Increased aerosol transmission for B.1.1.7 (alpha variant) over lineage A variant of SARS-CoV-2

Airborne transmission, a term combining both large droplet and aerosol transmission, is thought to be the main transmission route of SARS-CoV-2. Here we investigated the relative efficiency of aerosol transmission of two variants of SARS-CoV-2, B.1.1.7 (alpha) and lineage A, in the Syrian hamster. A novel transmission caging setup was designed and validated, which allowed the assessment of transmission efficiency at various distances. At 2 meters distance, only particles <5 μm traversed between cages. In this setup, aerosol transmission was confirmed in 8 out of 8 (N = 4 for each variant) sentinels after 24 hours of exposure as demonstrated by respiratory shedding and seroconversion. Successful transmission occurred even when exposure time was limited to one hour, highlighting the efficiency of this transmission route. Interestingly, the B.1.1.7 variant outcompeted the lineage A variant in an airborne transmission chain after mixed infection of donors. Combined, this data indicates that the infectious dose of B.1.1.7 required for successful transmission may be lower than that of lineage A virus. The experimental proof for true aerosol transmission and the increase in the aerosol transmission potential of B.1.1.7 underscore the continuous need for assessment of novel variants and the development or preemptive transmission mitigation strategies.

relationship between particle size and distance. The design consisted of two rodent cages connected via a polyvinylchloride (PVC) connection tube (76 mm inside diameter) which allowed air ow, but no direct animal contact, from the donor to the sentinel cage. The distance between donor and sentinel cage could be varied (16.5, 106, or 200 cm) by exchanging the PVC connection tube (Supplemental Figure 1 A, B). Directional air ow from the donor to the sentinel cage was generated by negative pressure. The air velocity generated by the air ow through the connection tube averaged at 327, 370, and 420 cm/min for the 16.5, 106, and 200 cm distances, respectively (Supplementary Table 1). This allowed for 30 cage changes per hour.
We next validated the caging design using an aerodynamic particle sizer to analyze the aerodynamic size of particles (dynamic range from <0.5-20 µm) traversing from donor to sentinel cage. Droplets and aerosols were generated in the donor cage (20% (v/v) glycerol solution, sprayed with a standard spray bottle) and the particle size pro le was determined at the beginning and end of the connecting tube to study the potential for size exclusion of the respective cage setups. The reduction of particles was size and distance dependent. At a distance of 16.5 cm between cages, relatively limited size exclusion of the generated particles was observed; ≥6.9% of particles 5-10 µm and ≥42.8% of particles ≥10 µm did not travel into the sentinel cage (Fig 1 A/D). At the intermediate distance of 106 cm between cages an increased reduction of number of particles and size exclusion was observed; ≥70% of particles ≥5 µm did not traverse into the sentinel cage and no particles ≥10 µm were detected. Hence, while in the donor cage 4.86% of detected particles were >5 µm, in comparison the particle pro le in the sentinel cage contained only 2% particles >5 µm (Fig 1 B/E). At the longest distance of 200 cm, we observed an almost complete size exclusion of particles ≥5 µm; ≥95% of particles 5-10 µm did not traverse and no particle ≥10 µm were detected in the sentinel cage. The composition pro le of particles in the sentinel cage comprised only 0.5% particles ≥5 µm (Fig 1 C/F). These combined results demonstrate that we have developed a novel caging system to effectively investigate the impact of distance and particle size exclusion on the transmission of SARS-CoV-2. The overall absence of particles ≥10 µm and extensive reduction of particles 5-10 µm indicate that the caging system with the distance of 200 cm is suitable to study true aerosol transmission; whereas, the 16.5 and 106 cm set-ups are suitable to study airborne transmission occurring via droplet, aerosols or a combination thereof.

SARS-CoV-2 aerosol transmission over 2 meters distance
Experimental SARS-CoV-2 airborne transmission has been demonstrated over short distances in the Syrian hamster model 14,15 . Using the validated caging system, we rst investigated short-distance airborne transmission. For each distance, four donor animals were inoculated intranasal (I.N) with 8x10 4 TCID 50 SARS-CoV-2 lineage A. After 12 hours, the infected animals were placed into the donor (upstream) side of the cages and four sentinels were placed into the downstream cages (2:2 ratio) and were exposed for 72 hours.  Table 2). These data demonstrate the ability of SARS-CoV-2 to transmit over long and short distances and suggest that transmission e ciency may be distance dependent. Therefore, we compared the in-silico binding e ciency of spike receptor binding domain (RBD) of a lineage A and B.1.1.7 variant with human and hamster ACE2. At position 501 of the B.1.1.7 spike RBD, the asparagine residue is substituted by tyrosine, which allows increased interactions with residues on ACE2 through stacking of aromatic sidechains and hydrogen bonding, hence higher a nity binding to human ACE2 19 . A sequence alignment between human and hamster ACE2 revealed variation at the amino acid level, however only two residues differ within the interface with SARS-CoV-2 RBD. At positions 34 and 82, histidine and methionine are replaced by glutamine and asparagine, respectively, in the hamster ACE2 (Figure 3 A, B). These substitutions are not located in the immediate vicinity of the N501Y mutation, suggesting that B.1.1.7 should also exhibit higher a nity binding to hamster ACE2.
To con rm that the observed enhanced binding a nity of B.1.1.7 to human ACE2 was also present for hamster ACE2 we directly compared viral entry using a VSV pseudotype entry assay. No signi cant difference in entry between human and hamster ACE2 with either lineage A or B.  Next, the same experiment was repeated with the exposure time limited to one hour. gRNA was detected 24 hours post exposure in oropharyngeal swabs of three out of four sentinels exposed to B.1.1.7 and two out of four sentinels exposed to lineage A, respectively (Figure 4 G/H). sgRNA was detected in oropharyngeal swabs of two out of four sentinels exposed to B.1.1.7 and two out of four sentinels for lineage A, respectively. At 3 DPE, whilst all sentinels exposed to B.1.1.7 were positive for gRNA and sgRNA, viral RNA was only detected in two of four sentinels exposed to lineage A. Viral loads in swabs did not differ signi cantly between the two variants. To ensure the differences observed in transmission were not due to increased donor shedding, we compared viral loads in oropharyngeal swabs taken  (Figure 5 A). Immediately after, the eight sentinels were cohoused with eight new sentinels (Sentinels 2) (2:2 ratio) for 24 hours and donor animals were relocated to normal rodent caging. This sequence was repeated for Sentinels 3. For each round, the previous sentinels were housed in the upstream cage and became the new donors. We assessed transmission by measuring viral RNA in oropharyngeal swabs taken from all animals at 2 DPI/DPE. While all donor animals (median gRNA = 7.3 copies/mL (log 10 ), median sgRNA = 7.0 copies/mL (log 10 )) and all Sentinels 1 (median gRNA = 7.0 copies/mL (log 10 ), median sgRNA = 6.8 copies/mL (log 10 )) demonstrated robust shedding, viral RNA could only be detected in four out of eight Sentinels 2 (median gRNA = 2.5 copies/mL (log 10 ), median sgRNA = 1.8 copies/mL (log 10 )), and in one Sentinels 3 animal ( Figure 5 B/C). We compared infectious virus titers in the swabs. While all donor animals (median = 4.25 TCID 50 (log 10 ) and all Sentinels 1 had high infectious virus titers (median = 4.5 TCID 50 (log 10 )), infectious virus could only be detected in four Sentinels 2 (median = 0.9 TCID 50 (log 10 )), and no Sentinels 3 animals (Figure 5 D). We then proceeded to compare the viral loads in the lungs of these animals at 5 DPE. As expected, viral RNA was only detected in animals that were positive for SARS-CoV-2 in their corresponding oropharyngeal swab. While all donor animals (median sgRNA = 10.0 copies/g (log 10 )) and all Sentinels 1 had high gRNA levels in the lung (median sgRNA = 10.0 copies/mL (log 10 ), viral RNA could only be detected in four Sentinels 2 (median sgRNA = 4.3 copies/mL (log 10 ), and no RNA was detected in any Sentinels 3 ( Figure 5 E). We compared the gross pathology of these lungs. Lungs from Sentinels 1 demonstrated SARS-CoV-2 infection associated pathology as previously described 14,20,21 . Pathology was only seen in three Sentinels transmission studies in hamsters and ferrets have been performed, none of these studies were able to differentiate between large and small droplet transmission 14,15,[26][27][28] . In our previous work, speci cally designed cage dividers were used to generate an air ow system minimizing large droplet cross-over. While the number of large droplets was markedly reduced, aerosol transmission could not be conclusively demonstrated 14 . Within the currently described transmission caging only 2% and 0.5% of particles found in the sentinel side were ≥5 µm at 106 and 200 cm distance, respectively, strongly suggesting that the transmission observed in these cages is by true aerosols. This is an important nding in two regards. First, epidemiological conclusive evidence for aerosol transmission of SARS-CoV-2 is currently still lacking, because it is di cult to determine with certainty the route or combination of routes of transmission.
Second, particles <5 µm are expected to reach the respiratory bronchioles and alveoli. While respirable aerosol (<2.5 µm), thoracic aerosol (<10 µm) and inhalable aerosol in general 29 all may be relevant to infection with SARS-CoV-2 30 , it has been suggested that direct deposition into the lower respiratory tract may decrease the necessary infectious dose 31 . Indeed, our previous work has demonstrated that aerosol inoculation in the Syrian hamster is highly e cient (25 TCID 50 , particles <5µm 32 ) and is linked to increased disease severity due to direct deposition of the virus into the lower respiratory tract 14 .
The data presented here need to be considered in the context of inherent differences between the Syrian hamster model and human behavior. Experimentally, animals were exposed to a unidirectional air ow at timepoints chosen for optimal donor shedding, which likely contributed to the high e ciency of aerosol transmission even at 200 cm distance after only one hour of exposure. However, this approximates human exposure settings such as restaurants or o ce spaces.
Increased risk of airborne transmission is an important concern in the context of VOCs. VOC B.1.1.7 was rst detected in the United Kingdom and has been shown to exhibit increased transmission with signi cantly increased reproduction number and attack rates 33,34 . Transmission e ciency is a function of donor shedding, exposure time, sentinel susceptibility, and potential environmental factors effecting stability during transmission. One experimental study in preprint has found no difference in contact, fomite or short distance airborne transmission between D614G and B.1.1.7 in the Syrian hamster, however, numbers were low and no aerosol transmission was compared 35
Receptor transfection BHK cells were seeded in black 96-well plates and transfected the next day with 100 ng plasmid DNA encoding human or hamster ACE2, using polyethylenimine (Polysciences). All downstream experiments were performed 24 h posttransfection.
Pseudotype production and Luciferase-based cell entry assay Pseudotype production was carried as described previously 48 . Brie y, plates pre-coated with poly-L-lysine (Sigma-Aldrich) were seed with 293T cells and transfected the following day with 1,200 ng of empty plasmid and 400 ng of plasmid encoding coronavirus spike or no-spike plasmid control (green uorescent protein (GFP)). After 24 h, transfected cells were infected with VSVΔG seed particles pseudotyped with VSV-G, as previously described 48,49 . After one hour of incubating with intermittent shaking at 37 °C, cells were washed four times and incubated in 2 mL DMEM supplemented with 2% FBS, penicillin/streptomycin and L-glutamine for 48 h. Supernatants were collected, centrifuged at 500xg for 5 min, aliquoted and stored at −80 °C. BHK cells previously transfected with ACE2 plasmid of interest were inoculated with equivalent volumes of pseudotype stocks. Plates were then centrifuged at 1200xg at 4 °C for one hour and incubated overnight at 37 °C. Approximately 18-20 h post-infection, Bright-Glo luciferase reagent (Promega) was added to each well, 1:1, and luciferase was measured. Relative entry was calculated normalizing the relative light unit for spike pseudotypes to the plate relative light unit average for the no-spike control. Each gure shows the data for two technical replicates.

Structural interaction analysis
Structure modeling was performed using the human ACE2 and SARS-CoV-2 RBD crystal structure, PDB ID 6M0J 50 .
Mutagenesis to model the residues that differ in the B.1.1.7 RBD and hamster ACE2 was performed in COOT 51  The remaining 5 animals for each route were euthanized at 14 DPI for disease course assessment and shedding analysis. Hamsters were weighted daily, and oropharyngeal swabs were taken on day 1, 2, 3 and 5. Swabs were collected in 1 mL DMEM with 200 U/mL penicillin and 200 µg/mL streptomycin. Hamsters were observed daily for clinical signs of disease. Necropsies and tissue sampling were performed according to IBC-approved protocols.

Aerosol cages
The aerosol transmission system consisted of two 7" X 11" X 9" plastic hamster boxes (Lab Products, Inc.) connected with a 3" diameter tube (Supplementary Figure 1). The boxes were modi ed to accept a 3" plastic sanitary tting To ensure the system was able to contain aerosols the airtightness of the system was validated with a negative pressure smoke test and a positive pressure leak test prior to moving into a containment laboratory. To perform the negative pressure test the air ow was adjusted to exhaust the system at 30 cage changes/hour, smoke was generated in the donor cage with a WizardStick and escaped particulate was measured with a TSI DustTrak DRX. To test the system under pressure the air ow was reversed, and the joints were tested using a gas leak detector.

Particle sizing
Transmission cages were modi ed by introducing an inlet on the side wall of the infected hamster side, and sample ports on each end of the connection tube for measurement of particles in the air under constant air ow condition.
Particles were generated by spraying a 20% (v/v) glycerol solution with a standard spray bottle through the donor cage inlet. The particle size was measured using a Model 3321 aerodynamic particle sizer spectrometer (TSI). First, the donor cage was coated with three sprays at an interval of 30 seconds (s). The sample port was opened, and a sample was analyzed. Every 30 s a new spray followed, and ve samples were analyzed (5 runs, each 60 s) for both donor side (primary infected side) and sentinel side.

Aerosol Transmission experiments
All transmission studies were conducted at a 2:2 ratio between donor and sentinels for each transmission scenario tested and virus variant with 2 separate transmission cages (N = 4 donors / 4 sentinels). To ensure no crosscontamination, the donor cages and the sentinel cages were never opened at the same time, sentinel hamsters were not exposed to the same handling equipment as donors and after each sentinel the equipment was disinfected with either 70% ETOH or 5% Microchem.
Initially, transmission was studied assessing distance. Donor hamsters were infected intranasally as described above with 8x10 4 TCID 50 SARS-CoV-2 (lineage A or B.1.1.7 variants). After 12 hours donor animals were placed into the donor cage and sentinels were placed into the sentinel cage (2:2). Air ow was generated between the cages from the donor to the sentinel cage at 30 changes/h. Hamsters were co-housed at 16.5 cm, 106 cm or 200 cm distance. Regular bedding was replaced by alpha-dri bedding to avoid the generation of dust particles. Oropharyngeal swabs were taken for donors at 1 DPI and for sentinels daily after exposure began. Swabs were collected in 1 mL DMEM with 200 U/mL penicillin and 200 µg/mL streptomycin. Exposure continued until respiratory shedding was con rmed in sentinels on three consecutive days. Then donors were euthanized, and sentinels were monitored until 14 DPE (days post exposure) for seroconversion.
Second, transmission was studied assessing duration of exposure. Donor hamsters were infected intranasally as described above with 8x10 4 TCID 50 SARS-CoV-2. After 24 hours (1 DPI) or 72 hours (3 DPI) donor animals were placed into the donor cage and sentinels were placed into the sentinel cage (2:2). Hamsters were co-housed at 200 cm distance for 1 or 4 hours at an air ow rate of 30 changes/h. Oropharyngeal swabs were taken for donors at day of exposure and for sentinels for three days after exposure.

Variant competitiveness transmission chain
Donor hamsters (N = 8) were infected intranasally as described above with 1x10 2 TCID 50 SARS-CoV-2 at a 1:1 ratio of lineage A and B.1.1.7 mixture. After 12 hours donor animals were placed into the donor cage and sentinels (Sentinels 1, N = 8) were placed into the sentinel cage (2:2) at 16.5 cm distance at an air ow of 30 changes/h. Hamsters were cohoused for 24 h. The following day, donor animals were re-housed into regular rodent caging and Sentinels 1 were placed into the donor cage of new transmission set-ups. New sentinels (Sentinels 2, N = 8) were placed into the sentinel cage (2:2) at 16.5 cm distance at an air ow of 30 changes/h. Hamsters were co-housed for 24 h. Then, Sentinels 1 were re-housed into regular rodent caging and Sentinels 2 were placed into the donor cage of new transmission setups. New sentinels (Sentinels 3, N = 8) were placed into the sentinel cage (2:2) at 16.5 cm distance at an air ow of 30 changes/h. Hamsters were co-housed for 24 h. Then, both Sentinels 2 and Sentinels 3 were re-housed to regular rodent caging and monitored until 5 DPE. Oropharyngeal swabs were taken for all animals at 2 DPI/DPE. All animals were euthanized at 5 DPI/DPE for collection of lung tissue.

Viral RNA detection
Swabs from hamsters were collected as described above. Then, 140 µL was utilized for RNA extraction using the QIAamp Viral RNA Kit (Qiagen) using QIAcube HT automated system (Qiagen) according to the manufacturer's instructions with an elution volume of 150 µL. For tissues, RNA was isolated using the RNeasy Mini kit (Qiagen) according to the manufacturer's instructions and eluted in 60 µL. Sub-genomic (sg) viral RNA and genomic (g) was detected by qRT-PCR 52,53 . RNA was tested with TaqMan™ Fast Virus One-Step Master Mix (Applied Biosystems) using QuantStudio 6 or 3 Flex Real-Time PCR System (Applied Biosystems). SARS-CoV-2 standards with known copy numbers were used to construct a standard curve and calculate copy numbers/mL or copy numbers/g.

Viral titration
Viable virus in tissue samples was determined as previously described 54 . In brief, lung tissue samples were weighted, then homogenized in 1 mL of DMEM (2% FBS). VeroE6 cells were inoculated with ten-fold serial dilutions of homogenate, incubated 1 hours at 37°C and the rst two dilutions washed twice with 2% DMEM. After 6 days cells were scored for cytopathic effect. TCID 50 /mL was calculated by the method of Spearman-Karber.

Serology
Serum samples were analyzed as previously described 55 . In brief, maxisorp plates (Nunc) were coated with 50 ng spike protein (generated in-house) per well. Plates were incubated overnight at 4°C. Plates were blocked with casein in phosphate buffered saline (PBS) (ThermoFisher) for 1 hours at room temperature (RT). Serum was diluted 2-fold in blocking buffer and samples (duplicate) were incubated for 1 hours at RT. Secondary goat anti-hamster IgG Fc (horseradish peroxidase (HRP)-conjugated, Abcam) spike-speci c antibodies were used for detection and visualized with KPL TMB 2-component peroxidase substrate kit (SeraCare, 5120-0047). The reaction was stopped with KPL stop solution (Seracare) and plates were read at 450 nm. The threshold for positivity was calculated as the average plus 3 x the standard deviation of negative control hamster sera.

Next-generation sequencing of virus
For sequencing from swabs, total RNA was depleted of ribosomal RNA using the Ribo-Zero Gold rRNA Removal kit

Statistical Analysis
Signi cance test were performed as indicated where appropriate. Statistical signi cance levels were determined as follows: ns = p > 0.05; * = p ≤ 0.05; ** = p ≤ 0.01; *** = p ≤ 0.001; **** = p ≤ 0.0001.      1 ratio), and three groups of sentinels (Sentinels 1, 2 and 3) were exposed subsequently at 16.5 cm distance. A. Schematic visualization of the transmission chain design. Animals were exposed at a 2:2 ratio, exposure occurred on consecutive days and lasted for 24 hours for each chain link. B/C. Respiratory shedding measured by viral load in oropharyngeal swabs; measured by gRNA and