was waster

8/22/2019 Was Waster

1/12

O R I G I N A L A R T I C L E

Comparative evaluation of the microbial community inbiological processes treating industrial and domesticwastewaters

A.P. Degenaar, A. Ismail and F. Bux

Centre for Water and Wastewater Technology, Department of Biotechnology, Durban University of Technology, Durban, South Africa

Introduction

Vegetable oil refining 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 effluent (Steffen et al.

1989). Because of large volumes of this effluent being

released into sewer systems, treatment to an acceptable

standard is required prior to discharge (Horan 1990).

Discharge of poor quality final effluents impacts nega-

tively on natural water sources resulting in eutrophica-

tion of natural ecosystems. Vegetable oil effluent (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 final effluents being produced, which do

not satisfy municipal discharge standards (Eroglu et al.

1990).

There are two methods of effluent treatment com-

monly employed by vegetable oil refineries in South

Africa; physical separation of oil and grease via dissolved

air floatation and pH correction (Lilley et al. 1997).

Previous studies have shown that effluents from food

Keywords

alpha-proteobacteria, beta-proteobacteria,

FISH, vegetable oil effluent 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: [email protected]

20070603: received 16 April 2007, revised

23 July 2007 and accepted 23 July 2007

doi:10.1111/j.1365-2672.2007.03563.x

Abstract

Aims: Comparison of the microbial composition and process performance

between laboratory scale processes treating domestic and vegetable oil waste-

waters.

Methods and Results: Two laboratory scale modified LudzackEttinger pro-

cesses were operated under similar operating conditions. One process was feddomestic wastewater and the other an industrial wastewater, vegetable oil efflu-

ent. 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 pro-

cess. The alpha-proteobacteria was the dominant community in the industrial

process (31% of EUBmix), while the beta-proteobacteria dominated the domes-

tic process (33% of EUBmix).

Conclusions: The findings served to establish a difference in the microbial pop-

ulation between the processes. Therefore, the class alpha-proteobacteria could

play a primary role in the degradation of vegetable oil effluent.

Significance and Impact of the Study: This research will aid in process design

and retrofitting of biological processes treating vegetable oil effluent.

Journal of Applied Microbiology ISSN 1364-5072

2007 The Authors

Journal compilation 2007 The Society for Applied Microbiology, Journal of Applied Microbiology 104 (2008) 353363 353


2/12

industries containing fatty materials are readily bio-

degradable and are therefore amendable to biological

treatment methods (Eroglu et al. 1990). Conventional

treatment of edible oil effluent primarily involves physi-

co-chemical processes (Bettazzi et al. 2007). However,

the application of biological treatment is gaining much

attention, with focus on the development of anaerobictreatment processes (Beccari et al. 2001). It is advanta-

geous 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 treat-

ment as an alternative has not been fully investigated,

therefore a lack of knowledge of the microbial communi-

ties responsible for VOE degradation exists. In this inves-

tigation, 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 perfor-

mance with a modified LudzackEttinger (MLE) process

treating domestic WW.

Mixed liquor volatile suspended solids (MLVSS) analy-

sis is the conventional means of measuring the active bio-

mass 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 acti-

vated sludge flocs 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 reflects the rate at which micro-

organisms utilize oxygen (Lilley et al. 1997). Although it

is not common practice to characterize the transforma-

tion 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, specifically hybridiza-

tion 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 fer-

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. How-

ever, only 15% of the indigenous bacteria in activated

sludge could be cultivated (Wagner et al. 1993; Kampferet al. 1996). These limitations have lead to techniques

using the 16S rRNA approach. In particular, fluorescent

in situ hybridization (FISH; Amann et al. 1995), poly-

merase chain reaction (PCR) and denaturing gradient gel

electrophoresis (DGGE) (Muyzer et al. 1993) have been

used extensively to conduct microbial community analy-

sis. The comparative analysis of rRNA molecules has

revolutionized our view of microbial taxonomy and

evolution (Woese 1987). Ribosomal RNA sequences are

perfect targets for fluorescently labelled oligonucleotide

probes, because they are highly conserved and naturally

amplified, 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 oligonu-

cleotides, probe specificity to individual phyla or species,

can be freely adjusted. In addition, DeLong et al. (1989)

showed that probe binding varied with ribosomal con-

tent and reflected cell growth rate, viz., metabolically

active cells will produce intensified fluorescence, 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-diami-dino-2-phenylindole (DAPI; Hicks et al. 1992) gives not

only an indication of the metabolic activity of bacteria,

but also that cells had sufficient rRNA for detection,

were permeabilized for probes by standard fixation pro-

cedures. Therefore, a high EUB : DAPI ratio in activated

sludge would indicate a highly metabolically active bacte-

rial population.

In probing COD removing activated sludges from vari-

ous municipal plants with oligonucleotide probes specific

for proteobacteria, Wagner et al. (1993) demonstrated

the dominance of proteobacteria, which together com-

prised 6075% microbial cells stained with DAPI.

Wagner and Amann (1997) reported members of the

beta-proteobacteria as playing a major role in the micro-

bial 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 treat-

ment of VOE.

Evaluation of the microbial community in wastewater A.P. Degenaar et al.

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


3/12

Materials and methods

Configuration and operation of laboratory scale

processes

The domestic and industrial laboratory scale biological

treatment processes were modelled upon the MLE processand 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 pro-

cess was obtained from Southern WW Works (Durban,

South Africa). The WW (primarily domestic) served as

influent 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 configuration of the domestic process con-

sisted of the following: an anoxic reactor (6 l), an aerobic

reactor (9 l) and a clarifier (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 influent flow rate. A sludge-recycle was

setup between the clarifier and the anoxic zone at a ratio

of 1 : 1 with respect to the influent flow rate. The influ-

ent flow rate was set at 24 l day)1 and a sludge age of

10 days was maintained by wasting 15 l day)1

of mixedliquor 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 auto-

mated 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 Purification

Works (Pietermaritzburg, South Africa). VOE used as

the influent for the industrial process, was collected in

25 l drums from the drain at the end of the refinery

process of a local edible oil refinery, situated in the

direct vicinity of Darvill WW Treatment Purification

Works. The 25 l drums of effluent 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 flocculation, fol-

lowed by biological treatment using activated sludge.

The VOE was pretreated using a commercial flocculent

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 filled with VOE

and allowed to reach room temperature (20C). Com-

pound C40 was added to the VOE with slow stirring

(3000 rev min)1

) at a final concentration of 8 g l)1

.However, the amount of C40 required for complete floc-

culation varied amongst effluent batches, because of the

inconsistent nature of the refinery process. Clarification

was reached after c. 10 min. The supernatant (floccu-

lated effluent) remained in the vessel for 2448 h to

facilitate efficient removal of the emulsified FOG. The

clear supernatant was transferred to a clean vessel in a

cold room at 4C. The initial pH of the effluent was

acidic (pH 3040) but on addition of the flocculent,

the pH turned basic (pH 90100). The final pH was

adjusted to pH 74 by the addition of concentrated sul-

furic acid, followed by adjustment of the COD concen-

tration to 1000 mg l)1 with tap water. Nitrogen and

phosphorus were found to be limiting in the pretreated

effluent. To maintain the integrity of the biological sys-

tem, 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 aero-

bic 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) analy-

ses, on influent and effluent samples and MLVSS analysis

on aerobic mixed liquor samples. All analyses were per-

formed according to standard methods (Clesceri et al.

1998).

Sampling and cell fixation

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 Gram-

positive bacterial cells were fixed immediately as follows;

Gram-negative cells were rendered permeable to probes

A.P. Degenaar et al. Evaluation of the microbial community in wastewater

2007 The Authors



4/12

by paraformaldehyde fixation (Amann 1995). Three vol-

umes of fresh 4% (wv) paraformaldehyde solution was

added to one volume of washed sample in a 15-ml poly-

propylene centrifuge tube and held at 4C for 15 h. Fixa-

tive was then removed after centrifugation at 5000 g and

cells were resuspended in 50% (vv) ethanol in PBS.

Gram positive cells were fixed by addition of ice-cold98% ethanol to samples at a final concentration of 50%

(vv) (Roller et al. 1994). Fixed samples were stored in

50% (vv) ethanol in PBS at )4C until required for

hybridization.

Sonication and slide preparation

Floc disruption was achieved by sonication of 15 ml of

fixed 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 final concentration of 01% (vv) and vortexed briefly.

A volume of 10 ll of treated sample was applied to each

well on a Teflon coated microscope slide, pretreated with

1 : 10 poly-l-lysine solution (Sigma) according to manu-

facturers instructions. Spots were allowed to air dry

before dehydrating through an ethanol series of 60, 80

and 98% (vv) 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),modified on the 5 end, with either a tetramethylrhod-

amine-5-isothiocyanate (TRITC) or 5(6)-carboxyfluores-

cein-N-hydroxysuccinimide (FLOUS) ester and HPLC

purified. Table 1 illustrates the oligonucleotide probes,

formamide percentages (FA) and sodium chloride con-

centrations used for FISH in this study. Hybridization

was carried out in a 50 ml polypropylene tube, isotoni-

cally equilibrated with hybridization buffer as outlined by

Amann (1995). A volume of 10 ll of hybridization

bufferprobe mix containing; 50 ng probe (5 ng ll)1),

09 mol l)1 NaCl, 001% SDS, 20 mmol l)1 TrisHCl, pH

72 and X% (vv) FA was applied to each dehydrated

spot (Specific 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 calledEUBmix according to Yeates et al. (2003). Probes BET42a

and GAM42a were hybridized simultaneously to increase

specificity 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 TrisHCl, 001% SDS, 5 mmol l)1 EDTA

and Y M NaCl (Specific molarities of sodium chloride are

given in Table 1) prewarmed to 48C. Slides were trans-

ferred to a 50 ml polypropylene tube filled with pre-

warmed wash buffer and incubated for 20 min at 48C.

Buffer salts were removed by dipping the slides briefly 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

microscope (Carl Zeiss, Gottingen, Germany) fitted for

epifluorescence with a 50 W high pressure mercury lampand filter sets 02, 09 and 15. Images were captured using

a CCD camera (Hamamatsu, Japan) and stored as tiff

files. From each hybridization, 30 random fields under

400 magnification 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 field.

Table 1 Details of probes, probe sequences, their specificities and hybridization conditions used in this study

Probe name Probe Sequence (53) Specificity % FA* NaCl (mol l)1) Reference

EUB338 GCTGCCTCCCGTAGGAGT Bacteria 20 019 Daims et al. (1999)

EUB338-II GCAGCCACCCGTAGGTGT Planctomycetales 20 019 Daims et al. (1999)

EUB338-III GCTGCCACCCGTAGGTGT Verrucomicrobiales 20 019 Daims et al. (1999)

ALF1b CGTTCGYTCTGAGCCAG Alpha-proteobacteria 20 019 Wagner et al. (1993)

BET42a GCCTTCCCACTTCGTTT Beta-proteobacteria 35 008 Yeates et al. (2003)

GAM42a GCCTTCCCACATCGTTT Gamma-proteobacteria 35 008 Yeates et al. (2003)

HGC69a TATAGTTACCACCGCCGT Actinobacteria 25 015 Roller et al. (1994)

*Percentage of formamide (%vv) in the hybridization buffer.

Molarity of sodium chloride in the wash buffer.




5/12

Statistical analysis

Statistical analyses were performed using Microsoft Excel

spreadsheet software including the Analysis Toolpak add-

in. The paired t-test was used to determine the effect of

VOE on process parameters and bacterial populations.

The level of significance was set at P < 5%, to differenti-

ate between the two sets of data. The Pearson product-

moment correlation coefficient (r) was used to investigate

the association between the EUB : DAPI ratio and

MLVSS concentration in each process. Analysis of vari-

ance (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 MLEprocesses are presented in Figs. 15. The domestic pro-

cess 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 aver-

age 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 con-

centration 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

(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

WW batch no.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

CODremoval(%)

75

80

85

90

95

100

105

Figure 1 Capacity of the domestic and industrial wastewater treat-

ment processes for the removal of organic material. There was a sta-

tistically significant difference in the organic removal capacities of the

processes (P < 0001%, Paired two-tailed t-test). (m) Domestic MLE

and (n) industrial MLE.

OUR

(mgOl

1h1)

0

10

20

30

40

50

60

WW batch no.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Figure 3 Oxygen utilization rates of the domestic and industrial

wastewater treatment processes. There was a statistically significant

difference between the measured oxygen utilization rates of the pro-

cesses (P < 0001%, Paired two-tailed t-test). (d) Domestic MLE and

() industrial MLE.

TKNrem

oval(%)

60

70

80

90

100

110

120

WW batch no.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Figure 2 Capacity of the domestic and industrial wastewater treat-

ment processes for the removal of nitrogen. There was no statistically

significant difference between the nitrogen removal capacities of the

processes (P = 776%, Paired two-tailed t-test). (.) Domestic MLE and

(,) industrial MLE.

A.P. Degenaar et al. Evaluation of the microbial community in wastewater

2007 The Authors



6/12

cells hybridizing to probe ALF1b accounted for 26% of

the cells hybridizing to EUBmix. Bacteria affiliated to the

beta- and gamma- subclasses of Proteobacteria repre-

sented 33% and 15% of cells detected by EUBmix respec-

tively. With probe HGC69a specific 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).

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

DAPI(%)

0

10

20

30

40

50

60

70

80

90

100

110

120

MLVSS(mgl1)

0

1000

2000

3000

4000

5000

6000

7000

8000

WW batch no.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

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 treat-

ment process. There was no correlation between EUB : DAPI ratios

and MLVSS concentrations of the domestic process (r = 068, Pearson

correlation coefficient). ( ) EUBDAPI and ( ) MLVSS.

DAPI(%)

0

10

20

30

4050

60

70

80

90

100

110

120

MLVSS(mgl1)

0

1000

2000

3000

4000

5000

6000

7000

8000

WW batch no.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

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 treat-

ment process. There was no correlation between EUB : DAPI ratios

and MLVSS concentrations of the domestic process (r = 037, Pearson

correlation coefficient). ( ) EUBDAPI and ( ) MLVSS.

EUBm

ix(%)

0

10

20

30

40

50

60

WW batch no.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Figure 6 Percentages of group-specific probes relative to EUBmix

counts in aerobic mixed liquor samples from the domestic wastewater

treatment process. There was a significant difference amongst the

percentages of the four group-specific probes within the domestic

process (P

< 0001%, anova single factor). ( ) ALF 1b; ( ) BET 42a;( ) GAM 42a and ( ) HGC 69a.

E

UBmix(%)

0

10

20

30

40

50

60

WW batch no.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Figure 7 Percentages of group-specific probes relative to EUBmix

counts in aerobic mixed liquor samples from the industrial wastewater

treatment process. There was a significant difference amongst the

percentages of the four group-specific probes within the industrial

process (P < 0001%, anova single factor). ( ) ALF 1b; ( ) BET 42a;

( ) GAM 42a and ( ) HGC 69a.




7/12

of the VOE using compound C40 removed 73% of the

COD. Research by Bettazzi et al. (2007) confirmed that

pretreatment using physico-chemical processes improved

the efficiency of biological treatment. Previous research

conducted by Sengul (1989) using coagulants such as

ploy-electrolytes when treating sunflower oil effluents,

showed COD removal rates of 76% and demonstratedpretreatment as a necessary step prior to biological treat-

ment. As shown in Fig. 1, the industrial process achieved

a mean COD removal capacity of 84%, which was signifi-

cantly lower than the domestic process which averaged

91% COD removal. The difference in COD removal

capacities between the processes could have been attrib-

uted 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 con-

firmed that VOEs such as olive mill WW comprise toxic

compounds, which negatively impact on biological treat-

ment. The COD removal capacity of the industrial pro-

cess support the work of other researchers such as

Mulligan and Sheridan (1975), who demonstrated the

capability of activated sludge to treat emulsified lipids

and provide removal efficiencies as high as 80%. Ozturk

et al. (1989) using a laboratory scale activated sludge pro-

cess to treat VOE, achieved a COD removal capacity of

72%. Mkhize et al. (2000) achieved 70% COD removal

using an anaerobicaerobic 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 findings sub-stantiated 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 influent was 250 mg l)1. In addition, other bio-

logical treatment processes using biomass rich in fungi

showed a COD removal capacity of 86% under aerobic

conditions (Caffaz et al. 2007). The TKN removal capac-

ity 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 deter-

mine 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 bear-ing probe conferred fluorescence are metabolically active;

hence, only active cells are counted. The EUB : DAPI

ratio was therefore used as a direct measure of the meta-

bolic 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 significantly

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 floc to become coated with hydro-

phobic material, thereby limiting oxygen transfer effi-

ciency 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 pro-

cess reflected 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 hybridiza-tions of each process were combined. As shown in Figs 4

and 5, Pearson correlation coefficient values of 037 and

068 were calculated for the domestic and industrial pro-

cesses, 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 there-

fore 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 signifi-

cantly 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 significantly greater

active biomass concentration of 3000 mg l)1 (P

was waster

Documents