A GI.1 norovirus positive fecal specimen was provided by Dr. Christine Moe at Emory University and a GII.4 norovirus-positive fecal specimen was obtained from a 2009 cruise ship outbreak of viral gastroenteritis. Bacteriophage MS2 (ATCC No. 15597-B1) was cultivated and assayed using Escherichia coli F_amp (ATCC No. 700891) following the previously reported protocol [29]. The titer of the purified MS2 stock was approximately 10¹⁰ PFU mL⁻¹. MNV (strain CW3), provided by Dr. Skip Virgin, Washington University School of Medicine (St Louis, Mo, USA), was propagated and assayed in RAW 264.7 cells (ATCC No. TIB-71, Manassas, VA). FCV strain F9 ATCC No.VR-782 was propagated and assayed in Crandell Reese Feline Kidney cells (CRfK ATCC No. CCL-94), as previously described [30]. The infectivity titers of MNV and FCV stocks were approximately10^8.0 and 10^8.9 PFU mL⁻¹, respectively. Virus stocks were stored in aliquots at −80 °C, and each aliquot was thawed and further purified by centrifugation and/or filtration for each experiment.

TiO₂ nanoparticles (P25, Degussa Co., Germany) were added to 500 g L⁻¹ suspension of polyethylene glycol (PEG, 500 g L⁻¹) at 10:1 w/w ratio and mixed for 30 min using a mortar and pestle. The resulting coating reagent was spread onto a glass plate, dried at room temperature, and calcined at 450 °C for 30 min to mineralize PEG. Surface-fluorinated TiO₂ was prepared by immersing the TiO₂-coated glass plate in a 30 mM NaF solution (pH 3.5, adjusted with HCl) for 30 min and then air-drying [24].

Virus-contaminated surface coupons were prepared by depositing 25 μL of pooled (MNV/FCV) or each individual virus suspension (MS2, GI.1 or GII.4) on TiO₂ or F-TiO₂ films, and dried for 1.5 h at room temperature under dark. The sample surface was then exposed to a commercial fluorescent lamp (4 W, GE Co., USA). The light intensity was controlled by adjusting the distance between the lamp source and the sample and measured at a representative UVA wavelength of 365 nm using a UVX radiometer, equipped with a 365 nm UVA detector (UVP Co., Upland, CA). Visible range experiments were conducted by using UV cut-off filter (Optivex^® UV glass filters, Pegasus lighting Co., USA) which blocks irradiation of UV fractions in lamp emission spectrum (<400 nm). After a predetermined time (up to 120 min), the coupons were vortexed for 1 min in 2 mL elution buffer (5% fetal bovine serum and 0.05% Tween 80^® in distilled water). Eluted viruses were plaque-assayed (MS2, MNV, and FCV) or the RNA copy number was quantified (MS2, GI.1, and GII.4). To evaluate oxidative damage of the capsid proteins during virus inactivation, the concentration of oxidized proteins was assayed using the OxyELISA oxidized protein quantitation kit (Millipore Co.), following the manufacturer’s instructions. Protein concentration was standardized to 10 lg mL⁻¹ by concentrating the harvested viruses using a stirred cell equipped with 1500 MWCO UF membrane (Amicon^® Ultra, Millipore Co., Billerica, MA), followed by a Bradford quantification. Control experiments were performed using uncoated glass plates with light and coated glass without light. All experimental values reported herein were from averaging at least four replicate analyses of two independent experiments. Significant differences in viral reductions between different test conditions were assessed using the Student’s t-test.

Sample aliquots (100 μL) were incubated with 1 μL of RNase One^™ Ribonuclease (Promega, Madison, WI) for 1 h at 37 °C, after which the reaction was stopped by adding 1 μL of RNase inhibitor (Invitrogen, Carlsbad, CA). Viral RNAs were then extracted using the MagMAX^™−96 Viral RNA Isolation Kit (Ambion, Austin, TX) and the KingFisher^® magnetic particle processor, according to the manufacturer’s instructions. Oligonucleotide primers and probes for MS2, GI.1 norovirus, and GII.4 norovirus used in this study are listed in Table 1 [31–33]. Viral RNAs were quantified by Taq-Man-based real-time RT-PCR using the QuantiTect Probe RT-PCR kit (Qiagen, Valencia, CA) on an ABI 7500 platform (Applied Biosystems, Foster City, CA) [8]. Reduction of viral RNA copy numbers was determined by calculating log₁₀ (Ct_T − Ct₀)/k, where Ct_T is the threshold crossing value (Ct) of viral RNA of a treated sample; Ct₀ is the initial Ct-value of viral RNA of an untreated sample, and k is the slope of the linear regression for Ct-value versus the logarithm of the viral RNA copy numbers. Average slopes of the standard curves for MS2, GI.1, and GII.4 were −2.99 ± 0.10 (r² ⩾ 0.950), −3.9 ± 0.10 (r² ⩾ 0.969), and −2.96 ± 0.3 (r² ⩾ 0.928), respectively.

Fig. 1 shows the results of experiments performed to evaluate the effects of incident UVA intensity, fluoride concentration (i.e., in solution used for F-TiO₂ preparation), and temperature on MS2 inactivation by the F-TiO₂ film. Values of 10 μW cm⁻², 30 mM, and 24 °C, respectively, were chosen as the baseline conditions for these parameters throughout this study. As seen in Fig. 1a, MS2 was inactivated more rapidly at higher light intensity. When the F-TiO₂ coated coupons were irradiated with fluorescent light at 3.5 μW cm⁻² in the UVA range, 1 log₁₀ (90%) MS2 inactivation was achieved in approximately 42 min, whereas a similar level of inactivation was obtained after exposure for 15 min at 15 μW cm⁻². The reduction of MS2 infectivity thus correlated with the square root of UVA intensity (r² = 0.97; inset in Fig. 1a). Increasing the temperature from 12 °C to 32 °C resulted in a decrease in infectivity of MS2 on F-TiO₂ coated coupons by 1.5–2.6 log₁₀ PFU mL⁻¹ (Fig. 1b). Additionally, altering the fluoride concentration of the NaF precursor solution from 10 to 30 mM led to a higher reduction of MS2 titers (1.89–2.6 log₁₀ PFU mL⁻¹ after 60 min exposure). When a cutoff filter was used to remove UVA radiation, exposure to the fluorescent light for 60 min did not have any effect on MS2 infectivity (Fig. 1d).

The positive effects of UVA intensity on photocatalytic inactivation of MS2 (Fig. 1a) coupled with the result obtained using a UV filter (Fig. 1d) confirm that the viruses are affected by UVA-initiated reactions. The observation of faster inactivation at higher temperatures (Fig. 1b) was consistent with previous studies which have shown that higher temperatures result in an increased production of (and reactivity of) ^·OH [34]. Although the virucidal action of free radicals is not well documented in the literature, in bacterial inactivation, ^·OH is known to damage cell walls/membranes or cause oxidation of coenzymes, thereby killing the microorganism [19,34–36]. Similar oxidative damages on viruses by ^·OH are expected and, consistently, we observed that surface fluorination of the TiO₂ coatings enhanced the photocatalytic MS2 inactivation by 4 times and resulted in over 2-log₁₀ reduced infectivity of the three tested NoV-surrogate viruses after 60 min. According to previous studies, fluorination of TiO₂ is thought to accelerate desorption of ^·OH from the TiO₂ surface, which would help to eliminate the diffusion limitations associated with degradation at the catalyst surface. This difference is speculated to be especially beneficial for virus inactivation, wherein a relatively substantial exposure to oxidants is required to rapidly affect the viral infectivity. It is also worth noting the results of Janczyk et al., who reported evidence of singlet oxygen (¹O₂) production by irradiated F-TiO₂ [37]. While most literature has attributed photocatalytic enhancement upon surface fluorination to enhanced ^·OH generation, photocatalytically produced ¹O₂ might have also contributed to the observed enhanced kinetics, as it is known to react strongly with proteins and to effectively inactivate viruses [38].

Inactivation kinetics of MS2, MNV and FCV on an unmodified TiO₂ film is shown in Fig. 2a, along with the measured emission profile of the fluorescent lamp employed in this study. Analysis of the light-emission profile of fluorescent light showed that less than 5% (Relative light intensity: footnote of Fig. 2a) of fluorescent light had a wavelength of ⩽400 nm, attributed mainly to the mercury emission line at 365 nm. Exposure to fluorescence light for 80 min on TiO₂ film reduced infectivity titers of MS2, FCV, and MNV by 0.85, 0.72, and 0.53 log₁₀ PFU mL⁻¹, respectively (Fig. 2a). Identical experiments conducted using F-TiO₂ surfaces showed much faster kinetics, with an infectivity reduction of 2.6, 2.0, and 2.6 log₁₀ PFU mL⁻¹ for MS2, FCV and MNV, respectively, after 60 min (Fig. 2b). The results suggest that the F-TiO₂ coating exhibited approximately 4 times faster inactivation performance against NoV surrogates compared to the pristine TiO₂ coating. The extent of reduction of FCV at 60 min was 20% (as slope value in Fig. 2b) less than that of MNV and MS2 (p-values of 0.041 and 0.035, respectively).

When GII.1 and GII.4 were in contact with F-TiO₂ film for 120 min under fluorescent light irradiation, reductions in RNA titers for GI.1 and GII.4 were 0.0 and 1.1 log₁₀ PCR RNA copy number, respectively (Table 2). Similarly, even though the infectivity of MS2 on F-TiO₂ film decreased by 3.6 log₁₀ PFU mL⁻¹ after 120 min at 10 μW cm⁻² UVA, there were no significant reductions for MS2 RNA copy titer. In contrast, capsid proteins of MS2 and GI.1, which had been in contact with fluorescent light-activated F-TiO₂ coated coupons for 60 min, were oxidized by 64% and 59% of total proteins, respectively, as shown in Fig. 3. In control tests performed on fluorescent light-exposed cover glass (no TiO₂ coating), or on F-TiO₂ film in dark conditions, oxidized protein concentrations were measured to be 0.5% and 0.6% of total proteins for MS2 and GI.1, respectively.

The lack of reduction in viral RNA for MS2 and GI.1 suggests that infectivity loss was not directly related to RNA damage within our experimental time range. Rather, our observation that a > 59% of capsid proteins of these viruses were oxidized after 120 min indicates that the likely cause of inactivation is denaturation of capsid proteins, which are primarily responsible for the integrity of the virus and attachment to host cells. Consistently, Sano et al. reported that carbonyl-group formation, resulting from oxidation of capsid proteins by free chlorine, was correlated with the level of protein oxidation and possibly infectivity loss [39]. Overall, the broad spectrum of antiviral activity of activated F-TiO₂ film and the similarity in the extent of oxidation damage of capsid proteins between GI.1 and MS2 suggest that F-TiO₂ film is likely to result in rapid inactivation of human NoV as well via the photochemical production of ^·OH and ¹O₂.

Experiments were also performed to evaluate the virucidal activity of F-TiO₂ film in a realistic office setting, wherein the UVA intensity from typical room lighting at the samples, placed on top of an office room table, was 2.4 μW cm⁻². Initial infectious titer of coliphage MS2 was 8.2 log₁₀ PFU mL⁻¹. After 12 h of exposure, the infectivity of MS2 on F-TiO₂ coated coupons declined below a lower level of quantification of plaque assay (3.1 log₁₀ PFU) indicating inactivation great than 5.1 log₁₀ PFU, while coliphage MS2 on cover glasses (no TiO₂) remained significantly higher with only a reduction of 1.3 log₁₀ from initial titers.

In fact, common fluorescent bulbs frequently produce minor emissions at 365 nm, with peaks at 378 and 313 nm also present in some products. The photocatalytic surface inactivation results obtained in a realistic office setting further validate the potential for F-TiO₂ to prevent virus transmission in the indoor environment. However, antimicrobial efficacy is reliant upon the residual UVA emission commonly emitted by mercury-based fluorescent lamps. Other types of lighting that have no UVA emission lines, such as white light LED bulbs, are not expected to result in surface inactivation of viruses by F-TiO₂. Nonetheless, given that most commercial buildings presently employ fluorescent room lighting, the photocatalytic F-TiO₂ coatings demonstrated herein may offer a sustainable and effective means of reducing NoV transmission on fomites.