1. Introduction

Viruses can cause major threat to human health, especially the extremely dangerous species of RNA virus like HIV, H5N1, SARS and etc. More recently, the coronavirus disease 2019 (COVID-19) has caused a huge negative impact on all aspects of the modern human society [1]. Since its outbreak, the SARS-CoV-2 virus has undergone five mutations, namely Alpha, Beta, Gamma, Delta, and Omicron [2]. This virus spreads from person to person mainly via respiratory droplet transmission. It can also be transmitted through aerial droplets and contact [3]. In the process of diagnosis, treatment and research in hospitals and scientific institutions, contaminated instruments, blood and tissue samples can cause further infections to the personnel handling them. Therefore, such objects must be disinfected frequently. Disinfectant agents and UV-irradiation are most commonly used means to inactivate viruses [4].

Viruses are vulnerable to UV irradiation, especially the SARS-CoV-2 virus [5]. Thus, ultraviolet (UV) light has become a promising non-thermal technology employed for microbial inactivation [6,7]. Moreover, UV irradiation can effectively kill viruses that survive in the air or water [8], in an environmentally friendly and efficient manner. The UV light spectrum can be further divided into UVA (320-400nm), UVB (280-320nm) and UVC (200-280nm). The effects of different bands have been reported in various disinfection studies. For instance, the UVA light at high dosages was proven to be effective in inactivating many species of viruses and bacteria [6,9]. The photobiological effects of UVA are mainly attributed to indirect effects. The absorbed UVA energy can lead to reactive oxygen species (ROS), which can cause oxidation of not only purine and pyrimidine bases, but also membrane lipids. Moreover, UVA radiation can also lead to single-strand breaks, and transit cytosine to thymine, which cause DNA/RNA damage [10]. On the other hand, exposure to UVB light primarily leads to direct DNA/RNA damage. Absorption of UVB energy by DNA/RNA can break the covalent bonds between two adjacent pyrimidine (cytosine and thymine) bases in the DNA strand, and thus results in pyrimidine dimers [11]. UVC light can cause even more damages to the viral DNA or RNA, so that they cannot reproduce [12]. In comparison, UVA and UVB light have relatively limited germicidal capability, since viruses and bacteria have adapted for millions of years to them [13]. According to the International Ultraviolet Association, it is generally accepted that UVC light can kill almost “any pathogenic microorganism” [14]. Therefore, UVC is naturally the most effective wavelength in the battle against viruses.

However, the shorter the wavelength, the weaker the penetrating potential of the light into non-aerial media. This phenomenon weakens the effect of UVC light in disinfecting fluidic and solid objects. UVA light with relatively better penetration can compensate UVC in this aspect. Thus, coupling UVA and UVC can pair the germicidal UVC effect and the greater penetrating power of UVA, which makes the coupled UVC and UVA irradiation a promising disinfection solution. Some previous researches have demonstrated that the pair can achieve microbial reduction in waste water more efficiently than any single one of the two bands. For instance, in [15], the authors showed that the most effective UV wavelengths in terms of inactivating mesophilic bacteria were not either 254nm or 365nm, but both 280nm and 365nm coupled. Moreover, the combined UVC and UVA (280nm and 350nm) irradiation also induced synergistic bactericidal effects on saprophytic bacteria [16]. Besides, applying extended UVA (365nm) exposure before UVC (265nm) significantly improved E. coli inactivation [17].

In terms of UV light sources, although the current disinfection equipment mainly employs low-pressure mercury lamps, the trend is clearly shifted to replace them with solid-state light sources, i.e., LEDs, due to environmental concerns. The materials used in fabricating UV LEDs range from GaN in UVA band to AlGaN in UVC band, with increasing aluminum composition. However, due to the intrinsic properties of aluminum-rich group III nitride materials, UVC LEDs still possess much lower external quantum efficiencies, than both the LEDs in the visible range and even their UVA counterparts [18]. Besides, UVC LEDs also suffer from much higher cost and lower reliability, than those in the other wavebands. Therefore, research efforts have also been made to alleviate the required output power of UVC LEDs for viral disinfection. Along this track, UVC LEDs have been combined either with other physical factors, e.g., heat [19,20], or with UVA LEDs. In [19], the radiation by UVC light at a temperature of 70$^{\circ}C$ for 15 minutes was shown to be able to inactivate viruses and bacteria, in a more effective way than the individual usage of the UVC and heat. Moreover, the cumulative UVA radiation doses required for sufficient viral inactivation could be greatly reduced, when the UVA light was combined with a heating temperature of 60$^{\circ}C$ [20]. Especially, another advantage of combining UVC with UVA LEDs is the significant reduction of the cost, compared to that of the disinfection solutions purely depending on UVC LEDs. However, the aforementioned work on studying the effect of coupled UVC and UVA light mainly focused on disinfecting bacteria. To the best of our knowledge, the effects of coupled UVC and UVA LEDs in inactivating viruses, especially the RNA types, has not yet been studied.

On the other hand, the SARS-CoV-2 virus relates to the family of Betacoronavirus ($\beta CoV$), a positive-sense single-stranded RNA virus with an envelope [21]. For safety concerns, the research on SARS-CoV-2 and other highly pathogenic viruses is restricted to biosafety-level-3 (BSL-3) facilities. Such BSL-3 classification makes the studies of disinfecting such viruses impossible in the majority of functioning research laboratories worldwide. Therefore, there is an urgent need for a viral model similar to the SARS-CoV-2.

In fact, some efforts have already been made toward finding suitable surrogates. For instance, bacteriophages are frequently used as models of mammalian viruses [22], and were also applied as a model in disinfection [23]. However, most bacteriophages contain double-stranded DNA without envelopes. DNA is even harder to be broken by UV light than RNA. Besides, enveloped viruses such as coronaviruses appear to be more sensitive to UV light than non-enveloped species [24]. Hence, bacteriophages are not very suitable to replace RNA viruses to study UV disinfection. On the other hand, feline infectious peritonitis virus [25], murine hepatitis virus [26] and influenza A virus [27] are RNA viruses with envelope, whose feasibility of replacing the SARS-CoV-2 in disinfection experiments has also been reported in the literature. Although these three types of viruses are similar to the SARS-CoV-2, they are still highly pathogenic, and thus also require high safety precautions. Besides, in our previous work, we also used the recombinant adeno-associated virus (rAAV) as a model [20]. However, rAAV takes a double-stranded DNA structure, which is also more suitable to replace DNA viruses. In comparison, a lentiviral vector is double-stranded with an envelope, which also shares considerable morphological and biological properties with the SARS-CoV-2 virus [28]. It has been known that airborne viruses with single-stranded nucleic acid (ssRNA) are more susceptible to UV inactivation than those with double-stranded ones (dsRNA) [29]. In other words, the UV light at a certain fluence is more likely to inactivate SARS-CoV-2 than lentivirus. Moreover, lentivirus has been successfully employed as a surrogate for SARS-CoV-2 in neutralization assays [30]. Additionally, lentivirus is safe and efficient, so it is widely accessible and applicable in normal biological laboratories [31]. Therefore, lentivirus is suitable as a surrogate in studying UV disinfection. Nevertheless, no disinfection research has ever applied this virus as a surrogate yet.

The contributions of this work are threefold. Firstly, lentivirus was used as a new surrogate for the SARS-CoV-2 virus in UV disinfection experiments for the first time. Secondly, we studied the viral inactivation efficacy of 280nm, 365nm and their combinations at various dosages, and found an equivalent solution of the combined two bands to replace the design purely with the UVC LEDs. To the best of our knowledge, this is the first attempt to study the performance of the coupled UVC and UVA LED light sources in inactivating viruses. Thirdly, to disclose the mechanism of viral inactivation by the UVA and UVC LEDs, real-time reverse transcription polymerase chain reaction (RT-qPCR) experiments were conducted to detect the levels of damage to the viral RNA.

2. Materials and ethods

2.1 Cell culture

Human embryonic kidney 293T cells (Catalogue number GNHu17) were obtained from the Cell Bank of Chinese Academy of Sciences (Shanghai, China) and maintained in Dulbecco’s Modified Eagle Medium (DMEM, Gibco, Catalog number C11995500BT, USA) supplemented with 10% Fetal Bovine Serum (FBS, EVERY GREEN, 11011-8611, China) and 1% of penicillin-streptomycin solution (HyClone, SV30010, USA). The cells were added to a 100mm tissue culture dish and cultured at 37$^{\circ}C$ in a 5% $\textrm {CO}_{2}$ humidified incubator.

2.2 Virus preparations

Lentiviral vectors have been increasingly used as a tool in the experimental study, since they are genetically modified to be safe. At present, the three-plasmid and four-plasmid expression systems are widely used. The packaging signal in this system is deleted, so the packaging sequence cannot be integrated into the virus genome. Therefore, the virus cannot replicate after infecting the host cell, nor can it use the host cell to produce new virus particles [32]. Moreover, the virulence genes in the commonly used lentiviral vectors have been deleted and replaced by exogenous target genes, which belong to pseudotyped viruses [33]. In this work, a second-generation lentiviral vector system was used to produce stable viruses [34]. The lentiviral vectors were generated by the transient cotransfection of HEK 293T cells with a three-plasmid system detailed as follows.

2.2.1 Plasmid constructs

The system consisted of psPAX2 (P0261) as packaging plasmid, pMD2.G (P0262) as VSV-G envelope expressing plasmid, as well as pLVX-AcGFP1-N1 (P0251) as transfer plasmid. They were purchased from the Miaoling Plasmid Sharing Platform (MLPSP). Three plasmids were transformed into separate STBL3 cell cultures. Amplified recombinant plasmids were extracted by EndoFree Mini Plasmid Kit II (TIANGEN, DP118, China), according to the manufacturer’s protocol. Then, the quality and quantity of the extracted DNA samples were measured by using the NanoDrop2000 spectrophotometer.

2.2.2 PEI-mediated transfections for generating lentiviral vectors

HEK 293T cells were seeded in 100mm tissue culture dishes at the amount of ${1\times 10}^{7}$ cells. The cultures were maintained at 37$^{\circ}C$ in a 5% ${\rm CO}_{2}$ throughout the virus production period. On the following day, when the cultures reached 60-70% confluence, the growth medium was replaced with 5$ml$ of the DMEM medium without FBS. At the same time, the lentiviral transfer vector plasmid (pLVX-AcGFP1-N1; 8$\mu g$), packaging plasmid (psPAX2; 6$\mu g$) and envelope plasmid (pMD2.G; 3$\mu g$) DNA were mixed in 200$\mu l$ DMEM without FBS. Concurrently, 50$\mu l$ polyethyleneimine (PEI; 1$mg/ml$) stock was diluted in 150$\mu l$ DMEM without FBS for 5 minutes on standing. Then, the DNA mixture was slowly added to the PEI solution; and the DNA/PEI mixtures were incubated at room temperature for 15 minutes. Next, 400$\mu l$ of DNA/PEI mixtures were added drop-wise to the tissue culture dish, which was incubated for 8 hours. Then, the medium in the culture dish was replaced with a 12$ml$ fresh growth DMEM medium. The cells were finally cultured for 48 hours.

2.2.3 Virus concentration

After a slightly yellowish medium was observed in the tissue culture dish, the medium was collected into a 15$ml$ centrifuge tube, and then cleared by centrifugation at 1000$rpm$ for 5 min. Next, the supernatant was passed through a 0.45$\mu m$ pore PVDF Millex-HV filter (Millipore). 3$ml$ lentiviral enrichment reagent (GM-040801-50, Genomeditech, China) was added to the virus suspension, which was then kept at 4$^{\circ}C$ overnight, and centrifuged (Eppendorf Centrifuge 5810R) at $4000\times g$ for 25 minutes. Then, 100$\mu l$ DMEM medium was used to re-suspend the viruses.

2.3 Virus titer

Since lentiviruses carry the GFP reporter gene, they express green fluorescent proteins (GFP) after invading the HEK 293T cells. Therefore, we determined viral titers by counting the 293T cells emitting fluorescence [35]. In brief, HEK 293T cells were seeded onto 96-well tissue culture-treated plates at a density of 1$\times {10}^{6}$ cells/well in 100$\mu l$ medium one day before the experiment. Then, 10$\mu l$ lentivirus supernatant was thawed on ice. Six EP tubes were prepared to contain 90$\mu l$ DMEM medium. The first tube was added with 10$\mu l$ lentivirus supernatant and mixed well. 10$\mu l$ of this mixture was taken from it and fed into the second tube. Therefore, the second tube was 10-fold diluted, and contained 1$\mu l$ of the original virus supernatant. This tenfold serial dilution process was repeated five times, as depicted in Figure S1 in the supplementary material. Finally, the old medium in the 96-well plate was removed, and replaced with the sequence of mixtures with six different viral concentrations in the EP tubes. Therefore, the virus capacity corresponding to each well was respectively 10, $10^{0}$, $10^{-1}$, $10^{-2}$, $10^{-3}$, $10^{-4}$ in a descending order. The corresponding dilution factor was thus 10$^{0}$, 10$^{1}$, 10$^{2}$, 10$^{3}$, 10$^{4}$, 10$^{5}$.

After 8 hours of infection, each well was supplemented with 200$\mu l$ of fresh medium containing FBS, and was incubated for 72 hours at 37$^{\circ}C$ in 5% $\textrm {CO}_{2}$. The virus titers were then quantified via the fluorescence counting assay performed as follows. The cells in each well were first observed using a Zeiss Axio Observer fluorescence microscope (Germany), and captured at 5X magnification with a camera. The number of cells in the last picture with fluorescent expression, i.e., taken from the well with the lowest concentration of the virus after the light irradiation, was then counted via the labeling and measure functions of ImageJ. The functional transducing units $\xi \, [TU/ml]$ can thus be calculated as follows.

Here, $n$ stands for the number of cells; and $x$ represents for the dilution factor.

2.4 UV disinfection apparatus

The disinfection experiments were performed in the UV device described in [20]. The experimental setup is schematically shown in Fig. 1. The UVC light source was composed of 6 light bars, each of which was evenly arranged with 8 UVC LEDs (LTPL-G35UV275GC-E, LITE-ON Technology, China), with a 4$cm$-interval between each LED. The UVA channel was a 4$\times$6 array, composed of 4 light bars, each of which was evenly arranged with 6 UVA LEDs (SST-10-UV-X130-F365-00, Luminus Devices, USA), with a 6$cm$-interval between each LED. Each channel is driven by two constant current drivers (TPS92512, TI, USA) to keep the magnitude of the current at a constant level. These drivers can maintain the output power of both channels at a relatively constant level, after the LEDs reach their thermal equilibrium states. Specifically, the drive current and voltage of each driver in the UVC channel were 0.6 A and 28.8 V; and those of the UVA channel are 1.2 A and 22.2 V respectively. These parameters are within the nominal operating range of the TPS92512 drivers.

 

Fig. 1. The experimental apparatus. (a) The schematic diagram. Blue represents 365nm LEDs; and purple represents 280nm LEDs. (b) The power spectral density curves of the UV LEDs are shown respectively by the dotted line for 365nm and the solid line for 280nm.

Download Full Size | PDF

In the best two cases of the viral inactivation, as will be detailed later, the energy consumption of the apparatus is as follows. Firstly, when the UVC channel worked at its full power for 60 minutes, totally 124.4$KJ$ electricity was consumed. Secondly, in the coupled case, when the UVC channel worked at the half of its maximum power and the UVA channel worked in its full power simultaneously for 45 minutes, the energy consumption was 190.5$KJ$.

The actual spectra of the LED light source were measured using a Maya2000Pro spectrometer (Ocean Optics, US), which is shown in Fig.  1(b). The measured peak wavelengths were 280nm and 365nm respectively. The viral samples were placed on the target plane at a distance of $20cm$ away from the LED array. The output power density and the uniformity of each UV channel on the target plane were measured by a PM100D power meter with an S120VC probe (Thorlabs, USA). The irradiance of the 280nm light was calibrated to 200$\mu W/cm^{2}$ on the target area; while that of the 365nm light was set at 13$mW/cm^{2}$. The UV irradiance was adjusted by tuning the current to the LEDs. Moreover, the UVA and UVC LEDs could be switched on at the same time.

2.5 Inactivation experiments

The virus suspension was irradiated by the aforementioned LED light sources. The experiments were arranged into nine groups and labeled by Roman numbers, as detailed in Table 1. In the experiments of the individual UV band, the virus was respectively irradiated by the 365nm light with an irradiance of 13$mW/cm^{2}$ for 60 minutes (I), and with an irradiance of 200$\mu W/cm^{2}$ by 280 nm respectively for 10 minutes (II), 30 minutes (III), 45 minutes (IV), and 60 minutes (V).

Table 1. Parameters of the UV irradiance. (- represents nothing; F represents that the virus was sequentially irradiated by 365nm and then by 280nm for either 30 or 45 minutes of each individual light (thus respectively 60 or 90 minutes in total); C represents the case that the virus was simultaneously irradiated by 365nm and 280nm for either 30 or 45 minutes. Roman numbers represent the experimental groups.)

View Table | View all tables in this article

In order to thoroughly investigate the inactivation effects of the coupled UVA and UVC irradiation on the virus, various combinations were examined. The UVC irradiance was set at 50%, i.e., 100$\mu W/cm^{2}$ in all the coupled irradiation experiments. More specifically, in the VI group, the virus was sequentially irradiated by 365nm for 30 minutes, and then by 280nm for 30 minutes. In the VIII group, the virus was sequentially irradiated by 365nm for 45 minutes, and then by 280nm for 45 minutes. In the VII group, both 280nm and 365nm were simultaneously used to irradiate the virus for 30 minutes; while in the IX group, both the light sources simultaneously worked for 45 minutes. Note that the motivation of first using 365nm to irradiate the virus in the VI and VIII groups was to test the hypothesis that UVA light can damage membrane proteins, and thus makes it easier to inactivate microorganisms [17].

On the other hand, in order to verify whether the thermal effect generated by the UV device had an adverse impact on the virus titer, we measured the temperature changes in the culturing medium (without the virus) with a thermal couple (DS18B20), after individually switching on each of the 280nm and 365nm LED light sources, or both of them simultaneously. In all the three cases, the working light sources were set to their full power for 50 minutes. Then, the changes in virus titers were evaluated after putting the virus in an environment heated to the recorded highest temperature for an hour, without UV irradiation.

The inactivation efficacy of the UVA, UVC, and their various combinations on the virus was defined as the logarithm to the base 10 of the ratio between the titers of the viruses after and before the UV light irradiation, i.e.,

As a summary, the overall experimental scheme can be shown in Fig. 2.

 

Fig. 2. The experimental flow chart.

Download Full Size | PDF

2.6 RT-qPCR of viruses

In order to verify the effect of UV irradiation on viral genes, RT-qPCR was used to detect the viral RNA. At the same time, it was verified via the comparisons with the virus titer reductions.

2.6.1 Primer design

Retroviruses are expressed under the control of long terminal repeats (LTRs), which contain all signals for transcriptional initiation as well as transcriptional termination [36]. Primers/probe location and sequences related to lentiviral-derived transfer vector genomes can be designed in endogenous DNA elements, such as LTR [37]. Due to the stability of LTRs in the virus, the copy numbers of the LTR region were determined to analyze the viral RNA. The primers were designed by software Primer Premier 5, and were synthesized by the Sangon Biotech company. The sense primer (LTR-F) was 5’-aggccaataaaggagagaacac-3’; and the antisense primer (LTR-R) was 5’-cttgaagtactccggatgcag-3’.

2.6.2 Viral RNA extraction

The viruses were processed to extract the viral RNA with a MiniBEST Viral RNA/DNA Extraction Kit ver. 5.0 (9766-1, Takara, Japan), according to the manufacturer’s directions. Briefly, the process entailed extracting the RNA chemically from 10$\mu l$ virus and loading it on a small chromatographic column in a micro-centrifuge tube. After two rounds of washing, 30$\mu l$ RNA was eluted, and was determined using Nanodrop.

2.6.3 Viral cDNA synthesis

The viral cDNA was synthesized with ReverTra Ace qPCR RT Master Mix with gDNA remover (FSQ-301, TOYOBO, Japan). An 8$\mu l$ mixture containing 500ng total RNA and 2$\mu l$ gDNA remover were incubated at 37$^{\circ}C$ for 5 minutes. Then, 2$\mu l$ 5x RT Master Mix II was added to the reaction, and sequentially incubated at 37$^{\circ}C$ for 15 min, 50$^{\circ}C$ for 10 min, and finally 98$^{\circ}C$ for 5 min.

2.6.4 Real-time polymerase chain reaction

Real-time PCR experiments were carried out using an ABI 7500 Fast Real-time PCR Detection System (Applied Biosystems, US). A total of 30$\mu l$ reacting volume contained 9.6$\mu l$ water, 15$\mu l$ SYBR Green Real-time PCR Master Mix (QPK-201, TOYOBO, Japan), 1.2$\mu l$ (10$\mu M$) of each primer, and 3$\mu l$ cDNA templates. The conditions of the PCR were set as 95$^{\circ}C$ for 60 seconds, 40 cycles (sequentially 95$^{\circ}C$ for 15s, 60$^{\circ}C$ for 15s, and 72$^{\circ}C$ for 40s). Each sample was run in triplicate. Product specificity was evaluated by using melting curve analysis and agarose gel electrophoresis. 10$\mu l$ of the amplified product was loaded in agarose gel (2.5%) containing 0.5$\mu g/ml$ ethidium bromide. The electrophoresis was conducted in the TAE buffer. After the electrophoresis, the DNA bands were visualized by UV transillumination. The relative cDNA expression levels were calculated by the $2^{- \triangle \triangle Ct}$ method. The measured amount of cDNA was equal to the amount of the viral RNA. We used the same method to detect the threshold cycle (CT) of LTR in the genomic DNA in the pLVX-AcGFP1-N1 plasmid. Then, the standard curve of the real-time PCR was obtained by the copy numbers of the plasmid with CT. This means that the copy numbers of viral RNA can be obtained from the standard curve. The $[log_{10}]$ reduction of viral RNA was calculated as follows:

2.7 Confocal laser scanning microscopy

In order to more clearly observe the effect of virus inactivation by the UV light, viruses were observed on an AIR HD25 confocal microscope (Nikon, Japan). The virus in a suspension was transfected into cells in a 35mm glass-bottom dish with a 20mm microwell. The cells transfected by both the un-irradiated and the post-irradiated viruses could be clearly observed by the laser confocal microscope.

2.8 Statistical analysis

Each experiment was repeated for three times. The software GraphPad Prism 6 (GraphPad Software, San Diego, US) was used in the statistical analysis and depiction of the results. All the values were expressed in the format of “mean$\pm$SEM”. The differences between the means of the two groups were tested for significance by Student’s t-test. The p values less than 0.05 were considered significant.

3. Results and Discussion

3.1 Inactivation by individual UVA and UVC LED light source

Traditional biological methods for detecting virus titers include plaque-forming assay (PFU) and $\textrm {TCID}_{50}$ assay. However, in our preliminary experiments, we found that the lentiviral vector did not form plaques, or generate a severe cytopathic effect (CPE) significantly, because the lentivirus could hardly harm the cells. Consequently, PFU assay and $\textrm {TCID}_{50}$ assay are not applicable, when using such a viral surrogate. However, transducing unit (TU) can represent the infectious events, and is also a more essential parameter for each recombinant vector tested resulting from transgene expression by host cells [38]. Therefore, we used TU to determine the virus titers. The fluorescence intensity of the 293T cells observed under a fluorescence microscope is shown from high to low respectively in Figure S2 in the supplementary material, where the last picture in the sequence with the fluorescent expression was selected to determine the viral titer.

In order to compare the effective doses of UVA and UVC to kill the virus, the $[log_{10}]$ reductions of the inactivated virus were measured at various doses, and depicted in Fig. 3(a). The irradiation at a fluence of 46.8$J/cm^{2}$ (13$mW/cm^{2}$ for 60min) by the 365nm LEDs resulted in 0.911 $[log_{10}]$ in the viruses. On the other hand, at the irradiance of 200$\mu w/cm^{2}$, the 280nm light irradiation was applied to the virus for 10, 30, 45, and 60 minutes respectively. The results showed that the $[log_{10}]$ reductions of the virus were respectively 1.825 (0.12$J/cm^{2}$), 2.042 (0.36$J/cm^{2}$), 2.873 (0.54$J/cm^{2}$) and 3.343 (0.72$J/cm^{2}$), as listed in Table 1. It can be seen from the data that the UVC light can better kill the virus, compared with the UVA light. The correlation coefficient between the fluence of the UVC light and the titer reduction is 0.9636; that is, they are highly correlated. As the dose of the UVC light increases, so does its capability to inactivate the virus.

 

Fig. 3. (a) The $[log_{10}]$ reductions in the viral titers by the UVA and UVC irradiation. The virus suspension was irradiated by the 365nm light at 13$mW/cm^{2}$ for 60 minutes, i.e., the experiment group I; and was irradiated by the 280nm light at 200$\mu w/cm^{2}$ respectively for 10 (II), 30 (III), 45 (IV), and 60 (V) minutes. The control was the virus free of UV irradiation. The Roman numbers correspond to the experimental numbers defined in Table 1. (b-c) Fluorescent images of the cells containing GFP observed by a confocal microscope: (b) the cells were directly infected by untreated viruses; and (c) the cells were infected with the viruses that had been irradiated by the 280nm light for 30 minutes.

Download Full Size | PDF

On the other hand, the results presented in Table 1 and Fig. 3 showed that inactivating the virus by the UVA light alone required a very high dose to reach a 1 $[log_{10}]$ reduction. Obviously, it is not an effective wavelength for virus inactivation. Therefore, UVA light combined with psoralen (PUVA) was used in photocatalytic processes to inactivate pathogens [39]. Psoralen has the right structure and shape that can be inserted between two strands of DNA in its double helix structure, and can thus induce the formation of inter-strand covalent cross-links between the reverse complementary nucleic acid strands via photocatalysis.

In this study, we demonstrated that the viruses could be effectively inactivated by the UVC irradiation with high virus titer reductions; whereas, the UVA irradiation was much less effective. Although this is not a new finding, it is still relevant to compare the effects of the two individual UV bands with those of their combinations, especially in terms of inactivating the RNA viruses. Moreover, in order to better reflect the inactivation effect of UVC, we observed via the confocal microscopy that the cells infected with the virus, which had already been irradiated by the 280nm light for 30 minutes (0.54$J/cm^{2}$), emitted much less fluorescence than the cells infected with the untreated virus did, as illustrated in Fig. 3(b) and (c).

On the other hand, to discover the potential of low power (and thus lower cost) UVC light sources for disinfecting RNA viruses, the irradiance of the 280nm LEDs was kept as low as 200$\mu w/cm^{2}$ in this work. In comparison, the previously reported dosage by other authors appeared to be higher. For instance, as reported in [39], the fluence of 0.055$J/cm^{2}$ (5.5$mw/cm^{2}$, 280nm) was capable of decreasing the PFUs of an H1N1 subtype by 3 $[log_{10}]$. In another work, the fluence of 25$mJ/cm^{2}$ (2.5$mw/cm^{2}$, 280nm) was reported to reduce the herpes simplex virus type 1 by around 3 $[log_{10}]$ [40]. Although the irradiance applied in our study was not that high, when the energy density (fluence) reached 0.72$J/cm^{2}$, the 280nm light was still capable of reducing the viruses by more than 3 $[log_{10}]$. Therefore, the UVC light was confirmed as an effective wavelength to inactivate RNA viruses.

3.2 Combinations of UVA and UVC LEDs for viral inactivation

In order to verify whether UVA can reduce the required power of UVC LEDs, combinations of the UVA light and the UVC with reduced power were applied in the inactivation experiments to investigate whether they can achieve the comparable effect to the case of using the full-power UVC light alone. Specifically, the UVA and UVC radiation were combined with their irradiance respectively set at 13$mW/cm^{2}$ for the 365nm LEDs and 100$\mu w/cm^{2}$ for the 280nm LEDs, to either sequentially or simultaneously irradiate the virus.

As shown in Fig.  4(a) and Table 1, after being sequentially irradiated with the 365nm for 30 minutes and then with the 280nm for 30 minutes, the virus titer was reduced by 1.643 $[log_{10}]$ (VI). The reduction of the virus titer after being simultaneously irradiated by the 280nm and 365nm for 30 minutes turned out to be 2.1 $[log_{10}]$ (VII). In comparison, irradiating the virus with the full power of 280nm (200$\mu w/cm^{2}$ ) for 30 min resulted in a reduction in the virus titer by 2.042 $[log_{10}]$ (III). A comparison of the titer reductions resulted in the III group and VI group showed no significant difference. Also, the III group and the VII group had no significant difference. This indicates that the effect of the coupled UVC at a dose of 0.18$J/cm^{2}$ with UVA at a dose of 23.4$J/cm^{2}$ is roughly equivalent to that of the UVC alone at a dose of 0.36$J/cm^{2}$, in terms of a similar viral inactivation of 2 $[log_{10}]$.

 

Fig. 4. Titer reductions by the combined UVA and UVC irradiation. The irradiance of the 365nm light was 13$m w/cm^{2}$ in all cases. When irradiated by the 280nm alone, its irradiance was 200$\mu w/cm^{2}$. When coupled with the 365nm, the irradiance of 280nm was 100$\mu w/cm^{2}$. (a) The $[log_{10}]$ reductions of the virus after 30-minute irradiation by the UVC alone, and 30-minute irradiation by both the UVA and UVC, either sequentially or simultaneously. (b) The corresponding results of 45 minutes. Roman numbers corresponds to the experimental numbers defined in Table 1. $\ast \ast$ represents $p \leq 0.01$; and $ns$ represents no significant difference.

Download Full Size | PDF

On the other hand, after being irradiated by the 365nm for 45 min and then with the 280nm for 45 min (VIII), the virus was reduced by 3.176 $[log_{10}]$, as shown in Fig.  4(b). Besides, the coupled irradiation by the 280nm and 365nm for 45 minutes (IX) reduced the virus by 3.04 $[log_{10}]$. By the UVC alone, a dose of 0.54$J/cm^{2}$ (IV) managed to reduce the virus by 2.873 $[log_{10}]$. The IV and VIII group, and the IV and the IX group were not significantly different. Therefore, the result from the combined use of the UVC at a dose of 0.27$J/cm^{2}$ and the UVA at a dose of 35.1$J/cm^{2}$ was equivalent to using UVC alone at 0.54$J/cm^{2}$. Therefore, the combined use of the UVA and the low-power UVC can replace the use of high-power UVC alone. These findings coincide with the previously reported results that the effect of combined UVA/UVC irradiation was as high as UVC irradiation alone on various targets, including saprophytic bacteria and pure bacterial cultures or the micro-organisms in a more complex medium, e.g., a wastewater effluent [15,16,41]. Thus, coupling UVA with UVC can indeed be a promising solution for virus disinfection.

It is worth mentioning that there are generally three realizations of combined UVA and UVC irradiation. The first is simultaneous irradiation. The second is sequentially switching on UVA and then UVC LEDs. The third is also sequentially, but in a reverse order. It has been shown in the literature that these different realizations had different effects on inactivating bacteria. In [16], it was reported that first inactivating bacteria with UVC followed by UVA could even reactivate the bacteria. This was attributed to the fact that UVC irradiation induces fatal lesions that restrain DNA duplication in the bacteria, which can be repaired by either photolyase (photoreactivation) or glycosylases and polymerases (dark repair). On the other hand, applying extended UVA exposure before UVC can significantly improve E. coli inactivation, but did not affect much virus MS2 [17]. However, viruses do not contain complex biological structures, lack cellular functions such as repair enzymes, and do not possess photoactivation and dark repair functions like bacteria. So theoretically, the order of applying the two UV bands may not have additional effects on viruses. In fact, in the results from this work, no significant difference was found between the VIII and IX group, which confirms this hypothesis. However, a comparison of the titer reductions in the VI and the VII group showed a significant difference (p=0.0015). Therefore, the inactivation effect of simultaneous irradiation is better than that of successive irradiation in 30 minutes.

3.3 Effect of temperature on virus titers

As shown in a previous work, a temperature of 60$^{\circ}C$ can enhance the disinfection effect of the UV light [20]. As shown in Fig.  5(a), the highest temperature of the virus suspension after being heated by the light sources could reach 35$^{\circ}C$ from the normal room temperature within 50 minutes. In order to verify that the inactivation of the virus had nothing to do with the raised temperature, we detected the virus titer after being left at an environment at a constant temperature of 37$^{\circ}C$ for one hour. The results showed that the virus titer was 8.7$\times 10^{6}$ $TU/ml$, which was not significantly lower than that of the control group 9$\times 10^{6}$ $TU/ml$, as shown in Fig.  5(b). Therefore, the temperature did not cause a drop in the viral viability.

 

Fig. 5. The effect of temperature on virus titers. (a) The temperature changes in the virus suspension within 50 minutes under 280nm, 365nm and their combined irradiation. (b) The virus titer after being left in an ambient temperature of 37$^{\circ}C$ for one hour, with the titer of the virus at a normal room temperature as the control group.

Download Full Size | PDF

3.4 Mechanism of UV disinfection

The inactivation mechanisms of UVA and UVC are different. Generally speaking, UVA mainly destroys viral proteins so that the virus cannot invade cells; while UVC causes irreparable damage to the viral genome. As reported in [42], the RNA genome in non-enveloped feline calicivirus was damaged, as shown by the decreased copy number, following 281nm radiation. To explore the mechanisms of UVA and UVC inactivation, RT-qPCR was performed to measure the genomic damage in this work. The combination of RT-qPCR and viral titer results indirectly verified whether UV destroys viral proteins. The genomic changes were visualized after irradiating the virus by the 365nm for 60 minutes (I), and by the 280nm for either 30 minutes (III) or 45 minutes (IV). In the RT-qPCR analyses, the serial 10-fold dilution curves were developed to correlate the PCR cross-threshold (CT) with the viral concentration and to measure the $log_{10}$ changes in the concentrations of the gene copies. From Fig. 6(a), the following linear equation was fitted and used in calculating the $log_{10}$ copy numbers, in the RT-qPCR analysis of the LTR region of the pLVX-AcGFP1-N1 stock, i.e., the standard curve of real-time PCR:

 

Fig. 6. RT-qPCR results. (a) Standard Real-time PCR curve shows the relationship between CT and copy number. (b) The copy number of the control group, that of the virus after being irradiated by the 365nm light for 60 minutes (I), and by the 280nm for both 30 minutes (III) and 45 minutes (IV). (c) The Melt Curve. (d) The agarose gel electrophoresis. The Marker is 500bp ladder molecular size marker. Roman numbers represents experimental numbers in the Table 2.

Download Full Size | PDF

Table 2. The results of RT-qPCR experiments

View Table | View all tables in this article

As shown in Fig. 6, the qPCR results were adjusted to reflect the UV damage to the entire genome by using Eq. (4). The copy numbers of the control group and those after being irradiated respectively by the 365nm light for 60 minutes (I) and by the 280nm light for 30 minutes (III) or 45 minutes (IV) were respectively 1.06$\times 10^{9}$, 2.33$\times 10^{8}$, 9.5$\times 10^{7}$, and 1.7$\times 10^{7}$, as depicted in Fig. 6(b). As a result, the 365nm irradiation for 60 minutes, and the 280nm irradiation for both 30 minutes and 45 minutes respectively reduced the viral RNA loads by 0.658, 1.147, and 1.793$copies/ml$, as listed in Table 2. Furthermore, the RNA reduction was highly correlated with the reduction in the virus titers, with a correlation coefficient of 0.986. The data indicated that both the UVA and UVC can cause damage to viral RNA. However, the UVA light indirectly damages the viral RNA, as aforementioned. A more specific reason can be attributed to the fact that the virus was in the DMEM medium, which contains a small amount of photosensitizer. Therefore, the UVA light also induced ROS, and thereby destroyed the viral RNA.

It is worth noting that in the qPCR experiments, the threshold cycle (CT) also appeared in the blank group. Therefore, we further examined the Melt curve through the 7500 Software as shown in Fig. 6(c), where the blank groups and the experimental groups were different. Moreover, we subjected the blank product to agarose gel electrophoresis experiments. There were no DNA bands in the blank group; while the control group had obvious DNA bands (Fig. 6(d)). This implies that the blank group was free of contamination.

4. Conclusion

In this study, we first demonstrated the feasibility of using a lentiviral vector as a surrogate for RNA viruses, e.g., SARS-CoV-2, to study UV disinfection. This surrogate is safe and reliable, and can thus be used in ordinary laboratories. With the lentivirus, the disinfection effects of the UVA (365nm) and UVC (280nm) LED light sources and their combinations at various doses were studied and compared. The results showed that the effect of the coupled 365nm and low-power 280nm light is equivalent to the individual high-power 280nm light, in terms of the reductions in the virus titers. Therefore, the combination of the UVA and UVC LEDs in viral inactivation can achieve the same disinfection effect as that of applying UVC LEDs alone. Moreover, the inactivation mechanism of the ultraviolet light was studied by detecting the damage in the viral RNA. The results showed that both the UVA and UVC light managed to damage the viral RNA, though their effects turned out to be significantly different. In summary, this study provides important evidence of applying coupled UVA and UVC light to effectively inactivate viruses, especially the RNA type. The benefits of coupling these two bands in disinfection include the reduction of hardware cost and the improvement of both the hardware reliability and the penetration capacity of the light into non-aerial media, e.g., waters and glassware.

As direct future extensions of the current work, the effects of more combinations of UV wavelengths in the range of 200-400nm are also worth investigating, as well as more variations of the applied light power density and exposure time.