Shedding Light on the Complex of Internal Carbon Driven Denitrifiers in Biofilm & Floc

Introduction Achieving energy-neutrality is an ongoing driver for new wastewater treatment approaches[1]. Since the successful application of anaerobic ammonia oxidation (anammox) in sidestream processes, concerted efforts have been devoted to integrating anammox into mainstream processes. Both partial nitritation (PN) and partial denitrification (PdN) produce the necessary electron donor, nitrite, for anammox bacteria to oxidize ammonia. Mainstream PN with anammox (PNA) requires the least amount of aeration energy and carbon addition, but stability of PN in mainstream treatment is still difficult to achieve. In contrast, the stability of PdN treating low carbon influent (C/N ~ 2-3) has been found to be robust and attracts increasing interest for implementing PdN with anammox (PdNA) in mainstream processes[2]. In 2020, pilot work at the Hampton Roads Sanitation District's (HRSD) James River Treatment Plant (JRTP) confirmed that stable PdN could be achieved in a second anoxic PdNA integrated fixed film activated sludge (IFAS) process[3]. Using feedback-controlled COD dosing to maintain an effluent nitrate residual, the ratio of consumed COD to removed total inorganic nitrogen was reduced to 1.08 to 2.183. In addition to PdN, these reductions in excess carbon could potentially be attributed to denitrification driven by bacterial intracellular stored carbon (here termed as Internal carbon-driven denitrification, ICD). However, the functional group (flocs vs. biofilm) that contributes to this phenomenon has not been identified. In addition, there is a knowledge gap for which conventional denitrifiers or other bacteria contribute to intracellular stored carbon to perform denitrification. To address this gap, this study employed flow activated cell sorting (FACS) and Single Cell Raman Spectrum (SCRS) techniques in combination with 16S sequencing to track the internal carbon accumulating organisms a second anoxic IFAS tank, with the aim to elucidate the carbon sources and organisms that may support ICD. Methods Sampling for this study was conducted at HRSD's JRTP, an A2O plant, where there were two parallel pilot PdNA IFAS reactors treating mixed liquor from JRTP's aerobic effluent. One IFAS reactor was acclimated to glycerol as an external carbon source, and the other was acclimated to methanol.[4] Specific denitrification rate (SDNR) tests were conducted on both IFAS reactors as well as just the suspended flocs using glycerol, methanol, or no external carbon during batch tests. The concentrations of nitrite, nitrate, and ammonia were measured throughout the test using HACH tubes. Floc samples were sampled directly while biofilm samples were collected by washing and shaking biofilm off vigorously in 0.1% NaCl solution. Sludge samples were fixed immediately and stored at 4 oC. After the batch test, fixed samples went for Fluorescence activated sludge sorting (FACS). SYBR Green and Nile Red were used in FACS process to bind with DNA and PHA (and similar lipid structures), respectively. Samples before and after sorting were sent out for 16S sequencing. SCRS detection was prepared according to Majed and Gu (2010).[5] Results 1. PdN and ICD were contributed by biofilm and flocs, separately. The SDNR rate in each IFAS reactor and the suspended sludge were measured separately. After subtracting the nitrogen removal contributed by flocs, the SDNR in the biofilm with glycerol and methanol were 0.36 and 0.49 g-N/gVSS/day, respectively. The SDNR in the flocs were 0.06 and 0.07 g-N/gVSS/day, respectively. The nitrite accumulation rates in the flocs were only 1-4% of that in the biofilm, indicating that the PdN in the IFAS system was mainly contributed to the biofilm. When no external carbon was provided, the biofilm SDNR would rapidly decrease to less than 10% of the original rate (Figure 1). In contrast, the SDNR in flocs without carbon addition decreased only 17%~29%. Additionally, flocs alone did not significantly consume external carbon. These results indicate that nitrite accumulation was mostly attributed to the biofilm through external carbon driven PdN, while organisms in flocs seemed to be more prone to internal carbon accumulation, and they performed additional ICD. 2. PHA-like internal carbon accumulators changed at the end of denitrification process. To further elucidate the potential internal carbon-driven denitrifiers, flocs samples were fixed before and after SDNR tests. It should be mentioned that Nile Red would stain not only PHA, but also lipids with a similar structure. Therefore, cells stained by SYPR Green and Nile Red were considered PHA-like internal carbon accumulators and sorted via FACS for microbial community analysis (Figure 2). Without sorting, the community was unchanged. However, the PHA-containing community changed significantly after 3h carbon-free denitrification (Figure 3). Genera labeled in red represent PHA accumulators that disappeared after SDNR tests, linking them to the observed ICD. To further analyze the intracellular carbon compounds, SCRS analysis showed ~20-30% of floc cells contained PHA, though no trends were evident (Figure 4). Undetectable GC-MS PHA corroborates that PHA is likely not the main carbon source in ICD.Though greater than PHA, the limited glycogen-containing cells (~10%) suggest glycogen may also not be the primary ICD substrate. Further research into identifying the specific carbon source(s) in post-denitrification is warranted.