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
8/22/2019 Was Waster
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
8/22/2019 Was Waster
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
Journal compilation 2007 The Society for Applied Microbiology, Journal of Applied Microbiology 104 (2008) 353363 355
8/22/2019 Was Waster
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.
Evaluation of the microbial community in wastewater A.P. Degenaar et al.
356 Journal compilation 2007 The Society for Applied Microbiology, Journal of Applied Microbiology 104 (2008) 353363 2007 The Authors
8/22/2019 Was Waster
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
Journal compilation 2007 The Society for Applied Microbiology, Journal of Applied Microbiology 104 (2008) 353363 357
8/22/2019 Was Waster
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.
Evaluation of the microbial community in wastewater A.P. Degenaar et al.
358 Journal compilation 2007 The Society for Applied Microbiology, Journal of Applied Microbiology 104 (2008) 353363 2007 The Authors
8/22/2019 Was Waster
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