Environ. Sci. Technol. 2007, 41, 3945-3951
Metabolic Footprinting: A New Approach to Identify Physiological Changes in Complex Microbial Communities upon Exposure to Toxic Chemicals I N Eˆ S D . S . H E N R I Q U E S , † D I A N A S . A G A , ‡ PEDRO MENDES,§ SEAMUS K. O’CONNOR,‡ AND N A N C Y G . L O V E , †,|,* Department of Civil and Environmental Engineering, Virginia Tech, 418 Durham Hall, Blacksburg, Virginia 24061, Department of Chemistry, The State University of New York at Buffalo, 611 Natural Sciences Complex, Buffalo, New York 14260, Virginia Bioinformatics Institute, Virginia Tech, Bioinformatics Facility and Department of Biological Sciences, Virginia Tech, 2125 Derring Hall, Blacksburg, Virginia 20641
Metabolic footprinting coupled with statistical analysis was applied to multiple, chemically stressed activated sludge cultures to identify probable biomarkers that indicate community stress. The impact of cadmium (Cd), 2,4dinitrophenol (DNP), and N-ethyl-maleimide (NEM) shock loads on the composition of the soluble fraction of activated sludge cultures was analyzed by gross biomolecular analyses and liquid chromatography-mass spectrometry (LCMS). Fresh mixed liquor from four distinct treatment plants was each divided in four different batches and was subjected to no chemical addition (control) and spike additions of the stressors Cd, DNP, or NEM. The results indicate that chemical stress caused a significant release of proteins, carbohydrates, and humic acids from the floc structure into the bulk liquid. Using discriminant function analysis (DFA) with genetic algorithm variable selection (GADFA), the samples subjected to the different stress conditions plus control could be differentiated, thereby indicating that the footprints of the soluble phase generated by LC-MS were different for the four conditions tested and, therefore, were toxin-specific but communityindependent. These footprints, thus, contain information about specific biomolecular differences between the stressed samples, and we found that only a limited number of m/z (mass to charge) ratios from the mass spectra were needed to differentiate between the control and each stressed sample. Since the experiments were conducted with mixed liquor from four distinct wastewater treatment plants, the discriminant m/z ratios may potentially be used as universal stress biomarkers in activated sludge systems.
Introduction Biological wastewater treatment systems rely on a complex microbial consortium structurally organized in biological * Corresponding author phone: (540)231-3980; fax: (540)231-7916; e-mail:
[email protected]. † Department of Civil and Environmental Engineering, Virginia Tech. ‡ The State University of New York at Buffalo. § Virginia Bioinformatics Institute, Virginia Tech. | Department of Biological Sciences, Virginia Tech. 10.1021/es062796t CCC: $37.00 Published on Web 05/05/2007
2007 American Chemical Society
flocs to carry out the degradation of soluble and insoluble components present in raw sewage. These systems are susceptible to toxic shock loads of industrial chemicals, which can impact the treatment efficiency severely (1). When challenged with a toxin, activated sludge communities may undergo different physiological and structural modifications that can result in macroscopic effects such as: floc disintegration or deflocculation (2-4); increased effluent soluble chemical oxygen demand (COD) (5); and nitrification inhibition (6). Experiments conducted in our laboratory showed that exposure of mixed liquor to different chemical toxins resulted in an increase of the soluble concentration of COD in the effluent of toxin-exposed reactors (5, 7). Although we were not certain about the origin of the increased soluble COD, we hypothesized that it partially resulted from the release of biofloc-associated compounds because of specific mechanisms occurring upon exposure of the bioflocs to the chemical toxins. Furthermore, we hypothesized that the materials released from activated sludge biological flocs during chemical stress events were toxin-specific, as the mechanisms responsible for the release effect were likely dependent on (1) the type of physical/chemical interactions established between the toxin and the floc extracellular polymeric substances (EPS) in which the bacteria are embedded and (2) the mode of action of the toxin and stress responses elicited by the bacterial populations within the flocs. The analysis of the soluble materials in the supernatant of chemically stressed mixed liquor could, therefore, reveal important physiological and structural changes induced by the stress condition. Metabolic footprinting consists of analyzing and comparing the “exome”, “exometabolome”, or extracellular matrix (spent culture medium in the case of bacterial cultures) produced under different conditions (8-10). It has been used to differentiate between different physiological states of wildtype yeast and different yeast single-gene deletion mutants (11), to discriminate between different Escherichia coli tryptophan mutants (12), and to distinguish between modes of action of several antifungal substances on yeast cells (13). Similarly, we believe that the application of metabolic footprinting to compare changes in soluble extracellular metabolite patterns in response to toxin exposure can be a powerful technique to detect physiological changes in activated sludge populations that correlate with specific modes of process deterioration. Furthermore, using an aggregate footprinting method allows one to use the method on indigenous cultures and does not rely upon knowing specifics about the composition of the community. Metabolic footprinting has not yet been applied to compare complex environmentally relevant microbial communities exposed to different environmental conditions. One study of activated sludge EPS used size exclusion chromatography and Fourier transform infrared spectroscopy (FTIR) to show that EPS fingerprints changed dramatically when upset events causing deflocculation of the biomass took place, indicating that alterations at the level of EPS composition occur in response to stress events in activated sludge and that those alterations may be identified through fingerprints generated through a spectroscopic technique (14). However, the soluble phase of the mixed liquor (supernatant) was not analyzed during the study. The objective of this study was to generate metabolic footprints of activated sludge mixed liquor subjected to cadmium (Cd), N-ethyl-maleimide (NEM), and 2,4-dinitrophenol (DNP) shock loads and to compare them to control conditions using liquid chromatography-mass spectrometry VOL. 41, NO. 11, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3945
(LC-MS) techniques. Cd and NEM are electrophilic (thiolreactive) chemicals, which have been shown to cause strong deflocculation events of activated sludge (4, 7), while DNP is an uncoupler of oxidative phosphorylation that we found to produce moderate to low increases in effluent total suspended solids (TSS) and soluble COD (7) in the effluent of sequencing batch reactors (SBRs). Both Cd and DNP are industrially relevant contaminants, commonly found in many industrial streams. For example, DNP has been detected in wastewaters from nitrobenzene-, explosives-, and dyemanufacturing plants (15), while cadmium release into the aquatic environment results mainly from discharges from mining operations, electroplating, and phosphate fertilizersmanufacturing facilities (16). In this study, we identify m/z (mass to charge) ratios for soluble materials that are released from activated sludge flocs during an upset event and determine those m/z ratios that occur most often in similarly shocked mixed liquors from four wastewater treatment plants. These m/z ratios and their associated compounds may function as toxin-specific biomarkers of wastewater treatment upset. Ultimately, these biomarkers may be used to develop methods that allow fast and reliable identification of toxins that cause upset events and early warning tools that can minimize or eliminate the impact of toxins in biological treatment systems.
Experimental Procedures Batch Experiments with Activated Sludge Mixed Liquor. Mixed liquor from four full-scale activated sludge facilities in Virginia was obtained fresh and was immediately transported to the laboratory. To stabilize the growth state of the biomass, the mixed liquor was aerated and fed synthetic wastewater composed of acetate, glycerol, nitrogen (as NH4Cl), and phosphorus (as Na2HPO4) for a period of at least 3 h before the start of an experiment. The synthetic wastewater simulated the organic load of the primary effluent of the plant from which the mixed liquor had been retrieved. After the stabilization period, four 1.5 L aliquots of the source mixed liquor were measured into separate containers. These four reactors were aerated and fed independently using the same synthetic wastewater composition and rate applied during the stabilization period. At time zero, three of the reactors were spiked with preselected amounts of Cd (100 mg/L Cd added as CdCl2), DNP (20 mg/L), and NEM (15 mg/L). The fourth reactor was a control to which no toxin was added. Sampling from all the reactors occurred within 5 min of toxin addition (
