Understanding Loss of Nitrification At A Refinery Wastewater Treatment Plant Using 16S rRNA Sequencing

The unexpected discovery of this research was the significant role that simultaneous nitrification-denitrification organisms play in a fully aerobic activated sludge system by using rRNA analysis as a window into the process. This observation highlights how metagenomics can be applied in a full-scale treatment process to better understand nitrification mechanisms with an emphasis on lost or partial nitrification events. Refinery biological systems are subject to highly variable influent characteristics that can adversely affect the ability to nitrify and meet ammonia discharge limitations. Typical analysis methods of mixed liquor volatile suspended solids (MLVSS) and ammonia / nitrite / nitrate testing often falls short in predicting loss of nitrification. Applications such as BioHealth* monitoring which measures cellular adenosine triphosphate (cATP) provides more detailed information on the health of the system, however it does not differentiate between nitrifying bacteria and more abundant heterotrophic bacteria. Advancements in DNA sequencing, such as 16S rRNA sequencing, has made quantitative speciation realistic for academic, industrial, and municipal end users. A refinery wastewater treatment plant (WWTP)†in southwestern USA was monitored over the course of several years (June 2017 to November 2020) to investigate nitrogen removal. Prior to, and throughout the study, the WWTP experienced frequent loss of complete nitrification. Loss of nitrification was a concern because the WWTP directly discharges into the receiving stream and the effluent exhibited acute toxicity, which in part, has been attributed to partial nitrification and 'nitrite lock'. The WWTP consists of dissolved air flotation, a single aerated activated sludge bioreactor (bioreactor 1), parallel aerated polishing bioreactors (bioreactor 2a and 2b) and secondary clarification. In theory, the WWTP is designed to achieve complete nitrification in bioreactor 1. The plant does not include designated anoxic zones dedicated for denitrification. †Use of name request pending In general, three different nitrification conditions were observed: complete nitrification (nitrite and ammonia removal), complete ammonia oxidation with incomplete nitrite oxidation (ammonia < discharge limit, nitrite > discharge limit) and no nitrification (ammonia > discharge limit). A full suite of process data was collected, including metagenomic analysis consisting of 16S rRNA sequencing, mixed liquor cATP levels, nitrogen loading, flow rate, and bioreactor effluent ammonia, nitrite and nitrate concentration. cATP is a measure of the quantity of active biomass in a sample. Samples were collected from bioreactor 1 for 16S rRNA sequencing (Illumina MiSeq) to identify the relative abundance of all bacteria and archaea in the sample to the genus level. The 16S gene is sequenced to identify microorganisms as it is a highly conserved gene present in all bacteria and archaea that contains variable regions unique to each genus. Organisms were identified by aligning the sequences to a comprehensive taxonomic rRNA sequence database (SILVA). By combining mixed liquor cATP measurements (BioHealth*) with the relative abundance output of 16S sequencing, a true estimate of an organism's concentration was calculated. Four functional groups of nitrogen-related organisms were identified: ammonia oxidizing, nitrite oxidizing, denitrifying and simultaneous nitrification-denitrification (SND). SND organisms are capable of heterotrophic nitrification as well as aerobic denitrification (Medhi et al. 2017). This contrasts with conventional theory that relies on separate autotrophic ammonia oxidation for nitrification and anoxic zones for denitrification. Nitrogen removal from wastewater can be simplified into the following steps: Nitrification Ammonia Oxidation: Ammonia (NH3) → Nitrite (NO2¯) Nitrite Oxidation: (NO2¯) → Nitrate (NO3¯) Denitrification Nitrate Reduction: (NO3¯) → (NO2¯) → NO → N2O → N2 The predominant ammonia oxidizing, nitrite oxidizing, and denitrifying organisms were the classic aerobic, autotrophic Nitrosomonas sp., and Nitrospira sp., and Thauera sp., a heterotroph which can grow under aerobic or anoxic conditions (Scholten et al., 1999), respectively. SND organisms showed more diversity however Pseudomonas sp. were the predominant SND organisms present. Figure 1 shows the abundance of each functional group over time. In general, denitrifying organisms were the most abundant. Ammonia oxidizing organisms had relatively small abundances compared to the other functional groups. It is important to note that, while the abundance of ammonia-oxidizing organisms is relatively low, it does not mean the population is insufficient to achieve process goals.
Preliminary analyses of metagenomics and process data was performed to better understand the role organism abundance played in treatment performance. Figures 2 and 3 show the abundance of Thauera over time, along with effluent nitrite and ammonia concentrations, respectively. Thauera, which is a denitrifying organism, tended to be high during periods of partial nitrification or 'nitrite lock'. It also appeared high just prior to loss of nitrification. The limited amount of metagenomic data relative to process data makes correlations difficult, however the trends are still valuable insights into treatment processes. A hypothesis was developed to explain the relationship between Thauera and nitrite lock. Thauera has been shown to perform denitrification in aerobic environments (Chen et al. 2018). Therefore, it is possible that concurrent nitrite oxidation and nitrate reduction was occurring, yielding minimal net nitrite reduction. Further testing is required to verify this hypothesis. The relationship between Thauera and ammonia oxidation is less clear. A counterintuitive, direct relationship was observed between Thauera and ammonia oxidizing organisms (Figure 4). Further investigation is required to understand these relationships; however, it may suggest elevated nitrite associated with high abundance of Thauera may be inhibiting ammonia oxidation. Certain strains of ammonia oxidizing organisms have been shown to be inhibited by nitrite (Cua and Stein, 2011). A Principle Coordinates Analysis (PCoA) plot was generated using the 16S rRNA sequencing data (Figure 5) where the microbial community is similar between samples that are clustered together and differs from samples further away.
Results
showed a clear separation between the microbial populations noted by Group 1 and Group 2 on the plot. Organisms in Group 1 are samples prior to and including January 2019 while organisms in Group 2 were from samples September 2020 onward. It is important to note that process improvements were made during this time period for overall process stability and issues related to nitrogen removal were lessened as seen in Figures 2 and 3. Although relationships between the results of the PCoA plot and 10 process parameters to explain the strong separation in microbial communities including F:M, TOC and DOC removal have been explored, further investigation into these relationships and specific changes to the microbial community is required.
This research highlights the synthesis between metagenomics, ATP and overall process monitoring which may provide an early warning of reduced nitrification, thus allowing intervention to prevent discharge exceedances and identifying why nitrification loss occurs. It is anticipated that further investigation into the changes in the microbial community and improved process control will provide insight into how this data can be used directly in process control to maintain nitrogen removal processes.