Diagnostic Microbiology and Infectious Disease xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Diagnostic Microbiology and Infectious Disease journal homepage: www.elsevier.com/locate/diagmicrobio

Overestimation of the Legionella spp. load in environmental samples by quantitative real-time PCR: pretreatment with propidium monoazide as a tool for the assessment of an association between Legionella concentration and sanitary risk Savina Ditommaso a,⁎, Elisa Ricciardi a, Monica Giacomuzzi a, Susan R. Arauco Rivera a, Adriano Ceccarelli b, Carla M. Zotti a a b

Department of Public Health and Pediatrics, University of Turin, Italy Department of Clinical and Biological Sciences, University of Turin, Italy

a r t i c l e

i n f o

Article history: Received 30 May 2014 Received in revised form 8 September 2014 Accepted 8 September 2014 Available online xxxx Keywords: Legionella spp PCR (polymerase chain reaction) Colony count Sensitivity and specificity Water microbiology Propidium monoazide

a b s t r a c t Quantitative polymerase chain reaction (qPCR) offers rapid, sensitive, and specific detection of Legionella in environmental water samples. In this study, qPCR and qPCR combined with propidium monoazide (PMA-qPCR) were both applied to hot-water system samples and compared to traditional culture techniques. In addition, we evaluated the ability of PMA-qPCR to monitor the efficacy of different disinfection strategies. Comparison between the quantification obtained by culture and by qPCR or PMA-qPCR on environmental water samples confirms that the concentration of Legionella estimated by GU/L is generally higher than that estimated in CFU/L. Our results on 57 hot-water-system samples collected from 3 different sites show that: i) qPCR results were on average 178-fold higher than the culture results (Δ log10 = 2.25), ii) PMA-qPCR results were on average 27-fold higher than the culture results (Δ log10 = 1.43), iii) propidium monoazide–induced signal reduction in qPCR were nearly 10-fold (Δ log10 = 0.95), and that iv) different degrees of correlations between the 3 methods might be explained by different matrix properties, but also by different disinfection methods affecting cultivability of Legionella. In our study, we calculated the logarithmic differences between the results obtained by PMAqPCR and those obtained by culture, and we suggested an algorithm for the interpretation of PMA-qPCR results for the routine monitoring of healthcare water systems using a commercial qPCR system (iQ-check real-time PCR kit; Bio-Rad, Marnes-la-Coquette, France). © 2014 Elsevier Inc. All rights reserved.

1. Introduction Contamination of hospital water with legionellae is a cause of nosocomial legionellosis, and rapid identification of the source of infection is essential to prevent further cases. Culture is essential for identifying and typing Legionella strains. However, culturing Legionella requires long incubation times (up to 10 days), and additional time can be needed if additional confirmation procedures are applied. Moreover, various factors can influence method accuracy: differences in the membrane, such as pore size, batches, fragility, crinkling, and electrostatic interactions, and differences in the procedures for washing the organisms from the membrane, such as using a shaker/vortex, ultrasound, or finger and thumb scraping. Quantitative polymerase chain reaction (qPCR) is an alternative tool that offers rapid, sensitive, and specific detection of Legionella bacteria in environmental water samples (Ballard et al., 2000; Behets et al., 2007;

⁎ Corresponding author. Tel.: +39-0116705841; fax: +39-0116705881. E-mail address: [email protected] (S. Ditommaso).

Dusserre et al., 2008; Joly et al., 2006; Wellinghausen et al., 2001; Yaradou et al., 2007). Currently, only culture methods (International Standards Organisation, 1998) and qPCR (Association Française de Normalisation, 2010; International Standards Organisation, 2012) are normalised. However, the major disadvantage of qPCR lies in its inability to distinguish between viable and membrane-compromised cells or extracellular DNA persisting in the environment (Josephson et al., 1993; Masters et al., 1994). The development of more rapid, cultureindependent methods capable of discriminating between live and dead cells is of major interest for measuring the risk of Legionella infection and preventing legionellosis. One of the most successful recent approaches used to detect viable cells utilised the technology based on sample treatment with the photoactivatable, cell membrane impermeable nucleic acid intercalating dyes, ethidium monoazide (EMA), or propidium monoazide (PMA), followed by light exposure prior to the extraction of DNA and amplification. Light activation of DNA-bound dye molecules results in irreversible DNA modification and subsequent inhibition of its amplification. The nucleic acid-binding dyes, EMA or PMA, used in combination with qPCR, are an attractive alternative for selectively detecting and

http://dx.doi.org/10.1016/j.diagmicrobio.2014.09.010 0732-8893/© 2014 Elsevier Inc. All rights reserved.

Please cite this article as: Ditommaso S, et al, Overestimation of the Legionella spp. load in environmental samples by quantitative real-time PCR: pretreatment with propidium mon..., Diagn Microbiol Infect Dis (2014), http://dx.doi.org/10.1016/j.diagmicrobio.2014.09.010

2

S. Ditommaso et al. / Diagnostic Microbiology and Infectious Disease xxx (2014) xxx–xxx

enumerating viable bacteria (Chang et al., 2010; Delgado-Viscogliosi et al., 2009; Nocker et al., 2007). This technique is not limited to use in bacteria but has also successfully been applied to fungi, protozoa, and viruses. Apart from samplespecific challenges, the 2 dyes both seem to have specific advantages and disadvantages. Whereas the greatest concern with EMA lies in its ability to penetrate bacterial cells with intact membranes causing underestimation of live cell population, the greatest concern with PMA is the generation of false-positive signals, due to incomplete signal suppression. Several comparative studies confirm that PMA outperforms EMA in selective removal of dead cells when combined with real-time PCR (Cawthorn and Witthuhn, 2008; Nocker et al., 2006; Pan and Breidt, 2007). For the successful application of qPCR, a few factors that can influence the outcome of the resulting data must be considered. These factors include, amongst others, the choice of the dye, its concentration, the contact mode (cells suspension or cells impinged on filters), the incubation conditions, the light source, the presence of a high number of dead cells, the presence of high levels of suspended solids or biomass in the analysed samples, the salt concentration in the reaction mix, the pH of the reaction mix, the length of the target gene, and the sequence of the target gene (Fittipaldi et al., 2012). In a previous study (submitted), we compared the results obtained by conventional qPCR and culture methods on spiked samples inoculated with Legionella pneumophila serogroup 1 (ATCC 33152), and we analysed the potential for the PMA-qPCR method to selectively quantify viable Legionella cells. In this study, we addressed the question of how efficiently signals would be suppressed by PMA in the presence of an environmental sample. For this purpose, the 2 methods (qPCR and PMA-qPCR) were both applied to hot-water system samples and compared to the culture technique. In addition, we evaluated the ability of PMA-qPCR to monitor the efficacy of different disinfection strategies. 2. Methods 2.1. Sample collection From April to June 2013, 57 hot water samples were collected from the in-building distribution systems of 3 healthcare facilities in the city of Turin. The healthcare facilities included in this study consisted of acute care hospitals that conducted environmental monitoring programmes for Legionella detection. These facilities were selected because they adopt different strategies for the control of Legionella contamination: facility A implemented continuous chlorination to ensure 2–10 ppm of free chlorine at the distal sites, facility B implemented continuous chlorination to ensure 1–2 ppm of free chlorine at the distal sites, and facility C implemented quarterly disinfection of the water supply by thermal shock. In facility C, we collected water samples 30 days after thermal shock. Each sample was collected in a sterile, 3-litre plastic bottle and was divided into 3 equal parts for culture, qPCR, and PMA-qPCR. The physical and chemical parameters of the water distributed in the facilities are almost identical for all water samples and in compliance with legal requirements, except for chlorine and temperature that are facility dependent; we decided to measure only chlorine and temperature parameters. 2.2. Quantification by the culture method Analyses for the detection and quantification of Legionella were carried out in accordance with the modified ISO 11731 method (Ditommaso et al., 2011) that recommended the use of different media for routine water tests in hospitals: BCYE (Buffered Charcol Yeast Extract with a-ketoglutarate, L-cysteine and ferric pyrophosphate) and MWY (modified Wadowsky and Yee medium supplemented with glycine, polymyxin B, vancomycin, anysomicin, bromthymol blue and bromcresol purple). BCYE gives a high recovery rate of positive

samples, a much greater yield of Legionella spp. than MWY, and is necessary for the detection of non–L. pneumophila species, which grew poorly on selective media. Selective media (MWY) were necessary for the recovery of Legionella spp. when the nonselective medium (BCYE) was difficult to interpret because contaminating background flora. The water samples were concentrated 100-fold by filtration through a 0.2-μm polycarbonate filter (Millipore, Billerica, MA, USA). The filter membrane was aseptically placed in one of the bottom corners inside the stomacher bag and rubbed with the finger and thumb of one hand for 1 min with 10 mL Page's solution (pH 6.8) to detach the bacteria. A 0.2-mL volume of the concentrated sample was spread on duplicate plates of BCYE agar and MWY agar (Oxoid, Wesel, Germany). Under these experimental conditions, the detection limit (LOD) was 50 CFU/L. 2.3. Quantification by qPCR and PMA-qPCR The second and third aliquots were filtered through a 0.45-μm polycarbonate filter (Millipore, Billerica, MA, USA) according to the manufacturer's instructions (Aquadien™; Bio-Rad, Marnes-la-Coquette, France). The first filter was directly added to the lysis solution for DNA extraction, whilst the second was first overlaid with 500 μl of PMA (50 μM) in 90 mm Petri dishes and incubated in the dark for 10 min followed by a 10 min exposure to a 500 W light on ice at a distance of 20 cm from the light source. After irradiation, the filter was added to the lysis solution for DNA extraction. In order to eliminate the bacteria resuspension step (which could generate bacterial loss), the extraction of DNA was performed directly from membrane filters. The experimental conditions were optimised in our previous study (submitted paper). Extracted genomic DNA was analysed for the presence of amplifiable sequences using qPCR. Analysis was performed by the iQ-Check™ Quanti Legionella spp. kit according to the manufacturer's instructions (Bio-Rad, Marnes-laCoquette, France). The iQ-Check™ Quanti Legionella spp. kit is NF VALIDATION certified (certificate numbers BRD07/15-12/15) and contains reagents to amplify and quantify a 100-bp fragment from the 5S rRNA gene of Legionella spp. This method allows the quantification of Legionella in water samples in less than 3 hours following the water sample filtration and DNA extraction steps. The LOD of this qPCR method is 5 GU per well; performing analysis in duplicate kits allows us to have a total method detection limit equal to 80 GU/L. The quantification limit was 10 GU/ 5 μL corresponding to 608 GU/L. The Amplification mix contains the internal control (IC), a linear plasmid, which is added to each PCR reaction and which should be amplified in all conditions. This control monitors the inhibitory effects that may take place in the reaction mix. The Legionella target and the IC are always amplified in the same PCR well. 3. Statistical analysis All data generated by qPCR were analysed by the Opticon Monitor Analysis Software version 3.4 (Bio-Rad). The positive predictive value (PPV) and negative predictive value (NPV) of the qPCR and PMA-qPCR techniques were calculated. PPV corresponds to the ratio between the number of culture positive samples and the number of positive samples Table 1 Comparison of the culture and qPCR results on 57 water samples from 3 hospital water supplies in the Piedmont region. Samples qPCR(+) qPCR(−) qPCR(+) qPCR(−) Total

Culture (+) Culture (−) Culture (−) Culture (+)

Number

%

26 19 12 0 57

46 33 21 0 100

Observed concordance = 79%; Cohen κ = 0.591.

Please cite this article as: Ditommaso S, et al, Overestimation of the Legionella spp. load in environmental samples by quantitative real-time PCR: pretreatment with propidium mon..., Diagn Microbiol Infect Dis (2014), http://dx.doi.org/10.1016/j.diagmicrobio.2014.09.010

S. Ditommaso et al. / Diagnostic Microbiology and Infectious Disease xxx (2014) xxx–xxx Table 2 Comparison of the culture and PMA-qPCR results on 57 water samples from 3 hospital water supplies in the Piedmont region. Samples PMA-qPCR (+) PMA-qPCR (−) PMA-qPCR (+) PMA-qPCR (−) Total

Culture (+) Culture (−) Culture (−) Culture (+)

Table 3 Comparison of the qPCR and PMA-qPCR results on 57 water samples from 3 hospital water supplies in the Piedmont region.

Number

%

Samples

24 21 10 2 57

42 37 18 3 100

PMA-qPCR (+) PMA-qPCR (−) PMA-qPCR (+) PMA-qPCR (−) Total

Observed concordance = 79%; Cohen κ = 0.586.

3

qPCR (+) qPCR (−) qPCR (−) qPCR (+)

Number

%

34 19 0 4 57

60 33 0 7 100

Observed concordance = 93%; Cohen κ = 0.85.

by both methods. NPV corresponds to the ratio between the number of culture negative samples and the number of negative samples by both methods. Agreement between the 2 methods was determined by comparing the culture versus the qPCR or PMA-qPCR on 2-by-2 contingency tables. Culture was taken as the reference method. To quantify the difference between methods, Δ log10 values were calculated. The statistical correlation was calculated with the Spearman rank correlation test. To analyse the results in relation to water temperature, we decided to disaggregate the results into 3 different temperature threshold: lower (b25 °C) medium (25–45 °C), and high (N45 °C). The bacteria are dormant below 25 °C (lower temperatures supported survival without loss of cultivability), multiply when temperatures are between 25 and 45 °C and nutrients are available. The cultivability declined with increasing temperature, although metabolic activity was observed at temperatures of up to 45 °C. Significant differences were evaluated by Student's t-test analysis and analysis of variance (ANOVA). 4. Results The performance of PMA-qPCR was investigated on 57 environmental hot-water samples and compared to the culture and qPCR methods. Twenty-one samples were taken from facility A; 15 samples, from facility B; and 21 samples, from facility C. During the analysis of environmental samples, we have never had any problems of inhibition of the IC amplification and for all samples was always satisfied the established range for the kit (28 b cycle threshold b 42). 4.1. Comparison between the culture and qPCR results When considering the culture method as the reference method for the detection of Legionella spp., sensitivity, specificity, PPV, and NPV by qPCR method reach 100%, 61%, 68%, and 100% respectively (Table 1). Legionella was detected in 67% and 46% of the samples with the qPCR test and the culture method, respectively; 46% tested positive and 33% tested negative with both methods; 21% that tested positive with the qPCR test were negative when tested with the culture method; and no samples that tested positive with the culture method were

negative with the qPCR test. Agreement between the 2 methods was 79%. Calculation of Cohen's kappa coefficient showed moderate concordance between the 2 methods (κ = 0.591). The comparison between the quantification obtained by culture and by qPCR on the 26 samples that were positive with both methods shows that the concentration of Legionella spp. estimated by GU/L is on average 178-fold higher than that estimated by CFU/L (Δ log10 = 2.25). 4.2. Comparison between the culture and PMA-qPCR results When considering the culture method as the reference method for the detection of Legionella spp., sensitivity, specificity, PPV, and NPV by the PMA-qPCR method reach 92%, 67%, 70%, and 91%, respectively. (Table 2) Legionella was detected in 60% of the samples by the PMA-qPCR test; 42% tested positive and 37% tested negative with both methods; 18% that tested positive with the PMAqPCR test resulted negative with the culture method; and 3% tested positive with the culture method resulted negative with the PMA-qPCR test. Agreement between the two methods was 79%. Calculation of Cohen's kappa coefficient showed moderate concordance between the 2 methods (κ = 0.586). The comparison between the quantification obtained by culture and by PMA-qPCR on 24 samples that tested positive with both methods shows that the concentration of Legionella spp. estimated by GU/L is on average 27-fold higher than that estimated in CFU/L (Δ log10 = 1.43). Overall, there was a weak correlation between the bacterial levels detected by culture and PMA-qPCR methods (r = 0.642, P = 0.001) (Fig. 1). 4.3. Effect of PMA treatment Legionella was detected in 67% of the samples by the qPCR test and in 60% of the samples by the PMA-qPCR test (Table 3). After treatment with PMA, 4 positive samples became negative (Δ log10 = 2.62). Overall, the comparison between the 34 samples that tested positive by both the qPCR (mean log = 5.16) and PMA-qPCR (mean log = 4.20) methods showed that the Legionella count dropped nearly 1 log (Δ log10 = 0.95) following the treatment with PMA. 4.4. Analysis of the results in relation to water chlorination Examination of the data indicated the greatest concordance between the results from the culture and PMA-qPCR methods when the water supply (facilities A and B) was adequately maintained with high doses of chlorine (from 0.5 ppm to 10 ppm). A low concentration of chlorine (0.1 ppm) did not affect the PMA-qPCR results: 13 out of the 17 samples

7,0

log PMA-qPCR by GU/L

6,0

5,0

4,0

3,0

2,0

1,0

0,0 0,0

1,0

2,0

3,0

4,0

5,0

log culture by CFU/L Fig. 1. Comparison of the Legionella culture results with the PMA-qPCR results for water samples analysed from 3 Turin hospitals. The dotted lines represent the lowest limit for detection (i.e., 50 CFU/L for culture and 80 GU/L for PMA-qPCR); r = 0.642 calculated with the Spearman rank correlation test.

Please cite this article as: Ditommaso S, et al, Overestimation of the Legionella spp. load in environmental samples by quantitative real-time PCR: pretreatment with propidium mon..., Diagn Microbiol Infect Dis (2014), http://dx.doi.org/10.1016/j.diagmicrobio.2014.09.010

4

S. Ditommaso et al. / Diagnostic Microbiology and Infectious Disease xxx (2014) xxx–xxx 6,0

12

5,0

10

4,0

8

3,0

6

2,0

4

1,0

2

0,0

Chlorine (ppm)

Log concentration

CFU/L GU/L CHLORINE

0 1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18 - 36

samples Fig. 2. Results obtained by culture and PMA-qPCR related to water chlorine concentration in facility A and B. were PMA-qPCR positive, whilst in 10 out of the 17 samples, Legionella lost its culturability (Fig. 2).

molecular methods that did not agree (culture negative and qPCR/PMA-qPCR positive). Disinfection by heating, when applied quarterly, did not decrease the Legionella count. We show 90% positive results by all 3 testing methods (Fig. 5).

4.5. Analysis of the results in relation to water temperature Data from 26 samples without chlorine treatment were analysed to investigate the influence of temperature on the results. The mean log differences between the results for PMA-qPCR as GU/L and culture as CFU/L at different temperatures ranges are shown in Table 4. Examination of the data indicated the greatest discrepancies between results from PMA-qPCR and culture occurred more frequently in samples taken from systems with a high (N45 °C, P = 0.0002) or low (b25 °C, P = 0.013) water temperature. 4.6. Results for each facility. 4.6.1. Facility A The contamination rate of the water in facility A was 14% (3/21) by culture and 19% (4/21) by PMA-qPCR. By all methods, 14/21 (66%) samples were negative (samples nos. 8–21), 2 samples tested positive by all methods (samples nos. 1 and 2), 3 samples became negative after PMA treatment (samples nos. 3, 6, and 7), and 2 samples were culture negative and qPCR/PMA-qPCR positive (samples nos. 4 and 5). For the 14 samples that tested negative by all methods, the chlorine concentration ranged from 1.5 to 10 ppm (Fig. 3). 4.6.2. Facility B The contamination rate of water in the facility B was 26% (4/15) by culture and 60% (9/15) by PMA-qPCR. By all methods, 5/15 (33%) samples were negative (samples nos. 7, 9, 13, 14, and 15), 3/15 samples tested positive (samples nod. 3, 4, and 10), 1 sample becomes negative after PMA treatment (sample no. 5), and 6 samples were culture negative and qPCR/PMA-qPCR positive (samples nos. 1, 2, 6, 8, 11, and 12). For 3 out of 5 samples negative by all methods, the chlorine concentration was between 0.5 and 1 ppm (Fig. 4). 4.6.3. Facility C The contamination rate of water in facility C was 79% (19/21) by culture and 100% (21/21) by PMA-qPCR. Two samples (nos. 11 and 18) had results by the culture and Table 4 Comparison of mean log differences (PMA-qPCR − culture) results for water samples at different temperature ranges. Temperature range (°C)

Mean log PMA-qPCR

Mean log culture

Δ log10 = PMAqPCR − culture

b25 25–45 N45

4.27 2.97 4.31

2.00 3.44 2.80

2.28a −0.47 1.50b

Results significantly different (t test) to values for lower temperature and high temperature (ANOVA, P = 0.0039). a P = 0.013. b P = 0.0002.

5. Discussion Legionella is ubiquitously found in aquatic habitats. From its natural reservoirs (e.g., lakes, rivers, thermal springs), it can enter and colonise man-made water supply systems (e.g., water distribution plants, tanks, cooling towers). Although no direct correlation has been demonstrated between Legionella load and risk of legionellosis, determination of the microbial load underlies the analysis of water samples taken from distribution systems. National (Conferenza stato regioni, 2000) and international guidelines for Legionella prevention and control (Health and Safety Commission, 2000; Ministère de la Santé et des Solidarietès, 2005; WHO, 2007) set the risk and intervention levels in water distribution systems based on the Legionella load detected in the samples. The most commonly used method for environmental surveillance of Legionella is the standard culture technique. Culture is essential for identifying and typing Legionella strains. However, culture method is complex with many steps including concentration, resuspension, and inoculation onto selective medium, and therefore, there may be considerable losses of legionellae during these processes. In addiction, their growth can sometimes be inhibited by the presence of the other organisms, and some species do not grow or grow only weakly at 36 °C, the temperature commonly used to isolate L. pneumophila and the other pathogenic species. The isolation medium is also not suitable for some of the other species that only grow poorly if at all on the BCYE medium, particularly when (as in MWY) selective agents are present (Ditommaso et al., 2011). To overcome the limitation of Legionella detection by culture, we compared conventional culture with a ready-to-use real-time qPCR system dedicated to routine quantification of Legionella spp. in water samples. We utilised Bio-Rad kits (AquadienTM and iQ-CheckTM), a PCR quantitative assay targeting the 5S rRNA of Legionella spp. To discriminate between the DNA of live and nonviable Legionella, we implemented a pre-treatment with PMA in conjunction with qPCR technique. Our results on 57 hot-water-system samples collected from 3 different sites show that i) qPCR results were on average 178-fold higher than the culture results (Δ log10 = 2.25), ii) PMA-qPCR results were on average 27-fold higher than the culture results

Please cite this article as: Ditommaso S, et al, Overestimation of the Legionella spp. load in environmental samples by quantitative real-time PCR: pretreatment with propidium mon..., Diagn Microbiol Infect Dis (2014), http://dx.doi.org/10.1016/j.diagmicrobio.2014.09.010

S. Ditommaso et al. / Diagnostic Microbiology and Infectious Disease xxx (2014) xxx–xxx

5

6,0 culture qPCR PMA-qPCR

Log Legionella

5,0

4,0

3,0

2,0

1,0

0,0 1

2

3

4

5

6

7

8-21

samples Fig. 3. Facility A: Quantification of Legionella by culture, qPCR, and PMA-qPCR in water supply with high chlorine concentration (0.5–10 ppm).

(Δ log10 = 1.43), and iii) PMA-induced signal reduction in qPCR were nearly 10-fold (Δ log10 = 0.95). Yáñez et al. (2011) also investigate the usefulness of PMA-qPCR for the detection of viable L. pneumophila in 40 water environmental samples using culture isolation, PMA-qPCR, and conventional qPCR. He reported that PMA-qPCR resulted in lower numbers of GU than qPCR with values being in better agreement with the Legionella numbers determined by culture isolation (Δ log10 = 0.62, equivalent to 4-fold). Nevertheless, he observed the differences between PMA-qPCR and culture isolation between samples. As also Yáñez et al. (2011) suggest, PMA-induced signal reduction in qPCR might indicate the presence of the membrane-compromised cells, whereas the difference between culture isolation and PMA-qPCR might indicate the presence of intact nonculturable cells. This finding is not surprising because the rates of recovery in culture are usually noticeably lower than 100% (Villari et al., 1998) due to fastidious growth requirements, overgrowth by other bacteria, and damage/loss to the legionellae by the concentration steps. In contrast, PCR detection methods also include nonculturable legionellae, Legionella living amoeba (Bates et al., 2000; Koide et al., 1993; Ng et al., 1997; Yamamoto et al., 1993), and Legionella doublets or chains, which are

counted as only 1 CFU by the culture method but quantified as individual cells by qPCR. Moreover, it is suggested the hypothesis (Delgado-Viscogliosi et al., 2009) that several Legionella cells may be aggregated and may give rise to a single colony, minimising the result. Lastly, some Legionella strains (e.g., Legionella longbeachae) could have more 5S gene copies than the references strain used for the PCR standard curve (Cazalet et al., 2010). Joly et al. (2006) observed variations dependent on the version of the PCR instrument running the same quantitative PCR technique for Legionella DNA quantification (in terms of extraction and the amplification target) in 2 different laboratories. Additional variability could be due to differences in the PCR tools and in the choice of the Legionella target gene(s) (Koide et al., 1993; Liu et al., 2003; Yamamoto et al., 1993). The differences observed amongst the 3 techniques may be attributed to both the sensitivity difference between the 2 methods and to the diverse physiological states in which Legionella cells can be found in the environment. These 3 analytical techniques are based on different cell properties (culturability, intact cells with amplifiable DNA, or damaged cells with amplifiable DNA). Nevertheless, the large amounts of legionellae detected by qPCR may also represent nonviable cells or only Legionella DNA not infectious to humans. Therefore, the high

6,0 culture qPCR PMA-qPCR

log Legionella

5,0

4,0

3,0

2,0

1,0

0,0 1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

samples Fig. 4. Facility B: Quantification of Legionella by culture, qPCR, and PMA-qPCR in water supply with low chlorine concentration (0.1 ppm).

Please cite this article as: Ditommaso S, et al, Overestimation of the Legionella spp. load in environmental samples by quantitative real-time PCR: pretreatment with propidium mon..., Diagn Microbiol Infect Dis (2014), http://dx.doi.org/10.1016/j.diagmicrobio.2014.09.010

6

S. Ditommaso et al. / Diagnostic Microbiology and Infectious Disease xxx (2014) xxx–xxx 8,0 culture qPCR PMA-qPCR

7,0

log Legionella

6,0

5,0

4,0

3,0

2,0

1,0

0,0 1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

samples Fig. 5. Facility C: Quantification of Legionella by culture, qPCR and PMA-qPCR in water supply treated by thermal shock.

qPCR signals should be critically interpreted and do not necessarily represent a health risk for exposed persons. Lee et al. (2011) reported that 1) when comparing qPCR with culture results for Legionella spp., the mean log difference was 1.05 (equivalent to 10-fold higher) for 275 domestic hot and cold water system samples; 2) when comparing the qPCR and culture results for L. pneumophila, the mean log difference was 0.61 (equivalent to 4-fold higher). On the basis of these results and taking into account alert levels suggested by the European Guidelines (1000 CFU/L) (Joseph et al., 2011), the author suggested algorithms for the interpretation of qPCR results for the routine monitoring of healthcare water systems, including an alert level of N104 GU/L for Legionella spp. and N4 × 103 GU/L for L. pneumophila. Unlike Lee et al. (2011), our qPCR results were on average 178-fold higher than the culture results (Δ log10 = 2.25). We attribute this difference (10 fold versus 178) to the different physiological states in which Legionella was in our samples, collected in facilities that implemented disinfection methods (chlorine or thermal shock) affecting the viability of the cells. Also, Lee et al. (2011) in the samples collected at temperature N45 °C (149/275) observed a high delta-log (ranging from 1.28 to 2.13) between the qPCR and culture (equivalent to 19–135 fold); he attributed this phenomenon to injured or killed cells. In our study, also, the lower temperatures could be a factor affecting the viability of the cells; as shown in Table 4, the mean log differences between the results for PMA-qPCR and culture were significantly higher both at lower (P = 0.013) and at higher (P = 0.0002) temperatures. In our previous work (submitted paper), we compared the quantification of live cells in 60 spiked samples by culture and by qPCR, and we demonstrated a mean log difference of 1.45 (therefore, 28 times greater for the qPCR). In this paper using PMA-qPCR on environmental samples, we obtained a mean log difference of 1.43 (equivalent to 27-fold higher), and taking in account alert levels suggested by Italian national guidelines (1 × 10 4 CFU/L), we calculated an alert level of N2.7 × 10 5 GU/L for Legionella spp. Moreover, in relationship to different disinfection method, we observed different degrees of concordance amongst the 3 methods: as expected, increased chlorine doses resulted in decreased culture counts. When chlorination was carried out using a concentration of free chlorine ranging from 0.5 to 10 ppm (facility A), the damages induced by chlorine seem to affect the nucleic acid integrity of Legionella cells and their cultivability and amplification with either molecular method;

66% (8/21) of these samples were negative by all methods. When chlorination was carried out using a concentration of free chlorine ranging from 0.1 to 0.5 ppm (facility B), the enumeration obtained using culture method dropped, and we obtain discordant results in 40% of samples (6/15 negative by culture and positive by qPCR and PMA-qPCR). Disinfection by heating, when applied quarterly (facility C), did not affect Legionella counts, and we showed that 90% of samples were positive by all 3 methods. In conclusion, qPCR is much faster and more sensitive than conventional culture for L. pneumophila detection (Ballard et al., 2000; Behets et al., 2007; Declerck et al., 2006; Levi et al., 2003; Wellinghausen et al., 2001), but the qPCR results are usually higher than those from the culture method, making it difficult to assess the real health risk (Joly et al., 2006; Wellinghausen et al., 2001; Yáñez et al., 2005). Here, we confirm that PMA-qPCR reduces the qPCR signal in samples containing dead cells (mean log difference = 0.95) making this method more suitable for the monitoring of environmental samples, especially after decontamination treatments, which can generate membranedamaged microorganisms. Several commercial real-time kits designed for detection and quantification of Legionella spp. in water by the real-time PCR are currently available. These kits include the iQ-check real-time PCR kit (Bio-Rad, France), the Aqua Screen Lp-qDual kit (Minerva Biolabs, Berlin, Germany), GeneDisc Legionella kit (GeneSystems, Bruz, France), and the Mericon Quant Legionella spp Kit (Qiagen, Hilden, Germany). However, none of these kits provide a solution for discriminating between viable and nonviable Legionella bacteria. There are many publications (Cawthorn and Witthuhn, 2008; Nocker et al., 2006; Pan and Breidt, 2007) that address this issue for various indicator organisms or environmental pathogens including for Legionella. The use of photoactivatable cell membrane impermeable nucleic acid intercalating dyes, such as EMA or PMA, followed by light exposure prior to DNA extraction and amplification, has been one of the most successful approaches to detect viable cells but, at present, does not exist algorithm to interpret the data on field. In our study, we calculated the logarithmic differences between the culture and PMA-qPCR (mean log difference = 1.43) and suggested an algorithm for the interpretation of PMA-qPCR results in the routine monitoring of healthcare water systems, using a specific qPCR system (iQ-check real-time PCR kit; Bio-Rad).

Please cite this article as: Ditommaso S, et al, Overestimation of the Legionella spp. load in environmental samples by quantitative real-time PCR: pretreatment with propidium mon..., Diagn Microbiol Infect Dis (2014), http://dx.doi.org/10.1016/j.diagmicrobio.2014.09.010

S. Ditommaso et al. / Diagnostic Microbiology and Infectious Disease xxx (2014) xxx–xxx

Recent scientific papers contributed to compare the molecular methods and the culture of Legionella; some authors (Fittipaldi et al., 2011; Joly et al., 2006; Whiley and Taylor, 2014) expressed doubts about the consistency of these 2 methods, but the needs to perform quickly, with a good sensitivity, managing large number of environmental samples, make it essential to find a consensus on how molecular methods results should be translated into culture-based limits contained in legislation and guidance documents. Acknowledgment This work was supported in part by grants from the University of Turin (ZOTCRILO 2012). References Association Française de Normalisation. Dètection et quantification des Legionella spp. et/ou Legionella pneumophila par concentration et amplification gènique par rèaction de polymerisation en chaìne en temps rèel (RT-PCR). T90-471 ; 2010 [Paris]. Ballard AL, Fry NK, Chan L, Surman SB, Lee JV, Harrison TG, et al. Detection of Legionella pneumophila using a real-time PCR hybridization assay. J Clin Microbiol 2000;38:4215–8. Bates MN, Maas E, Martin T, Harte D, Grubner M, Margolin T. Investigation of the prevalence of Legionella species in domestic hot water systems. N Z Med J 2000;113:218–20. Behets J, Declerck P, Delaedt Y, Creemers B, Ollevier F. Development and evaluation of a Taqman duplex real-time PCR quantification method for reliable enumeration of Legionella pneumophila in water samples. J Microbiol Methods 2007;68:137–44. Cawthorn DM, Witthuhn RC. Selective PCR detection of viable Enterobacter sakazakii cells utilizing propidium monoazide or ethidium bromide monoazide. J Appl Microbiol 2008;105:1178–85. Cazalet C, Gomez-Valero L, Rusniok C, Lomma M, Dervins-Ravault D, Newton HJ, et al. Analysis of the Legionella longbeachae genome and transcriptome uncovers unique strategies to cause Legionnaires' disease. PLoS Genet 2010;6(2). Chang B, Taguri T, Sugiyama K, Amemura-Maekawa J, Kura F, Watanabe H. Comparison of ethidium monoazide and propidium monoazide for the selective detection of viable Legionella cells. Jpn J Infect Dis 2010;63:119–23. Conferenza permanente per i rapporti tra lo stato le regioni e le province autonome di Trento e Bolzano. Linee guida per la prevenzione ed il controllo della legionellosi, n. 103. Gazzetta Ufficiale della Repubblica Italiana, serie generale; 2000. Declerck P, Behets J, Lammertyn E, Lebeau I, Anne J, Ollevier F. Detection and quantification of Legionella pneumophila in water samples using competitive PCR. Can J Microbiol 2006;52:584–90. Delgado-Viscogliosi P, Solignac L, Delattre JM. Viability PCR, a culture-independent method for rapid and selective quantification of viable Legionella pneumophila cells in environmental water samples. Appl Environ Microbiol 2009;75:3502–12. Ditommaso S, Gentile M, Giacomuzzi M, Zotti CM. Recovery of Legionella species from water samples using an internal method based on ISO 11731: suggestions for revision and implementation. Diagn Microbiol Infect Dis 2011;70:200–6. Dusserre E, Ginevra C, Hallier-Soulier S, Vandenesch F, Festoc G, Etienne J, et al. A PCRbased method for monitoring Legionella pneumophila in water samples detects viable but noncultivable legionellae that can recover their cultivability. Appl Environ Microbiol 2008;74:4817–24. Fittipaldi M, Codony F, Adrados B, Camper AK, Morató J. Viable real-time PCR in environmental samples: can all data be interpreted directly? Microb Ecol 2011;61(1):7–12. Fittipaldi M, Nocker A, Codony F. Progress in understanding preferential detection of live cells using viability dyes in combination with DNA amplification. J Microbiol Methods 2012;91:276–89. Health and Safety Commission. Legionnaires disease. The control of Legionella bacteria in water systems. Approved code of practice & guidance. United Kingdom: Her Majesty's Stationary Office; 2000.

7

International Standards Organisation. Water quality—Detection and enumeration of Legionella International standard. ISO 11731 ; 1998 [Geneva, Switzerland]. International Standards Organisation. Water quality—Detection and quantification of Legionella spp. and/or Legionella pneumophila by concentration and genic amplification by quantitative polymerase chain reaction (qPCR) ISO/TS 12869; 2012 [Geneva, Switzerland]. Joly P, Falconnet PA, André J, Weill N, Reyrolle M, Vandenesch F, et al. Quantitative realtime Legionella PCR for environmental water samples: data interpretation. Appl Environ Microbiol 2006;72:2801–8. Joseph C, Lee J, Surman Lee S, Drasar V, Crespi S, Briand E. EWGLI Technical Guidelines for the Investigation, Control and Prevention of Travel Associated Legionnaires' Disease; 2011 [Version 1.1]. Josephson KL, Gerba CP, Pepper IL. Polymerase chain reaction detection of nonviable bacterial pathogens. Appl Environ Microbiol 1993;59:3513–5. Koide M, Saito A, Kusano N, Higa F. Detection of Legionella spp. in cooling tower water by the polymerase chain reaction method. Appl Environ Microbiol 1993;59:1943–6. Lee JV, Lai S, Exner M, Lenz J, Gaia V, Casati S, et al. An international trial of quantitative PCR for monitoring Legionella in artificial water systems. J Appl Microbiol 2011;29: 1365–2672. Levi K, Smedley J, Towner KJ. Evaluation of a real-time PCR hybridization assay for rapid detection of Legionella pneumophila in hospital and environmental water samples. Clin Microbiol Infect 2003;9:754–8. Liu H, Li Y, Huang X, Kawamura Y, Ezaki T. Use of the dnaJ gene for the detection and identification of all Legionella pneumophila serogroups and description of the primers used to detect 16S rDNA gene sequences of major members of the genus Legionella. Microbiol Immunol 2003;47:859–69. Masters CI, Shallcross JA, Mackey BM. Effect of stress treatments on the detection of Listeria monocytogenes and enterotoxigenic Escherichia coli by the polymerase chain reaction. J Appl Bacteriol 1994;77:73–9. Ministère de la Santé et des Solidarietès. L’eau dans les Etablissements de Santé - guide technique; 2005 [France]. Ng DL, Koh BB, Tay L, Heng BH. Comparison of polymerase chain reaction and conventional culture for the detection of legionellae in cooling tower waters in Singapore. Lett Appl Microbiol 1997;24:214–6. Nocker A, Cheung CY, Camper AK. Comparison of propidium monoazide with ethidium monoazide for differentiation of live vs. dead bacteria by selective removal of DNA from dead cells. J Microbiol Methods 2006;67:310–20. Nocker A, Sossa-Fernandez P, Burr MD, Camper AK. Use of propidium monoazide for live/ dead distinction in microbial ecology. Appl Environ Microbiol 2007;73:5111–7. Pan Y, Breidt F. Enumeration of viable Listeria monocytogenes cells by real-time PCR with propidium monoazide and ethidium monoazide in the presence of dead cells. Appl Environ Microbiol 2007;73:8028–31. Villari P, Motti E, Farullo C, Torre I. Comparison of conventional culture and PCR methods for the detection of Legionella pneumophila in water. Lett Appl Microbiol 1998;27: 106–10. Wellinghausen N, Frost C, Marre R. Detection of legionellae in hospital water samples by quantitative real-time LightCycler PCR. Appl Environ Microbiol 2001;67:3985–93. Whiley H, Taylor M. Legionella detection by culture and qPCR: comparing apples and oranges. Crit Rev Microbiol 2014. (http://informahealthcare.com/doi/pdf/10.3109/ 1040841X.2014.885930). World Health Organization (WHO). Legionella and the prevention of legionellosis; 2007 [Geneva]. Yamamoto H, Hashimoto Y, Ezaki T. Comparison of detection methods for Legionella species in environmental water by colony isolation, fluorescent antibody staining, and polymerase chain reaction. Microbiol Immunol 1993;37:617–22. Yáñez MA, Carrasco-Serrano C, Barberá VM, Catalán V. Quantitative detection of Legionella pneumophila in water samples by immunomagnetic purification and real-time PCR amplification of the dotA gene. Appl Environ Microbiol 2005;71:3433–41. Yáñez MA, Nocker A, Soria-Soria E, Múrtula R, Martínez L, Catalán V. Quantification of viable Legionella pneumophila cells using propidium monoazide combined with quantitative PCR. J Microbiol Methods 2011;85:124–30. Yaradou DF, Hallier-Soulier S, Moreau S, Poty F, Hillion Y, Reyrolle M, et al. Integrated realtime PCR for detection and monitoring of Legionella pneumophila in water systems. Appl Environ Microbiol 2007;73:1452–6.

Please cite this article as: Ditommaso S, et al, Overestimation of the Legionella spp. load in environmental samples by quantitative real-time PCR: pretreatment with propidium mon..., Diagn Microbiol Infect Dis (2014), http://dx.doi.org/10.1016/j.diagmicrobio.2014.09.010