Diversity and seasonal dynamics of bacterial community in indoor environment

Abstract

Background: We spend most of our lives in indoor environments and are exposed to microbes present in these environments. Hence, knowledge about this exposure is important for understanding how it impacts on human health. However, the bacterial flora in indoor environments has been only fragmentarily explored and mostly using culture methods. The application of molecular methods previously utilised in other environments has resulted in a substantial increase in our awareness of microbial diversity.

Results: The composition and dynamics of indoor dust bacterial flora were investigated in two buildings over a period of one year. Four samples were taken in each building, corresponding to the four seasons, and 16S rDNA libraries were constructed. A total of 893 clones were analysed and 283 distinct operational taxonomic units (OTUs) detected among them using 97% sequence similarity as the criterion. All libraries were dominated by Gram-positive sequences, with the most abundant phylum being Firmicutes. Four OTUs having high similarity to Corynebacterium-, Propionibacterium-, Streptococcus- and Staphylococcus- sequences were present in all samples. The most abundant of the Gram-negative OTUs were members of the family Sphingomonadaceae, followed by Oxalobacteraceae, Comamonadaceae, Neisseriaceae and Rhizobiaceae. The relative abundance of alpha- and betaproteobacteria increased slightly towards summer at the expense of firmicutes. The proportion of firmicutes and gammaproteobacteria of the total diversity was highest in winter and that of actinobacteria, alpha- and betaproteobacteria in spring or summer, whereas the diversity of bacteroidetes peaked in fall. A statistical comparison of the libraries revealed that the bacterial flora of the two buildings differed during all seasons except spring, but differences between seasons within one building were not that clear, indicating that differences between the buildings were greater than the differences between seasons.

Conclusion: This work demonstrated that the bacterial flora of indoor dust is complex and dominated by Gram-positive species. The dominant phylotypes most probably originated from users of the building. Seasonal variation was observed as proportional changes of the phyla and at the species level. The microflora of the two buildings investigated differed statistically and differences between the buildings were more pronounced than differences between seasons.

Figure 1

Collector's curves for the combined sequence datasets for Buildings 1 and 2. Collector's curves illustrating the observed and estimated OTU richness and the number of sequences sampled in the pooled datasets of Building 1 (A) and Building 2 (B) at the evolutionary distance of 0.03. ▲, estimated number of OTUs; -, observed number of OTUs.

Figure 2

Abundance of clones and OTUs. Distribution of the clones and OTUs in taxonomic groups and their abundance in the individual samples are displayed. Number of clones per season in the largest phyla is indicated in the row with the name of the phylum. 1, Building 1; 2, Building 2; W, winter; Sp, spring; Su, summer; F, fall. Uncultured refers to sequences of species not yet cultured; unknown refers to a sequence similarity less than 95% to a database sequence, however if the clone was clearly affiliated to a certain group in the phylogenetic tree, it was assigned to that group. In addition to the clones listed in the figure, there was one clone in Building 2, summer sample affiliating with Planctomycetes, one clone in Building 1, spring sample affiliating with Chloroflexi, and one clone with unknown phylogenetic affiliation in Building 1, winter and fall samples. The grey shading indicates the number of clones in the sample belonging to the respective family. White indicates zero clones, the grey shadings 1, 2–5, or 6–14 clones, and black indicates 15 or more clones.

Figure 3

Cluster analysis of the clone libraries. The tree was created with Cluster environments analysis of the Unifrac program. Jackknife with 100 permutations was performed to assess the reliability of the tree. Jackknife values over 50 are given at the corresponding branches. 1, Building 1; 2, Building 2; W, winter; Sp, spring; Su, summer; F, fall.

Figure 4

Principal coordinate analysis. Unifrac PCA analysis of the eight clone libraries. First principal component (P1) represents the season and P3 represents the building. 1, Building 1; 2, Building 2; W, winter; Sp, spring; Su, summer; F, fall.