Particulate matter (PM2.5) as a potential SARS-CoV-2 carrier

The rapid spread of the SARS-CoV-2 in the COVID-19 pandemic had raised questions on the route of transmission of this disease. Initial understanding was that transmission originated from respiratory droplets from an infected host to a susceptible host. However, indirect contact transmission of viable virus by fomites and through aerosols has also been suggested. Herein, we report the involvement of fine indoor air particulates with a diameter of ≤ 2.5 µm (PM2.5) as the virus’s transport agent. PM2.5 was collected over four weeks during 48-h measurement intervals in four separate hospital wards containing different infected clusters in a teaching hospital in Kuala Lumpur, Malaysia. Our results indicated the highest SARS-CoV-2 RNA on PM2.5 in the ward with number of occupants. We suggest a link between the virus-laden PM2.5 and the ward’s design. Patients’ symptoms and numbers influence the number of airborne SARS-CoV-2 RNA with PM2.5 in an enclosed environment.


Methods
Sampling location and indoor air sampling. The layout and dimension of the wards are shown in  Table 2 (in the main text) and Fig. 1. Each ward was occupied by one to eighteen COVID-19 patients. As a caveat, during the measurement in this study, hospital's management staff has deployed three units of air purifiers at ward B, C, and D. During the air sampling measurement, air purifier (FANFIL AP510M, Aire-plus Technology, Singapore) was deployed at ~ 1 m distance in wards C and D, ~ 8 m in ward B, and no air purifier in single occupant room. PM 2.5 was sampled in a single-bed ward (31st March to 4th April 2020) and multiple bed wards (4th-29th April 2020) in a teaching hospital at Kuala Lumpur, respectively. Air sampling was conducted for 48 h during a 29 days sampling period using two types of instruments; an air quality sensor known as AiRBOXSense (AIR-BOXSENSE V3.0, UKM Tech. Sdn Bhd, Malaysia) 12,26 and a low volume sampler (LVS) (MINIVOL, AirMetrics, USA). Details of AiRBOXSense are described in 26 . Both instruments were operated side by side in wards occupied by SARS-CoV-2 positive patients. Instruments were treated using ultraviolet light for 20 min (UV) (UV-C 253.7 nm), further disinfected with 70% alcohol and calibrated before being translocated to the next wards. The same instruments were used to avoid variability during sampling.
AiRBOXSense was used to continuously measure PM 2.5 , while the LVS was used to determine the virus loading in PM 2.5 trapped on filter paper (WHATMAN glass microfiber filters, Grade GF/F) with a tight specification of 0.6-0.8 μm particle retention and pure borosilicate glass structure, GF/F. A 5 L min −1 of air was drawn into the AiRBOXSense. While, the Minivol's pump draws air at 5 L min −1 through a filter paper. The continuous concentration of PM 2.5 was recorded and stored in secure digital card (SD card) in the AiRBOXSense. The data synchronously retrieved via THINGSPEAK (The MathWorks Inc, USA) cloud storage and analysed using MAT-LAB software (The MathWorks Inc, USA). www.nature.com/scientificreports/ Each filter paper was collected after 48 h of sampling and stored in a sealed container and kept in − 80 °C laboratory freezer. The filter papers were extracted for viral load analysis using reverse transcription quantitative real time polymerase chain reaction (RT-qPCR) approach.
Calibration of AiRBOXsense. AiRBOXSense was calibrated according to 26 1 day before each sampling.
Calibration consists of setting a mathematical model describing the relationship between sensor data and reference instruments. The AiRBOXSense unit was calibrated in tandem with the GRIMM (as reference instrument) dust monitor model 1.108 (GRIMM Aerosol, Technik GmbH & Co. KG, Germany). The sensors measuring mass concentration were calibrated using GRIMM Aerosol, which was deployed at a clean area (laboratory) for lower concentration measurement and near to a car exhaust for high concentration measurement. The calibration equations are set by fitting a model during a calibration time interval when AiRBOXSense and GRIMM are co-located.

Viral nucleic acid extraction.
Prior to viral nucleic acid extraction, the membrane filter was processed according to 35 with slight modifications. The membrane was first divided into four parts and immersed in 1 mL sterile RNase-free water in separate tubes. Each part of the membrane was vortexed for 2 min in 30 s-intervals to release viral particles attached to the membrane. The tubes were then centrifuged at 500 rpm for 1 min to remove debris, and the supernatants were transferred into new microcentrifuge tubes for viral nucleic acid extraction. This process was repeated twice to ensure all virus particles were resuspended into the water. Subsequently, viral nucleic acid extraction was performed using a Viral Nucleic Acid Extraction Kit II (Geneaid Biotech Ltd., Taiwan) according to the manufacturer's protocol. The purified nucleic acid containing the samples was then kept at − 80 °C for further analysis.
Reverse-transcription quantitative real-time polymerase chain reaction (RT-qPCR) analysis. The primers and probes used in the detection of SARS-CoV-2 were 2019-nCoV_N1, and 2019-nCoV_ N2 combined primer/probe mixes purchased from Integrated DNA Technology (IDT). The information on primers and probes were included in Table 1. Human RNase P primer was not included as a control in this analysis because this study was not conducted using specimen from human. RT-qPCR was carried out using a THUNDERBIRD One-step RT-qPCR kit (Toyobo Co., Ltd., Japan) according to the manufacturer's protocol. The annealing temperature of the primers was set at 55 °C, as suggested by Centres for Disease Control and Prevention or CDC (2020) 28 . Detection of SARS-CoV-2 using the RT-qPCR approach with a BIORAD iQ5 Real-Time PCR machine (BIORAD, USA) as described by CDC (2020) with slight modifications. A standard curve was also generated using 2019-nCoV Positive Control (nCoVPC) with a series of tenfold dilutions from 2 × 10 5 to 2 copies/µL of the control template. The amplification efficiency and R 2 value were recorded, and the standard curve was used to estimate the viral RNA of SARS-CoV-2 on the membrane.

Results and discussion
Indoor PM 2.5 . All 48 h average of PM 2.5 concentration measurements and samplings were taken in COVID-19 wards as illustrated in Fig. 2 and Table 2. The highest concentration of indoor PM 2.5 was measured in general ward B (23.27 µg m −3 ) on the 4th April, while the lowest 48 h average concentration was measured in general ward D (6.23 µg m −3 ) on the 22nd April as shown in Fig. 2. The General ward B was occupied by a cluster of patients from the same institution and was observed to have the most activity among the patients. Higher PM 2.5 concentrations can be contributed by physical activities such as movements of health workers and patients 21,27,29 . The PM 2.5 concentrations measured in this study are slightly lower than reported in a European urban hospital 30 .
Virus RNA analysis. SARS-CoV-2 RNA was isolated from filter membranes collected from the LVS. Only the N1 nucleocapsid gene was successfully detected in RT-qPCR in certain wards. According to the Emergency Use Authorization (EUA), detection of either the N1 or N2 gene is considered positive for the presence of SARS-CoV-2 30 . We detected positive results for SARS-CoV-2 genes in the single room Ward A (74 ± 117.1 copies μL −1 ) and General Ward B (10 ± 7.44 copies μL −1 ). The viral genomes extracted from the filter paper were of heterogenous mixture. This contributes to the high standard deviation in the virus copy number as heterogeneous nucleic acid template was used in RT-qPCR and the presence of SARS-CoV-2 genome was relatively low. Nonetheless, the cycle threshold (CT) value was < 40 30 , confirming the positive detection of SARS-CoV-2 in our samples www.nature.com/scientificreports/ (Table 2). Due to operational restriction imposed by the hospital, the sample size was limited and replication was not possible. The uniqueness in the result is that viral RNA was still able to be detected in the single occupancy ward (Ward A). Ward A is a small enclosed room (22 m 2 ) with a lavatory attached. The frequent use of the lavatory by the symptomatic patient is likely to result in the increase of viral shedding activity in the room. We suspect that virus-laden PM 2.5 generated from the shedding activity circulated within the enclosed room despite low PM 2.5 concentration (11.25 µg m −3 ), thus explaining the spike in the data. The degree of viral shedding (from the patients) due to symptoms such as coughing, sneezing, diarrhoea, etc. has been reported to influence the number of virus particles in the environment 1,5 . It is suggested that the increased virus particles (due to shedding) in a poorly ventilated environment might increase the virus-PM 2.5 assemblage 9,19,31 . A study done by 5 reported that they were not able to detect SARS-CoV-2 in all of their tested air samples. However, they highlighted that their short sampling time of 15 min-4 h might not represent total air volume in the ward and the presence of SARS-CoV-2 might have possibly been diluted during air exchanges in the ward. In contrast, viral RNA was able to be detected in this study when air sampling duration was extended.
SARS-CoV-2 RNA was also detected in General Ward B. General Ward B is a larger room (~ 100 m 2 ) consisting of 18 occupied beds with two air purifying units installed at a distance of farther away from the LVS. The  Table 2. Summary of the data collected from a teaching hospital at Kuala Lumpur. a Selected wards that were sampled consisting of different patient clusters: Single room ward A, an executive ward that hosts only one COVID-19 patient; General ward B was occupied by an institutional cluster; General Ward C was occupied by patients arriving from overseas; and General Ward D was occupied by migrant workers. b Average 48 hourly concentrations (with standard deviation) of PM 2.5 measured in different wards. c Detection of SARS-CoV-2 RNA on captured PM 2.5 at different wards. www.nature.com/scientificreports/ amount of SARS-CoV-2 collected in the particulate matter is significantly lower than from Ward A despite the higher number of patients and concentration of PM 2.5 (17.58 µg m −3 ). Such a low viral load in the PM 2.5 could be attributed to the minimal viral shredding despite the high particulate matter. These particulate matters suspended in the air could be derived from floor and surfaces 32,33 as a result of the high occupants' activities in ward B. Virus-laden PM 2.5 was not detected in Wards C and D despite having similar ward size. The number of patients in Ward C is similar to Ward B, whereas the number of patients in Ward D is half of that of Wards C. The patients in Ward C and Ward D were also diagnosed with mild symptoms. The non-detection of the virus in these wards may be due to very low virus shedding from the patients. Another possible factor to explain the absence of SARS-CoV-2 RNA in PM 2.5 is that the LVS in Ward C (and also Ward D) was positioned adjacent to an air purifier. Although air-purifier's effectiveness in removing PM 2.5 remains unclear, air-filtration has been reported to reduce viral loading in air 9,32,33 .
Our results clearly indicated that SARS-CoV-2 RNA is present within sampling of the Ambient's particles. Hence, it is crucial to determine whether these RNAs came from intact virus particles or are merely RNA from non-infectious virus particles. The detection of SARS-CoV-2 viral RNA on surfaces was previously reported on a cruise ship, the Diamond Prince, even after 17 days after the evacuation of passengers 34 . In addition, the CDC pointed out that the infectivity of the detected particles was still uncertain. A study carried out in a CDC facility showed that SARS-CoV-2 could remain infectious up to 72 h on various types of surfaces 24 . Thus, it is suggested that infectious virus be determined by culturing of virus residing on the PM 2.5 onto appropriate cell culture. However, our study could not show a direct link between the concentration of PM 2.5 and SARS-CoV-2. We did find that PM 2.5 generated from human activities in healthcare facilities can influence the presence of SARS-CoV-2 RNA in indoor environments. Furthermore, the degree of viral shedding from symptomatic patients may also influence the presence of SARS-CoV-2 RNA on PM 2.5 . Therefore, we recommend that all possible precautions against airborne transmission in indoor environments should be taken seriously. www.nature.com/scientificreports/