Using proteomics for an insight into the performance of activated sludge in a lab-scale WWTP

Highlights

• The results of proteomic measurement of activated sludge using publicly available databases were described.

• The decrease in nitrificatian activity was related to the underexpression of the enzyme hydroxylamine oxidase (HAO).

• Identification using enhanced biological phosphorus removal (EBPR) database created bias towards EBPR microbial community.

• The majority of proteins identified were related to Candidatus Accumulibacter.

Abstract

The performance of a lab-scale wastewater treatment plant during the start-up phase was investigated. A period of varying pH resulted in the loss of ammonium removal efficiency together with a decrease in the specific autotrophic oxygen uptake rate (OUR) of the ammonium oxidizing bacteria (AOB). To understand how the AOB were affected at the molecular level, we used shotgun proteomics to quantify the relative protein levels in the system. Label-free quantification (LFQ) showed that during the time of poor ammonium removal, the marker enzyme hydroxylamine oxidase from Nitrosomonas sp. was massively reduced in intensity. The result of this study shows that quantitative proteomics has the potential to be used as a monitoring tool. Nevertheless, there are still some restrictions when measuring activated sludge using proteomic methods such as the availability of a suitable proteomic database. In this paper, we describe our experience of using publicly available databases for identification of activated sludge proteins.

Introduction

Activated sludge process is very sensitive to changes in influent quality and process conditions such as pH, temperature, oxygen level, and sludge loading (Gray, 2004). Nitrifiers, due to their low growth rate, are particularly easily affected by these changes (Bock and Wagner, 2013). The adverse effect can be seen directly in the loss of ammonia removal efficiency. Most of the time, there is little knowledge about the molecular mechanisms that lead to a decrease in ammonia degradation. As microbial protein expression is stimulated by their environment, proteomics has a great potential to be used as bioindicator of the present state of an activated sludge (Denecke, 2006). This calls for a more thorough investigation of the processes involved in activated sludge system using state-of-the-art molecular microbial tool of (meta)proteomics (Wilmes et al., 2008).

Metaproteomics refers to the characterisation of proteins in an environmental sample, a term first proposed by Wilmes and Bond (2004). This tool has since caught the interest of many researchers in the field of biological wastewater treatment such as Wilmes and Bond (2004) and Albertsen et al. (2013), who identified the main proteins in the activated sludge system of enhanced biological phosphorus removal (EBPR) system and investigated their functions in the purification process. Steiner et al. (2019) used proteomics methods to gather information on the malfunction of activated sludge treating landfill leachate. The main problem faced by researchers utilising proteomic method in environmental samples including activated sludge is the lack of availability of suitable database that represents the microbiology of their system (Wilmes et al., 2015). Creating a dedicated database is time consuming and expensive, which may not always be possible, particularly when measuring just a few samples. In such a case, for activated sludge sample, the choice is limited to utilising a database published from another plant, whose treatment process may not necessarily represent the sample to be analysed, or to handpick the sequences belonging to the microorganisms of interest from a public database like NCBI (Acland et al., 2014) or UniProtKB (The UniProt Consortium, 2017) and pool them together to create a new database for the identification, with the risk of missing the organisms or proteins which could be significant to their treatment system.

Shotgun proteomics is an indirect method of determining the presence of proteins by identifying peptides thereof after digestion with specific proteases e.g. trypsin (Zhang et al., 2013a). This method has been used in different areas of environmental research such as in the proteomic measurement of microbial communities in biogas plants (Heyer et al., 2013), in biofilm systems (Herschend et al., 2017) and for identifying proteins in the EPS of activated sludge (Zhang et al., 2015). Shotgun proteomics of environmental samples has been mostly qualitative in nature, although quantitative information of the key proteins is often important for an insight into the function of the microbial communities (Heyer et al., 2015). Quantitative proteomics can be done either as absolute quantification or relative quantification between samples. Absolute quantification requires protein or peptide sample to be spiked with a known quantity of standard, normally stable isotope-labelled peptides. Relative quantification is used to obtain information regarding changes in protein expression between two sample states. The proteins of interest need to be detected in both samples to allow the relative abundance of the proteins to be quantified (Silva et al., 2006). Label-free quantification (LFQ) is a relative quantification method based on either the comparison of the spectral count or peptide ion signal intensity (Bantscheff et al., 2012; Wang et al., 2008). The spectral count approach is based on the observation that the number of mass spectra detected correlates to the protein abundance. Comparing the number of spectra of a given peptide between samples allows the determination of its relative abundance. The relative quantification by ion signal intensity, on the other hand, is calculated by comparing the integrals of the peak area of the ion chromatograms from LC-MS/MS (Bantscheff et al., 2012). Both LFQ by spectral count and ion intensity are reported to have similar accuracy, proteome coverage and linear dynamic range (Bantscheff et al., 2007). LFQ has the advantage of relatively simple procedure in comparison to absolute quantification using isotope-labelled peptide. Furthermore, it allows whole proteome analysis, unlike isotope labelling which are limited to just a few selected peptides, as the quantification is not limited by costly and time-consuming labelling procedure. LFQ is also reported to provide higher dynamic quantification range compared to isotope labelling, although this method is less accurate as all the variations in the procedure prior to the measurement will be reflected in the mass spectral data (Bantscheff et al., 2012).

In this study, we aimed to look at the relevance and feasibility of proteomic measurement in wastewater treatment, particularly for the treatment plants without their own metagenomics/metaproteomic database. Furthermore, we investigated the use of quantitative proteomics of LFQ-MS to measure and monitor the proteins in activated sludge during the start-up phase of the lab-scale wastewater treatment plant. As the ammonium removal efficiency fluctuated during this period, the two enzymes of ammonium oxidizing bacteria (AOB) that catalyse ammonia oxidation process were of particular interest in this research: 1) ammonia monooxygenase (AMO), which oxidizes ammonia to the intermediate hydroxylamine (Eq. (1); Ensign et al., 1993) and 2) hydroxylamine oxidase (HAO), which catalyses the oxidation of hydroxylamine to nitric oxide (Eq. (2); Kuypers et al., 2018), before being further converted to nitrite by an as-yet-unidentified enzyme (Caranto and Lancaster, 2017). To the best of our knowledge, there is no previous research that has used LFQ-MS to monitor the long term changes in the expression of the main enzymes of AOB in the activated sludge process.

$$N{H}_3+O_2+2H^{+}+2e^{-}\stackrel{AMO}{\to }N{H}_{2}OH+H_{2}O$$

$$N{H}_{2}OH+H_{2}O\stackrel{HAO}{\to }NO+3H^{+}+3e^{-}$$