Universal and reusable virus deactivation system for respiratory protection

Abstract

Aerosolized pathogens are a leading cause of respiratory infection and transmission. Currently used protective measures pose potential risk of primary/secondary infection and transmission. Here, we report the development of a universal, reusable virus deactivation system by functionalization of the main fibrous filtration unit of surgical mask with sodium chloride salt. The salt coating on the fiber surface dissolves upon exposure to virus aerosols and recrystallizes during drying, destroying the pathogens. When tested with tightly sealed sides, salt-coated filters showed remarkably higher filtration efficiency than conventional mask filtration layer, and 100% survival rate was observed in mice infected with virus penetrated through salt-coated filters. Viruses captured on salt-coated filters exhibited rapid infectivity loss compared to gradual decrease on bare filters. Salt-coated filters proved highly effective in deactivating influenza viruses regardless of subtypes and following storage in harsh environmental conditions. Our results can be applied in obtaining a broad-spectrum, airborne pathogen prevention device in preparation for epidemic and pandemic of respiratory diseases.

Full text

1
Scientific RepoRts | 7:39956 | DOI: 10.1038/srep39956
www.nature.com/scientificreports
Universal and reusable virus
deactivation system for respiratory
protection
Fu-Shi Quan1,*, Ilaria Rubino2,*, Su-Hwa Lee3, Brendan Koch2 & Hyo-Jick Choi2
Aerosolized pathogens are a leading cause of respiratory infection and transmission. Currently used
protective measures pose potential risk of primary/secondary infection and transmission. Here, we
report the development of a universal, reusable virus deactivation system by functionalization of
the main brous ltration unit of surgical mask with sodium chloride salt. The salt coating on the
ber surface dissolves upon exposure to virus aerosols and recrystallizes during drying, destroying
the pathogens. When tested with tightly sealed sides, salt-coated lters showed remarkably higher
ltration eciency than conventional mask ltration layer, and 100% survival rate was observed in
mice infected with virus penetrated through salt-coated lters. Viruses captured on salt-coated lters
exhibited rapid infectivity loss compared to gradual decrease on bare lters. Salt-coated lters proved
highly eective in deactivating inuenza viruses regardless of subtypes and following storage in harsh
environmental conditions. Our results can be applied in obtaining a broad-spectrum, airborne pathogen
prevention device in preparation for epidemic and pandemic of respiratory diseases.
Aerosols take a prominent role in airborne transmission of respiratory diseases. Droplets with aerodynamic size
(da) < 10 μ m and 10 < da < 100 μ m are known to infect the alveolar regions and upper respiratory tract, respec-
tively1,2. Notably, aerosols can also be a route of infection in diseases that, contrary to for instance inuenza, do
not specically target the respiratory tract, as it could be the case of Ebola virus3. While vaccination can greatly
reduce morbidity and mortality, during a pandemic or epidemic new vaccines matching the specic strain would
be available, at the earliest, six months aer the initial outbreak. Additionally, following development of an eec-
tive viral vaccine, several potential problems would remain, such as limited supply due to insucient production
capacity and time-consuming manufacturing processes. As a result, individuals close to the point of an outbreak
would be in imminent danger of exposure to infectious diseases during the non-vaccine period. In the absence
of vaccination, respirators and masks can be worn to prevent transmission of airborne pathogenic aerosols and
control diseases, such as inuenza4.
e main alternative, the N95 respirator, requires training prior to use, must be expertly tted to address the
risk of faceseal leakage at the face-mask interface, and must be disposed of as biohazard5. Due to these factors,
the use of N95 respirators on a large scale is impractical and expensive during an epidemic or pandemic. Past
experiences of severe acute respiratory syndrome (SARS), H1N1 swine u in 2009, and Middle East respira-
tory syndrome (MERS) indicate that surgical masks have been most widely adopted by the public as personal
protective measure, despite controversy on their eectiveness6–9. Currently, among other factors, ltration in
respirators and masks depends on lter characteristics, including ber diameter, packing density, charge of bers
and lter thickness, as well as particle properties, such as diameter, density and velocity10–14. However, in the lack
of a system to deactivate the collected pathogens, safety concerns naturally arise about secondary infection and
contamination from virus-laden lter media during utilization and disposal. Furthermore, since re-sterilization is
not possible without causing damage, respirators and masks are recommended for single use only9,15,16. Scientic
eorts have been focused on treatment of lters with materials possessing well-known antimicrobial properties,
such as iodine, chlorine and metals17–25, although with limited eectiveness against virus aerosols26–28. erefore,
a key challenge is the development of an easy-to-use, universal virus negation system, which is reusable without
1Department of Medical Zoology, Kyung Hee University School of Medicine, Seoul, 130-701, Korea. 2Department
of Chemical and Materials Engineering, University of Alberta, Edmonton, AB T6G 1H9, Canada. 3Department of
Biomedical Science, Graduate School, Kyung Hee University, Seoul, 130-701, Korea. *These authors contributed
equally to this work. Correspondence and requests for materials should be addressed to H.J.C. (email: hyojick@
ualberta.ca)
Received: 04 August 2016
Accepted: 30 November 2016
Published: 04 January 2017
OPEN

www.nature.com/scientificreports/
2
Scientific RepoRts | 7:39956 | DOI: 10.1038/srep39956
reprocessing and capable of deactivating pathogens, thereby reducing potential risk of secondary infection and
transmission.
Here, we report a simple but ecient virus inactivation system exploiting the naturally occurring salt recrys-
tallization. Our strategy is to modify the surface of the brous ltration layer within masks with a continuous
salt lm for virus deactivation via two successive processes: i) salt is locally dissolved by the viral aerosols and ii)
supersaturation is followed by evaporation-induced salt recrystallization. Consequently, viruses are exposed to
increasingly higher concentrations of saline solution during drying and physically damaged by recrystallization.
Results
Preparation and characterization of salt-functionalized lters. To demonstrate the concept of virus
deactivation system based on salt recrystallization, the middle layer of three-ply surgical mask, polypropylene
(PP) microber lter, was coated with NaCl salt as an active virus negation unit (see SupplementaryFig.S1 for
bare PP lter). e coating formulations contained surfactant to enhance wetting of saline solution on the sur-
face of hydrophobic PP bers. Bare PP lters (abbreviated as Filterbare) were pre-wet to contain about 600 μ L of
coating solution (abbreviated as Filterwet). e amount of NaCl salt (Wsalt in mg/cm2) coated on the lter per unit
area, considering that the lters thickness is constant, was easily controlled by changing the coating solution
volume (Vsalt in μ L) during drying of pre-wet lter (radius: 3 cm, Wsalt = 3.011 + 0.013 × V s at, n = 7) (Fig.S2).
Scanning electron microscopy (SEM) and energy dispersive X-ray (EDX) mapping analysis showed the for-
mation of homogeneous NaCl coating during drying, as also conrmed by X-ray diraction (XRD) (Fig.1a,b
and SupplementaryFig.S3). Both the formation of NaCl coating on PP bers and presence of surfactant in the
coating formulation appeared to alter the lter surface properties from hydrophobic (bare lter; contact angle,
θ
c = 133.0 ± 4.7°) to completely hydrophilic (salt-coated lter; θ
c ~ 0°, n = 10) (Fig.1c and SupplementaryFig.S4).
Hydrophilic nature of salt coating can greatly improve adhesion of viral aerosols to PP fibers compared to
Filterbare, as seen in Raman microscope images (Fig.1d and SupplementaryFig.S5).
Figure 1. Mask with salt-coated lter for prevention and deactivation of airborne pathogens. (a) SEM
image of Filterwet+600μL (top le) and EDX mapping images of Na (red), Cl (green), and NaCl (combination of
Na and Cl mapping images), showing the formation of NaCl coating, as also conrmed by XRD spectra (b)
of Filterbare, Filterwet, Filterwet+100μL, Filterwet+300μL, Filterwet+600μL, Filterwet+900μL and Filterwet+1200μL (labelled as
Bare, wet, wet+ 100 μ L, wet+ 300 μ L, wet+ 600 μ L, wet+ 900 μ L and wet+ 1200 μ L, respectively; miller indices
corresponding to NaCl crystal are shown at the top of XRD spectra for each position). (c) Optical microscope
images for contact angle measurements using 3 μ L DI water droplets on (i) Filterbare and (ii) Filterwet+600μL
(n = 10). (d) Microscope images of aerosol on (i) Filterbare and (ii) Filterwet+600μL (n = 10).

www.nature.com/scientificreports/
3
Scientific RepoRts | 7:39956 | DOI: 10.1038/srep39956
Filtration eciency against viral aerosols and protective ecacy in vivo. Filtration eciency of
salt-coated lters was tested against aerosols with volumetric mean diameter (VMD) of 2.5–4 μ m containing
H1N1 pandemic inuenza virus (A/California/04/2009, abbreviated as CA/09) at dierent pressure conditions
(see Fig.2a for transmission electron microscope (TEM) image of H1N1 virus). Interestingly, as shown in Fig.2b,
Filterbare did not exhibit any signicant level of resistance against penetration of virus under our experimental
conditions (i.e., 0% ltration eciency). Conversely, salt-coated lters showed substantially increasing ltra-
tion eciency with pressure and amount of coated salt. In particular, in the case of Filterwet+600μL, ltration e-
ciency varied from 43 to 70%, with increasing pressure from 3 to 17 kPa, and Filterwet+1200μL exhibited persistent,
high-level eciency (~85%) (one-way ANOVA, P = 0.85).
To probe the eects of ltration eciency on protective ecacy, in vivo experiments were performed using
mice intranasally (IN) infected with penetrated dosages of H1N1 virus under breathing pressure (~10 kPa)29.
As shown in Fig.2c, similarly to negative control groups (mice infected with lethal dose of virus stock and
aerosolized virus), mice exposed to a dose penetrated through the bare lter showed rapid body weight loss,
followed by death within 10 days aer infection, in good agreement with the observed 0% ltration eciency
(Fig.2b). In contrast, mice groups exposed to virus derived from salt-coated lters resulted in 100% survival rate
(Fig.2d). Furthermore, lungs of mice from negative control groups exhibited severe lung infection 4 days aer
challenge (Fig.2e). Conversely, mice groups exposed to virus derived from salt-coated lters showed signicantly
lower levels of lung viral titers (t-test, P < 0.005). is is consistent with lower levels of inammatory cytokines,
interferon-γ (IFN-γ ), from salt-coated lter groups compared to negative control and bare lter groups (t-test,
P < 0.001) (Fig.2f).
Deactivation of virus on salt-functionalized lters. Inuenza virus stability tests were performed to
investigate the eects of salt coating. e same amount of recovered viruses from the PP bers was used, and, in
the case of bare lters, viral aerosols exposure was conducted in the absence of pressure due to 100% penetra-
tion of viral aerosols. Unlike bare lters (Fig.S6a(i)), formation of micron-sized NaCl phase represents a typical
feature of salt-coated lters due to recrystallization of NaCl salt, following local dissolution upon aerosols expo-
sure (SEM images in Fig.S6a, ii to iv, and EDX mapping in Fig.S6b). In contrast to 8% HA activity loss of virus
adsorbed onto Filterbare, salt-coated lters exhibited almost complete HA activity loss within 5 min of incubation
(Fig.3a). Such dramatic virus destabilization on salt-coated lters is further supported by negligible levels of
viral titers compared to Filterbare with incubation time (t-test, P < 0.001) (Fig.3b). It is also noted that virus titers
exhibited signicant decrease with increase of incubation time and amount of coated salt (ANOVA general lin-
ear model, P < 0.001). TEM analysis showed that inuenza virus on Filterbare experiences morphological change
into non-spherical shape during aerosol drying (Fig.3c(i)). Notably, inuenza virus was severely damaged on
salt-coated lters even at 5 min of incubation (Fig.3c(ii)). From microscopic analysis, aerosol drying time was
about 3 min, indicating that destruction of virus observed at 5 min is associated with drying-induced salt crystal-
lization. Physical damage of virus due to crystallization was similarly reported as a major destabilizing factor of
inactivated inuenza virus30,31. Lower levels of native uorescence and nile red uorescence from virus recovered
from salt-coated lters accounted for more severe conformational change of antigenic proteins and destabili-
zation of viral envelope, respectively, consistent with TEM analysis (t-test, P < 0.001) (Fig.3d). In parallel, we
investigated the separate eect of salt concentration increase on virus stability during the aerosol drying process,
irrespective of crystal growth. As displayed in SupplementaryFig.S7, the materials collected in suspension from
Filterwet+600μL induced visible morphological transformation of the virus (SupplementaryFig.S7b) compared to
suspension of Filterbare (SupplementaryFig.S7a). is can be attributed to the high salt/surfactant concentra-
tion and osmotic pressure, which have been well-known to destabilize proteins and viruses31–33. erefore, the
marked virus destabilization on salt-coated PP bers can be explained by the combined eects of salt concentra-
tion increase during drying and evaporation-induced salt crystallization.
To verify in vitro virus stability on the lters, an in vivo study was performed by infecting mice with virus
incubated for 60 min on PP lters. As shown in Fig.3e, Filterbare group exhibited 5% body weight loss at day 9
post-infection, reaching a body weight lower than that of salt-coated lter groups by 10–15%. us, signicantly
higher lung virus titers in the negative control group were observed in contrast to no detectable titers in the
salt-coated lter groups (Fig.3f).
Strain-nonspecic virus deactivation and eects of storage under harsh environmental conditions
on salt coating stability. Broad-spectrum protection of salt-coated lters against multiple subtypes of
viral aerosols was evaluated by investigating both lethal infectivity by penetrated virus in vivo and infectivity by
virus collected on lters during ltration in vitro using A/Puerto Rico/08/1934 (PR/34 H1N1) and A/Vietnam/
1203/2004 (VN/04 H5N1). Similarly to CA/09 H1N1, 100% of mice survived viral infection (PR/34 and VN/04),
with no evidence of weight loss, due to higher ltration eciency of salt-coated lter than that of bare lter
(Fig.4a). is is supported by no signicant level of viral titer in the lung. In addition, as shown in Fig.4b,
salt-coated PP lters destroyed adsorbed inuenza viruses irrespective of both subtypes and amount of coated
salts.
e stability of salt coating on PP bers was tested under harsh environmental conditions. Incubation at 37 °C
and 70% relative humidity (RH) for 1 day did not cause any signicant dierence in ltration eciency (t-test,
P = 0.718) (SupplementaryFig.S8). As a result, all mice infected with dosage of penetrated virus through the lter
stored at high temperature and RH displayed 100% survival with 7% of body weight loss (Fig.4c,d). Even aer
15 days of incubation, salts remained to coat PP bers (Fig.4e, and SupplementaryFig.S9a,b), despite change in
grain orientation due to recrystallization (Fig.4f, and SupplementaryFig.S10a,b).

www.nature.com/scientificreports/
4
Scientific RepoRts | 7:39956 | DOI: 10.1038/srep39956
Figure 2. Filtration eciency of salt-coated lters. (a) TEM image of CA/09 H1N1 inuenza virus.
(b) Pressure-dependent ltration eciency (n = 8–10, mean ± standard deviation (SD)). (c–f) Eects of
ltration eciency on protective ecacy in vivo. Body weight change of mice aer infection with the dosages
of penetrated virus (n = 12, mean ± SD) (c), survival rates (mean; 100% means that all mice in the group
survived as penetrated dosages were lower than lethal dose) (d), lung virus titers (n = 4, mean ± SD) (e), and
lung inammatory cytokine (interferon-γ (IFN-γ )) assay (n = 11, mean ± SD) (f). Legends: lters are labelled
as in Fig.1b.

www.nature.com/scientificreports/
5
Scientific RepoRts | 7:39956 | DOI: 10.1038/srep39956
Figure 3. Inactivation of virus adsorbed on salt-coated lters. (a,b) HA activity (a) and virus titer
(b) displaying the eects of incubation time on the remaining activity of virus (n = 4–8, mean ± SD). (c) TEM
images of viruses reconstituted, aer incubation for 5 and 60 min, from (i) Filterbare and (ii) Filterwet+600μL.
(d) Native uorescence/nile red uorescence of viruses incubated for 60 min (n = 12, mean ± SD). (e,f) Body
weight change of mice aer infection with virus recovered from lters aer incubation for 60 min (n = 12,
mean ± SD) (e), and lung virus titers (n = 6, mean ± SD) (f). Asterisk (*): below detection limit. Legends: lters
are labelled as in Fig.1b.

www.nature.com/scientificreports/
6
Scientific RepoRts | 7:39956 | DOI: 10.1038/srep39956
Figure 4. Strain- and environment-dependent performance of salt-coated lters. (a) Body weight change
of mice infected with penetrated PR/34 H1N1 and VN/04 H5N1 viruses through Filterwet+600μL (n = 12,
mean ± SD). (b) Virus titers of recovered viruses from bare and salt-coated lters (n = 4, mean ± SD; data
for Filterwet, Filterwet+600μL and Filterwet+1200μL are overlapped). (c,d) Body weight change (c) and survival rate
(d) of mice infected with dosage of penetrated virus through Filterwet+600μL before and aer exposure to harsh
environmental conditions (37 °C and 70% RH) for 1 day (lled square and open square overlap in (d)). (e) EDX
mapping image of NaCl-coated Filterwet+600μL aer incubation for 15 days at 37 °C and 70% RH (combination of
Na (red) and Cl (green) mapping images). (f) XRD spectra of Filterwet+600μL before and aer incubation at 37 °C
70% for 1 day and 15 days. Legends: lters are labelled as in Fig.1b.

www.nature.com/scientificreports/
7
Scientific RepoRts | 7:39956 | DOI: 10.1038/srep39956
Discussion
Development of a universally applicable, low-cost, and ecient mechanism for virus negation is regarded as a
major challenge in public health against general airborne biological threats. is led us to propose a new con-
cept of personal/public preventive and control measures using salt-recrystallization against pathogenic aerosols
based on two hypotheses. e salt-coating can enhance adsorption of virus on the lter bers and inactivate
virus by the increase of osmotic pressure followed by the crystallization of salts. As shown in Fig.2b, salt-coated
lters exhibited signicantly higher levels of ltration eciency than bare lters. Notably, the bacterial ltration
eciency (BFE) reported by the mask manufacturer is 99%. e dierent value of ltration eciency for bare
lters obtained under our experimental conditions may be partially due to the use of aerosols with dierent bio-
logical origins. e FDA-recognized ASTM F2101 – 14 standard for evaluation of BFE exposes surgical masks
to Staphylococcus aureus aerosols, by employing S. aureus ATCC 653834, which has an average diameter of about
1 μm. In this study, ltration eciency was calculated following exposure of bare and salt-coated lters to inu-
enza virus, which exhibits a smaller diameter than that of S. aureus by one order of magnitude. Additionally,
whereas during BFE evaluation all three layers of surgical masks are used, in this work ltration eciency refers
to mask lters (middle layer). It is worth noting that the conditions for BFE standard evaluation (such as ow
rate and time of application of ow) do not coincide with the experimental procedure we used for measurement
of the ltration eciency, which may further contribute to the dierent result. e enhanced ltration eciency
of salt-coated lters against inuenza virus aerosols as compared to bare lters can be explained by the observed
wetting of aerosols, favoring greater adhesion to salt-coated lters. Furthermore, the signicant improvement in
ltration eciency resulted in complete protection of mice against lethal inuenza aerosols, which demonstrates
the high level of protection provided by salt-coated lters, outperforming currently used bare lters.
Rapid loss of HA activity and viral infectivity on salt-coated lters can be explained by physical destruction of
virus during recrystallization of coated salts. When the salt-coated lter is exposed to virus aerosols, salt crystals
below the aerosol droplet dissolve to increase osmotic pressure to virus. Due to evaporation, the salt concentra-
tion of the droplet signicantly increases and reaches the solubility limit, leading to recrystallization of salt. As a
consequence, virus particles are exposed to increasing osmotic pressure during the drying process and are phys-
ically damaged by crystallization. As shown in Fig.3e,f, the superior advantage of physically destroying the virus
adsorbed to the salt-coated PP lters through natural salt crystallization process was further conrmed in vivo.
According to previous reports, hyperosmotic stress (> 541 mOsm) and crystallization induce membrane pertur-
bation with irreversible deformation of the viral envelope and structural virus damage, respectively, resulting in
infectivity loss of virus30,31. erefore, our data support that the extensive level of infectivity loss associated with
a salt recrystallization process caused by physical contact between virus aerosols and salt coating can be used in
developing virus negation systems that are reusable without reprocessing.
Similarly to CA/09 H1N1 aerosols, increased protection in vivo due to higher ltration eciency of salt-coated
lters compared to bare lters and deactivation of virus on salt-coated lters were observed following exposure
to PR/34 H1N1 and VN/04 H5N1 (Fig.4a,b). is suggests that salt-coated lters prevent virus penetration
and destroy virus attached to the lter in a non-specic way. Furthermore, the performance of salt-coated l-
ters was not degraded by storage at 37 °C and 70% RH, demonstrating that salt recrystallization-based lters
can ensure protection even under harsh environmental conditions. Notably, for demonstration of the concept
of salt-recrystallization based virus deactivation system, NaCl salt was used, which has a critical RH of 75% at
30 °C35. However, salts with higher critical RH can be easily used, such as ammonium sulfate, potassium chloride
and potassium sulfate, which have critical RH of 80%, 84% and 96.3% at 30 °C, respectively35. is suggests that
salt-coated lters may be developed for specic environmental conditions.
In conclusion, we demonstrated that the developed salt-recrystallization based ltration system provides high
ltration eciency and successfully deactivates multiple subtypes of adsorbed viruses. Moreover, we have shown
that stability of the salt coating is not compromised by high temperature and humidity, which suggests safe use
and long-term storage/reuse at such environmental conditions. Although our tests are based on exposure to dif-
ferent types of inuenza virus, the signicance of these results for personal and public protective measures may
be generally extended to enveloped respiratory viruses where infection and transmission can occur by aerosol.
Our salt-coated lter unit can promise the development of long-term stable, versatile airborne pathogen negation
system, without safety concerns. In fact, the destruction mechanism of viruses solely depends on the simple, yet
robust naturally occurring salt recrystallization process, combining the destabilizing eects of salt crystal growth
and concentration increase during drying of aerosols. is idea can be easily applied to a wide range of existing
technologies to obtain low-cost, universal personal and public means of protection against airborne pathogens,
such as masks and air lters in hospitals. erefore, we believe that salt-recrystallization based virus deactivation
system can contribute to global health by providing a more reliable means of preventing transmission and infec-
tion of pandemic or epidemic diseases and bioterrorism.
Methods
Bare and salt-coated lter samples preparation. e commercial surgical masks had a three-ply
structure. e middle layer is the lter media, whereas the inner and outer layers provide support and protect
the lter against wear and tear. e metal nose clips and elastic ear loops were removed and circular samples
(radius: 3 cm) were cut from the masks. e PP lters (middle layer) were isolated by removing the inner and
outer protective layers (bare lters, Filterbare). e coating solution was prepared by dissolving sodium chloride
(NaCl; Sigma Aldrich, St. Louis, MO) in ltered DI water (0.22 μ m pore size; Corning, Tewksbury, MA) under
stirring at 400 rpm and 90 °C, followed by the addition of Tween 20 (Fisher Scientic) to a nal concentration of
29.03 w/v% of NaCl and 1 v/v% of Tween 20. To obtain the salt-coated lters, the mask bare PP lters were pre-wet
to contain approximately 600 μ L of coating solution by incubating overnight at room temperature. Any remain-
ing dry areas were removed by applying gentle strokes with tweezers to the lters while immersed in the coating

www.nature.com/scientificreports/
8
Scientific RepoRts | 7:39956 | DOI: 10.1038/srep39956
solution. Subsequently, the lters were deposited in the desired volume of coating solution (0, 100, 300, 600,
900 and 1200 μL, of which corresponding membranes are abbreviated as Filterwet, Filterwet+100μL, Filterwet+300μL,
Filterwet+600μL, Filterwet+900μL, and Filterwet+1200μL, respectively) on petri dishes (60 × 15 mm; Fisher Scientic) to
control the amount of NaCl per unit area and dried in an oven (Isotemp Incubator, Fisher Scientic) at 37 °C
for 1 day.
Inuenza virus preparation. Inuenza viruses A/California/04/2009 (CA/09, H1N1), A/Puerto Rico/8/34
(PR/34, H1N1) and A/Vietnam/1203/2004 (VN/04, H5N1) were grown in 10-day old embryonated hen eggs, in
which H5N1 virus was derived by reverse genetics from HPAI A/Vietnam/1203/200436. Inuenza viruses were
puried from allantoic uid using discontinuous sucrose gradient (15%, 30% and 60%) layers following the pre-
viously reported procedure37.
Aerosols exposure to lters. For experiments involving aerosols exposure, an aerosol chamber (L × W ×
H = 145 × 145 × 150 mm; Emka Inc., Middletown, PA) was used (Fig.S11). It has a connection to the vacuum
line and a circular aperture in the top wall (diameter: 22 mm) to exactly accommodate the cylindrical part (diam-
eter: 20 mm, height; 20 mm) of the nebulizer unit that is below the aerosol generator (Aeroneb Lab Nebulizer
System; Aerogen, Galway, Ireland). Bleach was used as trap between the chamber and the vacuum pump (Welch
2522C-10, 22 L/min; Niles, IL). e lters were placed on top of the chamber aperture and the nebulizer unit
was inserted, ensuring the tight seal of the lters against the side of the aperture. 5 μ L of virus stock were added
to the nebulizer unit, aerosols (VMD 2.5–4 μ m from manufacturer specications) were generated for 30 sec and
subsequently the desired vacuum level (3, 10 or 17 kPa) was applied, by manual control, three times in 1 sec cycles.
Notably, in the case of bare lters, pressure was only applied for ltration eciency tests.
For all assays and analysis, suspensions of the lters were prepared as follows, unless otherwise indicated. To
reconstitute virus adsorbed onto lters, virus-laden lters were immersed in 400 μ L of sterilized DI water for
about 5 min, and then removed aer vortexing from the suspension. e virus suspension was centrifuged at
19,800 g and 4 °C for 10 min (Centrifuge 5810 R, Eppendorf, Hauppauge, NY), followed by resuspension of pellets
in 70 μ L of DI water to eliminate any interference from materials in supernatant during assays.
Filtration eciency tests. e lters were exposed to the virus aerosols at 3, 10 and 17 kPa and suspen-
sions of the lters were obtained, as described above. e ltration eciency was calculated as the ratio of the
amount of virus (i.e., total proteins measured from the virus) reconstituted from the lter to that from the virus
in the exposure aerosols. e concentration of virus in aerosols was determined by generating viral aerosols into
a 15 mL centrifuge tube, containing 1 mL of DI water. Aer vortexing, virus concentrations (i.e., total protein
concentration) were measured with bicinchoninic acid assay (BCA protein assay kit; ermo Fischer scientic,
Waltham, IL) with bovine serum albumin as a standard. In the case of virus reconstituted from salt-coated lters,
virus-laden lter suspension was replaced with DI water prior to BCA assay.
In vivo infection tests. Lethal infectivity of inuenza viruses (CA/09 H1N1) was examined in 8 week old
female inbred BALB/c mice (Nara Biotech; Seoul, Korea) by using the intranasal route. For bare and salt-coated
lters, 12 mice per group were infected with individual penetration dosage of inuenza virus through each lter.
e penetration dosage of the virus through the lters (Filterbare, Filterwet, Filterwet+600μL, and Filterwet+1200μL) was
calculated from the ltration eciency at 10 kPa (near breathing pressure) using the relationship: penetration dos-
age = virus dosage in lethal aerosol × penetration eciency (%)/100, where penetration eciency (%) = 100 −
ltration eciency (%). To examine the eects of the aerosolization process on the viral infectivity change, two
mice groups were infected with a lethal dose of virus before and aer aerosol formation, which served as negative
control groups. Body weight changes and survival rate of mice were monitored daily for 15 days. Mice with body
weight loss greater than 25% were euthanized. All animal protocols were approved by the Kyung Hee University
(KHU) Institutional Animal Care and Use Committee (IACUC). All animal experiments and husbandry involved
in this work were conducted under the approved protocols and guidelines of KHU IACUC. KHU IACUC oper-
ates under National Veterinary Research and Quarantine Service (NVRQS), and animal welfare law and regula-
tions of the WOAH-OIE (World organization for animal health).
To test strain-dependent lethal infection behavior, mice (12 per group) were infected with the penetrated dos-
age of viral aerosols (PR/34 H1N1 and VN/04 H5N1 viruses) through Filterwet+600μL at 10 kPa. Time-dependent
body weight change was monitored in the same manner described above.
Lung viral titer and lung inammatory cytokine assays after infection. On day 4 aer infection 6
mice of each group were sacriced for the collection of lung samples. Lung virus titers were measured on six-well
plates containing conuent MDCK cell monolayers. Inammatory cytokines (IFN-γ ) were determined using BD
OptEIA mouse IFN-γ ELISA kit (BD Biosciences, San Jose, CA) following the manufacturer’s procedure.
Test of viral infectivity change on lters. To investigate the eects of salt-coating on viral infectivity
loss, lethal inuenza aerosols were exposed to four dierent types of lters (Filterbare, Filterwet, Filterwet+600μL, and
Filterwet+1200μL). Since Filterbare exhibited almost complete penetration upon pressure application, aerosols were
exposed to the bare lter in the absence of pressure and samples were carefully handled to prevent mechanical
agitation. To measure time-dependent stability change of virus, virus-laden lters were incubated at ambient
conditions for 0, 5, 15, and 60 min aer aerosol exposure, and suspended in DI water to reconstitute virus at each
time point. In vitro stability of virus was characterized by measuring hemagglutinin activity (HA) and virus titers
at the same concentration as lethal dose30. e conformational stability of antigenic proteins was characterized
by measuring intrinsic uorescence using 0.1 mg/mL of virus suspension38. To investigate morphological change
of virus, lipid stability of viral wall was characterized by nile red uorescence (Sigma Aldrich), a uorescent lipid

www.nature.com/scientificreports/
9
Scientific RepoRts | 7:39956 | DOI: 10.1038/srep39956
stain, following manufacturer’s protocol39. A decrease in uorescence intensity can be used to examine the level
of disintegration of the virus. Both intrinsic and nile red uorescence were measured by using a uorimeter (LB
50B; PerkinElmer, Waltham, MA). Intensity changes of uorescent spectra were compared relative to those of a
control from virus stock.
To test infectivity dierence observed from in vitro ndings, in vivo study was performed for the virus recon-
stituted from the lters (Filterbare, Filterwet, Filterwet+600μL, and Filterwet+1200μL) aer incubation for 60 min at RT
(aerosol exposure at 10 kPa, except for Filterbare). 12 mice per group were infected with a lethal dose of virus
collected from each type of lter. Body weight change and lung virus titers were measured as described above.
Eects of environmental conditions on the performance of salt-coated lter. Salt-coated lters
(Filterwet, Filterwet+600μL, and Filterwet+1200μL) were stored at 37 °C, 70% RH in an incubator (Maru Max; Rcom,
Gyeonggi-do, South Korea) for 15 days. Every day, the lters were collected and incubated at ambient conditions
for 5 min. At 1-day incubation, ltration eciency was measured at 10 kPa from Filterwet+600μL, followed by in vivo
infection test. Lethal infectivity between two dierent lter groups (before and aer incubation at 37 °C, 70% RH)
was compared by measuring body weight change and survival rate of mice aer exposure to lethal CA/09 H1N1
aerosols. XRD analysis was performed to salt-coated lters incubated for 1 and 15 days, and SEM/EDX mapping
analysis for 15-day incubated samples.
Contact angle measurements and imaging of aerosols. e bare and salt-coated lters were xed
with carbon tape (Ted Pella, Inc., Redding, CA) to a metal, at substrate and 3 μ L of DI water were added on the
surface of the lters. e contact angles were measured from images collected with an optical microscope (10×
lens, Motic SMZ-140; Motic, Richmond, Canada) at RT. Images of aerosols on lter bers were obtained using a
dispersive Raman microscope (Nicolet Almega XR; Fisher Scientic).
Aerosol drying time on lters. e bare and salt-coated lters were xed with carbon tape to a metal, at
substrate and exposed to aerosols generated from 5 μ L of Sulforhodamine B Dye solution (1 mM, Sigma-Aldrich).
Aerosol drying time was determined with timer by observation with optical microscope.
Electron microscopy analysis. For virus stability tests, bare and salt-coated lters were exposed to CA/09
H1N1 aerosols and, aer 5 and 60 min incubation, virus was recovered by suspension of the lters, as described
above. To study the eects of the coating formulation during aerosol drying independently from crystal growth,
bare and salt-coated lters were immersed in DI water and removed aer 60 min. Subsequently, virus was incu-
bated in the obtained suspension for 60 min. Additionally, the virus suspension was centrifuged at 19,800 g and
4 °C for 10 min to collect the samples and suspend them in DI water. For TEM analysis (200 kV, JEOL JEM 2100;
JEOL, Peabody, MA), samples were deposited on copper grid (Electron Microscopy Sciences, Hateld, PA) and
negatively stained with solution comprised of phosphotungstic acid hydrate (1.5 w/v%, pH = 7.0; Sigma-Aldrich,
Oakville, Canada).
To identify the morphology of salt-coated lters and recrystallized salts, SEM/EDX analysis was performed
for bare and salt-coated lters aer coating with 10 nm thick gold layer. Scanning electron microscopy analysis
(Hitachi S-3000N; Hitachi, Toronto, Canada) was operated in secondary electron mode at 20 kV and EDX analy-
sis was obtained with EDX detector (Oxford Instruments, Concord, MA).
XRD analysis. To conrm the formation of crystalline NaCl coating during drying process and its stability
during storage at 37 °C and 70% RH, XRD analysis (BRU-1098; Bruker, Billerica, MA) was performed at dierent
coating conditions. Filters (1 × 1 cm) were mounted on a slide glass for XRD analysis (θ –2θ mode) using a CuKα
radiation.
Statistical analysis. To compare multiple conditions, Student’s t-test, One-way analysis of variance
(ANOVA), and general linear model were used (Minitab release 14; Minitab, State College, PA). P value of less
than 0.05 was considered to be signicant.
References
1. Tellier, . eview of aerosol transmission of inuenza A virus. Emerging Infect. Dis. 12, 1657–1662 (2006).
2. Weber, T. P. & Stilianais, N. I. Inactivation of inuenza A viruses in the environment and modes of transmission: a critical review.
J. Infect. 57, 361–373 (2008).
3. Duy, J. et al. Circulating microNA proles of Ebola virus infection. Sci. ep. 6, 24496 (2016).
4. Morse, S. S., Garwin, . L. & Olsiewsi, P. J. Next u pandemic: what to do until the vaccine arrives? Science 314, 929–929 (2006).
5. CDC. Laboratory performance evaluation of N95 ltering facepiece respirators, 1996. MMW, Morbidity and Mortality Weely
eport 47, 1045 (1998).
6. Seto, W. H. et al. Eectiveness of precautions against droplets and contact in prevention of nosocomial transmission of severe acute
respiratory syndrome (SAS). Lancet 361, 1519–1520 (2003).
7. Bałazy, A. et al. Do N95 respirators provide 95% protection level against airborne viruses, and how adequate are surgical mass? Am.
J. Infect. Control 34, 51–57 (2006).
8. Loeb, M. et al. Surgical mas vs N95 respirator for preventing inuenza among health care worers: a randomized trial. JAMA 302,
1865–1871 (2009).
9. Bunyan, D., itchie, L., Jenins, D. & Coia, J. E. espiratory and facial protection: a critical review of recent literature. J. Hosp. Infect.
85, 165–169 (2013).
10. Zhu, C., Lin, C. H. & Cheung, C. S. Inertial impaction-dominated brous ltration with rectangular or cylindrical bers. Powder
Tec hno l. 112, 149–162 (2000).
11. Lee, . W. & Liu, B. Y. H. eoretical study of aerosol ltration by brous lters. Aerosol Sci. Tec h . 1, 147–161 (1982).
12. Duarte Fo, O. B., Marra, W. D., achan, G. C. & Coury, J. . Filtration of electried solid particles. Ind. Eng. Chem. es 39, 3884–3895
(2000).

www.nature.com/scientificreports/
10
Scientific RepoRts | 7:39956 | DOI: 10.1038/srep39956
13. raemer, H. F. & Johnstone, H. F. Collection of aerosol particles in presence of electrostatic elds. Ind. Eng. Chem. 47, 2426–2434
(1955).
14. anz, W. E. & Wong, J. B. Impaction of dust and smoe particles on surface and body collectors. Ind. Eng. Chem. 44, 1371–1381
(1952).
15. Siegel, J. D., hinehart, E., Jacson, M. & Chiarello, L. 2007 guideline for isolation precautions: preventing transmission of infectious
agents in health care settings. Am. J. Infect. Control 35, S65–S164 (2007).
16. Coia, J. E. et al. Guidance on the use of respiratory and facial protection equipment. J. Hosp. Infect. 85, 170–182 (2013).
17. atnesar-Shumate, S. et al. Evaluation of physical capture eciency and disinfection capability of an iodinated biocidal lter
medium. Aerosol Air Qual. es. 8, 1–18 (2008).
18. Lee, J. H., Wu, C. Y., Wysoci, . M. & Farrah, S. & Wander, J. Ecacy of iodine-treated biocidal lter media against bacterial spore
aerosols. J. Appl. Microbiol. 105, 1318–1326 (2008).
19. Badrossamay, M. . & Sun, G. Acyclic halamine polypropylene polymer: eect of monomer structure on graing eciency, stability
and biocidal activities. eact. Funct. Polym. 68, 1636–1645 (2008).
20. Zhao, N. & Liu, S. ermoplastic semi-IPN of polypropylene (PP) and polymeric N-halamine for ecient and durable antibacterial
activity. Eur. Polym. J. 47, 1654–1663 (2011).
21. Cerez, I., Worley, S. D., Broughton, . M. & Huang, T. S. Antimicrobial surface coatings for polypropylene nonwoven fabrics. eact.
Funct. Polym. 73, 1412–1419 (2013).
22. Davison, A. M. Pathogen inactivation and ltration ecacy of a new anti-microbial and anti-viral surgical facemas and N95 against
dentistry-associated microorganisms. International dentistry Australasian edition 7, 36–42 (2012).
23. Borow, G., Zhou, S. S., Page, T. & Gabbay, J. A novel anti-inuenza copper oxide containing respiratory face mas. PLoS One 5,
e11295 (2010).
24. Borow, G. & Gabbay, J. Putting copper into action: copper-impregnated products with potent biocidal activities. FASEB J. 18,
1728–1730 (2004).
25. Li, Y., Leung, P., Yao, L., Song, Q. W. & Newton, E. Antimicrobial eect of surgical mass coated with nanoparticles. J. Hosp. Infect.
62, 58–63 (2006).
26. Lee, J. H. et al. Assessment of iodine-treated lter media for removal and inactivation of MS2 bacteriophage aerosols. J. Appl.
Microbiol. 107, 1912–1923 (2009).
27. Lore, M. B. et al. Performance of conventional and antimicrobial-treated ltering facepiece respirators challenged with biological
aerosols. J. Occup. Environ. Hyg . 9, 69–80 (2012).
28. engasamy, S., Fisher, E. & Shaer, . E. Evaluation of the survivability of MS2 viral aerosols deposited on ltering face piece
respirator samples incorporating antimicrobial technologies. Am. J. Infect. Control 38, 9–17 (2010).
29. Windisch, W., Hennings, E., Sorichter, S., Hamm, H. & Criee, C. P. Pea or plateau maximal inspiratory mouth pressure: which is
best? Eur. espir. J. 23, 708–713 (2004).
30. Choi, H.-J. et al. Stability of inuenza vaccine coated onto microneedles. Biomaterials 33, 3756–3769 (2012).
31. Choi, H.-J. et al. Stability of whole inactivated inuenza virus vaccine during coating onto metal microneedles. J. Control. elease
166, 159–171 (2013).
32. Baldwin, . L. How hofmeister ion interactions aect protein stability. Biophys. J. 71, 2056–2063 (1996).
33. Choi, H.-J. et al. Eect of osmotic pressure on the stability of whole inactivated inuenza vaccine for coating on microneedles. PLoS
one 10, e0134431 (2015).
34. ASTM International. ASTM F2101 - 14 Standard Test Method for Evaluating the Bacterial Filtration Eciency (BFE) of Medical
Face Mas Materials, Using a Biological Aerosol of Staphylococcus aureus (2014).
35. Adams, J. . & Merz, A. . Hygroscopicity of fertilizer materials and mixtures. Ind. Eng. Chem. es. 21, 305–307 (1929).
36. Song, J.-M. et al. Protective immunity against H5N1 inuenza virus by a single dose vaccination with virus-lie particles. Virolo gy
405, 165–175 (2010).
37. Quan, F. S. et al. A bivalent inuenza VLP vaccine confers complete inhibition of virus replication in lungs. Vaccine 26, 3352–3361
(2008).
38. Choi, H.-J., Ebersbacher, C. F., im, M., ang, S. & Montemagno, C. D. A mechanistic study on the destabilization of whole
inactivated inuenza virus vaccine in gastric environment. PLoS one 8, e66316 (2013).
39. Greenspan, P., Mayer, E. P. & Fowler, S. D. Nile red: a selective uorescent stain for intracellular lipid droplets. J. Cell Biol. 100,
965–973 (1985).
Acknowledgements
is research was nancially supported by startup funds from University of Alberta (H.J.C.), and grants from
National Research Foundation of Korea (NRF) (NRF-2014R1A2A2A01004899) and Ministry of Health &
Welfare, Republic of Korea (HI15C2928).
Author Contributions
H.J.C. conceived and designed the experiments. F.S.Q., I.R., S.H.L., B.K., and H.J.C. performed the experiments.
F.S.Q., I.R., S.H.L., B.K., and H.J.C. analyzed the data. I.R. and H.J.C. wrote the manuscript. F.S.Q. and B.K. edited
the manuscript.
Additional Information
Supplementary information accompanies this paper at http://www.nature.com/srep
Competing nancial interests: e authors declare no competing nancial interests.
How to cite this article: Quan, F.-S. et al. Universal and reusable virus deactivation system for respiratory
protection. Sci. Rep. 7, 39956; doi: 10.1038/srep39956 (2017).
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional aliations.
is work is licensed under a Creative Commons Attribution 4.0 International License. e images
or other third party material in this article are included in the article’s Creative Commons license,
unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license,
users will need to obtain permission from the license holder to reproduce the material. To view a copy of this
license, visit http://creativecommons.org/licenses/by/4.0/
© e Author(s) 2017