ww.sciencedirect.com

wa t e r r e s e a r c h 6 6 ( 2 0 1 4 ) 1e1 1

Available online at w

ScienceDirect

journal homepage: www.elsevier .com/locate/watres

Comparison of biological and chemical treatmentprocesses as cost-effective methods for eliminationof benzoate in saline wastewaters

Martina Kiel, Daniel Dobslaw, Karl-Heinrich Engesser*

Institute of Sanitary Engineering, Water Quality and Solid Waste Management, University of Stuttgart, D-70569

Stuttgart, Germany

a r t i c l e i n f o

Article history:

Received 21 May 2014

Received in revised form

26 July 2014

Accepted 30 July 2014

Available online 7 August 2014

Keywords:

Benzoate

Elimination

Saline conditions

Waste water treatment

Chemical precipitation

Biodegradation

* Corresponding author. Department of BioWaste Management, University of Stuttgart,

E-mail address: karl-h.engesser@iswa.unhttp://dx.doi.org/10.1016/j.watres.2014.07.0450043-1354/© 2014 Elsevier Ltd. All rights rese

a b s t r a c t

Eight mixed cultures able to degrade benzoic acid under saline conditions were established

and kinetic parameters were determined in batch processes with cultures SBM002

(0.5 g d�1$g oDM�1), SBM003 (0.7 g d�1$g oDM�1) and SBM007 (2.2 g d�1$g oDM�1) showing

the highest degradation rates. Treatability of an industrial waste water (12 g L�1 benzoic

acid, 82 g L�1 NaCl) by these cultures was proven in a fed-batch system (SBM002 & SBM003)

and a continuous flow reactor (SBM007). The performance of the continuous flow reactor

was 15-times higher compared to the fed-batch system due to the change of inocula,

higher concentration of ammonia as nutrient and less accumulation of possibly toxic

catecholic compounds. Average DOC reduction was found to be 98% at 100 g L�1 NaCl and

1.2 g L�1 benzoic acid under these conditions. Pre-treatment of the waste water via

chemical precipitation by acidification to pH 3.5 diminished the concentration of benzoic

acid to 2.1 g L�1. In a combined chemical-biological process the volume of the bioreactor is

reduced to 15% compared to a pure biological process. A comparison of operational costs

for these three alternatives is presented.

© 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Benzoic acid is amajor product of the chemical industry being

used as starting material for chemical syntheses, artificial

flavors, insect repellents, food preservatives, personal care

products, medical products, antifreeze coolants and fungi-

cides (WHO, 2005; SCCNFP, 2002). Thus, it is an important

waste water component of speciality chemical industry (Li

et al., 1995). Furthermore, it is a pollutant in effluents of

logical Waste Air PurificaBandt€ale 2, D-70569 Stutti-stuttgart.de (K.-H. Enges

rved.

olive oil production and black ripe olive canneries (Novachis,

2005; GE, 2011). Such waste waters typically contain 5e9% of

NaCl and COD levels up to 220 g L�1 withmainlymethoxylated

and hydroxylated benzoic acids as contaminants (Etchells

et al., 1966; Zouari, 1998; Di Gioia et al., 2001; Benitez et al.,

2003; Fiorentino et al., 2004).

Waste water from chemical industry can contain similar

concentrations of NaCl and COD. For example, acid chlorides

are used during synthesis of aromaticealiphatic ketones by

Friedel-Crafts-Acylation, synthesis of esters via the Schotten-

tion, Institute of Sanitary Engineering, Water Quality and Solidgart, Germany. Tel.: þ49 0 711 685 63734; fax: þ49 0 711 685 63729.ser).

wat e r r e s e a r c h 6 6 ( 2 0 1 4 ) 1e1 12

Baumann process and within the production of aromatic

peroxides. In all three syntheses, sodium chloride is set free as

well as by-products of the reaction like benzoate in case of

benzoyl peroxide production. In the latter case, typical waste

water shows a concentration of up to 30 g COD L�1, equivalent

to 12 g L�1 of benzoic acid, and about 8% of NaCl.

Biological treatment of saline waste waters in general is

critical due to the high saline content inhibiting bioconver-

sion processes (Piubeli et al., 2012), requiring specialized

communities and treatment technologies. In contrast to non-

saline conditions, especially kinetic parameters for the

design of bio-treatment processes are sparsely found in

literature.

Only few bacterial species are known to use benzoate as

sole source of carbon and energy under saline conditions

containing more than 3.5% NaCl. Rosenberg (1983) reported a

Pseudomonas halodurans strain utilizing benzoate in media

containing up to 12% NaCl. Ventosa and Del Moral described

several benzoate degrading halophilic bacterial strains, but

gave no information about the concentrations of NaCl and

benzoate (Ventosa et al., 1982; Del Moral et al., 1988). Garcıa

et al. described Halomonas organivorans G-16.1 degrading up to

610 mg L�1 benzoate at 100 g L�1 NaCl (Garcıa et al., 2004).

Only one study dealing with the kinetics of saline benzoate

metabolism was published to date (Oie et al., 2007). Hal-

omonas campisalis DSMZ 15413 degraded 50 mg L�1$h�1 of

benzoate at 10% NaCl and a protein concentration of

25 mg L�1.

The purpose of this studywas to evaluate the treatability of

a hypersaline waste water of chemical industry highly

contaminated with benzoate, considering biological, chemical

and combined approaches. For biological treatment, conve-

nient bacterial communities and pure cultures were enriched

and characterized with special regard to high conversion rates

of benzoate and hence a compact design of a bio-treatment

plant. Due to highly discontinuous formation of waste

water, a batch process, a fed-batch process and a continuous

flow reactor were tested as three alternatives for a bio-

treatment process. Based on the low solubility of undissoci-

ated benzoic acid compared to sodium benzoate, chemical

Table 1 e Enrichment sources and conditions of mixed culture

Name Source

SBM001 Mixed sample from non-saline sewage sludge, soil and river

SBM002 Saltern near Col�onia de Sant Jordi on Mallorca, Spain

SBM003 Derived from SBM002

SBM004 Saltern “Salinas de Im�on” near Siguenza, Spain

SBM005 Saltern near Sant Jordi de ses Salines on Ibiza, Spain

SBM006 Saltern near Col�onia de Sant Jordi on Mallorca, Spain

SBM007 Saltern near Col�onia de Sant Jordi on Mallorca, Spain

SBM008 Saltern near Col�onia de Sant Jordi on Mallorca, Spain

SBP100 Isolated from SBM006

SBP110 Isolated from SBM002

SBP175 Isolated from SBM003

SBP310 Isolated from SBM001

Saltern samples used for enrichment were mixed samples of water and

culture designations translate as: SBM, salt tolerant, benzoate degrading

precipitation at low pH was examined as well. Finally, a sole

chemical process, a sole biological process and a combined

process were compared with particular regard to operational

costs.

2. Materials and methods

2.1. Waste water

The waste water from industrial application showed the

following parameters: COD, 30 g L�1; concentration benzoic

acid, 12e14 g L�1; concentration NaCl, 82 g L�1; pH value,

11.5e12; temperature, 25e35 �C; mass flow waste water,

2735 kg h�1; total volumetric flow, 21,500 m3 a�1; freight ben-

zoic acid, 32.5e37.5 kg h�1; operation time, 8000 h a�1.

2.2. Enrichment

Benzoate degrading mixed cultures were obtained from

samples of salterns, sediment and activated sludge (Table 1).

Cultivation conditions and sterile mineral medium with

additional 1.22 g L�1 benzoic acid and 5e10% NaCl were used

as described before (Dobslaw and Engesser, 2012). The pH

value was adjusted to 7.1 by addition of 2 M NaOH. Mixed

cultures were named SBM001 to SBM008.

For enrichment of single strains, mineral medium plates

containing 7% NaCl were used. Grown colonies were trans-

ferred to fresh media to yield pure strains. To eliminate dou-

bles, single strains were subjected to BOX-PCR and

comparison of the band-patterns in agarose gels (Martin et al.,

1992; Van Belkum et al., 1996). Pure cultures were named

SBP100, SBP110, SBP175 and SBP310 based on their enrichment

source (1: salterns; 3: activated sludge).

2.3. 16S rRNA gene sequencing

Genomic DNA of pure strains was obtained by thermal

cracking of cells and separation of cell fragments. 16S rRNA

geneswere amplified by PCR using the bacteria specific primer

s SBM001 to SBM008 and pure strains SBP100 to SBP310.

Cultured since Concentrations of

NaCl (%) Benzoate (g L�1)

sediment March 2010 5e10 1.22

March 2010 10 1.22

March 2010 10 2.14

September 2010 5e10 1.22

October 2010 10 1.22

October 2010 5e10 1.22

early June 2011 12 2.44

late June 2011 12 2.44

November 2010 7 1.22

November 2010 7 1.22

November 2010 7 1.22

November 2010 7 1.22

sediments from ponds of various salinities. The acronyms chosen as

mixed culture; SBP, salt tolerant, benzoate degrading pure strain.

wa t e r r e s e a r c h 6 6 ( 2 0 1 4 ) 1e1 1 3

27F (50-AGAGTTTGATCCTGGCTCAG-30) and the universal

primer 1492R (50-ACCTTGTTACGACTT-30). The amplicon was

purified using GenElute® PCR-CleanUpKit of SigmaeAldrich

with pure water as solvent and sequenced by GATC Biotech

AG. Closely matching sequences were determined using the

Ribosomal Database Project and NCBI BLAST.

2.4. TA cloning

Genomic DNA of mixed cultures was extracted by mechanical

cell disruption using a micro-MiniBeadbeater (BioSpec Prod-

ucts). PCR of 16S rRNA genes was carried out as described in

chapter 2.3. Ampliconswere ligated into a linear vector of pUC

series carrying resistances against ampicillin and kanamycin

and a lacZa genewith amultiple cloning site-operon using the

PCR Cloning Kit (Qiagen).

The plasmids were introduced into Escherichia coli JM109 by

electroporationemployingaMicroPulser (Bio-RadLaboratories).

After 1hof regeneration, cellswereplatedonLBagar containing

100 mg L�1 ampicillin, 30 mg L�1 kanamycin, 100 mg L�1 IPTG

and40mgL�1X-Galandincubatedat37 �Cfor24h.Arisingwhite

colonies were transferred to new plates and plasmids were rei-

solatedusing theGeneJETPlasmidMiniprepKit (Fermentas). 16S

rDNA inserts were sequenced using SP6 (50-ATTTAGGTGA-CACTATAGAA-30) and T7 (50-TAATACGACTCACTATAGGG-30)primers. In total, 20 transformants were evaluated.

2.5. Illumina sequencing of metagenomic 16S rDNA

Genomic DNA of mixed cultures was extracted by mechanical

cell disruption using a micro-MiniBeadbeater (BioSpec Prod-

ucts). The sequencing library was prepared by amplification of

partial 16S rRNA genes via PCR using 338R (50-GCTGCCTCCCGTAGGAGT-30) and barcode-tagged 27F primers.

Adaptor sequences for the Illumina flow cell, as well as a

second index sequence, were attached by two further PCR

runs. The library was immobilized on the flow cell surface,

amplified to yield local clusters of identical fragments and

sequenced. A total of 10552 sequences of 120 bp in lengthwere

obtained and evaluated using the Ribosomal Database Project.

Table 2 e Process parameters for batch cultures, fed-batch rea

Parameter Batch cultures

Dimensions (cm) 250 mL shaker flask

Operational volume (L) 0.05

Temperature (�C) 30

Inoculum SBM001 e SBM008; SBP100 e SBP310

Operation time (d) 1e14

Aeration rate (L min�1) Passive

Medium:

pH 6.5e10.0

NaCl concentration (g L�1) 50e160

Benzoate concentration (g L�1) 1.22e12.2

Dosage rate (mL h�1) e

OLR (mg benzoate L�1$h�1) e

HRT (d) e

F/M ratio (g COD d�1$g oDS�1) e

OLR: Organic Loading Rate; HRT: Hydraulic Retention Time; F/M Ratio: Fo

2.6. Substrate quantification

Concentrations of benzoic acid and metabolites were deter-

mined at 210 nm using a SpectraSeries HPLC-UV/VIS system

(Thermo Separation Products) with a ProntoSIL Eurobond C18

column (125 mm$4.0 mm, 5.0 mm particle size; Bischoff

Chromatography, Leonberg, Germany). As solvent a mixture

of methanol:water:H3PO4 ¼ 400:599:1 was used with a flow

rate of 1 mL min�1. DOC analyses of supernatants were per-

formed using a Sievers 900 Portable TOC Analyzer (GE

Water&Process Technology, Manchester, UK). All samples

were centrifuged (13.000 rpm; 10 min) and diluted before

measurement.

2.7. Chloride concentration

Concentration of chloride was analyzed using Metrohm 761

Compact IC ion chromatography system with Metrosep A

Supp4 (250 mm $ 4.0 mm) and Hamilton PRP-X110S (250 mm $

4.1 mm) as anion columns. As solvent a mixture of 106 mg L�1

Na2CO3 and 336 mg L�1 NaHCO3 was used with a flow rate of

1 mL min�1. All samples were diluted before measurement.

2.8. Benzoic acid degradation in batch cultures, fed-batch reactor and continuous flow reactor

For batch experiments, cultures harvested during exponential

phase were adjusted to an optical density at 546 nm (OD546) of

0.15e0.2 and defined concentrations of benzoic acid and NaCl

were established. OD546 and remaining concentrations of

benzoic acid were analyzed periodically by photometer and

HPLC. Experimentswere performed in triples at pH values and

concentrations of NaCl and benzoate as indicated in Table 2.

Conversion rates of benzoic acid were standardized to gram of

organic dry matter as biomass concentration (g oDM). The

highest value of each triple is presented in Table 3 for mixed

cultures SBM002, SBM003, SBM007 and all pure cultures. The

results for all cultures are provided in the Supplementary

Materials (Table S1).

Operation conditions for the fed-batch reactor are specified

in Table 2. When the maximum operational volume of 30 L

ctor and continuous flow reactor.

Fed-batch reactor Continuous reactor

Diameter: 28; Height: 53 Diameter: 4; Height: 40

5e30 0.2

15.5 ± 2.5 22.5 ± 2.5

Mixture of SBM002 and SBM003 SBM007

107 132

20 0.1

Influent:

7.0 7.0

100 100

12.2 1.22

2.5e220 6e30

3.9e177 36.6e183

2.9e129 0.28e1.4

0.01e1.47 0.87e2.61

od to Microorganism Ratio.

wat e r r e s e a r c h 6 6 ( 2 0 1 4 ) 1e1 14

was reached, 10e20 L was taken out of the system depending

on the concentrations of biomass and benzoate in the

effluent. OD546 and concentrations of benzoic acid aswell as of

degradation intermediates were analyzed two times per day.

pH, temperature, conductivity, air flow and O2 concentration

were measured continuously.

Operational parameters of the continuous flow reactor are

shown in Table 2 as well. The effluent was ejected by a GL18

thread at 300 mm height of the reactor and transferred to a

collecting basin for daily HPLC and pH analyses.

2.9. Chemical precipitation of benzoic acid

Elimination of benzoic acidwas achieved by pH shift aswell as

addition of FeCl3. In case of the pH shift, a 30 w% HCl solution

was used to adjust the pH value to 12.3; 12.0; 7.0; 6.0; 5.65; 5.0;

4.5; 4.3; 4.0; 3.8; 3.2 and 2.5. In latter case, FeCl3 concentrations

of 0; 0.25; 0.5; 0.75; 1.0; 1.5; 2.0 and 2.5 g L�1 were adjusted by

simultaneous addition of HCl solution and aliquots of a

100 g L�1 FeCl3 solution to the reaction vessel. The liquid phase

was stirred with 250 rpm. Tests were done twice and

remaining benzoic acid was quantified as described.

3. Results and discussion

3.1. Enrichment, kinetic parameters and composition ofthe communities in mixed cultures

Six mixed cultures (SBM001 to SBM006) able to grow in liquid

mineral salt media with benzoic acid as sole source of carbon

and energy and 100 g L�1 NaCl were enriched from different

habitats (non-saline sewage sludge vs. hypersaline brine) and

different locations (evaporation ponds of several salterns in

the Mediterranean). Since best results during the early stages

Table 3 e NaCl dependency of relative transformation rates ofSBM003 and SBM007 as well as pure cultures during batch cultspecified.

Name of culture ConcentrationNaCl (w%)

Max. growthrate m (d�1) 1.22

SBM002 5 2.19 ea

7.5 1.94 e

10 1.30 0.30

SBM003 10 2.02 0.82

12 0.60 0.36

14 0.10 0.06

16 0.02 0.06

SBM007 7.5 3.72 1.00

10 4.17 2.20

12 1.46 0.87

SBP100 10 4.29 2.00

SBP110 10 3.29 1.35

SBP175 10 4.41 1.49

SBP310 10 2.51 0.99

Bold type in columns 4e10 indicates the benzoic acid start concentration

shown in column 3.a Not analyzed.

of the investigation were obtained with cultures enriched

from a saltern on Mallorca, two more cultures (SBM007 and

SBM008) were enriched from this site the following year, using

higher concentrations of benzoate and NaCl (Table 1). All

cultures enriched in spring-time showed significantly higher

growth and degradation rates compared to the culture

enriched in autumn from the same saltern (Table S1). Saltern

pond communities have been shown to undergo substantial

changes during the course of the year even if salinity stays

nearly constant (Boujelben et al., 2012). The phenomenon is

mostly attributed to variations in temperature and may have

affected community structure and degradation efficiency in

cultures enriched at different points in time.

At first, batch cultures exhibited poor degradation kinetics

(maximum transformation rates of 0.2 g d�1$g oDM�1) and

reproducibility of growth, but revealed high diversity of mi-

croorganisms (see Supplementary Materials, Figure S1). Dur-

ing repeated sub-cultivation over a minimum of 8 months,

biodiversity declined, whereas reproducibility of growth and

kinetic parameters significantly increased (0.3 g d�1$g oDM�1

after five months and 2.2 g d�1$g oDM�1 after 12 months for

SBM007). The composition of the community of SBM007 was

analyzed in detail after four months of use in a continuous

flow reactor (12 months after isolation), assuming stable bio-

conditions by this time. About 80.4% of the reads obtained

by Illumina sequencing of 16S rRNA genes were assigned to

bacterial taxa. The composition of the community was (frac-

tion of community in %/similarity in %): Halomonas sp. (62.4/

100.0); Idiomarina sp. (7.4/100.0); uncultured a-Proteobacteria

(4.6/68.1); Bacillus sp. (3.6/100.0) and Nitratireductor sp. (1.2/

100.0). The missing 19.6% are assumed to predominantly

consist of Bacteriovorax sp. based on TA Cloning results (4 of 20

clones). Remaining cloned gene sequences largely matched

the results of Illumina sequencing. All 20 amplicons of 16S

rRNA gene sequences had a length between 1095 and 1494

benzoic acid (g d¡1·g oDM¡1) for mixed cultures SBM002,ivation at 30 �C and start concentrations of benzoic acid as

Initial concentration of benzoic acid (g L�1):

2.44 3.66 4.88 6.11 9.16 12.21

Benzoic acid conversion rates (g d�1$g oDM�1):

1.48 1.36 e e e e

1.38 0.82 e e e e

0.45 0.35 0.17 0.14 0.03 0.01

0.66 0.61 0.55 0.53 0.08 0.00

0.35 e e e e e

0.00 e e e e e

e e e e e e

1.00 1.09 0.80 0.91 e e

1.82 1.66 1.33 1.29 e e

0.79 0.40 0.74 0.21 e e

e e e e e e

e e e e e e

e e e e e e

e e e e e e

s and conversion rates corresponding to the maximum growth rates

wa t e r r e s e a r c h 6 6 ( 2 0 1 4 ) 1e1 1 5

base pairs with an identity of mainly 99% (12 of 16 amplicons)

and 97%, 95%, 93% and 92% in case of Halomonas sp. The four

amplicons of Bacteriovorax sp. showed identities of 94e95% to

subspecies GSL371.

Bacteriovorax sp. is known as a bacterial predator parasit-

izing other Gram-negative bacteria and devouring its prey.

Richards et al. (2012) described a diminution of Vibrio sp.

counts by 3 log units within 48 h in natural seawater by Vibrio

predatory bacteria like Bdellovibrio sp., Bacteriovorax sp. or

Micavibrio sp. A recent work by Chen et al. (2012) showed an

increase of Bacteriovorax sp. counts by 4e7 log units within

48 h usingVibrio cholerae, Pseudomonas putida and E. coli as prey.

Still representing approximately 20% of cells in a biocoe-

nosis recultivated over 12 months and especially 4 months in

a continuous flow reactor, this genus was likely to be

responsible for poor reproducibility of growth during the first

cultivation steps in all mixed cultures. Despite presence of

Bacteriovorax sp., mixed cultures were usable for operation of

bioreactors after ongoing subcultivation.

3.2. Enrichment of pure strains

In order to further increase reproducibility of growth by

elimination of predators, the enrichment of pure strains was

aspired. Four pure bacterial strains named SBP100, SBP110,

SBP175 and SBP310, which were predominant in these mixed

cultures, were obtained and revealed high degradation po-

tential up to 2.0 g d�1$g oDM�1 of benzoic acid. Strains SBP100,

SBP110 and SBP175 were enriched out of a saltern, whereas

SBP310 was obtained out of an activated sludge sample. Ac-

cording to 16S rRNA gene sequences, strains SBP100 and

SBP175 were closely related and differed in only 8 bp with no

gaps. However, both strains showed significantly different

DNA fingerprints using BOX-PCR (Martin et al., 1992; Van

Belkum et al., 1996) and SBP175, as opposed to the other

three strains, was able to mineralize benzene and phenol

(Table 4). The phylogenetic tree of SBP100, SBP110, SBP175 and

Table 4 e Characterization of obtained pure strains with respecsubstrates as sole source of carbon and energy (þ, growth; -, n

Strain SBP100

Partial 16S-rRNA gene sequence (in bp) 1067 10

Identity (in %) 99.3 99

Closest match Halomonas organivorans

G-16.1

H

D

Motility positive po

NaCl growth range (in LB medium) 2e30% 2e

NaCl growth range (in mineral medium

with benzoic acid, pH 7.1)

2e24% 2e

Growth on:

Salicylic acid þ þ3-Hydroxybenzoic acid þ e

4-Hydroxybenzoic acid þ þ2-Methoxybenzoic acid e e

4-Methoxybenzoic acid e e

Vanillic acid þ e

Phenol e e

Toluene e e

Benzaldehyde þ þBenzyl alcohol þ þ

other important Halomonas strains is presented in the Sup-

plementary Materials (Figure S2). Corresponding 16S rRNA

gene sequences were deposited in GenBank and are listed

under accession numbers KF500536 to KF500539. The strains

grew on various hydroxylated and methoxylated benzoic

acids as well as benzyl alcohol and benzaldehyde, but not on

toluene. The potential to degrade a large spectrum of aliphatic

and aromatic substrates is typical for Halomonas sp. (Garcıa

et al., 2004; Oie et al., 2007; Mnif et al., 2009) and promising

for treatment of further saline waste waters as cited in the

introduction, e.g. effluents from the olive industry.

3.3. Effect of the pH value

The pH value of the medium strongly influences the conver-

sion rates of benzoic acid as well as the growth rates of the

mixed cultures. Conversion of the substrate preferably should

take place at alkaline conditions as the waste water of the

industrial application described is highly alkaline. Different

strains described in literature are able to degrade aromatic

compounds in presence of NaCl at concentrations up to

310 g L�1 and pH values up to 9.5 (Ventosa et al., 1982;

Rosenberg, 1983; Emerson et al., 1994; Del Moral et al., 1988;

Di Gioia et al., 2001; Fiorentino et al., 2004; Garcıa et al., 2004;

Oie et al., 2007). In difference to literature data, conversion

rates of benzoic acid significantly declined when the pH value

was 8.5 or higher in all mixed cultures. Substrate conversion

and growth stopped when pH values fell below 6.5. The pH

optimum of all tested cultures was around 7.5 ± 0.4. Within

this range, kinetic parameters of batch processes, the fed-

batch process and the continuous flow reactor were nearly

independent of the pH value (data not shown).

3.4. Conversion of benzoic acid in batch processes

Resultsof conversionexperimentsofbenzoicacidare shownin

Tables 3 and S1. Concentrations of benzoic acid andNaCl were

t to NaCl tolerance and ability to grow on various aromatico growth).

SBP110 SBP175 SBP310

19 1348 1041

.7 100.0 99.7

alomonas halophila

SM 4770

Halomonas organivorans

G-16.1

Dietzia sp. M1T8B24

sitive positive negative

22% 2e28% 0e13%

19% 2e15% 0e13%

þ e

þ e

þ þe e

e þþ e

þ e

e e

þ þþ þ

wat e r r e s e a r c h 6 6 ( 2 0 1 4 ) 1e1 16

systematically varied. Conversion rates were highest at

1.22 g L�1 of benzoic acid, started to decline sharply at con-

centrations between 3.66 and 4.88 g L�1 of benzoic acid and

finally reached values below 5%ofmaximumrates at 12.2 g L�1

of benzoic acid. Likewise, increasing concentrations of NaCl

resulted in declining conversion rates of benzoic acid. Growth

rates showed a similar concentration dependency as conver-

sionrates.Allmixedcultureswereable toconvert thesubstrate

at NaCl concentrations of 100 g L�1, meeting the requirements

of the industrial waste water. Of the cultures enriched in 2010,

SBM002 and SBM003 showed the maximum growth and con-

version rates at 10%NaCl andwere thus used as inocula for the

fed-batch reactor. In 2012, rates more than twice as high were

determined in culture SBM007. For this reason, SBM007 was

chosen as inoculum for the continuous reactor.

During benzoate conversion in batch cultures, HPLC peaks

representing metabolites rarely exceeded the detection

threshold. Occasionally, small amounts of cis,cis-muconate

(<70 mg L�1) accumulated in the medium, but were further

metabolized after consumption of the substrate.

All pure strains excreted cis,cis-muconate as well as cate-

chol when exposed to high concentrations of benzoate

(>2.5 g L�1). These compounds are known as intermediates of

a degradation pathway in which benzoate is dioxygenated to

form a dienediol and, after decarboxylation and rear-

omatization, catechol (Reiner and Hegeman, 1971). The aro-

matic ring is then cleaved in ortho position, producing cis,cis-

muconic acid (Harwood and Parales, 1996). Besides its occur-

ence in many non-halophilic bacteria, this pathway has been

proposed for benzoate degradation by the haloalkaliphile

Halomonas campisalis (Oie et al., 2007). An overview of further

aerobic benzoic acid degradation pathways is given by

Hammann and Kutzner (1998).

0

5

10

15

20

25

30

0 10 20 30 40 50Conc

entra

tion

Benz

oic

acid

real

& c

umul

ativ

e, O

rgan

ic d

ry

mat

ter,

Reac

tor v

olum

e

Duratio

Volume reactor (L)Real concentration benzoic acid (g/L)Feed rate (mL/h)

Fig. 1 e Process conditions in a fedebatch process with an ope

concentration of benzoic acid without bacterial degradation wa

influent (filled triangles). The difference between this line and t

measure for biological activity. During the first 20 days of opera

final value of 10%. The reactor was partially emptied on days 1

3.5. Conversion of benzoic acid in a fed-batch process

Results of an operation period of 107 days are presented in

Fig. 1. In general, the concentration of benzoic acid within the

reactor was limited to 3 g L�1 and thus used as control

parameter for dosage of the influent. Additionally, the hypo-

thetical cumulative concentration of benzoic acid over time

without biological conversion is given in this figure. The dif-

ference between both curves represents the amount of ben-

zoic acid converted. Four phases can be distinguished: an

initial phase (1st e 14th day) and three process phases (15th e

48th day, 48th e 89th day, 89th e 107th day).

During the initial phase the liquid volume and the con-

centration of organic dry matter were increased from 5 to 12 L

and 0.8e15 g oDM L�1, respectively. The conversion rate of

benzoic acid was about 0.176 g d�1$g oDM�1. The first process

phase started when a biomass concentration of 15 g oDM L�1

was reached. During this phase an average conversion rate of

benzoic acid of 0.031 g d�1$g oDM�1 was observed. Process

phase 1 ended after 48 days achieving a volume of 30 L. After

removal of 10 L, the second process phase started. During this

phase conversion rates of benzoic acid proceeded to decline

gradually resulting in an average conversion rate of

0.003 g d�1$g oDM�1, hence 10 times lower than the results

during the first process phase. One reason for this behavior

was found in a limitation of nitrogen, although

0.68e60.2 mg h�1 of ammonia were supplied. After enhance-

ment of ammonia at days 89 and 99, causing concentrations of

15 mg N L�1, conversion rates increased 20e23 folds to

0.062e0.069 g d�1$g oDM�1 in phase 3.

Compared to batch experiments with SBM002 and

SBM003, maximum rates observed during operation of the

fed-batch system were lower by a factor of 5e12. In part, this

0

40

80

120

160

200

240

60 70 80 90 100 110

Feed

rate

n (d)

Cumulative concentration benzoic acid (g/L)Organic dry matter (g oDM/L)

ration time of 107 days. A hypothetical cumulative

s calculated from the feed rate and concentration of the

he measured real concentration (unfilled triangles) is a

tion, the NaCl concentration was elevated from 6% to the

5, 48, 89, and 99.

wa t e r r e s e a r c h 6 6 ( 2 0 1 4 ) 1e1 1 7

difference can be explained by the higher temperature in

batch processes (see Table 2). As a second reason for

temporarily reduced kinetic parameters, the accumulation of

2,3-dihydroxybenzoic acid as a metabolite was identified. Its

concentration exceeded 1.0 g L�1 during the first three phases

of the reactor's run and culminated in 3.0 g L�1 on day 51.

Addition of ammonia during the fourth phase resulted in a

decline of the 2,3-dihydroxybenzoic acid concentration below

the detection limit. Catecholic compounds are known to

exert toxic effects on microorganisms (Schweigert et al.,

2001). 2,3-Dihydroxybenzoic acid has been described as a

strong inhibitor for salicylate synthase (Manos-Turvey et al.,

2010). Members of this protein family are associated with

biosynthesis of siderophores and other secondary metabo-

lites (Kerbarh et al., 2005). Unlike other dihydroxybenzoates,

the 2,3 derivative is no common intermediate of benzoate

biodegradation. To date, only one study observed its forma-

tion from benzoic acid by a Bacillus strain (Spokes and

Walker, 1974).

The composition of mixed cultures SBM002 and SBM003

used for operation of the fed-batch system has not been

examined in detail. However, they are likely to contain similar

predatory species as SBM007 due to the same place of origin of

these three cultures (see Table 1). Since a fed-batch system

allows nowashing out of predators, Bacteriovorax sp.may have

accumulated and strongly influenced the concentration of

benzoic acid degrading bacteria. The concentration of dry

matter in the fed-batch process stayed nearly constant over

time, indicating a shift in the community to degraders with

both poor kinetic parameters and high predation resistance.

Adding ammonia to the system helped to increase the con-

centration of effective degraders like Halomonas sp., but they

still formed a minor fraction.

As a consequence a fed-batch system seems to be not

suitable for the industrial application described.

0

1

2

3

4

5

6

7

40 50 60 70 80

Ben

zoic

aci

d co

ncen

tratio

ns fe

ed &

out

let,

Org

anic

dry

mat

ter,

Dilu

tion

rate

, Deg

rada

tion

rate

Duratio

Benzoic acid concentration feed (g/L)Organic dry matter (g oDM/L)Benzoic acid degradation rate (g/(d*g oDM))

Fig. 2 e Operation conditions in continuous flow reactor of mix

maximum dilution rate without washout, feed rate was gradual

concentration as well as benzoate concentration in influent and

phenomenon occurred.

3.6. Conversion of benzoic acid in a continuous flowreactor and washout phenomena

On the basis of previous experiments and in difference to the

original concentration of 12 g L�1 of benzoic acid in the waste

water, an inlet concentration of 1.22 g L�1 of benzoic acid is an

appropriate concentration for operation of the continuous

flow reactor. A reactor of this type was operated for 132 days,

divided into two process phases. In a first phase of 40 days,

dilution rate and concentration of benzoic acid in the influent

were varied frequently to simulate differing amounts and

compositions of waste water over time and for approximate

adjustment of process parameters. The system showed stable

process conditions at a dilution rate of 2.88 d�1 and a washout

phenomenon at 3.60 d�1 (data not shown).

During the second phase (days 40e132), fine adjustments

of process parameters took place. The maximum growth rate

of the mixed culture SBM007 was determined by gradually

increasing the dilution rate in steps of 0.36 d�1 and 0.72 d�1,

respectively, at a constant feed concentration of 1.22 g L�1 of

substrate (Fig. 2). Concentrations of benzoic acid and biomass

as well as the pH value were monitored and showed stable

behavior up to 2.88 d�1 as a dilution rate, representing a

degradation rate of 0.9 g d�1$g oDM�1 of benzoic acid.

At a dilution rate of 3.24 d�1, foam formation and genera-

tion of bacterial agglomerates was observed. However, con-

centration of biomass after homogenization stayed nearly

stable and benzoate was still almost completely degraded.

During days 60e67 of operation, an increase in the dilution

rate to 3.60 d�1 resulted in a critical loss of degradation per-

formance as well as biomass concentration in the reactor due

to coagulation and washout of the microorganisms. There-

fore, avoiding cell agglomeration and foam formation, a

dilution rate of 2.88 d�1 corresponding to a residence time of

the waste water of 8.3 h was assumed to be the optimal

90 100 110 120 130

n (d)

Benzoic acid concentration outlet (g/L)Dilution rate (1/d)

ed culture SBM007 (test phase 2). For determination of the

ly increased and accompanied by measurement of biomass

effluent. At a dilution rate of 3.60 d¡1 a washout

Table

5e

Conce

ntrationofbenzo

icacid/benzo

ate

rem

ain

ingin

liquid

phase

afteradditionof30w%

HClso

lutionand10w%

FeCl 3,resp

ectively.a

FeCl 3-d

osa

ge

(ingL�1)

00.1

0.25

0.5

0.75

11.5

22.5

pH

Start

Conc.

(ingL�1)

pH

Conc.

(ingL�1)

pH

Conc.

(ingL�1)

pH

Conc.

(ingL�1)

pH

Conc.

(ingL�1)

pH

Conc.

(ingL�1)

pH

Conc.

(ingL�1)

pH

Conc.

(ingL�1)

pH

Conc.

(ingL�1)

pH

12.3

12.7

±0.1

12.3

12.0

12.4

±0.2

12.0

7.0

12.1

±0.1

7.07

11.7

±0.1

6.64

12.0

±0.0

5.90

12.7

±0.1

5.58

10.1

±0.0

5.20

11.0

±0.0

5.00

10.6

±0.1

4.76

8.9

±0.1

4.53

8.2

±0.1

4.48

6.0

12.6

±0.0

5.97

12.6

±0.1

5.62

12.4

±0.0

5.50

12.2

±0.1

5.10

11.2

±0.0

4.99

11.1

±0.0

4.86

10.8

±0.1

4.67

9.4

±0.0

4.52

7.8

±0.0

4.45

5.65

12.3

±0.2

5.65

5.0

12.5

±0.0

4.99

12.4

±0.1

4.96

12.1

±0.0

4.89

11.1

±0.0

4.68

9.7

±0.1

4.61

8.7

±0.1

4.58

8.0

±0.0

4.48

8.0

±0.0

4.47

7.3

±0.0

4.37

4.5

11.0

±0.2

4.52

11.3

±0.2

4.52

11.2

±0.2

4.52

10.3

±0.2

4.51

8.4

±0.2

4.47

8.3

±0.2

4.46

7.7

±0.1

4.37

7.1

±0.1

4.30

6.1

±0.0

4.22

4.2

5.8

±0.1

4.20

5.4

±0.2

4.16

5.1

±0.2

4.12

4.8

±0.1

4.05

4.6

±0.1

3.97

4.2

±0.1

3.85

3.3

±0.0

3.67

1.9

±0.0

2.90

1.9

±0.0

1.98

4.0

4.5

±0.1

4.00

4.7

±0.1

4.01

4.5

±0.1

3.96

4.6

±0.0

3.93

3.2

±0.0

3.63

2.1

±0.0

3.49

1.9

±0.0

2.58

1.9

±0.0

2.09

1.9

±0.0

1.81

3.8

4.0

±0.1

3.80

3.2

2.0

±0.1

3.20

2.5

1.9

±0.0

2.50

1.9

±0.0

2.41

1.9

±0.0

2.26

1.9

±0.0

1.57

1.8

±0.0

1.57

1.8

±0.0

1.57

1.8

±0.0

1.56

1.8

±0.0

1.57

1.8

±0.0

1.56

aFinalco

nce

ntrationofFeCl 3

isprese

ntedin

thefirstline,whilefinalpH

valuesare

prese

ntedin

each

unevenco

lumnfrom

3rd

to19th

column.

wat e r r e s e a r c h 6 6 ( 2 0 1 4 ) 1e1 18

operation point, providing both system stability and high

degradation rates.

After washout the reactor was reinoculated at day 76 and

operated for further 56 days. During the start-up period (days

76e82) the dilution rate was increased stepwise by 0.72 d�1.

The degradation performance and concentration of biomass

proved stable as long as the dilution rate did not exceed the

previously determined rate of 2.88 d�1, verifying it as the

maximum tolerable rate.

With the exception of an accumulation of cis,cis-muconate

(120 mg L�1) on day 77, neither benzoic acid nor metabolites

were detectable by HPLC during stable operation conditions.

The average reduction of DOC in the effluent was about 98%.

As a consequence of the conversion of benzoic acid (presented

as sodium benzoate under these test conditions) the pH value

shifted from 7.0 in the influent to 7.2e7.4 in the effluent.

Predatory effects of Bacteriovorax sp. appeared to be negligible

under these conditions. Compared to maximum trans-

formation rates in batch cultures of SBM007, conversion rates

were lower by a factor of 2.5. As before, the temperature dif-

ference between batch experiments and the continuous cul-

ture (see Table 2) has to be considered.

In contrast to the conversion rates of batch processes

representing short-time performance, the conversion rates

of the continuous flow reactor are average values of a long

operation phase. Thus, these kinetic parameters are more

reliable for the design of an industrial scale treatment

plant than those of batch processes. A continuous flow

reactor preferably with an operational flow rate higher

than the growth rate of Bacteriovorax sp. seems very

promising to wash out unwanted intermediates as well as

predators.

3.7. Chemical precipitation of benzoic acid

The precipitation of benzoic acid as well as benzoate, both out

of an organic phase and an aqueous phase, is a well estab-

lished industrial scale process. Exemplarily, benzoic acid can

be precipitated from an organic phase like benzene, toluene or

xylene via reaction with alkoxides, preferably methoxide

(Tedball, 1975), via addition of water as an antisolvent in case

of an ethanolic solution containing high concentrations of

benzoic acid (Ferguson et al., 2012), or via pH shift. In com-

parison to sodium benzoate (660 g L�1, 20 �C) the solubility of

benzoic acid is significantly lower (2.7 g L�1, 20 �C). Thus,

elimination of benzoate by acidification via addition of 30 w%

HCl and/or 10 w% FeCl3 is an effective elimination procedure.

In case of FeCl3 dosage, ferric ions capture hydroxyl ions by

precipitation of Fe(OH)3. The remaining protons shift pH value

to acidic conditions fortifying a co-precipitation of benzoic

acid (Table 5). Acidification correlated with reduced concen-

trations of benzoic acid as long as the pH valuewas equal to or

higher than 3.5 (equivalent to 2.1 g L�1 benzoic acid). Further

lowering of the pH value had no significant effect anymore.

3.8. Comparison of size and operational costs of thetreatment plant

The waste water of the industrial site presented is currently

treated by a municipal waste water treatment plant. For

wa t e r r e s e a r c h 6 6 ( 2 0 1 4 ) 1e1 1 9

treatment of yearly 21,500 m3 of water (V) a heavy pollution

charge (HPC) has to be paid. According to equation (Eq. (1)) the

total costs of HPC are 250,000 V per year. The correlation co-

efficient between the concentration of benzoic acid (in g L�1)

and the COD value (in mg L�1) is 1965.

HPC ðin VÞ ¼ V$�ðCOD� 1200Þ$1200�1$0;95$0; 51

�(1)

However, current results show a high treatment capability

via a biological, chemical or combined process. A detailed

comparison of the operational costs of these three alterna-

tives is presented in Table 6, showing highest saving potential

for the combined process.

3.8.1. Biological processData revealed best performance for a continuous flow reactor

system. During operational conditions of 1.22 g L�1 substrate

and 4e5 g oDM L�1, a conversion rate of 0.9 g d�1$g oDM�1 of

benzoic acid was observed. Hence, for conversion of

37.5 kg h�1 of benzoic acid as freight under the same condi-

tions a reactor volume of 190 m3 is required. While operating

an aerated sludge reactor, costs for electricity (pumps, aera-

tion, mechanical dewatering), nitrogen supply (i.e. urea;

100 mg N L�1) and sludge deposition occur. The HPC is

omitted.

3.8.2. Chemical processThe precipitation of benzoic acid or benzoate both out of an

organic phase and an aqueous phase is a well established

industrial scale process (Tedball, 1975; Ferguson et al., 2012).

At last, the pH shift to acidic conditions is the most suitable

method.

The pH shift needed for chemical precipitation can be

achieved by adding solutions of both HCl and FeCl3. The

former reagent is preferable due to half the costs of purchase

for the same proton concentrations and the ability to recover

benzoic acid in high quality. Precipitation took place within

Table 6 e Comparison of operational costs of three different trecost reference.a

Situation of today Biologic

Heavy pollution charge 250,000 V e

Electricity for aerationb e 44,400

Electricity for mechanical filtrationb e 6950 V

Sludge depositionc e 21,500e

Acidification (pH 3.5)d e e

Neutralizationd,e 22,900 V 23,900

Demand on nitrogenf e 1000 V

Sum 273,000 V 97,800e

Saving 0 V 154,000

a Costs based on the following costs of purchase in Euro per metric ton: 3

electrical energy, 0.086 V.b Demand of electric energy for aeration and mechanical press based o

0.80 kWh kg COD�1 for aeration and 0.125 kWh kg COD�1 for filtration.c Assumption: Precipitate of benzoic acid is reusable in industrial proce

(BMU, 2006).d Costs based on a titration curve with a demand of 7.4 mL of 36 w% HCl s

neutralization was reduced due to lower concentrations of benzoic acid.e Discharge requirement for waste water into sewage system: pH betweef Addition of 7.14 g urea per kg COD based on experimental results.

seconds and the concentration of benzoic acid in the super-

natant stayed constant for hours (data not shown). Thus, the

required reactor volume is specified by storage demands of

the pre-treated waste water before mechanical filtration

rather than kinetic parameters of precipitation. During oper-

ation process costs for electricity (pumps, mechanical dew-

atering), chemicals (acidification, partial neutralization) and

optional deposition of benzoic acid occur. Due to an outlet

concentration of 2.1 g L�1 of benzoic acid (equivalent to 4.1 g

COD L�1), the limit value of 1200 mg COD L�1 to omit HPC is

still exceeded. Remaining HPC declines to almost 25,000 V.

3.8.3. Combined processCombining both a chemical pre-treatment and a biological

treatment of residual benzoic acid is a cost effective method.

Chemical pre-treatment reduces concentration and freight of

benzoic acid by 85% to 2.1 g L�1 and 5.75 kg h�1, respectively.

Biological treatment of remaining benzoic acid is facilitated by

neutralization of the chemical stage effluent. In comparison to

a chemical process with subsequent neutralization, opera-

tional costs (without HPC) of the combined process increase

by 20% due to higher requirements in neutralization and the

biological process as a third stage. Costs of the additional

biological step were estimated assuming a reduction in

reactor volume by 85% compared to biological treatment alone

(see chapter 3.8.1.). HPC is omitted.

4. Conclusion

� An efficient bacterial removal of benzoic acid under con-

ditions of 100 g NaCl L�1 was demonstrated. Thus, dilution

of contaminated waste water for reduction of saline con-

centration was not required.

� Both conversion rate and process stability were higher in a

continuous flow reactor system than in a batch or fed-

atment alternatives described and the situation of today as

al treatment Chemical precipitation Combined process

25,200 V e

V e 6050 V

6000 V 6950 V

43,000 V e 3550e7100 V

54,000 V 54,000 V

V 7900 V 8650 V

e 150 V

119,000 V 93,100 V 79,400e82,900 V

e175,000 V 180,000 V 190,000e194,000 V

6 w% HCl solution, 150 V; 50 w% NaOH, 330 V; urea solid, 215 V; kWh

n 8-years average values of the WWTP, University of Stuttgart, with

ss. Disposal cost of the remaining sludge per metric ton: 150e300 V

olution to get a pH of 3.5. The demand of NaOH in case of subsequent

n 5.0 and 9.0.

wat e r r e s e a r c h 6 6 ( 2 0 1 4 ) 1e1 110

batch process. The increase of both parameters was ob-

tained a) by reducing potentially toxic intermediates like

2,3-dihydroxybenzoic acid as well as predatory bacteria

like Bacteriovorax sp. by wash-out and b) improved supply

of nutrients.

� Chemical precipitation via pH shift to acidic conditions is

an efficient method for bulk elimination of benzoic acid as

solubility of benzoic acid is two log units lower than the

solubility of sodiumbenzoate. Acidification correlatedwith

reduced concentrations of benzoic acid as long as the pH

value was equal or higher than 3.5. Further lowering of the

pH value had no significant effect anymore. However, the

concentration of residual dissolved benzoic acid was still

some grams per liter, which restricts the applicability of

chemical precipitation to pre-treatment only.

� A combination therefore of a primary chemical and a sec-

ondary biological stage for treatment of benzoic acid

proved to be the method of choice. The performance of the

biological stage was improved by neutralization of the

chemically pre-treated waste water. As a final result, waste

water only slightly contaminated could be treated by a

municipal waste water treatment plant.

Acknowledgment

We are very grateful for the continuous support, help and

advice from Prof. J. Lalucat (Universidad de las Islas Baleares)

and Dr. R. Rossell�o-M�ora (CSIC Departamento Ecologıa y

Recursos Marinos at Esporlas, Mallorca). We further want to

thank Dr. D.H. Pieper from the HZI, Braunschweig, for Illu-

mina analyses; Dr.-Ing. N. Strunk for his advice in TA cloning

and Dipl.-Ing. J. Hartmann for assistance with the fed-batch

process.

Appendix A. Supplementary data

Supplementary data related to this article can be found at

http://dx.doi.org/10.1016/j.watres.2014.07.045.

r e f e r e n c e s

Benitez, F.J., Acero, J.L., Leal, A.I., 2003. Purification of storagebrines from the preservation of table olives. J. Haz. Mat. 96(2e3), 155e169. http://dx.doi.org/10.1016/S0304-3894(02)00183-8.

BMU, 2006. Kl€arschlammmengen und Entsorgungskosten imVergleich zum europ€aischen Ausland. In: FachtagungKl€arschlamm des Emscher Genossenschaft Lippe Verband,Bonn, 6./7.12.2006. BMU - Bundesministerium fur Umwelt,Naturschutz und Reaktorsicherheit, Berlin.

Boujelben, I., Gomariz, M., Martınez-Garcıa, M., Santos, F.,Pe~na, A., L�opez, C., Ant�on, J., Maalej, S., 2012. Spatial andseasonal prokaryotic community dynamics in ponds ofincreasing salinity of Sfax solar saltern in Tunisia. Ant. VanLeeuwenhoek 101 (4), 845e857. http://dx.doi.org/10.1007/s10482-012-9701-7.

Chen, H., Young, S., Berhane, T.-K., Williams, H.N., 2012.Predatory Bacteriovorax communities ordered by various preyspecies. PLoS ONE 7 (3), e34174. http://dx.doi.org/10.1371/journal.pone.0034174.

Del Moral, A., Prado, B., Quesada, E., Garcıa, T., Ferrer, R., Ramos-Cormenzana, A., 1988. Numerical taxonomy of moderatelyhalophilic Gram-negative rods from an inland saltern. J. Gen.Microbiol. 134, 733e741.

Di Gioia, D., Bertin, L., Fava, F., Marchetti, L., 2001. Biodegradationof hydroxylated and methoxylated benzoic, phenylacetic andphenylpropionic acids present in olive mill wastewaters bytwo bacterial strains. Res. Microbiol. 152 (1), 83e93. http://dx.doi.org/10.1016/S0923-2508(00)01171-2.

Dobslaw, D., Engesser, K.-H., 2012. Degradation of 2-chlorotolueneby Rhodococcus sp. OCT 10. Appl. Microbiol. Biotechnol. 93 (5),2205e2214. http://dx.doi.org/10.1007/s00253-011-3543-5.

Emerson, D., Chauhan, S., Oriel, P., Breznak, J.A., 1994. Haloferaxsp. D1227, a halophilic Archaeon capable of growth onaromatic compounds. Arch. Microbiol. 161 (6), 445e452. http://dx.doi.org/10.1007/BF00307764.

Etchells, J.L., Borg, A.F., Kittel, I.D., Bell, T.A., Fleming, H.P., 1966.Pure culture fermentation of green olives. Appl. Microbiol. 14(6), 1027e1041.

Ferguson, S., Morris, G., Hao, H.X., Barrett, M., Glennon, B., 2012.In-situ monitoring and characterization of plug flowcrystallizers. Chem. Eng. Sci. 77, 105e111. http://dx.doi.org/10.1016/j.ces.2012.02.013.

Fiorentino, A., Gentili, A., Isidori, M., Lavorgna, M., Parrella, A.,Temussi, F., 2004. Olive oil mill wastewater treatment using achemical and biological approach. J. Agric. Food. Chem. 52 (16),5151e5154. http://dx.doi.org/10.1021/jf049799e.

Garcıa, M.T., Mellado, E., Ostos, J.C., Ventosa, A., 2004. Halomonasorganivorans sp. nov., a moderate halophile able to degradearomatic compounds. Int. J. Syst. Evol. Microbiol. 54,1723e1728. http://dx.doi.org/10.1099/ijs.063114-0.

GE Power & Water, Water & Process Technologies, 2011. Bell-Carter Olive Food Packing Facility. https://knowledgecentral.gewater.com/kcpguest/documents/Case%20Studies_Cust/Americas/English/CS-BELL-INDWW-EN%20-1106-NA%20GE%20Logo.pdf (last accessed 14.04.14.).

Hammann, R., Kutzner, H.J., 1998. Key enzymes for thedegradation of benzoate, m- and p-hydroxybenzoate by somemembers of the order Actinomycetales. J. Basic Microbiol. 38 (3),207e220.

Harwood, C.S., Parales, R.E., 1996. The b-ketoadipate pathway andthe biology of self-identity. Annu. Rev. Microbiol. 50 (1),553e590. http://dx.doi.org/10.1146/annurev.micro.50.1.553.

Kerbarh, O., Ciulli, A., Howard, N.I., Abell, C., 2005. Salicylatebiosynthesis: overexpression, purification, andcharacterization of Irp9, a bifunctional salicylate synthasefrom Yersinia enterocolitica. J. Bacteriol. 187 (15), 5061e5066.http://dx.doi.org/10.1128/JB.187.15.5061-5066.2005.

Li, Y.-Y., Fang, H.H.P., Chen, T., Chui, H.-K., 1995. UASB treatmentof wastewater containing concentrated benzoate. J. Environ.Eng. 121 (10), 748e751. http://dx.doi.org/10.1061/(ASCE)0733-9372(1995)121:10(748).

Manos-Turvey, A., Bulloch, E.M.M., Rutledge, P.J., Baker, E.N.,Lott, J.S., Payne, R.J., 2010. Inhibition studies of Mycobacteriumtuberculosis salicylate synthase (Mbtl). Chem. Med. Chem. 5(7), 1067e1079. http://dx.doi.org/10.1002/cmdc.201000137.

Martin, B., Humbert, O., Camara, M., Guenzi, E., Walker, J.,Mitchell, T., Andrew, P., Prudhomme, M., Alloing, G.,Hakenbeck, R., Morrison, D.A., Boulnois, G.J., Claverys, J.P.,1992. A highly conserved repeated DNA element located in thechromosomeof Streptococcuspneumoniae.Nucleic. AcidsRes.20 (13), 3479e3483. http://dx.doi.org/10.1093/nar/20.13.3479.

Mnif, S., Chamkha, M., Sayadi, S., 2009. Isolation andcharacterization of Halomonas sp. Strain C2SS100, a

wa t e r r e s e a r c h 6 6 ( 2 0 1 4 ) 1e1 1 11

hydrocarbon-degrading bacterium under hypersalineconditions. J. Appl. Microbiol. 107 (3), 785e794. http://dx.doi.org/10.1111/j.1365-2672.2009.04251.x.

Novachis, L., 2005. Membrane wastewater treatment plantscropping up in food industry. Environ. Sci. Eng. 7. www.esemag.com/archive/0705/zenon.html.

Oie, C.S.I., Albaugh, C.E., Peyton, B.M., 2007. Benzoate andsalicylate degradation by Halomonas campisalis, an alkaliphilicand moderately halophilic microorganism. Water Res. 41 (6),1235e1242. http://dx.doi.org/10.1016/j.watres.2006.12.029.

Piubeli, F., Grossman, M.J., Fantinatti-Garboggini, F., Durrant, L.R.,2012. Enhanced reduction of COD and aromatics in petroleum-produced water using indigenous microorganisms andnutrient addition. Int. Biodeterior. Biodegr. 68, 78e84. http://dx.doi.org/10.1016/j.ibiod.2011.11.012.

Reiner, A.M., Hegeman, G.D., 1971. Metabolism of benzoic acid bybacteria. Accumulation of (-)-3,5-cyclohexadiene-1,2-diol-1-carboxylic acid by mutant strain of Alcaligenes eutrophus.Biochemistry 10 (13), 2530e2536.

Richards, G.P., Fay, J.P., Dickens, K.A., Parent, M.A., Soroka, D.S.,Boyd, E.F., 2012. Predatory bacteria as natural modulators ofVibrio parahaemolyticus and Vibrio vulnificus in seawater andoysters. Appl. Environ. Microbiol. 78 (20), 7455e7466. http://dx.doi.org/10.1128/AEM.01594-12.

Rosenberg, A., 1983. Pseudomonas halodurans sp. nov., ahalotolerant bacterium. Arch. Microbiol. 136 (2), 117e123.http://dx.doi.org/10.1007/BF00404785.

SCCNFP, 2002. Opinion of the Scientific Committee on CosmeticProducts and Non-food Products Intended for Consumers

Concerning Benzoic Acid and Sodium Benzoate. EuropeanCommission, Brussels. http://ec.europa.eu/food/fs/sc/sccp/out166_en.pdf.

Schweigert, N., Zehnder, A.J., Eggen, R.I., 2001. Chemicalproperties of catechols and their molecular modes of toxicaction in cells, from microorganisms to mammals. Environ.Microbiol. 3 (2), 81e91. http://dx.doi.org/10.1046/j.1462-2920.2001.00176.x.

Spokes, J.R., Walker, N., 1974. A novel pathway of benzoatemetabolism in Bacillus species. Ann. Microbiol. Enzymol. 24,307e315.

Tedball, C.A. 1975. Precipitation of sodium benzoate. UnitedStates Patent No. 3,907,877.

Van Belkum, A., Sluijter, M., De Groot, R., Verbrugh, H.,Hermans, P.W.M., 1996. Novel BOX repeat PCR assay for high-resolution typing of Streptococcus pneumoniae strains. J. Clin.Microbiol. 34 (5), 1176e1179.

Ventosa, A., Quesada, E., Rodriguez-Valera, F., Ruiz-Berraquero, F., Ramos-Cormenzana, A., 1982. Numericaltaxonomy of moderately halophilic gram-negative rods. J.Gen. Microbiol. 128, 1959e1968.

WHO, 2005. Benzoic Acid and Sodium Benzoate; ConciseInternational Chemical Assessment Document 26; WHO:Geneva. www.who.int/ipcs/publications/cicad/cicad26_rev_1.pdf.

Zouari, N., 1998. Decolorization of olive oil mill effluent byphysical and chemical treatment prior to anaerobic digestion.J. Chem. Technol. Biotechnol. 73 (3), 297e303. http://dx.doi.org/10.1002/(SICI)1097-4660.