ENVIRONMENTAL BIOTECHNOLOGY

Continuous cultures of Pseudomonas putida mt-2 overcomecatabolic function loss under real case operating conditions

Raúl Muñoz & María Hernández & Ana Segura &

Joao Gouveia & Antonia Rojas & Juan Luis Ramos &

Santiago Villaverde

Received: 2 February 2009 /Revised: 20 February 2009 /Accepted: 20 February 2009 /Published online: 10 March 2009# Springer-Verlag 2009

Abstract The long-term performance and stability ofPseudomonas putida mt-2 cultures, a toluene-sensitivestrain harboring the genes responsible for toluene biodeg-radation in the archetypal plasmid pWW0, was investigatedin a chemostat bioreactor functioning under real caseoperating conditions. The process was operated at a dilutionrate of 0.1 h−1 under toluene loading rates of 259±23 and801±78 gm−3h−1 (inlet toluene concentrations of 3.5 and10.9 gm−3, respectively). Despite the deleterious effects oftoluene and its degradation intermediates, the phenotype ofthis sensitive P. putida culture rapidly recovered from a95% Tol− population at day4 to approx. 100% Tol+ cellsfrom day13 onward, sustaining elimination capacities of232±10 gm−3h−1 at 3.5 g Tol m−3 and 377±13 gm−3h−1 at10.9 g Tol m−3, which were comparable to those achievedby highly tolerant strains such as P. putida DOT T1E andP. putida F1 under identical experimental conditions. Onlyone type of Tol− variant, harboring a TOL-like plasmid witha 38.5 kb deletion (containing the upper and meta operonsfor toluene biodegradation), was identified.

Keywords Microbial damage . Plasmid pWW0 .

Process stability . Pseudomonas putida mt-2 .

Solvent tolerance . Toluene biodegradation

Introduction

Biodegradation of aromatic hydrocarbons has serious limi-tations due to their inherent toxicity and mutagenic nature (USEPA 1994; Zilli et al. 2000), which ultimately challengesmicrobial stability (Sikkema et al. 1995; Ramos et al. 2002).Aromatic hydrocarbons can cause irreversible damage to cellmembranes (loss of ions, metabolites, lipids, and proteins,dissipation of pH gradients and electrical potentials, etc.)followed by cell lysis and death (Segura et al. 1999; Sikkemaet al. 1995). Operational problems derived from microbialinstability in processes treating toluene have been reportedin literature (Revah and Morgan-Sagastume 2005; vanGroenestijn and Kraakman 2005; Shareefdeen and Singh2005). For instance, Song and Kinney (2005) reported adecline in the elimination capacity (EC) of biofilterssubjected to high toluene loadings due to a deterioration ofthe toluene degrading community although the mechanismsresponsible of this deterioration were not identified.Villaverde et al. (1997) recorded a decreasing fraction oftoluene degrading Pseudomonas putida 54G in both sus-pended and biofilm-based bioreactors during the off-gastreatment of 0.5–2.4 g toluene m−3. This reduction, measuredas cellular culturability on toluene, was more pronounced thehigher the toluene inlet concentration was. In this context, P.putida mt-2, a strain harboring the genes responsible fortoluene metabolism in the archetypical pWW0 (TOL)plasmid (Williams and Murray 1974), exhibits a relativelylow tolerance toward toluene (Segura et al., 2003). Moreover,benzyl alcohol and benzoic acid can cause a partial loss of

Appl Microbiol Biotechnol (2009) 83:189–198DOI 10.1007/s00253-009-1928-5

R. Muñoz :M. Hernández : J. Gouveia : S. Villaverde (*)Department of Chemical Engineering and Environmental Technology,Valladolid University,Paseo del Prado de la Magdalena, s/n,Valladolid, Spaine-mail: svillave@iq.uva.es

A. Segura : J. L. RamosEstación Experimental del Zaidín,Consejo Superior de Investigaciones Científicas,Profesor Albareda, 1,Granada 18008, Spain

A. RojasCalantia Biotech, CEEI Valencia/Parque Tecnológico,Avd. Benjamin Franklin 12,46980 Paterna, Valencia, Spain

the TOL plasmid, enzymatic pathway damages, and even acomplete loss of the ability to degrade toluene in pWW0bearing P. putida strains (Nakazawa and Yokota 1973;Williams et al. 1988; Brinkmann et al. 1994; Duetz and vanAndel 1991; Leddy et al. 1995). There is however a lack ofsystematic studies addressing the influence of these genotypeand phenotype changes on pWW0 bearing P. putida culturesover the overall process performance and long-term stability(i.e., EC, removal efficiency, CO2 production, etc.) underoperating conditions commonly encountered in biotreatmentfacilities of toluene-laden industrial emissions (inlet toluenegas concentration below 10 g m−3).

In this study, the long-term performance of a P. putidamt-2 culture was monitored during toluene biodegradation ina continuous suspended growth bioreactor (SGR) operated atinlet toluene gas concentration of 3.5 and 10.9 gm−3 in anattempt to correlate the macroscopic process performancewith toluene-mediated changes in bacterial genotype andphenotype. These changes were described through pWW0plasmid extraction and PCR analysis. It was also aimed tocompare the elimination capacity of P. putida mt-2 cultureswith values previously reported in cultures of highlyresistant strains such as P. putida F1 (Díaz et al. 2008)and P. putida DOT T1E (Hernandez et al., 2008) grownunder identical experimental conditions.

Materials and methods

Microorganisms and culture conditions

P. putida mt-2 was obtained from the German Collection ofMicroorganisms and Cell Cultures [DSM No. 3931]. Theculture was maintained at 4°C in agar slants prepared withmineral salt medium (MSM), 20 g agar l−1 and glucose at5 gl−1 as the sole carbon and energy source. To furnishfresh inoculum, 250-ml E-flasks were supplied with 100 mlof MSM and 1 gl−1 of glucose and incubated for 12 h in athermostated magnetic shaker at 300 rpm and 25°C.

The description of the MSM used for bacterial cultivationcan be found elsewhere (Abril et al. 1989).

Chemicals

All chemicals and reagents were purchased from PANREACwith a purity of at least 99% (Barcelona, Spain). Analyticalgrade benzyl alcohol (BA) and benzoic acid (BAc) werepurchased from Sigma-Aldrich (USA).

Experimental design

The influence of toluene concentration on process perfor-mance and long-term stability (evaluated via periodical

monitorization of toluene EC, CO2 production, specificcatechol-2,3-dioxygenase (C23O) activity, biomass con-centration, and fraction of toluene degrading and C23Obearing cells) was investigated in two series of experi-ments under sterile conditions in a magnetically stirred1-l glass bioreactor (Afora S.A, Spain) operated as achemostat (Bordel et al. 2007). The monitorization of thefraction of non-toluene degrading cells (Tol−) is crucial inthe elucidation of the potential relationships betweenprocess performance and changes in both bacterialgenotype and phenotype. The bioreactor was filled with900 ml of sterile MSM and inoculated with 40 ml ofP. putida mt-2. Temperature and agitation rate wasmaintained constant at 25°C and 500 rpm, respectively.Toluene was supplied in the gas phase via aeration(1,100 ml min−1 of air filtered through a 0.2μmMillex®-FG membrane filter) by mixing a toluene-saturated stream with a toluene-free air stream at differentproportions. Two series of continuous experiments werecarried out for 19 days at a dilution rate (D) of 0.1 h−1 andtoluene inlet concentrations of 3.5±0.3 and 10.9±1.1 g m−3,respectively. Gaseous toluene and CO2 concentrationswere periodically monitored by simultaneous withdrawinga 250µl gaseous sample with Gas-Tight Hamilton syringesthrough valves A and C. Excreted metabolites, extracellulartotal organic carbon (TOC), pH, biomass concentration, thespecific C23O activity, the fraction of toluene degrading andC230-bearing cells, and the specific toluene respiration ratewere also periodically recorded by withdrawing a 40 mlsterile liquid sample through valve B. In addition, dissolvedoxygen concentration (DOC) and temperature (T) weremonitored on line.

Calculations

Toluene EC, CO2 production, the fraction of toluenedegrading cells (Tol+), and the fraction of cells exhibitingC230 activity (C230+) were calculated as:

EC ¼ CinTol � Cout

Tol

� �� V:

g

VR¼½ �g Tolm�3h�1 ð1Þ

CO2 production ¼CoutCO2

� CinCO2

� �� V

:

g

VR¼½ �gCO2 m

�3h�1

ð2Þ

Tolþ ¼ CFU culturable in Toluene

Total CFU in Glucose� 100 ¼½ �% ð3Þ

190 Appl Microbiol Biotechnol (2009) 83:189–198

C230þ ¼ CFU with C230 activity

Total CFU in Glucose� 100 ¼½ �% ð4Þ

where CinTol, C

outTol are the inlet and outlet toluene concentra-

tion, respectively, [=] g m−3; CinCO2

, CoutCO2

are the inlet andoutlet CO2 concentration, respectively, [=] g m−3; CFU isthe colony forming units; VR is the reactor volume, [=] m3;and V

:

g is the gas flow rate, [=] m3 h−1.In the following, results are given as average±standard

deviation.

Analytical methods

Chemical analysis

Toluene quantification was performed in a Gas Chromato-graph (Hewlett-Packard 6890, Palo Alto, CA, USA)coupled with a Mass Spectrometer Detector (Hewlett-Packard 5973 MSD, Palo Alto). CO2 concentrations weremeasured using a GC-TCD (Agilent Technologies 6890N,Palo Alto; 2007). BA and BAc were quantified by HPLC-UVusing a Supelcosil LC-PAH column (Supelco, Bellefonte,USA). These analyses were performed according to Bordelet al. (2007).

Dissolved total organic carbon was measured using aTOC analyzer (Shimadzu TOC-5050A, Japan) according tothe manufacturer. Liquid samples were centrifuged at6,000 rpm during 10 min prior to analysis. The DOC andT in the bioreactor were determined using an O2 transmitter4100 (Meter Toledo GmbH, Urdolf, Germany). A CRISONmicropH 2002 (Crison Instruments, Barcelona, Spain) wasused for pH determination. Turbidity at 650 nm was used asindicator for microbial growth and was measured using aHITACHI U200 UV/visible spectrophotometer (Hitachi,Tokyo, Japan). A correlation between turbidity at 650 nmand P. putida mt-2 dry weight (DW) was performedaccording to Bordel et al. (2007).

Specific respiration rate

The specific toluene respiration rate was determined in aSTRATHKELVIN STRATHOX respirometer (StrathkelvinInstruments Limited, Glasgow, UK). Respirometric reactionvessels were filled (in duplicate) with 16 ml of MSM and2 ml of P. putida mt-2 was withdrawn from the bioreactor(centrifuged at 6,000 rpm during 10 min and resuspendedin fresh MSM) and supplied with toluene at 25 mgl−1 (froma toluene-saturated MSM stock solution). The total reactionvolume was 19 ml. Tests in the absence of carbon sourcewere also carried out under similar conditions to serve ascontrols for endogenous respiration.

Tol+ and C230+ CFU plate counts

The determination of the fraction of cells Tol+ and C230+

was performed by spreading serial dilutions of liquidsamples periodically withdrawn from the reactor (understerile conditions) on agar plates with glucose at 5 gl−1 asthe sole carbon and energy source. After 1 day of growth,colonies were transferred and grid-arranged in triplicate intoagar plates under non-selective (two plates of glucose at5 gl−1) and selective (one plate grown on toluene vapors)conditions. Colonies grown for 2 days at 30ºC on toluenevapors represented the fraction of Tol+ cells. The presenceof the enzyme C23O (C230+ cells) was assessed in coloniesgrown for 1 day at 30ºC under non-selective conditions byspraying the plates with a 250 mM catechol solution,according to Duetz and van Andel (1991). Positive colonies(those exhibiting basal C23O activity) turned bright yellowalmost immediately.

C230 activity assay

Liquid samples of 20 ml were periodically withdrawn fromthe bioreactor under sterile conditions and centrifuged at6,000 rpm for 5 min at 4ºC. The cell pellet was washedtwice with 5 ml of a 100 mM phosphate buffer (pH7.5 at4ºC) and recentrifuged. The washed pellet was finallyresuspended in 2 ml of a 100 mM phosphate buffer (pH7.5at 4ºC) containing 10% acetone. Biomass concentrationwas then determined via OD650 measurement, and the cellsuspension was lysed by sonication on ice (6 cycles of 30 sON/60 s OFF at 100 W in a Vibra Cell 75185 sonicator).Cell debris was removed by centrifugation at 10,000 rpmfor 5 min, and the supernatant (cell-free extract) wasdecanted and used for enzyme assays. Cell-free extractswere assayed in duplicate in a total volume of 770µl (70µlof cell-free extract+685µl of a 100 mM phosphate buffer(pH7.5 at 25ºC) containing 10% acetone+15µl of a 25 mMcatechol stock solution) at 25ºC on an thermostatedHITACHI U200 UV/visible spectrophotometer. Catechol2,3-dioxygenase was assayed by continuous monitorizationof bioreaction medium absorbance at 375 nm for 60 s. Oneenzyme unite (U) was defined a 1 unit increase in theabsorbance at 375 nm/min.

Plasmids and PCR analysis

Genotype changes of P. putida mt-2 cultures were describedby extraction of plasmid pWW0 and PCR analysis.Enzymes for the degradation of toluene are encoded intwo operons in the plasmid pWW0. The enzymes encodedin the upper pathway transform toluene into benzoate, andthe enzymes of the meta pathway degrade benzoate intointermediates of the Krebs cycle (Harayama et al. 1986). To

Appl Microbiol Biotechnol (2009) 83:189–198 191

analyze the presence/absence of the plasmid pWW0 or thegenes that encode for the toluene degradative enzymes,plasmids of 12 Tol+ and 12 Tol− colonies obtainedthroughout the experiment at low toluene loadings wereanalyzed. Plasmids from 21 Tol+ and 38 Tol− colonies werealso analyzed when operating the system at high tolueneloadings. Extraction of plasmid pWW0 was done by thealkaline lysis method (Birnboim and Doly 1979) with smallmodifications. Briefly, 1 ml of culture was harvested bycentrifugation. Cells were gently resuspended in 500μl ofTENS (20 mM Tris–HCl pH=8, 5 mM EDTA pH=8,100 mM NaOH and 0.5% SDS) and kept on ice for 5 min.Then, 150μl of 3 M sodium acetate pH=4.8 was added andcarefully mixed. The tubes were centrifuged 10 min at13,000 rpm to eliminate the cellular debris, and thesupernatant was phenolized twice before precipitation with0.6 volumes of isopropanol. The remaining pellet waswashed once with 70% ethanol and resuspended in 40μl ofwater. Plasmid suspension (17μl) was digested with SmaI,and the digestion was visualized in a 0.5% (w/v) agarosegel.

Oligonucleotides XylM1 (5′ ATGGACACGCTTCGTTATTACC 3′) and XylM6rev (5′ TTTATCGCATCTTTGACGG 3′) and oligonucleotides XylZ1 (5′ CTGGATGCCCTGTGCCCGGAT 3′) and XylZ2 (5′ TGCAAATGCTTCGCTGAGAG 3′) were used to amplify a fragment of the xylM andxylZ genes, respectively. PCR was performed using anannealing temperature of 60ºC. Oligonucleotides orf144 (5′GTCTTGGATAAGCACACGCT 3′) and dgkA (5′ AACGCCAGCATCGGCACGAG 3′) were used to map the deletionin the pWW0 plasmid.

Results

Process operation at low toluene loadings

When the system was operated at average toluene inletconcentrations of 3.5±0.3 g m−3, the process was charac-terized by a rapid increase in toluene EC up to 134±12 gm−3h−1, biomass concentration (277±3 mg DW l−1), andCO2 production (277±17 g m−3h−1) within the first 24 h ofoperation (Fig. 1a, b). This increase was concomitant with adecrease in cultivation pH and DOC down to 6.55 and1.02 mg O2 l

−1, respectively (Fig. 1a). Process performancehowever severely declined 33 h following inoculationreaching a pseudo-steady state for approx 3.5 days at ECof 24±11 g m−3h,

−1 CO2 productions of 116±46 g m−3h−1,pH levels of 6.8±0.1, and biomass concentrations of65±7 mg DW l−1. Microbial activity rapidly recoveredwithin the seventh day of cultivation reaching steady ECs,CO2 productions, pH, and DOC of 232±10 g m−3h−1,552±28 g m−3h−1, 5.6±0.1, and 0.5±0.3 mg O2 l−1,respectively. Despite stationary state ECs were attainedfrom day7, biomass concentration gradually increased from792±11 mg DW l−1 at day7 to stable levels of 1,616±270 mg DW l−1 at day12. In this context, NH4

+ concen-tration was inversely correlated to biomass concentration,with total nitrogen depletion recorded from day10. TOCwas also correlated to biomass concentration, exhibiting asharp increase during process start-up up to 66 mg C l−1 tofinally stabilize at 186±33 mg C l−1 during the last stagesof the experimentation period (Fig. 1c). Likewise, BA andBAc also paralleled biomass time course, reaching concen-

0

50

100

150

200

250

300

0 4 8 12 16 20

EC

(g m

-3h-1

)

5

6

7

8

pH

0

200

400

600

800

0 4 8 12 16 20

CO

2 (g

m-3

h-1)

0

500

1000

1500

2000

Bio

mas

s (m

g l-1

)

b

0

40

80

120

160

200

0 4 8 12 16 20

Time (days)

TO

C (m

g l-1

)

0

2

4

6

8

10

12

14

BA

c / B

A (m

g l-1

)

c

0

20

40

60

80

100

0 4 8 12 16 20

Time (days)

Tol

+ /

C23

0+ (

%)

0.0

0.2

0.4

0.6

0.8

1.0

Res

pira

tion

(gO

2 g D

W-1

h-1)

d

aFig. 1 Time course of EC(solid squares), pH (pointedline), CO2 production rate (soliddiamonds), biomass concentra-tion (empty triangles), TOC(empty squares), benzyl alcohol(solid circles), and benzoic acid(empty diamonds) concentra-tions, specific toluene associatedrespiration rate (solid triangles),fraction of toluene degradingcells (empty circles), andfraction of cells exhibitingC230 activity (asterisks) duringtoluene biodegradation byP. putida mt-2 in a sterile SGRoperated as a chemostat at3.5 g Tol m−3 and D=0.10 h−1.Vertical bars represent standarddeviations

192 Appl Microbiol Biotechnol (2009) 83:189–198

trations of 12 and 7 mg l−1, respectively, by the second dayof operation (Fig. 1c). However, the monitorization of bothBA and BAc by HPLC-UV was hindered from day6 due tothe presence of unidentified compounds eluting at nearbyretention times to our target metabolites and thereforeinterfering their quantification. These compounds, interfer-ing BA and BAc quantification could be cellular inter-mediates released to the extracellular medium as a result ofthe partial cell disruption caused by the low pH prevailingduring this experimentation time.

The specific O2 consumption associated to toluenerespiration rapidly increased during the first day of cultiva-tion up to 0.54±0.01 g O2 g DW−1h−1, declining to values of0.1 g O2 g DW−1h−1 during process performance deteriora-

tion (days3–4) to further increase concomitantly with therecovery of microbial activity observed (Fig. 1d). Fluctuatingrespiration rates ranging from 0.3 to 0.6 g O2 g DW−1h−1

were then recorded during the pseudo-steady state achievedfrom the seventh day of operation. On the other hand, thefraction of toluene degrading cells and cells bearing theenzyme catechol-2,3-dioxygenase were always well correlat-ed (R2>0.999) throughout the entire experimentation period(Fig. 1d). These fractions remained approx. constant duringthe first 33 h ranging from 94% to 100% and graduallydecreased to 51% by the fourth day of cultivation. Regardlessof the stabilization of EC from day7, the population of Tol+

and C230+ cells periodically fluctuated from 63% to 93%.Furthermore, the activity of the enzyme catechol-2,3-dioxygenase rapidly increased up to 64±3 U mg DW−1

within the first 10 h of cultivation and gradually decreased tostabilize at 29±4 U mg DW−1 from day4 (Fig. 2).

Process operation at high toluene loadings

When operating at 10.9±1.1 g Tol m−3, toluene mineralizationreadily increased within the first 20 h of operation attainingECs, CO2 production, and biomass concentrations of 37±15 g m−3h−1, 250±22 g m−3h−1, and 184±3 mg DW l−1,respectively (Fig. 3a, b). This increase was also followed by adecrease in culture pH (down to 6.6), DOC (down to 0.9 mgO2 l

−1), and NH4+ concentration (from 190 to 120 mg Nl−1).

A short pseudo-stationary state characterized by low ECs,CO2 production rates, and biomass concentrations wasrecorded from days1 to 4.5. A sudden recovery occurred atday5, followed by a rapid process performance deterioration

0

20

40

60

80

0 4 8 12 16 20Time (days)

C23

0 (U

mg

DW

-1)

Fig. 2 Time course of catechol-2,3-dioxygenase activity in P. putidamt-2 during cultivation in a chemostat SGR operated at 3.5±0.3 g Tol m−3

(empty squares) and 10.9±1.1 g Tol m−3 (solid squares). Vertical barsrepresent standard deviations

0

100

200

300

400

0 4 8 12 16 20

EC

(g m

-3h-1

)

5

6

7

8

pH

0

200

400

600

800

1000

0 4 8 12 16 20

CO

2 (g

m-3

h-1)

0

500

1000

1500

2000

Bio

mas

s (m

g l-1

)

b

0

100

200

300

400

0 4 8 12 16 20

Time (days)

TO

C (m

g l-1

)

0

100

200

300

400

AcB

/ B

A (m

g l-1

)

c

0

20

40

60

80

100

0 4 8 12 16 20

Time (days)

Tol

+ /

C23

0+ (%

)

0.0

0.2

0.4

0.6

0.8

1.0

Res

pira

tion

(gO

2 g

DW

-1 h

-1)

d

aFig. 3 Time course of EC(solid squares), pH (pointedline), CO2 production rate (soliddiamonds), biomass concentra-tion (empty triangles), TOC(empty squares), benzyl alcohol(solid circles), and benzoic acid(empty diamonds) concentra-tions, specific toluene associatedrespiration rate (solid triangles),fraction of toluene degradingcells (empty circles), and frac-tion of cells exhibiting C230activity (dashed line) duringtoluene biodegradation byP. putida mt-2 in a sterile SGRoperated as a chemostat at10.9 g Tol m−3 and D=0.10 h−1.Vertical bars represent standarddeviations

Appl Microbiol Biotechnol (2009) 83:189–198 193

at day6. An steady and consistent increase in processperformance was however observed from day8 onward,characterized by the rising in toluene ECs, CO2, and biomassproductions and a decrease in culture pH and N–NH4

+

concentrations (total N depletion from day9). Final ECs of377±13 g Tol m−3h−1 concomitant with CO2 productions andbiomass concentrations of 806±35 g m−3h−1 and 1,774±197 mgl−1, respectively, were recorded by the end of theexperimentation period. Both TOC and BAwere correlated tobiomass concentration, exhibiting a sharp increase at day8 upto 280 and 260 mgl−1, respectively (Fig. 3c). Theseconcentrations gradually decreased to finally stabilize at182±17 mg C l−1 and 17 mg BA l−1. As pointed out above,interference with unidentified extracellular compounds elut-ing at nearby BAc retention time hindered the quantificationof this metabolite from day7.

The rapid increase recorded during process start-up inthe specific toluene associated respiration rate (up to 0.60±0.0 g O2 g DW−1h−1) was followed by a significant declineduring the two episodes of process deterioration recorded(0.14±0.1 g O2 g DW−1h−1 at the second day ofexperimentation and 0 g O2 g DW−1h−1 at day6; Fig. 3d).

However, process recovery at day8 resulted in a sharp risein bacterial respirations up to 0.6±0.01 g O2 g DW−1h−1,gradually stabilizing at 0.31±0.03 g O2 g DW−1h−1 by thelast week of experimentation. The fraction of Tol+ andC230+ rapidly declined from start-up values of 100% to 6%within the first 4 days of experimentation (Fig. 3d). Asudden increase in Tol+ and C230+ recorded at day5(concomitant with an increase in EC) was followed bylow percentages (3–8%) from days6 to 7. Despite thefluctuations observed from day8 to day13 (Tol+ and C230+

ranged from 15% to 69%), the fraction of cells capable togrow on toluene exhibiting C230 activity stabilized fromday14 at values higher than 93%. In addition, the activityof the enzyme catechol-2,3-dioxygenase steadily raised upto 37±5 U mg DW−1 within the first 2 days, stabilizingafterward at stationary values of 16±4 U mg DW−1 (Fig. 2).

Molecular analysis of the pWW0 plasmid in Tol+

and Tol− cells

All 50 Tol− colonies analyzed lost a plasmid fragment ofapprox. 39 kb (Fig. 4b). None of the Tol− colonies analyzed

Pu xylUWCMABN Pm xylXYZLEGFJQKIH

xylMA xylB xylC xylXYZ

xylL xylE

Krebs cycleintermetiates

OH

OCOOH

CO2OH

OHOHH

HOOC OH

a

tol-

tol+

tol+

tol-

15.4Kb14.5Kb13.4Kb12.6Kb

7.4Kb

3.2+2.9Kb

pWW0 (116599 bp)

2636

9

-34

30

1613

1

2161

5

2163

222

562

2933

232

975

3425

941

701

4495

3

4655

3

6108

662

129

7561

7

9575

3

1112

44

13.4

Kb

5.4

Kb

3.8

Kb

3K

b

3.6

Kb

7.4

Kb

3.2

Kb

15.4

Kb

12.6

Kb

20K

b

14.5

Kb

8.8

Kb

meta upper

b

CH2OH COOHCHOCH3

Fig. 4 a Schematic representation of the reactions involved in toluenedegradation and gene organization of the degradation pathway. bUpper part In silico analysis of the restriction fragments generated bycutting with SmaI. Numbers in the upper part of the drawing indicatedthe position at which the SmaI target sequence is located within thepWW0 plasmid; numbers in the lower part indicated the predicted

size of the restriction fragments. Lower part Visualization in a 0.5%(w/v) agarose gel of the restriction fragments originated from Tol− andTol+ cultures. The first lane was loaded with the molecular marker λ-HindIII. The size of the bands that are not present in Tol− cells areindicated by bolded numbers

194 Appl Microbiol Biotechnol (2009) 83:189–198

lost the entire pWW0 plasmid. To confirm the absence ofthe genes that encoded for the enzymes responsible fortoluene degradation in the Tol− cells, the xylM gene (thatencoded for the xylene monooxygenase and is locatedwithin the upper operon) and the xylZ gene (that encodedfor one of the toluate 1,2 dioxygenase subunits and ispresent in the meta operon) were amplified using astemplate plasmid DNA from ten different Tol+ and Tol−

colonies. All colonies exhibiting a 39 kb plasmid deletiondid not show the corresponding amplification of bands,while plasmids from Tol+ colonies did amplify (data notshown). Primers dgkA and orf144 (nucleotides 36541–36561 and 77950–77930, respectively, in the pWW0sequence) did not amplify when pWW0 plasmid was usedas template but an amplification band of around 2.3 kb wasobserved when plasmids carrying the deletion were used astemplate. Deletion was the product of the recombinationamong direct repeats up- and downstream the TnpA genesof the IS1246. Reconstituted sequence can be obtained bydeleting the 39042 base pairs between positions 38112 and77154 of the pWW0 plasmid (Meulien et al. 1981).

Discussion

A rapid decline in P. putida mt-2 activity was observedfrom the first day of cultivation, regardless of the tolueneloading applied. This initial deterioration of processperformance, characterized by a gradual decline in EC,CO2, and biomass production, was correlated with adecrease both in the fraction of Tol+ and C230+ cells andin specific toluene respiration rates. pH values remainedhigher than 6.5 during this period, which suggests that pHwas not likely to mediate this deterioration in P. putida mt-2activity (Figs. 1a and 3a). Indeed, pH-induced inhibitions inPseudomonas cultures have been observed at pHs below 5(Muñoz et al. 2008; Hernandez et al. 2008; Pini et al.personal communication). A preliminary analysis of thecorrelations existing among process parameters suggeststhat the loss of the ability to utilize toluene was the mainresponsible of process deterioration. Thus, a lower fractionof toluene degrading cells in the culture will ultimatelyresult in lower toluene associated O2 consumption rates (asshown by the respirometric assays) and in lower tolueneutilization rates during process operation.

This deterioration of microbial activity, leading to poorprocess performance, was however not surprising as severalstudies have reported instability problems in toluene removalprocesses, mainly in systems with plasmid-encoded degrada-tion pathways. Thus, based on the extensive work carried outregarding the deleterious effects of benzyl alcohol andbenzoate on plasmid-encoded pathways, it was hypothesizedthat the appearance of Tol− cells in this study was mediated

by the above-mentioned metabolites. However, it must behighlighted that the main goal and novelty of the hereinpresented study was not the elucidation of the mechanismsresponsible of Tol− generation rather than the determinationof the connection between the loss of toluene degradationcapacity and the long-term overall bioreactor performance. Inthis context and to the best of our knowledge, the studiespublished so far has focused on the mechanisms mediatingplasmid deleting, with no research paper addressing thepractical implications of catabolic function. For instance,benzyl alcohol and benzoic acid can cause a partial loss ofthe TOL plasmid, enzymatic pathway damages, and even acomplete loss of the ability to degrade toluene in pWW0bearing P. putida strains (Nakazawa and Yokota 1973;Willians et al. 1988; Brinkmann et al. 1994; Duetz and vanAndel 1991; Leddy et al. 1995). Only mutants with a TOL-like plasmid of reduced size (38.5 kb deletion) were recordedin our experimental system, regardless of the toluene loadingrate applied. These variants failed to amplify in PCR assayswith xylM- and XylZ-specific oligonucleotides, confirmingthe absence of the upper and meta operons in Tol− cells.These two catabolic operons are included in transposonTn4651 (56 kb), and recombination between two directrepeats located within the transposon has been recognized asthe cause of the 39 kb deletion in the plasmid pWW0(Meulien et al. 1981). In our particular study, BA and BAcrapidly accumulated in the cultivation medium duringprocess start-up, and the larger the accumulation, the higherthe toluene inlet concentration was (Figs. 1c and 2c). Thisaccumulation was however unanticipated based on previousfindings reported by Ramos et al. (1997), who observed thatenzyme induction within the TOL pathway occurred insequential manner in order to avoid BAc accumulation. Inthe process herein investigated, benzyl alcohol, the mainexcreted metabolite identified (with concentrations of up to260 mg l−1), might have been directly involved in thearchetypal reported 38.5 kb pWW0 plasmid deletion.Consequently, process operation at higher toluene concen-trations resulted in a larger loss of both toluene degradationcapacity and C230− activity (minimum Tol+ and C230+

fractions of 5–6% at 10.9 g Tol m−3 vs. 51% at 3.5 g Tol m−3,respectively). These results are in agreement with thosereported by Villaverde et al. (1997) who observed thattoluene harmful effects on P. putida 54 G activity were moresevere the higher the toluene concentration and longer theexposure time were. On the contrary, Bordel et al. (2007)showed an accumulation of BA as a result of a side-chaintoluene monooxygenation in cultures of P. putida F1operated at toluene inlet concentrations of 9 g m−3, butneither inhibitory nor mutagenic effects were reported at thedetected concentrations. Therefore, this suggests that onlyTOL plasmid-based degradation systems are prone to theBA-mediated harmful effects.

Appl Microbiol Biotechnol (2009) 83:189–198 195

Microbial activity was rapidly recovered within the seventhday of experimentation, regardless of the toluene loading rateapplied, resulting in a process performance enhancementcharacterized by a rapid increase in toluene ECs, CO2, andbiomass productions, toluene associated respirations, andTOC and metabolite accumulation. It must be also high-lighted that despite the depletion of ammonium in thecultivation medium, a significant bacterial growth wasobserved from day10 in the experiment carried out at lowtoluene loading rates. These experimental findings suggest adecrease in bacterial nitrogen content throughout the exper-iment. In this context, Hernandez et al. (2008) also recordedsignificantly higher nitrogen contents in wild-type P. putidaDOT T1E culture (0.2±0.03 gN g DW−1) compared to P.putida DOT T1E mutants (0.13±0.01 gN g DW) undercomparable experimental conditions. A control experimentthat carried out a low toluene concentration (3.5 g m−3) for8 days reported similar results (process deterioration causedby a gradual decrease in Tol+ and C230+ population,followed by a rapid recovery of P. putida mt-2 by theseventh day of cultivation), which highlights the reproduc-ibility and consistency of the herein reported findings (datanot shown). In both series of experiments, the enhancementon toluene biodegradation performance was correlated withan increase in the fraction of Tol+ and C230+ cells. At lowinlet toluene concentrations, the fraction of Tol+/C230+ cellsfluctuated from 63% to 93% while maintaining high andsteady toluene EC and CO2 production rates. However, whenthe process was operated at 10.9 g Tol m−3, the fraction ofTol+ and C230+ cells underwent a period of intense and rapidoscillations (15–69%) from day7 to day13, supportingtoluene ECs and CO2 productions comparable to thoseachieved at low toluene loadings during steady-state con-ditions. These severe fluctuations, much larger than thoseobserved at low toluene loadings, might have been inducedby the high concentrations of the mutagenic BA present inthe system after process recovery. From day13, the biodeg-radation process at high toluene loadings was neverthelesscharacterized by the establishment of an increasingly robustTol+ and C230+ culture (93–100%) concomitant with asignificant decrease in BA concentration and by an enhancedtoluene biodegradation performance (EC of up to 377±13 gTol m−3h−1). This is, to the best of our knowledge, the firstreport of a progressive cellular adaptation to a toxic VOC,yielding a more stable and efficient process capable todegrade toluene at rates comparable to those achieved byhighly resistant strains such as P. putida F1 or P. putida DOTT1E (see below). At this point, it must be also stressed thatstable toluene removal rates were achieved despite thefluctuation in the fractions of Tol+ and C230+ cells, whichhighlights the versatility of the enzymatic machinery har-bored by the tested bacterial strain. These results are in

agreement with those reported by Muñoz et al. (2008), whoobserved that P. putida F1 was capable to rapidly cope (lessthan 30 min) with sudden increases in toluene inletconcentration (from 3 to 6 g Tol m−3) at unaffected removalefficiencies. On the other hand, the specific C230 activitywas not correlated to any of the process parametersmonitored, increasing during the initial stages of thebiodegradation process and stabilizing from day10 onward.The reasons underlying the lower specific C230 activityrecorded under operation at high toluene loadings remainedstill unclear.

The presence of a sustained population of Tol−/C230−

population throughout the entire experimentation periodsuggest either a rapid generation of Tol−/C230− cells withinthe cultivation or the establishment of a mutant populationliving at the expenses of the dissolved organic carbonexcreted by toluene degrading cells. Villaverde and Fernández(1997) reported a positive correlation between non-toluene-associated respiration and the fraction of viable stressed andnon-toluene degrading cells in biofilms, suggesting growth onleakage and lysis products. Likewise, growth on theaccumulated benzoate via the ortho pathway encoded in theP. putida mt-2 chromosome could also explain the growth ofTol− cells in the bioreactor (Brinkmann et al., 1994). From apractical viewpoint, while the establishment of a significantpopulation of Tol− cells during the initial stages of experi-mentation did correlate with a deterioration in toluenebiodegradation, process performance remained unaffected bythe fluctuation in mutants population once the culture wasproperly adapted.

This study also differentiates bacterial tolerance frombacterial degradation capacity, confirming the results pre-viously reported by Mosqueda et al. (1999). Thus, whileP. putida DOT T1E was classified as a highly resistant strainpossessing three solvent efflux pumps and P. putida F1 as amoderately resistant strain harboring only two solventextrusion pumps, P. putida mt-2 was classified as a sensitivestrain harboring 1 solvent efflux pump (Mosqueda et al.1999). However, the ECs achieved by P. putida mt-2 attoluene inlet concentration of 3.5 g m−3 (232±10 g m−3h−1)were comparable to those achieved by P. putida F1 andP. putida DOT T1E under similar experimental conditions(216±15 g m−3h−1 and 284±22 g m−3h−1; Diaz et al. 2008)and only slightly lower (22%) than those recorded at tolueneinlet concentrations of 10 g m−3 under identical experimentalconditions. It must be also stressed that the low pH values(≈5.5) recorded in the herein reported study during these“high performance” periods could have exerted a significantinhibitory effect on P. putida mt-2 catabolic activity. HigherECs could be expected in processes operated at pH close toneutrality. Unfortunately, no specific tests evaluating bacte-rial activity at different pHs were performed to elucidate this

196 Appl Microbiol Biotechnol (2009) 83:189–198

anticipated enhancement. In addition, total carbon recoveries([CCO2+CTOC+CTolout+CBiomass]/[CTolin]) of 104±9%, and103±6% were recorded in the processes operated at low andhigh toluene loading, respectively.

In brief, this study illustrated the progressive self-adaptation of the sensitive bacterial strain P. putida mt-2toward the disrupting and toxic effects of toluene and itsintermediates (BA and BAc), yielding a more robust andefficient process within only 69 cell generations. While theinitial deterioration of microbial activity was correlated to adecrease in the fraction of Tol+/C230+ cells (likely mediatedby the presence of either benzyl alcohol or benzoic acid),efficient and consistent process performance at fluctuatingTol+/C230+ fractions was achieved by the end of theexperimentation period, regardless of the toluene inletconcentrations applied. Despite being considered as a highlysensitive strain, this strain exhibited toluene ECs comparablethan those reported by tolerant strains such as P. putida DOTT1E and P. putida F1 under identical experimental con-ditions even at toluene loadings of 800 g m−3h−1, whichsuggested that solvent tolerance might not be directlyinvolved in the mechanisms governing the bacterial catabolicpotential under the operational conditions herein evaluated(based on the assumption that this tolerance solely involvesthe solvent efflux pumps currently reported).

Acknowledgments This research was supported by the SpanishMinistry of Education and Science (RYC-2007-01667, PPQ2006-08230, CONSOLIDER-C BIO2006 05668, and CONSOLIDER-CSD2007-00055). Araceli Crespo and Isabel Jimenez are also gratefullyacknowledged for their practical assistance.

References

Abril MA, Michán C, Timmis KN, Ramos JL (1989) Regulator andenzyme specificities of the TOL plasmid-encoded upper pathwayfor degradation of aromatic hydrocarbons and expansion of thesubstrate range of the pathway. J Bacteriol 171:6782–6790.

Birnboim HL, Doly J (1979) A rapid alkaline extraction procedure forscreening of recombinant plasmid DNA. Nucleic Acids Res7:1513–1523

BrinkmannU, Ramos JL, ReinkeeW (1994) Loss of the Tol meta-cleavagepathway functions of Pseudomonas putida strain PaW1 (pWW0)during growth on toluene. J Basic Microbiol 34(5):303–309

Bordel S, Diaz LF, Muñoz R, Villaverde S (2007) New insights ontoluene biodegradation by Pseudomonas putida F1: Influence ofpollutant concentration and excreted metabolites. Appl MicrobiolBiotechnol 74:857–866

Díaz LF, Muñoz R, Bordel S, Villaverde S (2008) Toluenebiodegradation by Pseudomonas putida F1: targeting culturestability in long term operation. Biodegradation 19:198–207

Duetz WA, van Andel JG (1991) Stability of TOL plasmid pWW0 inPseudomonas putida mt-2 under non-selective conditions incontinuous culture. J Gen Microbiol 137:1369–1374

Harayama S, Leppik RA, Rekik M, Mermod N, Lehrbach PR,Reineke W, Timmis KN (1986) Gene order of the TOLcatabolic plasmid upper pathway operon and oxidation of bothtoluene and benzyl alcohol by the xylA product. J Bacteriol167:455–461

Hernandez M, Gouveia J, Segura A, Muñoz R, Villaverde S (2008)Addressing the role of the extrusion pump-bearing pGRT1plasmid on toluene biodegradation by Pseudomonas putidaDOT-T1E under real case scenarios. Process Biochemistry(submitted for publication)

Leddy MB, Phipps DW, Ridgway HF (1995) Catabolite-mediatedmutations in alternate toluene degradative pathways in Pseudomonasputida. J Bacteriol 177(16):4713–4720

Meulien P, Downing RG, Broda P (1981) Excision of the 40 Kbsegment of the TOL plasmid from Pseudomonas putida mt-2involves direct repeats. Mol Gen Genet 184:97–101

Mosqueda G, Ramos-Gonzalez MI, Ramos JL (1999) Toluenemetabolism by the solvent-tolerant Pseudomonas putida DOT-T1 strain, and its role in solvent impermeabilization. Gene232:69–76

Muñoz R, Díaz LF, Bordel S, Villaverde S (2008) Response ofPseudomonas putida F1 cultures to fluctuating toluene loads andoperational failures in suspended growth bioreactors. Biodegra-dation 19:897–9085

Nakazawa T, Yokota T (1973) Benzoate metabolism in Pseudomonasputida (arvilla) mt-2: demonstration of two benzoate pathways.J Bacteriol 115:262–275

Ramos JL, Marques S, Timmis K (1997) Transcriptional control of thePseudomonas TOL plasmid catabolic operons is achievedthrough an interplay of host factors and plasmid-encodedregulators. Annu Rev Microbiol 51:341–372

Ramos JL, Duque E, Gallegos MT, Godoy P, Ramos-Gonzalez MI,Rojas A, Terán W, Segura A (2002) Mechanisms of solventtolerance in gram-negative bacteria. Annu Rev Microbiol56:743–768

Revah S, Morgan-Sagastume JM (2005) Methods for odor and VOCcontrol. In: Shareefdeen Z, Singh A (eds) Biotechnology forodour and air pollution. Springer, Heidelberg, pp 29–64

Segura A, Duque E, Mosqueda G, Ramos JL, Junker F (1999)Multiple responses of Gram-negative bacteria to organic solvents.Environ Microbiol 1:191–198

Segura A, Rojas A, Hurtado A, Huertas MJ, Ramos JL (2003)Comparative genomic analysis of solvent extrusion pumps inPseudomonas strains exhibiting different degrees of solventtolerance. Extremophiles 7:371–376

Shareefdeen Z, Singh A (2005) Biotechnology for odour and airpollution. Springer, Heidelberg

Sikkema J, de Bont JAM, Poolman B (1995) Mechanisms ofmembrane toxicity of hydrocarbons. Microbiol Rev 59:201–222

Song JH, Kinney KA (2005) Microbial response and eliminationcapacity in biofilters subjected to high toluene loadings. ApplMicrobiol Biotechnol 68:554–559

U.S. EPA (1994) Ranking of pollutants with respect to human health.Technical background document to support rulemaking pursuantto clean air. Act. section 112 (g). US Environment ProtectionAgency, Research Triangle Park, N.C.

van Groenestijn JW, Kraakman NJR (2005) Recent developments inbiological waste gas purification in Europe. Chem Eng J 113:85–91

Villaverde S, Fernandez MT (1997) Non toluene-associated respira-tion in a Pseudomonas putida 54G biofilm grown on toluene in aflat-plate vapor-phase bioreactor. Appl Microbiol Biotechnol48:357–362

Appl Microbiol Biotechnol (2009) 83:189–198 197

Villaverde S, Mirpuri RG, Lewandowski Z, Jones WL (1997)Physiological and chemical gradients in a Pseudomomas putida54G biofilm degrading toluene in a flat plate vapor phasebioreactor. Biotechnol Bioeng 56:361–371

Williams PA, Murray K (1974) Metabolism of benzoate and themethylbenzoates by Pseudomonas arvilla (arvilla) mt-2: evidencefor the existence of a TOL plasmid. J Bacteriol 120:416–423

Williams PA, Taylor SD, Gibb LE (1988) Loss of the toluene-xylenecatabolic genes of TOL plasmid pWW0 during growth ofPseudomonas putida on benzoate is due to a selective growthadvantage of “cured” segregants. J Gen Microbiol 134:2039–2048

Zilli M, Del Borghi A, Converti A (2000) Toluene removal vapor in alaboratory-scale biofilter. Appl Microbiol Biotechnol 54:248–254

198 Appl Microbiol Biotechnol (2009) 83:189–198