PCR typing of tetracycline resistance determinants (Tet A–E) in Salmonella enterica serotype Hadar and in the microbial community of activated sludges from hospital and urban wastewater treatment facilities in Belgium

Abstract

The distribution of tetracycline resistance determinants Tet A–E was studied by PCR in 40 tetracycline-resistant Salmonella enterica serotype Hadar (S. hadar) isolates collected from human patients in 1996 and 1997, as well as in the microbial community originating from activated sludges of hospital and urban wastewater treatment facilities. A fast DNA extraction and purification method from activated sludges was used to provide amplifiable DNA. The method is based on the direct lysis of bacteria improved by bead-beating followed by DNA purification on polyvinylpolypyrrolidone spin columns to remove PCR inhibitors. The purified DNAs from salmonellae and activated sludges were characterized for the presence of tetracycline determinants with specific primer pairs designed on the basis of published sequences. The Tet A determinant was present in all clinical isolates and DNAs extracted from the bacterial community of the selected activated sludges. The Tet C determinant was identified in only one of the 40 clinical isolates and in six of the seven environmental samples. No signal was detected for Tet B, D and E determinants. This study revealed a high and stable prevalence of the Tet A determinant in both salmonellae clinical isolates and the microbial community of activated sludges from hospital and urban wastewater treatment facilities over a 2-year period.

Tetracycline resistance, PCR, Activated sludge, DNA purification, Salmonella

1 Introduction

Tetracyclines are broad-spectrum antimicrobials that are active against a wide range of Gram-negative and Gram-positive bacteria [1]. They inhibit protein synthesis by preventing the binding of aminoacyl-tRNA molecules to the 30S ribosomal subunit [2]. Their wide use for the past 40 years as therapeutic agents in human and veterinary medicine but also as growth promotor and for prophylaxis in animal husbandry has provoked the selection and emergence of bacterial resistances to these antibiotics, which has severely limited their use in recent years [3,4].

The two most widely distributed mechanisms of tetracycline resistance are active efflux and ribosome protection. These mechanisms are observed in both aerobic and anaerobic Gram-negative or Gram-positive bacteria demonstrating their wide distribution among the bacterial kingdom [5,6]. Tetracycline resistance is generally encoded by plasmids or transposons, several of which are conjugative, but some resistance determinants have also been found on the chromosome.

Resistance to tetracyclines in Gram-negative bacteria is mainly due to an active efflux of the antibiotic which is achieved by an export protein from the major facilitator superfamily [7], although the ribosome protection mechanism is also found [5,6]. Several genes encoding the export protein responsible for tetracycline efflux in Gram-negatives have been identified and sequenced. These closely related genes were classified on the basis of hybridization studies or sequence analysis [8,9]. So far, 12 classes encoding the 12 transmembrane segments export protein have been distinguished: classes A–E, G, H, J, P, Y, Z and 30 [9]. Each of these classes is represented by a determinant called Tet (Tet A–E, Tet G, Tet H, Tet J, Tet P, Tet Y, Tet Z and Tet 30).

While the distribution of efflux-mediated tetracycline determinants in clinical isolates has been extensively studied [10–14], their prevalence and distribution in environmental habitats has received less attention. Environmental studies on the distribution of tetracycline resistance determinants were confined to catfish culture ponds [15,16] and marine sediments [17]. All were conducted with the five most widely distributed determinants among aerobic, enteric Gram-negative bacteria, namely, Tet A–E. These determinants are not uniformly distributed, and they seem to be associated with specific genera and species [5,6]. They were mainly observed in enterics and within the gamma subclass of Proteobacteria.

The objectives of this study were (i) to determine the distribution of tetracycline resistance determinants (Tet A–E) in clinical isolates represented by 40 Tc^RSalmonella hadar clones isolated in Belgium from patient stools in 1996 and 1997 and in man-made ecosystems represented by the microbial community of activated sludges originating from hospital and urban wastewater treatment facilities; (ii) to compare the distribution observed in clinical isolates with that in total community DNA extracted from activated sludges; and (iii) to assess the determinant distribution on a long-term basis (2 years). For a sensitive and rapid typing of these determinants, we designed specific primer pairs for PCR amplification of tetracycline resistance determinants and used a DNA extraction method for the microbial community from activated sludges to detect the determinants of both culturable and non-culturable bacteria.

2 Materials and methods

2.1 Bacterial strains and tetracycline resistance determinants

The salmonellae strains were isolated in 1996 and 1997 by different microbiological diagnostic laboratories in Belgium (Table 1) and serotyped with a slide agglutination kit (Sanofi-Diagnostics Pasteur, Marnes-la-Coquette, France) at the Belgian Reference Laboratory for salmonellae (Institute of Public Health, Louis Pasteur, Brussels, Belgium). Among 26 247 strains of human origin, 1271 (4.8%) were serotyped as Hadar.

1

Source, antibiogram and presence of Tet determinant in 40 Tc^RSalmonella enterica (serotype Hadar) strains isolated from human stools in 1996 and 1997

Isolate	Source	Other resistances	Determinant
96 A1416/2	Vosselaar	SPT-AMC-NAL	Tet A
96 A5411/2	Thuin	SPT-AMC-NAL	Tet A
96 A5444	Boom	SPT-AMC-NAL	Tet A
96 BR489	Brussels	SPT-AMC-NAL	Tet A
96 BR1148	Brussels	SPT-AMC	Tet A
96 BR5414/1	Brussels	SPT-AMC-NAL	Tet A-Tet C
96 HN636	Namur	SPT-AMC-NAL	Tet A
96 HN721/1	Namur	SPT-NAL	Tet A
96 HN903	Tournai	SPT-AMC-NAL	Tet A
96 Lg792	Liège	SPT-NAL	Tet A
96 Lg5433	Chênée	SPT-AMC-NAL	Tet A
96 Li582/2	Hasselt	SPT-AMC-NAL	Tet A
96 Li5719	Genk	SPT-NAL	Tet A
96 Lv643/3	Leuven	SPT-NAL	Tet A
96 Lv526/1	Leuven	SPT	Tet A
96 VL428/2	Dendermonde	SPT-AMC-NAL	Tet A
96 VL749/1	Waregem	SPT-AMC-NAL	Tet A
96 VL963/5	Kortrijk	SPT-AMC-NAL	Tet A
96 VL1029/1	Brugge	SPT-AMC-NAL	Tet A
96 VL1506/3	Gent	SPT-AMC-NAL	Tet A
97 A1221/1	Herentals	SPT-AMC-NAL	Tet A
97 A1225/1	Geel	SPT-AMC-NAL	Tet A
97 A1356/1	Antwerpen	SPT-AMC	Tet A
97 BR1089/8	Brussels	SPT-AMC-NAL	Tet A
97 BR1089/11	Brussels	SPT	Tet A
97 BR1322/2	Brussels	SPT-AMC-NAL	Tet A
97 BR1328	Brussels	SPT-AMC-NAL	Tet A
97 HN1096/2	Auvelais	SPT-AMC-NAL	Tet A
97 HN1320	Belgrade	SPT-AMC-NAL	Tet A
97 HN1370/2	Dinant	SPT-AMC-NAL	Tet A
97 HN1370/3	Dinant	SPT-AMC-NAL	Tet A
97 HN1427/2	Dinant	SPT-AMC-NAL	Tet A
97 Lg1165	Yvoir	SPT	Tet A
97 Lg1282	Alleur	SPT-AMC-NAL	Tet A
97 Lg1376	Alleur	SPT-AMC-NAL	Tet A
97 Lg1388	Verviers	SPT-NAL	Tet A
97 VL1088/3	Kortrijk	SPT-AMC-NAL	Tet A
97 VL1234/1	Oudenaarde	SPT-NAL	Tet A
97 VL1303/1	Zwijnaarde	SPT-AMC-NAL	Tet A
97 VL1365	Dendermonde	SPT-AMC-NAL	Tet A
SPT: resistance to streptomycin; AMC: resistance to ampicillin; NAL: resistance to nalidixic acid. All isolates are sensitive to chloramphenicol, gentamicin and trimethoprim-sulfamethoxazole.

The tetracycline resistance determinants representing classes A–E were extracted from different Escherichia coli strains harboring plasmids carrying the different tetracycline resistance genes (Table 2) (kindly provided by Prof. Marylin C. Roberts, Department of Pathobiology, University of Washington, Seattle, WA, USA; Prof. Martine Thilly-Couturier, Department of Molecular Biology, Free University of Brussels, Brussels, Belgium; Prof. Takashi Aoki, Department of Aquatic Biosciences, Tokyo University of Fisheries, Tokyo, Japan; Prof. Dwight Hirsh, Department of Veterinary Microbiology and Immunology, School of Veterinary Medicine, University of California, Davis, CA, USA). Plasmids were prepared using the Qiagen kit for midi-plasmid preparation (Qiagen).

2

Reference DNA for testing the specificity of the different primer pairs against the tetracycline determinants Tet A–E, Tet G and Tet H and for the construction of Tet probes

Tc resistance determinant	E. coli strain	Plasmid	Reference
Tet A	HB101	RP1	[18]
Tet A	JM83	pSL18	[10]
Tet B	HB101	pRT11 (2.7-kb HpaI fragment from lambda::Tn10 cloned into pVH51)	[19,20]
Tet C	DO-7	pBR322	[21]
Tet D	C600	pSL106 (3.05-kb HindIII–PstI fragment from RA1 cloned into pACYC177)	[10]
Tet D	JM109	4.3-kb HindIII fragment from pSP9350 of Pasteurella piscida in pUC119	[22]
Tet E	HB101	pSL1504 (pSL1456 subcloned into pACYC177)	[23]
Tet G	HB101	9-kb HindIII fragment from pJA8122 of Vibrio anguillarum in pUC119	[24,25]
Tet H	DH5α	pVM111	[26]

2.2 Susceptibility testing

The susceptibility of 273 S. hadar isolates to the following antibiotics was determined by the disc agar diffusion method [27] performed on Mueller–Hinton plates (Sanofi-Diagnostics Pasteur, Marnes-la-Coquette, France). The antibiotics tested were ampicillin (10 μg), chloramphenicol (30 μg), gentamicin (10 μg), nalidixic acid (30 μg), streptomycin (10 μg), tetracycline (30 μg) and trimethoprim-sulfamethoxazole (1.25 μg+23.75 μg).

2.3 Sampling in the wastewater treatment plants

The activated sludge samples originated from three different wastewater treatment facilities. Samples #1, 2 and 5 were taken from an aeration tank treating hospital sewage (Erasmus Hospital, Brussels, Belgium) with a daily load of about 600 m³. The other samples originated from aeration tanks treating urban sewages. Samples #3 and 7 come from the treatment plant of the Dyle valley in Wavre (165 000 population equivalents (PE)). Samples #4 and 6 from the treatment plant of Lasne in Rosière (125 000 PE).

Samples (150 ml) of the mixed liquor were taken during different seasons at three different locations in the aeration tank of the wastewater treatment plant over a 2-year period. They were collected in 500-ml presterilized bottles, stored at 4°C and processed within 3 h. In the laboratory, the samples representing different locations of the aeration basin were combined and 10 aliquots of 0.2 ml withdrawn for the DNA extraction. The viable plate count was determined on Plate Count Agar (PCA) (E. Merck, Darmstadt, Germany) after 2–4 days incubation at 28°C. All platings were performed in triplicate.

2.4 DNA purification from Tc^RS. hadar

As tetracycline resistance genes can be located on both plasmids and the chromosome, DNA was extracted from Salmonella isolates according to the method described by Chen and Kuo [28]. This method allows the simultaneous extraction of both plasmid and chromosomal DNA.

2.5 Community DNA extraction from activated sludges

The DNA extraction protocol used in the present study was derived from the methods described by Van Vaerenbergh et al. [29] and Haynes et al. [30] with slight modifications. Aliquots (0.2 ml) of the liquor of the activated sludge were mixed with 0.6 ml of lysis buffer (Tris–HCl 50 mM, pH 8.0, EDTA 50 mM, sodium dodecyl sulfate 3%) in a 2-ml screw-capped polypropylene microtube (BioSpec) containing conditioned glass beads (half a tube; 0.1 mm diameter), and 0.6 ml of Tris–HCl-saturated phenol–chloroform–isoamyl alcohol (50:49:1) was added. The tubes were shaken for 2 min at high speed on a Mini-Beadbeater (BioSpec Products, Stratech Scientific, Luton, UK) to lyse the more resistant cell structures and incubated for 30 min at 65°C. After centrifugation for 15 min at 14 000 rpm in a microcentrifuge, the supernatant was transferred to a sterile microtube. The nucleic acids were precipitated in ethanol in the presence of potassium acetate 0.3 M for 1 h at −80°C and pelleted at 14 000 rpm for 15 min at 4°C. The pellet was washed with cold 70% ethanol, dried under vacuum and dissolved in 50–100 μl autoclaved MilliQ water.

Glass beads (BioSpec) of 0.1 mm diameter were incubated overnight in concentrated H₂SO₄, washed with MilliQ water, boiled in KHCO₃ 5%, washed with autoclaved MilliQ water and desiccated at 80°C under vacuum.

The DNA solutions were purified by centrifugation (5000 rpm, 5 min) through a spin column filled with conditioned polyvinylpolypyrrolidone (PVPP). PVPP was dispensed in a Mobicol column (MoBiTec, Göttingen, Germany) previously washed with 0.1 M HCl, rinsed with distilled water, autoclaved and enclosed by two filters (pore size: 35 μm). The purified extracts (clear and colorless) were ethanol-precipitated in the presence of 0.3 M potassium acetate for 1 h at −80°C and pelleted at 14 000 rpm for 15 min at 4°C. The pellet was washed with cold 70% ethanol, dried under vacuum and dissolved in 50 μl TE buffer (pH 8.0). The DNA was then stored at −20°C until further use.

PVPP was conditioned by repeated washes of 100 g in 1 l of distilled water until the upper phase was clear. The PVPP was then washed with 0.1 M NaOH until the effluent had a pH of 4–6 and neutrality restored by washing with 20 mM phosphate buffer (pH 7.5). The conditioned PVPP was air-dried and stored at room temperature.

2.6 PCR conditions

DNA amplification was performed in a Perkin-Elmer Cetus Thermal Cycler 9600. The amplification reaction mixture contained the amplification buffer (Tris–HCl pH 8.3, 50 mM; KCl 50 mM; 0.1 mg ml⁻¹ bovine serum albumin (Sigma Cohn fraction V)); MgCl₂ 3 mM; 200 μM of each dNTP (Pharmacia Biotech); 20 pmol of each amplimer, for the DNA extracted from Salmonella 1 U of Taq polymerase (Perkin-Elmer) and 20 ng of the template DNA, and for the DNA extracted from environmental samples 2 U of Taq polymerase and 100 ng of the template DNA, and autoclaved MilliQ water up to 50 μl final volume. After denaturation at 94°C for 1 min, the samples were submitted to 30 cycles (35 cycles for the environmental samples) of denaturation at 94°C for 1 min, annealing at 55°C for Tet A, B, D and E or 58°C for Tet C for 1 min (Table 3), extension at 72°C for 2 min and by an additional extension step of 10 min at 72°C.

3

Primer systems designed by the programs GeneWorks^® (IntelliGenetics, Campbell, USA) and Oligo5 (National Biosciences, Plymouth, UK) on the basis of published sequence data and sequence alignments and used for the typing of the different tetracycline resistance determinants Tet A–E

Determinant	Oligo name	Primer sequence	Product size (bp)	Primer position	Tm (°C)	Accession no.
Tet A	TETAFW4	GGCCTCAATTTCCTGACG	372	511	59	X0006
	TETARV1	AAGCAGGATGTAGCCTGTGC		882	60
Tet B	TETBFW3	GAGACGCAATCGAATTCGG	228	586	62	J01830
	TETBRV1	TTTAGTGGCTATTCTTCCTGCC		813	60
Tet C	TETCFW3	TGCTCAACGGCCTCAACC	379	503	63	J01749
	TETCRV1	AGCAAGACGTAGCCCAGCG		881	64
Tet C	TETCFW2	TCCTTGCATGCACCATTCC	635	472	63	J01749
	TETCRV2	AACCCGTTCCATGTGCTCG		1106	64
Tet D	TETDFW2	GGATATCTCACCGCATCTGC	436	459	61	X65876
	TETDRV1	CATCCATCCGGAAGTGATAGC		894	62
Tet E	TETEFW3	TCCATACGCGAGATGATCTCC	442	607	62	L06940
	TETERV2	CGATTACAGCTGTCAGGTGGG		1048	63

PCR products were detected by electrophoresis on 1.5% agarose gels using as size marker a 123-bp ladder (Life Technologies, Merelbeke, Belgium).

2.7 Hybridization conditions

DNA–DNA hybridizations were performed with the Alkphos kit (Pharmacia, Uppsala, Sweden). The 771-bp EcoRI–BamHI fragment of pSL18 [10], the 1275-bp HincII fragment of pRT11 [20], the 600-bp BamHI–NruI fragment of pBR322 [21], the 951-bp AgeI fragment of pUC119::tetA(D) tetR(D) [22] and the 1050-bp ClaI–PvuI fragment of pSL1504 [23] were used as probes for the Tet A, B, C, D and E determinants, respectively. PCR-amplified fragments were loaded on the gel and transferred to a nylon membrane (Hybond N+, Amersham, Les Ullis, France) by vacuum blotting. The membrane was subsequently air-dried and the DNA UV cross-linked for 5 min. Hybridizations and filter incubations were performed as described by the supplier. Autoradiography was done by exposing the filters to hyperfilms ECL (Amersham, Les Ullis, France) during 1 h.

3 Results

3.1 Design and evaluation of the specificity of oligonucleotide primer pairs

The strategy followed to design specific primers for each determinant of classes A–E was the following. Multiple alignments of the published DNA sequences of tetA genes representing each class (A: RP1, B: Tn10, C: pBR322, D: pIP173, E: pSL1456, G: pJA1822, H: pVM111) were carried out using the pileup program supplied with the GCG software package. Previous studies have shown that the classes A, C and G are more closely related to each other than to classes B, D, E and H, which form another branch in the phylogeny of efflux protein [26,31]. Therefore, the tetA genes of RP1, pBR322 and pJA1822 were aligned (52% DNA sequence identity) as were the tetA genes of Tn10, pIP173, pSL1456 and pVM111 (34% DNA sequence identity). The specific primer pairs were designed using the programs GeneWorks^® (IntelliGenetics, Campbell, USA) and Oligo5 (National Biosciences, Plymouth, UK) in regions displaying a low sequence conservation (Table 3). The specificity of primer pairs was tested for each class using a number of purified plasmids harboring the different determinants Tet A–E, Tet G and Tet H as reference (Table 2). All primer pairs tested resulted in PCR products of the predicted size only, demonstrating their high specificity (data not shown).

3.2 Epidemiological data and distribution of the tetracycline resistance determinants within the S. hadar isolates

Of the 1271 S. hadar isolates (years 1996 and 1997), 273 were screened for their resistance to tetracycline, ampicillin, chloramphenicol, gentamicin, nalidixic acid, streptomycin and trimethoprim-sulfamethoxazole. Except for four isolates, they all exhibited resistance to at least one antibiotic (data not shown). Resistance to tetracycline was detected in 251 (92%) of the 273 isolates. In order to determine the bacterial distribution of the tetracycline resistance determinants, 20 Tc^R isolates from 1996 and 20 isolates from 1997 were selected at random (Table 1). Their total DNA was extracted and analyzed by gel electrophoresis for the presence of both chromosomal and plasmid DNA. Plasmid profiles revealed similarities to some extent for both small and large plasmids as shown in Fig. 1 for 13 isolates. The total DNA of each isolate was submitted to PCR amplification with the primer pairs designed for each class. All 40 isolates gave a positive signal with the expected size for class A. In contrast, no positive signal was obtained for classes B, C, D or E, except for the isolate 96 BR5414/1 which was positive for class C. Southern blotting of the PCR product amplified with the specific primer set for Tet C hybridized with the internal 600-bp BamHI–NruI fragment of tetA(C) of pBR322 (Fig. 2).

1

Gel electrophoresis of non-digested total DNA extracted from different tetracycline-resistant S. enterica (serotype Hadar) isolated from human stools. Lanes: M, λHindIII; 1, isolate 97 A1225/1; 2, isolate 96 BR1148; 3, isolate 97 HN1370/2; 4, isolate 96 VL963/5; 5, isolate 97 VL1088/3; 6, isolate 97 Lg1388; 7, isolate 96 LV526/1; 8, isolate 96 HN636; 9, isolate 97 VL1234/1; 10, isolate 96 Lg792; 11, isolate 97 A1356/1; 12, isolate 96 A5444; 13, isolate 97 BR1089/11.

2

Southern blot hybridization of PCR products resulting from amplification of community DNA extracts from the seven activated sludge samplings and S. hadar isolate 96 BR5414/1 with specific primer pairs for Tet C determinant (FW2–RV2). The probe is the 600-bp BamHI–NruI fragment of pBR322, internal to the tetA(C) gene. Lane 1: positive control; lane 2: negative control; lane 3: sample #1 (Erasmus Hospital, 1996); lane 4: sample #2 (Erasmus Hospital, 1997); lane 5: sample #4 (Rosière, 1997); lane 6: sample #3 (Wavre, 1997); lane 7: sample #5 (Erasmus Hospital, 1999); lane 8: sample #6 (Rosière, 1999); lane 9: sample #7 (Wavre, 1999); lane 10: isolate 96 BR5414/1.

A S. hadar strain phenotypically sensitive to tetracycline (96 VL1039/2: AMC^R, NAL^R) was also analyzed by PCR amplification for the presence of tet genes of classes A–E. All PCR reactions were negative for the expected band demonstrating the absence of a non-functional or unexpressed tetracycline resistance gene in the bacterial genome.

3.4 Distribution of tetracycline resistance determinants in the microbial community of activated sludges

3.4.1 Percentage of tetracycline-resistant isolates in the total culturable population of activated sludges

The total bacterial count on PCA and the resistant quotient (rq) values of tetracycline-resistant bacterial colonies are shown in Table 4. The rq values on PCA agar supplemented with 10 mg l⁻¹ tetracycline hydrochloride varied from 6 to less than 0.01. The highest percentage of tetracycline-resistant isolates was found in the activated sludges from the hospital wastewater treatment facility. Increasing the tetracycline concentration on PCA agar plates from 5 to 20 mg l⁻¹ slightly affected the number of colony forming units (CFU), although no tetracycline was added before plating to induce the resistance mechanism.

4

Tetracycline rq of culturable bacteria from activated sludges and distribution of Tet determinants (Tet A–E) after screening of DNA extracts with the different primer pairs

Period	Sample no.; sludge source	No. of culturable bacteria (CFU on PCA plates ml−1 sludge)	Percentage of resistant bacteria to Tc 5 mg l−1	Percentage of resistant bacteria to Tc 10 mg l−1	Percentage of resistant bacteria to Tc 20 mg l−1	PCR results Tet determinants
December 96	#1; Erasmus Hospital	1.5×107	12	6	4	A and C
July 97	#2; Erasmus Hospital	3×107	0.6	2.1	0.22	A and C
July 97	#3; Wavre	6.5×108	0.018	0.0066	0.0057	A
August 97	#4; Rosière	4.5×108	0.027	0.027	0.027	A and C
March 99	#5; Erasmus Hospital	1.1×108	0.73	0.83	0.5	A and C
April 99	#6; Rosière	2.5×108	ND	1.96	ND	A and C
May 99	#7; Wavre	2×108	ND	0.9	ND	A and C
ND: not determined.

3.4.2 Analysis of directly extracted community DNA

DNA was extracted and purified from samples collected from three wastewater treatment facilities at different times during a 2-year period (Table 4). DNA extracts were either pale yellow or brownish, indicating the presence of impurities, likely humic acid-like substances. These were removed after the PVPP purification. Purity and concentration of the purified extracts, representing the total microbial DNA, were assessed using spectrophotometry and agarose gel electrophoresis. The DNA yield obtained ranged from 20 to 40 μg ml⁻¹ sludge, which is in accordance with methods previously reported for soil DNA extraction [32–34]. DNA purity of each sample was also tested by PCR amplification on different DNA dilutions with universal primers for 16S rRNA genes of eubacteria. All samples gave positive results with non-diluted DNA (data not shown), demonstrating that the purification step results in a DNA extract pure enough to allow PCR amplification.

Since all clinical isolates revealed the presence of the Tet A determinant and one the presence of the Tet C determinant, we amplified sample #1 with the primer pairs designed for these two determinants at different DNA concentrations (Fig. 3). A 35-cycle PCR amplification resulted in detectable products with the predicted size for Tet A and Tet C at DNA concentrations ranging from 1.6 ng to 160 ng of total environmental DNA. For further PCR amplifications with environmental purified DNA, we used 100 ng of DNA as template. All purified DNA samples originating from the different activated sludges were subsequently screened by PCR amplification with primer pairs for each class of tetracycline determinants. They all produced a band with the expected size for classes A and C, except for the sample #4 (Rosière, 1997) which was only positive for class A. In contrast, we were unable to detect a positive signal on gels when we amplified with primer pairs for classes B, D and E.

3

Gel electrophoresis of PCR products resulting from an amplification of community DNA extract. DNA extraction was performed from the activated sludge of the hospital treatment plant collected in December 1996. Lanes 2–6 correspond to the amplification of the community DNA with specific primer pairs for Tet A (FW4–RV1; product size: 372 bp) and lanes 8–12 to the amplification with specific primer pairs for Tet C (FW3–RV1; product size: 379 bp). Lanes: M, 123-bp ladder; 1, negative control (deionized water); 2, positive control (RP1: 0.06 ng); 3, extracted DNA (160 ng); 4, extracted DNA (16 ng); 5, extracted DNA (1.6 ng); 6, extracted DNA (0.16 ng); 7, negative control (deionized water); 8, positive control (pBR322: 0.04 ng); 9, extracted DNA (160 ng); 10, extracted DNA (16 ng); 11, extracted DNA (1.6 ng); 12, extracted DNA (0.16 ng).

All amplification reactions were analyzed by DNA–DNA hybridization after Southern blotting, to confirm specificity and to improve sensitivity. Amplification products obtained for Tet A and Tet C (Fig. 2) hybridized with the tetA(A) and tetA(C) probes, respectively, confirming a high DNA sequence homology. No amplification products of expected size became visible when hybridized with the respective probe for Tet B, D or E (data not shown).

4 Discussion

To allow a sensitive and rapid detection of tetracycline resistance determinants in both clinical isolates and total community DNA of environmental samples, the PCR technique was chosen. In contrast to DNA hybridization, PCR amplification enables sensitive and rapid detection of any class of antibiotic resistance determinant for which a DNA sequence is available, including those present in environmental samples in low copy number or in non-culturable bacteria. In the study of Andersen and Sandaa [17], hybridization of total DNA extracted from marine sediments with probes for the determinant classes A–E resulted in no signal, although the presence of tetracycline-resistant bacteria was confirmed by cultivation. Tc^R bacteria were likely present under the detection limit for DNA probes.

Several methods of DNA extraction from activated sludges have recently been described [35–37]. Since a protocol based on a cell lysis improved by bead-beating followed by a purification trough PVPP columns to remove Taq polymerase inhibitors was set up in our laboratory to type fungi in clinical and environmental samples [29,30], the same method was used with slight modifications for DNA purification from activated sludges. Removal of humic acid contaminants from soil extracts in the presence of PVPP was also reported by Holben et al. [38], Bertelet et al. [32] and Cullen et al. [33]. In our study, the DNA extracted from sludges was pure enough to allow amplification with universal 16S rDNA primers for eubacteria and the DNA yield ranged from 20 to 40 μg ml⁻¹ sludge, which is in accordance with methods previously reported for soil DNA extraction [32–34].

To analyze the distribution of tetracycline resistance determinants Tet A–E in clinical isolates and in man-made ecosystems, we have chosen to examine 40 Tc^RS. hadar strains isolated from human stools in 1996 and 1997, and seven samples originating from one hospital and two urban wastewater treatment plants. Salmonellae have been shown to contain tetracycline resistance of classes A–D [5,6], and a recent publication reports a chromosomally integrated Tet G determinant in Salmonella typhimurium DT104 [39]. The incidence of tetracycline resistance in S. hadar, which is the third predominant serotype in humans in Belgium, has doubled from 1991 (48%) to 1994 (96%) [40]. This frequency remained stable at least until 1996 (92%). Of particular interest has been a similar dramatic increase of ampicillin (16 to 79% from 1991 to 1994) and nalidixic acid resistance (4 to 79%) in avian S. hadar strains [40]. Our results indicate that the Tet A determinant was present in all of the 40 Tc^RS. hadar isolates. Tet C was also detected in one of the 40 Tc^RS. hadar isolates, already positive for Tet A. Tet B, D and E were not detected. Previous studies, however, identified Tet B and C most frequently in bacteria of the genus Salmonella[41,42]. More recently, four tetracycline-resistant Salmonella dublin strains isolated from cattle in Germany were shown to contain an identical 47-kb plasmid called pGFT1 which contained a tetA(A) gene located on an integrated copy of Tn1721-analogous transposon [14].

To study the tetracycline resistance determinant distribution in the microbial community of activated sludges without relying on culturing isolates, a molecular approach based on direct DNA extraction from sludge was adopted. It permits the detection of target DNA sequences in non-culturable bacteria and avoids the pitfalls of preferential selection of certain genera or species by plate cultivation. Indeed, it was shown that plating of activated sludges on rich medium favors the growth of members of the gamma subclass of Proteobacteria and selects against the growth of the beta subclass and Gram-positive bacteria with a high G+C content [43]. Selection of Tc^R isolates from activated sludges by plating on rich media could therefore influence the results about Tet distribution since several cultivation-independent studies indicated dominance of the beta subclass of the Proteobacteria in sludges, which could therefore be underestimated [44,45].

All purified DNA samples originating from the different activated sludges collected over a 2-year period were screened by PCR amplification, and PCR products were confirmed by hybridization after Southern blotting. Our results demonstrated that Tet A was the most dominant determinant in the microbial community of activated sludges from aerobic wastewater treatment plants, independent of their origin (hospital or urban). Tet C was also detected in six from seven samples (not in sample #3, Wavre, 1997). However, it should be noted that sample #3 gave the smallest frequency of Tc^R isolates and that the experimental conditions (35 cycles PCR amplification) may not have been sensitive enough to allow detection. As with salmonellae, Tet B, D and E determinants were not found. This distribution was observed on a 2-year survey indicating a stable prevalence of Tet A. This proves a positive correlation between clinical isolates and the microflora of activated sludges for tetracycline resistance determinants.

Previous environmental studies have documented various distributions among tetracycline determinants. Hybridizations performed with DNA probes for Tet A and Tet E on environmental tetracycline-resistant isolates were first conducted by DePaola et al. [15]. It was shown that over 90% of Tc^RAeromonas strains isolated from catfish intestinal contents and the water and sediment of catfish culture ponds in the southeastern USA were positive for both determinants; Tet E was twice as prevalent as Tet A. Another study on Gram-negative bacteria isolated from Norwegian and Danish marine sediments showed that Tet E was the predominant determinant among the five classes A–E [17]. Nineteen transferable plasmids encoding oxytetracycline resistance isolated from Aeromonas salmonicida strains recovered from Atlantic salmons affected by furuncolosis in Scotland were shown to hybridize only with the Tet A determinant [16]. In contrast, all clinical studies on Gram-negative isolates collected in the USA revealed Tet B as the most prevalent determinant, although Tet A and C were also found at lesser frequencies [10–12,41,42]. These results demonstrate that distribution patterns of Tet determinants vary with the species studied or the sampling origin, suggesting ecosystem-specific reservoirs for tetracycline resistance genes.

The present approach could be of general use to assess the prevalence of any antibiotic resistance gene in nature and the impact of human and agricultural activities on their selection and transfer. For instance, we are currently assessing whether the use of activated sludges as soil amendment in agriculture could result in the increase of resistance determinants in nature by genetic transfer from antibiotic-resistant bacteria of sludge to indigenous populations in terrestrial habitats.

Acknowledgements

This work was supported by Grant NO/50/012 of the Belgian Federal Office for Scientific, Technical, and Cultural Affairs, and by Grant BIO4 CT98 0053 from the EU-Biotechnology Programme (1994–1998). We are grateful to Elizabeth M.H. Wellington and Stefan Wuertz for critically reading and discussing the manuscript. We also thank Ms. Braun, Mr. Lambert and Flahaut, and staff of the different wastewater treatment plants.