Detection of Legionella pneumophila by PCR-ELISA Method in Industrial Cooling Tower Water

INTRODUCTION

Legionella pneumophila (LP), the causative agent of Legionnaires’ disease, was first recognized in 1976 following an epidemic of acute pneumonia in Philadelphia Lisby and Dessau (1994) and Fraser et al. (1977). Since then, a total of 41 Legionella species containing 62 serogroups have been characterized (Stranbach et al., 1989; Atlas, 1999). Twenty-one of Legionella species have been reported as pathogens in humans (Stranbach et al., 1989; Wilkinson et al., 1986; Fang et al., 1989). Legionnaires’ disease is normally acquired by inhalation or aspiration of Legionella from a contaminated environmental source. Several reports have shown a clear association between the presence of LP in hot water systems and the occurrence of legionellosis (Fiume et al., 2005; Hiroshi and Hiroyuki, 1997; Fields and Benson, 2002; Vincent-Houdek et al., 1993). Pneumonia caused by Legionella has a poor prognosis unless it is diagnosed early and treated with specific antibiotics. Eighty-five percent of these infections are caused by LP (Jonas et al., 1995). Cultivation of Legionella organism from appropriate samples represents the definitive method for diagnosis and has a sensitivity of 50 to 90.5% (Lisby and Dessau, 1994; Fraser et al., 1977; Thacker and Robert, 1991). However, colonies become macroscopically visible after 3 to 4 days of culture (Hiroshi and Hiroyuki, 1997; Bej and Mahbubani, 1991; Aslani et al., 1997). More recently, PCR has been used to amplify LP DNA. Amplification products are detected in stained agarose gels or by hybridization with specific oligonucleotide probes. Different DNA sequences have been selected for amplification (Declerck et al., 2006; Hiroshi and Hiroyuki, 1997; Jonas et al., 1995; Reggam and Leitner, 2002; Edelstein et al., 1987). The DNA sequences of the macrophage infectivity potentiator (mip) gene as a determinant of pathogenicity have been used for detection of LP and different Legionella species (Fiume et al., 2005; Reggam and Leitner, 2002; Michio and Astushi, 1993; Bej et al., 1990). Amplification of genes coding for r-RNA is also feasible and published protocols employ the 5S rRNA gene alone or combined with mip gene (Hiroshi and Hiroyuki, 1997; Jonas et al., 1995; Michio and Astushi, 1993). Recently, amplification of the 16S rRNA gene has enabled detection of 1 cfu mL‾¹ in water sample or 10 cfu mL‾¹ in stimulated bronchial fluid (Joly et al., 2006; Jonas et al., 1995; Wellinhause and Cathrin, 2001; Michio and Astushi, 1993). Amplification of the 16S rRNA gene has well defined advantages. In particular, a large number of 16S rRNA sequences are now available and suitable primers can thus be selected for gene amplification (Horng et al., 2006; Lisby and Dessau, 1994; Jonas et al., 1995). Here we describe a simple and generally applicable method for molecular diagnosis of LP using specific 16S rRNA PCR amplification in combination with Enzyme Linke Immunosorbent Assay (ELISA). DNA was purified from CTW samples. A 301 bp sequence of the 16S rRNA gene was amplified using Digoxigenin (Dig) labeled nucleotides. The PCR products were hybridized to a biotinlated specific probe immobilized onto streptavidin coated ELISA plates. The reaction was developed using Peroxidase Labeled anti-Dig antibodies.

MATERIALS AND METHODS

This study was conducted and supported by grant from avesina research institute (Thran, Iran) from July 2003 to May 2004.

Bacterial strains and culture condition: Serogroup 1 strain of L. pneumophila (ATCC 33152) was used as a standard LP strain. To achieve pure colonies of LP, bacteria were grown on enrichment broth and were spread on to αBCYE agar and MWYαBCYE agar (Hiroshi and Hiroyuki, 1997; Jonas et al., 1995; Aslani et al., 1997). Oligonucleotide probe was synthesized and 5� biotinylated by Cybergene, AB (Stockholm, Sweden). Oligonucleotide primers were synthesized by Cinagene (Tehran, Iran). A PCR core kit including PCR buffer, MgCl₂, dNTPs and Taq DNA polymerease were purchased from Roche (Roche Diagnostics, Mannheim, Germany). Culture media were purchased from Himedia (Tehran, Iran); different chemicals were from Sigma, BBL, Merck and Roche.

Handling of CTW samples for culture and identification of LP: CTW samples used in this study were taken mainly from the industrial units around Tehran, south and central Parts of Iran. After treatment with acid reagents (0.2 M KCl/HCl pH 2.2) by mixing equal volume of samples and acid treatment reagent and centrifugation (6000 rpm 30 min), samples were plated on specific (αBCYE agar) and selective [αBCYE agar+Legianella 040 (selective agar containing Glycine, polymixin B, Vanccomycine and Amisomycin)] media. Nine different biochemical tests (mobility, Hippurate hydrolysis, urease, carbohydrate fermentation, gram staining, Sudan Black B staining, gelatin liquefaction, nitrate reduction and catalase) were used to identify colonies primarily (Jonas et al., 1995; Michio and Astushi, 1993).

DNA preparation from CTW samples: The chromosomal DNA extraction was performed according to Nejadmoghaddam et al. (2007) procedure with slight alteration: in that 10 mL on samples were centrifuged for 10 min at 5000 rpm. The pellets were washed twice with PBS and finally resuspended in 120 μL ice-cold SNE lysis buffer (10% Sucrose, 0.1 M NaCl, 0.1 M EDTA. pH = 8, containing 4 mg mL‾¹ Lysozyme) and incubated for 30 min at 0°C and then 30 μL TESS lysis buffer (10 mM Tris/HCl. pH = 7.4, 1 mM EDTA, 100 mM Sodium acetate, 50 μL of 10% SDS) was added. It was incubated for 15 min at 70°C. After addition of 20 μL distilled water, the lysate was incubated for 30 min at 37°C with 10 μL RnaseA (1 μg μL‾¹-Roche). Three microliter of proteinase K (20 mg mL‾¹-Roche) was added to the solution and incubated at 37°C for 30 min. Phenol/chloroform extraction was carried out and upper aqueous phase transferred into a fresh microtube. Then sodium acetate (pH 5.2) equal to 0.1 volume and cooled Isopropanol equal to 2 volume of the transferred fraction were added and mixed gently. Chromosomal DNA was precipitated by centrifugation for 10 min at 14000 rpm (Eppendorf). DNA pellet washed with 70% ethanol and dried for 30 min and resuspended in 50 μL distilled water.

DNA electrophoresis: One microliter of extracted DNA was loaded onto ethidium bromide stained 1% agarose gel in 1χTBE buffer pH 8.0 and the electrophoresis was performed at 100 volts for 1 h. The size of extracted DNA was estimated by comparison with size marker (marker III from Roche) Fig. 1.

Nested PCR amplification of DNA: Oligomers were selected from the published full-length sequence of the LP 16S rRNA gene (NCBI) nucleotide sequence database: (Accession No. M59157). For the first step PCR, the sense oligonucleotide primer pl.2 (5�-AGGGTTGATAGGTTAAGAGC-3�) was located at position 451 to 470 and the antisense primer cp3.2 (5�- CCAACAGCTAGTTGACATCG-3�) was complementary to position 836 to 817.

Fig. 1:

Agarose gel electrophoresis of extracted DNA. Extracted DNA were electrophoretically separated on 1% agarose gel and visualized by ethidium bromide staining. Lane 1, Negative control was from D.D water, Lane 2, positive control (ATTC 33152), Lanes 3,4,5, three CTW samples (Parsoil refinery, Tehran oil refinery respectively), Lane 6 size marker III from Roche, Lane 7, 8, two CTW samples (Arak petrochemical company, Behran oil refinery respectively) and Lanes 9, 10, Non Legionella bacteria (Streptococcus H46A and Chlamydia pneumonia, respectively)

For the second step nested PCR, Lpnest F (5’- GCTGATTAACTGGACGTTACCC-3') sense primer was located at position 469 to 490 and LpnestR (5�-CTTTCGTGCCTCAGTGTCAG-3�) antisense primer was complementary to position 769 to 749. The nested PCR amplified a 301 bp fragment of the 16S-rRNA gene. The 5�-biotimylated 20- mer cp₂ (5�-CAACCAGTATTATCTGACCG-3�), complementary to positions 630 to 649 was used as the probing oligomer. In the first step PCR; DNA samples (1 μL each) were added to microtubes containing 24 μL of PCR mixture. The PCR mixture contained 1U of taq DNA polymerase (Roche), 20 pmol of each primer (p1.2, cp3), 2.5 μL 10X PCR buffer, 1.5 mM MgCl₂, 0.4 mM dNTPs (each). The PCR profile included an initial denaturation at 95°C for 5 min followed by 40 cycle of denaturation at 94°C for 30 sec, annealing at 57°C for 30 sec, extension at 72°C for 90 sec. The last extension step was extended for another 10 min at 72°C. In the second-step PCR samples were diluted 1:2000 and were added (1 μL each) to microtubes containing 24 μL of PCR mixture as in the first PCR but containing 20 pmol of nested PCR primers (LpnestF, LpnestR). The nested PCR was run with the same profile as the first PCR for 40 cycles except that the dNTP mixture contained 0.04 mM Dig-dUTP as well as the same amount of the other dNTPS as above.

Gel electrophoresis: Ten microliter of the PCR product was loaded onto ethidium bromide stained 2% agrose gel in 1χTBE buffer pH 8.0 and the electrophoresis was performed at 100 volts for 45 min. The sizes of DNA fragments were estimated by comparison with size markers (100 bp, Fermentas and marker VIII from Roche).

Enzyme-Linked Immunosorbent Assay (ELISA) for detection of PCR products: Sterptavidin-coated microplates were washed with 100 μL PBS-TWEEN 20. Hundred μL of 20 nmol L‾¹ of biotinylated probe was pipetted into each well. After 1 h incubation at 37°C, the wells were washed three times with PBS-Tween. Next 100 μL hybridization solution (33 μL 10χ SSC, 0.5 μL 10% SDS, 10 μL 20% PEG 1500, 56.5 μL D water) was added to each well. The double stranded DNA products were denatured at 95°C for 5 min and were added to each well as quickly as possible in 2 μL volumes and incubated at 50°Cfor 3 h. The wells were then washed by 3χwashing with PBS-Tween. Two microlitres Horseradish peroxidase conjugated sheep anti Digoxigenin F(ab�)² fragments (150 μmol μL‾¹) were then added to the wells and incubated at 37°C for 1 h in 100 μL conjugation solution (100 μL Tris/HCl 1M pH 8.3, 100 μL NaCl 1M). The unbound conjugate was removed by three washes with 100 μL of BPS-Tween. Then 100 μL of TMB solution were added and the OD was measured in an ELISA reader at 440 nm.

RESULTS

Isolation of LP colonies from CTW samples by bacterial culture: To clarify the morphology of LP colonies, standard LP (ATCC33152) was cultured on Legionella selective supplement agar (Fig. 2A). CTW samples were then cultured on Legionella specific agar plates. Among the different colonies that grew (Fig. 2B), those that morphologically resembled the LP standard colonies on Legionella selective supplement agar were selected and recultured on Legionella selective supplement agar (Fig. 2C). The resulting colonies were biochemically characterized using different tests (Table 1) and proved that all the CTW samples were contaminated with LP.

Specificity of the nested PCR: Specificity of the LP nested PCR was investigated for standard LP (ATCC33152) and isolated colonies from 5 industrial cooling towers as well as 4 other non-LP bacterial strains by running the primary and the nested PCRs.


Fig. 2:	(A) Pure colonies of standard LP (ATCC33152) on Legionella selective supplement agar, (B) Different colonies of Arak petrochemical company grown on Legionella specific agar plate and (C) Pure colonies of Arak petrochemical company on Legionella selective supplement agar

Table 1:	Biochemical characterization of colonies selected on Legionella selective supplement agar

G^¯: Gram Negative, +: Test is Positive, ¯ : Test is Negative

Table 2:	PCR amplification of DNA from standard LP, LP isolated from CTW and non-Lp bacteria

+: PCR is Positive, ¯ : PCR is Negative, ^a: PCR was performed using primers p1.2 and cp3.2 followed by detection of the amplified product by electrophoresis on an ethidium bromide-stained agarose gel, ^b: PCR was first performed using primers p1.2 and cp3.2 followed by nested PCR primers LpnestF and LpnestR. The detection of the final amplified product was performed by electrophoresis on an ethidium bromide stained agarose gel

Table 2 shows, results of these PCRs on isolated colonies of LP and some non-LP strains. While some of the LP that was isolated from CTW did not show any amplification in the primary PCR, they were all proved to be positive in the nested PCR. None of the non-LP bacteria showed amplification by the nested PCR. However, the SRB bacteria was the only non-LP that showed amplification in the primary PCR which was not confirmed by the nested PCR (Table 2).

Several PCR product sizes resulted from the first-step PCR in most of the samples (Fig. 3). However, in nested PCR the specific band (301 bp) was detected only in standard LP as well as in our isolated LP colonies and none of the non-LP bacteria produced such products (Fig. 4).

Specific LP detection in CTW by PCR-ELISA: The 16Sr-RNA based PCR amplification products were detected after hybridization with a 5'end-biotinylated oligomer probe by PCR-ELISA. The result of ELISA where equal amounts of nested PCR products (4 ng) were used. All CTW samples as well as the standard LP strain were positive in ELISA whereas all non-LP bacteria were negative (Table 3).

DISCUSSION

Water supply and CTW are the most common sources of LP contamination. Recently several approaches have been devised to detect these bacteria from water samples. Radioactively labeled nucleotide is hazardous and also requires the presence of relatively large numbers of cells in samples. Cultivation takes a long time (7-14 days) to respond (Jonas et al., 1995; Aslani, 1997; Maiwald et al., 1994). PCR, therefore, represents the method of choice and different target sequences have been proposed for amplification (Declerck et al., 2006). The gene encoding macrophage infectivity potentiator of L. pneumophila (mip) and the mip like gene of other Legionella sp. and 5S rRNA gene have been used for PCR amplification and LP detection (Fiume et al., 2005; Jonas et al., 1995; Raggam and Leitner, 2002; Edelstein et al., 1987; Wellinhause and Cathrin, 2001; Michio et al., 1993).

Fig. 3:

Agarose gel electrophoresis of the LP firs-step PCR products including along with water samples from cooling tower and 4 other bacteria. PCR products were separated on 2%-agarose gels and visualized by ethidium bromide staining. Lane 1: Negative control was from DD water. Lanes 2 and 3: Two cooling tower samples (Arak petrochemical company ad Bandar Imam Petrochemical Company, respectively), Lane 4: size marker 8 (Roche), Lanes 5 and 6: Two other cooling tower samples (Behran oil refinery and Pars oil Company respectively) and Lane 7 positive control (standard LP). Lanes 8, 9, 10 and 11: include 4 non-LP bacteria (Streptococcus HubA, Chlamyelia pneumonia SRB bacteria, Entrobacteriacea, respectively)

Selection of 16S rRNA gene sequences, on the other hand, harbors many advantages (Joly et al., 2006; Jonas et al., 1995; Raggam and Leitner, 2002; Pasculle et al., 1989; Michio et al., 1993; Wellinhause et al., 2001). Large databases of 16S rRNA gene are available, facilitating selection of appropriate sequences and it has been used as a fundamental molecular marker in bacterial taxonomy. Large amplification products are generated that can easily be identified by gel electrophoresis. There are potential obstacles in PCR like presence of inhibitors, low levels of microbial contamination and the efficiency of the primers. To clarify the negative PCR resulted in amplification of CTW samples; we designed a nested PCR, which clearly showed positive amplification of a specific 301 bp PCR product. Such a strategy has also been used by others (Fiume et al., 2005; Hiroshi and Hiroyuki, 1997; Jonas et al., 1995).

Fig. 4:

Agarose gel electrophoresis of the LP specific nested PCR products of different samples as follows. PCR products were separated on % 2-agarose gels and visualized by ethidium bromide staining. Lane 1: Negative control was from DD water, Lane 2: Positive control (standard LP, ATCC33152), Lanes 3 and 4: 2 cooling tower water samples (Arak Petrochemical Company and Bandar Imam Petrochemical Company, respectively), Lane 5 size marker 100 bp (Fermentas) (V. Graiciuno, LITHUANTA), Lanes 6,7,8 and 3 cooling tower samples (Behran oil refinery, Pars oil refinery and Tehran oil refinery, respectively). Lanes 9, 10,11 and 12: 4 non-LP bacteria (Streptococcus H4bA, Chlamydia pneumonia SRB bacteria and Entrobacteriacea culture, respectively)

Table 3:	Detection of LP by PCR ELISA

^a: OD readings were performed at 440 nm

To verify the specificity of the amplified product, some researchers have used radioactive probes (Jonas et al., 1995; Bej et al., 1990; Michio et al., 1993). However, nonradioactive techniques are often preferable and in this regard we chose to use biotin-avidin system to set up a specific probe-based ELISA technique for specific detection of the amplified product. All CTW samples were found to have LP contamination by this method. The presences of PCR inhibitors have also been documented (Horng et al., 2006; Hiroshi and Hiroyuki, 1997). Such inhibitors may even be concentrated during DNA extraction, which may contribute to lowering of the PCR efficiency. The negative PCR results from CTW samples shown in Table 1 may have resulted from such inhibitors. The second PCR (nested PCR) run on 2000 fold diluted primary PCR products may have effectively diluted the potential inhibitors yielding clear specific bands of 301 bp and providing a platform for increasing PCR efficiency. An important issue when PCR is used for detection of microorganisms is the potential power of PCR in amplifying DNA from dead bodies. (Hiroshi et al., 1997; Maiwald et al., 1994). To rule out such a possibility, all the CTW samples in our study were subjected to Lp specific and selective bacterial culture and all were found to contain live LP.

In summary, we have developed a PCR-ELISA technique to effectively amplify LP 16S rRNA from CTW samples. Present findings also prove the presence of Lp contamination in all the five oil industry plants raising the necessity of especial health care for their employees.

ACKNOWLEDGMENTS

We thank Dr. M. Akhondi and Dr. M.R. Sadeghi for their invaluable support. We are grateful to Roya Ghods and Pouneh Dokouhki for excellent technical assistance. Special thanks to Mr. A.A. Bayat for his kind cooperation in the experiments. This work was supported by grants from Avesina Research Institute.