This study investigated the bacterial abundance and diversity in size-resolved fractions of indoor bioaerosols in urban and semi-urban residences, in parallel with the outdoor bioaerosols, in the UK, throughout the year. Particles that were most loaded with bacteria were found in the sizes > 1 µm, and were dispersed across multiple sizes ranges, with maximum levels for the ranges 1–2 µm and 2–4 µm indoors and > 8 µm outdoors. Our results are in agreement with previous findings showing that for bacteria-laden particles the size distribution in the air is mostly not represented by the size ranges of their cells/spores and bacteria tend to attach on big particles (Clauβ, 2015). Previous qPCR-based investigations have found that the highest concentrations occurred in the coarse fraction, based on sampling using a Harvard high-volume cascade impactor in a suburban house, indoors and outdoors (Sippula et al., 2013), in the 3.3–4.7 µm and 4.7-7 µm ranges, based on samples collected with a six-stage non-viable Andersen in classrooms (Hospodsky et al. 2015), and in the 3.3–4.7 µm range, based on collection with an eight-stage non-viable impactor in a university classroom (Qian et al. 2012). The occurrence of bacteria aggregated or adhered to large-sized particles in outdoor air has been reported to provide a protection to the micro-organisms from harsh environmental conditions (Lighthart & Shaffer, 1997). Determination of large bioaerosols is important as particles may be fragmented upon deposition and get re-aerosolised as smaller aerosols, capable of entering deeper into the human respiratory system. It must be noted, however, that particle bounce in the impactor may potentially bias the reported size distributions recovered.
The higher indoor air bacterial concentrations compared with outdoors (i.e. I/O ratio greater than 1) for most size ranges during winter and spring suggests that the origin of those particles was most likely associated with in-house sources, whereas the shift of the I/O ratio towards values closer or lower to unity observed in summer indicates that the impact of indoor sources was less pronounced. Owing to the higher air exchange rates typically occurring during summer in naturally ventilated spaces, the indoor concentrations tend to track the outdoor levels more closely. Overall, I/O ratios < 1, denoting a potential outdoor origin, were observed for the largest particles, whilst bacterial aerosols that exhibited the highest ratios were within the fine fraction for both residences. In addition, the shift of the indoor concentration peak from 1–2 µm in winter to 2–4 µm in summer, observed for both houses, suggests a potential mixing with the outdoor air large-sized particles.
In terms of bacterial composition, the particle size fraction played a weak but significant role in structuring the microbiome of fine (< 2 µm) and larger (> 2 µm) aerosols, with bacterial taxa dispersed across all examined sizes. The environment and site (location) combined appeared to be the major drivers of community structure, with the variation occurring between indoors and outdoors more pronounced than the variation between the two residential settings, indicating that the key determinants shaping the indoor and outdoor air microbiomes are different. In addition, the compositional differences between indoor and outdoor aerosols were mostly influenced by the larger particles. Despite the significant differences in bacterial diversity (how many types of bacteria are present) found in indoor air between the two residences, outdoor aerosols exhibited similar levels of bacterial richness. Although the two study sites were characterised by different levels of urbanisation, the aerosol bacterial communities outdoors were found to vary less compared to indoors, most likely related to the fact that the spatial scale was not large enough to provide a location-specific pattern.
Considering that the two flats chosen to be sampled in the current study were as similar as possible in terms of various parameters known to structure the indoor microbiome including features such as ventilation type, occupancy patterns, as well as the external environment in the immediate vicinity of the housing buildings, suggests that the factors contributing to the variation in aerosol microbiomes between the two residences, which was more pronounced for the fine particles, were driven by in-house sources. Indeed, airborne bacteria were found to be predominantly sourced from residential surfaces in both flats. Household surfaces harbour diverse microbial communities which could be a result of either transfer through contact or deposition of microorganisms that were previously airborne (Barberán et al., 2015; Meadow et al., 2014b). Settled microbes can re-enter the air due to resuspension of dust particles induced by indoor activities and thus surfaces (e.g., furnishing and flooring) can act as secondary sources (Prussin & Marr, 2015; Ferro et al., 2004). Previous studies have demonstrated that humans leave a distinct microbial signal in the environments they occupy on both surfaces and in the surrounding indoor air (Meadow et al., 2015; Fierer et al., 2010). Although various common human-associated taxa (e.g., Staphylococcus, Micrococcus, Clostridia) were found to dominate aerosols in both houses, several of those exhibited substantial different relative proportions. For instance, Acinetobacter was more abundant in the air of the urban flat and was also found to be a lot more enriched on the urban residence surfaces, whereas Corynebacterium and Streptococcus were detected in noticeably higher abundance in both the air and the surfaces of the semi-urban flat. Paracoccus, known to typically inhabit a wide variety of outdoor environments, was also among the predominant airborne bacteria found to be more enriched indoors for both sites. The particular taxon was detected on furniture surfaces (bookcase, dresser) and the floor in both houses, too, in agreement with previous dust-associated microbiome investigations (Kim et al., 2022; Thompson et al., 2021). Moreover, Pseudomonas, which is ubiquitous in natural settings, exhibited mostly larger relative proportions in indoor air. Interestingly, Pseudomonas was also found on the bathtub surfaces sampled, in line with studies showing that Pseudomonas spp. colonise plumbing/water distribution systems and toilet bowls, and can be, therefore, aerosolised and settle on surrounding surfaces (Bédard et al., 2016; Pitts et al., 1998). Both Paracoccus and Pseudomonas comprise skin-associated species (Cosseau et al., 2016), too, and their presence in the residential air and surfaces could be also the result of skin shedding. It has to be noted, though, that the microbiome in indoor environments is also influenced by the occupants’ behaviour, such as cleaning frequency. In addition, micro-organisms can be carried inside by the occupants via footwear or clothing, followed by subsequent resuspension. For example, outdoor environment-associated taxa such as Alcanivorax, Microbulbifer, Colwellia and Pseudophaeobacter, which were highly abundant during spring in the semi-urban residence, exhibited I/O ratios > 1 indicating an indoor origin, which could not be specified.
The aerosol bacterial communities exhibited significant compositional variation across seasons, both indoors and outdoors, with seasonality being a stronger predictor of community structure compared to site. The seasonal variation observed outdoors is in line with previous HTS-based long-term monitoring surveys of outdoor bacterial aerosols (e.g., Du et al., 2018; Gandolfi et al., 2015). In our study, the seasonal factor for the outdoor environment had a larger effect on particles > 2 µm. It has been previously reported that large bioparticles in the atmosphere show strong seasonal variations due to their dependence on the plant phenology, whereas the abundance of smaller bioaerosol particles (< 2 µm) exhibit minimal variation (Matthias-Maser & Jaenicke, 2000; 1995).
Results to date have pointed out that the seasonal dynamics of the outdoor airborne bacteria might not necessarily be reflected indoors, mainly attributed to the dominance of human occupancy-associated bacteria, which have a homogenising effect on bacterial aerosols typically observed over time in enclosed spaces (Prussin et al., 2016; Adams et al., 2014; Gaüzère et al., 2014). Despite the strong contribution of the residents as a source for bacteria in the current study, one major parameter that contributed to the distinct seasonal differences, indoors, was the natural ventilation occurring throughout the year in both houses, which has been shown to increase the similarity between indoor and outdoor air community composition (Meadow et al., 2014; Kembel et al., 2012). Naturally ventilated buildings cannot limit the influx of outdoor coarse particles, which carry a more distinct seasonal signal.
The compositional dissimilarity between indoor and outdoor bacterial aerosols was higher during winter that windows mostly remain shut. In contrast, a pronounced increase of bacteria indicative of the outdoor environment was observed in indoor air during the warmer periods, when natural ventilation rates are highest, including taxa such as Rhodococcus, and Hymenobacter, typically recovered from soil and aqueous environments, Streptomyces, which is abundant in soil, and Pedobacter, also related with soil and sediments, reflecting the stronger influence of outdoor sources. Several environmental taxa found increased in both sites during summer, such as Alphaproteobacteria, Betaproteobacteria, Cytophagia, have been previously detected in river water and freshwater systems (e.g., Jordaan & Bezuidenhout, 2016). However, despite the fact that both study flats were situated in riverside blocks, source tracking analysis estimated that the river water accounted for a very low contribution. Bacteroidetes, which were found outdoors throughout the year, also became distinctly increased in the air of the urban house in summer. The occurrence of Bacteroidetes, which generally comprise gut mammalian microbiota, in outdoor air in cities has been previously linked to dog faecal material (Bowers et al., 2011).
It should be noted that the two sites exhibited different seasonal patterns, with the highest degree of similarity between indoor and outdoor aerosol microbiomes observed in summer for the urban site and in spring for the semi-urban site. The particular patterns were also reflected in the increase of the contribution of outdoor air as a source during summer and spring for the urban and semi-urban houses, respectively, compared to winter. Although establishing a clear temporal pattern is residence-dependent, due to the impact of house-specific sources and ventilation patterns, this difference might be the result of intra-season variability as sampling for each residence took place on different dates per each season. Within-season community variability could be also related to the local plant phenological periods, as noticed by Weikl et al. (2016), which could be easily overlooked due to limited time-point samples per season. In addition, seasonal variation associated with large-sized particles, which, in our study, exhibited higher bacterial richness, might be also related to the presence of pollen grains in the air (Matthias-Maser & Jaenicke, 1995), which may act as carriers of highly diverse bacterial communities (Ambika Manirajan et al., 2016; Obersteiner et al., 2016).