Diversity of Bacterial Biofilm Communities on Sprinklers from Dairy Farm Cooling Systems in Israel

Abstract

On dairy farms in hot climates worldwide, cows suffer from heat stress, which is alleviated by the use of water cooling systems. Sprinklers and showerheads are known to support the development of microbial biofilms, which can be a source of infection by pathogenic microorganisms. The aim of this study was to investigate the presence of microbial biofilms in dairy cooling systems, and to analyze their population compositions using culture-independent technique, 16S rRNA gene sequencing. Biofilm samples were collected on eight dairy farms from 40 sprinklers and the microbial constituents were identified by deep sequencing of the 16S rRNA gene. A total of 9,374 operational taxonomic units (OTUs) was obtained from all samples. The mean richness of the samples was 465 ± 268 OTUs which were classified into 26 different phyla; 76% of the reads belonged to only three phyla: Proteobacteria, Actinobacteria and Firmicutes. Although the most prevalent OTUs (Paracoccus, Methyloversatilis, Brevundimonas, Porphyrobacter, Gp4, Mycobacterium, Hyphomicrobium, Corynebacterium and Clostridium) were shared by all farms, each farm formed a unique microbial pattern. Some known potential human and livestock pathogens were found to be closely related to the OTUs found in this study. This work demonstrates the presence of biofilm in dairy cooling systems which may potentially serve as a live source for microbial pathogens.

Fig 1. Biofilm in the water sprinklers of dairy farm cooling systems.

(a) Typical cooling system in the holding area of a milking parlor. Water sprinklers above the cows are used to soak/wet the cows. (b) Dismantled sprinkler and its filter. (c) Biofilm sampled from sprinkler filter stained with Acridine orange and visualized by epifluorescence microscopy or (d) cultured on blood agar. Scale bar for panel c is 20 μm.

Fig 2. Dairy farms and their microbial abundance.

Names and locations (geographic coordinates) of the dairy farms and relative abundance of the most common phyla as revealed by 16S rRNA gene sequencing analysis. The map was constructed using ArcMap 10.0 software (Esri, Redlands, CA).

Fig 3. Rarefaction curves of 40 filter samples at a cutoff level of 3%.

The rarefaction curve, plotting the number of observed OTUs (sharing ≥97% identity) as a function of the number of sequences, was computed using the RDP Pyrosequencing Pipeline Rarefaction tool.

Fig 4. Overall microbial abundance.

Mean (red line) and maximum (blue bar) abundance of the most abundant genera in the sprinkler samples.

Fig 5. Comparison of bacterial communities from sprinkler samples.

NMDS ordination plot comparing sprinkler bacterial communities from different dairy farms (1–8); farm 6a and farm 6b denote samples obtained from the feeding bank and holding area, respectively, on the same farm. Each data point in the NMDS plot represents the bacterial community identified from a single sprinkler sample. Convex hulls are drawn for each farm representing the smallest convex polygon containing all points. Comparison using MRPP revealed a significant difference between the farm samples (P = 0.005).