Multidrug-Resistant Pseudomonas aeruginosa Strain That Caused an Outbreak in a Neurosurgery Ward and Its aac(6′)-Iae Gene Cassette Encoding a Novel Aminoglycoside Acetyltransferase

Abstract

We characterized multidrug-resistant Pseudomonas aeruginosa strains isolated from patients involved in an outbreak of catheter-associated urinary tract infections that occurred in a neurosurgery ward of a hospital in Sendai, Japan. Pulsed-field gel electrophoresis of SpeI-, XbaI-, or HpaI-digested genomic DNAs from the isolates revealed that clonal expansion of a P. aeruginosa strain designated IMCJ2.S1 had occurred in the ward. This strain possessed broad-spectrum resistance to aminoglycosides, β-lactams, fluoroquinolones, tetracyclines, sulfonamides, and chlorhexidine. Strain IMCJ2.S1 showed a level of resistance to some kinds of disinfectants similar to that of a control strain of P. aeruginosa, ATCC 27853. IMCJ2.S1 contained a novel class 1 integron, In113, in the chromosome but not on a plasmid. In113 contains an array of three gene cassettes of bla_IMP-1, a novel aminoglycoside resistance gene, and the aadA1 gene. The aminoglycoside resistance gene, designated aac(6′)-Iae, encoded a 183-amino-acid protein that shared 57.1% identity with AAC(6′)-Iq. Recombinant AAC(6′)-Iae protein showed aminoglycoside 6′-N-acetyltransferase activity by thin-layer chromatography. Escherichia coli expressing exogenous aac(6′)-Iae showed resistance to amikacin, dibekacin, isepamicin, kanamycin, netilmicin, sisomicin, and tobramycin but not to arbekacin, gentamicins, or streptomycin. Alterations of gyrA and parC at the amino acid sequence level were detected in IMCJ2.S1, suggesting that such mutations confer the resistance to fluoroquinolones observed for this strain. These results indicate that P. aeruginosa IMCJ2.S1 has developed multidrug resistance by acquiring resistance determinants, including a novel member of the aac(6′)-I family and mutations in drug resistance genes.

Pseudomonas aeruginosa is intrinsically resistant to many antibiotics; however, it is sensitive to a limited number of drugs, including some β-lactams, such as ceftazidime and imipenem, and aminoglycosides, such as amikacin and tobramycin. However, recent studies have shown that several strains of P. aeruginosa that are resistant to these antibiotics have emerged and are becoming widespread (21, 28).

In Japan, the major mechanism of resistance to aminoglycosides is production of aminoglycoside-modifying enzymes (43). The aminoglycoside 6′-N-acetyltransferases [AAC(6′)s] are of particular interest because they can modify a number of clinically important aminoglycosides including amikacin, gentamicin, netilmicin, and tobramycin. The AAC(6′)-I type confers resistance to amikacin through acetylation of the drug, whereas the AAC(6′)-II type acetylates gentamicin. To date, several different genes, designated aac(6′)-Ia to aac(6′)-Iad, that encode the AAC(6′)-I enzymes have been cloned and characterized (42, 50). Genes encoding aminoglycoside-modifying enzymes are often located on integrons (15), sequences that can integrate gene cassettes through site-specific recombination (17), in both plasmid and genomic DNA (15). Class 1 integrons participate in multidrug resistance in P. aeruginosa (27, 28, 37). Class 1 integrons contain two conserved segments (CS) that flank the antibiotic resistance gene cassettes. The 5′-CS contains the intI1 gene, which encodes integrase, the enzyme responsible for catalysis of site-specific recombination (8). The 3′-CS contains the qacEΔ1 and sul1 genes and an open reading frame (ORF), orf5 (13, 16).

We describe here the genotypic and phenotypic properties of a new multidrug-resistant P. aeruginosa strain that caused a nosocomial outbreak of infection at a hospital in Japan. The isolate carries a class 1 integron that contains an array of three gene cassettes, including one encoding a novel aminoglycoside acetyltransferase.

MATERIALS AND METHODS

Bacterial strains.

Seven clinical isolates of P. aeruginosa, including P. aeruginosa IMCJ2.S1, were obtained from seven patients with urinary tract infections in a neurosurgery ward of a hospital in Japan. P. aeruginosa ATCC 27853 was obtained from the American Type Culture Collection (Manassas, Va). Escherichia coli strains DH5α (Takara Bio, Shiga, Japan) and BL21-AI (Invitrogen, Carlsbad, Calif.) were used as hosts for recombinant plasmids and for expression of aac(6′)-Iae, respectively. The rifampin-resistant P. aeruginosa mutant ATCC 27853 RFP^r was used. P. aeruginosa GN17203 (51) was provided by S. Iyobe (Kitasato University, Sagamihara, Japan).

Antibiotics and disinfectants.

The antibiotics amikacin, cefoxitin, and imipenem were from Banyu Pharmaceutical Co. (Tokyo, Japan). Arbekacin and dibekacin were from Meiji Seika Kaisha (Tokyo, Japan), aztreonam was from Eizai (Tokyo, Japan), cefotaxime was from Aventis Pharma (Tokyo, Japan), and cefpodoxime and ceftazidime were from Glaxo Smith Kline (Tokyo, Japan). Cefepime was from Bristol Pharmaceuticals (Tokyo, Japan); ciprofloxacin and levofloxacin were from Daiichi Pharmaceutical (Tokyo, Japan); gentamicin, isepamicin, netilmicin, and sisomicin were from Schering-Plough (Osaka, Japan); kanamycin A and B mixture, neomycin B and C mixture, and streptomycin were from Nacalai Tesque (Kyoto, Japan); and meropenem was from Sumitomo Pharmaceutical (Osaka, Japan). Tetracycline was from Lederle Japan Co. (Tokyo, Japan); piperacillin and piperacillin-tazobactam were from Tomiyama Pure Chemical Industries (Tokyo, Japan); moxalactam, tobramycin, and sulfamethoxazole-trimethoprim were from Shionogi and Co. (Osaka, Japan); and kanamycin A, polymyxin B, and silver sulfadiazine were from Sigma Chemical (St. Louis, Mo.). The disinfectants alkyldiaminoethylglycine hydrochloride and povidone iodine were from Yoshida Pharmaceutical Co. (Tokyo, Japan); benzalkonium chloride was from Wako Pure Chemical Industries (Osaka, Japan); and chlorhexidine gluconate was from Ishimaru Pharmaceutical (Osaka, Japan).

In vitro susceptibility to antibiotics and disinfectants.

MICs of antibiotics, except polymyxin B and silver sulfadiazine, were determined by the microdilution method. The MICs of polymyxin B and silver sulfadiazine were determined by the agar dilution method according to the protocols recommended by the CLSI (formerly NCCLS), standard M7-A6 (33).

Bactericidal activities of disinfectants were evaluated by time- and dose-dependent killing studies in 96-well microplates. Briefly, 10⁵ microorganisms were incubated at 35°C for 0.5 min to 60 min in 160 μl disinfectants diluted serially twofold. To neutralize the bactericidal activities of the disinfectants, a 10-μl aliquot of each suspension was transferred to 200 μl Trypticase soy broth (Becton Dickinson, Franklin Lakes, NJ) containing 15% Tween 80 (Sigma), 1% soybean lecithin (Nacalai Tesque), and 0.5% sodium thiosulfate (Nacalai Tesque) and then cultured for 24 h. The minimum bactericidal concentrations (MBCs) of disinfectants were recorded relative to the duration of incubation with bacteria.

Transfer of drug resistance among bacteria.

Transfer of the drug resistance from P. aeruginosa clinical isolates to a rifampin-resistant mutant of P. aeruginosa, ATCC 27853 RFP^r, was examined with the broth mating method (25). After mating, transconjugants were selected on Mueller-Hinton agar plates containing rifampin (200 μg/ml) and imipenem (16 μg/ml) or amikacin (20 μg/ml). Plasmid DNAs from the clinical isolates were purified either with a QIAprep kit (QIAGEN, Tokyo, Japan), by Kado and Liu’s (24), or method by the method of Domenico et al. (11). With the QIAprep kit or Kado and Liu's method, the bacteria were lysed at different temperatures, 22°C for 5 min or 60°C for 70 min for each method.

PCR of class 1 integrons.

To identify the presence of a class 1 integron and to determine the size of any inserted gene cassettes, PCR amplification was performed as described previously (29) with primers 5′-cs and 3′-cs, which are specific for 5′-CS and the 3′-CS of class 1 integrons, respectively, and an Expand High Fidelity PCR system (Roche Diagnostics GmbH, Penzberg, Germany). To determine the content and order of genes in the integron, PCR amplification of the variable region of class 1 integrons was carried out with the primers listed in Table 1. All PCRs were performed with a GeneAmp PCR system 9700 thermal cycler (Applied Biosystems, Foster City, Calif.). Genomic DNAs extracted as described by Sambrook and Russell (41) were used as templates. Amplification conditions were 30 cycles of 94°C for 1 min, 50°C for 1 min, and 72°C for 3 min or 5 min. PCR for amplicons longer than 1 kb was performed with 1.25 U of Z-Taq polymerase (Takara Bio) and 30 cycles of 95°C for 1 s and 68°C for 120 s according to the manufacturer's instructions.

TABLE 1.

PCR primers

Primer	Sequencea (5′→3′)	Expected size of amplicon (bp)	Position (nt)b	Reference or GenBank accession no.
5′-cs	GGCATCCAAGCAGCAAG			29
3′-cs	AAGCAGACTTGACCTGA			29
int1-F	TGCGTGTAAATCATCGTCGT	838	Downstream of intI1	AF071413
int1-R	CGAAGTCGAGGCATTTCTGT		177-196 in intI1	AF071413
IMP-Fc	DTTYCTAAACAYGGYTTGGT	362	145-164 in blaIMP-1	AB070224
IMP-Rc	YTTTYAGGYARCCAAACYACT		486-506 in blaIMP-1	AB070224
aacS1-F	CGCAAGCTGCAGAAATTCTAT	465	47-67 in aac(6′)-Iae	This study
aacS1-R	TCCCATTTGCATTAGGAATCA		491-511 in aac(6′)-Iae	This study
aadA1-F	TGATTTGCTGGTTACGGTGA	451	144-163 in aadA1	AF071413
aadA1-R	TACTGCGCTGTACCAAATGC		575-594 in aadA1	AF071413
qacEdelta-F	TGAAAGGCTGGCTTTTTCTT	286	2-21 in qacEΔ1	AF071413
qacEdelta-R	GCAATTATGAGCCCCATACC		268-287 in qacEΔ1	AF071413
sul-F	TCACCGAGGACTCCTTCTTC	759	29-48 in sul1	AF071413
sul-R	GGGTTTCCGAGAAGGTGATT		768-787 in sul1	AF071413
int1imp1-F	AGCACCTTGCCGTAGAAGAA	695	262-281 in intI1	AJ640197
int1imp1-R	TTTTATAGCCACGCTCCACA		243-262 in blaIMP-1	AJ640197
imp1aacS1-F	AAAGGCAGCATTTCCTCTCA	737	265-284 in blaIMP-1	This study
imp1aacS1-R	GACGGCCAAGAATCGAAAT		89-107 in aac(6′)-Iae	This study
aacS1aadA1-F	ATTGTGTGGTTGGGTTGGAT	691	186-205 in aac(6′)-Iae	This study
aacS1aadA1-R	GGAGAATCTCGCTCTCTCCA		231-259 in aadA1	This study
aadA1qacEd-F	TGATTTGCTGGTTACGGTGA	873	144-163 in aadA1	AF071413
aadA1qacEd-R	ATGCGGATGTTGCGATTACT		42-61 in qacEΔ1	AF071413
qacEdsul-F	TCGGTGTTGCTTATGCAGTC	306	167-186 in qacEΔ1	AF071413
qacEdsul-R	ACATCCACGACGTCTGATCC		112-131 in sul1	AF071413
int-R	TGCGTGTAAATCATCGTCGT	3,172	Downstream of intI1	AF071413
sul-R	GGGTTTCCGAGAAGGTGATT		768-787 in sul1	AF071413
sul-F	TCACCGAGGACTCCTTCTTC	6,474	29-48 in sul1	AF071413
tniB-R	ATCATCGACCTGTCCCACCT		16-35 in tniBΔ1	AF071413
tniB-F	CAGAGCCAGTTGCTCCATTT	1,749	395-414 in tniBΔ1	AF071413
tniA-R	CTTTCACCGCGAAGTCACTC		384-403 in tniA	AF071413
GyrA1	TTATGCCATGAGCGAGCTGGGCAACGACT	366	147-176 in gyrA	26
GyrA2	AACCGTTGACCAGCAGGTTGGGAATCTT		484-512 in gyrA	26
GyrB1	GCGCGTGAGATGACCCGCCGT	390	1162-1182 in gyrB	31
GyrB2	CTGGCGGTAGAAGAAGGTCAT		1531-1551 in gyrB	31
PARC1	ATGAGCGAACTGGGGCTGGAT	210	166-187 in parC	21
PARC2	ATGGCGGCGAAGGACTTGGGA		354-375 in parC	21
ParE1	CGGCGTTCGTCTCGGGCGTGGTGAAGGA	592	1223-1250 in parE	1
ParE2	TCGAGGGCGTAGTAGATGTCCTTGCCGA		1787-1814 in parE	1
aacS1-FC	ATGAAATACAACATTGTTAATATTA	552	1-25 in aac(6′)-Iae	This study
aacS1-RC	TTACATTATATTTTTCCACATTAAT		528-552 in aac(6′)-Iae	This study

^aD stands for adenine, thymine, or guanine; R stands for adenine or guanine; Y stands for cytosine or thymine.

^bNucleotides are numbered according to deposited sequences.

^cPrimer designed to amplify bla_IMP-1 (accession no. AB070224) or homologous genes, including bla_IMP-2 (AJ243491), bla_IMP-3 (AB010417), bla_IMP-4 (AF445082), bla_IMP-5 (AF290912), bla_IMP-6 (AB040994), bla_IMP-7 (AF416736), bla_IMP-8 (AF322577), bla_IMP-9 (AY033653), bla_IMP-10 (AB074434), and bla_IMP-11 (AB074437).

PCR of QRDRs.

The gyrA, gyrB, parC, and parE quinolone resistance-determining regions (QRDRs) of P. aeruginosa were amplified by PCR with the primers listed in Table 1 according to methods described previously (1, 21, 26, 31). PCR products were sequenced with the same primers.

DNA sequencing.

DNA sequences were determined by the dideoxy chain termination method with an ABI PRISM 3100 sequencer (Applied Biosystems). Homology searches of nucleotide and deduced protein sequences were performed by FASTA and BLAST screens of the DDBJ, GenBank, and EMBL databases. Multiple-sequence alignments and searches for ORFs were performed with GENETYX-WIN software (Genetyx, Tokyo, Japan). The dendrogram for AACs was calculated with the CLUSTAL W Program (49).

Cloning of the aac(6′)-Iae gene.

The coding region of aac(6′)-Iae (Fig. 1) was amplified by PCR with 2.5 U of Ex Taq DNA polymerase (Takara Bio) and primers aacS1-FC and aacS1-RC (Table 1). The PCR products were cloned into pCRT7/NT (Invitrogen) downstream of the region encoding a six-His tag. Then plasmid pAAC6, which contains aac(6′)-Iae, or plasmid pREVAAC6, which contains aac(6′)-Iae in the reverse direction, was transformed into E. coli DH5α cells by the CaCl₂ method (6). DNA sequences of these cloned fragments were verified by sequencing of both strands as described above.

FIG. 1.

Structure of In2 (GenBank accession no. AF071413) and In113. Gene cassettes are represented as open boxes with an adjacent vertical bar (59-be), shown as heavy solid vertical bars. The novel ORF found in In113 is shown as a hatched box. Genes are indicated by horizontal arrows. IS are represented as gray boxes and are labeled. The sites of the 5′-CS, gene cassettes, 3′-CS, and tni module are indicated just below the construct. IRi and IRt are shown as vertical lines labeled i and t, respectively, and the attI1 sites are shown as open vertical bars toward the left of the constructs.

Purification of recombinant AAC(6′)-Iae.

E. coli BL21-AI harboring plasmid pAAC6 was grown to an A₆₀₀ of 0.2 to 0.3 in LB medium containing 50 mg/liter ampicillin at 37°C. After addition of arabinose (final concentration, 0.02%) to induce expression of AAC(6′)-Iae, the E. coli strain was cultured for another 18 h at 25°C. The bacterial cells were collected, resuspended in 50 mM HEPES buffer (pH 7.5) containing 0.1% Triton X-100, and lysed by sonication on ice for 15 s 40 times and then for 20 s 100 times. After centrifugation to remove the debris, the solubilized protein was applied to an AKTA Prime (Amersham Biosciences, Piscataway, NJ) system equipped with a HiTrap Chelating HP column (Amersham Biosciences) loaded with Ni²⁺. The column was washed with 20 mM Tris-HCl (pH 7.9) containing 60 mM imidazole and 0.5 M NaCl and was eluted with the same buffer containing 1 M imidazole. The eluted proteins were collected and dialyzed in 50 mM HEPES buffer (pH 7.5). The protein preparation yielded a single band upon sodium dodecyl sulfate-polyacrylamide gel electrophoretic analysis (data not shown).

Acetylation of aminoglycosides by recombinant ACC(6′)-Iae.

Enzymatic acetylation of aminoglycosides was done as described previously (53). Recombinant AAC(6′) from actinomycete strain #8 was provided by J. Ishikawa (National Institute of Infectious Diseases, Tokyo, Japan). Various aminoglycosides were incubated with recombinant AAC(6′)-Iae or AAC(6′) as a positive control in the presence of acetyl coenzyme A, and the acetylated derivatives were detected by thin-layer chromatography. The reaction was carried out at 37°C for 30 min to 12 h.

Pulsed-field gel electrophoresis (PFGE).

Genomic DNA from P. aeruginosa was prepared by the procedure of Grundmann et al. (14) and digested overnight with 10 U of SpeI, XbaI, or HpaI (Takara Bio). The DNA fragments were separated on 1.0% agarose gels in 0.5× Tris-borate-EDTA buffer with a CHEF Mapper system (Bio-Rad Laboratories, Hercules, Calif.) at 6 V/cm for 20 h.

Southern hybridization.

We performed Southern blotting to identify the location of In113. A 465-bp segment of aac(6′)-Iae and a 362-bp segment of bla_IMP-1 amplified by PCR were labeled with horseradish peroxidase and used as probes.

Nucleotide sequence accession number.

The nucleotide sequence of In113 reported here has been deposited in the EMBL/GenBank/DDBJ databases and assigned accession number AB104852.

RESULTS

Epidemiologic analysis of a nosocomial outbreak of P. aeruginosa.

From June 2002 to November 2002, a P. aeruginosa outbreak occurred in a neurosurgery ward of a 500-bed hospital in Japan. Three patients developed catheter-associated urinary tract infections with multidrug-resistant P. aeruginosa in June 2002. Various measures for infection control were undertaken, but four patients subsequently developed similar catheter-associated urinary tract infections with multidrug-resistant P. aeruginosa over the next 5 months. Seven P. aeruginosa isolates from these patients were analyzed by PFGE. The PFGE patterns of SpeI-, XbaI-, or HpaI-digested genomic DNAs from the isolates were identical, indicating that the isolates were all from monoclonal expansion of a single multidrug-resistant P. aeruginosa strain. This clone was named P. aeruginosa IMCJ2.S1. PFGE patterns of SpeI-, XbaI-, and HpaI-digested genomic DNAs from IMCJ2.S1 are shown in Fig. 2A.

FIG. 2.

(A) PFGE of SpeI-, XbaI-, and HpaI-digested genomic DNA from multidrug-resistant P. aeruginosa IMCJ2.S1. (B) Southern blotting of the same gels with an aac(6′)-Iae probe. Lanes M, HindIII-digested λ phage DNA as a size marker.

Susceptibility of P. aeruginosa IMCJ2.S1 to antibiotics and disinfectants.

The MICs of various antibiotics, including potent active β-lactams, against IMCJ2.S1 were compared with those against a reference strain, P. aeruginosa ATCC 27853 (Table 2). IMCJ2.S1 was resistant to all antibiotics tested except for arbekacin and polymyxin B. Strain ATCC 27853 was sensitive to all of the antibiotics tested except cefoxitin, flomoxef, and kanamycin. Thus, IMCJ2.S1 was classified as a multidrug-resistant strain of P. aeruginosa.

TABLE 2.

In vitro susceptibilities of P. aeruginosa IMCJ2.S1 and P. aeruginosa ATCC 27853 to various antimicrobial agents

Antibiotic	MIC (μg/ml) for:
	P. aeruginosa IMCJ2.S1	P. aeruginosa ATCC 27853
Piperacillin	>128	<4
Piperacillin-tazobactam	64	4
Cefotaxime	>128	8
Ceftazidime	>128	<1
Cefepime	>64	2
Cefoxitin	>64	>64
Flomoxef	>128	>128
Moxalactam	>128	16
Imipenem	128	4
Meropenem	128	1
Aztreonam	128	2
Amikacin	128	2
Arbekacin	2	<0.5
Dibekacin	>128	<0.5
Gentamicin	16	<1
Isepamicin	128	<4
Kanamycin	>128	>128
Netilmicin	>128	<0.5
Sisomicin	>128	<0.5
Streptomycin	>64	<4
Tobramycin	64	<0.5
Tetracycline	32	16
Sulfamethoxazole-trimethoprim	128	32
Levofloxacin	64	<0.5
Ciprofloxacin	32	<0.5
Polymyxin B	2	2
Silver sulfadiazine	64	64

To test whether IMCJ2.S1 showed increased resistance to disinfectants, the MBCs of four disinfectants, povidone iodine, alkyldiaminoethylglycine hydrochloride, benzalkonium chloride, and chlorhexidine gluconate, were determined for both IMCJ2.S1 and ATCC 27853. Both strains were resistant to chlorhexidine gluconate but sensitive to povidone iodine (MBC, <0.001% [wt/vol]), alkyldiaminoethylglycine hydrochloride (MBC, <0.001% [wt/vol]), and benzalkonium choride (MBC, <0.005% [wt/vol]). The MBC patterns of these strains were identical. These results indicate that the sensitivity of IMCJ2.S1 to disinfectants is not different from that of the P. aeruginosa reference strain.

Detection of an integron in P. aeruginosa IMCJ2.S1.

To determine if strain IMCJ2.S1 carried a class 1 integron, PCR analysis specific for class 1 integrons was performed (29). Strain IMCJ2.S1 yielded a 2.5-kbp PCR product, whereas E. coli CSH2 harboring plasmid NR1 (32), which carries In2 (30), yielded a 1.0-kbp PCR product. P. aeruginosa ATCC 27853 did not yield PCR products. These results suggest that strain IMCJ2.S1 and E. coli CSH2 each carry a class 1 integron and that this integron contains additional sequences that are not present in In2.

The class 1 integron frequently contains the tniB and tniA genes downstream of the 3′-CS (13, 16). To confirm the presence of a class 1 integron in IMCJ2.S1 and to elucidate the structure downstream of the 3′-CS, we performed PCR specific for intI1, qacEΔ1, sul1, and their spanning or marginal regions. PCRs yielded the expected products (Table 1), with the exception of a 4.7-kbp fragment after amplification with int-R and sul-R and a 2.5-kbp fragment after amplification with tniB-F and tniA-R. These data show that IMCJ2.S1 carries a class 1 integron and that this integron contains intI1-sul1 in a 4.7-kbp region, sul1-tniB in a 6.5-kbp region, and tniB-tniA in a 2.5-kbp region (Fig. 1).

Identical results were obtained for the other six isolates from the outbreak.

Structure of the class 1 integron found in P. aeruginosa IMCJ2.S1.

We analyzed the sequences of the PCR products to determine the structure of the class 1 integron of IMCJ2.S1. The 5′-CS contained intI1, the attI1 recombination site with a 7-bp core site sequence of GTTAGAA (45), and the TGGACA (−35) and TAAACT (−10) hexamers separated by 17 bp, which is characteristic of the Pc promoter (7, 45). Although TTGTTA (−35) and TACAGT (−10) hexamers separated by 14 bp were present again downstream of the Pc promoter, this region is not likely to act as the P2 promoter, because there is no GGG sequence (7, 45).

Between the 5′-CS and 3′-CS, there were three gene cassettes (Fig. 1). The 880-nucleotide (nt) cassette contained the metallo-β-lactamase gene bla_IMP-1 (35) and a 127-nt 59-base element (59-be) site, a site for site-specific cointegration events (Fig. 3), and this cassette was identical to one described previously (2, 35). The 647-nt cassette contained an ORF and a 68-nt 59-be site (Fig. 3). The sequence of this 647-nt cassette was not found in any database, and therefore, we named this integron In113 (Fig. 1). The ORF in the 647-nt cassette encoded a 183-amino-acid (aa) product that was 55.2% identical to a 6′-N-aminoglycoside acetyltransferase, AAC(6′)-Ia (48), and 57.1% identical to AAC(6′)-Iq of Klebsiella pneumoniae (4). We named the predicted protein AAC(6′)-Iae according to the standard nomenclature (42).

FIG. 3.

Structures of 59-be of In113. Seven-base-pair putative core sites in the left-hand (LH) and right-hand (RH) consensus sequences were designated 1L and 2L and 2R and 1R, respectively. The putative recombination event occurs between the G and the first T in the 1R core site and is indicated by vertical arrows (see reference 45). The relative orientations of 1L, 2L, 2R, and 1R are indicated by arrows under the sequence. An extra base in 2L is marked with an asterisk. Inverted repeats are underscored with arrows.

AAC(6′)-Iae was relatively similar to a subfamily of AAC(6′)-I enzymes that includes AAC(6′)-Ia (48), AAC(6′)-Iq (4), and AAC(6′)-Im (19) [which is not the AAC(6′)-Im reported by Chow et al. (5) and has also been referred to as AAC(6′)-Ip, by Centrón and Roy (4)] (61.7% identity in a 149-aa overlap) and to AAC(6′)-Ii (9) (40.3% identity in a 166-aa overlap) (Fig. 4). On the basis of the work of Neuwald and Landsman (34), four motifs in the amino acid sequences of the subfamily proteins belonging to AAC(6′)-Iae were designated motifs C, D, A, and B (Fig. 5). Comparison of amino acid sequences of members of the AAC(6′)-I subfamily with that of AAC(6′)-Iae revealed that motifs C, D, A, and B, which are found in most GCN5-related N-acetyltransferases (GNATs) (12, 34), were conserved in AAC(6′)-Iae (Fig. 5). A large motif at the C terminus, motif B (12), was 63.3% identical between AAC(6′)-Im (19) and AAC(6′)-Iae. The third cassette was 856 nt long and contained the aminoglycoside 3"-adenyltransferase gene aadA1 (18, 22) and a 60-nt 59-be site (Fig. 3). This cassette was similar to one reported previously (30, 36) except for a silent C-to-T substitution at nt 135.

FIG. 4.

Dendrogram of aminoglycoside 6′-N-acetyltransferases for comparison with AAC(6′)-Iae. The dendrogram was calculated with the CLUSTAL W program. Branch lengths correspond to the number of amino acid exchanges for AAC proteins. EMBL/GenBank/DDBJ accession numbers of AAC proteins are as follows: AAC(6′)-Ia, M18967-1; AAC(6′)-Ib, M23634; AAC(6′)-Ic, M94066; AAC(6′)-Id, X12618; AAC(6′)-Ie, M13771; AAC(6′)-If, X55353; AAC(6′)-Ig, L09246; AAC(6′)-Ih, L29044; AAC(6′)-Ii, L12710-1; AAC(6′)-Ij, L29045; AAC(6′)-Ik, L29510; AAC(6′)-Il, Z54241 and U13880; AAC(6′)-Im, Z54241-2; AAC(6′)-Iq, AF047556-1; AAC(6′)-Ir, AF031326; AAC(6′)-Is, AF031327; AAC(6′)-It, AF031328; AAC(6′)-Iu, AF031329; AAC(6′)-Iv, AF031330; AAC(6′)-Iw, AF031331; AAC(6′)-Ix, AF031332; AAC(6′)-Iy, AF144880; AAC(6′)-Iz, AF140221; AAC(6′)-Iaa, NC_003197; AAC(6′)-Iad, AB119105; AAC(6′)-IIa, M29695; AAC(6′)-IIb, L06163; AAC(6′)-29a, AF263519; AAC(6′)-29b, AF263519.

FIG. 5.

Alignment of the AAC(6′)-Iae amino acid sequence with those of four members of the AAC(6′)-I subfamily. Identical residues are marked with black boxes. Four motifs, including the highly conserved motif B, are underlined. A conserved region of 21 amino acids, described by Shmara et al. (44), is indicated by a dotted line. GenBank accession numbers are given in brackets to the right of AAC names. C. diversus, Citrobacter diversus; C. freundii, Citrobacter freundii.

The 3′-CS included qacEΔ1 (39), sul (47), and orf5 (30, 37). There were three inserted sequences (IS), IS1326 (3), IS1353 (3), and IS26 (38), in the region downstream of the 3′-CS (Fig. 1). IS26 is known to be inserted into the tniA coding region of the tni transposition module (30).

Drug resistance mediated by the AAC(6′)-Iae enzyme.

To examine the role of AAC(6′)-Iae in aminoglycoside resistance, a recombinant plasmid, pAAC6, carrying aac(6′)-Iae from strain IMCJ2.S1 was transformed into E. coli DH5α. E. coli harboring pAAC6 showed significantly lower susceptibility to amikacin, dibekacin, isepamicin, kanamycin, netilmicin, sisomicin, and tobramycin than the parent strain and the negative control. MICs for other aminoglycosides, including arbekacin, gentamicin, and streptomycin, were unchanged (Table 3). These results indicate that aac(6′)-Iae is involved in aminoglycoside resistance.

TABLE 3.

Aminoglycoside resistance patterns of E. coli DH5α alone or harboring plasmids with or without aac(6′)-Iae

Strain	MIC (μg/ml)aof:
	AMK	ABK	DIB	GEN	ISE	KAN	NEO	NET	SIS	STR	TOB
E. coli DH5α(pAAC6)b	8	0.5	32	0.25	8	64	4	32	16	4	8
E. coli DH5α(pREVAAC6)c	0.5	0.5	0.5	0.25	0.25	1	2	0.25	0.25	4	0.5
E. coli DH5α(pCRT7/NT)d	0.5	0.25	0.5	0.25	0.25	1	2	0.25	0.25	4	0.25
E. coli DH5α	0.5	0.5	0.5	0.25	0.25	1	2	0.25	0.25	4	0.25

^aAMK, amikacin; ABK, arbekacin; DIB, dibekacin; GEN, gentamicin; ISE, isepamicin; KAN, kanamycin; NEO, neomycin; NET, netilmicin; SIS, sisomicin; STR, streptomycin; TOB, tobramycin.

^bRecombinant plasmid constructed by cloning aac(6′)-Iae into pCRT7/NT.

^cRecombinant plasmid constructed by insertion of DNA fragment with reverse sequence of aac(6′)-Iae into pCRT7/NT.

^dCloning vector (ABPC).

To examine potential acetylase activity of AAC(6′)-Iae, we assessed the purified recombinant AAC(6′)-Iae against aminoglycosides by thin-layer chromatography (53). As shown in Fig. 6, kanamycin, amikacin, tobramycin, netilmicin, sisomicin, isepamicin, arbekacin, neomycin, and gentamicin were acetylated by AAC(6′)-Iae and AAC(6′). Acetylation by AAC(6′)-Iae was complete for all of these aminoglycosides except gentamicin, which showed incomplete acetylation. These aminoglycosides all have 6′-NH₂. The present results, therefore, suggest that AAC(6′)-Iae is a functional acetyltransferase that modifies the 6′-NH₂ position of aminoglycosides.

FIG. 6.

Thin-layer chromatogram of aminoglycosides incubated with AAC(6′)-Iae protein (I) or with AAC(6′) from Streptomyces lividans TK21 as a control (II) (53) in the presence (+) or absence (−) of acetyl coenzyme A. KAN, kanamycin; AMK, amikacin; TOB, tobramycin; ABK, arbekacin; GEN, gentamicin; NEO, neomycin; DIB, dibekacin; NET, netilmicin; SIS, sisomicin; ISE, isepamicin.

Location of In113.

Clinical isolates of P. aeruginosa frequently possess the R plasmid, which carries a class 1 integron. Therefore, we screened our seven P. aeruginosa clinical isolates for the presence of this plasmid. P. aeruginosa GN17203 was used as a positive control for bla_IMP-1, since it has been shown to harbor pMS350, which contains a bla_IMP-1 gene. Genomic DNA from IMCJ2.S1 was used as a control for aac(6′)-Iae and bla_IMP-1.

The extracts from the seven clinical isolates and P. aeruginosa GN17203 were separated by agarose gel electrophoresis, and Southern blotting with aac(6′)-Iae or bla_IMP-1 as a probe was performed. A plasmid that contained bla_IMP-1 but not aac(6′)-Iae was detected in P. aeruginosa GN17203. Despite repeated attempts (three times per procedure), we did not detect this plasmid by ethidium bromide staining or Southern blotting in any of the clinical isolates (data not shown). In contrast, Southern hybridization of SpeI-, XbaI-, and HpaI-digested genomic DNAs of the seven clinical isolates revealed 50-kb, 250-kb, and 60-kb aac(6′)-Iae-positive fragments, respectively (Fig. 2). These fragments were also positive for bla_IMP-1 (data not shown). To examine whether the drug-resistant phenotype of P. aeruginosa IMCJ2.S1 can be transferred by conjugation, IMCJ2.S1 was incubated with P. aeruginosa ATCC 27853 RFP^r. Carbapenem resistance was transferred from P. aeruginosa GN17203 to P. aeruginosa ATCC 27853 RFP^r, consistent with the results reported by Watanabe et al. (51). In contrast, resistance to amikacin or carbapenem was not transferred from IMCJ2.S1 to ATCC 27853 RFP^r. These results suggest that In113 is located in the chromosome, and not on a plasmid, of P. aeruginosa IMCJ2.S1.

Resistance of IMCJ2.S1 to fluoroquinolones.

IMCJ2.S1 was highly resistant to fluoroquinolones (Table 2). This resistance is typically associated with mutations in the QRDR within gyrA, gyrB, parC, and parE, which encode DNA gyrase or topoisomerase IV in P. aeruginosa (1, 21, 26, 31). Therefore, we screened IMCJ2.S1 mutations within the QRDR. Compared to the gyrA sequence of strain PAO1 (46), the gyrA sequence of IMCJ2.S1 contained an ACC-to-ATC mutation in codon 83 that causes a Thr-to-Ile change in the A subunit of DNA gyrase. IMCJ2.S1 also had a TCG-to-TTG mutation in codon 87 of parC that causes a Ser-to-Leu substitution in the C subunit of topoisomerase IV. IMCJ2.S1 had four mutations in gyrB: CGC to CGT in codon 396, AAA to AAG in codon 408, GAA to GAG in codon 484, and TTG to CTG in codon 513. There were four mutations in parE: GAA to GAG in codon 448, GGT to GGC in codon 472, AGT to AGC in codon 474, and GCC to GCT in codon 477. These mutations in gyrB and parE did not lead to amino acid changes in the proteins encoded (1, 31). Identical results were obtained with the other six clinical isolates. Together, these results indicate that IMCJ2.S1 contains mutations in gyrA and parC that are associated with its fluoroquinolone resistance.

DISCUSSION

A variety of aminoglycoside 6′-N-acetyltransferases have been described (Fig. 4) and classified into three subgroups (42, 50). Recently, a new enzyme, AAC(6′)-Iad, which is a member of the largest subfamily, was isolated from an Acinetobacter genospecies 3 strain in Japan (10). In the present study, we identified AAC(6′)-Iae, which shows considerable phylogenetic distance from members of the largest subfamily, which includes AAC(6′)-Iad and its divergents (Fig. 4). AAC(6′)-Iae belongs to the subfamily comprising AAC(6′)-Ia, -Ii, -Im, and -Iq (4, 9, 19, 48). There was only a low level of homology between the 59-be site of aac(6′)-Iae and those of the genes encoding other members of the aac(6′)-I family. Furthermore, aac(6′)-Iae has a low G+C content (26.8%) (data not shown), whereas the average G+C content of the P. aeruginosa PAO1 genome is 66.6% (46). Therefore, aac(6′)-Iae may be derived from an environmental species with an intrinsically low G+C content.

AAC(6′)-Iae from P. aeruginosa strain IMCJ2.S1, which was responsible for an outbreak of catheter-associated urinary tract infections, acetylated all of the aminoglycosides with 6′-NH₂, and acetylation of arbekacin and neomycin appeared to be complete (Fig. 6I). However, E. coli DH5α(pAAC6), expressing exogenous AAC(6′)-Iae, was sensitive to arbekacin and did not show reduced susceptibility to neomycin. Arbekacin and neomycin were shown to retain their antibiotic effects even after they were acetylated by AAC(6′) from an arbekacin-resistant actinomycete strain at the 6′ positions (53). Enterococcus faecium producing AAC(6′)-Ii was susceptible to neomycin even though AAC(6′)-Ii acetylated neomycin (52). These results suggest that acetylation of arbekacin and neomycin at 6′ positions does not affect the antimicrobial activities of these drugs. We cannot exclude the possibility that the antimicrobial activity observed after treatment with AAC(6′)-Iae is due to residual arbekacin or neomycin that was not acetylated.

E. coli DH5α expressing AAC(6′)-Iae was sensitive to gentamicin (Table 3), although AAC(6′)-Iae showed only partial acetylation of gentamicin (Fig. 6I). The sensitivity of these bacteria to gentamicin appears to be due to incomplete acetylation of gentamicin, which was observed with AAC(6′) from an arbekacin-resistant actinomycete strain (53)(Fig.6II). Commercially available gentamicin is a mixture of a number of derivatives of gentamicin, such as gentamicin C₁, C_1a, C₂, and C_2b, that have modifications of position 6′. Gentamicin C₁ and C_2b carry a methyl group on N-6′ and are refractory to AAC(6′)-I enzymes (42, 50). We cannot exclude the possibility that acetylated gentamicin components, which are more susceptible to AAC(6′)-I enzymes, retain antibiotic activity.

In the present study, we identified In113, a class 1 integron that contains a novel aminoglycoside resistance gene, aac(6′)-Iae. Several classes of integrons have been categorized on the basis of the structure of integrase (15, 40). The most common integrons in P. aeruginosa are those of class 1 (27, 28, 37). Because their structures are very similar to each other, the direct origin of In113 could be from In2 (30), which was originally isolated from Shigella flexneri in Japan in the late 1950s (32) (Fig. 1).

IMCJ2.S1 was resistant to all antibiotics tested except arbekacin and polymyxin B (Table 2). However, the presence of In113 and the mutations in gyrA and parC of the QRDR are not sufficient to explain the multidrug resistance of this strain. Alterations of gyrA and parC are known to contribute to fluoroquinolone resistance (1, 21, 26, 31). The bla_IMP-1 gene cassette, which encodes the IMP-1 metallo-β-lactamase, confers resistance to all β-lactams except monobactams (2, 27, 35). The aac(6′)-Iae gene cassette, which encodes AAC(6′)-Iae, confers resistance to amikacin, dibekacin, isepamicin, kanamycin, netilmicin, sisomicin, and tobramycin (Table 3). The variant aadA1 gene cassette, which encodes aminoglycoside 3"-adenylyltransferase, confers resistance to streptomycin (18, 22). The sul1 gene, which encodes dihydropteroate synthetase type I, confers resistance to sulfamethoxazole (47). Thus, the resistance of IMCJ2.S1 to aztreonam, gentamicin, tetracycline, trimethoprim, and silver sulfadiazine appears to be related to another, unidentified resistance factor(s).

In conclusion, we describe here a novel aminoglycoside 6′-N-acetyltransferase gene contained on a class 1 integron in a P. aeruginosa strain that caused a nosocomial outbreak of urinary tract infections. In113 may spread across Japan, because β-lactams, including carbapenems and aminoglycosides, are frequently used as therapeutic agents against P. aeruginosa and methicillin-resistant Staphylococcus aureus (20, 23). Surveillance for multidrug-resistant P. aeruginosa containing In113 is under way at several medical care facilities in the Sendai area of Japan.