Heterogeneity in susceptibility of viruses with different structures to various reactive oxygen species: Kinetics and biological mechanisms

Abstract

Waterborne viruses have caused outbreaks of related diseases and threaten human health, and advanced oxidation processes (AOPs), as clean and efficient technologies, have received widespread attention for their excellent performance in inactivating viruses. However, heterogeneity in susceptibility of structurally distinct viruses to various reactive oxygen species (ROS) is unclear. This study first measured the heterogeneity in inactivation kinetics and biological mechanisms of four typical viral surrogates (MS2, phi6, phix174, and T4) to various ROS by visible light catalysis. Notably, the second-order inactivation rate constants of four viruses by hydroxyl radicals (·OH), singlet oxygen (¹O₂), and superoxide radicals (·O₂⁻) were quite different: 10⁹–10¹⁰, 10⁷–10⁸, and about 10⁵ ​M⁻¹ ​s⁻¹, respectively. The susceptibility of four viruses to ROS varied significantly, in the order of phi6 ​> ​MS2 ​> ​phix174 ​> ​T4. More importantly, ¹O₂ can better oxidize capsid proteins. ·O₂⁻-induced RNA damage was significantly greater than that to the DNA genome, indicating that RNA viruses are more susceptible. ·OH can strongly inactivate the four structurally distinct viruses. Furthermore, the resistance of the ROS-inactivated virus to environmental interference was assessed in detail. This study advanced the understanding of heterogeneity in susceptibility of structurally distinct viruses to various ROS and provided a valuable theoretical basis for the application of AOPs in water disinfection.

Graphical abstract

Highlight

• •Heterogeneity in susceptibility of four viruses to various ROS was significant.
• •The second-order rate constants of virus inactivation by ROS were determined.
• •OH is universal for the inactivation of viruses with different structures.
• •¹O₂ can precisely and efficiently target and destroy the viral proteins.
• •Resistance of ROS-inactivated virus to environmental interference: ¹O₂ ​> ​·O₂⁻ >·OH.

1. Introduction

Humans around the world remain at risk of waterborne diseases, which are linked to waterborne pathogens [1,2]. Among these pathogens, viruses are more difficult to inactivate due to their unique biological structure and physiological characteristics, posing a great threat to human health [3]. Efficient waterborne virus inactivation is urgent, and advanced oxidation processes have received widespread attention in recent years as powerful and effective technologies for water disinfection.

Advanced oxidation processes (AOPs) can effectively inactivate viruses by reactive oxygen species (ROS), such as hydroxyl radicals (·OH), superoxide radicals (·O₂⁻), and singlet oxygen (¹O₂), which oxidize and destroy the structures and components of viruses [4]. Previous studies have established a mechanistic model of ROS-mediated virus inactivation [5], based on which the present study quantitatively assesses the contribution of various ROS to virus inactivation. In addition, the susceptibility of viruses to ROS is fundamentally dependent on viral biological components, including genome, lipid, and protein [6]. DNA (ssDNA and dsDNA) and dsRNA viruses have been reported to be more resistant than ssRNA viruses [7,8], but the susceptibility of viral genome types to different ROS is unknown. Enveloped and non-enveloped viruses also differ in their reactivity to ROS-induced inactivation because enveloped viruses contain an extra lipid bilayer [9,10]. Moreover, viral capsid proteins are targets of some inactivating factors [11], and their different reactivity may lead to significant heterogeneity in susceptibility of different viruses. The heterogeneous susceptibility of virus structure and components to different ROS remains to be clarified.

The generation and utilization of ROS are the foundation for virus inactivation by AOPs, and how specific ROS are generated is key to exploring the heterogeneous mechanisms of virus susceptibility. Ultraviolet (UV)/oxidant systems are common water disinfection techniques, but both UV and oxidants themselves can inactivate viruses [12,13]. Traditional Fenton processes have been carried out at low acidic pH to obtain greater efficiency, but the involvement of Fe²⁺ and H₂O₂ makes it impossible to consider the inactivation of ROS alone [14]. In ozone-mediated virus inactivation, the reaction kinetics of ozone are rapid, and the concentration of ·OH produced is relatively low, so the role of ROS in virus inactivation is often overlooked [15]. The above common homogeneous systems focus on the establishment and efficiency of catalytic systems [16], but the mechanism by which ROS causes virus inactivation is unclear. In heterogeneous AOPs, visible light catalytic technology has promising applications in environmental remediation [17]. By precisely selecting photocatalysts with distinct band structures [18], this approach can realize the establishment of a ROS system dominated by ·O₂⁻, ·OH, and ¹O₂, and visible light alone cannot inactivate viruses [19]. Based on this, three nanomaterials, g-C₃N₄, nano TiO₂, and C₆₀, with clear photocatalytic mechanisms, low cost, and easy large-scale preparation, were selected as model catalysts for generating specific ROS (Table S1).

In addition, the inactivation performance in natural water backgrounds is inhibited during advanced oxidation processes compared to clean matrices [[20], [21], [22]]. However, the efficiency of virus inactivation by ROS during AOPs is affected by the complexity of the actual wastewater sources, fluctuations in water quality and quantity, and variations in pre-disinfection treatment processes [23]. Studies on the resistance of ROS-inactivated viruses to environmental interference would provide a theoretical basis for ensuring the effective inactivation of viruses in different complex water matrices.

This study is the first to propose the heterogeneity in susceptibility of structurally distinct viruses to various ROS by exploring the kinetics and biological mechanisms of virus inactivation. We selected four typical bacteriophages with different structures (MS2, phi6, phix174, and T4) as virus surrogates, and systematically evaluated the susceptibility of structurally distinct viruses by using three ROS (·O₂⁻, ·OH, and ¹O₂ generated by g-C₃N₄, nano TiO₂, and C₆₀, respectively, in visible light catalysis systems). The study aimed to (1) explore the inactivation kinetics of structurally distinct viruses to obtain their second-order rate constants; (2) reveal the biological mechanism of the heterogeneity of inactivation efficiency of structurally distinct viruses; (3) assess the resistance of virus inactivation by various ROS to environmental interference.

2. Materials and methods

2.1. Chemicals and materials

Details of all chemicals and materials are listed in Text S1.

2.2. Virus propagation and plaque assay

Bacteriophages MS2, phi6, phix174, and T4 were selected as virus surrogates for this study. The characteristics of four model bacteriophages are provided in Table S2. Details are included in Text S2. Using published methods, four bacteriophages were propagated and enumerated [24]. In short, the bacterial host of phi6 was grown in Trypticase Soy Broth (TSB) medium at 26 ​°C, whereas that of MS2, phix174, and T4 was cultured in Luria–Bertani (LB) medium at 37 ​°C. The virus particles were resuspended in the Suspension Medium buffer (SM buffer) [25] and stored at 4 ​°C. Then, using the double-layer technique, bacteriophages were propagated and detected within the host.

2.3. Visible light photocatalytic inactivation experiments

Virus inactivation experiment was carried out in a 50 ​mL photochemical reactor with a cold trap (Fig. S1). Photocatalysts (i.e., g-C₃N₄, nanoscale TiO_2, and C₆₀ powder) were added to various water matrices (i.e., SM buffer, tap water, and secondary clarifier effluent) at a concentration of 0.5 ​g/L. The suspensions were magnetically stirred (800 ​rpm) for 10 ​min to ensure uniform dispersion of the catalyst. Virus stock solutions were subsequently introduced into the pre-dispersed catalyst suspensions to achieve an initial concentration of 1 ​× ​10⁷ ​PFU/mL. A 300 ​W xenon lamp with a cutoff filter (420 ​nm ​≤ ​λ ​≤ ​780 ​nm) was used to provide the light source. The light intensity was recorded by an optical power meter (CEL-NP2000). The light intensity was fixed at 100 ​mW/cm² by modifying the electrical current. Details of experimental conditions parameters are listed in Table S3. The reaction suspension was stirred with a magnetic stirrer at a constant speed throughout the experiment, and the reaction temperature was controlled at room temperature (20 ​± ​2 ​°C). Before irradiation, the system was left in the dark under continuous stirring for 30 ​min to establish an adsorption–desorption equilibrium.

Viruses were sampled following 0, 30, 60, 90, 120, and 180 ​min of illumination. The photocatalyst was removed by filtration through a 0.22 ​μm PES membrane after extraction of 1 ​mL of reaction solution. Viral suspensions without photocatalytic materials completed the reaction under the same conditions and served as a light control. The photocatalyst was added to the solution in the dark to complete the reaction as a dark control. All experiments were performed in triplicate. Means and standard errors are presented herein.

2.4. Quenching experiments and ROS quantitative experiments

Previous experiments confirmed that g-C₃N₄, nano TiO₂, and C₆₀ could inactivate viruses under visible light irradiation with ·O₂⁻, ·OH, and ¹O₂ as dominant ROS, respectively (Table S4). This study was further verified by quenching experiments and ROS quantitative experiments.

To ascertain the contribution of various ROS to virus inactivation during photocatalysis, quenching experiments were carried out. ·O₂⁻, ·OH, and ¹O₂ were scavenged by the addition of 100 ​mmol/L of 4-Hydroxy-TEMPO (Tempol), Isopropanol, and L-histidine to the system [26].

The decay kinetics of probe compounds can be calculated to determine the steady-state concentration of ROS. Using the second-order rate constants that were published, the concentrations of ·O₂⁻, ·OH, and ¹O₂ were measured using nitro-blue tetrazolium (NBT), nitrobenzene (NB), and furfuryl alcohol (FFA) as chemical probes [27,28].

2.5. Characterization and analysis

After negative staining with uranyl acetate, the samples were loaded onto copper grids and dried, and the morphology of the viruses before and after photocatalytic inactivation was observed by Transmission Electron Microscope (TEM, HITACHI HT7800). The total protein concentration of the virus suspension was measured by BCA protein assay kit (Beyotime, P0010S). The DNA and RNA concentrations were measured with a Nanodrop spectrophotometer (Thermo Nanodrop 2000, USA) after extraction using a DNA/RNA Viral Genome Extraction Kit (Solarbio Life Sciences, D2410). The level of malondialdehyde (MDA) was quantified using the MDA content assay kit (Beyotime, BC0020).

2.6. Kinetic modeling

The efficiency of structurally distinct viruses by photocatalytic inactivation (L) was estimated using Eq. 1. The inactivation rate constant of viruses was calculated using a quasi-first-order kinetic model (Eq. 2). The damage efficiency (E) was determined for viral DNA/RNA, proteins, and lipids, as shown in Eq. 3.

$$L={\log }_{10}\left(\frac{N_t}{N_0}\right)$$ (1)

$$\ln \left(\frac{N_t}{N_0}\right)=-k\times t$$ (2)

$$E=\frac{C_t}{C_0}$$ (3)

The virus inactivation by photocatalysis can result from the simultaneous effects of multiple different ROS, expressed as Eq. 4 the corresponding rate constants can be shown through Eq. 5.

$$-\ln \left(\frac{N_t}{N_0}\right)=-{\ln \left(\frac{N_t}{N_0}\right)}_{·\text{OH}}-{\ln \left(\frac{N_t}{N_0}\right)}_{{{·\mathrm{O}}_2}^{-}}-{\ln \left(\frac{N_t}{N_0}\right)}_{{{}^1\mathrm{O}}_2}-{\ln \left(\frac{N_t}{N_0}\right)}_{\text{other}}$$ (4)

$$k=k_{·\text{OH}}+k_{{{·\mathrm{O}}_2}^{-}}+k_{{}^1{\mathrm{O}}_2}{+k}_{\text{other}}$$ (5)

The contribution of various ROS to virus inactivation (R_contribution) was calculated according to Eq. 6. The virus inactivation rate constants by ROS (k_ROS) were corrected by Eq. 7.

$${\mathrm{R}}_{\text{contribution}}=\left(1-\frac{k_{\text{scavenger}}}{k}\right)\times 100\%$$ (6)

$$k_{\text{ROS}}={\mathrm{R}}_{\text{contribution}}\times k$$ (7)

A time response curve was used to fit the virus inactivation response to various ROS. Using Eq. 8, the inactivation rate constant was calculated by fitting the linear part of the following model [5]:

$$\ln \left(\frac{N_t}{N_0}\right)=-k_{\text{virus}-\text{ROS}}{\int }_0^t\left[\text{ROS}\right]\left(t\right)\mathrm{d}t$$ (8)

where N_t and N₀ are the virus titer (PFU/mL) at times t and 0, respectively, and C_t and C₀ are the viral DNA/RNA and protein concentrations (ng/μL) at times t and 0, respectively. [ROS](t) is the time-dependent ROS concentration, and ${\int }_0^t\left[\text{ROS}\right]\left(t\right)\mathrm{d}t$ is the time-integrated concentration of ROS.

3. Results and discussion

3.1. Heterogeneity in inactivation efficiency of structurally distinct viruses

The light control confirmed that visible light irradiation alone (without photocatalysts) was not effective in inactivating the four viruses (Fig. 1). In addition, the four viruses could not be adsorbed and removed by adding only three photocatalysts without illumination (Fig. S2). g-C₃N₄, TiO_2, and C₆₀ effectively inactivated four structurally distinct viruses under visible light. For MS2, g-C₃N₄, TiO₂, and C₆₀ resulted in an inactivation of 3.38-log₁₀, 4.61-log₁₀, and 3.07-log₁₀ (Fig. 1a), respectively. For phi6, g-C₃N₄, TiO₂, and C₆₀ resulted in an inactivation of 3.67-log₁₀, 5.16-log₁₀, and 3.29-log₁₀ (Fig. 1b), respectively. For phix174, g-C₃N₄, TiO₂, and C₆₀ resulted in an inactivation of 0.97-log₁₀, 1.38-log₁₀, and 0.67-log₁₀ (Fig. 1c), respectively, which were higher than that of T4 (0.89-log₁₀, 1.19-log₁₀, and 0.39-log₁₀, respectively) (Fig. 1d). The inactivation rate constants were calculated using Eq. 1 (Table S5, Fig. S4). For MS2, phi6, phix174, and T4, the order of the k values was: k(TiO₂) ​> ​k(g-C₃N₄) ​> ​k(C₆₀). The main reason for the heterogeneity in inactivation kinetics was that different photocatalysts had different ROS that played a dominant role in the inactivation process. Moreover, the results indicated that the inactivation resistance of the four viruses was in the order of T4 ​> ​phix174 > MS2 ​> ​phi6. Viruses are complete, structurally specific infectious particles. Their ultimate inactivation efficiency depends largely on their overall structural characteristics. These structural characteristics influence the effectiveness of ROS attacks on key targets, resulting in the significant differences in inactivation efficiency observed between different viruses.

Fig. 1.

Heterogeneity in inactivation efficiency of structurally distinct viruses. Inactivation of (a) MS2, (b) phi6, (c) phix174, and (d) T4 by different photocatalysts in 180 ​min under visible light; (e) Contribution of ·O₂⁻, ·OH, ¹O₂, and other ROS to inactivation rate constants of structurally distinct viruses by different photocatalysts. Conditions: [Photocatalysts] ​= ​0.5 ​g/L, initial concentration of virus ​= ​1 ​× ​10⁷ ​PFU/mL, pH ​= ​7.0 ​± ​0.2, T ​= ​20 ​± ​2 ​°C, I ​= ​100 ​mW/cm² (λ ​> ​420 ​nm).

3.2. Identification and quantification of ROS

3.2.1. Identification of ROS in photocatalysis

Quenching experiments were used to identify the ROS involved in photocatalysis to evaluate the effects of different scavengers on viral inactivation, thus further revealing the inactivation mechanism (Fig. S5). The four viruses could not be inactivated by adding quenchers alone (Fig. S6).

The contribution of ·O₂⁻, ·OH, and ¹O₂ to viral inactivation by g-C₃N₄, TiO₂, and C₆₀ photocatalysts was quantified over 180 ​min (Figs. 1e and S7). Quenching experiments confirmed that ·O₂⁻ and ·OH dominated virus inactivation for g-C₃N₄ and TiO₂ photocatalysts, respectively. Besides, it was shown that ¹O₂ was responsible for eliminating viruses in C₆₀ photocatalysts. The experimental results obtained were consistent with their band structure and redox potentials [[29], [30], [31]]. Consequently, these three photocatalysts were selected as media for generating different ROS-dominated photocatalytic systems to explore the heterogeneity in susceptibility of structurally distinct viruses.

3.2.2. Second-order rate constants for virus inactivation by ROS

The inactivation efficiency of viruses by ROS was assessed using the second-order inactivation rate constants, which provided a more precise comparison of the results from different studies [31]. Three probe chemicals, nitro-blue tetrazolium (NBT), nitrobenzene (NB), and furfuryl alcohol (FFA), were measured for degradation to determine the steady-state concentrations of dominant ROS (·O₂⁻, ·OH, and ¹O₂) for the three photocatalytic systems (Tables S6 and S7, Fig. S8) [32]. Based on the calculated steady-state concentrations of three ROS, we quantified the inactivation of the four viruses using the Wolf-linear decay model (Eq. 8). Using this equation, combined with data, we calculated the second-order inactivation rate constants (Table 1, Fig. S9). These k values directly reflect how the structural differences between the four viruses and the intrinsic oxidative capacity of ROS affect inactivation efficiency. Below, the findings for each ROS are discussed in detail.

Table 1.

Second-order rate constants between viruses and reactive oxygen species (M⁻¹ ​s⁻¹).

	·O2−	·OH	1O2
kMS2-ROS	4.71 ​× ​105	2.97 ​× ​1010	1.78 ​× ​108
kphi6-ROS	5.31 ​× ​105	3.46 ​× ​1010	1.80 ​× ​108
kphi174-ROS	1.20 ​× ​105	7.40 ​× ​109	2.80 ​× ​107
kT4-ROS	1.17 ​× ​105	5.32 ​× ​109	1.85 ​× ​107

For experiments with ·O₂⁻, the second-order rate constants were 3–5 orders of magnitude smaller than those of the other two ROS (Table 1). This difference is not difficult to understand, as there is heterogeneity in susceptibility of the structural components of the virus to various ROS. Although ·O₂⁻ is not as reactive as ·OH, the relatively high stability of ·O₂⁻ allows it to search for target molecules for reactions over a longer period [33]. Moreover, comparing the inactivation rate constants of structurally distinct viruses, they are very close to each other, all within a factor of 4. This suggested that ·O₂⁻-induced oxidation has little difference in its effect on the inactivation mechanisms of viruses with different structures.

For ·OH, the second-order inactivation rate constant was 2 orders of magnitude larger than ¹O₂ and much higher than that of ·O₂⁻. ·OH, as an extremely reactive free radical, has a strong oxidizing capacity with an oxidation potential of 2.8 ​V [34,35]. This high reactivity allows ·OH to quickly destroy viral protein capsids and genomes, thereby achieving efficient virus inactivation [6]. For the four viruses, phi6 was the most easily inactivated, with a k_phi6−·OH of 3.46 ​× ​10¹⁰ ​M⁻¹ ​s⁻¹, followed by MS2, which had a k_MS2−·OH of 2.97 ​× ​10¹⁰ ​M⁻¹ ​s⁻¹. This result was anticipated because ·OH has a strong oxidizing ability, and phi6 and MS2 are easily oxidized. The k_MS2−·OH calculated here is close to the data reported by Mamane and Rosado [12,36]. k_{phix174−·OH} was about four times smaller than k_MS2−·OH. This corresponded well with the information reported by Sommer et al., who found that ionizing radiation producing ·OH inactivated MS2 six times more efficiently than phix174 [37]. Lastly, among the four viruses, the susceptibility of T4 to ·OH was lowest, with a k_T4−·OH of 5.32 ​× ​10⁹ ​M⁻¹ ​s⁻¹. In general, ·OH can strongly inactivate four structurally distinct viruses with broad-spectrum inactivation.

Although the reactivity of ¹O₂ is lower than that of ·OH, it has high selectivity and stability [37]. The lifetime of ¹O₂ is relatively long, which allows it sufficient time to react with the target molecule [38]. Compared with other viruses, ¹O₂ inactivated MS2 and phi6 significantly more efficiently, with corresponding second-order rate constants 6 and 10 times higher than those of phix174 and T4, respectively. This also suggested that structural features (i.e., envelope, protein, genome) modulate ¹O₂-mediated inactivation. While the heterogeneity in susceptibility to ¹O₂ was not as great as expected among viruses, this contrasts with the indiscriminate reactivity of ·OH and highlights the unique potential of ¹O₂ for targeted virus inactivation strategies.

3.3. Heterogeneity in biological mechanisms of structurally distinct viruses

3.3.1. Morphological changes of structurally distinct viruses

To discern the underlying cause of virus inactivation, the micro-morphology of four structurally distinct viruses before and after 180 ​min treatment using different ROS-dominant photocatalysts, (i.e., g-C₃N₄ with ·O₂⁻ as the dominant, TiO₂ with ·OH as the dominant, and C₆₀ with ¹O₂ as the dominant) was examined by TEM. MS2 before inactivation had an icosahedral shape and was approximately less than 30 ​nm in size (Fig. 2a1). After 180 ​min of irradiation with visible light, its shape was distorted, and the capsid protein was damaged (Fig. 2a2-4). The shape of phi6 was close to spherical, with a diameter of 75 ​nm, and exhibited a layered structure (Fig. 2b1). Under any ROS role, the phi6 morphology was modified, and the diameter became smaller (Fig. 2b2-4), probably due to the absence of the virus envelope. As shown in Fig. 2c1, the morphology of phix174 is intact before inactivation, with a diameter slightly larger than MS2 (Fig. 2a1). After the reaction for 180 ​min, the morphology of phix174 under TiO₂ inactivation dominated by ·OH changed significantly, the shape was distorted, and the capsid was broken completely (Fig. 2c3). In contrast, the ROS generated by the other two photocatalysts had weaker damage to phix174, with slightly distorted shapes and insignificant capsid rupturing (Fig. 2c2,4). Similarly, before inactivation, T4 was structurally intact, with head, tail, and tail fiber-like structures (Fig. 2d1). Under photocatalytic attack dominated by ·O₂⁻ and ¹O₂, the tail and head of the T4 bacteriophage separated, but its head caused no significant damage, which displayed an irregular shape (Fig. 2d2,4). However, T4 bacteriophage treated with ·OH not only experienced head-tail separation but also suffered damage to the head capsid protein (Fig. 2d3). Comparing the results, T4 has the strongest resistance to ROS attack among the four viruses.

Fig. 2.

Heterogeneity in morphological changes of structurally distinct viruses. TEM images of MS2, phix174, phi6, and T4: (a1–d1) before inactivation; after 180 ​min of inactivation in (a2–d2) ·O₂⁻-dominated g-C₃N₄, (a3–d3) ·OH-dominated TiO₂, and (a4–d4) ¹O₂-dominated C₆₀ photocatalytic systems. The experimental conditions are the same as those described in Fig. 1.

3.3.2. Protein damage of structurally distinct viruses

Since virus survival during inactivation depends on the susceptibility of viral biological components, including proteins, genomes, and lipids (enveloped viruses) [39,40], viruses with different structures show different inactivation resistance in photocatalysis [41]. Therefore, to better understand the biological mechanism of the four viruses, we measured the dynamic changes in total protein concentration, DNA/RNA concentration, and lipid responses of four structurally distinct viruses during the photocatalytic process using different ROS-dominant photocatalysts (i.e., g-C₃N₄ with ·O₂⁻ as the dominant, TiO₂ with ·OH as the dominant, and C₆₀ with ¹O₂ as the dominant) (Fig. 3). We also used this model (Eq. 7) to correct the damage of viral biological components by ROS.

Fig. 3.

Heterogeneity in the biological components of structurally distinct viruses. Protein and RNA/DNA damage efficiency of (a,e) MS2, (b,f) phi6, (c,g) phix174, and (d,h) T4; (i) Level of lipid peroxidation (MDA detection) when phi6 was inactivated by different photocatalysts. The experimental conditions are the same as those described in Fig. 1.

From Fig. 3a and c, the total protein concentration of MS2 and phix174 showed a similar decreasing trend, but further analysis revealed that the rate constant of total protein loss in MS2 was greater than that in phix174. The reason why MS2 protein was more easily damaged than phix174 might include the structure of a single protein type and a smaller number of subunits [42]. In contrast, the capsid of the phix174 contains multiple proteins, making its structure more complex and stable, and therefore may have better resistance to ROS [39]. In addition, the total protein quantity of phi6 and T4 only showed a slight decrease, and their reaction rate constants were also lower than those of the other two viruses (Fig. 3b–d). The outer layer of phi6 is composed of a triple-layered structure including a lipid membrane and polyhedral capsid, making its proteins difficult to degrade [41]. In addition, T4 has a more complex structure, with its capsid and tail composed of multiple different proteins, forming a sturdy outer shell, which makes T4 more resistant to inactivation [39]. On the other hand, the damage to the viral protein was also closely related to the type of ROS. ·OH acts as a non-selective oxidant, indiscriminately attacking any amino acid residue to induce broad-spectrum damage. Compared to the other two ROS, ¹O₂, with its microsecond-level lifetime, can fully diffuse to protein targets, while selectively oxidizing key residues such as tryptophan, tyrosine, and methionine [[43], [44], [45]], significantly enhancing damage efficiency. Therefore, ¹O₂ can precisely and efficiently target and destroy the viral proteins.

3.3.3. Genome damage of structurally distinct viruses

As shown in Fig. 3e and f, all three ROS caused damage to the RNA of MS2 and phi6, and the RNA concentrations of MS2 and phi6 showed a decreasing trend within 180 ​min, with ·OH being the most effective with k_{RNA degradation} of 3.35 ​× ​10⁻³ ​min⁻¹ and 3.03 ​× ​10⁻³ ​min⁻¹. Moreover, phi6 was more resistant than MS2, and the reaction kinetics of the double-stranded genome was slower than the single-stranded genome (Fig. 4a and b). This was because the base groups in the single-stranded genome are more easily oxidized than those in the double-stranded genome. In addition, phix174 exhibited stronger resistance compared to RNA viruses because the genome of phix174 is a single-stranded circular DNA molecule (Fig. 3g), whereas the genomes of MS2 and phi6 are single-stranded RNA and double-stranded RNA molecules, respectively. DNA is deoxyribonucleic acid, whereas RNA is ribonucleic acid. Deoxyribose lacks a hydroxyl (-OH) group, which makes the sugar-phosphate backbone of DNA more stable and less prone to reacting. As shown in Fig. 4c, ·O₂⁻ was unable to damage the DNA of phix174, but ·OH and ¹O₂ caused some damage, with k_{DNA degradation} of 2.13 ​× ​10⁻³ ​min⁻¹ and 0.66 ​× ​10⁻³ ​min⁻¹, respectively. Neither ·O₂⁻ nor ¹O₂ oxidation could damage the genome of T4, and even with ·OH, the damage was not very effective, with k_{DNA degradation} being 0.14 ​× ​10⁻³ ​min⁻¹ (Fig. 3, Fig. 4d). The genome of T4 is a large double-stranded DNA molecule, approximately 1.66 ​× ​10³ kbp in length, encoding over 250 proteins [46]. Therefore, the external structure and the double-stranded DNA genomic structure made T4 more resistant to attack by ROS.

Fig. 4.

Heterogeneity in biological mechanisms of structurally distinct viruses inactivation by various ROS. Degradation of (a) MS2, (b) phi6, (c) phix174, and (d) T4 proteins, genomes, lipid, and inactivation by ROS; (e) Proposed heterogeneous regularity in susceptibility of structurally distinct viruses to various ROS.

3.3.4. Lipid damage of enveloped virus

The viral envelopes are composed of lipid bilayers, and ROS generated by photocatalysis are more likely to damage the envelope of phi6 [24]. In the photocatalytic system, all three ROS can cause lipid peroxidation (Fig. 3i). We examined the MDA levels of different ROS oxidizing phi6. ·OH could induce viral lipid peroxidation by increasing the MDA level in liposomes to 6.71 ​nmol/mL ¹O₂ led to increased MDA levels to 5.17 ​nmol/mL, and ·O₂⁻-induced lipid peroxidation to 3.65 ​nmol/mL.

3.3.5. ROS-structure susceptibility in virus inactivation

The heterogeneity in the susceptibility of structurally distinct viruses to various ROS depends on the structure and genome type of virus, as well as the oxidative capacity of ROS. To determine the importance of biological component damage to virus inactivation, we compared the relative contributions of protein damage, genome damage, and lipid damage to the second-order rate constants of the four viruses by ROS (Fig. 4e).

Differences in the composition, morphology, and stability of the protein capsid affect the susceptibility of viruses to ROS [47,48]. The more complex the protein structure, the stronger the viral resistance to ROS. The genome type of the virus also affects its susceptibility to ROS [49]. RNA is more susceptible to photocatalytic effects than DNA, and single-stranded genomes are more fragile than double-stranded genomes. There is a general trend of virus susceptibility in genomic type: ssRNA ​> ​dsRNA ​> ​ssDNA ​> ​dsDNA. For enveloped viruses, the lipid bilayer is rich in unsaturated fatty acids, which are highly susceptible to peroxidation induced by ROS [50,51]. Therefore, enveloped viruses are more easily inactivated by ROS than non-enveloped viruses.

ROS exhibits significant variations in oxidative capacity, which consequently affects the disruption efficiency of viral structural components. The viral genome was the primary target of ·OH attacks, rather than the overall structure and components of viral particles. The strong oxidant ·OH can cause broad-spectrum damage to the genomes of RNA and DNA viruses, leading to inactivation. In addition, the lipid envelope of enveloped viruses is more susceptible to ·OH. ·O₂⁻-induced inactivation of RNA viruses (MS2 and phi6) mainly through RNA degradation, but had little effect on the genomes of DNA viruses (phix174 and T4). Moreover, ¹O₂ was more reactive with viral proteins than with the genome, and can achieve precise and efficient specific destruction of viral proteins [52]. This ROS-structure interaction explains why phi6 ​> ​MS2 ​> ​phix174 > T4 have different susceptibilities, emphasizing that predictive inactivation models must integrate damage to the protein, lipid, and genome, rather than relying only on total ROS production.

3.4. Resistance of various ROS-inactivated viruses to environmental interference

The photocatalytic process is significantly impacted by actual environmental factors [52]. Tap water and secondary effluent water were chosen as the actual water to further examine the inactivation efficiency of viruses in real water environments. The effects of virus inactivation using the three photocatalysts in the face of environmental interference are shown in Figs. S11-S16, corrected for Eq. 7, to obtain the effects of ROS on virus inactivation. As shown in Fig. 5a–d, the inactivation rate constants of the four viruses were decreased in the real water. Real water has a more complex background than SM buffer because it contains higher concentrations of metal ions, dissolved organic matter (DOM), and other substances that consume oxidants and reduce virus inactivation efficiency [53]. Moreover, the photocatalytic effect is also influenced by the pH value [52]. The inactivation rate constants decreased as the solution increased, suggesting that acidic conditions were better for the photocatalytic inactivation of viruses [54]. The photocatalytic inactivation efficiency of viruses was considerably decreased when the water contained organic matter. The effects of humus (humic acid) were more pronounced than those of sucrose and tryptone [23]. The high concentration of humic acid significantly inhibited virus inactivation. It seemed reasonable to interpret that humic acid was an efficient scavenger for ROS [55,56]. As shown in Fig. 5e, among the three ROS, the most crucial is ·OH, which possesses a strong oxidation capacity with an oxidation potential of 2.8 ​V [57]. It can efficiently degrade most viruses, attracting widespread attention. However, because of its high reactivity, ·OH exists only for a very short period, and once formed, it reacts rapidly with the surrounding environment, so it is subject to large external influences and poor stability [58]. Compared with ·OH and ·O₂⁻, ¹O₂ has the advantages of a long lifetime, wide pH tolerance, strong resistance to interference by environmental media, and excellent selectivity, which makes it more suitable for virus disinfection in actual water environments [59].

Fig. 5.

Resistance of various ROS-inactivated viruses to environmental interference. Inactivation rate constants of (a) MS2, (b) phi6, (c) phix174, and (d) T4 by various ROS at different influencing factors; (e) Different matrix effects on the photocatalytic inactivation of the four viruses by various ROS (·O₂⁻, ·OH, and ¹O₂ were generated by g-C₃N₄, TiO₂, and C₆₀, respectively, and were corrected by Eq. (7)). Conditions: [Sucrose] ​= ​[Tryptone] ​= ​50 ​mg/L, [Humic acid] ​= ​10 ​mg/L, other conditions are the same as those described in Fig. 1.

4. Conclusion

This study is the first to propose the heterogeneity in susceptibility of structurally distinct viruses to various ROS by exploring the kinetics and biological mechanisms of virus inactivation. The second-order inactivation rate constants of four viruses by ·OH, ¹O₂, and ·O₂⁻ were quite different: 10⁹–10¹⁰, 10⁷–10⁸, and about 10⁵ ​M⁻¹ ​s⁻¹, respectively. The susceptibility of four viruses to ·OH, ¹O₂, and ·O₂⁻ varied significantly in the order of phi6 ​> ​MS2 ​> ​phix174 > T4, which was attributed to the susceptibility of the viral structures and biological components to ROS. The external structure and dsDNA genomic structure make T4 more resistant to ROS attack, whereas phi6 is more susceptible to inactivation due to the presence of the envelope. In addition, the biological mechanism of virus inactivation was elucidated by comparing the relative contribution of damage to viral biological components to inactivation by ROS. ¹O₂ better oxidized capsid proteins. ·O₂⁻-induced RNA damage was significantly greater than that to DNA genome, indicating that RNA virus is more susceptible to ·O₂⁻. ·OH universally degraded both structural and genomic components of viruses. Moreover, complex water matrix characteristics largely affected the virus inactivation by ROS. Acidic conditions enhanced virus inactivation by stabilizing ROS, while dissolved organic matter quenched ·OH and ·O₂⁻ more effectively than ¹O₂. This stability places ¹O₂-mediated advanced oxidation processes as a reliable candidate for real-world water disinfection. This study advanced the understanding of heterogeneity mechanism in susceptibility of structurally distinct viruses to various ROS and provided a valuable theoretical basis for the application of advanced oxidation technology in water disinfection.

CRediT authorship contribution statement

Zihan Fu: Writing – original draft, Formal analysis, Conceptualization, Methodology, Data curation. Pingfeng Yu: Writing – review & editing, Supervision, Validation, Investigation. Mengyao Wang: Visualization, Investigation, Methodology, Formal analysis. Lu Zhang: Methodology, Validation, Formal analysis. Ying Li: Investigation, Validation. Cong Lyu: Validation, Resources, Funding acquisition, Writing – review & editing, Supervision, Project administration.

Declaration of competing interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article: