Zinc chloride inhibition ofNitrosococcus mobilis

ARTICLE

Zinc Chloride Inhibition of Nitrosococcus mobilis

T.S. Radniecki,1 R.L. Ely1,2

1Environmental Engineering Program, Chemical Engineering Department, Yale University,

New Haven, Connecticut 06520-82862116 Gilmore Hall, Department of Biological and Ecological Engineering,

Oregon State University, Corvallis, Oregon 97331; telephone: 541-737-9409;

fax: 541-737-2082; e-mail: [email protected]

Received 4 June 2007; revision received 19 September 2007; accepted 19 September 2007

Published online 10 October 2007 in Wiley InterScience (www.interscience.wiley.com

ABSTRACT: Nitrosococcus mobilis, a halophilic nitrifier,plays an important role in global nitrogen cycling and inthe removal of nitrogen from wastewater treatment plants.However, ammonia oxidation is sensitive to a wide varietyof inhibitors, including the heavy metal, zinc. Using ametabolic-steady-state reactor, shotgun DNA microarrays,and quantitative polymerase chain reaction (qPCR), thisresearch looked at the dynamic physiological and transcrip-tional responses of N. mobilis to 1 and 10 mM ZnCl2. Byoxygen uptake rate measurements, zinc was determined toact directly on the ammonia monooxygenase (AMO)enzyme. The addition of excess copper prevented the inhibi-tion of AMO by ZnCl2 suggesting that zinc and coppercompete for placement in the metal active site in AMO.Shotgun DNA microarrays identified four previously unse-quenced genes that were up- or down-regulated in responseto 10 mM ZnCl2. Genes up-regulated in response to zincinhibition include methionine synthase I, UbiA prenyltrans-ferase and a recG-like helicase. RuBisCO was the lone down-regulated gene identified. qPCR was used to track the geneexpression of the identified genes over the course of the 4-hexperiment for both ZnCl2 concentrations. Because of theirphysiological importance, the expressions of AMO andhydroxylamine oxidoreductase (HAO) were also monitoredvia qPCR. The qPCR results showed general agreement withthe shotgun DNA microarray results for metH, UbiA, recGand RuBisCO, and revealed that AMO and HAO expressionlevels were maintained or modestly up-regulated duringZnCl2 inhibition.

Biotechnol. Bioeng. 2008;99: 1085–1095.

� 2007 Wiley Periodicals, Inc.

KEYWORDS: ammonia-oxidizing bacteria; shotgun DNAmicroarrays; qPCR; stress response; heavy metal toxicity;nitrification inhibition

T.S. Radniecki’s present address is School of Chemical, Biological and Environmental

Engineering, Oregon State University, Corvallis, OR 97331.

Correspondence to: R.L. Ely

� 2007 Wiley Periodicals, Inc.

Introduction

Ammonia-oxidizing bacteria (AOB) play a vital role inglobal nitrogen cycling by oxidizing ammonia to nitrite in atwo-step process; the ammonia monooxygenase (AMO)enzyme oxidizes ammonia to hydroxylamine, which is thenoxidized to nitrite by the hydroxylamine oxidoreductase(HAO) enzyme. This process, the sole source of energy forAOB, generates four electrons in the oxidation of hydro-xylamine, two of which are needed to support AMO activity;the remaining two electrons enter the electron transportchain (Kowalchuk and Stephen, 2001).Nitrosococcus mobilisis a halophilic AOB that has been reported to be either thedominant or typical AOB species in many nitrifying activat-ed sludge systems (Chen and Wong, 2004; Juretschko et al.,1998; Limpiyakorn et al., 2006; Mota et al., 2005; Purkholdet al., 2000; Rowan et al., 2003; Wagner et al., 1998).

AOB are generally considered to be the most sensitivemicrobes in the nitrification process, being readily inhibitedby various environmental factors and chemical compounds,including zinc, found in wastewater or contaminated sites(EPA, 1993; Hu et al., 2003; Juliastuti et al., 2003). Inhibitionor failure of biological nitrification in wastewater treatmentplants can cause high ammonia discharges and contribute toeutrophication of the receiving water bodies (EPA, 1993).Because current nitrification monitoring methods cannotidentify sources of inhibition, very sensitive methodscapable of detecting the initial stages of nitrificationinhibition and of identifying the source of the inhibitioncould help wastewater treatment plant operators controlnitrification process upsets. Such amethod potentially couldbe based on the detection of ‘‘signal genes’’ that are onlyexpressed in the presence of a particular contaminant orclass of contaminants.

Zinc has been used widely in applications including woodpreservation, catalysis, galvanization, batteries, paints,plastics, rubbers, fertilizers, and herbicides (EPA, 2005).The primary anthropogenic sources of Zn2þ in theenvironment are from metal smelters and mining activities(EPA, 2005). However, Zn2þ enters the environment

Biotechnology and Bioengineering, Vol. 99, No. 5, April 1, 2008 1085

throughmany routes, including from transportation-relatedsources such as car emissions (Hildemann et al., 1991),deicing salts, galvanized parts and railings (Barbosa andHvitved-Jacobsen, 1999; Legret and Pagotto, 1999), fuel andoil (Davis et al., 2001), brake linings and rubber tires(Hildemann et al., 1991). Construction materials such ascement, roofing and siding have been shown to leachZn2þ into the environment upon exposure to syntheticrain water (Davis et al., 2001). In studies examining theheavy metal composition of urban runoff, Zn2þ was themost prevalent heavy metal found (Davis et al., 2001;Rule et al., 2006). Since Zn2þ cannot be biodegraded, it isconsidered a permanent pollutant and increased additionsof Zn2þ from the sources listed above may impede nitrogencycling.

Because of its important role in nitrogen cycling and thelack of literature on its physiological and transcriptionalresponses to nitrification inhibitors, we undertook tocharacterize the inhibitory effects of Zn2þ on the physiologyof N. mobilis and to identify possible ‘‘signal genes’’ for zincinhibition. Real-time, dynamic physiological responses ofaxenic cultures to varying concentrations of ZnCl2 wereexamined in a metabolic-steady-state batch reactor. Nitriteproduction rates, specific ammonia-dependent oxygen up-take rates (AMO-SOUR) and hydrazine-dependent specificoxygen uptake rates (HAO-SOUR) were measured over a4-h ZnCl2 exposure period.

The genome of N. mobilis has not been sequenced.Therefore, to screen for signal genes, total RNA wasextracted both before and after exposure to 10 mM ZnCl2and hybridized to shotgun DNA microarrays. Microarrayshave been used previously to identify dominant AOB speciespresent in wastewater treatment plants and to characterizetranscriptional regulation of genes in response to ammoniaand carbon starvation, as well as to exposure to chloroformand chloromethane (Gvakharia et al., 2007; Kelly et al., 2005;Wei et al., 2006). Prior studies all were conducted usingorganisms with sequenced genomes. However, for organ-isms with unsequenced genomes, shotgun DNAmicroarrayscan be useful in identifying phase-specific gene regulation(Hayward et al., 2000; Hwang et al., 2003; Zaigler et al.,2003). We used quantitative reverse transcriptase polymer-ase chain reaction (qPCR) to confirm the shotgun DNAmicroarray results. Identified genes were analyzed at severaltime points throughout the experiment for exposure to both1 and 10 mM ZnCl2.

This work extends the knowledge of the physiologicaland transcriptional short-term responses of N. mobilis,an important AOB species, to the heavy metal zinc, arepresentative divalent cation. Four potential signal geneswere identified from shotgun DNA microarrays (methioninesynthase I, UbiA prenyltransferase, a RecG-like helicase andRuBisCO) and may help elucidate defense mechanismsemployed by N. mobilis. Because of their importance to theoxidation of ammonia, qPCR measurements of AMO andHAO gene expression in response to ZnCl2 inhibition wereperformed as well.

1086 Biotechnology and Bioengineering, Vol. 99, No. 5, April 1, 2008

Materials and Methods

Species and Culture Conditions

N. mobilis cells (courtesy of HP Koop, U. Hamburg,Germany, ATCC 25380) were cultured axenically at 308Cshaken (150 rpm) in the dark using a minimal mediapreviously described (Hyman et al., 1985). Once nitriteconcentrations reached 10–15 mM, about 1 week afterinoculation, the cells were determined to be in the mid-loggrowth phase, according to previously generate growthcurves, and were harvested by tangential flow filtration.

Experimental Set-Up

The reactor vessel, a Wheaton Double-Sidearm Cellstir(Fig. 1), had a total volume of 1.67 L, with 375 mL of cellsuspension. The top of the reactor was sealed with a Teflon-lined, gasketed cap. The two upper reactor sidearms weresealed with screw-cap, septum vial caps fitted with Teflon-gasketed Mininert valves. The lower sidearm was sealed witha hypo-vial Mininert valve (Supelco, Sigma-Aldrich,St. Louis, MO). A 3-inch, Teflon-coated magnetic stirringbar operated at medium-high speed by a magnetic stirrerprovided uniformmixing and aeration of the cell suspensionusing headspace O2.

To achieve metabolic-steady-state in the reactor, anammonium sulfate solution [5.45 mM (NH4)2SO4, 30 mMHEPES buffer (pH 8.0)] was fed at a very low rate through a21 ga needle in the upper sidearm using a Cole ParmerMasterflex Synchronous Drive Pump (5 rpm) and Tygontubing (size 13). The final feed rate was 3.4 mM (NH4)2SO4/min. Replicate reactors were started from freshN. mobilis seed for each experiment.

Determination of Metabolic-Steady-State in a BatchReactor and Sampling Methods

Nitrite assays were used to monitor N. mobilis nitriteproduction. At the start of an experiment, 4 mL sampleswere taken from the bottom sidearm of the reactor every20 min. Nitrite concentrations and OD600 were measured,and the specific nitrite production rate was calculated bytaking the difference in NO�

2 concentrations and dividingby the OD600. Metabolic-steady-state was indicated bya constant specific nitrite production rate over threeconsecutive time-points (100% nitrification activity), afterwhich a pulse of ZnCl2 was added through the bottomsidearm of the reactor (t¼ 0). Nitrite accumulation wasminimal (<2 mM) and volume changes due to NHþ

4 feedwere offset by samples taken for NO�

2 and SOUR assays. Thetotal exchange of reactor volume was less than 10% duringan experiment. Nitrite concentration and OD600 weremeasured at the following time points: t¼ 10 min (10 min

Figure 1. Experimental setup for conducting metabolic-steady-state reactor experiments.

after the ZnCl2 pulse addition), 60, 105, 150, 195, and240 min.

Nitrite Concentration Assay

One-milliliter cell suspensions were centrifuged at 16,000rpm for 1 min. Triplicate 10 mL aliquots of supernatant wereremoved and analyzed for nitrite concentration spectro-photometrically (Hageman and Hucklesby, 1971).

Ammonia-Dependent Specific Oxygen Uptake Rate(AMO-SOUR) Assay

Two milliliter cell suspensions were removed from eachreactor and injected directly into the evacuated 1.8 mLchamber of an O2 electrode apparatus consisting of a water-jacketed glass cell (Gilson Medical Electronics, Inc.,Middleton, WI) fitted with a Clark microelectrode(Model # 5331, Yellow Springs Instrument Co., YellowSprings, OH) attached to a YSI Model 5300 BiologicalOxygen Monitor (Yellow Springs Instrument Co.) and aflatbed recorder. The glass cell was held at 308C using aheated circulating water bath. The consumption of O2 wasmeasured over time by slope of the line generated on theflatbed recorder and normalized to the OD600. Units ofAMO-SOUR are mM O2/min-OD600. The AMO-SOURmeasured during the metabolic-steady-state, as definedabove, was considered to be 100% AMO-SOUR activity.

Hydrazine-Dependent Specific Oxygen Uptake Rate(HAO-SOUR) Assay

To evaluate whether zinc effected only AMO or othermetabolic processes, experiments were conducted at selectedtime points by removing a 2 mL cell suspension from the

R

reactor, adding it directly to the SOUR measurementapparatus, blocking AMO activity with allyl thiourea (ATU;100 mM), and adding hydrazine (750 mM) as an alternativesubstrate for HAO. Hydrazine-dependant specific oxygenuptake rates (HAO-SOUR) were determined as describedabove (Ely et al., 1995). Units of HAO-SOUR are mMO2/min-OD600. The HAO-SOUR measured during themetabolic-steady-state, as defined above, was considered tobe 100% HAO-SOUR activity.

Zinc Concentration Measurements

Because Zn2þ is non-biodegradable and there was nooutflow from the reactor, small changes (<10%) in the Zn2þ

concentration occurred during an experiment because ofremoval of samples for analyses. At the end of eachexperiment, 10mL of cell suspension were filtered through a0.2 mm syringe filter and the filtrate zinc concentration wasmeasured using the Zincon Method Quick-n-Easy TestOutfit Kit (Orbeco Analytical Systems, Inc., Farmingdale,NY) per manufacture’s instructions.

CuSO4 Protection of N. mobilis to ZnCl2 Inhibition

To determine if excess CuSO4 would protectN. mobilis fromZnCl2 inhibition, a series of AMO-SOUR experiments wereconducted. Batch-grown N. mobilis cells were harvestedduring the mid-exponential growth phase, washed in 30mMHEPES buffer, pH 7.8 and placed into the O2 electrodeapparatus, as described above. Control experiments con-tained 25 mM (NH4)2SO4 and selected amounts of CuSO4;treatment experiments contained 10 mM ZnCl2 and thesame amounts of (NH4)2SO4 and CuSO4 as in the controlexperiments.

adniecki and Ely: Zinc Chloride Inhibition of Nitrosococcus mobilis 1087

Biotechnology and Bioengineering

Gene Library Construction

N. mobilis genomic DNA was extracted using a PromegaWizard Genomic DNA Purification Extraction kit (Madi-son, WI) per manufacturer’s instructions. The DNA waspartially digested with Sau3AI and size fractionated on a1.2% agarose gel. One kilobase fragments were gel purifiedand ligated into a BamHI-cut Strategene pBluescript skIIþvector (La Jolla, CA). The 1 kb fragment size was chosen tolimit the likely number of gene sequences to be 1 or 2 for anygiven fragment. The ligation mixture was size fractionatedon a 1.2% agarose gel and vectors containing 1 kb insertswere gel purified and electroporated into TOP10 Electro-comp Escherichia coli (Invitrogen Co., Carlsbad, CA). Blue/white screening on LB agar plates containing 100 mg/mLampicillin and 40 mg/mL X-Gal was used to selecttransformed E. coli containing a pBluescript vector withan insert. Colonies containing a vector with an insert werepicked and placed into a 96-well plate containing 100-mLLB, 7-mL glycerol, and 10-mg ampicillin, shaken and grownovernight at 378C. The 96-well plates were placed at �808Cfor long-term storage.

Colony PCR Amplification and Purification

One microliter of transformed E. coli was placed in a PCRtube and lysed by microwaving on high power for 60 s. Afterthe cells were lysed, 30 mL of PCR master mix containing0.25 U Taq polymerase, 1� Taq Polymerase Buffer, 6 nMdNTPs, 5 nM M13 Forward Primer (50-GTAAAACGA-CGGCCAG-30), and 5 nM M13 Reverse Primer (50-CAGGAAACAGCTATGAC-30) were added to the PCRtube. The M13 primers were designed to hybridize to theStrategene pBluescript skIIþ vector on either side of theinserted DNA fragment. Samples were amplified in aThermo Hybaid thermocycler (Franklin, MA) as follows:60 s at 378C, 1 min at 958C, then 30 cycles at 958C for 30 s,508C for 1 min, and 728C for 3 min followed by the finalcycle of 728C for 7 min. Two-microliter aliquots of all PCRproducts were analyzed on a 1.2% agarose gel for qualitycontrol, PCR products were purified by ethanol/isopropanolprecipitation, and the purified PCR product pellet wassuspended in 30 mL of TE buffer at pH 8.0.

Shotgun DNA Microarray Construction

To help ensure that every gene was represented at least once,microarrays in this current research were designed forfivefold coverage of the N. mobilis genome and contained30,000 features. Colony PCR products were precipitated viaisopropanol precipitation and suspended in 30 mL ofmicroarray spotting buffer containing 3� SSC (sodiumchloride and sodium citrate) and 1.5 M betaine. Alien DNAcontrols (Stratagene, La Jolla, CA) were added in a 10-folddilution series in three places on the microarray. ABioRobotics MicroGrid II (Ann Arbor, MI) robot was used

1088 Biotechnology and Bioengineering, Vol. 99, No. 5, April 1, 2008

to print the 29,913 features onto aminosilane coatedCorning Ultragap microarray slides (Corning, NY). Thespot diameter was 150 mm and the spot-to-spot centerspacing was 175 mm. The microarray features were cross-linked to the microarray by placing in an UV Stratalinker2400 (Stratagene) at 225 mJ.

Total RNA Extraction and Preparation

Ten milliliters of N. mobilis cell suspension were extracted att¼ 0 min, t¼ 10 min, t¼ 60 min and t¼ 240 min for eachbioreactor experiment. The cells were immediately added to20 mL of RNAlater RNA Stabilization Reagent (QiagenSciences, Valencia, CA) and pelleted by centrifugation at4,000 rpm for 10 min. The resulting RNAlater supernatantwas decanted and the total RNA was extracted from thecell pellet using the Qiagen RNeasy Kit (Valencia, CA) permanufacturer’s instructions. The extracted RNA wasquantified and diluted to 1 mg total RNA/mL in RNase-freewater. RNA quality was analyzed using the AgilentBioanalyzer 2100 (Palo Alto, CA).

cDNA Synthesis, Labeling, and Hybridization

Labeled cDNA was synthesized from 3 mg of total RNA,extracted at t¼ 0 min and t¼ 60 min, from each 10 mMZnCl2 reactor experiment and spiked with Alien DNA(Stratagene) using an indirect labeling method provided inthe Genisphere 3DNA Array 900 MPX kit (Genisphere, Inc.,Hatfield, PA) using total RNA with control sample labeledwith Cy3 and experimental sample labeled with Cy5. Allsteps were according to the Genisphere protocol with thecDNA hybridization at 428C and the fluorescent dyehybridization at 478C. Replicate microarrays were hybri-dized at the same time under the same conditions to helpminimize variations in efficiencies.

Analysis of Shotgun DNA Microarray Data andSelection of Features to Sequence

Microarrays were scanned using a Packard ScanArray 4000scanner (Perkin Elmer Inc., Wellesley, MA) and analyzedusing QuantArray software v. 3.0.0.0 (Packard Biosciences,Perkin Elmer, Inc., Wellesley, MA). The BASE database v.1.2.15 (Saal et al., 2002) archived all information generatedby the microarray experiments. Mean feature intensitiesnormalized by the Lowess normalization method (Clevelandand Devlin, 1988; Yang et al., 2002) were used for featureexpression pattern comparisons. The Lowess normalizationmethod is a global normalization method that uses a locallyweighted least squares regression and is a useful method formicroarray analysis when housekeeping genes in unse-quenced organisms are unknown. The top ten up- anddown-regulated features, showing at least a twofold changein gene expression, homogeneous spot hybridization, round

spot morphology as well as signal intensities above 1,000units, and signal-to-noise ratios�3, were selected for furtheranalysis.

Sequencing of Selected ShotgunDNA Microarray Features

Clones of the selected features were picked from the genelibrary and grown overnight in deep-well 96-well blocksshaken at 378C in 1 mL of LB containing 100 mg/mLampicillin. The cells were pelleted at 1,000 rpm for 10 minand the resulting supernatant was decanted. Plasmids wereextracted from the pelleted cells using the Qiagen Bio Robot3000 (Valencia, CA) and the Macherey-Nagel NucleoSpinRobot-96 Plasmid Kit (Duren, Germany) per manufac-turer’s instructions. Four hundred nanograms of purifiedplasmid were sequenced using an ABI 3730 capillarysequence machine (Foster City, CA) and the previouslydescribed M13 forward and reverse primers. Full-lengthsequences are available upon request.

BLASTx Analysis and E-Value Homologs

For the initial annotation of shotgun DNA microarrayfeatures, 150 bases were deleted from the 30 and 50 ends ofthe sequences to remove the vector sequence. The sequencesgenerated were compared against available protein databasesusing BLASTx (Gish and States, 1993). Features thatprovided the lowest E-values (the probability due to chancethat there is another alignment with a similarity greater thanthe one currently being examined) for both down and upregulated genes were selected for qPCR confirmation(Table I).

Quantitative PCR Assays

Gene Runner v. 3.00 (Hastings Software, Inc., Hastings-on-Hudson, NY) was used to generate qPCR primers (Table II)from DNA sequences generated from shotgun DNAmicroarray results. The qPCR primers were optimized forconcentration, annealing temperature, and MgSO4 con-centrations via PCR to achieve high efficiency with only oneproduct detected. cDNA was generated from 1 mg of totalRNA using the iScript cDNA Synthesis kit (Bio-Rad,

Table I. BLASTx results.

Microarray feature namea Relative regulation BL

NmoOSU0015622 Up UbiA prenyltransferase

NmoOSU0024465 Up metH Methionine synt

NmoOSU0022499 Up RecG-like helicases

NmoOSU0025101 Down RuBisCO

aName of feature spotted onto the shotgun DNA microarray slides.bAccession number for BLASTx homologs.

R

Hercules, CA) per manufacturer’s instructions and diluted100-fold in TE buffer at pH 8. qPCR was carried out intriplicate on an ABI Prism 7000 Sequence Detection System(Applied Biosystems, Foster City, CA) using the iQ SYBRGreen Supermix kit (Bio-Rad). The 50-mL reactionscontained 1X SYBR Green Supermix, 1� ROX referencedye (Invitrogen), 0–10 mM of additional MgSO4 (asoptimized for each primer pair), 500 nM Forward andReverse primers and 10 mL of 1/100 diluted cDNA. Alltriplicate reactions were carried out in duplicate for each ofthe bioreactor runs analyzed. The following cycle conditionswere used for qPCR; 2 min at 958C, then 50 cycles at 958Cfor 30 s, 46.0–53.48C (as optimized for each primer pair) for45 s and 728C for 45 s. At the end of the reaction,dissociation curves of the products were generated bybringing the temperature to 608C and raising thetemperature by 0.58C every 20 s until a final temperature of958C was attained.

Reactions showing more then one product in theirdissociation curves were not analyzed further. The relativeexpression values for the remaining reactions were deter-mined using DART-PCR analysis (Peirson et al., 2003)taking into account the efficiency of the reaction andnormalizing the data to the amount of 16S RNA present ineach reaction.

Results

Nitrite Production and Specific Oxygen Uptake Rates

N. mobilis cells in metabolic-steady-state exposed to 10 mMZnCl2 showed an immediate decrease in activity and did notexhibit noticeable recovery over 4 h of exposure, resulting ina 100% decrease in nitrite production rates and an 86%decrease in AMO-SOUR (Figs. 2 and 3). In contrast, HAOactivity measurements showed a decrease in activity of only24% over the 4-h experiment (Fig. 4).

N. mobilis cells in metabolic-steady-state exposed to 1mMZnCl2 showed a slight decrease of 11% in the nitriteproduction rate over the 4 h experiment (Fig. 2). However,the AMO-SOUR of the cells appeared to increase steadilyafter exposure to 1 mMZnCl2, showing an apparent increaseof 26% in 4 h (Fig. 3). HAO activity measurements showed adecrease in activity of 25% over the 4 h experiment (Fig. 4).

ASTx results E-value Accession numberb

3.00E�91 NP_841083

hase I cobalamin-binding domain 3.00E�78 NP_841658

2.00E�69 NP_841872

3.00E�50 AAG39458

adniecki and Ely: Zinc Chloride Inhibition of Nitrosococcus mobilis 1089

Biotechnology and Bioengineering

Table II. qPCR primers.

Primer name Primer sequence

16S For 50-CGCATCGAAAGATGTGCTAA-30

16S Rev 50-GTTGGTCGTCCTCTCAGACC-30

UbiA For 50-CACGGTCGGTATTTCTCTGGTAG-30

UbiA Rev 50-GGTAACGGTCTGCCCTTTGTAC-30

metH For 50-ATCTTTCAGGGCAGACACCG-30

metH Rev 50-GGCGACACTTGGNAACTCTTCG-30

recG For 50-CCGCAGTCAAAGTAAACGCAAG-30

recG Rev 50-CGCCAGCATTATCACCAACG-30

RuBisCO For 50-GTTACCAATAATGGAAGCGGTTAAG-30

RuBisCO Rev 50-TATCGTGCCAAAGCATATAAGTCC-30

AMO For 50-TGGCGACATACCTGTCACAT-30

AMO Rev 50-ACAATGCATCTTTGGCTTCC-30

HAO For 50-CAAACTTGCCGAAATGAACC-30

HAO Rev 50-GCTGGTGATGTTCTCTGCAA-30

Figure 3. AMO-SOUR activities of N. mobilis. At t¼ 0 min, ZnCl2 was injected

into the reactor. Error bars indicate 95% confidence intervals.

Figure 5 displays the difference in SOUR activityversus NO�

2 activity. An uncoupling of the NO�2 :O2

stoichiometry is observed after N. mobilis is exposed toboth 1 mM ZnCl2 and 10 mM ZnCl2 with the differencein NO�

2 :O2 stoichiometry being greater at 1 mM ZnCl2.

Protective Properties of CuSO4

To determine if copper could protect N. mobilis from zinc-mediated inhibition, competition experiments were con-ducted. N. mobilis was exposed to 10 mM ZnCl2 and a rangeof CuSO4 concentrations from 0.65 to 130 mM and theirAMO-SOURs were measured. These results were comparedwith controls containing N. mobilis and the various CuSO4

concentrations but no ZnCl2. As can be seen in Figure 6, a100-fold increase in CuSO4 from 0.65 to 65 mM resultedin an increase in AMO-SOUR activity from 44% to 88%.A further increase in CuSO4 concentrations to 130 mM

Figure 2. Nitrite production activities of Nitrosococcus mobilis. At t¼ 0 min,

ZnCl2 was injected into the reactor. Error bars indicate 95% confidence intervals.

1090 Biotechnology and Bioengineering, Vol. 99, No. 5, April 1, 2008

resulted in an increase in AMO-SOUR to 98% activity. Incontrast, the addition of 130 mM CuSO4 in the controlsresulted in a 17% decrease in activity over cells exposed to0.65 mM CuSO4 during this experiment. The protectiveproperty of CuSO4 against ZnCl2 inhibition suggeststhat Cuþ and Zn2þ may compete for the same bindingsites on AMO.

Identification and Selection of Differentially ExpressedFeatures From 10 mM ZnCl2 Reactor Experiments

Although labeling and hybridization conditions were main-tained as identical as possible, replication of up- and down-regulated features was low (11 of 139 up-regulated and 6 of33 down-regulated features) in duplicate microarray hybri-dizations. Part of the reason for this phenomenon may bethe difficulty of cost-effectively purifying 30,000 colonyPCR products. A simple isopropanol/ethanol precipitation

Figure 4. HAO-SOUR activities of N. mobilis. At t¼ 0 min, ZnCl2 was injected into

the reactor. Error bars indicate 95% confidence intervals.

Figure 5. Uncoupling of NO�2 :O2 stoichiometry after exposure to ZnCl2, t¼ 0 min.

& SOUR at 1 mM ZnCl2. & NO�2 production rate at 1 mM ZnCl2. * Difference in SOUR

and NO�2 production rate at 1 mM ZnCl2. ~ SOUR at 10 mM ZnCl2. ~ NO�

2 production

rate at 10 mM ZnCl2. ^ Difference in SOUR and NO�2 production rate at 10 mM ZnCl2.

method was used to purify the colony PCR products. Duringthe purification procedure, macromolecules (proteins, cellwall, cell membranes, etc.) may have precipitated with thecolony PCR product. The spotting of these macromoleculesmay have affected the spot morphologies and hybridizationefficiencies. To score microarray data and more objectivelyselect features for subsequent experiments, a simplealgorithm was developed (Eq. (1)) that multiplies eachsignal intensity of duplicate features by a subjective factorbased on the quality of the feature:

Score ¼ I1 � Q1 þ I2 � Q2 (1)

where I1 is the signal intensity for a feature from microarray#1; I2 the signal intensity for a feature from microarray #2;Q1 the quality score given to a feature from microarray #1;Q2 the quality score given to a feature from microarray #2.

Quality scores were based on four factors: featuremorphology, homogeneity, ‘‘moons’’ (high intensity small

Figure 6. Results of the effect of CuSO4 concentrations on AMO-SOUR of

N. mobilis cells exposed to 10 mM ZnCl2.

R

spots, probably from either residual genomic DNA orproteins contaminating the feature), and background noise(Fig. 7). Each factor was assigned a value of 0 (for poormorphology or poor homogeneity of signal or the presenceof ‘‘moon’’ or the presence of high background noise) or0.25 (for nice round and contained morphology or highhomogeneity of signal or the absence of ‘‘moons’’ or theabsence of background noise), yielding a Q-score for eachfeature of 0 (for no feature), 0.25, 0.5, 0.75, or 1 (for a perfectfeature). Equation (1) gave positive values for up-regulatedfeatures and negative values for down-regulated features.The 10 features with the most positive scores and the tenwith the most negative scores were selected for BLASTxanalysis.

BLASTx Analysis

Selected features were purified from the shotgun gene libraryand sequenced, and BLASTx was used to find homologs ofthe sequenced features. Although most of the top BLASTxhits were from N. europaea, the only AOB sequenced at thetime of this study, a wide variety of genes from otherprokaryotes appeared as well. Up- and down-regulatedfeatures with the lowest E-values (Table I) were selected forqPCR analyses.

Quantitative PCR Assay Analysis

Using segments of selected up- or down-regulated featuresthat showed high similarity to potential homologs, qPCRprimers were designed to produce small PCR products,ranging from 150 to 250 base pairs (Table II), for methio-nine synthase I (metH), UbiA prenyltransferase (UbiA), arecG-like helicase (recG), and ribulose-1,5-bisphosphatecarboxylase/oxygenase (RuBisCO). Changes in transcrip-tional regulation of the four genes, as well as of AMO andHAO, varied with ZnCl2 concentration and exposure time(Fig. 8). metH (a zinc binding ligand), UbiA (an electrontransport chain protein), recG (generally involved in DNArepair), and HAO showed generally higher transcriptionafter 60 and 240 min of exposure to 10 mM ZnCl2 than with1 mM ZnCl2, indicating that the cells could respondtranscriptionally although their nitrite production wasseverely limited. UbiA and recG were the most up-regulatedgenes tested at 4.4� 0.49 and 4.9� 0.39, respectively, after 4h of exposure to 10 mM ZnCl2. metH was significantly up-regulated after 4 h of exposure to both 1 mM and 10 mMZnCl2, 2.3� 0.10 and 2.9� 0.49, respectively. RuBisCO wasthe only down-regulated gene selected from the shotgunDNA microarrays and showed significant down-regulationto both 1 and 10 mM ZnCl2, �4.9� 0.49 (at 240 min) and�3.4� 0.48 (at 60 min), respectively. AMO showed up-regulation after 4 h of exposure to both concentrations ofZnCl2 tested, 2.2� 0.2 and 2.1� 0.45, respectively, althoughit was initially down regulated in response to 10 mM ZnCl2,

adniecki and Ely: Zinc Chloride Inhibition of Nitrosococcus mobilis 1091

Biotechnology and Bioengineering

Figure 7. A: The perfect feature with high intensity, good morphology, good homology, no moons, and a low background. B: A feature containing a moon. C: A feature showing

poor homology. D: A feature showing poor morphology (most likely due to contaminating proteins or DNA). E: A feature with a high background present.

while HAO showed a greater up-regulation to 10 mMZnCl2,3.2� 0.01, than to 1 mM ZnCl2, 1.8� 0.1.

Discussion

N. mobilis displayed a high sensitivity to the divalent cationheavy metal Zn2þ with little or no recovery of activityobserved after 4 h of exposure, consistent with reports formixed cultures or other nitrifying species (Braam andKlapwijk, 1981; Hu et al., 2003; Lee et al., 1997). Dramaticdecreases in nitrite production were seen in pure cultures ofN. mobilis in response to very low amounts of Zn2þ (11%decrease at 1.0 mM ZnCl2 and 100% decrease at 10 mM

Figure 8. qPCR results of selected N. mobilis genes after 10 min &, 60 m

1092 Biotechnology and Bioengineering, Vol. 99, No. 5, April 1, 2008

ZnCl2). This sensitivity to Zn2þ is much higher thanpreviously reported with activated sludge (Jonsson et al.,2001; Lee et al., 1997; Panswad and Polprucksa, 1998;Tomlinson et al., 1966), but is in the reported range for testsconducted on pure cultures of N. europaea (Juliastuti et al.,2003; Loveless and Painter, 1968; Meiklejohn, 1954; Satoet al., 1988; Skinner and Walker, 1961; Tomlinson et al.,1966).

The shotgun DNA microarrays identified four previouslyunsequenced genes that were transcriptionally regulatedin response to ZnCl2 inhibition (metH, UbiA, recG, andRuBisCO). qPCR allowed the relative quantification of thesegenes to various ZnCl2 concentrations over the course ofthe 4-h experiment. The regulation of these genes and

in , and 240 min & of exposure to 1 mM ZnCl2 (A) and 10 mM ZnCl2 (B).

the physiological data generated during the experimentssuggested how Zn2þ inhibits N. mobilis and how the cellscombat Zn2þ inhibition.

The decrease in hydrazine-dependent oxygen uptake byHAO in response to ZnCl2 appeared to be independent ofthe ZnCl2 concentrations used in this work (1–10 mM). Ascan be seen in Figure 4, a 10-fold increase in ZnCl2 caused nosignificant increase in HAO inhibition even though it causeda 10-fold decrease in the rate of nitrite production. Zn2þ isknown to be a potent inhibitor of the electron transportchain (Beard et al., 1995; Chvapil, 1973), consistent with thepossibility that decreases in hydrazine-dependent oxygenuptake activity by HAO after exposure to ZnCl2 could haveresulted from Zn2þ inhibition of the electron transportchain of N. mobilis. This hypothesis is further supported bythe observation of the up-regulation of UbiA and HAO inresponse to 10 mM ZnCl2 (Fig. 8). Production of excessUbiA and HAOmay be a method employed by N. mobilis toovercome zinc-mediated damage to the electron transportchain.

Excess Zn2þ is also known to damage DNA. Thesignificant up-regulation of recG in response to both slightand severe ZnCl2 inhibition (Fig. 8) suggests that Zn2þ maydamage N. moblis DNA and be partially responsible for theobserved inhibition.

Because the inhibition of HAO appeared to reach amaximum after the addition of 1 mMZnCl2 while inhibitionof nitrite production and AMO-SOUR continued toincrease with the addition of more ZnCl2, three distinctinhibitionmechanismsmay be suggested for ZnCl2. The firstmay be via disruption of the electron transport chain, thesecond by damage to DNA, and the third from Zn2þ

competion with Cuþ for the metal active site in AMO.Several heavy metals have been shown capable of displacingessential divalent cations in biologically important enzymes,rendering the enzyme nonfunctional (Chvapil, 1973; Nies,1999). Zn2þ is unable to undergo redox reactions (Amoret al., 2001; Nies, 1999) and thus the replacement of Cuþ

by Zn2þ in the active site of AMO would make AMO unableto oxidize NH3 to NH2OH.

To examine this possibility, whole-cell Cuþ/Zn2þ

experiments were conducted. The results in Figure 6 showthat increases in CuSO4 concentration did protectN. mobilisfrom ZnCl2 inhibition, consistent with the hypothesisthat Zn2þ was replacing Cuþ in the binding site of AMO.Surprisingly, even though AMO was being renderedbiologically useless by Zn2þ, N. mobilis appeared to onlyslightly up-regulate AMO transcription after 4 h of exposure(Fig. 8). It seems unlikely that Cuþ would similarly protectN. mobilis activity from other heavy metals unless they werecompeting with Cuþ for binding sites on AMO.

After exposure to 1 mM ZnCl2, N. mobilis demonstratedan unexpected increase in oxygen consumption even thoughthe nitrite production rate decreased (Figs. 2 and 3). This issurprising as NO�

2 production and AMO-SOUR should bestoichiometrically related; O2 would be expected only to beconsumed in substantial quantities for direct oxidation

R

of NH3 to NO�2 . A comparison of O2:NO

�2 stoichiometry

showed changes after exposure to ZnCl2 at both concentra-tions tested (Fig. 5). Thus, exposure to ZnCl2 resultedin uncoupling of NO�

2 :O2 stoichiometry at both low andhigh inhibitory levels. Abiotic consumption of oxygen dueto Zn2þ hydroxide complexes was measured and found to beinsignificant (data not shown). Amass balance of NHþ

4 /NH3

and NO�2 indicated that neither O2 uptake for oxidation

of NO�2 nor conversion of NHþ

4 to another product couldhave accounted for the observed shift in O2:NO

�2 stoichio-

metry. An ostensible uncoupling of oxygen uptake fromammonia oxidation, such as observed in these experiments,suggests uncertainty and/or possible imprecision in methodsthat rely exclusively on respirometric measurements toassess the health and activity of nitrification processes.

A mechanism that could contribute to increased O2

uptake would be metal detoxification by complexation.Complexation of heavy metals with glutathione in gram-negative bacteria results in the formation of bisglutathionatecomplexes (Kachur et al., 1998). The complexes then canreact with molecular O2, resulting in oxidized bisglu-tathione, the heavy metal cations, and hydrogen peroxide.The heavy metal cations are free once again to react withanother glutathione molecule, repeating the above processwhile consuming O2 and exerting stress on the cell (Kachuret al., 1998).

For slow-growing bacteria, complexation can be anefficient defense mechanism against low concentrations ofheavy metals (Nies, 1992, 1999). N. mobilis grows slowly,with a doubling time of 12–24 h under ideal conditions(Watson et al., 1981), and the metal concentrations used inthese experiments, 1 and 10 mM, were relatively low. Thesefactors, together with the observed up-regulation of metH, azinc binding ligand, could suggest that N. mobilis relies, atleast partly, on complexation for heavy metal detoxification,and that the reaction of O2 with ligand-metal complexescould account for at least some of the observed uncouplingof O2 uptake and NO�

2 production.RuBisCO was the only down-regulated gene selected from

the shotgun DNA microarray data for further analysis.N. europaea has been shown previously to down regulate68% of its genome, particularly energy-intensive processessuch as carbon sequestration by RuBisCO, during starvationexperiments (Wei et al., 2004, 2006) and to transcribe amoand hao, two key enzymes in energy harvesting, preferen-tially during periods of energy starvation (Hyman and Arp,1995; SayavedraSoto et al., 1996). Together, these transcrip-tional responses may allow N. europaea to conserve energyduring periods of starvation while remaining ready tometabolize NH3 immediately when it becomes available.Our data suggest that most of the ZnCl2-mediated inhi-bition in N. mobilis affected AMO directly (Figs. 3 and 4),causing energy availability to be limited in the presenceof ZnCl2. In response, Figure 8 shows that N. mobilisdown-regulated RuBisCO while generally maintaining orslightly up regulating AMO and HAO and Zn2þ-resistancemechanisms, such as complexation proteins (metH), DNA

adniecki and Ely: Zinc Chloride Inhibition of Nitrosococcus mobilis 1093

Biotechnology and Bioengineering

repair (recG), and electron transport chain repair (UbiA).Therefore, the approach employed by N. mobilis duringZnCl2 inhibition appears to be partly an energy limitationresponse, similar to that of N. europaea, but also a moregeneral heavy metal response.

This study disclosed some potential signal genes ofZnCl2-related stress in N. mobilis cells. Evaluating the extentto which these genes reflect specific responses to Zn2þ

exposure, more general heavy metal inhibition responses, oran even more general stress responses will require moreextensive research. Interestingly, a recent whole-genomemicroarray study with N. europaea inhibited by two similarchlorinated methanes, chloromethane and chloroform,resulted in 37 shared up-regulated and 82 shared down-regulated genes (Gvakharia et al., 2007). However, exposureto chloroform resulted in 138 up-regulated genes not up-regulated in chloromethane-exposed cells, and cells exposedto chloromethane up regulated 30 genes not up regulated inchloroform-exposed cells. Likewise, cells exposed to chloro-form down regulated 419 genes not down regulated in cellsexposed to chloromethane, while chloromethane-exposedcells down regulated 66 cells not down regulated inchloroform-exposed cells. These results suggest that AOBcan have gene expression patterns unique to inhibitors withsimilar chemical structures and, therefore, that specificsignal genes may be able to detect and identify particularnitrification inhibitors. With the signal genes identified,reporter strains could be constructed, perhaps using greenfluorescent protein (GFP) or another indicator, for example,that could be used as stress-indicating biosensor. Alter-natively, perhaps an instrument could be developed to usethe signal genes and a reporter, perhaps in a diode arrayarrangement with other signal genes, to serve a similarfunction. Such biosensors could be used to provide sensitiveand early detection and diagnoses of impending problems innitrification processes.

Thanks are due to Brad Dubbles for his help with the growth and

harvesting of the large volumes of N. mobilis necessary for the

experiments. Thanks are also due to Norman Hommes and Luis

Sayavedra-Soto for their valuable discussions and general expertise on

AOB. Special thanks are due to Daniel J. Arp for his valuable expertise

on AOB and discussions regarding the content and presentation of

this manuscript. Thanks are due to Caprice Rosato and Scott Givan at

the Oregon State University Center for Genome Research and Bio-

computing for help in design, construction, implementation, and

analysis of the shotgun DNA microarrays. This research was sup-

ported by funding provided to R. Ely by the Yale Faculty of Engineer-

ing, the Hellman Family Fellowship (Yale University), and Dean’s

Award (Yale University) and by the Department of Biological and

Ecological Engineering (Oregon State University). This material is also

based upon work supported under a National Science Foundation

Graduate Fellowship awarded to Tyler Radniecki.

References

Amor L, Kennes C, Veiga MC. 2001. Kinetics of inhibition in the biode-

gradation of monoaromatic hydrocarbons in presence of heavy metals.

Bioresour Technol 78(2):181–185.

1094 Biotechnology and Bioengineering, Vol. 99, No. 5, April 1, 2008

Barbosa AE, Hvitved-Jacobsen T. 1999. Highway runoff and potential for

removal of heavy metals in an infiltration pond in Portugal. Sci Total

Environ 235(1–3):151–159.

Beard SJ, Hughes MN, Poole RK. 1995. Inhibition of the cytochrome bd-

terminated NADH oxidase system in Escherichia coli K-12 by divalent

metal cations. FEMS Microbiol Lett 131(2):205–210.

Braam F, Klapwijk A. 1981. Effect of copper on nitrification in activated

sludge. Water Res 15(9):1093–1098.

Chen GH, Wong MT. 2004. Impact of increased chloride concentration on

nitrifying activated sludge cultures. J Environ Eng-ASCE 130(2):116–

125.

Chvapil M. 1973. New aspects in biological role of zinc—Stabilizer of

macromolecules and biological-membranes. Life Sci 13(8):1041–1049.

Cleveland WS, Devlin SJ. 1988. Locally weighted regression—An approach

to regression-analysis by local fitting. J Am Stat Assoc 83(403):596–

610.

Davis AP, Shokouhian M, Ni SB. 2001. Loading estimates of lead, copper,

cadmium, and zinc in urban runoff from specific sources. Chemo-

sphere 44(5):997–1009.

Ely RL, Hyman MR, Arp DJ, Guenther RB, Williamson KJ. 1995.

A cometabolic kinetics model incorporating enzyme inhibition, inac-

tivation, and recovery. 2. Trichloroethylene degradation experiments.

Biotechnol Bioeng 46(3):232–245.

EPA US. 1993. Process design manual: Nitrogen control. Washington, DC:

U.S. Environmental Protection Agency. Report nr 625/R-93/010.

EPA US. 2005. Toxicological review of zinc and compounds.

Washington, DC: U.S. Environmental Protection Agency. Report nr

7440-66-6.

Gish W, States DJ. 1993. Identification of protein coding regions by

database similarity search. Nat Genet 3(3):266–272.

Gvakharia BO, Permina EA, Gelfand MS, Bottomley PJ, SayavedraSoto LA,

Arp DJ. 2007. Global transcriptional response of Nitrosomonas euro-

paea to chloroform and chloromethane. Appl Environ Microbiol

73(10):3440–3445.

Hageman RH, Hucklesby DP. 1971. Nitrate reductase from higher plants.

Methods Enzymol 23:491–503.

Hayward RE, DeRisi JL, Alfadhli S, Kaslow DC, Brown PO, Rathod PK.

2000. Shotgun DNA microarrays and stage-specific gene expression in

Plasmodium falciparum malaria. Mol Microbiol 35(1):6–14.

Hildemann LM, Markowski GR, Cass GR. 1991. Chemical composition of

emissions from urban sources of fine organic aerosol. Environ Sci

Technol 25(4):744–759.

Hu ZQ, Chandran K, Grasso D, Smets BF. 2003. Impact of metal sorption

and internalization on nitrification inhibition. Environ Sci Technol

37(4):728–734.

Hwang L, Hocking-Murray D, Bahrami AK, Andersson M, Rine J, Sil A.

2003. Identifying phase-specific genes in the fungal pathogen Histo-

plasma capsulatum using a genomic shotgun microarray. Mol Biol Cell

14(6):2314–2326.

Hyman MR, Arp DJ. 1995. Effects of ammonia on the de novo synthesis of

polypeptides in cells of Nitrosomonas europaea denied ammonia as an

energy source. J Bacteriol 177(17):4974–4979.

Hyman MR, Sansome-Smith AW, Shears JH, Wood PM. 1985. A kinetic

study of benzene oxidation to phenol by whole cells of Nitrosomonas

europaea and evidence for the further oxidation of phenol to hydro-

quinone. Arch Microbiol 143:302–306.

Jonsson K, Aspichueta E, de la Sota A, Jansen JL. 2001. Evaluation of

nitrification-inhibition measurements. Water Sci Technol 43(1):201–

208.

Juliastuti SR, Baeyens J, Creemers C, Bixio D, Lodewyckx E. 2003. The

inhibitory effects of heavy metals and organic compounds on the net

maximum specific growth rate of the autotrophic biomass in activated

sludge. J Hazard Mater 100(1–3):271–283.

Juretschko S, Timmermann G, Schmid M, Schleifer KH, Pommerening-

Roser A, Koops HP, Wagner M. 1998. Combined molecular and

conventional analyses of nitrifying bacterium diversity in activated

sludge: Nitrosococcus mobilis and Nitrospira-like bacteria as dominant

populations. Appl Environ Microbiol 64(8):3042–3051.

Kachur AV, Koch CJ, Biaglow JE. 1998. Mechanism of copper-catalyzed

oxidation of glutathione. Free Radic Res 28(3):259–269.

Kelly JJ, Siripong S, McCormack J, Janus LR, Urakawa H, El Fantroussi S,

Noble PA, Sappelsa L, Rittmann BE, Stahl DA. 2005. DNA microarray

detection of nitrifying bacterial 16S rRNA in wastewater treatment

plant samples. Water Res 39(14):3229–3238.

Kowalchuk GA, Stephen JR. 2001. Ammonia-oxidizing bacteria: A model

for molecular microbial ecology. Annu Rev Microbiol 55:485–529.

Lee YW,Ong SK, Sato C. 1997. Effects of heavymetals on nitrifying bacteria.

Water Sci Technol 36(12):69–74.

Legret M, Pagotto C. 1999. Evaluation of pollutant loadings in the runoff

waters from a major rural highway. Sci Total Environ 235(1–3):143–

150.

Limpiyakorn T, Kurisu F, Yagi O. 2006. Development and application of

real-time PCR for quantification of specific ammonia-oxidizing bac-

teria in activated sludge of sewage treatment systems. Appl Microbiol

Biotechnol 72(5):1004–1013.

Loveless JE, Painter HA. 1968. Influence of metal ion concentrations and

pH value on growth of a Nitrosomonas strain isolates from activated

sludge. J Gen Microbiol 52:1–14.

Meiklejohn J. 1954. Some aspects of the physiology of the nitrifying bacteria.

London: Cambridge University Press. p 67–83.

Mota C, HeadMA, Ridenoure JA, Cheng JJ, de los Reyes FL. 2005. Effects of

aeration cycles on nitrifying bacterial populations and nitrogen

removal in intermittently aerated reactors. Appl Environ Microbiol

71(12):8565–8572.

Nies DH. 1992. Resistance to cadmium, cobalt, zinc, and nickel in microbes.

Plasmid 27(1):17–28.

Nies DH. 1999. Microbial heavy-metal resistance. Appl Microbiol Biotech-

nol 51(6):730–750.

Panswad T, Polprucksa P. 1998. Specific oxygen uptake, nitrification and

denitrification rates of a zinc-added anoxic/oxic activated sludge

process. Water Sci Technol 38(1):133–139.

Peirson SN, Butler JN, Foster RG. 2003. Experimental validation of novel

and conventional approaches to quantitative real-time PCR data

analysis. Nucleic Acids Res 31(14):e73.

Purkhold U, Pommerening-Roser A, Juretschko S, Schmid MC, Koops HP,

Wagner M. 2000. Phylogeny of all recognized species of ammonia

oxidizers based on comparative 16S rRNA and amoA sequence analysis:

Implications for molecular diversity surveys. Appl Environ Microbiol

66(12):5368–5382.

Rowan AK, Moser G, Gray N, Snape JR, Fearnside D, Curtis TP, Barer MR,

Head IM. 2003. A comparitive study of ammonia-oxidizing bacteria in

R

lab-scale industrial wastewater treatment reactors. Water Sci Technol

48(3):17–24.

Rule KL, Comber SDW, Ross D, Thornton A, Makropoulos CK, Rautiu R.

2006. Diffuse sources of heavy metals entering an urban wastewater

catchment. Chemosphere 63(1):64–72.

Saal LH, Troein C, Vallon-Christersson J, Gruvberger S, Borg A, Peterson C.

2002. BioArray software environment: A platform for comprehensive

management and analysis of microarray data. Genome Biology 3(8):

software 0003.1–0003.6.

Sato C, Leung SW, Schnoor JL. 1988. Toxic response of Nitrosomonas

europaea to copper in inorganic medium and wastewater. Water Res

22(9):1117–1127.

SayavedraSoto LA, Hommes NG, Russell SA, Arp DJ. 1996. Induction of

ammonia monooxygenase and hydroxylamine oxidoreductase mRNAs

by ammonium in Nitrosomonas europaea. Mol Microbiol 20(3):541–

548.

Skinner FA,Walker N. 1961. Growth ofNitrosomonas europaea in batch and

continuous culture. Arch Mikrobiol 38(4):339–349.

Tomlinson TG, Boon AG, Trotman CNA. 1966. Inhibition of nitrification

in activated sludge process of sewage disposal. J Appl Bacteriol

29(2):266–291.

Wagner M, Noguera DR, Juretschko S, Rath G, Koops HP, Schleifer KH.

1998. Combining fluorescent in situ hybridization (FISH) with cultiva-

tion and mathematical modeling to study population structure and

function of ammonia-oxidizing bacteria in activated sludge. Water Sci

Technol 37(4–5):441–449.

Watson SW, Valois FW, Waterbury JB. 1981. The family nitrobacteraceae.

In: Starr MP, Stolp H, Truper HG, Balows A, Schlegel HG, editors. The

prokaryotes: A handbook on habitats, isolation, and identification of

bacteria. New York: Springer-Verlag. p 1005–1022.

Wei XM, Sayavedra-Soto LA, Arp DJ. 2004. The transcription of the cbb

operon in Nitrosomonas europaea. Microbiol Sgm 150:1869–1879.

Wei XM, Yan TF, Hommes NG, Liu XD, Wu LY, McAlvin C, Klotz MG,

Sayavedra-Soto LA, Zhou JZ, Arp DJ. 2006. Transcript profiles of

Nitrosomonas europaea during growth and upon deprivation of ammo-

nia and carbonate. FEMS Microbiol Lett 257(1):76–83.

Yang YH, Dudoit S, Luu P, Lin DM, Peng V, Ngai J, Speed TP. 2002.

Normalization for cDNAmicroarray data: A robust composite method

addressing single and multiple slide systematic variation. Nucleic Acids

Res 30(4):e15.

Zaigler A, Schuster SC, Soppa J. 2003. Construction and usage of a one-fold

coverage shotgun DNA microarray to characterize the metabolism of

the archaeon Haloferax volcanii. Mol Microbiol 48(4):1089–1105.

adniecki and Ely: Zinc Chloride Inhibition of Nitrosococcus mobilis 1095

Biotechnology and Bioengineering