Using propidium monoazide to distinguish between viable and nonviable bacteria, MS2 and murine norovirus



This article is corrected by:

  1. Errata: ERRATUM Volume 55, Issue 5, 397, Article first published online: 27 September 2012

G.P. Ko, Department of Environmental Health, School of Public Health, Institute of Health and Environment, Seoul National University, Kwanak-gu Kwanak-ro 1, Seoul 151-752, Korea. E-mail:


Aims:  The ability to distinguish between viable and/or infectious micro-organisms and inactivated cells is extremely important for correctly performing microbial risk assessments. In this study, we evaluated whether propidium monoazide (PMA)-qPCR could distinguish between viable and nonviable bacteria and viruses.

Methods and Results:  A PMA-qPCR combined assay was applied to viable and inactivated bacteria (Escherichia coli and Bacillus subtilis) and viruses (MS2 and murine norovirus [MNV]). PMA, a DNA-intercalating agent, in combination with PCR was better able to distinguish between viable and nonviable bacteria and viruses than conventional PCR.

Conclusions:  These results suggest that a combined PMA-qPCR assay can be used to measure the viability of bacterial cells and bacteriophage MS2, but not MNV.

Significance and Impact of the Study:  PMA-qPCR could potentially be used to measure the viability of some micro-organisms, including virus. However, a thorough evaluation should be performed prior to measuring the viability of micro-organisms by PMA-qPCR in a quantitative way.


Molecular techniques such as PCR and real-time quantitative PCR (qPCR) are commonly used to detect micro-organisms in various environmental media. Although these techniques are rapid, sensitive and specific, they are unable to distinguish between living and dead cells (Yanez et al. 2011). The viability of micro-organisms is one of the most critical factors in the fields of food and water hygiene and microbial risk assessment (Nocker et al. 2006). Previous studies have shown that the detection of nucleic acids using molecular methods does not prove that a micro-organism is viable and infectious (Josephson et al. 1993; Kramer et al. 2009). The inability to differentiate between live and dead micro-organisms can lead to overestimates of the risk to human health (Yanez et al. 2011).

Cultivation is the most commonly used method for analysing micro-organisms; however, to cultivate a target micro-organism, the proper media, culture conditions and relatively long incubation times are generally required (Kramer et al. 2009). The presence of viable but nonculturable (VBNC) organisms suggests that many viable bacteria cannot be cultivated in vitro (Kell et al. 1998). In fact, most micro-organisms (>99%) cannot be cultivated (Leckie 2005). Thus, cultivation may significantly underestimate viable cell numbers during the quantitative detection and disinfection of micro-organisms (Keer and Birch 2003). Nevertheless, cultivation remains the standard method for detecting viable micro-organisms. Other molecular methods for detecting only infectious micro-organisms have been investigated (Louie et al. 2000; Ko et al. 2003).

One feature used to distinguish between living and dead cells is the presence of an intact membrane (Nocker et al. 2006). Viable cells should possess an intact membrane, while dead cells are more likely to be membrane-compromised. DNA-intercalating dyes such as ethidium monoazide (EMA) and propidium monoazide (PMA) only penetrate dead or membrane-damaged cells (Nocker et al. 2006; Fittipaldi et al. 2010). These intercalating chemicals have an azide group that is photo-activated and covalently binds to DNA upon light exposure. This covalent interaction blocks DNA amplification by PCR (Lin et al. 2011). Recent studies have indicated that PMA treatment prior to DNA extraction and PCR results in the exclusive amplification of DNA from living cells (Nocker et al. 2006; Bae and Wuertz 2009). However, the PMA-qPCR method has been demonstrated only for bacteria and fungi; few studies have been performed with viruses (Nocker et al. 2006; Fittipaldi et al. 2010). Therefore, we evaluated whether PMA-qPCR can distinguish between viable and nonviable bacteria and RNA viruses. It is particularly important to quantify important nonculturable viruses, including norovirus, because no other available method allows the measurement of the infectivity of these viruses in vitro.

Materials and methods

Bacterial strains and culture

The bacterial strains used in this study included a Gram-negative bacterium (Escherichia coli C3000) and a Gram-positive bacterium (Bacillus subtilis). Escherichia coli C3000 (ATCC no. 15597) was grown in tryptic soy broth (Difco Laboratories, Detroit, MI, USA) overnight at 37°C with shaking and then stored at −70°C until use. The stock was thawed on ice and centrifuged at 2000 g for 10 min. The supernatant was discarded, and the pellet washed with 1 ml of PBS (2·7 mol l−1 NaCl, 54 mmol l−1 KCl, 86 mmol l−1 Na2HPO4, and 28 mmol l−1 KH2PO4). The stock concentrations were estimated to be ∼107 CFU ml−1. Bacillus subtilis (ATCC no. 6633) was grown in Luria–Bertani broth (Difco Laboratories) overnight at 37°C with shaking and then stored at −70°C until use. The concentrations of the stocks were estimated to be ∼107 CFU ml−1.

Virus culture and stock

Murine norovirus (MNV) was cultured in Raw264·7 cells in Dulbecco’s modified Eagle’s medium (DMEM; 10% foetal bovine serum, 10 mmol l−1 HEPES, 10 mmol l−1 sodium bicarbonate, 50 μg μl−1 gentamicin and 10 mmol l−1 nonessential amino acids) (Lee et al. 2008). Viruses were propagated and cultured in Raw264·7 cell monolayers. Inoculated cells were frozen and thawed three times and then mixed with equal volumes of chloroform. Next, they were centrifuged at 2000 g for 10 min. The viruses were concentrated using an Amicon filter (Ultra-15; Millipore, Billerica, MA, USA) at 5000 g for 10 min at 4°C. The supernatant was stored at −70°C until use. The MNV concentration, evaluated using a plaque assay, was approximately 108 PFU ml−1.

A bacteriophage MS2 plaque assay was conducted using the single agar method as described previously (EPA 2001). Host E. coli C3000 and bacteriophage MS2 (ATCC no. 15597-B1) were mixed and cultivated at 37°C in a shaking incubator overnight. Next, chloroform was added at a volume equal to the culture suspension and centrifuged at 4000 g for 20 min. The supernatant was recovered and stored at −70°C. The titre of the stock was approximately 1 × 1011 PFU ml−1.

Inactivation conditions

The bacteria were killed by the exposure to 70% isopropanol for 10 min with shaking. The isopropanol was then discarded, and the cells were centrifuged at 5000 g for 5 min. The pellets were resuspended in 1 ml of PBS, solutions with four different ratios of live to dead cells were made (100 : 0, 75 : 25, 25 : 75 and 0 : 100), and the final volume was adjusted to 500 μl. MNV was heat-treated for 10 min at 72°C in a water bath, while MS2 was inactivated at 80°C for 20 min. The viruses were mixed with three different ratios of untreated to heat-treated cells (100 : 0, 10 : 90 and 0 : 100), in a final volume of 500 μl. Inactivation of the bacteria and viruses was confirmed using a culture assay.

PMA treatment prior to PCR amplification

Propidium monoazide (Biotium Inc., Hayward, CA, USA) was dissolved in 20% dimethyl sulphoxide (DMSO; Sigma Aldrich Co., St Louis, MO, USA) to a final concentration of 50 mmol l−1 and stored in the dark at −20°C. For the bacteria, 1·25 μl of PMA was added to 500 μl of sample to produce a final concentration of 125 μ mol l−1 and incubated in the dark for 5 min with shaking. The samples were kept on ice and exposed to light for 10 min using a 500 W halogen lamp at a distance of 30 cm. Next, the bacteria were centrifuged at 5000 g for 5 min, resuspended in 500 μl of PBS and subjected to both cultivation and PMA-qPCR assays. MNV was treated with 125 or 250 μ mol l−1 PMA, while MS2 was treated with 10, 50 or 125 μ mol l−1 PMA. After treatment for 5 min in the dark followed by 10 min of light exposure, the samples were analysed by both plaque and RT-PCR assays.

Nucleic acid extraction and real-time qPCR

Bacterial (E. coli and B. subtilis) samples were extracted using an Ultraclean soil kit (Mobio Laboratories, Carlsbad, CA, USA) according to the manufacturer’s instructions. Real-time TaqMan PCR was performed using the universal eubacterial primers 1055f/1392r and the TaqMan probe 16STaq1115 with an Applied Biosystems 7300 real-time PCR instrument. The reaction mixture contained 2·5 μl of template, 10 pmol l−1 each primer, 0·7 μl of probe and 12·5 μl of 2× TaqMan Universal PCR Master Mix (Applied Biosystems, Foster City, CA) in a final volume of 25 μl. The reaction conditions were as follows: 2 min at 50°C and 10 min at 95°C, followed by 45 cycles of 95°C for 30 s, 50°C for 60 s and 72°C for 30 s. All assays included a negative control to which no template was added.

Murine norovirus and bacteriophage MS2 were extracted using a QIAamp Viral Mini Kit (Qiagen, Valencia, CA, USA) according to the manufacturer’s protocol. MNV real-time PCR was performed in a volume of 25 μl including 12·5 μl of 2× master mix, 1 μl of 25× RT-PCR enzyme mix, 50 pmol l−1 each primer, 0·06 μl of probe and 2·5 μl of template. The cycling conditions were as follows: 30 min at 48°C and 10 min at 95°C, followed by 45 cycles of 15 s at 95°C and 1 min at 60°C (Lim et al. 2010).

Bacteriophage MS2 real-time PCR was performed in a total volume of 25 μl, including 12·5 μl of 2× master mix, 1 μl of 25× RT-PCR enzyme mix, 50 pmol l−1 each primer, 0·06 μl of probe and 2·5 μl of template. The cycling conditions were as follows: 10 min at 45°C and 15 min at 95°C, followed by 40 cycles of 15 s at 95°C and 1 min at 60°C (Lim et al. 2010).

Long-template RT-PCR (LT-PCR) for MNV

In addition to the real-time RT-PCR assay, MNV was quantified using LT-PCR. Viral titres were estimated by end-point dilutions. PCR was performed in a total volume of 25 μl containing 5× RT-PCR buffer, 0·4 mmol l−1 dNTPs, 20 U of RNase inhibitor, 50 pmol l−1 each primer and 2·5 μl of template. Viral RNA was amplified for 30 min at 42°C and 15 min at 95°C, followed by 40 cycles of 1 min at 94°C, 1 min at 55°C and 1 min at 72°C with a final 10 min extension at 72°C. The product (880 bp) was visualized by 1% agarose gel electrophoresis and ethidium bromide staining. Table 1 shows the nucleic acid sequences and locations of the primers and probes used in this study. Using the RNAfold program (, we estimated the minimal free energy and possible secondary structure of the bacteria, MNV and bacteriophage MS2 primer target regions (Table 1).

Table 1.   Summary of real-time and LT PCR assays for bacteria, MNV and MS2
 AssayOligonucleotideSequence (5′→3′)Location*MFE† (kcal mol−1)
  1. *Bacterial sequence locations based on GenBank accession no. AF420301 (Dionisi et al. 2003); MNV based on GenBank accession nos. DQ285629 for real-time PCR (Lee et al. 2008; Lim et al. 2010) and AY228235 for LT-PCR (Hsu et al. 2007); bacteriophage MS2 based on GenBank accession no. NC001417 (O’Connell et al. 2006).

  2. †Minimal free energy.

BacteriaReal-time PCR1055FATGGCTGTCGTCAGCT1055–1070 
Bacteriophage MS2Real-time PCRMS2-ST-fGTCGCGGTAATTGGCGC632–648−23·20

Statistical analysis

A statistical analysis of our results was performed using spss ver. 10.0 (SPSS Inc., Chicago, IL, USA). The effects of PMA on live and dead cells were analysed using the Mann–Whitney U-test. Differences between viable and nonviable cells were evaluated using the Kruskal–Wallis test; t-tests were performed using sigmaplot ver. 10.0 (SPSS Inc.).


The ability of PMA treatment (125 μ mol l−1) to distinguish between the viability of E. coli and B. subtilis was evaluated using culture and qPCR (Fig. 1). The bacterial concentrations are presented as both absolute and relative values. As the fraction of inactivated bacteria increased, there was a significant decrease in the DNA copy number measured by PMA-qPCR (E. coli [< 0·05], B. subtilis [=0·001]). PMA-qPCR indicated a 3·04 log reduction in DNA amplification from dead Ecoli cells when measured by qPCR. For B. subtilis, PMA treatment resulted in a 3·48 log decrease in DNA amplification from dead cells. In contrast, as expected, quantification using conventional qPCR remained constant. These results suggest that conventional qPCR amplified DNA from dead cells regardless of the increased fraction of inactivated bacteria. In contrast, PMA-qPCR could distinguish between viable cells and dead cells, at least to some extent. However, Fig. 1 also shows that our PMA-qPCR assay results were not completely proportional to the fraction of viable cells.

Figure 1.

 Determination of viability by cultivation and real-time PCR with and without PMA [qPCR (inline image); PMA-qPCR (inline image); Culture (inline image); PMA-culture (inline image); Culture (inline image); qPCR (inline image) and PMA-qPCR (inline image)]. Cells were inactivated by the treatment with isopropanol (final concentration 70%) for 10 min with shaking. Various ratios of inactivated to live cells (100 : 0, 75 : 25, 25 : 75 and 0 : 100) were used for (a) Escherichia coli and (b) Bacillus subtilis. The mean values and standard deviations (error bars) were calculated from three independent replicates. *Significant difference between the reduction rates as identified by a t-test, depending on the live/dead cell ratio. NS, > 0·05; *< 0·05; **< 0·01.

The real-time PCR cycle threshold (Ct) values increased significantly as the concentration of dead cells increased (Table 2). For example, the average Ct value for non-PMA-treated dead E. coli was 18·01, whereas that for PMA-treated dead E. coli was 32·66. The increased Ct value suggests that DNA amplification was inhibited. The Ct value for PMA-treated dead cells increased by 14 and 17 for E. coli and B. subtilis, respectively, compared to live bacterial cells. Alternatively, the Ct values for the non-PMA-treated dead bacterial samples increased by only two and three for E. coli and B. subtilis, respectively. Therefore, information regarding bacterial cell viability can be obtained using PMA treatment.

Table 2.   Comparison of qPCR Ct values with and without PMA treatment for Escherichia coli and Bacillus subtilis
C t value E. coli P-valueb B. subtilis P-valueb
Cell ratio (live/dead)PMA-qPCRqPCRPMA-qPCRqPCR
  1. *Statistical significance (< 0·05).

  2. aStatistical analyses were performed using the Kruskal–Wallis test.

  3. bStatistical analyses were used to evaluate the differences between the PMA-qPCR and qPCR results using the Mann–Whitney U-test.

100 : 019·14 ± 2·1916·38 ± 2·420·12717·60 ± 1·2415·50 ± 0·620·127
 75 : 2519·37 ± 1·3416·37 ± 2·030·12716·30 ± 1·8116·07 ± 1·890·827
 25 : 7521·03 ± 0·8817·16 ± 1·170·0520·27 ± 1·7417·20 ± 1·860·127
  0 : 10032·66 ± 1·5618·01 ± 1·010·0534·26 ± 1·3618·99 ± 0·370·05
P-valuea0·0550·468 0·022*0·091 

The ability of PMA to determine MS2 viability was also evaluated. MS2 viability had a 2 log reduction when exposed to higher PMA concentrations (e.g. >125 μ mol l−1) (unpublished data). Therefore, a low PMA concentration (10 μmol l−1) was used in this study (Fig. 2). When MS2 was fully heat-inactivated, it was undetectable using either the plaque or the combined PMA-qPCR assay. Thus, the detection of inactivated MS2 was significantly lower than the detection of live MS2 by PMA-qPCR (< 0·01). However, more than 80% of the inactivated MS2 was detected using conventional real-time PCR without PMA treatment. These results strongly suggest that combined PMA-qPCR could differentiate between heat-inactivated and noninactivated MS2.

Figure 2.

 The ability of the combined PMA-RT-PCR assay to determine MS2 viability [plaque assay (inline image); qPCR (inline image); and PMA-qPCR (inline image)]. The PMA concentration used was 10 μmol l−1. Cells were killed by heat exposure (20 min at 80°C). Mixed ratios of untreated and heat-treated cells (100 : 0, 10 : 90, and 0 : 100) were used. Error bars indicate the standard deviations from three independent replicates. *Significant difference between reduction rates as identified by a t-test, depending on the live/dead cell ratio. **< 0·01.

The ability of PMA (125 or 250 μmol l−1) to determine MNV viability was also evaluated (Fig. 3). Because MNV was significantly inactivated, viable MNV detected by plaque assay was dramatically reduced when 100% of MNV was heat-treated (< 0·01). The detection of PMA-treated MNV decreased significantly using both qPCR and LT-PCR. However, by PMA-qPCR, inactivated MNV was still detected after treatment with either 125 or 250 μ mol l−1 PMA. LT-PCR detected slightly less MNV nucleic acid than the other methods, but more than 73% of MNV nucleic acid was detected with or without PMA treatment. These results indicate that the combined PMA-RT-PCR assay could partially distinguish between viable and inactivated MNV; however, the assay is limited. Improvements to the technique are necessary for the detection of MNV, and the quantification of viable virus remains problematic.

Figure 3.

 Measurement of MNV by plaque assay, qPCR with and without PMA, and LT-PCR with and without PMA [plaque assay (inline image); qPCR (inline image); PMA-qPCR (inline image); LT-PCR (inline image); and PMA-LT-PCR (inline image)]. (a) 125 μmol l−1 PMA. (b) 250 μmol l−1 PMA. Cells were killed by heat exposure (10 min at 72°C). Mixed ratios of untreated and heat-treated cells (100 : 0, 10 : 90, and 0 : 100) were used. Error bars indicate the standard deviations from three independent replicates. *Significant difference between reduction rates as identified by a t-test, depending on the live/dead cell ratio. NS, > 0·05; **< 0·01.


In this report, we evaluated whether a PMA-qPCR assay could distinguish between viable and nonviable bacteria and viruses. The viability of micro-organisms is crucial for environmental health microbiology and associated human health risk assessments. It was previously shown that molecular assays, including PCR, cannot distinguish between infectious and noninfectious micro-organisms (Josephson et al. 1993). To detect infectious micro-organisms exclusively, many different approaches have been considered. For example, the integrated cell culture-PCR assay (Shieh et al. 2008), detection of negative strand nucleic acid from a positive-sense virus (Lerat et al. 1996) or the detection of mRNA from a DNA virus (Ko et al. 2003) may increase the probability of identifying viable virus.

Recently, DNA-intercalating agents such as PMA and EMA have been used to detect micro-organisms with intact membranes (Nocker et al. 2006, 2007; Vesper et al. 2008). PMA cannot penetrate lipid bilayers easily; however, it can integrate into the DNA of membrane-compromised cells, covalently cross-link to the DNA and block PCR amplification. Therefore, PMA-qPCR could be used to distinguish between viable and nonviable bacterial cells based on the integrity of the cell membrane (Nocker et al. 2006, 2007; Kobayashi et al. 2009; Varma et al. 2009). The peptidoglycan-containing cell walls of Gram-positive and -negative bacteria provide structural integrity, while lipid membranes function mainly as permeability barriers. Viable cells with undamaged membranes can specifically exclude DNA-intercalating agents. Our study evaluated a limited number of microbial species; however, our results indicate that PMA-qPCR could at least partially distinguish between viable and inactivated bacteria.

All culturable bacteria should be viable; thus, cultivation was used as a standard method to measure viable micro-organisms in our study. However, not all viable cells are culturable. VBNCs, as well as damaged but still viable cells, cannot be cultivated in vitro. In our study, stocks of cells of a known concentration were prepared and used for subsequent experiments. VBNCs can exist in samples during experimental and storage procedures, which should be considered when using a PMA-qPCR assay to detect viable cells. Currently, cultivation is considered the gold standard for detecting viable micro-organisms. However, cultivation may underestimate the number of viable cells (Jannasch and Jones 1959). Therefore, the concentration measured by PMA-combined RT-PCR could be caused by either false positives (inactivated bacteria) or true positives (viable bacteria), which are not culturable in vitro. Further evaluation should be performed in the future.

In addition to bacterial cells, we evaluated single-stranded RNA viruses lacking a lipid membrane (MNV and MS2). Because of differences in the biological structures and other characteristics of the viruses and bacteria, the ability of PMA to penetrate and intercalate in the nucleic acids is likely very different. For example, MS2 and MNV are naked viruses, the coat of which consists of capsid proteins. This may increase the permeability of the capsid region to PMA, or the rigid structure of the capsid protein may make the incorporation of PMA into nucleic acid even harder in activated virus. These possible conflicting effects require careful investigation. In the present study, we found that a PMA-qPCR combined assay could distinguish between inactivated and infectious MS2 (Fig. 2). This assay was also able to differentiate between viable and nonviable MNV. However, there was a discrepancy between the results of cultivation and PMA-qPCR. Previous studies have demonstrated that the effects of DNA-intercalating agents differ depending on the micro-organism (e.g. avian influenza virus, T4 bacteriophage and Norwalk virus) (Fittipaldi et al. 2010; Graiver et al. 2010; Parshionikar et al. 2010), light exposure time and status of viral particles (Kim et al. 2011). In addition, the secondary structure of the nucleic acid can affect the binding affinity of PMA (Parshionikar et al. 2010). Thus, our conflicting results for MS2 and MNV may be due to differences in the size and secondary structure of the RT-PCR-targeted nucleic acids (Table 1). The numbers of plaques in plaque assay did not follow the exact ratio of untreated and heat-treated samples (Figs 2 and 3). It is unclear whether it was caused by either different efficiencies of plating (EOP) or other possible mechanisms (e.g. breaking of viral clumps or increase of viral transfection). This should be further investigated in future.

Differences between micro-organism inactivation procedures may also affect the outcome. For example, isopropanol and heat treatment involve different inactivation mechanisms and cause different structural changes in the micro-organism. In our study, we inactivated MNV and MS2 by exposing them to 72 and 80°C, respectively. Such high temperatures can damage the viral capsid (Fittipaldi et al. 2010; Parshionikar et al. 2010). In conclusion, combined PMA-qPCR assays can partially distinguish between viable and nonviable micro-organisms. PMA offers the usefulness of PCR for the quick determination of the presence of potentially viable micro-organisms in environmental and clinical samples without culture enrichment. PMA treatment requires only a short incubation prior to performing a typical DNA extraction procedure. However, the effect of particulate matter in the samples should be fully characterized prior to performing a PMA assay using natural water or stool samples. However, a thorough evaluation should be performed prior to measuring the viability of micro-organisms by PMA-qPCR in a quantitative way because the PMA-combined RT-PCR assay can determine the viability of micro-organisms such as norovirus with much greater sensitivity. Thus, our results suggest that further investigation using different conditions and micro-organisms is warranted.


This research was supported by a National Research Foundation of Korea grant funded by the Korean government (2011-0029826) and by the Agriculture Research Center programme of the Ministry of Food, Agriculture, Forestry and Fisheries,.