former

is thought to dominate such communities of Acinetobacter(Kampfer et al. 1992; Knight et al. 1995; Guardabassi et al.1999). However, none of the activated sludge isolates

analysed here clustered closely with the BG7 type strain

(Fig. 5), so, with a few exceptions, it was not possible to

identify con®dently any of the unknown strains by compar-

ing their RAPD patterns with those of the known genomic

species. This technique can distinguish between different

strains of a single species with some bacteria (e.g. Soto et al.1999), although whether that is the case with Acinetobacterspp. is not known. The dendrogram (Fig. 5) would suggest

Fig. 5 Dendrogram of RAPD-PCR patterns

of all Acinetobacter isolates examined by

RAPD-PCR after their numerical analysis

using UPGMA and SSM. Results from ana-

lysis of duplicate strains are asterisked

316 E. CARR ET AL .

ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 309±319

that although some environmental strains cluster closely

enough to be considered members of a single known species

(allowing for some pattern variation within each species),

others are suf®ciently distant to be new species. This

opinion is supported by other published data on some of

these strains (Soddell et al. 1993; Maszenan et al. 1997).

The view expressed by Guardabassi et al. (1999), based as it

was on phenotypic identi®cations, that the current taxo-

nomic groupings of Acinetobacter probably embrace most of

the species likely to exist in aquatic habitats, appears

improbable. Further taxonomic characterization work,

including DNA-DNA hybridization, is required before the

systematics of these strains are resolved, but unpublished

data using 16S±23S spacer region polymorphism lend

further support for this hypothesis of the existence of novel

strains (Carr, unpublished). If true, then the current range

of 16S rRNA targeted probes used for FISH studies of

Acinetobacter which began to question their role in phos-

phate removal (e.g. Wagner et al. 1994; Snaidr et al. 1997)

may be inadequate for detecting all the strains that could be

there.

Whether plant-speci®c populations of Acinetobacter exis-

ted for the four systems examined was not totally clear,

although the data presented here suggest that their biodi-

versity based on RAPD patterns was lower within isolates

from two of the plants (e.g. Ballarat and Wodonga) than

from all four (Fig. 5). All plants were removing phosphate

during the sampling period and treating mainly domestic

wastes, but the sludge ages differed, e.g. 5 days in Ballarat

compared with 25 days in Bendigo, and the levels of

biodiversity were possibly higher in Bendigo and Albury

than in the other two plants. The reasons for this are

unclear. No other similar study using genomic ®ngerprint-

ing has looked at inter-plant variations in Acinetobacterpopulations, whether such methods might be used to

monitor phosphorus removal performance in such plants,

or whether the level of biodiversity is related to the

phosphate removing capacity of a plant. As mentioned

above, molecular studies continue to add support to the view

of little or no role for Acinetobacter in phosphorus removal

(e.g. Wagner et al. 1994; Bond et al. 1995, 1999). Further-

more, both Hesselman et al. (1999) and Crocetti et al. (2000)

have provided additional evidence in favour of the involve-

ment of a Rhodocyclus like organism (i.e. a b-Proteobacteria,

not c-Proteobacteria). Studies here using a combination of

¯uorescent in situ hybridization (FISH) with 16S rRNA

targeted probes (Wagner et al. 1994) and 33P microautora-

diography (Lee et al. 1999) have failed to support a strong

case for Acinetobacter in this process. Instead, the data

further support a role for the b-Proteobacteria as the PAOs

in the Bendigo plant (Carr and Liu unpublished). However,

it is unlikely that only a single bacterial population is

responsible, and individual plants may differ.

It was not surprising to detect multiple antibiotic

resistance in some of these isolates, in the light of the

widespread occurrence of antibiotic resistance in members of

this genus (Bergogne-Berezin and Towner 1996) and other

surveys of environmental strains of Acinetobacter (e.g.

Guardabassi et al. 1999). The patterns seen here were

similar to those of most hospital isolates screened in

Australia, a ®nding consistent with these resistant strains

being of human origin. Activated sludge systems have not

been examined previously for antibiotic-resistant Acineto-bacter strains, and only a small number of isolates was typed

in this study. Thus, how frequently they occur in such

ecosystems is not well understood. It is likely to be high, and

strains resistant to other antibiotics are probably present

there too. The opportunities for horizontal gene transfer

must be present (Ravatn et al. 1998), especially with plasmid

DNA readily detected in activated sludge (e.g. Bauda et al.1995). Acinetobacter spp. seem capable of taking up naked

DNA, which is plentiful in activated sludge (Frolund et al.1995), and being transformed under natural conditions

(Nielsen et al. 2000). The possibility that horizontal gene

transfer and/or mutational change might also explain the

levels of biodiversity revealed by RAPD, which appear to

exist in environmental strains of this organism, cannot be

ruled out, and that clonal populations of Acinetobactersimilar to those suspected for Helicobacter pylori (Wang et al.1999) may exist.

Table 3 Antibiotic resistance (expressed as percentage) among

Acinetobacter isolates from activated sludge using the Vitek GNS

406 system. Clinical data from the Australian Antimicrobial

Resistance Surveillance Program Annual Report 1993

Number of isolates/number identi®ed

Bendigo Ballarat Albury Wodonga

Clinical 15/2 7/0 8/2 4/0

Amikacin 10á7 0 0 0 0

Amoxicillin/ 27 0 0 0 0

clavulanic acid

Ampicillin 88á5 6á7 14á3 0 100

Aztreonam 100 100 88 50 100

Cefotaxime 30á3 13á3 0 0 0

Ceftazidine 11á5 6á7 0 0 0

Cephalophin 98á6 100 85á7 87á5 100

Cipro¯oxacin 18á6 0 0 0 0

Gentamicin 42á1 0 0 0 0

Imipenam 0 0 0 0 0

Nitrofurantoin 97á8 100 85á7 87á5 100

Tricarcillin/

clavulanic acid 12á6 0 0 0 0

Trimethoprim/

sulfa 6á3 0 0 0 0

Trimethoprim 20 100 85á7 87á6 100

ACINETOBACTER FROM ACTIVATED SLUDGE 317

ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 309±319

REFERENCES

Arenal, F., Platas, G., Martin, J., Salazar, O. and Pelaez, F. (1999)

Evaluation of different PCR-based DNA ®ngerprinting techniques

for assessing the genetic variability of isolates of the fungus

Epicoccum nigrum. Journal of Applied Microbiology 87, 898±906.

Bauda, P., Lalhement, C. and Manem, J. (1995) Plasmid content

evaluation of activated sludge. Water Research 29, 371±374.

Beacham, A.M., Seviour, R.J. and Lindrea, K.C. (1992) Polyphosphate

accumulating abilities of Acinetobacter isolates obtained from a

biological nutrient removal pilot plant. Water Research 26, 121±122.

Beacham, A.M., Seviour, R.J., Lindrea, K.C. and Livingston, I. (1990)

Genospecies diversity of Acinetobacter isolates obtained from a

biological nutrient removal pilot plant of a modi®ed UCT

con®guration. Water Research 24, 23±29.

Bergogne-Berezin, E. and Towner, K.J. (1996) Acinetobacter spp. as

nosocomial pathogens: microbiological, clinical and epidemiological

features. Clinical Microbiology Reviews 9, 148±165.

Bernards, A.T., Dijkshoorn, L., Van Der Toorn, J., Bochner, B.R. and

Boven, C.P.A. (1995) Phenotypic characterization of Acinetobacter

strains of 13 DNA-DNA hybridization groups by means of the

Biolog system. Journal of Medical Microbiology 42, 113±119.

Bond, P., Hugenholtz, P., Keller, J. and Blackall, L. (1995) Bacterial

community structures of phosphate-removing and non-phosphate-

removing activated sludges from sequencing batch reactors. Applied

and Environmental Microbiology 61, 1910±1916.

Bond, P.L., Erhart, R., Wagner, M., Keller, J. and Blackall, L.L. (1999)

Identi®cation of some of the major groups of bacteria in ef®cient and

nonef®cient biological phosphorus removal activated sludge systems.

Applied and Environmental Microbiology 65, 4077±4084.

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 johnsonnii sp.

nov. & Acinetobacter junii sp. nov. & emended description of

Acinetobacter calcoaceticus and Acinetobacter lwof®i. International

Journal of Systematic Bacteriology 36, 228±240.

Bouvet, P.J.M. and Jeanjean, S. (1989) Delineation of a new proteolytic

species in the genus Acinetobacter. Research Microbiology 140,

291±299.

Christensson, M., Blackall, L.L. and Welander, T. (1998) Metabolic

transformations and characterization of the sludge community in an

enhanced biological phosphorus removal system. Applied Microbiol-

ogy and Biotechnology 49, 222±234.

Crocetti, G.R., Hugenholtz, P., Bond, P.L. et al. (2000) Identi®cation

of polyphosphate-accumulating organisms and design of 16S rRNA-

directed probes for their detection and quantitation. Applied and

Environmental Microbiology 66, 1175±1182.

Dijkshoorn, L., Van Harsselaar, B., Tjernberg, I., Bouvet, P.J.M. and

Vaneechoutte, M. (1998) Evaluation of ampli®ed ribosomal DNA

restriction analysis for identi®cation of Acinetobacter genomic

species. Systematic and Applied Microbiology 21, 33±39.

Frolund, B., Griebe, T. and Nielsen, P.H. (1995) Enzymatic activity in

the activated-sludge ¯oc matrix. Applied Microbiology and Biotech-

nology. 43, 755±761.

Gerner-Smidt, P. (1992) Ribotyping of the Acinetobacter calcoaceticus±

Acinetobacter baumannii complex. Journal of Clinical Microbiology 30,

2680±2685.

Gerner-Smidt, P. and Frederiksen, W. (1993) Acinetobacter in Den-

mark: I. Taxonomy, antiobiotic susceptibility and pathogenicity of

112 clinical strains. APMIS 101, 815±825.

Gerner-Smidt, P. and Tjernberg, I. (1993) Acinetobacter in Denmark:

II. Molecular studies of the Acinetobacter calcoaceticus±Acinetobacter

baumannii complex. APMIS 101, 826±832.

Guardabassi, L., Dalsgaard, A. and Olsen, J.E. (1999) Phenotypic

characterization and antibiotic resistance of Acinetobacter spp.

isolated from aquatic sources. Journal of Applied Microbiology 87,

659±667.

Hesselmann, R.P.X., Werlen, C., Hahn, D., Van Der Meer, J.R. and

Zehnder, A.J.B. (1999) Enrichment, phylogenetic analysis and

detection of a bacterium that performs enhanced biological phos-

phate removal in activated sludge. Systematic and Applied Micro-

biology 22, 454±465.

Ibrahim, A., Gerner-Smidt, P. and Liesack, W. (1997) Phylogenetic

relationships of genus Acinetobacter. FEMS Microbiology Letters 124,

349±354.

Janssen, P. and Dijkshoorn, L. (1996) High resolution DNA ®nger-

printing of Acinetobacter outbreak strains. FEMS Microbiology

Letters 142, 194±194.

Jawad, A., Snelling, A.M., Heritage, J. and Hawkey, P.M. (1998)

Comparison of ARDRA and recA-RFLP analysis for genomic

species identi®cation of Acinetobacter spp. FEMS Microbiology

Letters 165, 357±362.

Jenkins, D. and Tandoi, V. (1991) The applied microbiology of

enhanced biological phosphate removalÐaccomplishments and

needs. Water Research 25, 1471±1478.

Juni, E. (1972) Interspecies transformation of Acinetobacter: genetic

evidence for a ubiquitous genus. Journal of Bacteriology 112, 917±931.

Kampfer, P., Bark, K., Busse, H.J., Auling, G. and Dott, W. (1992)

Numerical and chemotaxonomy of polyphosphate accumulating

Acinetobacter strains with high polyphosphate:AMP phosphotrans-

ferase (PPAT) activity. Systematic and Applied Microbiology 15,

409±419.

Kampfer, P., Erhart, R., Beimfohr, J., Wagner, M. and Amann, R.

(1996) Characterization of bacterial communities from activated

sludge: culture-dependant numerical identi®cation versus in situ

identi®cation using group- and genus-speci®c rRNA-targeted

oligonucleotide probes. Microbial Ecology 32, 101±121.

Knight, G.C., Seviour, E.M., Seviour, R.J. et al. (1995) Development

of the microbial community of a full-scale biological nutrient

removal activated sludge plant during start-up. Water Research 29,

2085±2093.

Koeleman, J.G.M., Stoof, J., Biesmans, D.J., Savelkoul, P.H.M. and

Vandenbroucke-Grauls, C.M.J.E. (1998) Comparison of ampli®ed

ribosomal DNA restriction analysis, random ampli®ed polymorphic

DNA analysis and ampli®ed fragment length polymorphism ®nger-

printing for identi®cation of Acinetobacter genomic species and

typing of Acinetobacter baumannii. Journal of Clinical Microbiology 36,

2522±2529.

Lee, N., Nielsen, P.H., Andreasen, K.H. et al. (1999) Combination of

¯uorescent in situ hybridization and microautoradiographyÐa new

tool for structure-function analyses in microbial ecology. Applied and

Environmental Microbiology 65, 1289±1297.

Maszenan, A.M., Seviour, R.J., McDougall, B.M. and Soddell, J.A.

(1997) Diversity of isolates of Acinetobacter from activated sludge

318 E. CARR ET AL .

ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 309±319

systems based on their whole cell protein patterns. Journal of

Industrial Microbiology and Biotechnology 18, 267±271.

Mino, T., Van Loosdrecht, M.C.M. and Heijnen, J.J. (1998)

Microbiology and biochemistry of the enhanced biological phosphate

removal process. Water Research 32, 3193±3207.

Moschetti, G., Blaiotta, G., Aponte, M. et al. (1998) Random ampli®ed

polymorphic DNA and ampli®ed ribosomal DNA spacer poly-

morphism: powerful methods to differentiate Streptococcus thermo-

philus strains. Journal of Applied Microbiology 85, 25±36.

Nielsen, K.M., Smalla, K. and van Elsas, J.D. (2000) Natural transfor-

mation of Acinetobacter sp. strain BD413 with cell lysates of

Acinetobacter sp., Pseudomonas ¯uorescens and Burkholderia cepacia in

soil microcosms. Applied and Environmental Microbiology 66, 206±212.

Nowak, A. and Kur, J. (1995) Genomic species typing of Acineto-

bacters by polymerase chain reaction ampli®cation of the recA gene.

FEMS Microbiology Letters 130, 327±332.

Pacheco, A.B.F., Guth, B.E.C., de Almeida, D.F. and Ferreira, L.C.S.

(1996) Characterization of enterotoxigenic Escherichia coli by random

ampli®cation of polymorphic DNA. Research Microbiology 147,

175±182.

Rainey, F.A., Lang, E. and Stackebrandt, E. (1994) The phylogenetic

structure of the twenty-one DNA groups of the genus Acinetobacter

as revealed by 16S ribosomal DNA sequence analysis. International

Journal of Systematic Bacteriology 47, 837±841.

Ravatn, R., Zehnder, A.J.B. and Van der Meer, J.R. (1998) Low-

frequency horizontal transfer of an element containing the chloro-

catechol degradation genes from Pseudomonas sp. strain B13 to

Pseudomonas putida F1 and to indigenous bacteria in laboratory-scale

activated-sludge microcosms. Applied and Environmental Micro-

biology 64, 2126±2132.

Rudant, E., Bouvet, P., Courvalin, P. and Lambert, T. (1999)

Phylogenetic analysis of proteolytic Acinetobacter strains based on

the sequence of genes encoding aminoglycoside 6¢-N- acetyltrans-

ferases. Systematic and Applied Microbiology 22, 59±67.

Snaidr, J., Amann, R., Huber, I., Ludwig, W. and Schleifer, K. (1997)

Phylogenetic analysis and in situ identi®cation of bacteria in

activated sludge. Applied and Environmental Microbiology 63, 2884±

2896.

Soddell, J.A., Beacham, A.M. and Seviour, R.J. (1993) Phenotypic

identi®cation of non-clinical isolates of Acinetobacter species. Journal

of Applied Bacteriology 74, 210±214.

Soto, S.M., Guerra, B., Gonzalez-Hevia, M.A. and Mendoza, M.C.

(1999) Potential of three-way randomly ampli®ed polymorphic DNA

analysis as a typing method for twelve Salmonella serotypes. Applied

and Environmental Microbiology 65, 4830±4836.

Tjernberg, I. and Ursing, J. (1989) Clinical strains of Acinetobacter

classi®ed by DNA-DNA hybridisation. APMIS 97, 596±605.

Tynkkynen, S., Satokari, R., Saarela, M., Mattila-Sandholm, T. and

Saxelin, M. (1999) Comparison of ribotyping, randomly ampli®ed

polymorphic DNA analysis and pulsed-®eld gel electrophoresis in

typing of Lactobacillus rhamnosus and L. casei strains. Applied and

Environmental Microbiology 65, 3908±3914.

Vaneechoutte, M., Dijkshoorn, L., Tjernberg, I. et al. (1995) Identi-

®cation of Acinetobacter genomic species by ampli®ed ribosomal

DNA restriction analysis. Journal of Clinical Microbiology 33, 11±15.

Vaneechoutte, M., Tjernberg, I., Baldi, F. et al. (1999) Oil-degrading

Acinetobacter strain RAG-1 and strains described as Acinetobacter

venetianus sp. nov. belong to the same genomic species. Research

Microbiology 150, 69±73.

Vickery, M., Smith, A., DePaola, A., Jones, D., Steffan, R. and Bej, A.

(1998) Optimization of the arbitraily-primed polymerase chain

reaction (AP-PCR) from intra-species differentiation of Vibrio

vulni®cus. Journal of Microbiological Methods 33, 181±189.

Vila, J., Marcos, M.A. and Jimenez De Anta, M.T. (1996) A

comparative study of different PCR-based DNA ®ngerprinting

techniques for typing of the Acinetobacter calcoaceticus±A. baumannii

complex. Journal of Medical Microbiology 44, 482±489.

Wagner, M., Erhart, R., Manz, W. et al. (1994) Development of an

rRNA-targeted oligonucleotide probe speci®c for the genus Acine-

tobacter and its application for in situ monitoring in activated sludge.

Applied and Environmental Microbiology 60, 792±800.

Wang, G., Humayun, M.Z. and Taylor, D.E. (1999) Mutation as an

origin of genetic variability in Helicobacter pylori. Trends in

Microbiology 7, 488±493.

Wiedmann-Al-Ahmad, M., Tichy, H.V. and Schon, G. (1994)

Characterization of Acinetobacter type strains and isolates obtained

from wastewater treatment plants by PCR ®ngerprinting. Applied

and Environmental Microbiology 60, 4066±4071.

Yamamoto, S., Bouvet, P.J.M. and Harayama, S. (1999) Phylogenetic

structures of the genus Acinetobacter based on gyrB sequences:

comparison with the grouping with DNA-DNA hybridization.

International Journal of Systematic Bacteriology 49, 87±95.

ACINETOBACTER FROM ACTIVATED SLUDGE 319

ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 309±319