Jose-Luis Sagripanti, Edgewood Chemical Biological Center, US Army, AMSRD-ECB-RT 5183 Blackhawk Rd. (Edgewood Area), Aberdeen Proving Ground, MD 21010-5424, USA. E-mail: email@example.com
Aims: To compare the inactivation rate of Venezuelan equine encephalomyelitis (VEE) virus in liquids to that of Sindbis virus (SV, another alphavirus) and to a bacteriophage (MS2) generally used as a viral simulant in the development of countermeasures in biodefense.
Methods and Results: Viruses were inoculated into liquids and viral titres were determined at various times postinoculation. The viruses were stable in distilled-deionized (dd) water at 4°C during the 21 days of the study. The inactivation rates of VEE and SV in dd water at 21 and 30°C were very similar (between 0·12 and 0·14 log10 per day), while MS2 was three-fold slower. In tap water (chlorine content between 4 and 5 ppm) at 21°C, VEE and SV were inactivated at twice the rate measured in dd water.
Conclusions: The inactivation rates of VEE and SV were similar to each other and faster than MS2 in all liquids tested.
Significance and Impact of the Study: VEE is likely to remain viable for many days after release into water, snow, or even chlorinated tap water. SV can be used to estimate the persistence of VEE in liquids, but using MS2 as a simulant would overestimate of the stability of VEE.
The alphavirus genus includes viruses that have been used as weapons in the past by both the U.S. and the former Soviet Union (Smith et al. 1997; Griffin 2001). Among these, of particular historical relevance is Venezuelan equine encephalomyelitis virus (VEE). VEE is pathogenic in equines, and normally cycles through rodents and a mosquito vector (Peters and Dalrymple 1990). Transmission of VEE can occur by the inhalation of infectious aerosols by respiratory route as well as by bites from mosquitoes. Mosquitoes are responsible for sporadic outbreaks in horses and man, generally occurring in North and South America. Symptoms can vary in severity and include headache, fever, confusion, seizures, paresis, dysphasia and cranial nerve palsies. Infection can be fatal and symptoms are often incapacitating with a prolonged period of convalescence (Griffin 2001).
It is extremely important to understand the ability of VEE to survive in the environment in order to provide accurate public health information and to initiate the necessary decontamination procedures after a biological attack. Thermal inactivation of VEE and the effect of sodium chloride on VEE inactivation have been previously described (Walder and Liprandi 1976). The stability of reconstituted VEE vaccine has also been reported (McManus and Robinson 1972).
Simulants allow researchers to avoid experiments with pathogenic VEE that must be performed in specialized biocontainment laboratories. Therefore, we compared the survival of VEE in liquids to the survival of two other viruses that could be used as simulants for VEE, particularly in biodefense research.
Sindbis virus (SV) is the prototype member of the alphavirus genus and has been used as a model for inactivation of enveloped viruses in plasma-derived products (Remington et al. 2004; Espindola et al. 2006). SV causes disease in birds and, like VEE, is normally transmitted through a mosquito vector (Taylor et al. 1955). Although SV can infect humans producing skin rash and arthritis (Griffin 2001), human infection usually occurs without recognized clinical disease (Peters and Dalrymple 1990). Strains of SV differing in virulence have been derived from independent isolates from Egypt, South Africa and Israel (Griffin 2001). SV has been used as a model virus in many studies on viral biochemistry and pathology in mice, where SV produces a variety of symptoms according to the strains of mice and SV used in the studies (Peters and Dalrymple 1990).
MS2 is an RNA phage with a genome of 3569 bp (Fiers et al. 1976). It does not cause disease in man, can be safely used at biosafety level 1, and can be released into the environment to study the effects of aerosol release of viruses. Previous studies on the survival of MS2 have shown that it is stable in a wide range of pH (Feng et al. 2003) and that the bacteriophage remains at practically constant titres for at least 30 h in buffered waste water (Kohn and Nelson 2007) or during 35 days in ground water (Gordon and Toze 2003). Other studies found that MS2 decayed at a rate comparable to feline calicivirus during 21-day incubation at 37°C in dechlorinated water, but the phage was approximately threefold more stable than feline calicivirus at 4 and 25°C (Allwood et al. 2003).
In this study, the survival of VEE in liquids should provide an indication of the persistence of this agent after an intentional or accidental release, and comparison with SV and MS2 should define any limitations in the use of either simulant in viral survival studies.
Materials and methods
Coliphage MS2 (Strauss and Sinsheimer 1963) and SV strain Ar-339 (Taylor et al. 1955) were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). The vaccine strain of VEE (TC-83), a biosafety level 2 tissue culture attenuated strain (Kinney et al. 1989) was provided by Michael Parker of USAMRIID. VEE and SV were propagated on African green monkey kidney (Vero) cells in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% foetal bovine serum, 100 U ml−1 penicillin, 100 mg ml−1 streptomycin and 2 mmol l−1l-glutamine, and the titre of VEE and SV quantified as previously described (Simizu et al. 1967; Bushar and Sagripanti 1990). MS2 was grown in E. coli strain 15597 (Strauss and Sinsheimer 1963) and titred by plaque assay (see below).
Liquids tested included: (i) distilled-deionized (dd) water that was sterilized by autoclaving, (ii) sterile phosphate buffered saline (PBS, containing 0·144 g l−1 KH2PO4, 0·795 g l−1 Na2HPO4, NaCl 9·0 g l−1, pH 8), (iii) tap water from the public drinking water supply system (City of Aberdeen, MD, USA). Three samples of tap water obtained at least a week apart were analysed by the National Testing Laboratories Ltd (Cleveland, OH, USA). All tap water samples showed a chlorine level (tested colorimetrically) between 4 and 5 ppm. The pH of tap water was 7·6, and the levels of chloride and sodium ions were 35 and 24 mg l−1, respectively. The levels of all heavy metals, as well as other inorganic and organic substances were below maximal levels acceptable in drinking water. Snow was collected immediately after falling and it was allowed to melt at room temperature inside a sterile beaker. Tap water and melted snow samples were sterilized by filtration (through 22 μm pore membranes) prior to inoculation with viruses. Each liquid was tested at least twice with each virus in experiments where each data point was titrated in duplicate.
Inoculation, incubation and sampling
The liquids under study were inoculated with 10 μl of viruses and mixed. In the experiments described for Fig. 1, the final concentration of SV in the liquids was between 7·0 and 8·4 × 103 PFU ml−1. For all other experiments, the final concentrations of VEE, SV, and MS2 in the liquids were between 1·5–3·5 × 106 PFU ml−1, 3·3–4·7 × 105 PFU ml−1, and 1·2–2·5 × 108 PFU ml−1, respectively. Samples of 0·1 ml were immediately removed and diluted for quantification by plaque assay as described below. All liquids were incubated at room temperature (21°C) after inoculation. For dd water, two additional tubes were inoculated, one of which was incubated at 4°C and the other at 30°C. At various time intervals, 0·1 ml samples were removed and serially diluted for quantification by plaque assay.
For liquids inoculated with VEE and SV, samples were diluted in 2 × modified Eagle medium (MEM) without phenol red, supplemented with 4% foetal bovine serum, 100 U ml−1 penicillin, 100 μg ml−1 streptomycin and 2 mmol l−1l-glutamine. Six-well tissue culture plates, seeded with 106 Vero cells per well the day before infection, were used for plaque assays. Tissue culture media was removed by aspiration and two wells were infected with 100 μl of each diluted virus suspension. Plates were swirled to mix and placed in a 37°C, 5% CO2 incubator for 1 h, swirling every 15 min. After the 1-h incubation period, wells were overlaid with 2 ml of a mixture of equal volumes of 2 × MEM and 2% agarose at 45°C. Once solidified, plates were returned to the incubator. After 2 days, wells were again overlaid with 2 ml of the overlay mixture containing 0·1 mg ml−1 Neutral Red. Once solidified, plates were returned to the incubator. Plaques were counted after 3–5 h of incubation.
For liquids inoculated with MS2, samples were diluted in 271 broth (1% w/w tryptone, 0·1% w/w yeast extract, 0·8% w/w sodium chloride, 1% w/w glucose, 0·2% v/v of 1 mol l−1 calcium chloride, and 0·1% v/v thiamine (10 mg/ml), pH 7·5), and then 100 μl of each of the dilutions were added to 100 μl of E. coli strain 15597 grown overnight at 37°C in 271 broth. Samples were incubated for approximately 5 min after which time 3 ml of 271 top agar was added to the sample and the mixture was poured over the surface of a 271 agar plate. Plates were incubated overnight at 37°C and plaques were counted.
Survival in various liquids
In a range setting series of experiments, the titre of SV in tap water fell below the detection level (<0·01 virus surviving fraction or <1% of original inoculum) within 3 days (Fig. 1). In contrast, SV survived with a relatively small decrease in titre for at least 8 days at 21°C in sterile dd water, PBS, or melted snow (Fig.1). This apparent stability of SV promoted further experiments where inactivation was followed through a wider range of virus titre (see below).
VEE levels were reduced (from 1·5–3·5 × 106 PFU ml−1) to undetectable levels (<10 PFU ml−1) after 3 weeks in PBS, while SV levels remained slightly (about 1 log10) higher than VEE (data not shown). In contrast, MS2 phage showed almost no reduction in viable virus even after the 3-week sampling period. Results similar to those obtained in PBS were obtained in experiments with viruses inoculated into the appropriate liquid culture media (SV and VEE in DMEM, and MS2 in 271 Broth).
Effect of temperature on virus survival
The effect of temperature on the titres of VEE, SV and MS2 in dd water was studied during 21 days and the results are shown in Fig. 2. None of the viruses showed considerable inactivation at 4°C during the 21 days of the study. We have now carried out the incubation of SV at 4°C for 194 days with less than a 2 log10 reduction in viral titre (data not shown). The rates of virus inactivation at 21°C or at 30°C were very similar for both VEE and SV, with virus titres declining steadily over the 3-week time period to levels 2–3 log10 below the starting titre. In contrast to the results obtained with the alphaviruses, at 21°C MS2 levels remained similar to the original inoculum level over the 3-week study span. An approximate 1 log10 decrease in survival of MS2 was observed after 21 days in water at 30°C (Fig. 2).
Virus inactivation rates
The mean decimal reduction value (D value), which is the number of days before a 90% reduction in virus titre is attained, has been used to measure viral decay and survival (Allwood et al. 2004). This parameter is also similar to the 1 log10 removal time (also known as T90) used in other studies (Gordon and Toze 2003). To characterize the virus inactivation we selected a value better reflecting the whole incubation period, since the first log10 reduction generally showed the greatest variation in our experiments.
The kinetics of viral inactivation were derived from experimentally determined survival data of the type depicted in the figures. The resulting kinetic graphs were fitted to a straight line of the form y = ax + b, where ‘a’ is the slope corresponding to the inactivation rate (expressed in log10 per day), and ‘b’ is the y-intercept. The inactivation rates for VEE, SV and MS2 in liquids of environmental interest are presented in Table 1 and the goodness of the fit of data to the inactivation kinetics curve is represented by the respective correlation coefficient (r). The inactivation rates in Table 1 demonstrate that the inactivation kinetics of VEE is similar to that of SV in tap water and in dd water (at any tested temperature). MS2 was more resistant than the mammalian viruses, with inactivation rates several-fold slower than VEE or SV. In PBS, inactivation of MS2 was over 10-fold slower than inactivation of VEE and SV (data not shown). In addition, Table 1 shows the log10 reduction to be expected for the studied viruses in different liquids and conditions after 21 days, and the time in days required for virus concentration to decrease by 3 log10 (99·9% kill or 0·1% survival of the initial viral load). These two parameters may be useful to quickly estimate environmental decay after a viral release in liquid environments. In dd water, SV required less time for a 3 log10 reduction than VEE (20 days vs 33 days at 21°C, 19 days vs 28 days at 30°C), while MS2 required greater than 50 days for a 3 log10 reduction at all temperatures (Table 1).
Table 1. Virus inactivation rate (IR)†, log reduction at 21 days (R21)‡, and time to reduce survival by 3 log (T3)§
†Inactivation rate (IR) was calculated as the slope obtained by plotting virus inactivation vs time, up to 21 days, and expressed as log10 inactivation per day. The correlation coefficient of the regression (r) is indicated in parentheses.
‡R21 was calculated by subtracting virus titre after 21 days from original virus titre in the inoculum at the beginning of the experiment. Average ± SE is displayed.
§T3 was graphically obtained as the value in the abscissa corresponding to a reduction of 3 log in virus survival in the ordinate.
¶NSC – no significant change from original inoculum.
0·18 ± 0·99
0·09 ± 0·16
2·59 ± 0·71
0·16 ± 0·23
2·94 ± 1·00
0·76 ± 0·25
5·33 ± 0·22
2·23 ± 0·22
Inactivation in tap water
The inactivation rates of VEE, SV and MS2 were determined at room temperature (21°C) in tap water from the public drinking system with a chlorine content between 4 and 5 ppm. The titres of VEE and SV in tap water declined steadily over time, being reduced by more than 5 log10 to below the limit of detection (<10 PFU ml−1) within 3 weeks (Fig. 3). In contrast, MS2 was still easily detectable at the end of the experiment (21 days incubation), albeit at a reduced concentration. We considered the inactivation of alphaviruses in tap water as resulting from two components, the effect of dd water and the killing by chlorine in tap water. Assuming that these two effects are independent and additive, then, the contribution of chlorination can be estimated by subtracting the virus inactivation rate in dd water from the inactivation rate measured in tap water (Table 1). After deducing the inactivation rate in dd (nonchlorinated) water, the chlorination level of 4–5 ppm produced an inactivation rate for VEE of between 0·11 and 0·14 log10 per day at room (21°C) temperature.
The stability of VEE, SV, and MS2 was tested over time in various liquids and at a range of temperatures that might be encountered outdoors, particularly in water supplies. The linear relationship shown in the figures indicates that VEE and SV in our study consisted of relatively homogeneous populations of viruses (without sensitive or more resistant subpopulations). In contrast, the inactivation of MS2 as a function of time appeared biphasic (Fig. 3), indicating the presence of a more resistant subpopulation of viruses, perhaps due to aggregation of virus particles in the liquids.
While the temperature had an effect on the survival of the alphaviruses, the concentration of these viruses decreased less than three log10 after 21 days at 21°C or 30°C (Fig. 2). Although a 3-log10 inactivation may appear substantial in the laboratory, this inactivation level may be small when compared to the levels likely to be encountered after an intentional or accidental release of viruses in the environment. Survival of both VEE and SV remained near the original inoculum levels when maintained at 4°C. Our long term incubation of SV at 4°C indicates that less than a 2 log10 reduction in viral titre should be expected for alphaviruses after 6 months in cold water. However, the results reported here do not include the effect of sunlight that should substantially increase virus inactivation, at least in shallow bodies of water where solar virucidal UV can penetrate (Lytle and Sagripanti 2005; Sagripanti and Lytle 2007). Our results also do not consider the potential interactive effect of other microbes on the viruses, since all of the liquids we used were sterilized before inoculation. Bacteria present in environmental liquids might utilize the virus particles as a food source resulting in an increased inactivation rate. Although we know of no studies demonstrating differences in the stability in liquids between pathogenic strains and the attenuated strain of VEE (TC-83) used in our studies, further studies are necessary to compare the stability of attenuated and virulent strains of VEE.
The relatively slow inactivation of alphaviruses in dd water or melted snow suggests that a very low level of ions stabilizes VEE and that the presence of salt (in PBS) does not further stabilize VEE. Our findings agree with the enhanced inactivation of VEE by salt previously reported at 55°C, but seem in apparent contradiction with observations from the same source where hypertonic (2 mol l−1 sodium chloride) stabilized VEE at temperatures below 32°C (Walder and Liprandi 1976).
The relatively rapid inactivation of VEE in tap water was likely due to chlorination (Hurst 2001). The tap water from our public distribution source had a chlorine level between 4 and 5 ppm, typical of public water sources (http://www.epa.gov/safewater/contaminants/index.html#primary). After deducing the inactivation rate in dd (nonchlorinated) water, the chlorination level of 4–5 ppm produced an inactivation rate for VEE of between 0·11 and 0·14 log10 decay per day at room (21°C) temperature (Table 1).
MS2 did not remain at high titres for long periods of time in tap water either. The long term survival of MS2 in dd water that we observed agrees with a previous study where the titre of MS2 remained practically unchanged for at least 30 h in phosphate buffer (Kohn and Nelson 2007). The inactivation rate that we determined for MS2 in dd water is approximately four-fold slower than reported in groundwater samples (Yates et al. 1985) and one order of magnitude slower than the reported inactivation of MS2 on lettuce and cabbage leaves (Allwood et al. 2004). This suggests that MS2 is more stable in the absence of salt, ions (as in dd water), or possible microbes, and more persistent in liquids than on some contaminated surfaces.
The stability of MS2 in groundwater has been found to be similar to that of Poliovirus 1 and Echovirus 1 (Yates et al. 1985; Gordon and Toze 2003). In water at 37°C, MS2 inactivation was similar to inactivation of feline calicivirus, but differences were noted at lower temperatures (Allwood et al. 2003). Poliovirus and Echovirus 1 are members of the Picornaviridae family, which together with members of the Caliciviridae family are all nonenveloped viruses (Murphy and Kinsbury 1990). MS2 is a nonenveloped bacteriophage that was more stable than the enveloped mammalian viruses (VEE or SV) studied here. VEE and other enveloped mammalian viruses may be inactivated more easily than nonenveloped mammalian viruses and bacteriophages due to damage to the lipid or proteins in the envelope.
VEE and SV (as other alphaviruses) replicate in the mammalian host and are transmitted through short-term exposure to the mosquito vector. This life cycle suggests that survival outside either host may not pose a strong evolutionary pressure in the alphaviruses. In contrast, bacteriophages like MS2 may have evolved to be more suited to survival in an aqueous environment after release from the infected bacterium, since a substantial amount of time might pass before the phage encounters another susceptible bacterial host. Different life cycles may have favoured the evolution of bacteriophages that are more stable than mammalian enveloped viruses in aqueous environments.
Whatever the reason for the observed differences in stability, the results that we obtained highlight the disparity in survival observed between MS2 and the alphaviruses. Thus, our data demonstrate that MS2 is inadequate to simulate the survival of VEE in liquids. Using MS2 as a viral simulant should lead to overestimation of the ability of VEE (and other alphaviruses) to survive in liquid environments. This overestimation would unnecessarily increase quarantine periods and other countermeasures, delaying access to water sources, snow fields, and other moist or watery environments naturally contaminated or infected after an intentional viral release. The relative survival of SV and VEE were not identical in all liquids tested, with SV decaying slightly (c. 20%) faster at 21 and 30°C, and nearly twice as fast at 4°C than VEE (according to corresponding R21 values in Table 1). Therefore, doubling of the T3 values obtained for SV in liquids may provide a safer estimate of VEE survival. These findings highlight the challenge faced to assure the safety of water supplies and to remediate environments contaminated with viral agents.
We gratefully acknowledge Dr Mike Parker of USAMRIID for supplying the VEE TC-83 virus. This work was supported by the US Department of Defense Chemical and Biological Defense program administered by the Defense Threat Reduction Agency and by In-House Laboratory Independent Research (ILIR) funds from the Research and Technology Directorate, Edgewood Chemical Biological Center, Research Development and Engineering Command, US Army.