Rapid Detection and Sequence-Specific Differentiation of Extended-Spectrum β-Lactamase GES-2 from Pseudomonas aeruginosa by Use of a Real-Time PCR Assay

Abstract

The LightCycler was compared to nested PCR for the detection of bla_GES/IBC genes from 100 Pseudomonas aeruginosa clinical isolates. The real-time PCR assay detected a bla_GES/IBC gene product from 83 isolates, exhibiting a sensitivity and specificity of 94.3 and 100% respectively, compared to nested PCR and DNA sequencing.

Detection of GES-2 and other extended-spectrum β-lactamase (ESBL)-producing Pseudomonas aeruginosa isolates in the clinical microbiology laboratory utilizing ordinary methods is notoriously difficult and laborious (20). Apart from the known distribution of GES/IBC-type ESBLs in P. aeruginosa and members of the family Enterobacteriaceae (18, 20), recent reports from Brazil and Japan show that these enzymes are present in P. aeruginosa and Klebsiella pneumoniae, respectively (3, 17). In addition, the integron genetic structures that support these ESBLs, such as GES-2, not only confer resistance to broad-spectrum β-lactam antibiotics but also to unrelated classes of antimicrobials (20), making these isolates very difficult to treat and control successfully. Apart from standard PCR and gene sequencing, molecular detection methods for ESBLs in general led to a wide array of applications most suited for the TEM- and SHV-type of ESBL families (1, 2, 6, 8, 10, 12, 16). This study describes the application of the LightCycler to rapidly and specifically detect bla_GES-2 in the GES/IBC ESBL family from South African P. aeruginosa isolates.

Bacterial isolates and DNA extraction.

One hundred P. aeruginosa clinical isolates, identified and collected from 1998 to 2001, as described previously (19), were included in this study. Control isolates are listed in Table 1. Disk susceptibility testing was conducted as described previously (11, 19), with ceftazidime resistance being a requirement for inclusion. Extraction of whole-cell DNA was accomplished with an ethanol precipitation-based method (5, 13), with DNA extraction being visually verified (19).

TABLE 1.

Well-characterized bacterial strains used for standardization of molecular methods

Strain	Relevant property	Source or reference
P. aeruginosa GW-1	blaGES-2-producing isolate	14
P. aeruginosa Pu21	blaGES-2 transconjugant isolate	14
K. pneumoniae ORI-1	blaGES-1-producing isolate	15
Enterobacter cloacae HT-9	blaIBC-1-producing isolate	4
E. coli brcpHT-8	blaIBC-1 transconjugate isolate	4
E. coli DH5α	blaIBC-2 transconjugant isolate	9
E. coli ATCC 25922	blaGES/blaIBC-negative strain	ATCCa

^aATCC, American Type Culture Collection, Manassas, Va.

LightCycler primer/probe design.

A fluorescence resonance energy transfer-mediated mutation assay was performed with the LightCycler as described previously (16), with reaction-specific modifications as necessary. Forward primer GES-C and previously described reverse primer GES-1B (Table 2) amplified a 689-bp product. A 3′ fluorescein isothiocyanate-labeled sensor probe was designed from the gene sequence of bla_GES-1 (15), spanning the mutation site of nucleotides (nt) 493 and 494 of bla_GES-2 (14) (Table 2). A 33-mer anchor probe (Table 2) was designed to bind at a distance of 1 nt directly downstream of the sensor probe. Design and placement of fluorogenic probes are depicted in Fig. 1.

TABLE 2.

Oligonucleotide sequences used in this study

Primer or probea	Sequenceb	Tm (°C)	Position (nt) on blaGES	Function	Source or reference
Nested PCR
GES-1A	ATG CGC TTC ATT CAC GCA C	56.6	01-19	Forward primer	14, 20
GES-1B	CTA TTT GTC CGT GCT CAG G	56.4	846-864	Reverse primer	14, 20
GES-C	GTT TTG CAA TGT GCT CAA CG	54.6	176-195	Forward primer	19
GES-D	TGC CAT AGC AAT AGG CGT AG	55.6	527-546	Reverse primer	19
Real-time PCR
GES-C	As above
GES-1B	As above
Probe 1	AGA TGG GCG ACA ACA CAC CT-FL	59.8	488-507	Sensor probe complementary to blaGES-1/blaIBC	This study
Probe 2	LC Red 640-GCG ACC TCA GAG ATA CAA CTA CGC CTA TTG CTA-PH	68.4	509-541	Anchor probe	This study

^aPrimers (GES designations) were synthesized and purified by integrated DNA Technologies, Coralville, Iowa.

^bSequences are delineated 5′ to 3′ Fluorophore-labeled probes were synthesized and purified by TIB Molecular Biology, Berlin, Germany. Boldface letters depict the relevant point mutation. FL, fluorescein; PH, phosphorylation; LC Red 640, LightCycler Red 640.

FIG. 1.

Alignment of real-time PCR primer and fluorogenic probe sequences with bla_GES-2. Stem-loop sequences appear in boxes, and arrows indicate the directions of stem-loop formation. The relevant nucleotide mismatch (nt 493 and 494) is depicted as a shaded box. Note the close approximation of the sensor probe to the stem-loop structure in addition to the 1-nt distance between the sensor and anchor probes. Also shown are forward and reverse primers (nt 488 to 507, sensor probe 3′ marked with fluorescein [FL]; and nt 509 to 541, anchor probe 5′ marked with LightCycler Red 640 [LCR640]).

LightCycler PCR.

The reaction mixture comprised a 0.5 μM concentration of each primer, a 0.2 μM concentration each of the sensor and anchor probes, 4 μl of LightCycler FastStart DNA Master^PLUS reaction mix (Roche Diagnostics, Penzberg, Germany), 5 μl of DNA template, and distilled water to a final reaction volume of 20 μl. DNA templates from control isolates (Table 1) and distilled water served as positive and negative controls, respectively. Amplification and melting curve analysis used in this study deviated from previously published data (16), as shown in Table 3. Data analysis was performed as previously described (16). Amplicons from control isolates (Table 1) were analyzed by agarose gel electrophoresis (16).

TABLE 3.

LightCycler amplification and melting curve protocol followed in this study

Program	No. of cycles	Target temp (°C)	Hold time (s)	Temp transition rate (°C/s)	Fluorescence acquisition modea
Polymerase activation	1	95	600	20	None
Three-step PCR	40
Denaturation cycle		96	1	20	None
Amplification cycle		52	10	20	Single
Extension cycle		72	15	20	None
Three-step melting curve	1
Denaturation cycle		95	10	20	None
Holding cycle		45	30	20	None
Melting cycle		95	0	0.2	Continuous

^aFluorimeter gains settings were F1 equal to 1, F2 equal to 15, and F3 equal to 30.

Nested PCR and DNA sequencing.

Primers GES-1A and GES-1B (Table 2) were used to amplify the entire bla_GES/IBC coding region. Second-round amplification utilized primers GES-C and GES-D (Table 2) and 2 μl of the amplification product from the initial PCR, targeting a 371-bp region. Reaction conditions and controls were similar to those of a previously published standard PCR method (19). Automated sequencing of 371-bp amplicons was performed as previously described (19).

Results and discussion.

Results obtained with real-time PCR, nested PCR, and DNA sequencing methods are shown in Table 4. Electrophoresis of amplicons obtained from control isolates (Table 1) showed that all exhibited a 689-bp bla_GES/IBC gene product except the Escherichia coli ATCC 25922 isolate (data not shown). No DNA template amplification was observed with distilled water controls. A stem-loop in the secondary DNA structure was noted as previously described (19). The LightCycler, based on real-time fluorimetric measurement of amplification products, proved ideal to analyze the “hot spot” area in the bla_GES gene for resistance mutations, as previously described for Candida albicans isolates (7). Previous real-time PCR work conducted for the detection of SHV-type ESBLs from Enterobacteriaceae clinical isolates (16) found a 100% sensitivity and specificity for the described LightCycler assay. During this study, the LightCycler and nested PCR assays detected a bla_GES/IBC product from 83 and 88 clinical isolates, respectively, exhibiting a sensitivity of 94.3% for the LightCycler compared to nested PCR and a 100% specificity compared to sequencing data. The noted decrease in sensitivity may have been influenced by a number of factors including the quality of the DNA template used in the experiments (20) and the choice of both DNA polymerase and polymerase buffer systems utilized in the LightCycler assay (21). Based on a previous study which demonstrated a ca. 6°C temperature shift per nucleotide mismatch (16) and due to the 2-nt mismatch between bla_GES-2 and the sensor probe, a ca. 10 to12°C difference in melting temperature (T_m) was expected between bla_GES-2 and bla_GES-1/IBC amplification products. However, T_m analyses of bla_GES-2 and bla_GES-1 products obtained from control isolates demonstrated a difference of 7.8 to 9.84°C, whereas bla_GES-2- and bla_IBC-type products exhibited a temperature difference of 8.64 to 10.42°C (Table 4). Analyses of 83 bla_GES-2 clinical isolates yielded a T_m of 55.63 ± 0.33°C, clearly distinguishing the 2-nt mismatch from bla_GES-1-type amplification products (Fig. 2). Where there was an exact match between the sensor probe and the template sequence, T_m values of 64.07 ± 0.72°C (bla_GES-1) and 64.78 ± 0.59°C (bla_IBC) were found (Table 4), in comparison with 66°C reported previously (16). One clinical isolate with a T_m of 66.83°C yielded a bla_GES-1 product on sequence analysis, making this the first report of β-lactamase GES-1 from a South African P. aeruginosa clinical isolate. Currently, molecular detection of ESBL-encoding genes from P. aeruginosa is the only reliable method available and, as such, the LightCycler has proved to be sensitive and highly specific in the rapid detection of ESBL genes of the GES/IBC family from P. aeruginosa.

TABLE 4.

Results obtained with nested PCR, real-time PCR, and DNA sequencing methods

Method	Isolate type	No. blaGES/IBC amplification positive	No. blaGES/IBC amplification negative	No. of amplicons analyzed	Tm (°C) ± SD	Sequencing result
Nested PCR	Clinicala	88	12	87		blaGES-2
				1		blaGES-1
Real-time PCR	Clinicala	83	17
Melting curve analyses	Clinical			82	55.63 ± 0.33	blaGES-2
	Clinical			1	66.83	blaGES-1
	GES-2 controlsb				55.25 ± 0.3
	GES-1 controlb				64.07 ± 0.72
	IBC controlsb				64.78 ± 0.59

^aAll isolates tested were resistant to ceftazidime.

^bControl isolates were shown in Table 1.

FIG. 2.

Melting peaks of bla_GES-1, bla_GES-2, and E. coli ATCC 25922 (bla_GES/IBC template-negative control). Amplification products are plotted as the negative derivative of fluorescence F2 [(−d(F2)/dT)] versus temperature. The T_m difference between an exact sensor probe match (GES-1) and a 2-nt mismatch (GES-2) is clearly visible. No melting peak was generated with the negative control isolate.