Influence of seasonality, wastewater treatment plant process, geographical location and environmental parameters on bacterial community composition in selected South African domestic activated sludge wastewater treatment plants

Abstract

Microbial communities are responsible for the bioremediation of wastewater (WW) in activated sludge (AS) wastewater treatment plants (WWTPs). Process performance is reliant on the metabolic activities of the microbial species present in the AS, which include, amongst other functional groups, chemoorganotrophic (heterotrophic) bacteria responsible for mineralization of organic carbon and denitrification, and nitrifying autotrophs and heterotrophs. Little is known about the underlying operational factors and geographical factors (ecophysiology) influencing the complex dynamics of the microbial populations in South African (SA) AS bioreactors, and how those dynamic interactions affect the systems’ functional stability. Prior to this study, there were no studies detailing the overall bacterial composition in SA wastewater treatment plants (WWTPs). In this study, we compared the bacterial communities in domestic WWTPs with different configurations from two different geographical locations in SA by analysing the bacterial metagenomes and physicochemical data. Samples (AS sludge and influent) and were collected from 3 WWTPs located in the city of Cape Town (CoCT) and 6 (WWTPs) located in the city of Ekurhuleni (CoE). The metagenomic bacterial community structures in 141 samples were first assessed using terminal restriction fragment polymorphism (T-RFLP). Based on the T-RFLP results, next generation sequencing (NGS) was performed by sequencing amplicons of the V4 region of the 16S rRNA gene using the Illumina MiSeq platform on a rationalised set of 50 samples. Univariate and multivariate statistical tools were used to analyze the output data and assess the influence of seasonality, WWTP process (configuration), geographical location (site) and environmental variables on bacterial community selection. Physiochemical analyses were performed in the bioreactors in-situ (temperature, pH, dissolved oxygen (DO)), and on AS and/or influent samples (dissolved sludge volume index (DSVI), volatile suspended solids (VSS), biological oxygen demand (BOD), chemical oxygen demand (COD), total phosphate (TP), ortho phosphate (σ-P), sulfate as sulfur (SO42-_S), ammonia as nitrogen (NH4+-N), nitrates and nitrites as nitrogen (NO3-/NO2-N), total alkalinity (T.alk.) and volatile fatty acids (VFA) concentrations). Good separation of the data points representing each geographical location (factor ‘site’) was noted on non-metric multidimensional scaling (nMDS plots) obtained using both the T-RFLP and NGS data. Highly significant differences (P=0.001) were found between the 2 sites using 3-way fully nested analysis of similarity (ANOSIM), validating the nMDS results. As the geographical location played such a dominant role on the bacterial community selection, the factors ‘season’ and ‘configuration’ were analyzed separately for each geographical location. In the CoE WWTPs, no significant differences (P>0.05) were found for either of these factors when analysing the T-RFLP and NGS data using 2-way nested ANOSIM. In contrast, using the same methodology for the CoCT WWTPs, significant differences (0.05>p<0.01) were noted for the factor ‘configuration’. The highly significant differences in the overall bacterial structures from the 2 geographical locations could not be linked to climatic or WWTP operational differences. However, the there were highly significant differences (students t-test p<0.005) in some of the physicochemical parameters. With the influent from the CoE being generally more dilute in nature. It was hypothesized that this was the primary reason for the differences in the AS bacterial communities from each site. This was validated using BEST, 3-way ANOSIM and PCA. The BEST analysis indicated that the AS TP concentration was the primary driver for differences in the bacterial community structures from the 2 geographical locations. Concentrations of AS TP in the CoCT WWTPs and CoE WWTPs were 4.0 ± 6.4 and 9.5 ± 7.2, respectively. High throughput sequencing generated a total of 2 645 unique zero radius operational taxonomic units (zOTUs) with a frequency of 530 594 across the 50 samples with median of 10 101 per sample (range 7592 to 14 890). In total, 27 identified phyla were observed. Proteobacteria was the most abundant phylum in all AS samples, accounting for 34.75 - 86.42% relative abundance (RA) at the CoCT and 37.133 – 77.89% RA at the CoE. The other two most dominant phyla were Bacteroidetes (5.08 – 32.72% RA at CoCT and 10.50 – 45.88% RA at CoE), and Actinobacteria (2.82 – 21.35% RA at CoCT and 1.43 – 14.16% RA at CoE). At the family level, a total of 277 families were identified, of which 256 were common to both CoCT and CoE. At this level, the taxonomic profile differed between the two geographic locations. Pseudomonadaceae was the most dominant family at the CoE, accounting for 0.53 – 51.21% RA while Moraxellaceae dominated at the CoCT accounting for 0.25 – 39.83% RA. A total of 832 genera were identified, of which 749 were common to both the CoCT and CoE WWTPs. Pseudomonas was the most dominant genus, with a mean estimated absolute abundance (EAA) of 23 358 ± 19 720 cells/ngDNA.gsludge-1 in the CoE and 20 582 ± 39 028 cells/ngDNA.gsludge-1 in the CoCT.

Overall, this study showed that geographically related physicochemical differences were more significant than season or WWTP configuration on bacterial community selection in the WWTPs in the CoE and CoCT. It was also shown that community fingerprinting using T-RFLP and high throughput sequencing yielded similar results. It was concluded that although T-RFLP is an older and is a less sophisticated fingerprinting technique than NGS and does not provide direct information on the taxa that are present, it can still be a valuable and reliable screening tool in cases where it important to include large datasets and funding is limited. It is recommended that future work is directed towards qualitative and quantitative analyses of the functional microbial species, particularly the nitrifying bacterial populations in South African AS WWTPs, and comparison of the results with global studies.