You are on page 1of 12

Journal of Applied Microbiology ISSN 1364-5072

ORIGINAL ARTICLE

Comparative evaluation of the microbial community in biological processes treating industrial and domestic wastewaters
A.P. Degenaar, A. Ismail and F. Bux
Centre for Water and Wastewater Technology, Department of Biotechnology, Durban University of Technology, Durban, South Africa

Keywords alpha-proteobacteria, beta-proteobacteria, FISH, vegetable oil efuent treatment. Correspondence F. Bux, Centre for Water and Wastewater Technology, Department of Biotechnology, Durban University of Technology, PO Box 1334, Durban 4000, South Africa. E-mail: faizalb@dut.ac.za

Abstract Aims: Comparison of the microbial composition and process performance between laboratory scale processes treating domestic and vegetable oil wastewaters. Methods and Results: Two laboratory scale modied LudzackEttinger processes were operated under similar operating conditions. One process was fed domestic wastewater and the other an industrial wastewater, vegetable oil efuent. Nitrogen removal capacities of the processes were similar. The industrial process exhibited a lower COD removal capacity and oxygen utilization rate, although a greater mixed liquor volatile suspended solids concentration was observed in the industrial process. Fluorescent in situ hybridization (FISH) with probes EUBmix, ALF1b, BET42a, GAM42a and HGC69a revealed that 81% and 72% of total cells stained with 4, 6-diamidino-2-phenylindole (DAPI) within the domestic and industrial processes respectively bound to EUBmix. This indicated a slightly lower Eubacterial population within the industrial process. The alpha-proteobacteria was the dominant community in the industrial process (31% of EUBmix), while the beta-proteobacteria dominated the domestic process (33% of EUBmix). Conclusions: The ndings served to establish a difference in the microbial population between the processes. Therefore, the class alpha-proteobacteria could play a primary role in the degradation of vegetable oil efuent. Signicance and Impact of the Study: This research will aid in process design and retrotting of biological processes treating vegetable oil efuent.

2007 0603: received 16 April 2007, revised 23 July 2007 and accepted 23 July 2007
doi:10.1111/j.1365-2672.2007.03563.x

Introduction Vegetable oil rening industries within Southern Africa, consume almost two million cubic metres of water annually. Approximately 40% of this potentially potable water is discharged into sewers as efuent (Steffen et al. 1989). Because of large volumes of this efuent being released into sewer systems, treatment to an acceptable standard is required prior to discharge (Horan 1990). Discharge of poor quality nal efuents impacts negatively on natural water sources resulting in eutrophication of natural ecosystems. Vegetable oil efuent (VOE) has been found to contain relatively high concentrations

of fats, oils and greases (FOG), chemical oxygen demand (COD), phosphorus, sodium, sulfate and a variety of other pollutants. Untreated VOE is known for creating shock-loading problems for the receiving wastewater (WW) treatment installations, resulting in poor quality nal efuents being produced, which do lu et al. not satisfy municipal discharge standards (Erog 1990). There are two methods of efuent treatment commonly employed by vegetable oil reneries in South Africa; physical separation of oil and grease via dissolved air oatation and pH correction (Lilley et al. 1997). Previous studies have shown that efuents from food
353

2007 The Authors Journal compilation 2007 The Society for Applied Microbiology, Journal of Applied Microbiology 104 (2008) 353363

Evaluation of the microbial community in wastewater

A.P. Degenaar et al.

industries containing fatty materials are readily biodegradable and are therefore amendable to biological lu et al. 1990). Conventional treatment methods (Erog treatment of edible oil efuent primarily involves physico-chemical processes (Bettazzi et al. 2007). However, the application of biological treatment is gaining much attention, with focus on the development of anaerobic treatment processes (Beccari et al. 2001). It is advantageous for process engineers and scientists to characterize and quantify the active biomass in biological processes, to promote the growth of desired organisms or manage problematic organisms (Keith et al. 2005). Aerobic treatment as an alternative has not been fully investigated, therefore a lack of knowledge of the microbial communities responsible for VOE degradation exists. In this investigation, aerobic treatment using activated sludge was chosen as an alternative method for the treatment of VOE. The effect of VOE on measured process parameters was also determined by comparison of the process performance with a modied LudzackEttinger (MLE) process treating domestic WW. Mixed liquor volatile suspended solids (MLVSS) analysis is the conventional means of measuring the active biomass concentration in activated sludge mixed liquor, according to the engineering paradigm. However, it is an indirect method and provides a lumped indication of the active biomass present and represents not only the active biomass but endogenous residue (dead cellular material) and inert particulate COD. These endogenous residues and inert particulates become entrapped in activated sludge ocs and accumulate with increasing sludge age and contribute to the overall MLVSS concentration (Wentzel et al. 1995). For these reasons, MLVSS analysis is not sensitive to changes in activity of the biomass, but is more suitable in providing an indication of the amount of active biomass present in a process. The oxygen utilization rate (OUR) is a more direct measurement which reects the rate at which microorganisms utilize oxygen (Lilley et al. 1997). Although it is not common practice to characterize the transformation kinetics of lipids in activated sludge using OUR (Dueholm et al. 2001), it serves as a good measure of the metabolic activity and health of the activated sludge process. The use of molecular methods, specically hybridization with rRNA targeted oligonucleotide probes, provides novel insights with respect to the structure and dynamics of microbial communities in activated sludge (Daims et al. 2001). According to Activated sludge Model no.2 (Henze et al. 1995) heterotrophic organisms comprise several groups viz.; the ordinary heterotrophs, which grow aerobically and are responsible for COD removal, denitrifying organisms growing anoxically, and the fer354

menters, which grow anaerobically. Previously, culture dependant techniques such as most probable number (MPN) method and heterotrophic plate counts on enrichment media, have been used to characterize and enumerate these communities in activated sludge. However, only 15% of the indigenous bacteria in activated mpfer sludge could be cultivated (Wagner et al. 1993; Ka et al. 1996). These limitations have lead to techniques using the 16S rRNA approach. In particular, uorescent in situ hybridization (FISH; Amann et al. 1995), polymerase chain reaction (PCR) and denaturing gradient gel electrophoresis (DGGE) (Muyzer et al. 1993) have been used extensively to conduct microbial community analysis. The comparative analysis of rRNA molecules has revolutionized our view of microbial taxonomy and evolution (Woese 1987). Ribosomal RNA sequences are perfect targets for uorescently labelled oligonucleotide probes, because they are highly conserved and naturally amplied, and can therefore be used in determinative studies in microbiology (Amann et al. 1990). By using selected regions within larger rRNA molecules (16S and 23S rRNA) as hybridization targets for synthetic oligonucleotides, probe specicity to individual phyla or species, can be freely adjusted. In addition, DeLong et al. (1989) showed that probe binding varied with ribosomal content and reected cell growth rate, viz., metabolically active cells will produce intensied uorescence, because of their increased rRNA content. The application of FISH for microbial community analysis of activated sludge processes could be considered a novel approach with a comparatively higher degree of success. Dual staining of samples with probe EUB338 and 4,6-diamidino-2-phenylindole (DAPI; Hicks et al. 1992) gives not only an indication of the metabolic activity of bacteria, but also that cells had sufcient rRNA for detection, were permeabilized for probes by standard xation procedures. Therefore, a high EUB : DAPI ratio in activated sludge would indicate a highly metabolically active bacterial population. In probing COD removing activated sludges from various municipal plants with oligonucleotide probes specic for proteobacteria, Wagner et al. (1993) demonstrated the dominance of proteobacteria, which together comprised 6075% microbial cells stained with DAPI. Wagner and Amann (1997) reported members of the beta-proteobacteria as playing a major role in the microbial consortia of activated sludge plants and alpha- and gamma-proteobacterial classes being less abundant. In this study, the microbial composition of two laboratory scale processes, treating domestic WW and VOE were characterized and compared using FISH, to identify the bacterial communities implicated in the biological treatment of VOE.

2007 The Authors Journal compilation 2007 The Society for Applied Microbiology, Journal of Applied Microbiology 104 (2008) 353363

A.P. Degenaar et al.

Evaluation of the microbial community in wastewater

Materials and methods Conguration and operation of laboratory scale processes The domestic and industrial laboratory scale biological treatment processes were modelled upon the MLE process and focussed on carbon removal (Lilley et al. 1997). The units were designed and manufactured by the Department of Civil Engineering, University of Cape Town, South Africa. MLE process treating domestic WW Aerobic mixed liquor and WW used in the domestic process was obtained from Southern WW Works (Durban, South Africa). The WW (primarily domestic) served as inuent for the process. The WW was collected in 25 l plastic drums, transported to the laboratory and stored at 4C in a cold room. The COD concentration of the WW was adjusted to 500 mg l)1 using tap water and the pH maintained at 75, by alkalinity adjustment using CaCO3. The reactor conguration of the domestic process consisted of the following: an anoxic reactor (6 l), an aerobic reactor (9 l) and a clarier (15 l) positioned at a 60 angle to the horizontal. An aerobic-recycle between the anoxic and aerobic zones was setup at a 2 : 1 ratio, with respect to the inuent ow rate. A sludge-recycle was setup between the clarier and the anoxic zone at a ratio of 1 : 1 with respect to the inuent ow rate. The inuent ow rate was set at 24 l day)1 and a sludge age of 10 days was maintained by wasting 15 l day)1 of mixed liquor from the aerobic reactor. The process was operated at room temperature (20C) and the OUR of the mixed liquor in the aerobic reactor was measured using an automated technique (Randall et al. 1991) with the lower and upper dissolved oxygen limits set at 20 and 50 mg O l)1 respectively. MLE process treating industrial WW The aerobic mixed liquor used in the industrial process was obtained from Darvill WW Treatment Purication Works (Pietermaritzburg, South Africa). VOE used as the inuent for the industrial process, was collected in 25 l drums from the drain at the end of the renery process of a local edible oil renery, situated in the direct vicinity of Darvill WW Treatment Purication Works. The 25 l drums of efuent were transported to the laboratory and stored at 4C in a cold room. A two-stage approach was adopted to treat the VOE, i.e. a pretreatment step involving chemical occulation, followed by biological treatment using activated sludge.

The VOE was pretreated using a commercial occulent compound C40 (Chemserve Trio, South Africa) in order to prevent organic shock loading because of the high FOG content. A 300-l plastic vessel was lled with VOE and allowed to reach room temperature (20C). Compound C40 was added to the VOE with slow stirring (3000 rev min)1) at a nal concentration of 8 g l)1. However, the amount of C40 required for complete occulation varied amongst efuent batches, because of the inconsistent nature of the renery process. Clarication was reached after c. 10 min. The supernatant (occulated efuent) remained in the vessel for 2448 h to facilitate efcient removal of the emulsied FOG. The clear supernatant was transferred to a clean vessel in a cold room at 4C. The initial pH of the efuent was acidic (pH 3040) but on addition of the occulent, the pH turned basic (pH 90100). The nal pH was adjusted to pH 74 by the addition of concentrated sulfuric acid, followed by adjustment of the COD concentration to 1000 mg l)1 with tap water. Nitrogen and phosphorus were found to be limiting in the pretreated efuent. To maintain the integrity of the biological system, nitrogen and phosphorus were supplemented in the form of ammonium chloride and potassium dihydrogen orthophosphate salts, at a C : N : P ratio of 100 : 5 : 1. The industrial process was also setup as an MLE process in consonance with the domestic process, except for the following variations in reactor capacities; an 8 l anoxic reactor and two separate 10 l aerobic reactors, giving a total aerobic reactor volume of 20 l. A sludge age of 15 days was maintained by wasting 175 l day)1 of aerobic mixed liquor. Daily monitoring of process performance Daily analyses were conducted to determine steady-state conditions and to monitor process performance. These included COD and total Kjeldahl nitrogen (TKN) analyses, on inuent and efuent samples and MLVSS analysis on aerobic mixed liquor samples. All analyses were performed according to standard methods (Clesceri et al. 1998). Sampling and cell xation Grab samples of activated sludge were collected from the aerobic reactors of the domestic and industrial laboratory scale MLE processes. Samples were washed twice and resuspended in phosphate buffered saline [PBS; 130 mmol l)1 sodium chloride, 10 mmol l)1 sodium phosphate buffer (pH 72)]. Gram-negative and Grampositive bacterial cells were xed immediately as follows; Gram-negative cells were rendered permeable to probes
355

2007 The Authors Journal compilation 2007 The Society for Applied Microbiology, Journal of Applied Microbiology 104 (2008) 353363

Evaluation of the microbial community in wastewater

A.P. Degenaar et al.

by paraformaldehyde xation (Amann 1995). Three volumes of fresh 4% (w v) paraformaldehyde solution was added to one volume of washed sample in a 15-ml polypropylene centrifuge tube and held at 4C for 15 h. Fixative was then removed after centrifugation at 5000 g and cells were resuspended in 50% (v v) ethanol in PBS. Gram positive cells were xed by addition of ice-cold 98% ethanol to samples at a nal concentration of 50% (v v) (Roller et al. 1994). Fixed samples were stored in 50% (v v) ethanol in PBS at )4C until required for hybridization. Sonication and slide preparation Floc disruption was achieved by sonication of 15 ml of xed sample at 5 W for 5 min in a 2-ml micro test-tube using a probe sonicator (Virsonic 100; Virtis, Gardiner, NY). Following sonication, cell dispersion was facilitated by the addition of Igepal CA-630 (Sigma, St Louis, MO, USA), a nonionic, nondenaturing detergent, to samples at a nal concentration of 01% (v v) and vortexed briey. A volume of 10 ll of treated sample was applied to each well on a Teon coated microscope slide, pretreated with 1 : 10 poly-l-lysine solution (Sigma) according to manufacturers instructions. Spots were allowed to air dry before dehydrating through an ethanol series of 60, 80 and 98% (v v) ethanol for 3 min each (Amann 1995). FISH and DAPI staining Oligonucleotide probes used in this study were purchased from MWG-BIOTECH AG (Ebersberg, Germany), modied on the 5 end, with either a tetramethylrhodamine-5-isothiocyanate (TRITC) or 5(6)-carboxyuorescein-N-hydroxysuccinimide (FLOUS) ester and HPLC puried. Table 1 illustrates the oligonucleotide probes, formamide percentages (FA) and sodium chloride concentrations used for FISH in this study. Hybridization was carried out in a 50 ml polypropylene tube, isotonically equilibrated with hybridization buffer as outlined by

Amann (1995). A volume of 10 ll of hybridization buffer probe mix containing; 50 ng probe (5 ng ll)1), 09 mol l)1 NaCl, 001% SDS, 20 mmol l)1 Tris HCl, pH 72 and X% (v v) FA was applied to each dehydrated spot (Specic FA concentrations are given in Table 1) and hybridized at 46C for 15 h. Probes EUB338, EUB338-II and EUB338-III were used in a mixture called EUBmix according to Yeates et al. (2003). Probes BET42a and GAM42a were hybridized simultaneously to increase specicity because of the single mismatch at position 1033 between the target sequences of these probes (Yeates et al. 2003). Hybridization was stopped by rinsing unbound probe from slides with wash buffer containing; 20 mmol l)1 Tris HCl, 001% SDS, 5 mmol l)1 EDTA and Y M NaCl (Specic molarities of sodium chloride are given in Table 1) prewarmed to 48C. Slides were transferred to a 50 ml polypropylene tube lled with prewarmed wash buffer and incubated for 20 min at 48C. Buffer salts were removed by dipping the slides briey in deionized water, excess water was shaken off and slides were air dried. Cells were stained after hybridization with 10 ll of 025 lg ml)1 DAPI solution for 10 min in the dark, rinsed with deionized water and allowed to air dry. Slides were mounted in VECTASHEILD anti-fading mounting medium (Vector Laboratories, Burlingame, CA) and laminated with clear nail polish. Microscopy and image analysis Hybridizations were viewed under a Zeiss Axioplan ttingen, Germany) tted for microscope (Carl Zeiss, Go epiuorescence with a 50 W high pressure mercury lamp and lter sets 02, 09 and 15. Images were captured using a CCD camera (Hamamatsu, Japan) and stored as tiff les. From each hybridization, 30 random elds under 400 magnication were selected for enumeration, using Zeiss KS300 image analysis software (Carl Zeiss). Relative probe percentages were calculated by dividing the number of probe conferred cells by the number of bacterial cells binding to probe EUBmix in each eld.

Table 1 Details of probes, probe sequences, their specicities and hybridization conditions used in this study
Probe name EUB338 EUB338-II EUB338-III ALF1b BET42a GAM42a HGC69a Probe Sequence (53) GCTGCCTCCCGTAGGAGT GCAGCCACCCGTAGGTGT GCTGCCACCCGTAGGTGT CGTTCGYTCTGAGCCAG GCCTTCCCACTTCGTTT GCCTTCCCACATCGTTT TATAGTTACCACCGCCGT Specicity Bacteria Planctomycetales Verrucomicrobiales Alpha-proteobacteria Beta-proteobacteria Gamma-proteobacteria Actinobacteria % FA* 20 20 20 20 35 35 25 NaCl (mol l)1) 019 019 019 019 008 008 015 Reference Daims et al. (1999) Daims et al. (1999) Daims et al. (1999) Wagner et al. (1993) Yeates et al. (2003) Yeates et al. (2003) Roller et al. (1994)

*Percentage of formamide (%v v) in the hybridization buffer. Molarity of sodium chloride in the wash buffer.
2007 The Authors Journal compilation 2007 The Society for Applied Microbiology, Journal of Applied Microbiology 104 (2008) 353363

356

A.P. Degenaar et al.

Evaluation of the microbial community in wastewater

Statistical analysis Statistical analyses were performed using Microsoft Excel spreadsheet software including the Analysis Toolpak addin. The paired t-test was used to determine the effect of VOE on process parameters and bacterial populations. The level of signicance was set at P < 5%, to differentiate between the two sets of data. The Pearson productmoment correlation coefcient (r) was used to investigate the association between the EUB : DAPI ratio and MLVSS concentration in each process. Analysis of variance (anova) single factor, with Alpha set at 005 was used to determine differences amongst the four bacterial populations within each process. Results Steady-state performance of laboratory scale processes Steady-state results of the domestic and industrial MLE processes are presented in Figs. 15. The domestic process demonstrated an average COD removal capacity of 91%, while the industrial process achieved a slightly lower average COD removal capacity of 84% (Fig. 1). Overall, the domestic and industrial processes showed an average TKN removal capacity of 90% (Fig. 2). An average OUR of 31 and 19 mg O l)1 h)1 was measured in the aerobic mixed liquor of the domestic and industrial processes respectively (Fig. 3). The average MLVSS concentration calculated for the aerobic mixed liquor of the domestic process was 2053 mg l)1 (Fig. 4) and 3000 mg l)1 was calculated for the industrial process

120 110 TKN removal (%) 100 90 80 70 60

6 7 8 9 10 11 12 13 14 15 WW batch no.

Figure 2 Capacity of the domestic and industrial wastewater treatment processes for the removal of nitrogen. There was no statistically signicant difference between the nitrogen removal capacities of the processes (P = 776%, Paired two-tailed t-test). (.) Domestic MLE and (,) industrial MLE.

60 50 OUR (mg O l1 h1) 40 30 20 10

105 100 COD removal (%) 95 90 85 80 75 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 WW batch no.


Figure 1 Capacity of the domestic and industrial wastewater treatment processes for the removal of organic material. There was a statistically signicant difference in the organic removal capacities of the processes (P < 0001%, Paired two-tailed t-test). (m) Domestic MLE and (n) industrial MLE.

6 7 8 9 10 11 12 13 14 15 WW batch no.

Figure 3 Oxygen utilization rates of the domestic and industrial wastewater treatment processes. There was a statistically signicant difference between the measured oxygen utilization rates of the processes (P < 0001%, Paired two-tailed t-test). (d) Domestic MLE and ( ) industrial MLE.

(Fig. 5). Hybridization of aerobic mixed liquor samples revealed that on average 81% of DAPI stained cells in the domestic process (Fig. 4) and 72% of DAPI stained cells in the industrial process (Fig. 5) bound to probe EUBmix. Microbial community analysis Hybridization of aerobic mixed liquor samples from the domestic process with family level probes; ALF1b, BET42a, GAM42a and HGC69a, revealed that on average
357

2007 The Authors Journal compilation 2007 The Society for Applied Microbiology, Journal of Applied Microbiology 104 (2008) 353363

Evaluation of the microbial community in wastewater

A.P. Degenaar et al.

120 110 100 90 80 70 60 50 40 30 20 10 0

8000 7000
MLVSS (mg l1)

60 50 40 30 20 10 0

6000 5000 4000 3000 2000 1000 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 WW batch no. 0

DAPI (%)

EUBmix (%)

Figure 4 Percentages of EUBmix-hybridized cells relative to DAPI counts and mixed liquor volatile suspended solids concentrations in aerobic mixed liquor samples from the domestic wastewater treatment process. There was no correlation between EUB : DAPI ratios and MLVSS concentrations of the domestic process (r = 037, Pearson correlation coefcient). ( ) EUB DAPI and ( ) MLVSS.

6 7 8 9 10 11 12 13 14 15 WW batch no.

Figure 6 Percentages of group-specic probes relative to EUBmix counts in aerobic mixed liquor samples from the domestic wastewater treatment process. There was a signicant difference amongst the percentages of the four group-specic probes within the domestic process (P < 0001%, anova single factor). ( ) ALF 1b; ( ) BET 42a; ( ) GAM 42a and ( ) HGC 69a.

2000 1000 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 WW batch no. 0

EUBmix (%)

120 110 100 90 80 70 60 50 40 30 20 10 0

8000 7000 6000 5000 4000 3000 MLVSS (mg l1)

60 50 40 30 20 10 0

Figure 5 Percentages of EUBmix-hybridized cells relative to DAPI counts and mixed liquor volatile suspended solids concentrations in aerobic mixed liquor samples from the industrial wastewater treatment process. There was no correlation between EUB : DAPI ratios and MLVSS concentrations of the domestic process (r = 068, Pearson correlation coefcient). ( ) EUB DAPI and ( ) MLVSS.

DAPI (%)

6 7 8 9 10 11 12 13 14 15 WW batch no.

cells hybridizing to probe ALF1b accounted for 26% of the cells hybridizing to EUBmix. Bacteria afliated to the beta- and gamma- subclasses of Proteobacteria represented 33% and 15% of cells detected by EUBmix respectively. With probe HGC69a specic for Actinobacteria, 6% of cells hybridized by EUBmix were also detected (Fig. 6). Whereas, the mixed liquor from the industrial process showed that the alpha- subclass of Proteobacteria represented 31% of EUBmix cells. Cells which hybridized to BET42a and GAM42a accounted for 17% and 8% of cells detected by EUBmix respectively and 4% of EUBmix cells represented the Actinobacteria (Fig. 7).
358

Figure 7 Percentages of group-specic probes relative to EUBmix counts in aerobic mixed liquor samples from the industrial wastewater treatment process. There was a signicant difference amongst the percentages of the four group-specic probes within the industrial process (P < 0001%, anova single factor). ( ) ALF 1b; ( ) BET 42a; ( ) GAM 42a and ( ) HGC 69a.

Discussion Subsequent to a one month period of acclimation, the MLE processes were operated for the duration of 15 WW batches. During this period, process performances were typical of steady-state behaviour. COD and TKN removal capacities, OURs and MLVSS concentrations of both MLE processes were consistent (Figs 15). Pretreatment

2007 The Authors Journal compilation 2007 The Society for Applied Microbiology, Journal of Applied Microbiology 104 (2008) 353363

A.P. Degenaar et al.

Evaluation of the microbial community in wastewater

of the VOE using compound C40 removed 73% of the COD. Research by Bettazzi et al. (2007) conrmed that pretreatment using physico-chemical processes improved the efciency of biological treatment. Previous research conducted by Sengul (1989) using coagulants such as ploy-electrolytes when treating sunower oil efuents, showed COD removal rates of 76% and demonstrated pretreatment as a necessary step prior to biological treatment. As shown in Fig. 1, the industrial process achieved a mean COD removal capacity of 84%, which was signicantly lower than the domestic process which averaged 91% COD removal. The difference in COD removal capacities between the processes could have been attributed to the chemical composition of the VOE. The high organic load (1000 mg COD l)1) and the possibility of the presence of toxic compounds in the VOE, could have impacted on the COD removal capacity of the industrial process. Research by Boukchina et al. (2007) also conrmed that VOEs such as olive mill WW comprise toxic compounds, which negatively impact on biological treatment. The COD removal capacity of the industrial process support the work of other researchers such as Mulligan and Sheridan (1975), who demonstrated the capability of activated sludge to treat emulsied lipids ztu rk and provide removal efciencies as high as 80%. O et al. (1989) using a laboratory scale activated sludge process to treat VOE, achieved a COD removal capacity of 72%. Mkhize et al. (2000) achieved 70% COD removal using an anaerobic aerobic sequencing batch reactor to treat VOE. Reddy et al. (2003) demonstrated an average COD removal capacity of 81% using a similar laboratory scale MLE process to treat VOE. Current ndings substantiated previous research by Boukchina et al. (2007) who showed 84% COD removal using aerobic treatment of olive mill WW, although the initial COD concentration of the inuent was 250 mg l)1. In addition, other biological treatment processes using biomass rich in fungi showed a COD removal capacity of 86% under aerobic conditions (Caffaz et al. 2007). The TKN removal capacity of the industrial process was not affected by the VOE as demonstrated in Fig. 2, with TKN removal capacities of both processes averaging 90% removal. However, nitrogen removal was not the main focus of this study as nitrogen was found to be limiting in the VOE and was, therefore, supplemented prior to biological treatment. In this investigation, three methods were used to determine the active biomass concentration of the aerobic mixed liquor of the two laboratory scale processes namely; OUR measurement, MLVSS determination and FISH using probe EUBmix. OUR measurement and hybridization with probe EUBmix was used to asses the overall physiological state of the processes. The EUB : DAPI ratio provides an indication of the ratio of

total number of metabolically active eubacterial cells to the total number of cells and is a direct measure of the metabolic activity of bacterial cells in activated sludge biomass. DeLong et al. (1989) and Gourse et al. (1996) have shown that rRNA content within bacterial cells is directly proportional to growth rates. Therefore, it can be assumed that in activated sludge mixed liquor, cells bearing probe conferred uorescence are metabolically active; hence, only active cells are counted. The EUB : DAPI ratio was therefore used as a direct measure of the metabolic activity of the bacterial biomass. A high EUB : DAPI ratio in activated sludge would therefore indicate a highly metabolically active bacterial population. The OUR of the industrial process was signicantly lower than that of domestic process (Fig. 3). This could be attributed to the oily nature of the VOE resulting in lipid overloading of the activated sludge biomass. Banerji (1974) suggested that the high FOG loadings may cause the activated sludge oc to become coated with hydrophobic material, thereby limiting oxygen transfer efciency and reducing the OUR. The results of OUR measurement (Fig. 3) and EUB : DAPI ratios (Figs 4 and 5) were in agreement. Both sets of results indicated that the aerobic mixed liquor of the industrial process had reduced metabolic activity compared with the domestic process (Fig. 3). The depleted OUR in the industrial process reected stress on the microbial community as depicted by a decrease in EUB : DAPI ratios throughout the process (Fig. 4). To determine whether there was a correlation between the MLVSS concentration and EUB : DAPI ratio, the results of MLVSS determinations and EUBmix hybridizations of each process were combined. As shown in Figs 4 and 5, Pearson correlation coefcient values of 037 and 068 were calculated for the domestic and industrial processes, respectively. These values indicated that there was no correlation between MLVSS and EUB : DAPI ratios in either of the processes. Results of hybridization with EUBmix revealed that 72% of the total number of cells stained with DAPI in the aerobic mixed liquor of the industrial process and 81% of DAPI stained cells in the domestic process bound to probe EUBmix and can therefore be assumed to be metabolically active, belonging to the domain Eubacteria (Figs 4 and 5). The EUB : DAPI ratio determined for the industrial process was signicantly lower than the domestic process (P = 0009%) This could also indicate a slightly diminished contribution of the bacteria in the industrial process when compared with the domestic process. However, the opposite effect was observed according to MLVSS analysis, which showed that the industrial process had a signicantly greater active biomass concentration of 3000 mg l)1 (P < 0001%) (Fig. 5) when compared with 2053 mg l)1 calcu359

2007 The Authors Journal compilation 2007 The Society for Applied Microbiology, Journal of Applied Microbiology 104 (2008) 353363

Evaluation of the microbial community in wastewater

A.P. Degenaar et al.

lated for the domestic process (Fig. 4). The lack of correlation existing between MLVSS concentrations and EUB : DAPI ratios of both processes, demonstrates that MLVSS analysis is an indirect method which is not sensitive to changes in metabolic activity and therefore does not give a true reection of the metabolic activity of activated sludge biomass. These ndings validate the importance of molecular techniques such as FISH in accurately quantifying the active biomass concentration in activated sludge mixed liquors. The colour of the mixed liquor in activated sludge treatment processes treating domestic WW and operated optimally is generally dark brown (chocolate brown; Lilley et al. 1997). However, the effect of the VOE on the colour of the activated sludge, resulted in a transition from the original brown colour to cream over time. Washing of mixed liquor samples twice with PBS prior to xation proved successful in removing excess oils from the ocs which minimized background noise, thereby improving visualization during epiuorescent microscopy. Sonication parameters for adequate oc disruption were optimized at 5 W for 5 min. The composition and distribution of the bacterial populations as per phylogenetic classication system in the aerobic mixed liquor of the industrial and domestic processes are presented in Figs 6 and 7 respectively. The results are expressed as percentages of probe specic counts divided by EUB338mix counts. The microbial community prole of the domestic process served as a control to determine possible population shifts that may have occurred because of the degradation of the VOE by the aerobic mixed liquor of the industrial process. The community structure prole of the domestic process (Fig 6.) indicated the beta-proteobacteria to be the numerically dominant group of Eubacteria, i.e. 33% of Eubacterial cells were beta-proteobacteria (BET42a EUBmix), followed by alpha-proteobacteria at 26% (ALF1b EUBmix), gamma-proteobacteria at 15% (GAM42a EUBmix) and Actinobacteria at 6% (HGC69a EUBmix). Mudaly et al. (2000) showed a similar trend of bacterial predominance, with the betaproteobacteria comprising 22% of total cells, followed by alpha-proteobacteria (19%); gamma-proteobacteria (17%) and Actinobacteria (11%) in a laboratory scale EBPR sludge-treating domestic WW. These results reect a similar scenario in full scale processes treating domestic WW. Wagner et al. (1993) identied members of the beta-proteobacteria predominating in aerated activated sludge systems with 42% of cells binding to BET42a with respect to probe EUB. Wong et al. (2005) surveyed WW treatment plants in Japan and found the beta-proteobacteria to be the most abundant group in two treatment plants treating domestic WW and accounted for 20% and 30% of EUBmix-stained cells respectively.
360

Distinct differences between the microbial communities present within the aerobic reactors of the domestic and industrial processes were observed. Results of the paired t-test indicated there was a signicant difference in the alpha-proteobacteria (P = 13%), beta-proteobacteria (P < 0001%), gamma-proteobacteria (P < 0001%) and Actinobacteria (P = 11%) populations between the two processes. In the industrial process, it was evident that the alpha-proteobacteria were found to be numerically predominant binding to 31% of EUBmix detected cells, followed by beta-proteobacteria and gamma-proteobacteria constituting 17% and 8% of EUBmix stained cells, respectively (Fig. 7). The Actinobacteria represented only 4% of EUBmix stained cells (Fig. 7). A similar observation was made by Stoffels et al. (1998) using a trickle bed bioreactor to treat aromatic compounds whereby the seed inoculum comprising of gamma-proteobacteria when inoculated into the fermenter, resulted in predominance of members of the alpha-proteobacteria and beta-proteobacteria. In the current research, the response to pollutant adaptation resulted in the process selecting the appropriate bacteria to degrade the substrate, thereby enforcing dominance of the alpha-proteobacteria. The decrease in the beta-proteobacteria could be attributed to the toxicity of the substrate (VOE) to the bacteria within the subclass. Substrate uptake is a very important factor in the selection of microbial populations in activated sludge and determines dominance of one group over the other. Microbes that are not adapted to metabolize the substrates are usually washed out of the system (Satoh et al. 1998), explaining the decrease in the other Proteobacterial populations in relation to the alpha-proteobacteria. The overall comparative predominance of alpha-proteobacteria throughout the industrial process further supported the theory of acclimation, i.e. members of the alpha-proteobacteria had adapted to the substrate. Comparative analysis of the two processes using group-specic probes showed dominance of the beta-proteobacteria in the domestic process and the alpha-proteobacteria in the industrial process. It is common knowledge that bacterial populations in domestic processes are dominated by the beta-proteobacteria and therefore the change in the population dynamics in the industrial process could be attributed to the nature of the substrate, which comprised primarily VOE. A signicant nding by Kaewpipat and Grady (2002) showed that bacterial communities in identically operated activated sludge reactors became signicantly different over time, even though they started from a common community. In spite of differences in community composition of the processes, the TKN removal efciencies were similar, demonstrating that different microbial communities can be functionally similar.

2007 The Authors Journal compilation 2007 The Society for Applied Microbiology, Journal of Applied Microbiology 104 (2008) 353363

A.P. Degenaar et al.

Evaluation of the microbial community in wastewater

Current ndings were in agreement with research by Layton et al. (2000), who conducted a detailed microbial community prole of an activated sludge system treating WW from a chemical manufacturing process. On acclimation, the percentage of 16S rDNA cloned sequences showed representatives of the alpha-proteobacteria predominating, followed by the beta-proteobacteria and gamma-proteobacteria. The microbial community of a dairy WW plant was shown to be numerically dominated by spindle shaped bacteria belonging to the Flavobacter Bacteriodes phylum (Manz et al. 1994). These ndings reiterate that the microbial community prole of municipal WW plants treating domestic WW and comprising predominantly beta-proteobacteria (Wagner et al. 1993) certainly differ from those treating industrial efuents. Population shifts become evident once the seed sludges are exposed to the industrial WW (inuent) and can be attributed to substantial differences in the chemical composition of waste streams. Industrial WWs generally have a much higher total organic carbon load when compared to municipal (Layton et al. 2000). With reference to the current research, the VOE was a complex substrate comprising a range of lipids, which necessitated initial hydrolysis of the substrate by the bacteria before the substrate could be metabolized, resulting in a limiting step in the degradation process. FISH analysis conrmed that the gamma-proteobacteria and Actinobacteria were numerically low in both processes. Previous research has shown that the gamma-proteobacteria were less abundant in municipal activated sludge (Wagner et al. 1993; Manz et al. 1994). Low percentages of Actinobacteria were also veried by Wagner and Amann (1997) who found only 9% of all bacteria present in activated sludge samples originating from a municipal sewage treatment plant were Actinobacteria. The sums of the group-specic probes (ALF1b, BET42a, GAM42a and HGC69a) were 80% and 60% for the domestic and industrial processes respectively. The sum of 80% calculated for the domestic process and 60% calculated for the industrial process were lower than the EUB : DAPI ratios of 81% and 72% respectively. This indicated that the majority of microbial community of the domestic process was accounted for by the group-specic probes, whereas 11% of the bacterial biomass of the industrial process was not accounted for. These populations could belong to other divisions such as the CFB mpfer et al. 1996) and low G + C Gram-posiphylum (Ka tive bacteria (Snaidir et al. 1997) or even possibly hitherto undiscovered taxonomic ranks. It is also possible that upon acclimation, the industrial process could have selected for a highly specialized population that could not be detected by the group-specic probes used. The applied value of the research is in the accurate elucidation

of the microbial communities involved in the biological treatment of VOE, using novel molecular techniques. This information will aid plant operators and researchers to understand the microbial population dynamics involved, with the ultimate objective of trouble-shooting and optimizing biological treatment of vegetable oil efuent. Conclusion Biological treatment can be used to successfully treat VOE and results obtained compared favourable with previous research conducted in the eld. FISH is more sensitive and could be regarded as a more accurate technique than MLVSS analysis in determining the active biomass fraction of activated sludge mixed liquor. FISH analysis gives a more reliable result as it measures the active biomass fraction directly or in situ and is sensitive to changes in cellular rRNA content as only metabolically active cells are counted. The high organic strength of the VOE reduced the COD removal capacity, but did not affect the TKN removal capacity of the laboratory scale industrial process. The VOE had the following effects on process parameters of a laboratory scale MLE process: (i) a reduced OUR, because of possible lipid overloading and (ii) an elevated MLVSS concentration. In a laboratory scale MLE process treating VOE, a dominance of the alpha-proteobacteria over members of the beta-proteobacteria, gamma-proteobacteria and Actinobacteria in the aerobic mixed liquor was observed. This could be attributed to pollutant adaptation of the alpha-proteobacteria to the VOE, resulting in the dominance of the alphaproteobacteria. Acknowledgement The Durban University of Technology (DUT) and the National Research Foundation (NRF). References
Amann, R.I. (1995) In situ identication of micro-organisms by whole cell hybridization with rRNA-targeted nucleic acid probes. In Molecular Microbial Ecology Manual ed. Akkermans, A.D.L., van Elsas, J.D. and de Bruijn, F.J. Vol. 3.3.6. pp. 115. London: Kluwer Academic Publishers. Amann, R.I., Binder, B.J., Olsen, R.J., Chrisholm, S.W., Devereux, R. and Stahl, D.A. (1990) Combination of 16S rRNA targeted oligonucleotide probes with ow cytometry for analyzing mixed microbial populations. Appl Environ Microbiol 56, 19191925. Amann, R.I., Ludwig, W. and Schleifer, K.H. (1995) Phylogenetic identication and in situ detection of individual

2007 The Authors Journal compilation 2007 The Society for Applied Microbiology, Journal of Applied Microbiology 104 (2008) 353363

361

Evaluation of the microbial community in wastewater

A.P. Degenaar et al.

microbial cells without cultivation. Microbiol Rev 59, 143 169. Banerji, J.K. (1974) Grease problems in municipal wastewater treatment systems. In Proceedings of the 29th Industrial Wastewater Conference pp. 768. West Lafayette, IN: Purdue University. Beccari, M., Majone, M., Petrangeli Papini, M. and Torrisi, L. (2001) Enhancement of anaerobic treatability of olive oil mill efuents by addition of Ca(OH)2 and bentonite without intermediate solid liquid separation. Water Sci Technol 43, 275282. Bettazzi, E., Caretti, C., Caffaz, S., Azzari, E. and Lubello, C. (2007) Oxidative processes for olive mill wastewater treatment. Water Sci Technol 55, 7987. Boukchina, R., Choi, E., Kim, S., Yu, Y.B. and Cheung, Y.J. (2007) Strategy for olive mill wastewater and reuse with a sewage plant in an arid region. Water Sci Technol 55, 7178. Caffaz, S., Caretti, C., Morelli, M., Lubello, C. and Azzari, E. (2007) Olive mill wastewater biological treatment by fungi biomass. Water Sci Technol 55, 8997. Clesceri, L.S., Greenberg, A.E. and Eaton, A.D. (ed.) (1998) Standard Methods for the Examination of Water and Wastewater, 20th edn. Washington DC: APHA, AWWA and WEF. hl, A., Amann, R., Schleifer, K.H. and Wagner, Daims, H., Bru M. (1999) The domain-specic probe EUB338 is insufcient for the detection of all Bacteria: development and evaluation of a more comprehensive probe set. Syst Appl Microbiol 22, 434444. Daims, H., Ramsing, N.B., Schleifer, K.H. and Wagner, M. (2001) Cultivation-independent, semiautomatic determination of absolute bacterial cell numbers in environmental samples by uorescence in situ hybridization. Appl Environ Microbiol 67, 58105818. DeLong, E.F., Wickham, G.S. and Pace, N.R. (1989) Phylogenetic stains: ribosomal RNA-based probes for the identication of single cells. Science 243, 13601363. Dueholm, T.E., Andreasen, K.H. and Nielsen, P.H. (2001) Transformation of lipids in activated sludge. Water Sci Technol 43, 165172. ztu _ San, H.A. and Demir, I. _ (1990) Com lu, V., O rk, I., Erog parative evaluation of treatment alternatives for wastewaters from edible oil rening industry. Water Sci Technol 22, 225234. Gourse, R.L., Gaal, T., Bartlett, M.S., Appleman, J.A. and Ross, W. (1996) rRNA transcription and growth rate-dependant regulation of ribosome synthesis in Escherichia coli. Ann Rev Microbiol 50, 645677. Henze, M., Gujer, W., Mino, T., Matsuo, T., Wentzel, M.C. and Marais, G.v.R. (1995) Activated Sludge Model No.2. I.A.W.Q. Scientic and Technical Report No.3. London: Cambridge University Press. Hicks, R.E., Amann, R.I. and Stahl, D.A. (1992) Dual staining of natural bacterioplankton with 4, 6-diamidino-2-phenylindole and uorescent oligonucleotide probes targeting

kingdom-level 16S rRNA sequences. Appl Environ Microbiol 58, 21582163. Horan, N.J. (1990) Biological Wastewater Treatment Systems: Theory and Operation. Chinchester, England: John Wiley & Sons. Kaewpipat, K. and Grady, C.P.L. (2002) Microbial population dynamics in laboratory scale activated sludge reactors. Water Sci Technol 46, 1927. mpfer, P., Erhart, R., Beimfohr, C., Bo hringer, J., Wagner, Ka M. and Amann, R. (1996) Characterization of bacterial communities from activated sludge: culture-dependant numerical identication vs in situ identication using group- and genus-specic rRNA-targeted oligonucleotide probes. Microb Ecol 32, 101121. Keith, J.E., Boyer, J.G. and de los Reyes, F.L. III (2005) Changes in the rRNA levels of specic microbial groups in activated sludge during sample handling and storage. Lett Appl Microbiol 41, 208215. Layton, A.C., Karanth, P.N., Lajoie, C.A., Meyers, A.J., Gregory, I.R., Stapleton, R.D., Taylor, D.E. and Sayler, G.S. (2000) Quantication of Hyphomicrobium populations in activated sludge from an industrial wastewater treatment system as determined by 16S rRNA analysis. Appl Environ Microbiol 66, 11671174. Lilley, I.D., Pybus, P.J. and Power, S.P.B. (1997) Operating Manual for Biological Nutrient Removal in Wastewater Treatment Works. WRC Report No. TT 83 97. Pretoria, South Africa: Water Research Commission. Manz, W., Wagner, M., Amann, R. and Schleifer, K.H. (1994) In situ characterization of the microbial consortia active in two wastewater treatment plants. Water Res 28, 17151723. Mkhize, S.P., Atkinson, B.W. and Bux, F. (2000) Evaluation of a laboratory scale biological process for the treatment of edible oil efuent. Water SA 26, 555558. Mudaly, D.D., Atkinson, B.W. and Bux, F. (2000) Microbial community prole of a biological excess phosphorus removal (BEPR) activated sludge system using a cultivation-independent approach. Water SA 26, 343352. Mulligan, T.J. and Sheridan, R.P. (1975) Treatment of high strength fatty acid derivative wastewaters. In Proceedings of the 30th Industrial Waste Conference pp. 997. West Lafayette, IN: Purdue University. Muyzer, G., de Waal, E.C. and Uitterlinden, A.G. (1993) Proling of complex microbial populations by denaturing gradient gel electrophoresis analysis of polymerase chain reaction-amplied genes coding for 16S rRNA. Appl Environ Microbiol 59, 695700. ztu rk, I., San, H.A. and Erog lu, V. (1989) Pilot and full scale O treatability studies on wastewaters from an edible oil rening industry. In Proceedings of the 44th Industrial Waste Conference, May 911. pp. 769776. West Lafayette, IN: Purdue University. Randall, E.W., Wilkinson, A. and Ekama, G.A. (1991) An instrument for the direct determination of oxygen utilization rate. Water Sci Technol 17, 1118.

362

2007 The Authors Journal compilation 2007 The Society for Applied Microbiology, Journal of Applied Microbiology 104 (2008) 353363

A.P. Degenaar et al.

Evaluation of the microbial community in wastewater

Reddy, K., Drysdale, G.D. and Bux, F. (2003) Evaluation of activated sludge treatment and settleability in remediation of edible oil efuent. Water SA 29, 245250. Roller, C., Wagner, M., Amann, R., Ludwig, W. and Schleifer, K.H. (1994) In situ probing of Gram-positive bacteria with high DNA G + C content using 23S rRNA-targeted oligoucleotides. Microbiology 140, 28492858. Satoh, H., Mino, T. and Matsuo, T. (1998) Diversity of microorganisms in activated sludge processes in metabolisms of substrate uptake. In Proceedings of the Second International Symposium on Microbial Community and Functions in Wastewater Treatment Processes. pp. 102114. Department of Environmental and Urban Engineering, University of Tokyo, 10 March 1998. Sengul, F. (1989) A case study on sunower seed oil industries waste characterization, classication and treatment. Wat Sci Technol 22, 241248. Snaidir, J., Amann, R., Huber, I., Ludwig, W. and Schleifer, K.H. (1997) Phylogenetic analysis and in situ identication of bacteria in activated sludge. Appl Environ Microbiol 63, 28842896. Steffen, O., Roberts, A. and Kirsten, H. (1989) Water and Waste-Water Management in the Edible oil Industry. WRC Report No. TT 40 89. Pretoria, South Africa: Water Research Commission. Stoffels, M., Amann, R., Ludwig, W., Hekmet, D. and Schleifer, K.H. (1998) Bacterial community dynamics during start-up of a trickle-bed bioreactor degrading aromatic compounds. Appl Environ Microbiol 64, 930939.

Wagner, M. and Amann, R.I. (1997) Molecular techniques for determining microbial community structures in activated sludge. In Microbial Community Analysis: The Key to the Design of Biological Wastewater Treatment Systems. I.A.W.Q. Scientic and Technical Report no. 5. ed. Cloete, T.E. and Muyima, N.Y.O. pp. 6172. London: Cambridge University Press. Wagner, M., Amann, R., Lemmer, H. and Schleifer, K.H. (1993) Probing activated sludge with oligonucleotides specic for proteobacteria: inadequacy of culture-dependent methods for describing microbial community Structure. Appl Environ Microbiol 59, 15201525. Wentzel, M.C., Mbewe, A. and Ekama, G.A. (1995) Batch test for the measurement of readily biodegradable COD and active organism concentrations in municipal wastewaters. Water SA 21, 117124. Woese, C.R. (1987) Bacterial evolution. Microbiol Rev 51, 221 271. Wong, M.T., Mino, T., Seviour, R.J., Onuki, M. and Liu, W.T. (2005) In situ identication and characterization of the microbial community structure of full-scale enhanced biological phosphorous removal plants in Japan. Water Res 39, 29012914. Yeates, C., Saunders, A.M., Crocetti, G.R. and Blackall, L.L. (2003) Limitations of the widely used GAM42a and BET42a probes targeting bacteria in the Gammaproteobacteria radiation. Microbiology 149, 12391247.

2007 The Authors Journal compilation 2007 The Society for Applied Microbiology, Journal of Applied Microbiology 104 (2008) 353363

363

You might also like