A conceptual framework modeling of functional microbial communities in wastewater treatment electro-bioreactors

Various wastewater treatment processes have been developed to reduce water pollution and improve drinking water quality. One of these wastewater treatment technologies is the wastewater electro-bioreactor (EBR), which has been the focus of many research studies elucidating the microbial community present in the bioreactors and electro-bioreactors (ElNaker et al. 2018a, 2019). Several developments and efforts have been made to functionally characterize the microorganisms associated with the removal of organic and inorganic nutrients from wastewater (Juretschko et al. 1998; Daims et al. 2015). The integration of electrochemical processes into membrane bioreactors (MBRs) combines biodegradation, electrochemical and membrane filtration processes into one system, achieving high effluent quality compared to conventional MBRs and activated sludge processes (Giwa & Hasan 2015a, 2015b, 2017; Giwa et al. 2016a, 2016b; Ahmed & Hasan 2017). Bio-electrochemical systems have been reported to treat recalcitrant wastewaters (Xu et al. 2016; Chen et al. 2019). These studies have confirmed that coupling bio-electrochemical systems with up-flow anaerobic sludge blanket (UASB) has resulted in high removal efficiency.

Models are a standard tool for engineers to describe important processes and functionally defined populations in wastewater treatment plants (WWTPs) (Derco et al. 1990; Henze 2008; Valentín-Vargas et al. 2012). Many models have been developed, such as the activated sludge models: ASM1, which focused on the degradation of organic carbon material and nitrogen removal (nitrification and denitrification); ASM2, which included biological phosphorus removal; ASM2d, which included the denitrifying phosphorus accumulating organisms (PAOs); and ASM3, which recognizes the importance of storage polymers in the heterotrophic activated sludge conversions (Henze et al. 2015; Arif et al. 2018).

Other dynamic and simulation models have also been developed for the modification of the activated sludge process and to solve the complexity of several wastewater treatment processes, including moving bed bioreactors (Charfi et al. 2018a, 2018b; Leyva-Díaz et al. 2020), MBRs (Giwa et al. 2019; Mannina et al. 2020), membrane fouling control (Charfi et al. 2018a, 2018b; Song et al. 2020) and integrated MBR model for greenhouse gas emission (Mannina et al. 2018, 2020). These models are a sophisticated way to evaluate process alternatives, optimize design, and analyze and evaluate costs.

Conceptual modeling is a representation of a system that uses ideas and concepts to formulate such a schematic. Conceptual models are developed after process generalization or conceptualization. A recent research study proposed a conceptual ecosystem model of an enhanced biological phosphorus removal (EBPR) system based on the microbial communities in full-scale EBPR plants in Denmark (Nielsen et al. 2010). Their results revealed that a limited number of identical species constitute the majority of communities in all plants. Their model encompasses knowledge about core species in different microbial functional groups and their specific physiological properties.

Conceptual frameworks and designs were developed to facilitate the complexity of wastewater treatment processes. Previous models included either both the analysis and optimization of the process design or modeling of the full scale of microbial communities' structure in WWTPs (Yuan & Blackall 2002; Daims et al. 2006; Comas et al. 2008; Henze 2008). Hu et al. (2019) developed a neural network for an up-flow bio-electrochemical system for treating b-lactams pharmaceutical wastewater under different hydraulic retention times (HRTs) (Hu et al. 2019). Indeed, theoretical and mathematical modeling of EBRs with a focus on the system operational conditions, performance and optimization has been investigated in previous studies; yet there has been limited investigation of microbial community structure and function. However, these models do not integrate the electro-bioreactor operating conditions, performance, and microbial community structure and function.

Focusing on this approach, our laboratory investigated the impact of applying electric fields at different current densities (CD) of 3, 5 and 7 A/m² for EBRs operated in aerobic conditions. These operating conditions were selected based on our previous study (Zeyoudi et al. 2015), and industrial scale MBR WWTPs (Tchobanoglous et al. 2003). The operation of the electrochemical process under aerobic conditions in EBRs enhances the removal efficiency and diversifies the functional microbial community where oxygen, as an electron acceptor, has a major role in the process of oxidation of organic matter and bacterial metabolism.

The identification and analysis of microbial communities in control bioreactors (CBRs) and EBRs operated for 24 hours at CDs of 3, 5 and 7 A/m² and HRT of 6 hours has previously been published (ElNaker et al. 2018a, 2018b, 2019). Average abundance of major functional bacterial groups at the phylum level in both CBRs and EBRs at different CDs and HRTs are summarized in Figure 1. A quick summary of various bacterial phyla demonstrated that Proteobacteria and Bacteroidetes accounted for the largest fraction (33.2–35.3% and 30.4–32.3% average abundance, respectively, across all samples), followed by Chloroflexi (10.0–10.6%), Planctomycetes (9.9–10.5%), Verrucomicrobia (7.1–7.5%) and Nitrospirae (4.2–4.5%) (ElNaker et al. 2018a, 2018b).

The main objective of this study was to present a conceptual model for the microbial communities in wastewater EBRs operated at different CDs and HRTs with efficient removal of soluble chemical oxygen demand (sCOD), phosphorus and nitrogen. The model was based on our recent comprehensive quantitative data on the microbial community structure of EBRs as well as a review of data from other relevant literature. To the best of our knowledge, this study is the first microbial community conceptual model that extends simple community characterization approaches and provides a framework that can aid engineers in the design and operation of wastewater EBRs. The conceptual framework proposed in this study will be useful to assist in the management of the WWTPs treating wastewaters around the world.

The conceptual model presented here was based on quantitative data on microbial populations present in CBRs that did not use electric current as part of the treatment process and EBRs following the methodology shown in Figure 2 (ElNaker et al. 2018a, 2018b, 2019). All the experiments utilized activated sludge samples collected from the MBR plant in Masdar City (Abu Dhabi, United Arab Emirates), which was used as inoculum in batch EBRs operated for 24 hours. Fresh collected activated sludge was fed with synthetic wastewater (0.2% glucose, 1.5 mM ammonium sulfate, 270 μM potassium phosphate, 160 μM magnesium sulfate, 20 μM manganese sulfate, 1.47 μM iron (III) chloride, 20 μM calcium chloride, 330 μM potassium chloride, 300 μM sodium bicarbonate) in aerobic batch CBRs and EBRs. Dissolved oxygen (DO; mg/L), pH, temperature (T; °C) and electrical conductivity (EC; μS/cm) of the synthetic wastewater were analyzed using a HACH HQ40d single-input multi-meter (Hach Company, Loveland, CO, USA). An aluminum anode and a stainless steel cathode were used as the electrodes and were connected to DC power supplies. A 50-mL sample was collected from each reactor cell and centrifuged for 15 min at 3885xg. The supernatant was analyzed by measuring sCOD, phosphorus (as PO₄³⁻-P), ammonia (as NH₄⁺-N) and nitrate (as NO₃⁻-N) using HACH vials LCK 314-1014, LCK 348-349, LCK 303-304 and LCK 339-340, respectively, according to manufacturer's instructions. The EBRs were operated at CDs of 3, 5 and 7 A/m² and a HRT of 6 hours. The effect of varying HRT on CBR and EBR performance was evaluated at HRTs of 6, 10, 16, 24, 50 and 75 hours.

Conceptual models consist of four main components based on Robison's framework (Robinson 2002): objectives (experiment objective), inputs (experimental factors), outputs (responses/outcomes) and model content (scope and level of detail), identifying any assumptions and simplifications. These fundamentals were implemented to develop the conceptual model presented in this study. An overview diagram of how our model was developed is presented in Figure 3. Operating conditions, such as CD and HRT, were the inputs, with outputs being the microbial communities that developed under different parameter configurations and the pollutant removal efficiencies.

Biological wastewater treatment models are founded on the number of functional microbial groups with associated kinetics, and the concentrations of electron donors (such as COD) and electron acceptors (such as oxygen, nitrate, nitrite) (Henze 2008; Nielsen et al. 2010). This study is the first to present a model developed for EBRs, as previous conceptual models and designs focused on either the reactor performance or the microbial communities, with limited integration of the two. The conceptual model proposed is anticipated to provide an overview of the functional microbial communities in wastewater EBRs by including:

• (a)

  Microbial communities in CBRs and EBRs operated at different CDs and HRTs with efficient COD, P and N removal.

• (b)

  Functional microbial communities enriched in CBRs and EBRs.

The conceptual model included all functional microbial communities enriched and depleted in the EBRs at CDs of 3 A/m²-EBRshort (3-EBRshort), 5 A/m²-EBRshort (5-EBRshort) and 7 A/m²-EBRshort (7-EBRshort) and a short HRT at a CD of 3 A/m² (3-EBRshort) and a long HRT at a CD of 3 A/m² (3-EBRlong) in relation to nutrient removal is presented in Figure 4(a). The functional genera are listed in Table 1. Various concentrations of contaminants in the mixed liquor suspended solids (MLSS) and synthetic wastewater components were presented as input variables (Giwa et al. 2015). The mechanisms involved simultaneously are microbe-mediated biodegradation of organic pollutants and electrochemical coagulation via the application of an electric field (Giwa & Hasan 2015a, 2015b). In WWTPs, quantitative changes between autotrophic and heterotrophic microbial communities are affected by wastewater characteristics, and the type and operation of the technological system (Hu et al. 2012; Ju & Zhang 2014; Muszyński et al. 2015).

The observations of bacterial groups enriched or depleted in EBRs associated with sCOD, N and P removal are shown in Figure 4(a). A summary of the model is provided below:

• In wastewater EBRs, sCOD removal was high (95–98%) and improved as the CD increased. Functional bacterial species Neisseriaceae (Treimo et al. 2006; Porwal et al. 2015) associated with COD removal were enriched in both 3-EBR and 7-EBR. Pseudomonas species were depleted in 7-EBR and enriched in 3-EBRlong. Flavobacterium, Lactococcus and Vogesella (Nielsen et al. 2012) were enriched in 3-EBRshort.

• Efficient N removal (99%) was directly correlated with the abundance of Nitrospira, Xanthomonadaceae, Janthinobacterium and Nitrosomonadaceae (Weissbrodt et al. 2014; Wells et al. 2014). Populations of Nitrospira, Xanthomonadaceae and Nitrosomonadaceae were enriched in 7-EBR and 5-EBR, which might explain the high removal efficiency in both of these EBRs compared to 3-EBR. Nitrospira and Janthinobacterium sp. were enriched in 3-EBRshort specifically operated at an HRT of 10 hours. In general, bacterial species associated with N removal were enriched in EBRs with higher CDs and shorter HRTs.

• P removal (97%) was directly correlated with the abundance of Rhodocyclaceae, Dechloromonas, Rhodobacteraceae and Rhodobacter (Bond et al. 1995) bacterial communities, which were enriched in 3-EBR and 7-EBR compared to 5-EBR. Dechloromonas species (Ahn et al. 2007; Zhang et al. 2017) were enriched in 3-EBR, 7-EBR and 3-EBRlong operated at HRT of 75 hours. This suggests that Dechloromonas species favored a wide range of operating conditions of EBRs, which correlates to the high removal efficiency of P in EBRs.

• Sulfur reducing bacteria, Thiobacillus and Thiobacteriales (Madigan et al. 1997), were enriched in all EBRs, but were present at low levels (low OTU counts) (ElNaker et al. 2019). Thiobacillus species were enriched in 3-EBR and 7-EBR, while Thiobacteriales favored 7-EBR only. Both species favored 3-EBRshort operated at an HRT of 6 hours. The low abundance of these bacteria could be due to the aerobic conditions maintained in these reactors as they have been reported to be physiologically active during anaerobic periods or in sludge storage tanks (Nielsen et al. 2010).

• Methanotrophs (methane oxidizing bacteria) Methylococcaceae, Hyphomicrobium and Methylocaldum (Bowman 2006) were detected in our EBRs with low average relative abundance (0.26%) across EBRs in comparison to other bacterial functional groups. Methylococcaceae and Methylocaldum were enriched in 5-EBR and 7-EBR while Hyphomicrobium favored 3-EBRshort operated at an HRT of 6 hours and 3-EBRlong operated at an HRT of 50 and 75 hours.

• Antibiotic resistance associated bacterial species Enterobacter (Lee et al. 2014) were depleted in all EBRs except for 3-EBRlong. Beneficial bacterial species Bdellovibrio (El-Shanshoury et al. 2016) were enriched in 3-EBR at all short HRTs. It is possible that enrichment of Bdellovibrio, which are predators of Gram-negative microbes, might explain the decrease in Enterobacter species, but that cannot be determined conclusively from the data collected. The data do suggest that using EBR technology over conventional MBR technology could potentially lead to a decrease in the spread of wastewater mediated antibiotic resistance. Future experiments should be developed to confirm this possibility.

• Hydrolyzers such as Saprospiraceae, Sphingobacteriales and Cytophagaceae (Bond et al. 1995; Kong et al. 2007; Xia et al. 2008; Zhang et al. 2018) species were enriched in 3-EBR. Sphingobacteriales favored 3-EBR and was depleted in 5-EBR and 7-EBR. Hydrolyzers were depleted in 5-EBR and 3-EBRlong. Cytophagaceae species favored both 3-EBR and 7-EBR.

From the model discussed in Figure 4(a), it could be noted that EBRs operated at CDs of 3, 5 and 7 A/m² have impacted the bacterial groups in various ways, which is a major finding for future EBR design. The conceptual modeling approach allowed a clear simplification of the interaction between the multiple processes that occur in EBRs, showing the impact of applying an electric field to microbial community structure, function and organic pollutant removal efficiency. This model can be updated as more data are collected, which in turn could be implemented to predict and improve microbial community structure and function, thus effectively removing organic pollutants and nutrients from wastewater.

A conceptual model of functional microbial communities detected in CBRs operated at CBRshort (HRTs of 6, 10 and 16 hours) and CBRlong (HRTs of 24, 50 and 75 hours) is shown in Figure 4(b). A summary of the model is provided below:

• In CBRs, sCOD removal did not exceed 92% across all bioreactors. Two bacterial species associated with COD removal, Pseudomonas and Flavobacterium, (Xu et al. 2018a, 2018b), were enriched in both CBRshort and CBRlong.

• Efficient N removal in CBRs was correlated to Nitrospira species, which were enriched in CBRshort at HRTs of 6 and 16 hours and CBRlong at all HRTs tested. Nitrosomonas species were depleted in all CBR operating conditions.

• P removal in CBRs was associated with enrichment of ‘Candidatus Accumulibacter’ (Stokholm-Bjerregaard et al. 2017) in CBRshort at an HRT of 6 hours and CBRlong while Dechloromonas species (Terashima et al. 2016) were depleted in all CBRs tested.

• Enterobacter species, which are the natural microbiota of gastrointestinal tracts and tend to be associated with the spread of antibiotic resistance, were enriched in all CBRs operated.

• Sulfur reducing bacteria Thiobacillus species (Ma et al. 2015) increased in abundance in CBRshort at an HRT of 10 hours while they decreased in the other CBRs tested.

• Methanotrophs (methane oxidizing bacteria) Methylococcaceae (Hatamoto et al. 2014) were inactive in all CBRs tested.

• Of the hydrolyzers, only Sphingobacteriales species were stimulated and enriched in CBRshort at HRTs of 6 and 10 hours and were depleted at CBRlong.

In general, it was observed that the absence of an electric field in CBRs as compared to EBRs resulted in the development of communities with lower relative abundances of bacterial species that have been shown in many previous studies to correlate well with removal of N and P containing nutrients. This observation also supported the hypothesis that the mechanism of action in EBRs includes the development of a superior water treatment microbial community in addition to electrochemical coagulation reactions that remove nutrients from wastewater, as thoroughly discussed in the next section.

Operating conditions such as CD, HRT, aeration intensity and nutrients play a vital role in the microbial composition of the functional groups in EBRs. These parameters and variables most likely work together to affect the microbial community structure through different mechanisms. First, the electric field directly affects bacterial community structure through impacting the growth rates or survival rates of individual bacterial species in different ways, changing the available enzymatic milieu that then acts to degrade various impurities in the wastewater. Second, the presence of the current might alter substrate uptake kinetics as well as cellular uptake rates (ElNaker et al. 2019), which could also be a contributing factor for the enrichment and depletion of different bacterial species. Third, the bacterial community itself undergoes changes that result from interspecies interactions facilitated by the shift in environment (caused by the electric field or enzymatic action). An example of this is the depletion of Enterobacter species, which, as mentioned earlier, could result from an enrichment of Bdellovibrio species. These interactions do not have to be predator/prey interactions and can instead be interactions where one species produces molecules or byproducts that are either toxic or required for the growth of another species in the community.

Finally, the electrocoagulation process, where organic compounds form coagulates that precipitate out of the wastewater, indirectly affect the community by changing the concentrations of available nutrients in the wastewater. In combination with the electrocoagulation process, there is a biodegradation process in which microorganisms stimulated by the current oxidize or sequester these organic compounds. Aerobic conditions favor the formation of reactive oxygen species that are detrimental to living cells, but it appears that aerobic biofilms possess an arsenal of enzymes to protect them against oxidative stress (Erable et al. 2012; Sun et al. 2019). Accordingly, a slight layer of microbial biofilm could precipitate on the aluminum anode, which could have enhanced phosphorus removal. However, the EBRs operated in aerobic conditions reported an increase in permeate flux, enhancement of microbial growth and activity, and reduction of filtration resistance and sludge extracellular polymeric substances (EPS) content (Ensano et al. 2016). The forces exerted by the electric field, microbial community and nature of the wastewater work together to convert the wastewater into higher quality effluent water (ElNaker et al. 2018a, 2019) An illustration of the possible mechanisms for the removal of N, P, and COD when integrating the electrochemical process with biological treatment is shown in Figure 5. Key points from the illustration are summarized below:

• N removal in CBRs and EBRs could be achieved via two main mechanisms (Grady et al. 2011). The first is nitrification, conducted in a two-step process in which ammonia (NH₄⁺) is oxidized into nitrite (NO₂⁻) and nitrite is oxidized into nitrate (NO₃⁻). This is achieved in CBRs by the nitrifiers; bacterial family Nitrospiraceae; Nitrospira genera (ammonia oxidizing bacteria) and Nitrosomonas (nitrite oxidizing bacteria) (Juretschko et al. 1998; Daims et al. 2015). The second is denitrification, which involves the biological oxidation of many organic substrates in wastewater treatment using nitrate or nitrite as an electron acceptor instead of oxygen. This could be carried out by microbes such as Dechloromonas and other species present in both CBRs and EBRs, which in turn result in effective nitrogen removal. In EBRs, Nitrospira and Dechloromonas species have been stimulated and increased in abundance in the presence of an electric field. It has been reported that the genome of the chemolithoautotrophic nitrifying bacteria encodes the pathways both for ammonia and nitrite oxidation, which are consequently activated during growth by ammonia oxidation to nitrate (Juretschko et al. 1998; Daims et al. 2015).

• P removal in CBRs is biologically accomplished by PAOs (Stokholm-Bjerregaard et al. 2017), which proliferate and store large quantities of inorganic phosphate as polyphosphate, thereby allowing phosphorus removal from wastewater via bioaccumulation. Multiple scenarios have been postulated to explain PAO metabolism, and they differ primarily in the source of the reducing power needed to form poly-β-hydroxyalkanoate (PHA) (Ahn et al. 2007). When the wastewater and biomass are under aerobic conditions, they have oxygen as an electron acceptor, so the PAOs carry out normal aerobic metabolism for growth by using the stored PHA as their carbon and energy source, generating ATP through electron transport phosphorylation.

• Among PAOs, members of Dechloromonas have also been reported as denitrifying PAO that can uptake acetate and act as denitrifiers. During denitrification, these species can take up phosphorus with nitrate acting as an electron acceptor. This genus likely contributes to denitrification and P removal in the reactors tested. In EBRs, P removal most likely occurs due to electrocoagulation in addition to electric current stimulation effects on bacterial communities associated with P removal.

• COD removal is complex and occurs via various processes: biodegradation and electrochemical oxidation and absorption. Biodegradation could be carried out by Lactococcus and Neisseria species that have been stimulated and increased in abundance in EBRs; they can acquire energy from the electrochemical oxidation that occurs due to the presence of an electric field in aerobic conditions. Those species have been reported to reduce and degrade COD in wastewaters typically discharged by the dairy industry (Porwal et al. 2015). Alongside the biodegradation process, when the electric field is applied, aluminum ions (Al³⁺) are released in situ from the anode into the reactor, causing coagulation and flocculation of the negatively charged colloidal particles (Drews 2010; Bouamra et al. 2012). Consequently, long chains of Al hydroxide complexes are formed as larger flocs precipitating contaminants such as organic matter and P.

Altering the operating conditions of a bioreactor will contribute to the evolution of different functional microbial communities, which in turn will aid in the removal of different nutrients from wastewater. This means that if engineers can predict how a parameter affects the microbial community, protocols with regards to operating conditions can be set up to match the characteristics of the wastewater being treated. Based on the conceptual models we have developed in this study, we have created a first draft of such a protocol, where recommendations are made for how to treat various types of wastewater, as illustrated in Figure 6. The suggested protocols are summarized below:

• Industrial wastewater, containing petroleum hydrocarbons, oil and hazardous materials are generally degraded by Pseudomonas, Acinetobacter and Alteromonas species (Yu et al. 2017; Xu et al. 2018a, 2018b; Chung et al. 2020). These species were enriched at CD3/HRT6, CD7/HRT6 as well as CD3/HRT10 (HRT in hours). Aromatic hydrocarbons could also be removed by Shingomonas species (Zylstra & Kim 1997), which were slightly enriched at CD7/HRT6.

• Household municipal wastewater containing sewage is degraded by Pseudomonas and Vogesella species (Yan et al. 2015). Pseudomonas species were enriched at CD3/HRT50 and CD3/HRT75, and Vogesella species were enriched at CD3/HRT6. In addition, Rhodoplanes species (Ma et al. 2015) were enriched at CD7/HRT6 and are reported to degrade xenobiotic substances (Zhu et al. 2019), which belong to a relatively new class of contaminants in municipal wastewater called emerging pollutants.

• Pollution that is caused by hazardous compounds such as bisphenol A and heavy metals, such as cadmium, could be treated by Sphigobium and Plesiomonas species (Sasaki et al. 2005; Zielińska et al. 2014), which increased in abundance at CD3/HRT16 and CD3/HRT6, respectively.

• Agricultural wastewater, which typically contains fertilizers and pesticides, could be treated by bacterial species such as Flavobacterium, Pseudomonas and Rhodococcus (Huang et al. 2018), which were enriched in EBRs operated at CD3 and CD5 at HRTs of 6, 10 and 50 hours.

• Wastewater generated by hospitals, the pharmaceutical and the personal care products (PPCP) industries contain heavy metals and toxic chemical compounds such as Cu, Fe, Cd, Pb, Hg, Ni, Pt, cyanide, phenol and others. This type of wastewater could be biodegraded by Aeromonas and Alteromonas species (Cruz et al. 2007; Selvin et al. 2009; Matyar et al. 2010; Lee et al. 2014; Mohamed 2016), which favored EBRs operated at CD3/HRT6.

• Wastewater generated by the dairy industry could be treated by Lactococcus and Neisseria species, which are reported to degrade sCOD and dairy products (Treimo et al. 2006; Dong et al. 2011; Porwal et al. 2015). Both species increased in abundance at CD3/HRT10 and CD7/HRT6.

Taken together, operating an electro-bioreactor with specific combinations of CD and HRT (and under fixed pH, solids retention time, EC and DO as indicated earlier) can boost the electrocoagulation and bioactive process leading to evolution of numerous functional bacterial communities and result in a more efficient WWTP.

In this study, a new conceptual framework model for EBRs based on the detailed structure and function of microbial communities of EBRs operated at different CDs (3, 5, and 7 A/m²) and HRTs (short: 6, 10 and 16 hours; and long: 24, 50 and 75 hours) was proposed. Specific bacterial groups were associated with efficient sCOD, N and P removal. The model is a first attempt at analyzing data from many previous studies in order to provide guidance for future studies of the structure and function of microbial communities exposed to electric fields. The model also allows for future experiments to test if altering the parameters of an electro-bioreactor can be tailored to the type of wastewater being treated. The model provides different mechanisms for the integration of electrocoagulation and biodegradation processes in efficient nutrient removal from wastewater. This conceptual model provides a framework for developing mathematical models in order to validate and predict how the treatment efficiency of EBRs can be improved by varying operating conditions. Finally, the model lays the foundation for the development of a more comprehensive quantitative model when more data is collected and published by different groups operating in this field.

This study has been supported by Khalifa University of Science and Technology in Abu Dhabi (UAE). Also, the support provided by the Center for Membranes and Advanced Water Technology (CMAT) (Award No. RC2-2018-009) at Khalifa University of Science and Technology is greatly appreciated.

All relevant data are included in the paper or its Supplementary Information.