Detection of Legionella pneumophila Using a Real-Time PCR Hybridization Assay

Abstract

A real-time PCR hybridization assay for Legionella pneumophila is described; the assay uses LightCycler (Idaho Technology) methodology to specifically detect 2.5 CFU/reaction, equivalent to 1,000 CFU/liter of starting water sample. The assay, including DNA extraction and confirmation of product identity, is completed within 90 min of receipt of a sample.

Legionnaires' disease is normally acquired by inhalation or aspiration of Legionella pneumophila serogroup 1 from a contaminated environmental source. Rapid identification of the source of infection is essential to prevent further cases of disease, and a number of conventional PCR assays for the detection of Legionella spp. have been described previously (6–10, 12–17). The introduction of rapid thermal cyclers combined with microvolume fluorimeters (e.g., the LightCycler; Idaho Technology, Idaho Falls, Idaho) now enables >30 PCR cycles in <20 min, combined with immediate confirmation of PCR product identity. This paper describes a prototype real-time assay using LightCycler methodology that detects L. pneumophila within 90 min of receipt of water samples.

Primers mip-Lpn0901F (5′-AACCGATGACACATCATTA) and mip-Lpn1011R (5′-CTTGCATGACTTTAGCCA) were designed to amplify a 131-bp region at the 5′ end of the macrophage infectivity potentiator (mip) gene of L. pneumophila (3). These were used in conjunction with mip-specific hybridization probe mip-Lpn0941P (5′-Cy5-TCGGCACCAATGCTATAAGA-biotin).

Organisms used to assess the specificity of the assay are listed in Table 1. DNA was extracted from a single colony in 180 μl of lysis (ATL) buffer in a DNA Mini Kit (QIAgen Ltd., Crawley, United Kingdom) to a final eluate of 50 μl. One-liter natural water samples were concentrated (4) to a final volume of ca. 1 ml, of which 80 μl was added to 100 μl of ATL buffer and processed using the QIAgen kit to a final DNA eluate of 50 μl. DNA from laboratory-maintained water microcosms was extracted by centrifuging a 1-ml sample at 5,000 × g for 5 min, discarding 920 μl of supernatant, and, following resuspension of the pellet, processing the remaining 80 μl with the QIAgen kit as before. All DNA eluates were diluted 1:10 in 0.2% (wt/vol) bovine serum albumin (Sigma, Poole, United Kingdom) to minimize inhibition (11). For sensitivity testing and construction of a standard curve, samples from a 1-liter microcosm containing L. pneumophila serogroup 1 were prepared and the viable count was determined by plating out 100-μl portions of appropriate dilutions on buffered charcoal yeast extract agar plates (5).

TABLE 1.

Strains used for testing the specificity of the L. pneumophila LightCycler assay

Species	Serogroup	Designation		Original designation	Result with L. pneumophila-specific LightCycler assay
		NCTC	ATCC
L. pneumophila subsp. pneumophilaa	1	11192	33152	Philadelphia-1	Positive
	1	12024	43106	Allentown-1	Positive
	1	11404	43111	Bellingham-1	Positive
	1	12006	43108	Benidorm 030E	Positive
	2	11230	33154	Togus-1	Positive
	3	11232	33155	Bloomington-2	Positive
	5	11417		Cambridge-2	Positive
	6	11287		Oxford-1	Positive
	7	11984	33823	Chicago-8	Positive
	8	11985	35096	Concord-3	Positive
	9	11986	35289	IN-23-G1-C2	Positive
	10	12000	43283	Leiden-1	Positive
	11	12179	43130	797-PA-H	Positive
	12	12180	43290	570-CO-H	Positive
	13	12181	43736	82A3105	Positive
	14	12174	43703	1169-MN-H	Positive
	15	NDb		Portland-1	Positive
	16	ND		Jena-1	Positive
L. pneumophila subsp. fraseri	4	11233	33156	Los-Angeles-1	Positive
L. pneumophila subsp. pascullei		12273	33737	U8W	Positive
L. adelaidensis		12735	49625	1762-AUS-E	Negative
L. anisa		11974	35292	WA-316-C3	Negative
L. birminghamensis		12437	43702	1407-AL-H	Negative
L. bozemaniae	1	11368	33217	WIGA	Negative
L. brunensis		12240	43878	441-1	Negative
L. cherrii		11976	35252	ORW	Negative
L. cincinnatiensis		12438	43753	72-OH-H	Negative
L. dumoffii		11370	33279	NY-23	Negative
L. erythra		11977	35303	SE-32-A-C8	Negative
L. fairfieldensis		12488	49588	1725-AUS-E	Negative
L. feeleii	1	12022	35072	WO-44C	Negative
L. geestiana		12373	49504	1308	Negative
L. gormanii		11401	33297	LS-13	Negative
L. gratiana		12388	49413	8420412	Negative
L. hackeliae	1	11979	35250	Lansing-2	Negative
L. israelensis		12010	43119	Bercovier-4	Negative
L. jamestowniensis		11981	35298	JA-26-G1-E2	Negative
L. jordanis		11533	33623	BL-540	Negative
L. lansingensis		12830	49751	1677-MI-H	Negative
L. londiniensis		12374	49505	1477	Negative
L. longbeachae	1	11477	33462	Long Beach-4	Negative
L. lytica		ND	ND	L2	Negative
L. maceachernii		11982	35300	PX-1-G2-E2	Negative
L. micdadei		11371	33218	TATLOCK	Negative
L. moravica		12239	43877	316-86	Negative
L. nautarum		12375	49506	1224	Negative
L. oakridgensis		11531	33761	OR-10	Negative
L. parisiensis		11983	35299	PF-209C-C2	Negative
L. quateirensis		12376	49507	200/83-1335	Negative
L. quinlivanii	1	12433	43830	1442-AUS-E	Negative
L. rubrilucens		11987	35304	WA-270A-C2	Negative
L. sainthelensi	1	11988	35248	Mt.St.Helens-4	Negative
L. santicrucis		11989	35301	SC-63-C7	Negative
L. shakespearei		12829	49655	214	Negative
L. spiritensis		11990	35349	Mt.St.Helens-9	Negative
L. steigerwaltii		11991	35302	SC-18-C9	Negative
L. tucsonensis		12439	49180	1087-AZ-H	Negative
L. wadsworthii		11532	33877	81-716A	Negative
L. waltersii		13017	51914	2074-AUS-E	Negative
L. worsleiensis		12377	49508	1347	Negative
Bartonella hensalae					Negative
Bartonella doshiae					Negative
Bartonella quintana					Negative
Bordetella pertussis					Negative
Burkholderia sp.					Negative
Empedobacter brevis					Negative
Enterococcus sp.					Negative
Haemophilus influenzae					Negative
Klebsiella aerogenes					Negative
Pseudomonas aeruginosa					Negative
Salmonella enterica serovar Poona					Negative
Sphingomonas paucimobilis					Negative
Stenotrophomonas maltophilia					Negative
Streptococcus pneumoniae					Negative

^aL. pneumophila comprises three subspecies: subsp. pneumophila, fraseri, and pascullei. 

^bND, not yet deposited. 

Reaction mixtures contained 1 μl of DNA extract, 0.7 μl of LightCycler master mixture (Roche Diagnostics, Lewes, United Kingdom), 4 mM MgCl₂, 3 pmol each of the mip primers, 3 pmol of hybridization probe, SYBR green (Biogene, Cambridge, United Kingdom) at a final concentration of 1:10,000, 0.1 U of uracil–N-glycosylase (Roche Diagnostics), and PCR grade water to a final volume of 7 μl. Of this, 5 μl was added by brief centrifugation to a LightCycler capillary reaction cuvette and amplified in a model LC32 Idaho Technology LightCycler. Reaction conditions were 3 min at 95°C, followed by 50 cycles of 0 s (hold time on reaching temperature) at 95°C, 1 s at 60°C, and 2 s at 72°C. The double-stranded PCR product was measured during the 60°C annealing step by detection of fluorescence associated with the binding of SYBR green dye to the product. Product identity was confirmed by fluorescence resonance energy transfer from SYBR green to the Cy5-labeled hybridization probe (1). The product melt was as follows: 0 s (hold time on reaching temperature) at 95°C, 0 s at 50°C, and 0 s at 95°C. Temperature change rates were 20°C/s, except for the final step, which had a temperature change rate of 0.2°C/s. Reactions were monitored on-line in real time.

The L. pneumophila-specific PCR product had a characteristic melting curve (melting temperature [T_m], 80°C) as monitored by SYBR green dissociation, with a specific melt (T_m, 63°C) of the Cy5-labeled hybridization probe (Fig. 1). Only strains of L. pneumophila generated specific PCR product. No PCR product was produced with other Legionella spp. or any of the other organisms tested (Table 1). Figure 2 shows the results obtained with different dilutions of the known culture of L. pneumophila. Quantification was achieved at a lower limit of 2.5 CFU/LightCycler reaction, equivalent to ca. 1,000 CFU/liter in the original water sample. In contrast, the sensitivity limit of a conventional gel-based multiplex PCR assay using the mip primers and a set of genus-specific 5S ribosomal DNA primers (2) was 25 CFU/reaction (i.e., an order of magnitude less sensitive than the LightCycler assay).

FIG. 1.

Melting curves for L. pneumophila-specific LightCycler-based PCR. Shown are changes in SYBR green (A) and Cy5 (B) fluorescence (−dF/dT) versus temperature. ——, L. pneumophila serogroup 1; -----, negative control.

FIG. 2.

Quantification of L. pneumophila-specific double-stranded PCR product as measured by SYBR green fluorescence. (A) Log fluorescence versus cycle number (a, 1,000 CFU/reaction; b, 100 CFU/reaction; c, 25 CFU/reaction; d, 2.5 CFU/reaction); (B) crossing points (cycle numbers) of log linear correlations with the noise line band, plotted against the logarithmic concentration of the standards.

The prototype LightCycler assay was further assessed with 14 natural water samples and 10 laboratory microcosms, of which 11 water samples and all 10 microcosms were culture positive for L. pneumophila. All 10 microcosms were also positive for L. pneumophila with the LightCycler assay. Six of the 11 culture-positive natural water samples were positive with the LightCycler assay, but of the 5 culture-positive samples that were negative, 3 contained <200 CFU/liter (i.e., below the detection limit of the LightCycler assay). These samples yielded positive assay results after being spiked with the L. pneumophila DNA extract (equivalent to 1,000 CFU) used to prepare the standard curve. The remaining two culture-positive samples appeared to contain PCR inhibitors, as they yielded a negative result even after being spiked. The three culture-negative water samples were also negative with the LightCycler assay but did not contain PCR inhibitors, as they yielded a positive result after being spiked.

In conclusion, the prototype LightCycler assay is a promising quantifiable biprobe method that amplifies a 131-bp region at the 5′ end of the mip gene and appears to be specific for L. pneumophila. The sensitivity limit of 2.5 CFU/reaction (1,000 CFU/liter of starting water) probably represents ca. 25 copies of target DNA, since culture, following concentration by filtration and centrifugation, only detects about 10 to 30% of the bacteria present in the original sample, with the rest being either nonviable or lost during the concentration steps. No false-positive results compared with culture were obtained, but there were two false-negative results, apparently caused by PCR inhibition. The problem of inhibition, particularly that caused by iron compounds (e.g., rust) often present in environmental water samples, may limit the usefulness of PCR-based assays unless improved DNA preparation methods are developed and inhibition controls are routinely included. The sensitivity and specificity of the assay remain to be established with a larger number of varied samples, but the assay may be particularly useful in outbreak situations where a reservoir of infection normally contains >1,000 CFU/liter and there is a requirement to screen significant numbers of samples in as short a time as possible.