3.4. Effects of ENPs on nitrification during anaerobic ammonium oxidation
Anaerobic ammonium oxidation (anammox) process is a novel biological nitrogen removal technology that is gaining popularity for nitrogen removal in wastewater streams. In this process, ammonium is directly converted to dinitrogen gas using nitrite as the electron acceptor in the absence of oxygen (Eq. 5) for nitrogen removal from wastewater streams.
\begin{equation} \text{NH}_{4}^{+}+1.32\text{NO}_{2}^{-}+0.066\text{HCO}_{3}^{-}+0.13H^{+}\rightarrow 1.02N_{2}+0.26\text{NO}_{3}^{-}+0.066\text{CH}_{2}O_{5}N_{0.15}+2.03H_{2}O\ \ \ \ (Eq.5)\nonumber \\ \end{equation}
No addition of organic carbon source is required since CO2 is utilized as the only carbon source. In addition, it significantly reduces oxygen demand since ammonium is only required to be partially nitrified to NO2-instead of NO3-; thus leading to considerable saving in operational cost. Due to these and other advantages (low CO2 emission and low biomass yield), anammox processes have been widely regarded as an innovative and sustainable alternative to the classical activated sludge process (Zhang ZZ et al., 2016). There are more than one hundred full-scale anammox installations worldwide that are being applied for the treatment of side-stream wastewater (reject water).
A few studies have studied addressing the of ENPs effects on nitrogen removal by annamox processes (Table S4). Zhang et al (2017 ) studied the short-term (24 hrs) effects of CuNPs, CuONPs, ZnONPs and AgNPs in a batch study using anammox sludge. Their results showed that CuONPs, ZnONPs and AgNPs up to 50 mg/g suspended solid did not affect anammox activity, ROS generation or LDH release. By contrast, CuNPs at 1.25 and 2.5 mg/g SS resulted in severe inhibition of anammox activity, without inducing an increase in LDH release. Higher loads of CuNPs caused significant inhibition of anammox activity and increased LDH release. The toxicity was primarily attributed to dissolved Cu2+ ions. Another batch study (Zhang et al. 2017) demonstrated that the addition of EDTA or S2- could attenuate the adverse effects of CuNPs, presumably due to the chelation or sulfidation of Cu2+ ions. Later, the long-term effect of CuNPs was studied by adding CuNPs to an up-flow anaerobic sludge blanket (UFAB) reactor at 0.5 mg/L for 15 days, 1.0 mg/L for 15 days, and 5 mg/L for 30 days. Results showed that 5 mg/L of CuNP caused near complete inhibition of nitrogen removal and significant a decrease of the abundance of anammox bacteria. Withdrawing CuNPs from the influent permitted the recovery of nitrogen removal.
The long-term effects of ZnO NPs on annammox sludge was also studied using UFAB reactor (Zhang et al. 2018). ZnO NPs (~30 nm) were added to the bioreactor at 1.0 mg/L on day 31, increased to 5.0 mg/L on day 46 and 10 mg/L on day 61. Results showed that shock-load of 10 mg/L ZnONPs resulted in the deprivation of 90% of the nitrogen removal capacity within 3 days. Anammox activity was significantly inhibited without any significant increase in LDH release or intracellular ROS production. These effects were attributed to dissolved Zn2+ ions and complete recovery was observed within 40 days after withdrawing the NPs from the influent. Another study by the same group investigated the effects of other metal oxides NPs on the performance of anammox process (Zhang et al. 2018). SiO2 NPs (~30 nm), TiO2 NPs (~60 nm, hydrophilic), CeO2 NPs and Al2O3 NPs (30 nm, hydrophilic) on granular anammox sludge in lab-scale UFAB bioreactors. NPs were added to the bioreactors at 1, 50 and 200 mg/L in a step-wise fashion with a 30-day interval and lasted for an entire duration of 90 days. No adverse effects on nitrogen removal were observed, and this resilience was attributed to adaptation of the microorganisms through community shift and enhanced EPS production. Most recently, Li et al (Li et al. 2019 ) reported that exposure to graphene oxide (1 and 10 mg/L) resulted in acute toxicity and inhibition of annamox nitrogen removal. The effects disappeared by day 19 and reversed by the end of the study at day 61, with a TN removal efficiency higher than control. The same doses of AgNP caused long-term inhibition on TN removal, which did not recover. The long-term enhancement of TN removal by GO was accompanied by the relative high abundance of anammox bacteria C.Anammoxoglobus ; while the TN removal inhibition by AgNP was accompanied by the disappearance of some species with anammox ability. This observation seems to contradict with the findings by Zhang et al (2018), in which they reported no long-term adverse effects of AgNPs on anammox activity at concentrations of 1, 10 and 50 mg/L. This discrepancy may be related to differences in the type of sludge, bioreactor and particles used in these studies.
The effects of iron NPs seemed to be beneficial to anammox nitrogen removal. Li et al (2018) reported that adding Fe3O4 NPs (1, 10 mg/L) to an unplanted anammox subsurface flow constructed wetlands produced concentration-dependent acute toxic effects on ammonia removal; these effects disappeared overtime and by day 61, nitrogen removal rate were actually enhanced. Nano scale zero valent iron (nZVI) have also been proved to be beneficial for anammox bacteria growth and nitrogen removal (Erdim et al. 2019). In summary, these early studies have shown that in an anammox process, ENP toxicity was mainly caused by dissolved ions; the role of ROS generation was less significant than in the conventional activated sludge process, likely due to the lack of oxygen supply.
4. Mechanisms of ENP toxicity
Several mechanisms for ENP toxicity have been proposed based on experimental observations (Figure 1). Metal based ENPs are believed to exert their toxicity mainly through dissolved ions, in combination with the effects from nanoparticles. Metal ions bind with the negatively charged compounds in the bacteria cell wall, resulting in cell wall destabilization or collapse. Metal ions have high affinity to molecules containing –SH groups, such as cysteine; this binding can break S-S bond bridges that are necessary to maintain the integrity of folded proteins or directly disrupt the function of certain enzymes (Slavin et al. 2017). For example, the activity of most AMOs in N. Europaeaare inhibited by Co2+, Zn2+, Ni2+, and Fe2+ in a concentration-dependent manner (Ensign et al. 1993). The dissolution of metal-based NPs also generates reactive oxygen species, which could cause cell membrane damage via oxidation of membrane components such as lipids. The internalized metal ion can also react with mitochondrial H2O2 and produce intracellular ROS or affect DNA repair and cause mutation (Huangfu et al., 2019). Intracellularly-produced ROS may also damage the cell membrane via lipid oxidation or damage cellular DNA without visible cell membrane damage. Non-metal based ENPs such as carbon nanomaterials also produce ROS upon light illumination, a property that is shared by TiO2 and ZnO NPs.
ENPs may enter cells through direct penetration or endocytosis. Direct penetration is caused by non-specific binding forces (electrostatic, hydrophobic, van der Waals) between the particle and the cell membrane; while endocytosis involves specific receptor-ligand interactions. Once inside the cell, ENPs can bind with intracellular biomolecules, interact with mitochondria, induce ROS production, or damage cellular functions (Yang et al, 2019). Increased ROS production leads to enzyme inactivation and DNA damage, likely the reason for the observed inhibition of key enzymes involved with nitrification.
ENPs can cause physical damage to bacteria in various ways. Adsorption of ENPs onto cell wall/membrane leads to depolarization of the cell wall/membrane, which changes the negativity of the membrane and makes it more permeable. Carbon nanotubes can puncture the cell membrane like needles. Graphene nanosheets can both cut through cell membrane and also disruptively extract phospholipids from the membrane due to strong van der Waals interactions and hydrophobic effects. These types of physical damage disrupt and weaken the cell membrane, resulting in release of biomolecules such as LDH. NP aggregation onto the cell surface may also facilitate NP dissolution, releasing metal ions that can easily enter the cell (Slavin et al, 2017). Large graphene nanosheets and aggregates of smaller NPs can entrap bacteria and prevent them from taking up nutrients. Long CNTs can wrap around the bacteria and induce osmotic cell lysis.
EPS generally acts as a protective layer for the microorganisms by absorbing ENPs and the dissolved ions. On the other hand, the EPS may promote ENP dissolution after the absorption capacity has been reached. Under some circumstances, strong interactions between ENPs and the EPS may result in stripping of the protective EPS layer off sludge microorganisms, thus making the microorganisms more vulnerable. As mentioned earlier, the ENP-microorganism interactions also depend heavily on the properties of the latter including type (gram-negative vs. gram-positive) (Mocan et al, 2017), shape (rod-shaped vs. spherical) (Al-Jumaili et al, 2017), hardness (Liu et al, 2009), cell wall structure (lipopolysaccharides, phospholipids defects) (Hsu et al, 2016), and enzyme and metabolism activities (Krishnamoorthy et al. 2012). Therefore, it is expected that different microorganisms in the activated sludge will respond differently to the same ENP stress, resulting in microbial community shifts as observed by several studies cited in this review. These shifts however, may not always cause inhibition.
5. Issues and challenges
There are a number of issues and challenges associated with assessing the effects engineered nanomaterials on nitrification in activated sludge.
First, even though NP physical-chemical properties are critical factors causing microbial toxicity, current literature is frequently missing critical data regarding NP physical-chemical properties, including hydrodynamic size, shape, surface charge, hydrophobicity, surface roughness, deformability (soft vs hard NPs), surface chemistry, electronic structure and coating. The most frequently provided data is particle size, many of them determined with TEM, but the hydrodynamic size is more appropriate for NPs in an aqueous environment. Some papers reported the size information provided by the manufacturer, which is not always correct and requires verification. A limited number of papers provided zeta potential measurements, even fewer have included hydrophobicity, surface roughness, deformability (soft vs hard NPs), surface chemistry, or coating (Tables S1-S3). Lack of comprehensive physical-chemical characterization makes it very difficult to draw meaningful comparisons between studies, because NP toxicity and interactions with the microorganisms and the EPS are governed by these properties (Huangfu et al., 2019, Slavin et al., 2017).
Second, it is also necessary to understand fate and transformation of these particles post entry into the wastewater. Majority of these particles, especially those which are not stabilized with coatings, will undergo changes and take on a new physio-chemical identity. These changes could include biodegradation, dissolution, precipitation, aggregation, adsorption of naturally occurring substances and chemical transformation, depending on both particle-specific properties & particle state (free or matrix incorporated), and on the chemistry of the surrounding solution (pH, ionic strength, ionic composition) (Petosa et al., 2010). As a result, NPs identity in the wastewater could be vastly different from the original NPs. Clar et al (Clar et al., 2016) showed that aggregated CuO NPs in wastewater were about 1600 nm in diameter, about 40 times larger than the original NP (46 nm). Since it is these transformed particles that interact directly with the microorganisms, a thorough characterization of these transformed NPs is vital.
NP transformations are particularly relevant in the pipes that carry sewage to wastewater treatment plants. This underground network is anaerobic and contains soluble and insoluble constituents that may react with a wide range of ENPs (Metcalf and Eddy, 2014). The principle compound of interest is sulfide, which is biologically produced by a consortia of sulfate-reducing microorganisms. Sulfide forms complexes with Ag NPs (Kaegi et al., 2013), Cu NPs (Hatamie et al., 2014), and ZnO NPs (Lupitskyy et al., 2018). There are also coarse particles and colloids that can heteroaggregate with ENPs during transit in the sewer (Zhang et al., 2016). These processes tend to reduce the bioavailability and ecotoxicity of ENPs, however, the extent of NP transformation may be limited in sewer systems with short residence times.
There are other important, unresolved issues. The synergetic effects of multiple species of NPs are particularly relevant for WWTPs due to the presence of a vast variety of NPs found in sewage. The purity of the nanomaterials and variations among different preparations must also be considered; this is important because the toxicity may be affected by the impurities during the manufacturing process. For instance, CuO NPs that contains Cu NPs contaminants are expected to have a toxicity profile different from that of pure CuO NPs, since the release of dissolved Cu2+ions from CuO NPs is significantly slower than from CuNPs (Zhang et al., 2017 ). The majority of literature so far has used unfunctionalized or minimally functionalized ENPs in their studies, while in reality, many applications use functionalized ENPs because of the improved efficacy, usability or added functionality. Functionalized ENPs will have different surface structure, chemistry, and aggregation properties; all of which will result in a completely different toxicity profiles.
Lastly, there is considerable uncertainty about the values of ENP concentrations that are present in wastewater. There are a small number of published studies that present such data and the range and temporal variability of ENP concentrations in domestic wastewater are not yet understood. There is a major information gap because it is difficult to relate published findings to realistic operational scenarios. There is a need to determine environmentally relevant concentrations of ENPs using field sampling campaigns. Such work should be done with a well-organized series of grab samples, taken together with flow and water quality data.
6. Conclusions
There is an urgent need to understand the effects of ENPs on wastewater treatment because of the growing use of nanomaterials. Nitrification is a critical wastewater treatment process that may be disrupted under certain conditions. Studies have confirmed that short-term, environmentally-relevant concentrations generally did not inhibit nitrification in conventional activated sludge systems. Long-term exposure to relatively high concentrations of ENPs inhibited nitrogen removal and shifted the microbiological community structure in activated sludge. Some studies have shown resiliency of activated sludge systems. Physical & chemical properties of the ENP, properties of the microorganisms and their environment are all believed to contribute to the variabilities in toxicity results observed in literature. Several mechanisms may contribute to the ENP-induced toxicity, including physical disruption of the cell membrane, generation of ROS, inhibition of enzymes and metabolic processes, and intracellular accumulation of ENPs. The effects of non-metal and composites ENPs have not been well-studied and need to be thoroughly investigated in future studies; as these materials are gaining increasing popularity in real applications. Aggregation and transformation of ENPs in wastewater are common and thus the observed toxic effects are in fact caused by aggregates or transformed NPs, thorough characterization of these “transformed NPs” will help to better interpret the results and explain the variabilities among different studies. Early studies on the emerging anammox technology provided evidences of ENP-induced nitrification inhibition in these processes, however, the mechanisms are expected to different from that in activated sludge due to the lack of oxygen and differences in the nitrification microorganisms. Future research should also include the even more recent development of biological nitrogen removal processes that combine partial nitrification with anammox.