Food & Beverages

Biofilm population dynamics in a trickle-bed bioreactor used for the biodegradation of aromatic hydrocarbons from waste gas under transient conditions

Description
Biofilm population dynamics in a trickle-bed bioreactor used for the biodegradation of aromatic hydrocarbons from waste gas under transient conditions
Published
of 12
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
   Biodegradation  15:  133–144, 2004.© 2004  Kluwer Academic Publishers. Printed in the Netherlands.  133 Biofilm population dynamics in a trickle-bed bioreactor used for thebiodegradation of aromatic hydrocarbons from waste gas under transientconditions D. Hekmat ∗ , A. Feuchtinger, M. Stephan & D. Vortmeyer  Institute of Chemical Engineering, Technische Universität München, Boltzmannstrasse 15, 85748 Garching,Germany ( ∗ author for correspondence: e-mail: hekmat@lrz.tum.de) Accepted 12 November 2003 Key words:  aromatic hydrocarbons, biodegradation, multispecies biofilm, population dynamics, trickle-bedbioreactor, waste gas Abstract The dynamics of a multispecies biofilm population in a laboratory-scale trickle-bed bioreactor for the treatment of waste gaswasexamined.Themodelpollutantwas aVOC-mixtureofpolyalkylatedbenzenescalledSolvesso100  .Fluorescence in-situ hybridization (FISH) was applied in order to characterise the population composition. Thebioreactorwasoperatedundertransientconditionsbyapplyingpollutantconcentrationshiftsandastarvationphase.Only about 10% of the biofilm mass were cells, the rest consisted of extracellularpolymericsubstances (EPS). Theaverage fraction of Solvesso 100  -degrading cells during pollutant supply periods was less than 10%. About 60%of the cells were saprophytes and about 30% were inactive cells. During pollutant concentration shift experiments,the bioreactor performance adapted within a few hours. The biofilm population exhibited a dependency upon thedirection of the shifts. The population reacted within days after a shift-down and within weeks after a shift-up.The pollutant-degraders reacted significantly faster compared to the other cells. During the long-term starvationphase, a shift of the population composition took place. However, this change of composition as well as thedegree of metabolic activity was completely reversible. A direct correlation between the biodegradation rate of the bioreactor and the number of pollutant-degrading cells present in the biofilm could not be obtained due toinsufficient experimental evidence. Introduction Trickle-bed bioreactors have proven to be effectivesystems for waste gas treatment. This is especially thecase for relatively low pollutant concentration levelsof around 1 g C X H Y  per m 3 of waste gas. The biode-gradation of the organic pollutants takes place mainlyvia aerobic oxidation by a mixed population of mi-croorganisms. These microorganismsare immobilisedin a biofilm which covers the surface of a packingmaterial of the trickle-bed column. Previous studieshave confirmed the suitability of trickle-bed bioreact-ors for the treatment of volatile organic compounds(VOC’s) such as dichloromethane (Diks & Ottengraf 1991), toluene (Arcangeli & Arvin 1992; Møller et al.1996; Pedersen et al. 1997), and polyalkylated ben-zenes (Hekmat & Vortmeyer 1994). Further work wasperformed with regard to the biodegradation dynam-ics, since the industrial process under real conditionsis usually unsteady (Deshusses et al. 1996). This isbecause on one hand, the pollutant source is not con-stant over time and on the other hand, the bacterialculture exhibits an internal dynamic behavior depend-ing uponvariable physiologicalconditions. Therefore,the start-up behavior of trickle-bedbioreactorsand theeffect of feed pollutant concentration cycling was in-vestigated (Hekmat et al. 1997). For simplification,however, the biofilm was regarded as a homogen-eous mass in most studies. Thus, the properties of individual species was ignored (Siebel & Characklis  134 Figure 1.  Schematic diagram of the experimental set-up of the trickle-bed bioreactor (GC (FID) = gas chromatograph with flame ionizationdetector). Figure 2.  Left: segment of the packed column during continuous long-term operation in the trickle regime. Right: hydrophilised polypropyleneRalu  -ring packing material. The carrier material is covered with a thick biofilm.  1351991) and the biofilm was treated as a black box.However, it is well known that in any open environ-ment biological system, multispecies biofilms exist.In these biofilms, different bacterial populations in-teract with each other. These interactions influencethe structure and the physiology of the biofilm as itdevelops (James et al. 1995). The evidence of directmetabolic interactions between community memberswas reported by Møller et al. (1998) who studied abinary population biofilm of   Pseudomonas putida  and  Acinetobacter   sp. degrading toluene. In a later work,Christensen et al. (2002) analysed the metabolic inter-actions of the above mentioned two-species microbialconsortium degrading benzyl alcohol. It was observedthat the two organisms exhibited competition and/orcommensal interactions depending on their relativephysicalpositioningin the biofilm. Thus, it was shownthat multispecies biofilms represent quite differenti-ated systems with various complex processes takingplace simultaneously.In order to identify microorganismsand to determ-ine their quantity in biofilms, the fluorescence in-situhybridization (FISH) method was developed (Amannet al. 1990) and has been successfully applied in orderto detect individual microbial cells without cultiva-tion (Amann et al. 1995). The FISH-method was usedto characterise the microbial populations of a trickle-bed bioreactordegradingpolyalkylatedbenzenesfromsynthetic waste gas (Hekmat et al. 1998; Stoffels etal. 1998). In these studies, the population dynamicsduring the start-up phase of the trickle-bed bioreactorwas examined. However, no data of the long-termoperation was obtained.Due to the above mentioned complexity, currentunderstanding of biofilm systems for waste gas treat-ment is limited. Only little work exists on the de-scription of the population dynamics in a multispeciesbiofilm of trickle-bed bioreactors for the treatmentof waste gas. In fact, very few examinations of themicrobial population behavior during long-term oper-ation under transient conditions of such systems havebeen carried out. Therefore, the aim of the presentstudy was to perform laboratory experiments undertransientconditionswitha trickle-bedbioreactorusingpolyalkylated benzenes as a model pollutant. It wasplanned to determine the population dynamics duringpollutant concentration shift experiments and during along-term starvation period. Description of model pollutant and bioreactorsystem The model pollutant was a mixture of hydrophobicpolyalkylated benzenes (Solvesso 100  , DeutscheExxon Chemical GmbH, Cologne, Germany). Thisproduct is commonly used in industrial applicationsas a solvent. Solvesso 100  contains mainly aromatichydrocarbons such as trimethylbenzenes or ethyltolu-enes (composition given in Hekmat et al. 1997).The inoculum of the bioreactor srcinated from thewastewater of a car painting facility (BMW AG, Mu-nich, Germany). The experiments were performed ina laboratory-scale trickle-bed bioreactor under non-sterile conditions (Hekmat et al. 1997). A schematicdiagramof the experimentalset-up is presentedin Fig-ure 1. The synthetic waste gas entered the columncontinuously from the top. The recirculating mineralsalt solution flowed co-currently with the gas down-ward through the column. The inner diameter of thecolumn was 140 mm and the height was 0.7 m res-ulting in a column volume of 10.8 l. The column waspackedwith hydrophilisedpolypropyleneRalu  -rings(Raschig AG, Ludwigshafen, Germany). A segmentof the trickle-bedcolumnduringcontinuouslong-termoperation and Ralu  -ring carrier material are presen-ted in Figure 2. As can be seen, the carrier materialis coveredwith a thick biofilm. Detailed specificationsof the packing material, GC analytics, and media aregiven in Hekmat et al. (1997).Preliminary experiments were performed in or-der to examine the component-specificbiodegradationof Solvesso 100  . It was shown that the standarddeviation of the degree of conversion of six major C 9 -components of Solvesso 100  with fractions higherthan 5% was about 30%. However for reasons of sim-plicity, the pollutant was treated as one single carbonsource during the subsequent investigations. Description of analytical methods, isolation of Solvesso 100  -degrading strains, and design of Solvesso 100  -specific oligonucleotide probes The fixation of liquid samples for FISH measure-ments was performed as reported by Wagner et al.(1993). Solid samples were obtained from biofilmsgrown on glass coverslips and Ralu  -ring carriers.Solid sampling ports existed at four positions alongthe reactor height as indicated in Figure 1. From eachport, glass coverslips or Ralu  -rings being distributed  136along the reactor cross section were removed. Thebiofilm of the carriers was removed and disintegratedin 0.9 M phosphate buffered NaCl solution (pH 7.4)by a combination of scraping and vigorous shaking,subsequent vortexing for 5 min, and final treatmentin an ultrasonic bath for 2 min. The biofilm suspen-sion was cooled in an ice bath inbetween the steps.In order to be able to differentiate the biomass intocells and extracellular polymeric substances (EPS),a quantitative EPS extraction method using a cationexchanger resin (Dowex 50X8, 20–50 mesh, Fluka,Neu-Ulm, Germany) was developed (Linn 1999). Thebiofilm of the carriers was removed and disintegratedusingtheabovementionedprocedure.60mlofbiofilmsuspension was then separated in a Sorvall centrifuge(Kendro Laboratory Products, Langenselbold, Ger-many) at 12000 min − 1 for 15 min at 4  ◦ C. 20 mlphosphatebufferedextraction solution (pH 7) contain-ing 30 g Dowex was added to the pellet. After shakingthis solutionfor3hat 4 ◦ C, thecentrifugationstepwasrepeated. The pellet was then separatedfromthe resid-ual Dowexandtheweightofthepelletwas determinedgravimetrically. Two parallel samples from 10 carrierswere treated every time. The standard deviation was ± 4.5%.Total cell counts (active and inactive cells) fromhomogenizedbiofilmsamples fromthreecarriersweredetermined by a combination of membrane filtra-tion and staining with 4  ,6-diamidino-2-phenylindole(DAPI) as described by Wagner et al. (1993). De-pending on the cell concentration, 5–20  µ l samplesolution was incubated in the dark with 50  µ l of a1  µ g ml − 1 DAPI solution in 100  µ l purified H 2 Ofor 15 min at room temperature. Vacuum filtrationunder sterile conditions in the dark for 5 min wasperformed using a 0.2  µ m GTTP-polycarbonate fil-ter (Millipore, Schwalbach, Germany). After dryingin the dark for 5 min, the filter was embedded inCitifluor solution (Citifluor Ltd., London, UK) andexamined with an Axioplan microscope (Carl Zeiss,Oberkochen,Germany)usingthefilterset01.Foreachsample, 25 randomly chosen microscopic fields werecounted. Mean values and standard deviations weredetermined. Color photomicrographs were taken withKodak Ektachrome P1600x films. For the applicationof the FISH-method, the biofilm suspension sampleswere fixed with paraformaldehyde. This fixation pro-tocol was adequate since almost no Gram-positivebacteria were present in the bioreactor system (Linn1999). Combined FISH and DAPI measurements of fixedsamples fromthreecarriersimmobilisedon glassslides were obtained by hybridization and subsequentDAPI staining usingthe abovementionedmicroscopicprocedure with filter sets 01, 09, and 15. The se-quences, hybridization conditions, and references forthe oligonucleotideprobes used in this study are givenin Table 1. The oligonucleotideswere synthesized andlabeled at the 5  terminus with fluorescein, CY3 orCY5 (Interactiva, Ulm, Germany).The isolation procedure of Solvesso 100  -degrading bacteria was as follows: serial solutions inthe range of 10 − 1 to 10 − 10 of biofilm suspension andcirculation liquid were plated in duplicate on mineralmedium (Hekmat et al. 1997). A filter paper soakedwith Solvesso 100  was placed at the bottom of aninverted glass petri dish and served as the sole carbonsource. Then, the dishes were incubated in an air-tightmetal box at 30  ◦ C. The filter paper was replenishedwith Solvesso 100  every 3–4 days. Pure cultureswere isolated by striking out single colonies. Twodistinct new Solvesso 100  -degrading strains wereisolated from a mature biofilm after about 2 months of bioreactoroperation(Linn1999).Theywereidentifiedas  Pseudomonas  sp. by sequencing of the 16S rRNAgene. The new isolates were named  Pseudomonas  sp.strain w20 (accession number: Y18344) and  Pseudo-monas  sp. strain g24 (accession number: Y18345).The 16S rRNA sequences were used to design twonew isolate-specific oligonucleotide probes Pw20-586and Pg24-586, both complementary to position 586-605 according to  E. coli -numbering (Brosius et al.1981) (see Table 1). Oligonucleotides of the selec-ted regions were synthesized and CY3-labeled at the5  -end.Hybridizationconditionsforthenewoligonuc-leotide probes were optimised by gradually increasingthe formamide concentration in the hybridization buf-fer as described previously (Manz et al. 1992). Thenewly designed probes Pw20-586 and Pg24-586 wereused together with the probe Bcv13b to monitor theSolvesso 100  -degrading bacteria. The latter probewas characterised by Stoffels et al. (1998) to be spe-cific for  Burkholderia vietnamensis  and  Burkholderiacepacia , both representing Solvesso 100  degradingbacteria. The probes BET42a and GAM42a were usedwith competitor oligonucleotides as described earlier(Manz et al. 1992).During preliminary experiments, the gradients of biofilm density/composition along the reactor heightwere examined. However, no significant gradientswere observed. Thus, it was deduced that the biomassdistribution and the microbial community composi-tion were practically homogeneous and non-variant  137 Table 1.  Oligonucleotide probes used in this studyProbe Specificity Target site, 1 % FA ReferencerRNA position in-situ 2 (probe sequence 5  to 3  )ALF968 Alpha subclass of 16S, 968–985 20% Neef 1997 Proteobacteria Bcv13b  Burkholderia  23S, 255–277 30% Stoffels vietnamensis , 3 et al. 1998  Burkholderiacepacia 3 BET42a Beta subclass of 23S, 1027–1043 35% Manz Proteobacteria  et al. 1992EUB338 Bacteria 16S, 338–355 0% Amannet al. 1990GAM42a Gamma-subclass 23S, 1027–1043 35% Manzof   Proteobacteria  et al. 1992Pw20-586  Pseudomonas  sp. 16S, 586–605 40% This studyw20 3 (ACATCCAACTTGCTGAACC)Pg24-586  Pseudomonas  sp. 16S, 586–605 30% This studyg24 3 (ACCTTCAACTTGCTGAACC) 1  Escherichia coli  numbering (Brosius et al. 1981). 2 Percent formamide (FA) in in-situ hybridization buffer. 3 Specific for Solvesso100  -degrading isolate. along the column height. Hence, the chosen samplingmethod ensured representative biofilm samples of theentire reactor. For further experiments, solid samplesfrom only one port (port 4, see Figure 1) were ex-amined. Performance of the trickle-bed bioreactor duringoperation with pollutant concentration shifts andduring a long-term starvation period The performance of the trickle-bed bioreactor un-der transient conditions was examined for a periodof approx. 9.5 months. During this time period, thepollutantgas inlet concentrationwas alteredinstantan-eously and then kept constant for several weeks (shiftexperiments). The bioreactor performance is charac-terised by the three parameters: specific pollutant load(PL), specific elimination capacity (EC), and degreeof conversion (DC). These parameters are defined asfollowsPL = c 0 /τ  EC = (c 0 − c 1 )/τ  DC = (c 0 − c 1 )/c 0 , where  c 0  is the gas inlet concentration,  c 1  is the gasoutlet concentration, and  τ   is the mean gas resid-ence time. The operating conditions for the transientexperiments with pollutant concentration shifts anda long-term starvation period are given in Table 2.The time courses of the specific pollutant load andthe specific elimination capacity are presented in Fig-ure 3. After a quasi-steady-state at  c 0  =  600 mg m − 3 had been reached for several weeks (average specificelimination capacity EC  =  80 g m − 3 h − 1 ), the firstshift-down to  c 0  =  200 mg m − 3 took place on the38th day. The gas inlet concentration was then keptconstant for 3 weeks (EC  =  38 g m − 3 h − 1 ). Thefirst shift-up occurred on the 55th day and the gasinlet concentration was then kept constant again at c 0  =  600 mg m − 3 . It was observed that the EC rosewithin a few hours up to the srcinal value of approx.80 g m − 3 h − 1 and stayed relatively constant at thislevel. Since this observedfast responsewas surprising,theresponsetimes ofabioticphysicalmechanismsandof biological processes were compared (Roels 1983).
Search
Similar documents
View more...
Tags
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks