To our knowledge, this is the first study to compare the results of PCR, solid-phase cytometry, and standard methods for quantifying L. pneumophila
in graduated water samples. We confirm the large discrepancies between PCR and culture for L. pneumophila
quantification in treated water samples (2
). We show that L. pneumophila
DNA detected by PCR in water samples treated with chlorine for 24 h includes VBNC forms and dead bacteria. Indeed, culture was negative after the addition of 0.5 ppm chlorine for 24 h, whereas the PCR signal was preserved and the number of metabolically active bacteria (detected by TVC) was barely affected. With 1 ppm chlorine, PCR and immunodetection signals did not change but TVC values were halved. In the presence of 3 ppm chlorine, fewer than 10 bacteria were labeled by TVC, whereas the PCR signal was below the quantification limit and the immunodetection results were unaffected, suggesting that dead bacteria were detected by antibodies. All signals were below the quantification limit at 30 ppm chlorine. Kinetic studies of the effect of 0.5 ppm chlorine also showed that immunodetection and PCR efficiently detected both dead and viable bacteria, whereas culture and viability testing were both affected after, respectively, a few hours and 2 days of treatment. Viable bacteria detected by TVC labeling remained infective, since they were able to multiply in A. polyphaga
and recovered their culturability.
Chlorine is the most widely used disinfectant for drinking water, cooling towers, and wastewater. In France and elsewhere in Europe, concentrations of 1 to 2 (sometimes 3) ppm chlorine are used for continuous treatment, in order to obtain around 1 ppm at the point of use (DGS circular 2002/243 [22 April 2002]; 9a
). “Shock” treatment of drinking water networks uses 10 to 20 ppm chlorine, followed by washes. As much as 50 ppm chlorine can be used to decontaminate cooling towers. Although in natural environments, temperature, pH, and organic matter can affect the availability of the active form of chlorine and modify the resistance of the bacteria to chlorination (9
), here we used experimental concentrations compatible with those used for continuous treatment of drinking-water networks and for shock decontamination.
Chlorine treatment markedly influenced the results of the methods used here to quantify L. pneumophila
organisms. The concentration of 0.5 ppm chlorine abolished culturability (changing bacteria into VBNC forms) and reduced viability after 24 h. Culture was the least sensitive detection and quantification technique. The PCR method was less sensitive (detection limit, 170 GU/liter, according to the manufacturer's instructions) than immunodetection and TVC, which can detect 1 stained bacterium per filtered sample (17
). PCR is therefore less precise than solid-phase cytometry when the number of bacteria with amplifiable DNA is too low or the bacteria are unsuitable for any efficient DNA extraction and/or amplification due to their poor quality. These results imply that a small number of bacteria, probably including some viable cells, were present in some PCR-negative samples treated with 3 ppm chlorine. However, the discrepancy between the TVC signal (viable bacteria) and the immunodetection signal (intact bacteria) at the same chlorine concentration was even larger, suggesting that at 3 ppm chlorine, immunodetection detects mainly dead bacteria compared to TVC and PCR techniques. It has been shown that the bacterial membrane can be permeabilized by chlorine in distilled water (28
). As with other permeabilizing agents, such as Triton, antibodies can still detect the bacterial membrane or even potentially functional intracellular enzymes (5
). However, this permeabilization process, when used for prolonged periods of time, could allow the bacterial contents, including DNA, to escape or could allow chlorine to damage bacterial DNA. Indeed, HOCl, the more reactive form of chlorine in aqueous environments, causes lethal DNA damage even at the concentrations used in drinking water (9
). Bacterial “ghost” cells have been described as stable immunogenic bacterial cell envelopes that have been emptied of their cytoplasmic contents through a lysis pore created by expression of the cloned phage ϕX174 lysis gene E. Thus, an intact bacterial envelope does not necessarily imply bacterial viability or the presence of amplifiable DNA, as shown by the PCR, TVC, and immunodetection results in the presence of 3 ppm chlorine. Altogether, chlorine could act in the following sequence: (i) culturability is lost due to degradation of surface components (as in the presence of 0.5 ppm chlorine); (ii) chlorine gradually penetrates the cell and directly degrades intracellular esterases and/or allows the release of the fluorochrome after pore formation, thus explaining the gradual fall in the TVC signal with increasing concentrations of chlorine; (iii) chlorine degrades the bacterial DNA, leading to extinction of the PCR signal (as in the presence of 3 ppm chlorine); and (iv) chlorine abolishes immunodetection signals by destroying the physical integrity of the bacterial cell (as in the presence of 30 ppm chlorine). Taken together, these data show that PCR signals include viable bacteria, detected by TVC, and some dead bacteria. In contrast with PCR, whose signal is negative when the number of viable bacteria is null or very low, immunodetection detects mainly dead bacteria at 1 and 3 ppm chlorine.
enters the VBNC state in order to survive in starvation environments (4
). VBNC legionellae have been detected in natural water samples (8
). Chlorination results in the complete loss of culturability of both starved and nonstarved L. pneumophila
). Heat treatment also abolishes culturability without affecting bacterial integrity. Steinert et al. were the first to demonstrate that VBNC legionellae could be resuscitated by adding an amoeba, Acanthamoeba castellani
), and this finding was subsequently confirmed by other authors with VBNC legionellae obtained after starvation, heat, or treatment with chlorine and other biocides (4
). TVC with ChemChrome V6 detects intracellular esterase activity in various bacterial species, including L. pneumophila
). Nonspecific viability labeling methods such as TVC staining must be combined with a specific identification technique. Delgado-Viscogliosi et al. used anti-Legionella
antibodies and ChemChrome V6 staining with epifluorescence microscopy to detect total and viable L. pneumophila
bacteria in natural water samples. They also found that chlorine abolished viability staining without affecting antibody reactivity (8
In conclusion, our findings suggest that PCR results include viable (TVC-positive) but noncultivable bacteria. These viable bacteria were able to multiply in A. polyphaga
and to recover their culturability. This explains the large discrepancies between PCR and culture for L. pneumophila
quantification in chlorine-treated water samples (2
). Different environmental conditions and disinfection processes could have different effects on VBNC generation. Our results also suggest that dead bacteria were detected by PCR at 0.5 and 1 ppm chlorine in water together with VBNC forms, and not at higher concentrations (≥3 ppm chlorine), as PCR signals were not observed. Since the application domain of the viability and immunodetection quantification tools is still limited to filterable water, research should also be done to develop similar methods that can be applied to nonfilterable environmental or clinical samples.