Dr. Thérèse Sigstam's PhD public defense on May 23th, 2014

Abstract

Waterborne viruses are responsible for major outbreaks of diarrhea and other diseases throughout the world. Efficient virus inactivation during water and wastewater treatment is therefore important to prevent the contamination of water resources by inadequately treated wastewater, and to ensure the microbial safety of drinking and recreational water. This thesis discusses the kinetics and mechanisms of virus inactivation in commonly used homogeneous (chlorine dioxide, chlorine, UV and singlet oxygen) and heterogeneous (copper surfaces) disinfection systems, in order to obtain a better understanding of these processes.

In order to compare the efficiency of different disinfection treatments within a wide variety of viruses or among different disinfectants, the Chick-Watson first-order model is frequently used. It allows determination of virus inactivation rate constants as a function of disinfectant dose. However, virus inactivation by chlorine dioxide shows a deviation from this first-order model, namely a tailing curve. The mechanisms underlying this deviation are currently not understood. Tailing has been previously reported, but is typically attributed to the decay in disinfectant concentration. However, present results showed that tailing occurs even at constant ClO2 concentrations. Four working hypothesis to explain the cause of tailing were tested, specifically changes in the solution’s disinfecting capacity, aggregation of viruses, resistant virus subpopulations, and changes in the virus properties during disinfection. In experiments using MS2 as a model virus, it was possible to rule out the solution’s disinfecting capacity, virus aggregation and the resistant subpopulation as reasons for tailing. Instead, the cause for tailing is the deposition of an adduct onto the virus capsid over the course of the experiment, which protects the viruses. This adduct could easily be removed by washing, which restored the susceptibility of the viruses to ClO2. This finding highlights an important shortcoming of ClO2, namely its self-limiting effect on virus disinfection. It is important to take this effect into account in treatment applications to ensure that the water is sufficiently disinfected before human consumption.

Besides its deviation from first-order kinetics, ClO2 also exhibits drastically different inactivation kinetics for different viruses, even if viruses are closely related. These differences can only be rationalized if the underlying disinfection mechanisms are understood. Herein, we therefore determined how small differences in the composition of the viral genome and proteins impact the disinfection kinetics and mechanisms of ClO2 and other disinfectants. To this end, we investigated the inactivation of three related bacteriophages (MS2, fr, and GA) by UV254, singlet oxygen (1O2), free chlorine (FC), and ClO2. Genome damage was quantified by q-PCR, and protein damage was assessed by quantitative matrix-assisted laser desorption ionization (MALDI) mass spectrometry. ClO2 caused great variability in the inactivation kinetics between viruses and was the only treatment that did not induce genome damage. In contrast, the inactivation kinetics were similar for all viruses when treated with disinfectants possessing a genome-damaging component (FC, 1O2, and UV254). On the protein level, UV254 subtly damaged MS2 and fr capsid proteins, whereas GA’s capsid remained intact. Singlet oxygen oxidized a methionine residue in MS2 but did not affect the other two viruses. In contrast, FC and ClO2 rapidly degraded the capsid proteins of all three viruses. Protein composition alone could not explain the observed degradation trends; instead, molecular dynamics simulations indicated that degradation is dictated by the solvent-accessible surface area of individual amino acids. Finally, despite the similarities of the three viruses investigated, their mode of inactivation by a single disinfectant varied. This explains why closely related viruses can exhibit drastically different inactivation kinetics.

Compared to the homogeneous disinfection methods mentioned above, even less is known regarding the mode of action of heterogeneous disinfectants. In heterogeneous systems, disinfection is mediated by virus-surface interactions. A prominent example of a heterogeneous disinfection system is the copper jar, which is widely used in Indian homes to store water, and which has been shown to possess antiviral activity. However, it is not understood which virus-copper interactions lead to disinfection, or how the viruses are affected by these interactions. To better understand this system, adsorption-inactivation experiments were performed with two bacteriophages, MS2 and Qbeta, on metallic copper. Hereby, the contributions of both the metallic copper and the dissolved copper leached from the solid surface were investigated. MS2 was found to be inactivated by dissolved copper only, whereas the metallic copper had an important role in the inactivation of Qbeta. To shed light on which virus-surface interactions cause inactivation, adsorption-inactivation processes were studied on self-assembled monolayers (SAMs) which allow studying the effect of a single type of interaction at a time. The two viruses were found to differ in their adsorption behavior, but not their inactivation trends: hydrophobic and hydrogen-bonding surfaces led to efficient adsorption and inactivation of both viruses; positively charged surfaces led to adsorption via electrostatic attraction, but not to inactivation; and finally, pure gold surfaces (exhibiting van der Waals interactions) adsorbed Qbeta more efficiently than MS2, but did not inactivate either virus. Based on these results, it was suggested that the different inactivation behavior of the two viruses observed in presence of metallic copper was a result of their differing adsorption, rather than inactivation tendencies. The stronger surface interactions of Qbeta may furthermore be rationalized by the presence of disulfide bridges in its capsid protein, which can form thiolate bonds with the copper surface. Then concurrent reduction of the disulfide bridges may cause important distortions of Qbeta’s capsid’s conformation, or may result in its disintegration.

Overall this thesis has shed new light on the mechanisms involved in virus inactivation and on how to relate them to inactivation predictions of nonculturable viruses. The use of surrogates to this purpose showed to be reliable for genome-active disinfection treatments but care should be taken when predicting inactivation by protein-active treatments. This mechanistic insight into viral inactivation also allowed understanding the mechanisms causing tailing by protein-active disinfectants.

Keywords: Virus inactivation, chlorine dioxide, tailing effect, inactivation mechanisms by oxidants, genome degradation, protein degradation, virus-surface interaction, copper, self-assembled monolayers (SAMs)