University of Glasgow1, United Utilities2
Introduction
The application of next-generation sequencing techniques to the study of microbial communities that inhabit Drinking Water Systems (DWSs) has revealed the presence of a diverse and abundant microbiome. Different sampling strategies have been applied to study the drinking water microbiome (DWM). Several studies have focused on specific points in the system, characterizing the microbial ecology of finished water at the treatment plant1, water at the distribution system2,3 and tap water4. A comprehensive sampling strategy, looking simultaneously at the microbiome of the systems at different stages (source, treatment plant, distribution system, tap) has also been applied, revealing a reduction in microbial diversity caused mainly by filtration and chlorination1,5, and the influence of the filter in shaping the communities that inhabit the distribution system6. The variation of the DWM at different temporal scales has also been previously explored. Consistently, Proteobacteria has been reported as the dominant bacterial phylum in drinking water systems. Nonetheless, seasonal studies have found Alphaproteobacteria to be the dominant class in summer and fall, Betaproteobacteria to dominate in spring, and an equal distribution of both classes during winter7, while monthly studies have found Betaproteobacteria to be dominant during the summer months, while Alphaproteobacteria dominated during the winter months3. To our knowledge, bacterial diversity variation in DWSs at smaller time scales is still unknown and has not been reported. Similarly, the variation of the DWM in small spatial scales within a distribution system needs further exploration, since the reported studies so far include sampling points that are considerably spatially distributed3, or don't focus specifically on addressing bacterial diversity variability as a function of small spatial scales 8.
Methods and materials
The following sampling strategy was designed and applied to study the diurnal bacterial variation in DWSs. Triplicate samples of disinfected bulk water (15 to 18L) were collected from 5 residences located in Glasgow over 24 hours divided in six time periods (08-12h,12-16h,16-20h,20-00h,00-04h,04-08h). Four houses were located in the same distribution system, while one house was selected as an outlier control fed by a different system. The samples were filtered through sterile 0.22um pore size filters and subject to DNA extraction, triplicate PCR reactions and DNA sequencing in an Illumina MiSeq Sequencing Platform. In addition, four-hour composite samples of water, corresponding to the six time periods mentioned above, were collected and subject to physical and chemical analysis (total organic carbon, temperature, pH, total chlorine, phosphate, ammonia, nitrite, nitrate, turbidity, dissolved oxygen, and conductivity) using standard methods. Sequence processing was conducted using a previously reported protocol9. Statistical analyses were conducted using Mothur10 and R11.
Results and Discussion
Diurnal variation of alpha diversity (i.e. richness) was investigated by calculating the Chao Index for the 5 houses, for each time period. For houses A, B and E, Permutational ANOVA showed a significant difference (p<0.01) between the six time periods, while for houses C and D the difference was not significant. Similar analyses showed that the five houses were significant different from each other in terms of alpha diversity from 04-08Hr, 12-16Hr and 16-20Hr (higher water demand period), and a not significant difference for time periods 08-12Hr, 20-00Hr and 00-04Hr (lower water deman).
Beta diversity estimates (Bray Curtis and Jaccard distances) show two clusters of houses sharing similarities in community structure and membership (houses A, B and C; and houses D and E). Permutational t-tests show that, for houses A, B and C, beta diversity estimates are significantly different (p<0.001, using Bonferroni correction) for the greatest time differences (4-hr and 20-hr). In the case of houses D and E, significant differences were observed between 12-hr and 16-hr groups.
The bacterial communities of the five houses sampled were composed mainly of Proteobacteria. For houses A, B and C, Betaproteobacteria dominated over Alphaproteobacteria in all time periods, while for houses D and E the opposite was observed. This is consistent with the results obtained from the beta diversity estimations (clusters based on Bray Curtis and Jaccard distances), and is attributed to rezoning works carried out by the water company during the sampling period. The third most abundant class is Gammaproteobacteria for all houses sampled. The relative abundance of Alphaproteobacteria showed weak positive correlation with dissolved oxygen measurements (Pearson coefficient=0.52), and weak negative correlation with pH measurements (Pearson coefficient=-0.51). All the other correlation coefficients between mean Proteobacteria classes and water quality parameters were less than 0.50.
Conclusion
A significant change in bacterial diversity over a 24-hour period was been observed in three of the five residences included in this study, whereas significant differences in beta diversity estimates were found for the five houses studied. This is an indication that the structure and the membership of the microbial communities in the distribution system change as a function of time, even for a small time scale (hourly in this case), and this change should be accounted in order to capture the variability of the drinking water microbiome during sampling campaigns. This study provides important insights into the design of sampling campaigns to study drinking water microbial communities. 1. Poitelon, J. B. et al. Variations of bacterial 16S rDNA phylotypes prior to and after chlorination for drinking water production from two surface water treatment plants. J Ind Microbiol Biotechnol 37, 117-128, 2010).
2. Lu, P. et al. Phylogenetic diversity of microbial communities in real drinking water distribution systems. Biotechnology and Bioprocess Engineering 18, 119-124, (2013).
3. Pinto, A. J., Schroeder, J., Lunn, M., Sloan, W. & Raskin, L. Spatial-temporal survey and occupancy-abundance modeling to predict bacterial community dynamics in the drinking water microbiome. Mbio 5, (2014).
4. Holinger, E. P. et al. Molecular analysis of point-of-use municipal drinking water microbiology. Water Research 49, 225-235, (2014).
5. Lin, W., Yu, Z., Zhang, H. & Thompson, I. P. Diversity and dynamics of microbial communities at each step of treatment plant for potable water generation. Water Research 52, 218-230, (2014).
6. Pinto, A. J., Xi, C. & Raskin, L. Bacterial community structure in the drinking water microbiome is governed by filtration processes. Environmental Science & Technology 46, 8851-8859, (2012).
7. McCoy, S. T. & VanBriesen, J. M. Temporal variability of bacterial diversity in a chlorinated drinking water distribution system. Journal of Environmental Engineering-Asce 138, 786-795, (2012).
8. Sekar, R. et al. Bacterial water quality and network hydraulic characteristics: a field study of a small, looped water distribution system using culture-independent molecular methods. Journal of Applied Microbiology 112, 1220-1234, (2012).
9. Schloss, P. D. & Westcott, S. L. Assessing and improving methods used in operational taxonomic unit-based approaches for 16S rRNA gene sequence analysis. Applied and Environmental Microbiology 77, 3219-3226, (2011).
10. Schloss, P. D. et al. Introducing mothur: open-source, platform-independent, community-supported software for describing and comparing microbial communities. Applied and Environmental Microbiology 75, 7537-7541, (2009).
11. R: A language and environment for statistical computing (R Foundation for Statistical Computing, Vienna, Austria, 2014).