The role of bioaerosols in the dispersal of antimicrobial resistance genes (ARGs) and resistant microorganisms is poorly understood. In addition, bioaerosols are powerful composite samples representative of the surrounding environment and can be used as sentinels of many local habitats. Evidence suggests that using environmental DNA from dust collected on vehicle cabin air filters can define regional resistance profiles. Here, this method was used to investigate differences in resistance gene profiles, their underlying bacterial communities, and their links to anthropogenic and environmental variables across Canada. In total, 477 car filter samples were collected, with every province and territory being represented. DNA was extracted from filter dust. High‐throughput qPCR was used to detect and quantify a panel of 36 ARGs and 3 mobile genetic elements. Bacterial biomass was assessed using standard qPCR methods of the 16S rRNA gene, which was also used to assess bacterial biodiversity via metabarcoding. Results indicated that qepA dominates antimicrobial resistance profiles across Canada. However, after they were removed from the dataset, regional profiles were evident based on gene type and richness. Factors positively linked to total numbers of ARGs included human and livestock populations; whereas mean annual precipitation was negatively linked to resistance gene quantities. Measures of α‐diversity were generally greater in the western regions of Canada than in the east and the north. Community composition analyses showed similarities between the prairies and territories, which were separated from other regions. Finally, network analyses revealed a relatively stable group of core ARGs across regions, which were largely correlated with low‐abundance genera. Such findings suggest that rare taxa are key links in the diffusion of antimicrobial resistance in environmental contexts. Furthermore, this study highlights the potential application of vehicle air filters in building long‐term monitoring capacity of outdoor bioaerosols.

In total, 477 vehicle cabin air filters were collected from 51 locations across Canada, through a network of participating mechanics, individuals, and municipal governments. Two additional filters were collected as a methodological control. We asked that filters be collected during routine vehicle maintenance and placed in sterile resealable plastic bags for transport to the Institut universitaire de cardiologie et de pneumologie de Québec – Université Laval in Quebec City, QC, Canada for processing. We also asked participants to include the odometer reading (km) since last change and the forward sorting area of their postal code for environmental and population metadata collection. This data is not presented to preserve anonymity. Filters were categorised into the 6 geographical regions of Canada as defined by Statistics Canada: British Columbia, Prairies, Ontario, Quebec, Atlantic, Territories. Filters were collected between summer 2020 and winter 2021. During this time, travel restrictions within some Canadian regions were in place and long-distance travel was discouraged, which helped control for effects of interregional travel.

For each filter, metadata on the sampling location was collected from publicly available sources. Latitude, longitude (not presented to preserve anonymity) and elevation were recorded for each filter. Distance to the nearest major water body (i.e., ocean or Great Lake), was estimated using direct measurements from Google maps. Population and population density came from 2021 census data (Statistics Canada 2022b). Both mean annual temperature (MAT) and mean annual precipitation (MAP) were taken from the most recent records published by Environment and Climate Change Canada (2024). Estimates of local livestock populations for each location were taken from the 2021 census of agriculture (Statistics Canada 2022a) at the level of agricultural subdivision. Data on local hospital capacity based on number of beds in 2020-2021 was taken from Ministère de la Santé et des Services sociaux for Québec (2022) and from the Canadian Institute for Health Information (2022) for all other jurisdictions. These can be found in Table S2.

DNA Extraction

A section was cut out of each filter, including the two control filters, and weighed for DNA extraction. This section of material was placed in one half of a strainer bag and homogenised for 1 min in a 100 ml saline solution of 0.05 tween20 with a paddle mixer (Fisher Scientific). This was done to dislodge fine dust from the filter. Next, a 50 ml aliquot of this solution was collected from the opposite side of the filter bag and differentially centrifuged for 3 min at 250 rpm to remove large particles and plant matter. Then 3 ml aliquots of supernatant were spun at 10,000 rpm to form pellets for DNA extraction. All DNA extractions were made with DNEasy PowerLyzer PowerSoil Kit (Qiagen, Toronto, ON, CA) following manufacturer’s instructions.

qPCR Analyses

Total bacterial biomass per g of filter material was calculated by qPCR of the 16S rRNA gene following Bach et al (2002). Primer sequences are presented in Table S3. These analyses were conducted on a Bio-Rad CFX-384 Touch^TM Real-Time PCR Detection System (Bio-Rad, Montreal, QC, CA) with the following thermoprotocol: 95 °C for 3 min then 40 cycles of 95 °C for 20 s and 62 °C for 1 min. Data were used when efficiency curves were between 90–110%. As part of quantification, DNA extraction blanks were included with all 16S rRNA gene qPCR analyses. If reactive, the starting quantity (Sq) values of corresponding blanks were subtracted from those true samples. An upper limit of 36 cycles was used. There were no reactions in no template qPCR negative controls.

Total 16S rRNA gene copy numbers per gram of filter were estimated using the formula:

16S rRNA copy number per g filter = ((Sq*50)*(50/3.33))/W

In which, Sq is the starting quantity estimate from qPCR reactions using 2 ml of DNA taken from a 100 ml reaction. This extraction was performed on a 3.33 ml aliquot of the 50 ml extraction solution. W is the weight of the filter in mg. Copy numbers were log₁₀-transformed for statistical analyses.

Next, the ARG profile of each filter was assessed with SmartChip HT-qPCR (TakaraBio USA, San Jose, CA, USA) using the same aliquot. In this case, extracts were screen against a predefined panel of 36 ARGs and 3 MGEs as outlined in George et al (2022). Table S3 outlines the primer sequences and target genes of the 3 aminoglycoside, 10 beta-lactam, 1 colistin, 4 macrolide, 3 MGEs, 2 quinolone, 2 sulfonamide, 10 tetracycline, and 4 vancomycin genes used. The mode of action of these genes was taken from the Comprehensive Antibiotic Resistance Database (https://card.mcmaster.ca). All genes used SYBR Green dye fluorescence except for blaCTX-M-1 and mcr1, which used FAM probes (Table S3). The SmartChip thermoprotocol was: 95 °C for 3 min; then 45 cycles of 95 °C, 10 s; 60 °C, 30 s; 55 °C, 31 s; followed by a melting curve of 55 °C for 5 s + 0.5 °C/cycle. On each chip a no template control and a positive control containing a mix of all targets at concentrations of ~10⁵ copies ml ^-1 was included. To eliminate false positives, melt curves were verified against those of the positive control. Relative copy numbers of target genes were calculated with the comparative CT method (Schmittgen and Livak 2008):

𝑅𝑒𝑙𝑎𝑡𝑖𝑣𝑒 𝑔𝑒𝑛𝑒 𝑐𝑜𝑝𝑦 𝑛𝑢𝑚𝑏𝑒𝑟 = 2 ^{―(𝐶𝑡(𝐴𝑅𝐺) ― 𝐶𝑡(16𝑆))}

Relative abundance of ARGs/MGEs was normalised by multiplying the copy number by 16S rRNA gene copy number.

REFERENCES

Bach, H.‐J., J. Tomanova, M. Schloter, and J. C. Munch. 2002. “Enumeration of Total Bacteria and Bacteria With Genes for Proteolytic Activity in Pure Cultures and in Environmental Samples by Quantitative PCR Mediated Amplification.” Journal of Microbiological Methods 49: 235–245

Canadian Institute for Health Information. 2022. Trends in Hospital Spending, 2009–2010 to 2020–2021—Data Tables—Series D: Beds Staffed and in Operation by Functional Centre. Ottawa: Canadian Institute for Health Information.

George, P.B.L., F. Rossi, M. W. St‐Germain, et al. 2022. “Antimicrobial Resistance in the Environment: Towards Elucidating the Roles of Bioaerosols in Transmission and Detection of Antibacterial Resistance Genes.” Antibiotics 11: 974.