Mucin co-localizes with influenza virus and preserves infectivity in deposited model respiratory droplets

Abstract

The stability of influenza virus in respiratory particles varies with relative humidity (RH) and protein content. This study investigated the decay, or loss of infectivity, of influenza A virus (IAV) in 1-μL respiratory droplets deposited on a surface with varying concentrations of mucin, one of the most abundant proteins in respiratory mucus, and examined the localization of virions within droplets. IAV remained stable at 0.1% and 0.5% mucin in phosphate-buffered saline (PBS) over 4 hr at 20%, 50%, and 80% RH, with a maximum decay of 1.2 log₁₀/mL. In contrast, in pure PBS droplets, the virus decayed by at least 2.6 log₁₀/mL after 4 hr at 50% and 80% RH. Mucin’s protective effect was independent of its concentration, except at 80% RH after 4 hr. Confocal microscopy of the particles revealed that at 20% and 50% RH, mucin led to thicker coffee rings and dendritic patterns where virions co-localized with mucin. At 80% RH, no morphological difference was observed between PBS-only and mucin-containing droplets, but virions still co-localized with mucin in the center of droplets with 0.5% mucin. Analysis by digital droplet PCR showed that mucin helped maintain virus integrity. To our knowledge, this is the first study to localize influenza virus in model respiratory droplets. The results suggest that mucin’s co-localization with virions in droplets may protect the virus from environmental stressors, enhancing its stability.

Graphical Abstract

Introduction

Influenza virus is primarily transmitted in respiratory particles, including small aerosol particles and larger droplets.^1–4 Successful transmission requires that the virus remain infectious in the environment. The stability of influenza virus and other viruses in aerosols and droplets has been shown to vary with temperature, relative humidity (RH), and the chemical composition of the suspending fluid.^5–19 Some studies have found that enveloped viruses exhibit a “U-shaped” relationship between viability and RH, i.e., the greatest loss in infectivity at intermediate RH, in phosphate-buffered saline (PBS) or culture media.^{5, 9, 10, 15, 17, 20, 21} This U-shaped relationship has also been reported for virus suspended in saliva.^{6, 18, 19} Furthermore, studies have shown that the dependence of virus infectivity on RH is suppressed when the suspending medium is supplemented by proteins.^{5, 10, 14, 15, 22} Therefore, it is critical to explore the effects of proteins on virus infectivity to understand the underlying mechanisms of virus inactivation and transmission.

Various proteins, including bovine serum albumin (BSA), porcine, chicken, or human albumin, and fetal bovine or calf serum (FBS or FCS), have been shown to protect viruses from inactivation across a wide range of RH.^{5, 10, 14, 15, 22, 23} Proteins may facilitate the formation of a semisolid, viscous phase in respiratory aerosols and droplets at intermediate RH, slowing the diffusion of solutes that can damage viruses.^{14, 24} Proteins in Dulbecco’s modified Eagle’s medium (DMEM) may co-locate with virions at the thick edge of dried droplets on surfaces, in a phenomenon known as the “coffee-ring effect”, which may protect viruses from inactivation.²² However, the same effect may be detrimental when virions co-locate with lysozyme, an antiviral protein present in saliva.¹⁹ These disparate findings related to the coffee-ring effect indicate that different proteins may play distinct roles. Studies are needed that involve physiologically relevant proteins, such as mucin, a glycoprotein that is one of the most abundant organic compounds in respiratory fluid.^25–27 Additionally, a virion’s microenvironment, or its position relative to proteins and other components of the respiratory particle is important for understanding inactivation mechanisms.^{13, 28, 29}

Respiratory mucus serves as the major barrier against virus penetration to the airway epithelia.^{30, 31} It consists of mucin, electrolytes, lipids, and other non-mucin proteins such as enzymes and antimicrobial peptides. In respiratory mucus, the most prevalent types of mucin are 5AC (MUC5AC) and 5B (MUC5B), both of which form gels.^{32, 33} It is also found in saliva as mucin 7 (MUC7), which does not form a gel.³⁴ Mucin concentration typically ranges from 0.1% to 0.5% w/v.³³ By binding to viruses, mucin reduces their penetration through mucus to infect respiratory cells.^{30, 31}

Although ample evidence suggests that respiratory mucus can protect IAV in the environment,^{6, 10–12, 35, 36} few studies have investigated whether this protection is attributable to mucin or to other factors. A study of IAV in 100-μL droplets applied to respirators demonstrated that the virus decayed faster in 5 mg/mL mucin compared to 2% FBS, but slower than in DMEM at 20% RH.³⁷ On stainless steel, the effect of mucin varied by virus strain and absolute humidity (AH).³⁸ However, both these studies^{37, 38} employed 100-μL droplets that are much larger than real respiratory droplets.¹⁷ Mucin protected aerosolized mouse hepatitis virus (MHV) only transiently when added to DMEM at 0.5% w/v¹⁶ and protected aerosolized bacteriophages MS2³⁹ and phi6²³ when added to artificial saliva at 0.3% w/v. Given the abundance and importance of mucin in the human respiratory system, it deserves more detailed analysis of its influence on virus stability.

To clarify the role of mucin in IAV inactivation, this study aimed to define the impact of mucin on IAV infectivity in evaporating droplets and to explore the mechanisms behind this effect. We assessed the role of mucin in a simplified system consisting only of mucin and PBS to tightly control the protein and salt concentrations and facilitate interpretation of results. We first analyzed the infectivity of IAV by plaque assay in evaporating, 1-μL respiratory droplets at different initial concentrations of mucin over 4 hr at 20%, 50%, and 80% RH. We further explored droplet morphology and localization of fluorescently labeled virions in evaporating droplets through confocal microscopy. Finally, we analyzed the integrity of virions after exposure to different concentrations of mucin using RNase treatment. Our findings provide valuable insights into the location of IAV in evaporating droplets and the mechanisms of virus inactivation.

Methods and Materials

Virus preparation and quantification

We cultured Madin-Darby canine kidney (MDCK) cells in Minimum Essential Medium (MEM) (Gibco, 12360038) supplemented with 1% (v/v) L-glutamine (Gibco, 25030081), 10% gamma irradiated FBS (VWR, 97068–086), and 1% penicillin-streptomycin (Pen/Strep) (Gibco, 15140122), maintained at 37°C and 5% CO₂. We passaged influenza virus A/California/07/2009 by growing 1:50,000 CP1 virus stock on confluent MDCK cells in infection media, which was MEM containing 1% L-glutamine, 1% antibiotic-antimycotic (Gibco, 15240062), and 1 μg/mL TPCK trypsin (Thermo Fisher, 20233).⁶ We harvested virus in the supernatant when cytopathic effect (CPE) was apparent. To remove extracellular materials, we centrifuged the supernatant at 200×g for 10 min. Virus stocks were aliquoted and stored at −80°C. We quantified infectious virus in stocks and samples through plaque assay, as described in detail in the Supporting Information (SI).

Virus infectivity in mucin-containing evaporating droplets

We suspended porcine stomach mucin (type III, Sigma-Aldrich, M1778) in 1×PBS (diluted from 10×PBS, Ambion, AM9624) at a final concentration of 1 mg/mL (0.1% w/v) or 5 mg/mL (0.5%), representing healthy individuals or those with chronic respiratory diseases, respectively.³³ The detailed composition of PBS can be found in Table S1. Porcine stomach mucin (including both type II and type III) has been used frequently as a surrogate for gel-forming human mucin^{14, 16, 23, 29, 39} and we therefore employed it in this study to facilitate comparison with other results. PBS without mucin served as the control. We diluted virus stocks 1:10 in the mucin solutions and control, and we then pipetted five 1 μL droplets per well on six-well polystyrene plates (Thermo Scientific, 140675). These droplets were exposed to three different RHs (20%, 50%, 80%) at room temperature inside an environmental chamber (Electro-Tech Systems 5504). The temperature and RH were monitored using a HOBO logger (Onset, UX100–011). After exposing the droplets in the chamber for 0, 0.5, 1, 2, or 4 hr, we resuspended them in 500 μL of MEM by pipetting up and down 10 times and then froze the samples at −80°C for subsequent analyses, including plaque assay and virus integrity analysis. We conducted three independent experiments, each including two technical replicates for every combination of mucin concentration and RH. We diluted virus stocks into mucin solutions or PBS as a control right before depositing droplets on surfaces, and we verified that virus stocks did not decay throughout the course of each independent experiment.

Droplet drying kinetics

We deposited five 1 μL droplets of PBS, 0.1% mucin in PBS, or 0.5% mucin in PBS on a polystyrene petri dish (Falcon, 351008), and measured their weight every 30 sec for 4 hr at 20%, 50%, and 80% RH inside the environmental chamber using a microbalance (Sartorius, MSE3.6P-000-DM). The changes in weight of the droplets reflected the evaporation rate of water from them. At the beginning of the period, droplet weight declined linearly. It then plateaued at a moment that we defined as the drying time. The microbalance was sensitive to very small environmental fluctuations, so taring prior to droplet addition can lead to negative readings as the droplets evaporated. We normalized all readings by setting the minimum mass to 0.1 mg and adjusting all mass readings upward by the same offset. We then calculated the weight percentage by dividing the adjusted weight at each time point by the initial adjusted mass.⁶

Virus labeling and droplet visualization

We labeled virus with Alexa Fluor 594 NHS Ester dye (Invitrogen, A37572) using the methods described in Bradley et al.⁴⁰ with modifications to visualize virions in the drying droplets. Briefly, we filtered 20 mL of virus stocks using 0.2 μm pore size filters to remove small cellular debris. We then ultracentrifuged the filtered stocks at 24,000×rpm for 2 hr at 4°C through a 30% sucrose cushion. We resuspended the virus pellets in 100 μL of the same PBS used for the mucin solutions by vigorously pipetting and added Alexa Fluor 594 at a final concentration of 0.1 μg/μL. We incubated the mixture of virus and dye at room temperature in the dark for 1 hr and removed extra dye through extensive dialysis using the Slide-A-Lyzer MINI Dialysis Device Kit (10K MWCO, Thermo Fisher, 69576). We also prepared negative controls by adding infection media alone to the cells, allowing them to grow in parallel with the virus-infected cells, and collecting supernatant. The supernatant went through the same labeling process as the virus suspension. All the labeled virus and negative controls were aliquoted and stored at −80°C until analysis.

We pipetted 1-μL droplets of fluorescently labeled virus suspended in different concentrations of mucin solutions onto a hydrophobic¹⁴ chambered coverglass with cover (Nunc Lab-Tek II, 155409). After incubating the droplets uncovered at 20%, 50%, or 80% RH at room temperature for 4 hr, we covered and sealed them with parafilm in the environmental chamber before moving them for imaging. We visualized the droplets immediately after the incubation period using a confocal laser scanning microscope (Zeiss, LSM 880) at 10× magnification. Mucin auto-fluoresced at an excitation wavelength of 488 nm, so there was spectral separation between labeled virus and mucin that allowed us to image both. We noticed that the energy emitted from the laser can heat the droplets quickly and thus alter the RH and droplet morphology within the chambered slides. Therefore, we chose the parameters to achieve a balance between fast scanning speed and image resolution, and did not scan droplets at higher resolution. For comparison, we also imaged fluorescently labeled virus in pooled human saliva (Innovative Research, IRHUSL250ML) at a dilution ratio of 1:20. Detailed imaging parameters are listed in Table S2. All the images were taken using the same parameters.

Virion integrity analysis by RNA extraction and ddPCR

We analyzed virion integrity following a protocol based on the hypothesis that an intact envelope and capsid protect viral RNA from being hydrolyzed by RNase.^{7, 41} Following RNase treatment and extraction, we would detect a higher concentration of RNA if the virions were intact and a lower concentration of RNA if they were damaged, because the RNase could access and degrade RNA. Briefly, we prepared RNA digestion buffer containing 10 mM Tris-HCl and 1 mM EDTA (Supelco, 93302) supplemented with 50 mM NaCl. We analyzed samples collected at time zero and after incubation for 4 hr at 50% RH for comparison. We mixed 15 μL of sample, 3 μL of RNase T1/A mix (Thermo Fisher, EN0551), and 132 μL of RNA digestion buffer, and incubated the mixture at 37 °C for 30 min. We generated “no RNase” controls by replacing RNase with an equal amount of PBS. We then added 1 μL of SUPERase-In RNase Inhibitor (Invitrogen, AM2694) to each sample and incubated it at room temperature for 20 min.

We immediately extracted RNA from the samples using the Qiagen QIAamp Viral RNA Mini kit according to the manufacturer’s protocol. The elution volume was 60 μL. We quantified RNA by droplet digital PCR (ddPCR) using the primers and probes reported by the Centers for Disease Control and Prevention,⁴² which were purchased from Integrated DNA technologies. Droplets were generated by the AutoDG Automated Droplet Generator (Bio-Rad) and analyzed by the QX200 Droplet Reader (Bio-Rad). Each plate contained duplicates of samples and no-template controls. Detailed sequences and experimental conditions can be found in SI (Table S3).

Statistics

Using R 4.3.3, we conducted one-way analysis of variance (ANOVA) followed by Tukey’s test to compare virus infectivity in droplets with different mucin concentrations, and Student’s t-test to examine virus integrity. We defined a threshold for significance of p < 0.05.

Results and Discussion

Virus infectivity in mucin-containing evaporating droplets

We analyzed the loss of IAV infectivity in 1-μL droplets containing PBS only (i.e., 0% mucin), 0.1% mucin in PBS, or 0.5% mucin in PBS at 20% RH, 50% RH, and 80% RH (Figure 1A–C). In general, the addition of mucin protected IAV from decay at all RHs. Decay is defined as the difference in titer, usually between a particular time point and time zero, unless specified otherwise. At 20% RH, virus decayed by an average of 1.4 log₁₀-PFU/mL after 0.5 hr for virus in PBS, and 2 log₁₀-PFU/mL after 4 hr (Figure 1A). However, in 0.1% mucin and 0.5% mucin, the average decay was <0.7 log₁₀-PFU/mL and <0.4 log₁₀-PFU/mL, respectively, after 4 hr (Figure 1A). Although the decay in 0.1% mucin was greater than in 0.5% mucin, the difference was not statistically significant. At 50% RH, IAV in PBS decayed on average by 1.9 log₁₀-PFU/mL during the first 0.5 hr and 2.6 log₁₀-PFU/mL after 4 hr (Figure 1B). Yet in mucin, the decay was much less: ~1.2 log₁₀-PFU/mL in 0.l% mucin and <1 log₁₀-PFU/mL in 0.5% mucin after 4 hr, with no significant differences between the two conditions (Figure 1B). At 80% RH, there was very little decay when mucin was present. The average decay over the first 0.5 hr was 1.3, 0.1, and 0.1 log₁₀-PFU/mL in PBS, 0.1% mucin, and 0.5% mucin, respectively (Figure 1C). This decay was less than that observed at 50% RH but similar to the decay at 20% RH. After 4 hr, the average decay increased substantially in PBS and less so with mucin, and there was significantly less decay in 0.5% mucin than in 0.1% mucin at the 4-hr time point (Figure 1C). This was the only case where we observed a difference in decay as a function of mucin concentration.

Figure 1.

IAV infectivity in 1-μL droplets containing PBS, 0.1% mucin in PBS, or 0.5% mucin in PBS at (A) 20% RH, (B) 50% RH, and (C) 80% RH. The y-axis represents log₁₀-PFU/mL, or plaque forming units per mL of sample after log transformation. Data points are jittered at each time point to improve readability. Each dot represents an independent replicate, which is the average of two technical replicates. Error bars represent mean ± standard deviation. Compact letter display is used to represent statistical differences. Different letters at the bottom of each group represent significant differences between groups, while the same letters indicate no significant difference. Dashed gray lines represent the limit of detection (LOD).

The observed protective effect of mucin is consistent with findings in the literature, where mucin was reported to protect aerosolized MHV,¹⁶ MS2,³⁹ and phi6²³ from decay, although the protective effect disappeared after 5 min for MHV. The protective effects of mucin alone were similar to those of respiratory mucus, which can limit the loss of IAV titer in 1-μL droplets to approximately 1 log₁₀-PFU/mL or less across various RHs after incubation for 1–2 hr.^{5, 6, 11} The mucin-PBS matrix was more protective than saliva.⁶ We further compared the mucin-PBS matrix to other protein-salt matrices from previous studies in terms of slowing IAV decay. In PBS with 3.5 mg/mL of FCS, the IAV titer decreased by ~2 log₁₀ at 84% RH after 3 hr of incubation.¹⁰ IAV decayed less (from ~1.5 log₁₀ to <1 log₁₀) when the mass ratio of human albumin to NaCl increased from 0.55 to 2.1 after 1 hr at 60% RH; the initial NaCl concentration was 8 mg/mL.⁵ Mucin appeared to offer similar protection as albumin and greater protection than FCS. However, these comparisons are limited, as the referenced studies utilized different virus strains and surface materials with different testing conditions and air compositions, negating direct comparison with our results.

We noticed that the degree of protection offered by mucin was independent of its initial concentration except at 80% RH after 4 hr. This result aligns with the finding by Alexander et al.¹⁶ Rockey et al. also found no association between the total concentration of proteins in respiratory mucus or saliva and IAV infectivity.⁶ Schaub et al. further reported a lack of association between infectivity and the initial salt or protein content of droplets.⁵ However, the researchers found that IAV inactivation was inversely correlated with the ratio of organics to salts, which was ultimately related to the efflorescence state of the droplets.⁵ In the following section, we investigated whether mucin protects IAV by altering the drying kinetics, or physical state (e.g., liquid or effloresced), of droplets.

The effect of droplet drying kinetics on virus infectivity

Several studies have suggested that the decay of IAV in evaporating droplets is biphasic: faster when the droplets are still wet and evaporating and slower after the droplets have equilibrated with their surroundings and, in many cases, dried completely.^{5–7, 17, 43} Therefore, a shorter evaporation period (a faster drying time) would be associated with less reduction in infectivity. We compared the drying time of 1-μL droplets containing PBS, 0.1% mucin in PBS, or 0.5% mucin in PBS, and found no significant difference as a function of mucin concentration at the same RH (Table 1, Figure S1), consistent with previous literature.⁴⁴ The drying time of all droplets ranged from 10.9 to 12.4 min, 15.9 to 18.9 min, and 50.1 to 57.9 min at 20%, 50%, and 80% RH, respectively. All the droplets effloresced at 20% and 50% RH around the drying time, but they did not effloresce at 80% RH (Figure 2). The drying time of the mucin-PBS matrices was shorter than that of respiratory mucus or saliva droplets of the same size measured on the same surface by 3–4 min at 20% RH, ~10 min at 50% RH, and ~20 min at 80% RH.⁶ We observed that the drying time of droplets was influenced by their shape, with spread-out, flattened droplets drying more quickly than compact, taller droplets, despite having the same initial volume (Figure S2). To ensure consistency in our measurements, we focused exclusively on the drying time of the compact droplets.

Table 1.

Drying time of 1-μL droplets containing PBS, 0.1% mucin in PBS, or 0.5% mucin in PBS. The numbers represent the averages of two independent replicates, with the standard deviations shown in parentheses. There was no significant difference in drying time for different concentrations of mucin at the same RH (p > 0.05, Student’s t-test).

Drying time (min)	PBS (0% mucin)	0.1% mucin	0.5% mucin
20% RH	12.4 (0.7)	10.9 (0.7)	11.4 (0.7)
50% RH	18.9 (1.8)	19.7 (1.4)	15.9 (0.4)
80% RH	54.3 (6.6)	57.9 (0.3)	50.1 (0.7)

Figure 2.

Confocal images showing localization of fluorescently labeled virions in drying droplets with different mucin concentrations at 20%, 50%, or 80% RH after 4 hr of incubation. For each combination of concentration and RH, pictures from left to right represent brightfield droplet morphology, fluorescently labeled virions (red, Alexa Fluor 594), and mucin (green autofluorescence). The scale bars represent 200 μm. One representative image from duplicate droplets is shown for each condition.

In our experiments, IAV exhibited the greatest decay within the first 0.5 hr in PBS at all three RHs (Figure 1). As the droplets were evaporating, the concentration of solutes within the droplets rose and reached a maximum, or even supersaturation, right before droplets effloresced.^{7, 9} This could lead to increasing osmotic pressure and subsequent loss of virus integrity in the droplets.^{7, 45} Thus, many studies have reported rapid virus inactivation before efflorescence of droplet or aerosols.^{6, 7, 13, 17, 43} A similar drying time among PBS, 0.1% mucin, and 0.5% mucin solutions indicated that IAV was exposed to enriched solutes for a similar duration in all three matrices, yet the presence of mucin significantly protected IAV in droplets at all three RHs. We thereby speculated that mucin, like many other organic compounds, assisted in forming a semisolid state in droplets, thus reducing the diffusivity of solutes, and potentially reducing the exposure of virions to enriched salts.^{14, 24} Furthermore, mucin concentrations increased during the evaporation period and could reach saturation before the droplets effloresced, regardless of initial concentration, as long as it was above a certain threshold. Therefore, there was no significant difference between 0.1% and 0.5% mucin in protecting IAV at 20% and 50% RH. However, virus decay continued after the evaporation period during the equilibrium period, particularly in PBS at 80% RH (Figure 1C), suggesting that other factors besides drying kinetics affected IAV inactivation.

Droplet morphology and virus localization after evaporation

After the droplets exited the linear evaporation period, they effloresced at 20% and 50% RH, or reached equilibrium as a liquid at 80% RH. The droplet morphology and localization of virus in PBS were similar between 20% and 50% RH (Figure 2). Each droplet shrank into a large, square-shaped salt crystal, accompanied by numerous smaller satellite crystals within the original droplet area.

Images suggest that virions, which were labeled with a fluorescent red dye, accumulated around the edges of the large crystal and in the center of small crystals, similar to the morphology of dried salt droplets supplemented with nanoparticles.⁴⁶ This drying pattern can result from counteracting effects of “Marangoni flow”, which results from the concentration gradient within droplets and transports non-volatile solutes from the edges to the center, and “capillary flow”, which drives solutes to the water-surface contact line due to surface tension and leads to the formation of a “coffee-ring”.^{47, 48} The images suggest that as the PBS droplets evaporated, the virions were either driven by Marangoni flow towards the large salt crystal, or served as nuclei of crystallization and scattered on the surface.

We observed a faint coffee-ring pattern in dried PBS droplets, and this effect became much more pronounced with the addition of mucin to the PBS at 20% and 50% RH (Figure 2). The mucin-containing droplets developed dendritic, feather-like patterns with a thick coffee ring at the edge. The ring became thicker as the mucin concentration increased. There were several small salt crystals in 0.1% mucin droplets at 20% RH but only one large salt crystal at 50% RH. All distinct salt crystals disappeared in 0.5% mucin droplets at both 20% and 50% RH. The addition of mucin caused most of the virions to cluster around the coffee ring and dendritic patterns close to the ring, which were primarily composed of mucin, as evidenced by green fluorescence images of the samples and the negative controls (Figure 2 and Figure S3). These results indicated that mucin might suppress the Marangoni flow by increasing the viscosity of the solution,⁴⁹ and this reduction in Marangoni flow allowed for the formation of a thicker coffee ring due to capillary flow.^{48, 50} The co-localization of mucin and virions in the coffee ring may have protected IAV from exposure to high concentrations of salts near the salt crystals. This finding is consistent with previous studies on other organic compounds.^{14, 22}

At 80% RH, the droplets did not fully evaporate or effloresce; instead, they formed a thin layer of liquid for all matrices (Figure 2). The fluorescence intensity was weaker in the droplets at 80% RH. This may be due to a higher water content at 80% RH, resulting in less concentrated solutes and virions. In PBS and 0.1% mucin droplets, the virions were nearly evenly dispersed throughout the thin liquid. In contrast, they accumulated slightly in the center of the 0.5% mucin droplets. Additionally, there was a notable co-localization of virions with mucin in the 0.5% mucin droplets, whereas the effect was less evident in 0.1% mucin droplets. This difference in mucin-virion localization may explain the significantly lower reduction in IAV infectivity observed with 0.5% mucin compared to 0.1% mucin at 80% RH after 4 hr of incubation (Figure 1C). Additionally, merged brightfield and green fluorescence images showed that much undissolved mucin (the black dots in the brightfield) was deposited in the center of the 0.5% mucin droplets, while few deposits were observed in the 0.1% mucin droplets (Figure 2). This suggests that the mucin concentration might have reached saturation in 0.5% mucin droplets but not in 0.1% mucin droplets during the equilibrium stage at 80% RH, producing differences in droplet viscosity and thus salt diffusivity.¹⁴

The drying morphology of mucin-PBS droplets was similar to that of respiratory mucus with thick coffee rings and dendritic patterns, which was not surprising given that they showed similar protective effects on IAV infectivity.⁶ This morphology has been reported for most protein-salt matrices.^{14, 51, 52} Several studies used fluorescent nanoparticles as virus surrogates in proteinaceous matrices to determine the localization of virus in droplets.^{6, 14} At 20% and 50% RH, nanoparticles successfully mimicked the localization of virions, clustering at the coffee ring and dendritic patterns.^{6, 14} However, at 80% RH, nanoparticles clustered at the edge of the thin liquid,⁶ whereas in our droplets, virions accumulated in the center with mucin (Figure 2 and Figure S5). We hypothesize that this difference was due to mucin binding to hemagglutinin on the IAV envelope,^{30, 34} leading to co-localization of mucin and virions. This co-localization was also found between mucin and coronavirus OC43 (a virus that binds to mucin) after droplets effloresced, where virions were almost coated by mucin as evidenced by transmission electron microscopy.⁵¹ Likely, IAV became coated with mucin in our experiments, offering another layer of protection to the virus envelope. However, co-localization and findings from past literature do not necessarily confirm biological interactions between mucin and IAV. Further research is required to elucidate the nature of mucin-IAV interactions.

We also compared the drying patterns of mucin-containing droplets with those of pooled human saliva, a real respiratory fluid, to determine whether the coffee-ring effect and co-localization remained consistent in saliva. The morphology differed between saliva and mucin-containing droplets (Figure 2 and Figure S4). At 20% and 50% RH, saliva showed line-like patterns, whereas mucin-containing droplets exhibited dendritic patterns. This observation is consistent with findings reported in previous studies.⁶ We added fluorescently labeled virus to saliva at a 1:20 dilution, resulting in much weaker red fluorescence compared Figure 2, making it harder to localize virions. Yet, we still observed faint red fluorescence in the coffee ring at 20% and 50% RH. Furthermore, we observed strong green fluorescence, possibly due to autofluorescence emitted by salivary proteins, which was particularly intense in the coffee ring area (Figure S4). This indicates a co-localization of virions and salivary proteins. However, another study found that co-localization of vesicular stomatitis virus and lysozyme in the coffee ring of saliva led to virus inactivation at intermediate (50%) RH, attributed to the antiviral activity of lysozyme.¹⁹ Therefore, multiple components in saliva may have counteracting effects that influence virus infectivity. It is challenging to determine whether the green fluorescence primarily originates from mucin or other salivary proteins. Additional evidence is needed to ascertain the composition of the coffee ring in salivary droplets.

We expect 1-μL droplets to behave differently from levitated and settled aerosols. Studies have suggested that mucin facilitated the formation of inclusions and thus resulted in a heterogenous phase change during evaporation of levitated droplets.^{16, 44} Coffee-ring effects and dendritic patterns were not observed in the aerosols after they dried and settled. Instead, mucin was found in the gaps between salt crystals; it also contributed to more spherical shapes in dried aerosols. Groth et al. found that polycrystalline or monocrystalline structures could be either embedded in mucin or at the air-particle interface in mucin-NaCl aerosols.²⁴ Moreover, studies found that either nanoparticles⁵³ or fluorescently labeled phi6²⁹ were homogeneously distributed in settled, effloresced saliva or proteinaceous aerosols, respectively. The discrepancies between droplets on surfaces and suspended aerosols suggest that their internal flow fields differ, preventing the formation of a coffee ring and dendritic patterns in suspended aerosols. The observations from droplets on surfaces may not directly apply to aerosols. Although the coffee-ring effect is not present in aerosols, it remains unclear whether mucin co-localizes with IAV in aerosols and if such co-localization promotes virus stability. Further research is needed to determine the role of mucin in protecting IAV from decay in aerosols and to elucidate the underlying mechanisms.

Recovery and virus integrity analysis

We analyzed IAV recovery from the polystyrene substrate by quantifying viral RNA within 1-μL droplets of PBS, 0.1% mucin in PBS, or 0.5% mucin in PBS at time zero and after 4 hr at 50% RH (Figure 3), the RH at which we observed the greatest loss of virus infectivity. The total RNA level (red dots) was ~5.5 log₁₀-gene copies/mL for all conditions. There was no difference in RNA level between time zero and 4 hr, suggesting that the difference in infectious IAV (Figure 1B) and RNA is not due to diminished recovery after extended incubation. To assess damage to virion structure, we measured the RNA signal following RNase treatment in the same droplets. After RNase treatment, the amount of RNA recovered at time zero was 1 log₁₀-gene copies/mL lower, suggesting that ~1 log₁₀-gene copies/mL of RNA was outside of the virions, referred to as “free” RNA (Figure 3). After 4 hr, the RNA level in PBS with RNase treatment dropped to the limit of detection. This indicates that virions were no longer intact in PBS, as damage to the capsid and/or envelope allowed the RNase to hydrolyze free RNA that was previously protected inside virions. There was no difference in RNA level between time zero and 4 hr for 0.1% mucin and 0.5% mucin with RNase treatment, suggesting that the IAV structure was still intact after 4 hr of exposure without releasing more free RNA. These results show that mucin at a concentration of at least 0.1% helps maintain virus integrity.

Figure 3.

Log₁₀-gene copies/mL of recovered RNA after IAV was exposed to 50% RH at time zero and after 4 hr in PBS, 0.1% mucin, or 0.5% mucin. Red dots represent the RNA level without any RNase treatment, corresponding to the total RNA. Blue dots represent the RNA level after RNase treatment, corresponding to the RNA within the envelope and capsid. Each dot represents an independent replicate, which is the average of two technical replicates. Error bars represent mean ± standard deviation. *** p < 0.001. The gray dashed line represents the limit of detection (LOD).

We compared the reduction in RNA level to IAV infectivity for each condition. There was at least 1.5 log₁₀ decay of internal RNA and 2.6 log₁₀ decay of virus infectivity in PBS after 4 hr. A smaller reduction in RNA compared to infectivity may be attributed to a poorer detection limit for RNA (see SI). The loss in virion integrity seemed to be one of the major reasons for the loss in infectivity for IAV in PBS droplets, as found in a previous study.⁷ However, we observed no loss of internal RNA in either 0.1% or 0.5% mucin, but ~1 log₁₀ decay of infectivity. Since RNase treatment at time zero reduced the same amount of RNA in both PBS and mucin-containing droplets, mucin likely does not inhibit RNase activity. Therefore, the loss of virus infectivity likely occurred prior to the disruption of the envelope and capsid. We speculate that either the major structural proteins, such as hemagglutinin and neuraminidase, or the lipid membrane layer were damaged even in the presence of mucin. However, it is unclear whether the observed reduction is primarily due to the denaturation of proteins or lipids. Further research is needed to identify which components are most vulnerable and to understand the underlying mechanisms of their susceptibility.

Limitations and implications

This study has several limitations. We conducted our analyses in a simplified mucin-PBS matrix, which can be informative but does not fully recapitulate real respiratory fluids. In addition, the interactions between respiratory proteins and their potential to preserve or inactivate virus remain uncertain. Further exploration with real respiratory fluids is essential to better understand their impact on virion localization. The addition of dye was not expected to damage virus integrity,⁵¹ but the labeling protocol can reduce the titer of virus by 2 log₁₀ due to extensive dialysis. Therefore, we concentrated the virus stock by ultracentrifugation to obtain a sufficient titer of infectious labeled virions, though this concentration may exceed physiologically relevant levels. We acknowledge differences between deposited droplets and suspended aerosols in terms of drying morphology and virus localization, and we should be cautious about extending the findings of this study to aerosols. We did not analyze the impact of pH, which has been shown to affect virus infectivity.^{13, 54} However, pH is highly dependent on air composition, so measurements can vary depending on laboratory conditions.^{13, 55, 56} Lastly, co-localization does not confirm biological interactions between mucin and IAV. It remains to be investigated whether mucin binds to the envelope proteins of IAV and thus offers more protection from other environmental stressors.

To date, our study is the first to localize IAV in mucin-containing droplets at various RHs. Our findings suggest that mucin protects IAV from decay in part by co-localizing with IAV, which reduces the exposure of IAV to high concentrations of solutes. We further validated that mucin helps preserve the integrity of virions compared to pure PBS solutions over 4 hr at 50% RH. This protective effect is independent of the initial concentration of mucin. The results of this study provide valuable information on the protective role of mucin, one of the most abundant proteins in respiratory fluids, and on virus infectivity in droplets under different environmental conditions. These findings contribute to the understanding of virus inactivation mechanisms, transmission of influenza virus through indirect contact with large droplets and fomites, and design of mitigation strategies.

Supplementary Material

Details of plaque assay procedures; composition of PBS; confocal imaging parameters; details of digital droplet PCR assays and procedures; figures of droplet drying kinetics showing weight percentage changes over time; figures of drying kinetics of spread-out, flattened droplets versus compact, taller droplets; confocal images of negative controls; confocal images of saliva droplets; fluorescent images of red nanoparticles in 0.5% mucin-PBS droplets; flow chart of procedures.

Synopsis.

Mucin protects influenza A virus from inactivation in evaporating model respiratory droplets on surfaces, possibly due to co-localization with virus.