|Home | About | Journals | Submit | Contact Us | Français|
Our research on the mechanisms of action of chlorine-based oxidants on Cryptosporidium parvum oocysts in water revealed a dual-phase effect: (i) response to oxidative stress, which was demonstrated by induced expression of the Hsp70 heat shock gene, and (ii) oocyst inactivation as a result of long-term exposure to oxidants. The relative biocidal effects of sodium hypochlorite (bleach) and electrolytically generated mixed oxidant solution (MOS) on C. parvum oocysts were compared at identical free chlorine concentrations. Oocyst inactivation was determined by quantitative reverse transcription-PCR (qRT-PCR) amplification of the heat-induced Hsp70 mRNA and compared with tissue culture infectivity. According to both assays, within the range between 25 and 250 mg/liter free chlorine and with 4 h contact time, MOS exhibits a higher efficacy in oocyst inactivation than hypochlorite. Other RNA-based viability assays, aimed at monitoring the levels of β-tubulin mRNA and 18S rRNA, showed relatively slow decay rates of these molecules following disinfection by chlorine-based oxidants, rendering these molecular diagnostic viability markers inappropriate for disinfection efficacy assessment.
Cryptosporidia, the etiological agents of cryptosporidiosis outbreaks worldwide, are known for their remarkable ability to withstand chlorination even at free chlorine levels far exceeding those normally employed in water treatment processes (3, 7, 17, 19).
Research on protozoan oocyst inactivation by chemical disinfectants has been somewhat hindered by the lack of reliable and rapid quantitative assay methods for assessing inactivation and viability in Cryptosporidium oocyst populations (4). Undoubtedly, oocyst infectivity is the best way to measure disinfection efficacy (12, 25), but this technique requires specialized laboratory setup and has a relatively long turnaround time.
The initial goal of this study was to evaluate the applicability of different molecular markers for the development of a rapid and potentially field-deployable quantitative PCR (qPCR) assay for assessing oocyst viability directly, without an intermediate cell culture or animal infection step, as well as to demonstrate the applicability of such a rapid assay for evaluating the biocidal efficacy of chlorine-based oxidants and of other disinfection techniques. As such, in the water treatment practice, a qPCR-based assay could be applied (i) when immediate decisions are needed to respond to a detectable water threat and mitigate the potential health consequences, (ii) for rapid assessment of disinfection efficacy when it needs to be quantified within the time frame of a few hours following chemical or UV disinfection, and (iii) to rapidly adjust and optimize disinfection dose rates to reduce the use of disinfection chemicals when possible and therefore reduce disinfection by-product formation in the finished water. At the same time, the qPCR assay may provide a valuable research tool for better understanding the molecular and cellular nature of the chlorine resistance of protozoa.
The viability assays included quantitative reverse transcription-PCR (qRT-PCR) to detect and quantify changes in gene expression by measuring the level of the heat-induced Hsp70 mRNA, as well as other molecular viability indicators, β-tubulin mRNA and 18S rRNA, in Cryptosporidium parvum oocysts treated with chlorine-based disinfectants. Oocyst inactivation, quantified by qRT-PCR, was compared with oocyst infectivity in cell culture.
Oocysts of the Iowa II isolate of C. parvum have been routinely obtained from the Sterling Parasitology Laboratory, Department of Veterinary Science and Microbiology, University of Arizona. The oocysts were used within 45 days after the time of harvesting from infected calves and initial purification. The excystation rate remained about 95% within this period, indicating a negligible change in oocyst viability.
The oocyst inactivation experiments were carried out in EPA type 1 water (i.e., dechlorinated tap water) at 21 to 22°C in the presence of different oxidant disinfectants: sodium hypochlorite (Clorox bleach) and MIOX mixed oxidant solution (MOS), generated by salt brine electrolysis in a MIOX/MSR purifier. The oxidants were applied at identical concentrations and for identical exposure times for comparison of their biocidal efficacies. The pH was 7.3 ± 0.1. The concentrations of hypochlorite and MOS were determined as free available chlorine (FAC) by using the DPD (N,N-diethyl-p-phenylenediamine) free chlorine reagent (product no. 1407028; Hach). Negative controls included samples of 2 × 105 oocysts killed by exposure to 10% ammonia solution for 1 h at 65°C (13; G. Bajszar, unpublished data). Upon oocyst treatment at temperatures higher than 65°C, loss of the PCR template was observed, most likely due to thermal degradation of a portion of the oocyst RNA (data not shown). Freeze-thawing cycles for oocyst inactivation may also contribute to relatively rapid RNA degradation (5).
Aliquots of 100 μl containing 2 × 105 oocysts were exposed to the chlorine-based disinfectants for various contact times, after which the disinfectants were quenched with 0.05 mM sodium thiosulfate for each 5-mg/liter FAC dose. The oocysts were then subjected to heat shock for 20 min at 45°C to induce the expression of the Hsp70 gene. The level of the Hsp70 mRNA was measured by qRT-PCR.
Total RNA was purified with the phenol-chloroform-guanidine isothiocyanate-based Tri-Reagent RT-LS (catalog no. 311; Molecular Research Center, Inc., Cincinnati, OH). The RNA purification protocol was performed essentially as suggested by the supplier and consistently yielded pure PCR-ready RNA free of DNA contamination. RNA yields were measured in a Picofluor fluorometer (Turner BioSystems) using Ribogreen fluorescent dye for RNA quantification. The RNA concentration was adjusted to 4 ng RNA per μl, and 5-μl aliquots, corresponding to 20 ng total oocyst RNA, were added to each qRT-PCR mixture.
GenBank sequence entries were explored for PCR primer and TaqMan probe sequences, which corresponded to the published genome sequences (1, 32). A number of PCR primer and TaqMan probe sequences for the amplification of fragments within the Hsp70 (heat shock) mRNA, the β-tubulin mRNA, and the 18S small-subunit (SSU) rRNA were designed with Beacon Designer software (Premier Biosoft, Palo Alto, CA). The primers and probes were purchased from Biosearch Technologies (Novato, CA).
For Hsp70 (GenBank accession no. XM_625373.1), the set was as follows: forward primer (FP10), 5′TTGAACGTATGGTTAATGATGCTG3′; probe, 5′FAM-AGTCTGTTCTGCTCATCCTCACCC-BHQ3′; reverse primer (RP10), 5′CGAGCCAGTCAAGAGCATCC3′.
This primer-probe set amplifies a 91-bp fragment within the Hsp70 mRNA of the C. parvum Iowa II strain (GenBank accession no. XM_625373.1). A BLAST search on this sequence revealed also the homologous Hsp70 genes of C. hominis (XM_661662.1) and C. muris (XM_002140816.1) as the GenBank nucleotide entries with 100% matching scores for the same primer and probe sequences. Other primer-probe sets used in this study are listed below.
For the β-tubulin gene (GenBank accession no. Y12615), the primers were as follows: forward primer FP1βT, 5′TCTGGCTATTGGGGAACTTG3′; probe1βT, 5′FAM-GGGAACCAGATTGGTGCTAA-BHQ3′; reverse primer RP1βT, 5′ATCCCGTGCTCATCAGAAAT3′; amplicon, 199 bp.
For the SSU (18S) rRNA gene (GenBank accession no. AF093493.1), primers were as follows: forward primer FP118S, 5′ACTACCTCCCTGTATTAGGATTGG3′; reverse primer RP118S, 5′ACGGGGAATTAGGGTTCGATTC3′; probe18S, 5′CGCGCCTGCTGCCTTCCTTAGATG3′; amplicon, 106 bp.
Each reaction mix contained 20 ng of total oocyst RNA in a 20-μl volume. The qScript one-step fast qRT-PCR enzyme kit (Quanta Biosciences) was used according to the manufacturer's instructions. The concentration of the forward and reverse primers was 500 nM, and that of the 5′FAM-3′BHQ TaqMan probe was 100 nM. The thermal cycler (Eppendorf Realplex2 Mastercycler EP) was programmed as follows: cDNA synthesis, 10 min at 50°C; initial denaturation, 1 min at 95°C; and PCR cycling (40 to 45 cycles), 95°C for 5 s and 60°C for 30 s (data collection step). The controls included RNA samples isolated from C. parvum oocysts not subjected to heat shock, RNA samples from dead oocysts (killed by exposure to 65°C for 1 h in 10% ammonia solution ), and qPCR amplifications without the reverse transcriptase. Use of the last control served to verify the lack of genomic DNA contamination in the RNA preparations.
The focus detection method (FDM) was used essentially as described by Slifko et al. (25) with only minor modifications. The Sporo-Glo kit containing fluorescent antibodies and reagents for the FDM assay was purchased from Waterborne, Inc. (New Orleans, LA). HCT-8 human ileocecal adenocarcinoma cells (ATCC CCL-244) were purchased from the American Type Culture Collection, Manassas, VA, and grown in monolayer cultures according to ATCC instructions. Confluent HCT-8 cells were infected with serial 10-fold dilutions of C. parvum oocysts per well, starting with 2 × 105 oocysts. The infection foci were visualized with fluorescein-conjugated anti-sporozoite antibodies, which react with sporozoites, merozoites, and all other reproductive forms of C. parvum but not with the intact oocysts. After 24 h of incubation at 37°C, the bright-green fluorescent infection foci were counted using a Zeiss Axioscope 20 epifluorescence microscope. The results were computed on the basis of oocyst dilutions. Statistical evaluation was limited to the calculation of standard deviation.
Statistical analyses of data were performed with JMP statistical software (SAS Institute, Inc., Cary, NC).
The level of the heat-induced stress response was considered to be a quantitative indicator of oocyst viability because the qRT-PCR amplification plots for heat-induced Hsp70 mRNA showed a strong correlation with the number of oocysts present in the test samples (Fig. 1A and B). The induction of Hsp70 mRNA by elevated temperature (45°C, 20 min) was measured in C. parvum oocysts by qRT-PCR. Typical qRT-PCR amplification of the Hsp70 mRNA from 10-fold oocyst dilutions is shown in Fig. Fig.1A.1A. Because the threshold cycle (CT) values slightly vary from one experiment to another, a series of mRNA amplification plots for 10-fold dilutions of oocyst samples (in the concentration range between 105 and 102 oocysts per sample) were included in each experiment as an internal calibration standard.
The control experiments with oocysts not subjected to heat shock (Fig. (Fig.1B)1B) indicate that a significant base level of Hsp70 mRNA is present in oocysts naturally, without heat treatment. We proved that this base level of Hsp70 mRNA is not an artifact from the amplification of residual genomic DNA contamination in the RNA preparation isolated from the oocysts: the control samples, which were amplified without reverse transcriptase, revealed only a negligible PCR amplification, which would correspond to Hsp70 mRNA amplification from a 104-fold oocyst dilution. The uninduced base level of Hsp70 mRNA is 10 to 15 times lower than the heat-induced level (Fig. (Fig.1B),1B), and Hsp70 mRNA was undetectable in oocysts killed at 65°C in 10% ammonia (not shown).
It seemed to be a straightforward assumption that the correlation between live oocyst numbers and induced Hsp70 mRNA levels can be used in developing a rapid PCR-based quantitative viability assay. However, when the oocyst inactivation was a result of chemical disinfection, the qRT-PCR assay workflow had to be modified because elevated Hsp70 gene expression is also induced by chlorine-based oxidants. The primary effect of either HOCl/OCl− or MOS on C. parvum oocysts manifests in the induction of a quasi-heat shock response triggered by oxidative stress, resulting in the upregulation of the nuclear Hsp70 gene. At FAC concentrations in the range between 5 and 125 mg/liter, increased Hsp70 mRNA synthesis was shown within the first 2 h of exposure to the chlorine oxidants (Fig. (Fig.2).2). Upon longer exposure, especially at higher FAC concentrations, the level of Hsp70 mRNA decreases, indicating oocyst inactivation.
In order to establish a correlation between heat shock response and oocyst inactivation, first the time course of Hsp70 mRNA synthesis and decay was determined. A batch of 3 × 106 oocysts was subjected to MOS at a 25-mg/liter FAC concentration for 1 h. The oxidant was quenched, and the oocysts were briefly centrifuged and resuspended in distilled water. The oocyst suspension was incubated at 22°C, during which aliquots of 105 oocysts were taken at different time intervals (1, 2, 4, 6, 8, 12, 16, and 24 h), and a second heat shock was applied at the 24-hour time point to measure the extent of the oocysts' heat shock response during a 4-h interval, using the qRT-PCR amplification of the Hsp70 mRNA. The controls were (i) the same number of oocysts subjected initially to heat shock at 45°C for 20 min but not treated with chlorine oxidants and also (ii) oocysts completely killed by exposure to 65°C for 1 h in 10% ammonia. The results are shown in Fig. Fig.33.
It can be concluded that the Hsp70 mRNA induced by either 20-min heat shock or oxidant shock decays within 12 to 16 h and returns to its base level when the oocysts are kept at 21 to 22°C. When the oocysts are exposed to MOS for longer times, a lower decline rate of the Hsp70 mRNA level is observed, and even after 12 h it was significantly higher than the base mRNA level. A heat shock applied at 24 h still induces a stress response, indicating that a portion of the oocysts is alive. The number of live oocysts depends on the initial FAC concentration (Fig. (Fig.3B)3B) and, as shown below in Fig. Fig.4,4, depends also on the type of the chlorine-based oxidant applied for disinfection; i.e., whether it is dissolved chlorine as hypochlorite or as mixed oxidant.
When the oxidant concentration and extended exposure time “overwhelm” the intracellular defense, the oxidant treatment eventually leads to oocyst death. We investigated the oocyst inactivation for a wide range of oxidant concentrations—from 5 up to 500 mg/liter FAC. The results indicate that the type of chlorine oxidant is important for achieving the same levels of C. parvum oocyst inactivation. Figure Figure44 shows a comparative oocyst inactivation plot for hypochlorite and MOS applied at identical FAC concentrations.
One can see that sodium hypochlorite solution exhibits less biocidal efficacy against C. parvum oocysts—when applied at similar FAC concentrations—than fresh MOS generated by salt brine electrolysis. In addition, oocyst inactivation does not show a linear correlation with the FAC concentration: at lower FAC concentrations there is an initial “lag phase” in which increasing FAC concentrations do not seem to significantly affect oocyst viability. This might be attributed to the passive protective action of the robust oocyst wall as well as to active protection by the oxidant-induced stress response.
Infectivity is the critical parameter for the characterization of oocyst viability. Oocysts were subjected to different concentrations of hypochlorite and MOS for 4 h and were then used to infect HCT-8 human ileocecal adenocarcinoma cells grown in monolayer cultures. Dilutions of the oocysts were applied, and the infectious foci were counted under an epifluorescence microscope. The results (Fig. (Fig.4B)4B) show a trend in oocyst inactivation similar to what has been observed by the qRT-PCR technique. Namely, a higher FAC concentration is needed when hypochlorite solutions are used to achieve the same level of inactivation that can be observed at lower FAC concentrations with MOS.
It should be noted that counts of infection foci on the HCT-8 cells from the initial 105 oocyst samples (as the untreated control and its subsequent dilutions) consistently yielded infection focus numbers corresponding to 3.0 × 104 to 3.3 × 104 live oocysts instead of 105. Therefore, to circumvent potential underestimation of the results, the normalized 100% value for tissue culture infectivity, which was compared with the induced Hsp70 levels in 105 oocysts in postdisinfection viability tests, corresponded to 3.0 × 104 to 3.3 × 104 infective oocysts.
β-Tubulin mRNA was first reported by Widmer et al. (29) as a molecular marker of oocyst viability. The authors applied semiquantitative end product RT-PCR to monitor β-tubulin mRNA levels. Using a new set of PCR primers and an improved technique for RNA isolation from C. parvum oocysts, the β-tubulin mRNA was amplified from 105 oocysts treated with MOS. The aim of this experiment was to follow the time course of oocyst inactivation by measuring the level of heat-induced Hsp70 mRNA and to monitor the level of β-tubulin mRNA at the same time—once again in a quest to find a rapidly quantifiable molecular marker for oocyst viability. Figure Figure55 shows that even when oocyst inactivation is in the 2- to 3-log range, i.e., when only ~0.1% of oocysts responds to heat shock, the levels of β-tubulin mRNA and of the 18S rRNA are still 60 to 70% of that of untreated oocysts.
To our knowledge, no published work has yet elucidated the specific and general damages caused by exposure of protozoan oocysts (or cysts) to free chlorine (as OCl− or HOCl) or electrochemically generated hypochlorite solution containing additional mixed oxidant components with potent biocidal properties, referred to as mixed oxidant solution (MOS). Therefore, it would be also difficult to differentiate between the primary, prosurvival, and cell defense responses at cellular and molecular levels and the secondary (biocidal) effects of chlorine-based oxidants on protozoa, especially at their oocyst or cyst stage. Most of the published work is limited to quantitative assessment of disinfection treatment of protozoan cysts and oocysts (4, 6, 12, 17, 19, 22, 25).
During the course of this study, we discovered that induced Hsp70 gene expression is also a primary biological response of the C. parvum oocysts to the oxidative stress caused by chlorine-based disinfectants. The observation that a chaperone response is involved in the primary reactions to oxidative stress is not surprising. By analogy, the mechanisms, which are a part of defenses against heat stress and hydrogen peroxide in bacteria, are also involved in free-chlorine resistance, indicating a possible overlap in the defense circuits (9, 10). The bactericidal effect of free chlorine involves, at least in part, the action of hydroxyl radicals generated by a Fenton-type reaction (18), and some of the more reactive hypochlorous acid targets seem to be membrane associated (28). In these bacterial defense mechanisms, the HSp33 redox-regulated chaperone plays an ultimate key role by protecting the cells from oxidative protein unfolding (30).
The results reported here point to the importance of further investigating the molecular mechanisms of responses to oxidative stress in protozoa. The analysis of the C. parvum and C. hominis genome sequences (1, 32) indicated a number of genes involved in the heat shock response. The list includes a 40-kDa heat shock protein, heat shock protein DnaJ (pfj4), heat shock protein DnaJ (Pfj2), organellar dnaK-type molecular chaperone Hsp70, heat shock protein 83, the heat shock 105-kDa family, HSp40-like heat shock protein, Hsp60, heat repeat-domain protein heat shock protein 70 (Hsp70 precursor), and Hsp 90. Other stress gene products, such as DNAJ, Pfj1, stress-induced sti1-like protein, and DnaJ, sense and respond to protein unfolding (33). Coexpression of selected, yet unidentified, genes in these groups may be induced by oxidative stress as well. To our knowledge, Hsp70 is the first such “stress” gene whose induced expression has been detected in C. parvum oocysts in response to oxidative stress.
As to other possible cellular defense mechanisms, researchers have investigated UV-induced DNA damage in C. parvum oocysts, but no evidence was found that DNA repair would be capable of restoring oocyst infectivity (2, 20, 21, 23).
When the higher oxidant concentration and extended exposure time “overwhelm” the cellular defense, the exposure to the oxidant eventually leads to oocyst death. Our efforts initially focused on developing fast molecular diagnostic assays to help quantify the biocidal efficacy and understand the mechanisms of the biocidal effects of different types of chlorine-based oxidants against waterborne pathogens in general and C. parvum oocysts in particular. In MOS, chlorine as hypochlorous acid or hypochlorite ion is the predominant component. However, the chemical and biocidal behavior of MOS, coupled with potentials measured at the anode of the electrolytic cell, suggests the presence of other oxidants, including reactive oxygen species and hydrogen peroxide, as have been shown in solutions of nonchloride brines electrolyzed under similar conditions (8, 14). These additional oxidant species in MOS are thought to be responsible for enhanced biocidal efficacy. The results reported here confirm the trend observed by a number of authors in the differences between the biocidal efficacies of hypochlorite and MOS against C. parvum oocysts (6, 22, 27), namely, that the MOS generated by salt brine electrolysis exhibits a stronger biocidal efficacy against the oocysts than sodium hypochlorite when applied at identical FAC concentration.
A number of molecular markers have been investigated in this study for their applicability as oocyst viability indicators. Neither a randomly picked β-tubulin mRNA nor the 18S rRNA was able to adequately reflect the actual viability of oocysts within hours or even days during and after treatment with oxidative disinfectants. The levels of these indicator molecules were relatively high even in oocysts, which were shown to contain less than 1% of the infective or heat shock-responsive oocysts (Fig. (Fig.5).5). It appears that upon treatment with chlorine-based disinfectants, the integrity of C. parvum oocysts is retained for hours and perhaps even days postdisinfection. These results may implicate that some of the currently available viability indicator assays, such as dye permeation (4), might be inappropriate for assessing inactivation of protozoan oocysts by common chemical disinfectant treatment, such as chlorination, chlorine dioxide, hydrogen peroxide, ozone, and UV light.
qRT-PCR, as a viability assay, has been shown to linearly correlate with the number of live oocysts. Cai et al. (5) applied 18S rRNA as the PCR amplification template in their studies on drug efficacy against oocytes. These authors concluded that the first-order decay rate of 18S rRNA in dead oocysts was 99% in ~3 h, which was significantly faster than what we observed. These authors applied repeated freeze-thaw cycles to obtain dead oocysts and sporozoites for the control assays of RNA decay. This approach leads to physical destruction of the oocysts, which, in turn, facilitates exposure of the cellular RNA to nucleases and, therefore, faster RNA decay. On the contrary, in our experiments, the oocysts were left intact after the disinfection treatment, and the oocyst lysis with Tri-Reagent RT, containing the chaotropic agent guanidium isothiocyanate, provided immediate protection against nucleases.
In earlier published work (15, 16, 26) end product RT-PCR was used to amplify the heat-induced Hsp70 mRNA from C. parvum oocysts recovered from environmental water samples. The DNA amplification products were then separated by gel electrophoresis, and DNA bands of characteristic molecular mass indicated the presence of live (heat shock-responsive) oocysts in the water sample. Measuring induced gene expression by the more sensitive and accurate qRT-PCR, as we report here, may offer a molecular diagnostic tool for rapid quantitative assessment of viable/nonviable protozoan oocyst counts, which can be further developed into a rapid field-applicable molecular diagnostic technology. However, the apparent interference between the stress response of the oocysts to the chlorine-based oxidant and the qRT-PCR assay for Hsp70 mRNA quantification may suggest that induced gene products which are not involved in cellular stress responses might prove more suitable for such a viability assay.
This study was supported in part by a grant (N68711-05-C-0065) from the U.S. Department of Homeland Security.
We are grateful to Craig Milroy, Kevin Schwartz, Matt Santillanes, and Justin Sanchez for their help with mixed oxidant generation and to Susan Rivera, Andrew Boal, and Wesley Bradford for reviewing the manuscript and for the insightful discussions throughout the project. We also thank the anonymous reviewers for their valuable comments.
Published ahead of print on 29 January 2010.