The ignored diversity: complex bacterial communities in intensive care units revealed by 16S pyrosequencing

Abstract

Indoor microbial communities play an important role in everyday human health, especially in the intensive care units (ICUs) of hospitals. We used amplicon pyrosequencing to study the ICU microbiome and were able to detect diverse sequences, in comparison to the currently used standard cultivation technique that only detected 2.5% of the total bacterial diversity. The phylogenetic spectrum combined species associated with the outside environment, taxa closely related to potential human pathogens, and beneficials as well as included 7 phyla and 76 genera. In addition, Propionibacterium spp., Pseudomonas spp., and Burkholderia spp. were identified as important sources of infections. Despite significantly different bacterial area profiles for floors, medical devices, and workplaces, similarities by network analyses and strains with identical molecular fingerprints were detected. This information will allow for new assessment of public health risks in ICUs, help create new sanitation protocols, and further our understanding of the development of hospital-acquired infections.

Figure 1. Cartoon illustration showing each sampling site of the intensive care unit (ICU).

Origin of the single samples is explained in Table 1.

Figure 2. Schematic drawing showing the detected bacterial genera found in the three different areas (floor, devices and workplaces) in the intensive care unit (ICU).

Overlaps between the facilities are indicated by the arrangement of the circles.

Figure 3. The bacterial communities of the floor environment (A), medical devices (B) and workplaces (C).

Relative clone composition of genera was determined by pyrosequencing of 16S rRNA from metagenomic DNA extracted from the intensive care unit. The identification of the closest strain based on 16S rRNA sequence similarity was achieved using the web server SnoWMAn 1.11 (http://snowman.genome.tugraz.at). Phylogenetic groups accounting for 1% of all quality sequences are summarized in the artificial group Other. Multi-colored charts at the legend are shown for each genus and sample correspondingly.

Figure 4. Bacterial communities associated with different areas of the intensive care unit by principal coordinate analysis (PCoA).

Two- (A) and three-dimensional (B) PCoA plot based on the weighted UniFrac distance matrix. Percentage of the diversity distribution explained by each axes is indicated on the figure. Samples associated with the floor (green triangle), medical devices (red rectangle) and workplaces (blue points) are shown as single points.

Figure 5. Profile clustering Cytoscape network visualize the 40 most abundant OTUs across the floor (green), medical devices (red) and workplaces (blue) of the ICU.

Node sizes correspond to the mean relative abundance between the three sampling areas where two (three) sampling groups feature the same relative abundance over the respective samples and the full nodes are subdivided into two (three) areas. A comparative node (grey) indicates the size of a node that would represent 100 reads in a sample.

Figure 6. BOX PCR fingerprints of genetically very similar Staphylococcus species isolated from nearly all sampling sites.

The similarity of BOX patterns between them was more than 80%.