Introduction
The seasonal variation of respiratory viral diseases, such as influenza, have long been recognized with the incidence of infection waxing in the wintertime and waning in the summertime (Moriyama et al., Reference Moriyama, Hugentobler and Iwasaki2020). Evidence of the same infectious behavior for the SARS-CoV-2 virus, the virus responsible for the coronavirus 2019 (COVID-19) pandemic, has also been presented, as shown in Figure S1 in the Supplementary Material (Shamsa et al., Reference Shamsa, Shamsa and Zhang2023), although the exact behavior in time has been difficult to determine because of the emergence of virus variants and the uneven preventative public health measures taken (Townsend et al., Reference Townsend, Hassler, Lamb, Sah, Alvarez Nishio, Nguyen, Tew, Galvani and Dornburg2023).
What causes this seasonality is not yet fully established, with three different theories put forward: the effect of climate conditions on (1) host resistance to infection, (2) host social behavior, and (3) virus survivability (Price et al., Reference Price, Graham and Ramalingam2019). Increasingly, evidence has pointed to the role of relative humidity of indoor air, with a strong correlation found for increased virus survival with low relative humidity (Choi et al., Reference Choi, Tuel and Eltahir2021; Liu et al., Reference Liu, Huang, Li, Zhao, Wang, Huang and Yang2021; Pineda Rojas et al., Reference Pineda Rojas, Cordo, Saurral, Jimenez, Marr and Kropff2021; Wang et al., Reference Wang, Tang, Feng, Lin, Lv, Chen and Wang2021; Nieto-Caballero et al., Reference Nieto-Caballero, Davis, Fuques, Gomez, Huynh, Handorean, Ushijima, Tolbert and Hernandez2022; Oswin et al., Reference Oswin, Haddrell, Otero-Fernandez, Mann, Cogan, Hilditch, Tian, Hardy, Hill, Finn, Davidson and Reid2022; Park et al., Reference Park, Lee, Oh, Nam, Barrett, Lee and Hwang2022; Ravelli and Martinez, Reference Ravelli and Martinez2022; Verheyen and Bourouiba, Reference Verheyen and Bourouiba2022; Yin et al., Reference Yin, Zhao and Pereira2022). It has been proposed that the change in indoor relative humidity with seasons is caused by the fact that we heat the indoor air in the colder months to keep the occupants warm, which dries out the indoor air (Lowen and Steel, Reference Lowen and Steel2014). Nevertheless, because the cause of this correlation between airborne viral infection and relative humidity has not been established, the control of relative humidity has not been adopted as a public health policy. This study advocates for a change in this viewpoint.
Several theories have been advanced to explain how relative humidity affects virus survivability. These involve pH change at the droplet surface, changes in strength of the virus protein structure, and efflorescence in which salt or carbohydrate coat and protect the virus upon evaporation which is accelerated at lower relative humidity (Nieto-Caballero et al., Reference Nieto-Caballero, Davis, Fuques, Gomez, Huynh, Handorean, Ushijima, Tolbert and Hernandez2022; Oswin et al., Reference Oswin, Haddrell, Otero-Fernandez, Mann, Cogan, Hilditch, Tian, Hardy, Hill, Finn, Davidson and Reid2022). Recent work has also shown that as the relative humidity increases, the concentration of reactive oxygen species (ROS) in water droplets increases, such as the hydroxyl radical (OH) and hydrogen peroxide (H2O2) (Dulay et al., Reference Dulay, Huerta Aguilar, Chamberlayne, Zare, Davidse and Vukovic2021; Mofidfar et al., Reference Mofidfar, Mehrgardi, Xia and Zare2024). Previous work has demonstrated that ROS in water droplets can act as a bactericide with high effectivity, much larger than if it were only caused by the presence of H2O2 (Dulay et al., Reference Dulay, Lee, Mody, Narasimhan, Monack and Zare2020). The present study demonstrates that water droplets surrounded by air at different relative humidity levels affect the infectivity of SARS-CoV-2 that is contained inside water droplets. Specifically, we show that as the relative humidity of the air is raised to 40% or more, there is a noticeable decline in the infection rate of SARS-CoV-2. The public health implications of these findings are briefly discussed.
Methods
Cell culture
An immortalized cell line of human colorectal adenocarcinoma cells with overexpression of SARS-CoV-2 viral N protein (Caco-2-N) were maintained in Dulbecco’s modified Eagle medium (DMEM; Gibco, China) supplemented with 10% (vol/vol) fetal bovine serum (FBS), and 50 IU/ml penicillin/streptomycin in a humidified 5% (vol/vol) CO2 incubator at 37°C. This cell lines were tested negative for mycoplasma.
SARS-CoV-2 GFP/ΔN virus production
A trans-complementation system for SARS-CoV-2 including viral nucleocapsid gene (N) deficiencies SARS-CoV-2 virus, and Caco-2-N cell line that overexpresses N protein was used to produce SARS-Cov-2 in a biosafety level-2 (BSL-2) cell culture system, which recapitulates authentic viral infection and replication but without virulence. In detail, Caco-2-N cells were infected with P0 virus of SARS-CoV-2 GFP/ΔN to amplify the virus. Briefly, Caco-2-N cells were first spread in a 6-well plate, and after the cells were adhered to the wall, the medium was changed to a maintenance medium (DMEM with 2% FBS and no P/S), followed by the addition of P0 virus to the wells of a cell plate that was transferred to an incubator for virus amplification. The cell supernatant was harvested after 48 h and centrifuged to remove impurities before ultracentrifuging to obtain the P1 virus.
Microdroplet generation and collection
The overall experiment was conducted in a closed environmental chamber with a controlled relative humidity in the range of 20%–70% and controlled temperature in the range of 10°C–40°C. Microdroplets were generated by spraying the aqueous viral suspension at a rate of 5 μl/min through a 100 μm inner diameter fused silica tubing with 120 psi air coaxial sheath gas in a sealed chamber at 25°C under relative humidity in range from 20% to 70%. The sprayed viral microdroplets were collected in a Petri dish containing PBS solution filtered by 0.22 μm syringe filter, which was located 15 cm from the nebulizer.
Nanoparticle tracking analysis (NTA)
The harvested virus solution was first transferred to a biosafety cabinet, then 100 μl was removed and diluted fivefold with PBS to form the sample. All samples were measured by Nanosight NS300 (Malvern Pananalytical Ltd., Malvern, UK) at 25°C. Each sample was measured for 60 s and repeated five times, and finally the mean value was calculated.
Infection by virus
After nanoparticle tracking analysis (NTA) measurement of the viral content of the collected viral solution, the same amount of virus was used to infect the 48-well plate that had been previously spread with Caco-2-N cells. Positive control wells were filled with the same amount of virus from untreated stock virus solution and negative controls were filled with the same volume of virus-free PBS. The plates were then incubated at 37°C with 5% CO2 for 48 h.
Confirmation of viral activity by flow cytometry
After 48 h of virus infection, cells were washed in cold PBS and digested by trypsin (Thermo Fisher Scientific, Waltham, MA, USA) before collected in 1.5 ml microcentrifuge tubes. The percentage of virus-infected cells and fluorescence intensity were analyzed by FACS by using CytoFLEX (Beckman Coulter, Brea, CA, USA). Data were analyzed using FlowJo V10 software (Becton Dickinson, Franklin Lakes, NJ, USA).
Immunofluorescence assay
After 48 h of virus infection, cell culture plates were transferred to an inverted fluorescence microscope CKX53 (Olympus, Tokyo, Japan) or ZEISS LSM 900 confocal laser scanning microscope (ZEISS, Jena, Germany), and photographed using the accompanying computers. Analysis of fluorescence light intensity was accomplished using ImageJ software.
Western blot
Viruses contained in the collected viral solution after nebulization were quantified by the numbers of viral particles using the NTA test, and the same amount of virus particles were ultracentrifuged. A lysis buffer (50 mM Tris–HCl, pH 7.4, 300 mM NaCl, 1% Triton X-100, 5 mM EDTA, and 10% glycerol) containing a protease inhibitor (10%, Roche, 04693116001) were used to resuspend the viruses. Lysates were electrophoresed on 10% SDS-polyacrylamide gel electrophoresis (PAGE) and transferred to a polyvinylidene fluoride (PVDF) membrane. Nonspecific bands were blocked by 5% skim milk for 2 h and then incubated with SARS-CoV-2 spike protein mAb (Cell Signaling Technology, Boston, MA, USA), SARS-CoV-2 envelope protein antibody (Cell Signaling Technology, Boston, MA, USA), and SARS-CoV-2 membrane polyclonal antibody (Abnova, Taipei, China) at 4°C overnight. Protein bands were visualized using a standard enhanced chemiluminescent substrate (Epizyme, Cambridge, MA, USA).
Gel electrophoresis
After determining the RNA content of the extracted nucleic acids solution, the same amount of RNA was removed for further use. A 0.9% agarose gel was prepared by dissolving agarose powder (Invitrogen, Thermo Fisher Scientific, Sunnyvale, CA, USA) in tris-acetate (TAE) buffer. Confirmed extracted nucleic acids solution with the same RNA amount were mixed with loading buffer (Thermo Fisher Scientific, Sunnyvale, CA, USA) before electrophoresed using 100 V for 60 min. Nucleic acid mobility was visualized using a GelDoc Go (Bio-Rad, Hercules, CA, USA).
ROS concentration
A H2O2 standard solution reacted with 0.1 M potassium titanium oxalate (PTO) solution in equal volume, and the absorbance at 400 nm was determined by UV–Vis spectroscopy. Then, by varying the H2O2 concentration, a standard absorption curve was constructed. The samples (sprayed DI water microdroplets) were collected at the different relative humidity levels (20%–70%), reacted with 0.1 M PTO solution in equal volume, and the absorbances at 400 nm were determined. The PTO responds not only to H2O2 but other reactive oxygen species, such as OH.
Results and discussion
To examine the effect of humidity on the activity of SARS-CoV-2 in microdroplets, we atomized the GFP-tagged virus-containing suspension in a sealed chamber and collected the viral microdroplets at 25°C under relative humidity of 20% and 70%, respectively (as shown in Figure 1a). We confirmed the virus concentration in each of the collected suspensions by NTA analysis and adjusted them to the same level. The collected viruses were then used to infect Caco-2 cells. We compared the differences in viral activity in three samples, which are the viruses in bulk aqueous suspensions, and in microdroplets collected at 20% and 70% relative humidity, by assessing the cellular infectivity. As shown in Figure 1b, Caco-2 cells were first observed with 4′,6-diamidino-2-phenylindole (DAPI). The cells with green fluorescence were infected by SARS-CoV-2. It is evident that the virus activity in the bulk aqueous suspension is higher than that in the microdroplets. More importantly, the activity of viruses collected at 70% relative humidity is lower than that of viruses collected at 20% relative humidity. We then analyzed the corresponding fluorescence intensity of infected Caco-2 cells using ImageJ software. As shown in Figure 1c, the intensity of the viruses in microdroplets decreased with the rise of ambient humidity. Therefore, we found that the activity of viruses in aqueous microdroplets changed with the change in ambient humidity, and the trend of the change was consistent with the previous reports of COVID-19 transmission (Ma et al., Reference Ma, Pei, Shaman, Dubrow and Chen2021).
We also examined how the viral activity changed with temperature (see Figure S2 in the Supplementary Material) but found little change over the range of 20°C to 40°C. We further analyzed the change in the infectivity of the viruses in the microdroplets as the relative humidity was varied. As shown in Figure 2a, the viral infectivity is 40% at 70% relative humidity and 62% at 20% relative humidity. When the humidity was higher than 60%, the decrease in virus infectivity was maintained at about 40%.
Previous work reported that ROS can be spontaneously generated in microdroplets (Lee et al., Reference Lee, Walker, Han, Kang, Prinz, Waymouth, Nam and Zare2019; Xia et al., Reference Xia, Li, Zhang, Yin, Chen, Liang, Jiang and Zare2023). Therefore, we speculated that the change of viral activity in microdroplets may be related to ROS in microdroplets. To verify our conjecture, we tested the concentration of ROS in microdroplets at different humidity levels as shown in Figure 2b. As the relative humidity increases from 20% to 70%, the concentration of ROS increases from 0.2 to 1.2 mg/L.
Given that the distance between the nebulizer and the collector in our experimental design was only 15 cm, the flight time of the droplets in the air was less than 10 ms. Thus, the data presented refer to what happens at short times, such as face-to-face contact of a healthy person with a sick person. It might be expected that the effect of relative humidity on viral infectivity would be even more pronounced for longer times before droplet evaporation.
Recent work suggests that change in viral activity in microdroplets may be related to evaporation of microdroplets and diffusion of carbon dioxide into the microdroplets in air (Nieto-Caballero et al., Reference Nieto-Caballero, Davis, Fuques, Gomez, Huynh, Handorean, Ushijima, Tolbert and Hernandez2022; Oswin et al., Reference Oswin, Haddrell, Otero-Fernandez, Mann, Cogan, Hilditch, Tian, Hardy, Hill, Finn, Davidson and Reid2022). The evaporation of the droplets and the diffusion of carbon dioxide can hardly explain the phenomena we observed because our data refer to droplets that only travel for such short times.
Destruction of viral proteins was confirmed through Western blot as shown in Figure 2c , d. Damage of viral membrane, spike and envelope proteins rises as humidity increases. While destruction of RNA can hardly be observed as the increase of humidity as shown in Figure S3 in the Supplementary Material. These results indicate that spontaneous generation of ROS in microdroplets disrupts proteins on the membrane of SARS-CoV-2, and thus leads to a decrease in viral infectivity. In addition, we have also used a standard ROS initiator (H2O2) to treat the virus solution and the results of the NTA test (see Figure S4 in the Supplementary Material), and we found that Coomassie brilliant blue staining (see Figure S5 in the Supplementary Material), and BCA assay (see Figure S6 in the Supplementary Material) indicated that the viral particles remain constant, but viral proteins gradually degrade with increased ROS. Therefore, the results further confirm that the change of ROS concentration in microdroplets in response to humidity changes may be one of the factors affecting viral infectivity. Our results suggest that the initial activity of viruses that enter the air by exhalation from a sick individual may exhibit seasonal variations caused by changes in the relative humidity of the air surrounding water droplets containing SARS-CoV-2.
Conclusion
In this study, we found that SARS-CoV-2 viral infectivity in microdroplets decreases as the relative humidity of air surrounding microdroplets increases. Moreover, the decrease in viral infectivity tends to be nearly constant when the humidity is greater than 60%. This trend is basically consistent with the observation that ROS concentration in microdroplets rises with humidity. This phenomenon can be attributed to the damage of proteins on the membrane of virus caused by ROS in microdroplets. Our findings provide a new perspective for understanding the seasonal spread of COVID-19 disease, and it is certainly expected to explain the seasonal variation observed for other infectious viral respiratory diseases. These findings strongly suggest the need to control indoor relative humidity for reducing the spread of viral respiratory infections, such as COVID-19.
Open peer review
To view the open peer review materials for this article, please visit http://doi.org/10.1017/qrd.2024.7.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/qrd.2024.7.
Data availability statement
All data may be requested from the corresponding authors.
Author contribution
Y.X., P.P., L.R., B.C., and R.N.Z. designed the research. Y.L. and L.C. performed the viral experiments. Y.X. analyzed the ROS concentration in microdroplets. B.C. and R.N.Z. wrote the manuscript. All authors provided critical commentary and contributed to editing the manuscript.
Financial support
This work represents a collaboration in which work at Jianghan University was financially supported by the National Natural Science Foundation of China (Grant No. 22376080), work at Shenzhen Bay Laboratory was supported by the National Natural Science Foundation of China (Grant Nos. 82222035 and 82372106), and work at Stanford University by the US Air Force Office of Scientific Research through the Multidisciplinary University Research Initiative (MURI) program (AFOSR FA9550-21-1-0170). Experiments related to authentic SARS-CoV-2 were conducted in Guangzhou Customs District Technology Center ABSL-3 Laboratory.
Competing interest
The authors declare none.
Comments
Please find a manuscript titled " As Air Relative Humidity Increases, Infectivity of SARS-CoV-2 Decreases Within Water Droplets” by Yu Liu, Lei Cao, Yu Xia, Lang Rao, Bolei Chen, and Richard N. Zare. This work represents an international collaboration between three different labs, one in the USA, and two in China. In a previous recent publication, M. Mofidfar, M. A. Mehrgardi, Y. Xia, and R. N. Zare, “Dependence on Relative Humidity in the Formation of Reactive Oxygen Species in Water Droplets,” Proc. Natl. Acad. Sci. (USA) 121, e2315940121 (2024), it was shown that reactive oxygen species, such as the hydroxyl radical and hydrogen peroxide, increase on the surface of water microdroplets when the relative humidity of the surrounding air increases. This publication suggested that this might explain the seasonality of infectious viral diseases, such as COVID-19 and influenza A. This study provides DIRECT PROOF that the infectivity of the SARS-CoV-2 virus in a water droplet increases as the relative humidity of the air surrounding the droplet decreases. Thus, this work should attract much attention because it has direct implications for a change in public health policy from that of stressing ventilation of indoor air to one that also asks that the relative humidity be controlled and raised to a healthy level. Relative humidity is not claimed to be the only important factor in reducing the spread of communicable viral respiratory infections, but this work does demonstrate that it should not be overlooked, as has been the case in the past.