The authors have declared that no competing interests exist.
Although human occupancy is a source of airborne bacteria, the role of walkers on bacterial communities in built environments is poorly understood. Therefore, we visualized the impact of walker occupancy combined with other factors (temperature, humidity, atmospheric pressure, dust particles) on airborne bacterial features in the Sapporo underground pedestrian space in Sapporo, Japan. Air samples (
Humans spends over 90% of their time in indoor environments such as houses, offices, schools, hospitals, restaurants, and other public spaces such as underground pedestrian walkways [
The features and rate of bacterial shedding from individuals could be reflected in the bacterial community structures of built environments; therefore, they may serve as markers for estimating healthy indoor air conditions. Although building ventilation systems and sanitation can influence the community structure of indoor air bacteria, the abundance of human-derived bacteria is more than two-fold greater in human occupied indoor environments than unoccupied areas [
Air also contains a significant number of dust particles composed of both inorganic and organic materials, including bacteria, and most airborne bacteria found indoors are on particles with diameters of 0.1–10 μm [
Therefore, we visualized impact of walker occupancy with other factors including temperature, humidity, atmospheric pressure, and dust particle numbers on changing airborne bacterial features in a built environment, the Sapporo underground pedestrian space, located in Sapporo, Japan. The results revealed an impact of walker occupancy on airborne bacteria that was influenced by temperature and humidity.
The study space was an underground pedestrian passageway 520 m in length and 16–18 m in width that was built in 2011. There is open access to the area from 5:45 h to 0:30 h throughout the year (
(A) Representative features showing the pedestrian with walkers. (B) Map showing the pedestrian structure with the location of public spaces or entrances/exits. The photo and the map were obtained with permission from the following website (
The weather differed on different sampling days (Japan Weather Association:
(A) Changes in temperature (°C). (B) Changes in humidity (%). (C) Changes in atmospheric pressure (hpa). Left and right panels show the daily and monthly changes, respectively. Asterisks indicate a significant difference (
Pedestrian and particle numbers were also measured at 18 time points from 8:00 h to 20:00 h on each regular sampling day, and at 6 time points from 5:50 h to 7:50 h and 22:15 h to 24:35 h on baseline sampling day. When compared among regular sampling days, there was no significant difference in the average number of walkers per 10 min among sampling days (
(A) Changes in walker numbers. (B) Changes in particle numbers (particle size Δ0.5: 0.3–0.5μm). (C) Changes in numbers of particles (particle size Δ1.0: 0.5–1.0μm). (D) Changes in numbers of particles (particle size Δ5.0: 1.0–5.0μm). Left and right panels show daily and monthly changes, respectively. Asterisks indicate a significant difference (
A total of 4,800 L of air was trapped on a filter every 2 hours on each regular sampling day from 8:00–20:00 h or every 1 hours on each baseline sampling day from 5:50 to 7:20 h and 22:15 to 24:15 h. Each of the filter-rinsed solution samples was divided into two tubes to determine the number of CFUs and for DNA extraction followed by 16S rDNA sequencing. Because the airborne bacteria originated from the built environments and the pedestrians themselves, CFUs were enumerated by culture on SCD (soybean-casein digested) (nutrient rich medium to reflect pedestrians) and R2A plates (nutrient poor medium to reflect the built environment). The average number of CFUs was 59.9±19 (May 2), 12,366±13,507 (June 1), 3,069±7,058 (July 5), and 36.8±23.4 (July 15). In contrast to the environmental factors shown above, there were no significant differences in CFUs observed among sampling days, although there appeared to be complicatedly but a change depending on up/down of temperature and humidity with walker numbers (
(A) Changes in CFU numbers estimated by the SCD plate culture (rich medium for “walkers”). (B) Changes in CFU numbers estimated by R2A plate culture (poor medium for “built environment”). (C) Total CFUs. Left and right panels show daily and monthly changes, respectively.
(A) Changes in OTU numbers estimated based on 16S rDNA sequence analysis and BLAST search. Left and right panels show daily and monthly changes, respectively. (B) Clustering of the phylum type defined by 16SrDNA sequencing. As mentioned in the text, the OTUs obtained from the mock control were removed from those obtained for each sample. These OTU data were similarly processed with a setting [>80% filtering and Pearson correlation (centered)]. After replaced to the files with a ‘cdt’ extension in the software of Cluster 3.0 for visualizing on the TreeViewX.
Pearson’s correlation analysis and cluster analysis were conducted to evaluate the relationship between airborne bacteria and the degree of walker occupancy, as well as changes in environmental factors (temperature, humidity) and the amount of dust particles. A significant positive correlation (correlation coefficient: >0.4 with
(A) Matrix showing comparison of Pearson's correlation coefficient among factors. Values show the correlation coefficient. Bold values showing positive and negative correlation with colors indicate statistical significance (a correlation coefficient value of >0.5 or <−0.05 with a
Because a large number of bacteria are emitted from skin or oral cavities through coughing, sneezing, talking, and breathing, bacteria shed by humans are considered a source of airborne bacteria in built environments [
Variations in walker numbers were similar among sampling days, with peaks occurring in the morning and evening, although there was no difference in the average number of walkers between sampling days. These results indicate that the pedestrian space is regularly used for commuting to school or work. Although a small peak was observed around noon on June 1, this was because people entered the walkway to avoid a short period of rain. In contrast to the number of walkers, minimal changes in particle numbers were observed throughout the day. Although it is common for indoor air particles to be critically influenced by outdoor air [
There was no significant difference in CFU numbers among sampling days. Moreover, the numbers of CFUs showed irregular variations throughout the day. Evaluation of OTUs revealed 22 phylum types, with the majority belonging to
No direct interaction of walker occupancy with airborne CFUs and OTU features was seen upon analysis by Pearson's correlation coefficient test. However, cluster analysis indicated an obvious lineage consisting of walker occupancy, CFU numbers, OTU types, dust particles with small size, and seasonal factors, including temperature and humidity. These results suggest that walker occupancy could be indirectly related to airborne bacterial features in the underground space. We proposed the following possible scenario to explain the changing bacterial features in response to walkers (
Upper panel (small number of walkers) showing the number of bacteria released from individuals is minimal. Middle panel (moderate number of walkers) showing that although increasing walker occupancy facilitates the number of floating small dust particles with bacteria released from individuals, the number of bacteria is still minimal. Lower panel (high number of walkers) showing that increasing temperature and humidity in the presence of a high concentration of small dust particles causes the bacteria released from walkers to reach the maximum level.
Taken together, the results of this study indicate that there is a chain of positive correlations of walker occupancy with airborne bacteria in response to increased temperature and humidity in the presence of airborne small particles in the Sapporo underground pedestrian space. Thus, the small airborne particles of dust at high temperature and humidity may be a crucial factor facilitating the release of bacteria from walkers in constructed environments. These findings advance our knowledge regarding complicated features of bacteria released from human individuals in built environments, which will improve public health in urban communities.
Air sampling with monitoring of environmental factors (temperature, humidity, atmospheric pressure, dust particle numbers) was performed at 8:00–20:00 on May 2, June 1, and July 5, 2016 (regular sampling); air analysis in few walkers was also performed at 5:50–7:50 and 20:15–24:15 on July 17, 2017 (baseline sampling). Samples were collected with air samplers (LV40BW, Shibata Scientific Technology Ltd.) placed approximately 3 m from the wall of the pedestrian walkway and 1.5 m from the floor to approximate pedestrian head, according to previous report [
At 2 h intervals (regular sampling) or at 1 h interval (baseline sampling), filters were aseptically removed, placed in extraction buffer (PBS with 0.05% Triton X-100) on site, and vortexed on laboratory, according to a previous report [
Amplicon sequence analysis of samples was conducted as follows. First, the 16S rDNA amplicon library was generated by PCR amplification with 35 cycles (for regular sampling) or 25 cycles plus enrichment of DNA with magnetic beads (Agencort, Beckman Coulter) using the Ion Plus Fragment Library Kit (Life Technologies). Amplicons were sequenced by the Ion S5 and Ion S5 XL Systems (Thermo Fisher). Amplicon libraries were sequenced on a 318 chip using the Ion Torrent Personal Genome Machine (PGM) system and the Ion PGM Hi-Q kit (Life Technologies). Raw reads were processed by the PGM software to remove low quality and polyclonal sequences. 16 rRNA sequences were analyzed using Metagenomics 16S w1.1 ver.5.2 with the Torrent Suite Software (Life Technologies) to perform OTU clustering. OTU annotation was based on applying the Basic Local Alignment Search Tool (BLAST) to data available in the Greengenes database with a baseline of >90% coverage. The analysis, including quality filtering, OTU production, identification of microorganisms, determination of their abundance in the sample and phylogeny generation, was conducted using the QUIIME softwar.
Airborne bacteria derived from the built environments and the walkers themselves were cultivated using SCD (soybean-casein diseated) (Nissui) (nutrient rich medium for walkers) and R2A plates (BD) (nutrient poor medium for built environment).
Cluster analysis was performed using Cluster 3.0 for Mac OS X (Clustering Library 1.52). Phylogenetic trees generated from aligned population structures were constructed and then visualized in Java TreeViewX (version 0.5.0). Specifically, as mentioned above, the OTUs obtained from the mock control were removed from those obtained for each sample. After that the OTU data were treated into a software, ‘Cluster 3.0’, Specifically, the OTU data were transformed to log scales, and then processed with a setting [>80% filtering and Pearson correlation (centered)]. Meanwhile, all data (temperature, humidity, particle number, atmospheric pressure, OUTs, CFUs, walker’s numbers) used for total cluster analysis were similarly converted to an equivalent range from ‘0’ to ‘1’, and then processed with a setting (>80% filtering and Spearman rank correlation). These data were then replaced to the files with a ‘cdt’ extension in the software of Cluster 3.0 for visualizing on the TreeViewX. In addition, the phylum contents of OTUs for mock controls obtained at May 5 and July 15 was shown as
Comparison of the values between factors was conducted using a Mann-Whitney U test. Correlations among factors (walker’s occupancy, temperature, humidity, atmospheric pressure, dust particle numbers, CFUs) were identified by Pearson’s correlation coefficient test. A correlation coefficient value of >0.5 or <−0.05 with a
The study reported in this manuscript did not involve any human participants, human data, human tissue, data pertaining to specific individuals, or animal experiments. Meanwhile, sampling location and times were selected based on guidance from the City of Sapporo, which maintains the pedestrian space.
The composition of phylum was very similar with a high correlation coefficient value (
(TIFF)
We thank the staff at the Department of Medical Laboratory Science, Faculty of Health Sciences, Hokkaido University, for their assistance throughout this study.