Water destined for personal and household consumption should be safe and acceptable in taste, odor and color. However, complaints about drinking water quality are a common issue among the Brazilian population. Also, due to the pollution of water bodies, social groups that are not supplied by treated water may be exposed to different contaminants. The aim of this study was to assess the efficiency of a water treatment tank coupled with UV light on the inactivation of enteric viruses and the reduction of chlorine concentration for use in residences, as well as in rural and isolated communities. Viral disinfection and chlorine concentration decay assays were performed in a tank with capacity of 300 L and a 36-W UV lamp coupled, with controlled temperature. Recombinant human adenovirus (AdHu5-GFP) and murine norovirus (MNV-1) infectivity were assessed after 0, 3, 6 and 12 h of water recirculation. 99.99% inactivation was reached after 12 h for AdHu5-GFP and before 6 h for MNV-1. Chlorine concentration had a decay of 0.77 mg/L after 12 h. Regarding the efficiency observed, a product model was designed. This tank model was efficient in ensuring viral inactivation as well as in reducing residual chlorine and can be adjusted to other scales.
The problem of poor water quality worsens when waste sanitation is not adequate. Worldwide, about 80% of waste generated is discharged into water bodies without receiving any treatment (UN 2015). The lack of proper treatment and disposal leads to the contamination of rivers, seas and groundwater. Water destined for personal use and household consumption should be safe and acceptable in taste, odor and color. Therefore, water must be free from pathogenic microorganisms and hazardous chemicals (WHO 2012). When such basic conditions are not met, people face serious health risks.
In Brazil, 85% of the population receives piped and chlorine-treated water supplied by municipal companies, and most households have one or more water storage tanks in their houses prior to consumption (IBGE 2015). However, complaints about water are a common issue among the Brazilian population, mainly because they are unsatisfied with the quality of water supplied, mostly due to taste and a lack of confidence in the safety of water.
Among those who are not supplied with treated water, the majority live in rural areas and isolated communities, such as indigenous and riverine areas. Typically, these social groups use water collected from wells, rivers and lakes for their daily needs, often without providing any treatment. Nevertheless, due to the constant pollution of water bodies, these communities may be exposed to different contaminants and are susceptible to develop possible diseases. Bofill-Mas et al. (2013) gathered several studies that detected human viral pathogens (human adenovirus and polyomavirus) in tap water, river water, lagoons and ground water worldwide.
Diarrhea remains the main disease related to the consumption of contaminated drinking water. 1.7 billion cases per year are reported by the World Health Organization (WHO), of which 88% are caused by inadequate hygiene, sanitation and water supply. Children are more vulnerable to infections produced by the consumption of contaminated water mainly due to their immature immune system. Diarrhea is estimated to cause 1.5 million child deaths per year, mostly among children under five living in developing countries (WHO 2013).
Chlorine is the disinfectant most widely used in water treatment plants. However, it has some disadvantages such as limited efficiency in protozoan-cyst inactivation, and its strong odor and taste when present in high doses, as well as the toxic disinfection by-products formed when it comes into contact with organic matter (LeChevallier & Au 2004). An alternative to chlorine treatment is the use of germicidal ultraviolet radiation (UV-C), since it has a high efficiency to inactivate various microorganisms in a reduced contact time, it forms fewer disinfection by-products and it does not interfere with taste and odor.
In general, bacteria are more susceptible to UV treatment when compared to enteric viruses. To achieve a decay of 4 logs (99.99%) of fecal bacteria, such as Escherichia coli and Streptococcus faecalis, a dose up to 30 mJ/cm2 is required. At the same dosage, some viruses are also inactivated, such as noroviruses and hepatitis A. However, other enteric viruses such as adenovirus require higher doses (around 170 mJ/cm2) to achieve similar levels of inactivation (LeChevallier & Au 2004; Hijnen et al. 2006).
In disinfection studies it is important to evaluate, whenever possible, the infectivity of viruses, since their genomes can be detected even when the virus is physically damaged and incapable of replicating. Thus, recombinant adenovirus type 5, which expresses green fluorescent protein (AdHu5-GFP), can be used as a model for adenovirus disinfection studies by counting fluorescent infected cells within 24 h post infection using fluorescence microscopy (Garcia et al. 2015). Murine norovirus (MNV-1) has been widely used as a surrogate for human norovirus, because it belongs to the same family and gender and has a similar structure and transmission route. In addition to that, it has a well-standardized plaque assay technique to detect viable norovirus within 48 h (Cannon et al. 2006; Gonzalez-Hernandez et al. 2012).
Thus, the aim of this study was to assess the efficiency of a water treatment tank coupled with UV-C radiation to inactivate enteric viruses and evaluate the reduction of chlorine concentration for use in residences as well as in rural and isolated communities, using AdHu5-GFP and MNV-1 as pathogen models.
Cell culture, viral production and viability assays
The cell lines HEK293A and RAW264.7 were used and they are permissive to AdHu5-GFP and MNV-1, respectively. Cells were grown in 180 cm2 flasks and maintained at 37 °C with 5% CO2 incubator. HEK293A cells were courtesy of Prof. Dr Aguinaldo Roberto Pinto from the Federal University of Santa Catarina. For cultivation, a DMEM medium was used with 10% fetal bovine serum (FBS), and 1% Hepes salt solution 1 M. RAW264.7 cells were kindly provided by Prof. Dr Rosina Girones Llop from the University of Barcelona, Catalonia, Spain. The cultivation was conducted in DMEM supplemented with 10% FBS, 1% l-glutamine 200 mM, 1.5% HEPES salt solution 1 M and 1% non-essential amino acids 10 mM 100×.
For viral production, 1.0 × 107 focus forming units (FFU) of AdHu5-GFP, kindly provided by the Wistar Institute of Anatomy and Biology (MTA N0 401150-6351), Pennsylvania, USA and 2.0 × 106 plaque forming units (PFU) of MNV-1 (provided by Prof. Dr Rosina Girones Llop from the University of Barcelona) were added to the respective cell cultures. After 1 h of incubation, 25 mL of maintenance medium was added, which differs in FBS concentration (2%). The cells were maintained at 37 °C and observed daily using a microscope for cytopathic effects.
The incubation time varied between 36 h and 48 h to obtain 90–100% cytopathic effect. At this stage, the bottles were frozen at −80 °C and defrosted at 25 °C three times in order to promote cell lysis and release the intracellular viruses. The medium was centrifuged at 3,500 × g for 5 min and the supernatant was collected and stored at −80 °C.
Experimental design of viral inactivation in tank
Viral disinfection assays with UV radiation were performed in a tank with a capacity of 300 L. A peristaltic bomb allowed water recirculation with a flow rate of 1,800 L/h (Figure S1, available online). The storage tank had a 36-W UV-C lamp (Atman II, low-pressure lamp, monochromatic light output at 254 nm) coupled, which applied an approximate dose of 44 mJ/cm² for each water passage. This dose was determined in a previous study performed by Souza et al. (2013), using the same tank. A chiller device was used to maintain the water temperature at 18 °C during the experiment.
The tank was filled with 300 L of tap water and the following physical-chemical parameters were measured in each replicate: temperature, pH, conductivity, and dissolved oxygen (before and during circulation of water). Free residual chlorine, turbidity and total coliforms were also assessed. After the measurement, 10 mL of a 10% (w/v) sodium thiosulfate solution was added in order to chelate all the residual chlorine, ensured by a new measurement. The water was artificially contaminated with AdHu5-GFP (4.0 × 106FFU/L) and MNV-1 (8.0 × 105 PFU/L). These viral quantities were planned in order to allow follow up to a decay of 4.0 log, which represents 99.99% inactivation.
After water homogenization, a sample of 10 L was collected, representing the initial time (0 h), then the UV lamp was turned on. After 3, 6 and 12 h of recirculation, new 10 L samples were taken. Samples were immediately concentrated with skimmed milk flocculation (Calgua et al. 2013). Briefly, the pH of the water sample was adjusted to 3.5 and 10 mL of 1% skimmed milk (Sigma®) solution was added at the same pH. After stirring and a rest period of 8 hours each, the bottom 500 mL was centrifuged at 3.800 × g, for 30 minutes at 4 °C. The pellet was then dissolved once again in a 0.2 M phosphate buffer solution at pH 7.5. The final eluate was inoculated into monolayers of HEK293A and RAW264.7 cells in order to assess viral infectivity using the fluorescence microscopy (AdHu5-GFP) and the plaque assay (MNV-1) techniques as described previously.
For comparison, the same experimental design was carried out, but without exposure to UV light. All the experiments were performed in three independent replicates. In each replicate, 10 L of dechlorinated tap water was artificially contaminated with 1.0 × 107 FFU of AdHu5-GFP and 2.0 × 106 PFU of MNV-1, as a positive control. This sample was concentrated and had its infectivity assessed, representing the viral recovery control.
Chlorine decay in water tank
Aiming the use of this storage tank for domestic purposes, an independent experiment was conducted, without virus addition and UV exposure, to assess the decay of chlorine concentration in water during the recirculation. The storage tank was filled with 300 L of tap water and 50 mL of a 2.5% sodium hypochlorite solution (2.38 mg/L of free chlorine) was added. This solution is suitable for disinfection and treatment of drinking water and distributed by the Brazilian Ministry of Health. Free residual chlorine was measured at the initial time (0 h) and after 1, 3, 6, 12, 24 and 36 h of water recirculation. Measurements were taken using a portable photometer HI96711C (Hanna Instruments) following the manufacturer's instructions. A chiller device was used to maintain the water temperature at 18 °C during the experiment. The experiment was performed in three independent replicates.
Experimental data were analyzed using repeated analysis of variance (ANOVA) measures and linear regressions. Levene's test was used to check the homoscedasticity. When necessary, Bonferroni was used as a post hoc test. All analyses were performed using IBM SPSS 19.0 software, and a p-value of <0.05 was considered significant.
Viral inactivation in water tank
Physical-chemical parameters measured before and after temperature stabilization are presented in Table S1 (available with the online version of this paper). All samples were negative for the presence of total coliforms.
Viral inactivation curves in the storage water tanks are presented in Figure 1(a) and 1(b). With UV radiation exposure, AdHu5-GFP decayed 1.14 log in the first three hours, reaching 4.03 log after 12 h (Figure 1(a)). The linear regression analysis showed a R2 = 0.9795. MNV-1 had a 3.58 log decay in 3 h with UV, and was not able to be detected at the following sampling times (Figure 1(b)). In tanks without UV treatment, both viruses remained stable and viable at all times evaluated.
One control sample was used in each replicate to evaluate the percentage of virus recovery using skimmed-milk concentration. The method applied in this study was able to recover an average of 53% of adenovirus and 26% of norovirus. These values represent high viral recovery rates and similar results were also obtained by Calgua et al. (2008; 2013), using the same protocol.
Chlorine decay in water tank
An independent experiment was performed to evaluate the chlorine concentration decay during water recirculation, without virus and UV exposure. After 12 h of water recirculation there was an average decay of 0.77 mg/L, reaching 2.11 mg/L in 36 h (Figure 1(c)). The linear regression analysis showed an R2 = 0.9564.
Model of water treatment tank
Regarding the efficiency of the tank attached to a UV lamp, a model of a product was designed (Figure 2), with the proposal that it be employed regularly. The model consists of a tank (40 × 40 × 75 cm) with water capacity of 100 L but maintaining the hydrological system and flow rate of the tank studied. There is a filter prior to the water entering the tank to remove larger particles and reduce turbidity. The water circulates via pipes using a peristaltic pump with an approximate flow of 30 L/min (1,800 L/h). During circulation, the water will pass through a reactor with a 36-W UV-C lamp, applying an estimated dose of 44 mJ/cm2 in each passage. The water returns to the tank and, after 4 h, this water would be treated and safe for consumption, ensuring the reduction of particles, pathogens and residual chlorine. This product can also be adjusted to other scales and operational values.
The water tank attached to a UV lamp (36 W) presented in this study proved to be able to inactivate 99.99% of adenovirus in 12 h and norovirus in less than 6 h. Although the norovirus inactivation was quick, it is not possible to ensure that a linear decay was observed, due to the single sampling point. The results were obtained in a 300 L tank. Also, there was a decay of free chlorine concentration during water recirculation. Regarding this efficiency, a model of a tank with a capacity of 100 L was proposed.
The efficiency of adenovirus inactivation in water using UV is well known. Hijnen et al. (2006) gathered several inactivation studies using UV showing that adenovirus is the most resistant virus, needing a dose of 170 mJ/cm2 for 4 log decay. Equivalent results were observed by Ryu et al. (2015) using integrated cell culture quantitative polymerase chain reaction (ICC-qPCR).
MNV-1 is the most resistant virus among the Calicivirus, although complete inactivation is achieved using an UV dose of 40 mJ/cm2 (Lee et al. 2008; Park et al. 2011). Park et al. (2015) demonstrated that in metallic surfaces, UVC was able to inactivate 3 logs of MNV-1 with a UV dose of 90 mJ/cm2.
The minimum UV dose required for water disinfection is dependent on the characteristics of the treatment system, exposure time, UV absorption coefficient by water and the species of microorganism studied. The variation of these factors, as well as the water source, quality and turbidity, could explain the differences in results in several studies.
Genome damage is the main factor for viral inactivation by UV. For human adenovirus, it was determined that DNA damage is higher when using UV at a wave length of 260 nm, generating 2 damages per kilobase when a dose of 50 mJ/cm2 is applied (Eischeid et al. 2009; Beck et al. 2014). Regarding RNA viruses, MNV-1 genome decay is intact when exposed to pulsed light (200–1,000 nm), with fragments smaller than 200 nucleotides remaining (Vimont et al. 2015). Beck et al. (2015) showed that MS2 phage (RNA single strand) is inactivated by UV mainly for genome damage, without distinctions of wave length.
UV also has action on viral protein, and more complex capsid structures are normally more resistant to denature, such as adenovirus (Beck et al. 2015). According to Eischeid & Linden (2011), human adenovirus hexon proteins and fibers are more susceptible to UV damage when exposed to a wavelength of 254 nm. For MNV-1, pulsed light (200–1,000 nm) can break VP1 proteins. Wigginton et al. (2012) suggest that UV treatment at 254 nm causes protein backbone cleavage in MS2 phages.
Adenovirus is composed of a double-strand DNA genome that can be repaired by the host cells' machinery during in vitro assays. Some studies showed that adenovirus exposed to UV are more susceptible when tested in cells that lack the DNA repair system (Day 1974; Rainbow 1980, 1989). Thus, the damage caused in AdHu5-GFP, in the present study, could be repaired, in part, by the host cell HEK293A. This fact can also contribute to a higher resistance to inactivation when compared with MNV-1.
As expected, free chlorine concentration has a decay during the recirculation of water in the treatment tank. Although the decay at 12 h (time required for 4 log viral inactivation in 300 L) was lower than 1 mg/L, this could be significant in reducing water taste and odor. Also, the assay was conducted at 18 °C and at higher temperatures the free chlorine decay could surpass the values observed (Monteiro et al. 2017).
The chlorine concentration for drinking water allowed by the Brazilian regulation is between 0.2 and 5.0 mg/L, but the use of 2.0 mg/L is recommended (Brazil Ministry of Health 2017). At this concentration, chlorine can influence the population's perception of water taste, odor and safety. Therefore, the reduction of free chlorine concentration due to water recirculation in the proposed treatment tank could improve the perception of these qualities by consumers.
The proposition of a smaller tank (100 L) allows easier installation for domestic use. Although the experiments were performed with a 300-L tank, if the tank hydrodynamics and UV dose are conserved, then the inactivation kinetics will be proportional at other scales.
Other water tanks with UV treatment preparing water for consumption have already been developed. In general, these tanks have a UV chamber over or inside the water. This kind of design does not allow a homogeneous irradiation of the contents, which might mean that not all the pathogens in the water are inactivated. Also, these tanks do not ensure the minimum contact time for safe consumption (Massholder 1997; Jung et al. 2005; Raymond & Engelhard 2008; Marchi 2015). Cary & Adam (2006) created a similar water treatment apparatus that also uses recirculation and a UV lamp, but our work shows the efficiency of viral inactivation and chlorine reduction on a larger scale. To our knowledge, this is the first study that proposes this kind of tank for drinking water disinfection focusing on the inactivation of enteric viruses and the reduction of chlorine for domestic use.
The 300-L tank can inactivate 99.99% of enteric viruses within 12 h and reduces the free chlorine concentration. Also, a model was proposed with one 100-L tank having the capacity to inactive pathogens while reducing the use of residual chlorine and being adjustable to other scales. Thus, this tank could be useful for houses that do not use tap water for drinking purposes or residences that are decentralized, such as in rural areas and isolated communities.
This work was supported by The Brazilian National Council for Scientific and Technological Development (CNPq) [grant numbers 420398/2016-3 and 400183/2014-5]. We thank Heather Louise Godwig for reviewing the English language. The authors declare they have no actual or potential competing financial interests.