Class 1 Integrons Increase Trimethoprim-Sulfamethoxazole MICs against Epidemiologically Unrelated Stenotrophomonas maltophilia Isolates

Abstract

Twenty-five plasmid-specified antimicrobial resistance determinants common to gram-negative bacilli from nosocomial infection were investigated from 31 Stenotrophomonas maltophilia isolates. Twenty-four clones were identified by pulsed-field gel electrophoresis, and in three clones that exhibited an increased trimethoprim-sulfamethoxazole MIC, the sul1 determinant was found. These results support not only the higher spread of class 1 integrons compared to other mechanisms but also the potential limitation of using trimethoprim-sulfamethoxazole for therapy of severe S. maltophilia infections.

In the last decade Stenotrophomonas maltophilia has emerged as a nosocomial pathogen related to a significant case/fatality ratio in certain patient populations (7).

S. maltophilia infections are difficult to treat due to their intrinsic antibiotic resistance (1). Although trimethoprim-sulfamethoxazole has been found to be active against almost all isolates and has also long been considered the antimicrobial therapy of choice for severe S. maltophilia infections (4, 7, 11), Muder et al. suggested that the combination of trimethoprim-sulfamethoxazole with ticarcillin-clavulanate or an extended-spectrum cephalosporin might be superior to trimethoprim-sulfamethoxazole used alone (13).

Though several antimicrobial determinants have been described as conferring trimethoprim-sulfamethoxazole resistance in gram-negative bacilli (10), such mechanisms have not been previously reported in S. maltophilia isolates. In addition, the antimicrobial resistance determinants frequently located on plasmids have not been studied yet in this species, even though it is expected that many of the acquired mechanisms found in Pseudomonas aeruginosa, Acinetobacter spp., and other gram-negative bacteria are likely to be present.

In this study, plasmidic antimicrobial determinants common in the gram-negative nosocomial bacterial population of Buenos Aires, Argentina (15, 18), were investigated to determine their role in the susceptibility profiles of unrelated S. maltophilia clinical isolates. During an 8-month period (from December 2000 to July 2001), 31 epidemiologically unrelated isolates were recovered from two hospitals in Buenos Aires. Bacterial isolates were grown in brain heart infusion agar (Difco Laboratories, Detroit, Mich.) at 30°C. Species identification was performed by phenotypic methods (9) and confirmed by rRNA-PCR which yields a specific 531-bp consensus sequence from the 23S rRNA gene from S. maltophilia (20). Macrorestriction was used for clone delineation. For this purpose, genomic DNA was digested with 10 U of SpeI (Promega Corporation, Madison, Wis.) according to the procedure described by the manufacturer. Endonuclease-digested genomic DNAs were separated by pulsed-field gel electrophoresis (PFGE) with a CHEF-DR III system (Bio-Rad, Richmond, Calif.) (16). Running conditions were 24 h at 14°C, with an initial switching time of 1 s and final time of 30 s, at 6 V/cm. Concatemers of DNA isolated from bacteriophage λ (New England Biolabs, Beverly, Mass.) were run next to DNA fragments for size comparison. DNA banding patterns were analyzed by visual examination and all loci were scored for the presence or absence of a band. Strain delineation was inferred in terms of the percentage similarity of banding patterns (17). Similarities were estimated by the Dice coefficient (8). Cluster analysis was carried out by the unweighted pair-group method with arithmetic averages and the software program NTSyS-PC version 1.40. Thirty-one S. maltophilia isolates were studied by PFGE and showed 25 distinguishable banding patterns. Similarity among 24 of them ranged from 20 to 65%, and they were considered to represent 24 different clones (Fig. 1A). Only one isolate showed 95% banding pattern similarity to another isolate from the same cluster, and it was classified as a subtype R1 (Fig. 1A). Figure 1B shows several of the banding patterns observed.

FIG. 1.

(A) SpeI macrorestriction clustering analysis of S. maltophilia isolate fingerprints. The cutoff level for clones was 75% of similarity. The R1 isolate showed a banding pattern similarity of 95% to another isolate of the same cluster and was classified as a subtype. (B) Macrorestriction fingerprint of nine S. maltophilia isolates digested with SpeI. Lane 1, isolate 13 (clone S); lane 2, isolate 1 (clone A); Lane 3, isolate 3 (clone B); lane 4, isolate 4 (clone C); lane 5, isolate 11 (clone R); lane 6, isolate 6 (clone D); lane 7, isolate 7 (clone E); lane 8, isolate 8 (clone F); lane 9, isolate 9 (clone G); lane 10, isolate 10 (clone H); M, lambda phage ladder DNA concatemers (band sizes are expressed on the right in kilobases).

Antibiotic susceptibility tests were performed according to an agar dilution method based on the methodology described by the National Committee for Clinical Laboratory Standards (14). Antibiotics were purchased from Sigma Clinical Co. (St. Louis, Mo). Inocula of 10⁴ bacteria per spot were deposited with a Steer inoculator. The MIC of trimethoprim, sulfamethoxazole, and trimethoprim-sulfamethoxazole was defined as ≥90% inhibition of growth compared to growth on control (no-drug) plates (5). P. aeruginosa ATCC 27853 was used as the reference strain.

Drug susceptibility tests gave the following MIC₉₀s (in micrograms/milliliter; ranges are given in parentheses): ticarcillin-clavulanic acid, 16 (1 to 256); gentamicin, 256 (16 to >256); amikacin, >256 (8 to >256); levofloxacin, 2 (<0.5 to 16); ciprofloxacin, 8 (<0.5 to 128); moxifloxacin, 1 (<0.5 to 16); tetracycline, 16 (0.5 to 64); sulbactam, 256 (64 to 256); cefepime, 32 (1 to 64); ceftazidime, 64 (1 to 128); imipenem, 128 (8 to 256); and meropenem, 64 (<0.5 to 256). These susceptibility test results were similar to others reported previously (1, 19). Table 1 shows that the sulfamethoxazole MICs of isolates 3, 11, and 13 were 3 to 8 dilutions higher than the MICs for the remaining isolates. Besides, these isolates exhibited trimethoprim-sulfamethoxazole MICs that were 4 to 8 dilutions higher than the MICs for other isolates, and the increase in the trimethoprim MIC was 2 to 4 dilutions.

TABLE 1.

General features of S. maltophilia isolates

Isolate no.	Hospital	Sampleb	Genotypec	Class 1 integrond	Susceptibility (MIC [μg/ml])e
					TMP	SUL	SXT
1	H1	Bronchial asp	A	−	16	8	0.12/2
2	H1	Bronchial asp	A	−	4	4	0.12/2
3	H1 outa	Urine	B	+	64	1,024	8/152
4	H1	BAL	C	−	16	8	0.12/2
5	H1	Tracheal asp	C	−	16	8	0.12/2
6	H1	Bronchial asp	D	−	4	4	0.12/2
7	H1	Blood	E	−	16	4	0.12/2
8	H1	Bronchial asp	F	−	16	8	0.12/2
9	H1	Urine	G	−	16	8	0.12/2
10	H1	Blood	H	−	16	4	0.12/2
11	H1 out	Ear effusion	R	+	64	32	2/38
12	H1	BAL	R	−	4	4	0.12/2
13	H1	Urine	S	+	64	1,024	16/304
14	H1	Urine	R1	−	8	8	0.25/4
15	H1 out	Soft tissue	T	−	8	8	0.12/2
16	H1	Tracheal asp	U	−	4	8	0.12/2
17	H1	Bronchial asp	V	−	8	8	0.12/2
18	H1	Catheter	W	−	4	8	0.12/2
19	H2	BAL	J	−	16	8	0.12/2
20	H2	Bronchial asp	L	−	16	8	0.25/4
21	H2	Blood	N	−	16	8	0.25/4
22	H2	BAL	P	−	16	8	0.12/2
23	H2	Bronchial asp	K	−	16	8	0.12/2
24	H2	BAL	M	−	16	8	0.12/2
25	H2	Bronchial asp	Q	−	4	8	0.12/2
26	H2	Tracheal asp	X	−	4	4	0.12/2
27	H2	Catheter	X	−	4	8	0.12/2
28	H2	Catheter	I	−	4	8	0.12/2
29	H2	Blood	O	−	4	4	0.12/2
30	H2	Blood	O	−	4	4	0.12/2
31	H2	Blood	O	−	4	8	0.12/2

^aOut, outpatients from hospital H1. The other samples are from nosocomial infection or colonization.

^basp, aspiration; BAL, bronchoalveolar lavage.

^cGenotype was determined by PFGE.

^d+, presence of class 1 integron; −, absence of class 1 integron. MICs for isolates with a class 1 integron are in boldface.

^eTMP, trimethoprim; SUL, sulfamethoxazole; SXT, trimethoprim-sulfamethoxazole. The ratios for SXT (trimethoprim/sulfamethoxazole) correspond to the concentration (1:19) of SXT tested as previously described (9).

Genomic DNA was analyzed by standard protocols using sodium dodecyl sulfate, lysosyme, and proteinase K, followed by phenol-chloroform-isoamyl alcohol extraction and ethanol precipitation. PCRs for plasmidic antimicrobial resistance determinants common in our gram-negative nosocomial bacterial population were performed as previously described (12). Twenty-five PCRs were carried out for each isolate with a set of primers specific for the nucleotide sequences, as previously described (15). These primers detect the presence of class 1, 2, and 3 integrons (Tn3, Tn1331, and Tn9) and orf513-bearing class 1 integrons (bla_FOX, bla_SHV, bla_PER-2 and bla_CTX-M-2). In some cases, PCR was performed by using the internal primers for the class 1 integrase gene (sulpro3 primer) and sul1, located in the 3′-conserved segment of class 1 integrons. PCR cartography was carried out with several combinations of the sulpro3 primer and primers of the cassette gene. Detection of the sulfamethoxazole resistance determinants described in gram-negative bacteria was made with specific primers for sulI (Sul1f [5′-GACGGTGTTCGGCATTCT-3′] and Sul1r [5′-TTTGAAGGTTCGACAGC-3′]) and sul2 (sul2f [5′-GCAGGCGCGTAAGCTGA-3′] and sul2r [5′-GGCTCGTGTGTGCGGATG-3′]) (GenBank accession number AJ313522).

PCR results were corroborated by sequencing analysis of amplimers obtained by using the sulpro3 and sulIr primers with several internal primers, which yielded the products of the cassettes of the variable region of class 1 integrons (Table 1). PCR fragments were purified with a QIAquick kit (QIAGEN Inc., Studio City, Calif.), according to the manufacturer's directions. Both DNA strand sequences were obtained by using an ABI 373 sequencer (Applied Biosystem Division). Internal oligonucleotide primers were used in order to ensure that both strands were sequenced. The nucleotide sequences were analyzed by Genetics Computer Group (GCG) software.

Investigation of common plasmid nosocomial-mediated trimethoprim and sulfamethoxazole determinants, such as class 1 integrons for trimethoprim family 1 and family 2, class 2 integrons (or Tn7) for dfrAI and dfrAIb, and sul1 and sul2 (10), demonstrated that the sul1 gene is present in isolates 3, 11, and 13. This result explained the highest sulfamethoxazole MIC observed in these isolates. The increased trimethoprim MIC for these isolates compared to the MICs for other isolates (64 versus 16 to 4 μg/ml, respectively) included in this study could not be explained by the presence of plasmid-specified trimethoprim resistance determinants, and the increase might rely on chromosomal mechanisms (10). Although the sul1 gene has never been shown to exhibit activity in folic acid synthesis, it could be also speculated that the presence of the sul1 gene could contribute to this MIC increase because it codes for a dihydropteroate synthase enzyme that acts in a previous step of the folic acid synthesis. Irrespective of the possible effect of the sul1 determinant on the trimethoprim MIC, its presence may affect the trimethoprim-sulfamethoxazole MIC since for isolates lacking this plasmidic determinant, the MICs ranged from 0.12 (trimethoprim)/2(sulfamethoxazole) to 0.25/4 μg/ml versus 2/38 to 16/304 μg/ml for those with the determinant. In this regard, recently a MIC of 8/152 μg/ml (19) was suggested as the resistance breakpoint of trimethoprim-sulfamethoxazole for S. maltophilia isolates, rather than the NCCLS recommendation of 4/76 μg/ml, based on the well-demonstrated bimodal distribution of the trimethoprim-sulfamethoxazole MIC. Our results, however, showed that a MIC of 2/38 μg/ml could be due to the presence of a sulfamethoxazole determinant, and, therefore, whether there is a correlation between the presence of sul1 and antimicrobial therapy failure should be determined.

Three different cassette rearrangements, catB2, aac(6′)-Ib-ant(3")-Ia, and aac (3)-Ia-orfP-orfP-orfQ-ant(3")-Ia, were found in isolates 3, 11, and 13 (Fig. 2). Isolates number 3 and 11 harbored a new order of cassettes, and the gene sequences have already been reported to GenBank [accession number AF047479 for catB2, accession number X60321 for aac(6′)-Ib, and accession number AJ487034 for aadA1]. Isolate number 13 harbored a cassette order already described (AF453999) (6).

FIG. 2.

Class 1 integrons found in S. maltophilia strains. Isolates 3 and 11 harbored new cassette rearrangements. Isolate 13 harbored a cassette order described previously.

Isolates 3, 11, and 13 were classified into three different clones (Table 1): B, R, and S, respectively; two of the isolates were recovered from outpatients, indicating that they are genetically and epidemiologically unrelated. Clone R of isolate number 12 and clone R1 of isolate number 14, which lack class 1 integrons, indicate also that the presence of the sulI gene was not clonally related. In addition, its presence determines the increase of the trimethoprim-sulfamethoxazole MIC.

Other plasmidic antimicrobial determinants investigated were not found in the studied isolates. Several recent findings such as the plasmid location of L1 and L2 β-lactamases (3) as well as a cluster of genes from gram-positive bacteria involved in antibiotic and heavy-metal resistance in S. maltophilia (2) suggested the ability of S. maltophilia to be involved in genetic transfer and in the acquisition of genetic determinants from other bacteria. Based on these findings, and on the fact that in recent years several integrons and transposons were found in almost all species tested, frequently acquired antimicrobial mechanisms were also expected to be present in S. maltophilia. However, only 0.9% of isolates harbored class 1 integrons. At least two hypotheses could be proposed to explain this fact. First, S. maltophilia is intrinsically resistant to a wide variety of antimicrobial agents; therefore, its survival does not depend on the selection of strains that harbor acquired mechanisms, as in the case of almost all the opportunistic species in the nosocomial environment. Second, intrinsic features may not allow frequent acquisition of antimicrobial determinants. In this regard, our results support the higher spread of the class 1 integrons compared to other mechanisms and also the potential limitation of trimethoprim-sulfamethoxazole therapy for severe S. maltophilia infections.