Bioresource Technology 102 (2011) 4196–4203

Contents lists available at ScienceDirect

Bioresource Technology

journal homepage: www.elsevier .com/locate /bior tech

Acclimation and activity of ammonia-oxidizing bacteria with respect to variationsin zinc concentration, temperature, and microbial population

Seungyong Lee, Kyungjin Cho, Juntaek Lim, Woong Kim, Seokhwan Hwang ⇑School of Environmental Science and Engineering, Pohang University of Science and Technology, San 31, Hyoja-dong, Nam-gu, Pohang, Kyungbuk 790-784, South Korea

a r t i c l e i n f o a b s t r a c t

Article history:Received 5 October 2010Received in revised form 6 December 2010Accepted 7 December 2010Available online 15 December 2010

Keywords:Ammonia-oxidizing bacteriaZincTemperatureResponse surface analysisAcclimation

0960-8524/$ - see front matter Crown Copyright � 2doi:10.1016/j.biortech.2010.12.035

⇑ Corresponding author. Tel.: +82 54 279 2282; faxE-mail address: shwang@postech.ac.kr (S. Hwang)

Activity of ammonia-oxidizing bacteria (AOB) to simultaneous variation in Zn2+ concentration(0.01–3.5 mg/L), temperature (23–33 �C), and AOB concentration (3–30 � 106 gene copies/mL) in a steelindustry wastewater treatment plant was evaluated. Two equations were developed to describe the lagperiod (i.e., AOB acclimation) and ammonia oxidation rate (i.e., growth of the AOB) depending on the vari-ables. AOB concentration and temperature both had significant effects on lag period and the ammoniaoxidation rate. Zn2+ concentration only had a significant effect on ammonia oxidation rate at 5% a-level.There was a significant interaction between AOB concentration and temperature for both lag period andammonia oxidation rate. The effects of the variables were not significant when AOB concentration washigher than 2.0 � 107 copies/mL. There was no visible shift or changes in AOB communities based onDGGE analysis with amoA gene primers.

Crown Copyright � 2010 Published by Elsevier Ltd. All rights reserved.

1. Introduction

Conventional integrated steel manufacturing primarily consistsof three unit processes: reduction of iron ores to steel in a blastfurnace, casting, and rolling. A large volume of cooling water isrequired in the processes and is inevitably contaminated withvarious materials (Kim et al., 2008). Major pollutants in the waste-water generated from steel manufacturing include suspended sol-ids, metallic compounds, oil and grease. The waste streams alsoinclude gasification products such as ammonia. The wastewatercreates a serious disposal problem for the steel processing industrydue to its large discharge amount as well as its heterogeneouscharacteristics (Lay-Son and Drakides, 2008; WSA, 2009).

A biological nitrogen removal (BNR) system employingautotrophic nitrification followed by heterotrophic denitrificationis currently an essential step in the overall management ofwastewater generated in steel manufacturing industries due tothe tightened effluent discharge limits in many nations (Kellyet al., 2004; Wyffels et al., 2004). Regardless of the type of BNRprocess, the overall treatment efficiency is primarily dependenton two different kinds of microbial activities in the nitrificationphase: oxidation of ammonia to nitrite by ammonia-oxidizingbacteria (AOB); and the subsequent oxidation of nitrite to nitrateby nitrite-oxidizing bacteria (NOB).

The first step of nitrification (i.e., ammonia oxidation) is gener-ally regarded as a rate-limiting step due to the low growth rate of

010 Published by Elsevier Ltd. All r

: +82 54 279 8299..

AOB and their high susceptibility to changes in growth conditions.Autotrophic nitrifying bacteria (nitrifiers) cannot acclimate asreadily to the inhibitory event as can heterotrophic denitrifyingbacteria (denitrifiers). This difference in adaptive capacity mightrender the overall nitrification process sensitive to changes in envi-ronmental conditions (Limpiyakorn et al., 2005).

In most BNR processes, a long sludge retention time (SRT),usually at least 10 days (Campos et al., 2007; Hao et al., 2009), isrequired due to the low growth rate of AOB. Consequently, theresidual ammonia concentration in the system is likely to be verylow. When an external supply of the limiting substrate (i.e., ammo-nia) for growth and/or energy is not available or is extremely low,microorganisms turn to internal substrate sources. For example, ifcertain enzyme proteins used in metabolizing the original exoge-nous organic substrates are no longer needed, then they may be-come substrates themselves and are primarily used to maintaincells in a viable state. When ammonia concentration is low duringthe ammonia oxidation process, AOB in such a system can be con-sidered as being in a maintenance energy dominating stage (Priceand Sowers, 2004; Vogeler et al., 2008). AOB under this conditionare likely to be stagnant in growth and are also highly susceptibleto variations in environmental factors such as temperature, pH,and presence of heavy metals (Limpiyakorn et al., 2005).

Zinc (chiefly Zn2+) is a heavy metal frequently found in waste-water generated from steel manufacturing using oxygen furnacesand has non-biodegradable properties. Depending on the concen-tration of Zn2+ and the length of exposure with AOB in wastewater,zinc causes an acute or chronic effect by binding to the active siteof ammonia monooxygenase (AMO). This displaces the functional

ights reserved.

Table 1Experimental design and observed lag periods and ammonia oxidation rates.

Trial Independent variables Dependent variables

Zn2+

concentration(mg/L)

AOBconcentration(106 copies/mL)

Temperature(�C)

Lagperiod(h)

AORb

ðNHþ4 -Nmg=L=hÞ

1 0.01 3 23 72 0.72 3.5 3 23 141 0.63 0.01 30 23 1.2 4.74 3.5 30 23 0.5 3.55 0.01 3 33 26 0.4

S. Lee et al. / Bioresource Technology 102 (2011) 4196–4203 4197

copper ligand and thus results in inhibition of AOB growth (Huet al., 2004; Radniecki et al., 2009). For example, Nitrosomonaseuropaea exposed to 3.4 lM ZnCl2 showed a 61% decrease in therate of nitrite production and a 69% decrease in the ammoniadependent specific oxygen uptake rate (AMO-SOUR) following30 min of exposure (Park and Ely, 2008). Nitrosococcus mobilisexposed to 10 lM ZnCl2 showed an immediate decrease in activityand did not exhibit noticeable recovery, resulting in a 100%decrease in nitrite production rates and an 86% decrease inAMO-SOUR (Radniecki and Ely, 2008). Thus, inhibition of growthof AOB by various heavy metals has been of great interest toresearchers (Hu et al., 2002; You et al., 2009).

It must be noted, however, that the degree of microbial toxicityis likely to depend on the interaction between the inhibitor(s) andother environmental factors, which might be stimulatory formicrobial growth and thus possibly complement the inhibitoryeffect. It has been reported, for example, that the mixed liquorvolatile suspended solids (MLVSS) level in a bioreactor treatingindustrial wastewater is an important factor in the resistance tozinc and copper toxicity (Sirianuntapiboon and Hongsrisuwan,2007). The addition of more sludge to methanogenic pure culturescan also significantly increase the tolerance of the microorganismsto zinc, copper, and nickel (Jarrell et al., 1987). It could be sug-gested, therefore, that the inhibitory effects of heavy metals couldbe alleviated by increasing the biomass concentration of the sys-tem. In this manner, the ratio of the inhibitor to the quantity ofAOB can be lowered. A lower ratio would improve the toleranceof AOB to inhibitory compounds in the treatment system.

Temperature is also one of the most influential factors on thegrowth of nitrifying bacteria. Nitrifiers are temperature sensitiveand, consequently, so is the rate of nitrification. The rate of nitrifi-cation falls sharply at lesser than 10 �C but increases almost di-rectly proportionally with increasing temperature over the rangeof 10–30 �C. This activity curve fits with evidence showing theoptimal temperature for nitrification in the BNR process is around30 �C in the absence of inhibitor(s) (Gerardi, 2002). Therefore, itwould be expected that the rate of nitrification would acceleratewith increasing temperature and, conversely, that the rate wouldsignificantly decline with decreasing temperature at a given Zn2+

concentration.The main objective of this research was to develop predictive

equations for a mechanistic understanding of the AOB activities(i.e., acclimation and growth) in steel industry wastewater treat-ment plants by incorporating the simultaneous effects of Zn2+,temperature, and AOB concentration into a response surface anal-ysis (RSA). Such models could be used to formulate more reliabletechnological guidelines and to serve as the basis for technicaldecisions in wastewater treatment plant operation.

Molecular assays targeting 16S rRNA and amoA genes of AOBwere used for the microbial quantification and community struc-ture analysis, respectively. The results from this work could beused as general guidelines on the capacity of heterogeneous AOBpopulations to accommodate expected magnitudes of disruptionand recovery times, and, where possible, controllable factors thatmay affect the response.

6 3.5 3 33 22.4 0.27 0.01 30 33 0.4 9.48 3.5 30 33 0.9 7.09a 1.755 16.5 28 0.8 ± 0.01 3.4 ± 0.0110 0.01 16.5 28 1.8 4.911 3.5 16.5 28 0.2 2.112 1.755 3 28 35.6 0.213 1.755 30 28 0.7 7.314 1.755 16.5 23 1.4 2.215 1.755 16.5 33 2.7 6.4

a Center point. The experiment was repeated 3 times and the response repre-sented average values.

b AOR, ammonia oxidation rate.

2. Methods

2.1. Inoculum system and seed preparation

A continuous stirred tank reactor (CSTR) with 7 L workingvolume was set up to produce a consistent seed to inoculate thesubsequent batch cultures for RSA. The reason for using theinoculum CSTR was to minimize any confounding effect associatedwith the use of inconsistent inoculum. Ammonia-containing

wastewater (i.e., 300 mg NHþ4 -N=L) obtained from a local steel-pro-cessing industry was used as a substrate for the inoculum CSTR,which was operated at 10 days hydraulic retention time (HRT)and 28 �C. pH and dissolved oxygen (DO) concentration was con-trolled to be above 7.5 and 2.0 mg DO/L, respectively. The steadystate effluent containing no detectable NHþ4 -N concentration (datanot shown) was collected for centrifugation at 14,000g for 5 minwhen needed. The supernatant was discarded and the pellet waswashed with 20 mM NaHCO3 to remove residual ammonia. Afteranother centrifugation and washing step, the concentrated seedwas prepared and aerated until it was used as the inoculum forthe subsequent batch trials. Steady state was assumed when theconcentration of the NHþ4 -N;NO�2 -N;andNO�3 -N did not vary bymore than 10%.

2.2. Experimental design and batch RSA tests

The RSA experiment was designed with a central compositecube design (Lee et al., 2008; Myers and Montagomery, 1995)and batch ammonia oxidation reactions were conducted at eachcondition (Table 1). The ranges of each independent variable weredetermined based on the results from monthly sampling of thefull-scale treatment plant for 1 year before the RSA experiment.

Each trial was performed with a modified ISO 9509 method (Doet al., 2008). Concentrated (tenfold) synthetic wastewater was ini-tially prepared with 2.65 g/L (NH4)2SO4 for a final concentration of56 mg/L NHþ4 -N. The concentrated solution also contained 16.8 g/LNaHCO3 as buffering agent, which could maintain the pH at 7.5during ammonia oxidation. A 100�-concentrated ZnCl2 stock solu-tion, the concentrated seed, and the synthetic wastewater wereadded to 250 mL flasks, with distilled deionized water (DDW) toa total working volume of 125 mL. The experimental conditionswere shown in Table 1. The flasks were kept in a shaking incubator,stirred at 130 rpm, at the desired temperature of 23, 28, or 33 �C.The cultures were periodically sampled until the ammonium ionwas completely depleted.

2.3. Real-time quantitative PCR

The AOB concentration in the inoculum effluent was measuredusing LightCycler 480 Real-time PCR System (Roche Diagnostics,

4198 S. Lee et al. / Bioresource Technology 102 (2011) 4196–4203

Germany) with the group-specific primer-probe sets for AOBdeveloped previously (Lim et al., 2008) and adjusted to give adesired initial concentration in each trial. The group-specificprobe-primer sets of AOB have been designed to specifically detectsubgroups of AOB for cluster of N. europaea, N. nitrosa, N. cryotoler-ans, and Nitrosospira multiformis). Plasmid, constructed using typestrains belonging to the clusters and normalized to be 106 copiesof corresponding 16S rRNA genes/lL, were used as target AOB tem-plates to construct standard curves (Do et al., 2008).

All qPCR runs were performed in duplicate, and each reactionmixture was prepared using the LightStart DNA TaqMan Masterkit (Roche) in a total volume of 20 lL:5.0 lL PCR-grade water,1.0 lL of each primer (final concentration 0.5 lM) and probe (finalconcentration 0.5 lM), 2 �Master mix 10 lL, and 2.0 lL templateDNA. The thermal cycling protocol was as follows: initial denatur-ation for 10 min at 94 �C followed by 40 cycles of 10 s at 94 �C,simultaneous annealing and extension 30 s at 60 �C. GenomicDNA was extracted from the samples using an automated nucleicacid extractor (Magtration System 6GC, PSS Co., Japan).

2.4. Denaturing gradient gel electrophoresis and microbialidentification

At the end of the batch experiments, denaturing gradient gelelectrophoresis (DGGE) was performed using the extracted DNAto investigate any changes in AOB community structure. PCRamplification of amoA gene fragment was performed with theprimers amoA-1F-GC and amoA-2R (Rotthauwe et al., 1997). ThePCR was conducted in a Veriti 96-well Thermal Cycler (AppliedBiosystem, Forest City, CA, USA) using following protocol: initialdenaturation at 94 �C for 2 min, 35 cycles of denaturation at94 �C for 30 s, annealing at 58 �C for 45 s, and extension at 72 �Cfor 105 s; and final extension at 72 �C for 7 min (Qin et al., 2008).

Amplified amoA gene fragments were loaded on 8% (w/v)acrylamide–bisacrylamide gels with a denaturant gradient30–60% made with a gradient maker. The gel was run at 60 �Cand 85 V for 16 h in 1� TAE buffer using the D-code system(BioRad, Hercules, CA, USA). Following electrophoresis, the gelwas stained with ethidium bromide solution for 20 min, rinsedfor 20 min in deionized water (DIW), and then photographedunder UV trans-illumination.

Cloning and sequencing were performed to identify the DGGEbands. Each band was excised directly from the DGGE gel, sus-pended in 40 lL of DIW, and incubated overnight at 4 �C. ElutedDNA solution of 2 lL was further amplified by using amoA-1Fand amoA-2R primers.

The PCR products were purified from a 1% agarose gel using a kit(General biosystem, Seoul, Korea) and cloned into the pGEM-T Easyvector (Promega, Madison, WI, USA). The clones containing inser-tions of the partial amoA genes were sequenced with a 3730XLDNA analyzer (Applied Biosystem, Foster City, CA, USA) by acommercial service (Solgent, Korea). The sequencing results werecompared with the reference sequences in the GenBank database.Sequence alignment and phylogenetic tree were performed usingMEGA ver. 4.0 (Tamura et al., 2007).

2.5. Analytical methods

Zn2+ content in the samples was measured using inductivelycoupled plasma spectrometer (IRIS/AP, Thermo Jarrell Ash,Franklin, MA, USA). The concentrations of ammonium ions wereanalyzed with an ion chromatography (Metrohm 790 PersonalIC), and were plotted to calculate the lag period and ammoniaoxidation rate (AOR) using the modified Gompertz model(Eq. (1)) (Do et al., 2008; Giotta et al., 2006).

y ¼ A exp � explmax � expð1Þ

A� ðk� tÞ þ 1

� �� �ð1Þ

y, the logarithm of the relative population size against time; A,asymptote, the maximum value of growth reached; lmax, maximumspecific growth rate, tangent to the growth curve at the inflectionpoint; k, lag time, t-axis intercept of the tangent at the inflectionpoint on the growth curve.

3. Results and discussion

3.1. Range of variables and RSA

The lag period is the amount of time a population of organismsrequires to adjust to new environmental conditions such as thepresence of inhibitors. Cells placed into a different environmentcan be expected to undergo a period of adjustment prior to theirexponential growth in the new growth conditions. Active growthof AOB in the exponential growth stages mostly coincides withan active ammonia oxidation rate. Therefore, the lag period andammonia oxidation rate were used as dependent variables to rep-resent acclimation and growth of the AOB despite the inhibitoryconditions.

Zinc concentration, AOB concentration, and temperature in thefull-scale BNR facility of a local steel-producing industry weremonitored monthly for a 1 year period and used as the experimen-tal ranges of the variables. The Zn2+ concentration varied from 0.04to 3.28 mg/L and the AOB concentration ranged from 5.8 � 106 to2.3 � 107 copies/mL. The lowest and highest temperatures of thewastewater in the treatment plant were 23.8 and 33.4 �C for theperiod.

The culture conditions at the center point were 1.76 mg Zn2+/L,1.65 � 106 copies of AOB/mL, and 28 �C. Repeated observations atthe center were used to estimate the experimental error in theRSA. A total of 17 trials, including the center points were run toapproximate the responses of the lag period and the AOR (Table1). Increasingly complex equations from linear to partial cubicwere tried sequentially to model the data in Table 1.

The residual NHþ4 -N concentrations varied at each trial. The cul-ture conditions at trial 2, for example, were the most unfavorablefor the growth of AOB in this research (i.e., highest Zn2+ concentra-tion, lowest AOB concentration, and temperature). Those of trial 7were the most favorable for AOB growth among all trials (i.e.,lowest Zn2+ concentration, and highest AOB concentration andtemperature). The lag period for trial 2 was 353 times longer andthe AOR 93% lower compared to trial 7. This would indicate acombined effect of the environmental conditions on the responseof AOB growth. Therefore, predictive models of AOB activity mustaccount for simultaneous changes in all of the environmentalvariables.

3.2. Modeling and analyzing the response of lag period

Out of the various possible regression equations from linear topartial cubic, lag period could be best estimated with a quadraticmodel (Eq. (2)). The P-value of the regression was significant at1% a-level and lack of fit (LOF) was not significant at 0.1% a-level.The regression coefficient and residual standard deviation of thequadratic model were 0.92 and 15.8, respectively. The residualplots for the model and data set showed no patterns or trends(Fig. 1). A test for the constant variance assumption was alsopassed because a random plot of residuals showed homogeneouserror variances across the observed values. Therefore, it was con-cluded that the model was able to accurately predict lag periodand was

Fig. 1. Residual plots of the quadratic model for the lag period (closed circle), the interaction model for the ammonia oxidation rate (opened circle).

S. Lee et al. / Bioresource Technology 102 (2011) 4196–4203 4199

g1 ¼ 446:5þ 27x1 � 1:5� 10�5x2 � 20:4x3 � 4:8� 10�7x1x2

� 0:7x1x3 þ 3:5� 10�7x2x3 þ 1:4x21 þ 1:2� 10�13x2

2 þ 0:2x23

ð2Þwhere g1 = lag period (h) and xi = independent variable i (i = Zn2+

concentration, AOB concentration, and temperature in order).

Fig. 2. Two- and three-dimensional contour plots of the models for the lag periods (h) wand temperature (b) within the design boundaries.

With this model equation, the change in the length of lag periodcan be calculated when the independent variables change simulta-neously. If Zn2+ concentration, AOB concentration, and tempera-ture were 3.5 mg/L, 3.0 � 106 copies/mL, and 23 �C, respectively,the length of the lag phase would be estimated as 124.8 h. Usingthe model, the lag period was expected to decrease to 34.9 h if

ith respect to AOB concentration and Zn2+ concentration (a) and AOB concentration

4200 S. Lee et al. / Bioresource Technology 102 (2011) 4196–4203

AOB concentration and temperature increased to 1.0 � 107 copies/mL and 28 �C, respectively. It could also be predicted that the lagperiod would not increase even if Zn2+ concentration increasedfrom 0.01 to 3.5 mg/L providing that the AOB concentration couldbe maintained above 2.0 � 106 copies/mL at 28 �C.

Fig. 2 shows two- and three-dimensional contour plots of themodel for the lag period with respect to the concentrations ofZn2+ and AOB, and temperature. The length of lag period was sig-nificantly affected by the AOB concentration and temperature,whereas Zn2+ had no significant effect on the lag period (Table 2)at 5% a-level. Sudden increases in the lag period could be seen in

Table 2ANOVA for independent variables.

Term Lag period AORa

Meansquare

DFb Pvalue

Meansquare

DF Pvalue

Zn2+ 410 2 0.260 4.5 1 0.022AOB 4249 2 0.002 88.8 1 0.000Temperature 1799 2 0.020 13.7 1 0.001Zn2+ � AOB 1002 1 0.085 1.4 1 0.168Zn2+ � Temperature 283 1 0.322 0.2 1 0.571AOB � Temperature 4434 1 0.004 9.9 1 0.003

a AOR, ammonia oxidation rate.b DF, degree of freedom.

Fig. 3. Two- and three-dimensional contour plots of the models for the ammonia oxidatiAOB concentration and temperature (b) within the design boundaries.

the response as AOB concentration and temperature decrease.When the AOB concentration and temperature are lower than1.5 � 107 copies/mL and 28 �C, the lag period was predominantlyinfluenced by the linear combination of the initial AOB concentra-tion and temperature (Fig. 2b).

However, it should be noted that the effect of all other variableson the lag period was insignificant when AOB concentration washigher than 2.0 � 107 copies/mL. Culture conditions of very highZn2+ concentration and a low temperature could be regarded asan unfavorable environment for acclimation of AOB to Zn2+ inhibi-tion in a full-scale treatment process. But results in this studyshowed that a desired level of AOB activity could be maintaineddespite temperature shock and Zn2+ inhibition if the AOB popula-tion were kept above a certain level in the treatment system. Theseresults support previous findings that the toxicity of Zn2+ is muchhigher on pure cultures of AOB than on activated sludge, where thetoxic effect of Zn2+ reduces with increasing sludge concentration(Juliastuti et al., 2003; Radniecki and Ely, 2008).

A lag is not necessarily a dormant period. The cells may be tak-ing in significant amounts of substrate, synthesizing new enzymes,and undergoing enlargement prior to division, but, in any event,they have not yet begun the orderly and steady replication charac-teristic of the succeeding phase (Augustin et al., 2000; Giotta et al.,2006). Avoidance of the effects of a lag period is one of the reasonswhy acclimated seed is stipulated. The length of a lag periodcaused by inhibitory material may also be decreased by

on rates (mg/L/h) with respect to AOB concentration and Zn2+ concentration (a) and

S. Lee et al. / Bioresource Technology 102 (2011) 4196–4203 4201

maintaining a large number of cells, which may be possible in thefull-scale treatment process.

The equation developed in this research could also be used topredict the time required for an AOB population of a certain sizeto acclimate to a level of Zn2+ in the wastewater at a given temper-ature. Thus, it would provide a mechanistic understanding for pro-cess engineers in the treatment plant that would allow them toaccommodate AOB recovery times.

3.3. Effects on the ammonia oxidation rate

The interaction equation (Eq. (3)) was selected to predict the re-sponse of the ammonia oxidation rate to simultaneous changes inthe environmental variables. The P-value of regression of the mod-el was significant and LOF was not significant at 0.1% a-level. Theregression coefficient and residual standard deviation of the modelwere 0.96 and 0.8, respectively. The residual plots showed no pat-terns or trends, which indicated there was no weakness of the se-lected model (Fig. 1).

g2 ¼ 0:2þ 0:4x1 � 2:1� 10�7x2 � 5:3x3 � 1:8� 10�8x1x2

� 1:9� 10�2x1x3 þ 1:6� 10�8x2x3 ð3Þ

where g2 = ammonia oxidation rate (mg/L/h) and xi = independentvariable i (i = Zn2+ concentration, AOB concentration, and tempera-ture in order).

The change in the ammonia oxidation rate could also be pre-dicted using the model equation. When Zn2+ concentration, AOBconcentration, and temperature were 3.5 mg/L, 3.0 � 106 copies/mL, and 28 �C respectively, the estimated AOR for NHþ4 -N was0.28 mg/L/h. Even if the Zn2+ concentration was kept at 3.5 mg/L,the AOR would increase to 7.54 mg NHþ4 -N=L=h when the AOB con-centration was changed from 3.0 � 106 to 3.0 � 107 mg/L, and thetemperature from 28 to 33 �C, respectively. Effects of Zn2+ concen-tration, AOB concentration, and temperature were significant onthe ammonia oxidation rate at 5%, 0.1%, 0.1% a-level, respectively.The interaction term of AOB concentration and temperature wasalso significant at 1% a-level (Table 2).

Fig. 4. DGGE profile of the PCR products amplified with amoA gene primers. The temploxidation reaction.

Fig. 3a and b show the effects of the Zn2+ concentration and AOBconcentration on the ammonia oxidation rate, and the effects ofthe initial AOB concentration and temperature on the ammoniaoxidation rate, respectively. It was clearly shown that the responsesurface of the AOR was nearly linear to the variation in the inde-pendent variables for the entire region of investigation. Zn2+ con-centration did not have much effect on the AOR (Fig. 3a), but therate of increase was proportional to increasing AOB concentrationand temperature. The combined effect of AOB concentration andtemperature on the AOR can be seen as the curvature region inthe upper left corner (Fig. 3b), where the AOB concentration andtemperature were approximately 2.0 � 107 copies/mL and 30 �Cor higher, respectively.

Compared to the lag period response, the response of AORseemed to follow the general rules in biochemical reaction with re-spect to biomass concentration, temperature, and presence ofinhibitor. As the reaction rate is directly proportional to biomassconcentration, membrane bioreactor systems are often preferredover conventional clarifiers as a means to increase the concentra-tion of biomass (Klimiuk and Kulikowska, 2006; Xue et al., 2009).Biochemical reactions roughly double for each 10 �C increase intemperature. This positive effect of temperature is similar to whatoccurs for all reactions (Rittmann and McCarty, 2001). Nitrificationbiokinetics are sensitive to elevated temperatures, with the rates ofnitrification increasing to an apparent maximum at around 30 �C,beyond which the overall rate decreases (Jiao et al., 2010; Zhanget al., 2009). In the presence of an inhibitor in the treatment sys-tem, the positive effect on microbial growth from increasing tem-perature and biomass concentration would decrease based on thequantity and toxicity of the inhibitor. The model developed herecan predict the combined effect of Zn2+ concentration, AOB concen-tration, and temperature on ammonia oxidation rate in the waste-water treatment process.

3.4. Community structure of ammonia oxidizing bacteria

Genomic DNA was extracted after the cessation of activeammonia, and then PCR-DGGE was performed with amoA gene

ate DNA was extracted from each experimental trials at the end point of ammonia

Fig. 5. Neighbor-joining tree based on amoA gene sequences obtained from DGGE band.

4202 S. Lee et al. / Bioresource Technology 102 (2011) 4196–4203

primers to investigate the AOB community structure. A total of sixdifferent DGGE bands, designated as Band 1–6, were visually de-tected and excised from the gel for subsequent sequencing analysis(Fig. 4). The affiliations of the amoA gene sequences were deter-mined by comparison with the GenBank database.

From the DGGE profile, there was no visible shift of AOB com-munities. Three- (bands 1–3) and two- DGGE bands (bands 4–5)were closely related to N. europaea (94–95% similarity) and uncul-tured Nitrosomonas sp. (95% similarity), respectively. The sixthband was related to N. multiformis (93% similarity). The phyloge-netic affiliation can also be seen in Fig. 5. The partial sequence ofamoA genes of bands 1–5 is closely grouped with the amoAsequence of N. europaea while the sequence of band 6 forms agroup with the sequence of N. multiformis. Due to the low similar-ity (i.e., lower than 97%), it was concluded that further qPCRanalysis with proper primers and probe targeting the AOB wouldbe required to identify the AOB bands appeared in the batch tests.

4. Conclusion

Quadratic and interaction equations were respectively devel-oped to describe the changes in lag period and ammonia oxidationrate with respect to the simultaneous effects of Zn2+ concentration(0.1–3.5 mg/L), AOB concentration (3.0 � 106–3 � 107 copies/mL),and temperature (23–33 �C). The lag period was significantly af-fected by the AOB concentration and temperature, whereas Zn2+

concentration had no significant effect. Effects of all variables weresignificant on the ammonia oxidation rate. Based on DGGE analysiswith amoA gene, there was no visible shift in AOB communities.

Acknowledgements

This work was supported by the Korea government Ministry ofKnowledge Economy as the New & Renewable Energy of the KoreaInstitute of Energy Technology Evaluation and Planning (KETEP)grant (Grant No. 20101T100100366), by the Korea Ministry ofEnvironment (MOE) through Human Resource Development forEnergy from Waste & Recycling, and by the Korea Research Foun-dation Grant funded by Korea Government (MOEHRD).

References

Augustin, J.C., Brouillaud-Delattre, A., Rosso, L., Carlier, V., 2000. Significance ofinoculum size in the lag time of Listeria monocytogenes. Appl. Environ.Microbiol. 66, 1706–1710.

Campos, J.L., Garrido, J.M., Mosquera-Corral, A., Mendez, R., 2007. Stability of anitrifying activated sludge reactor. Biochem. Eng. J. 35, 87–92.

Do, H., Lim, J., Shin, S.G., Wu, Y.J., Ahn, J.H., Hwang, S., 2008. Simultaneous effect oftemperature, cyanide and ammonia-oxidizing bacteria concentrations onammonia oxidation. J. Ind. Microbiol. Biotechnol. 35, 1331–1338.

Gerardi, M.H., 2002. Nitrification and Denitrification in the Activated SludgeProcess. John Wiley and Sons, New York.

Giotta, L., Agostiano, A., Italiano, F., Milano, F., Trotta, M., 2006. Heavy metal ioninfluence on the photosynthetic growth of Rhodobacter sphaeroides.Chemosphere 62, 1490–1499.

Hao, X., Wang, Q., Zhang, X., Cao, Y., Mark Loosdrecht, C.M.v.,, 2009. Experimentalevaluation of decrease in bacterial activity due to cell death and activity decayin activated sludge. Water Res. 43, 3604–3612.

Hu, Z., Chandran, K., Grasso, D., Smets, B.F., 2002. Effect of nickel and cadmiumspeciation on nitrification inhibition. Environ. Sci. Technol. 36, 3074–3078.

Hu, Z., Chandran, K., Grasso, D., Smets, B.F., 2004. Comparison of nitrificationinhibition by metals in batch and continuous flow reactors. Water Res. 38,3949–3959.

Jarrell, K.F., Saulnier, M., Ley, A., 1987. Inhibition of methanogenesis in pure culturesby ammonia, fatty acids, and heavy metals, and protection against heavy metaltoxicity by sewage sludge. Can. J. Microbiol. 33, 551–554.

Jiao, Y., Zhao, Q., Jin, W., Hao, X., You, S., 2010. Bioaugmentation of a biologicalcontact oxidation ditch with indigenous nitrifying bacteria for in situremediation of nitrogen-rich stream water. Bioresour. Technol. in Press, doi:10.1016/j.biortech.2010.09.061.

Juliastuti, S.R., Baeyens, J., Creemers, C., Bixio, D., Lodewyckx, E., 2003. Theinhibitory effects of heavy metals and organic compounds on the netmaximum specific growth rate of the autotrophic biomass in activatedsludge. J. Hazard. Mater. 100, 271–283.

Kelly, R.T., Henriques, I.D.S., Love, N.G., 2004. Chemical inhibition of nitrification inactivated sludge. Biotechnol. Bioeng. 85, 683–694.

Kim, Y.M., Park, D., Jeon, C.O., Lee, D.S., Park, J.M., 2008. Effect of HRT on thebiological pre-denitrification process for the simultaneous removal of toxicpollutants from cokes wastewater. Bioresour. Technol. 99, 8824–8832.

Klimiuk, E., Kulikowska, D., 2006. The influence of hydraulic retention time andsludge age on the kinetics of nitrogen removal from leachate in SBR. Pol. J.Environ. Stud. 15, 283–289.

Lay-Son, M., Drakides, C., 2008. New approach to optimize operational conditionsfor the biological treatment of a high-strength thiocyanate and ammoniumwaste: pH as key factor. Water Res. 42, 774–780.

Lee, S., Bae, H., Song, M., Hwang, S., 2008. Bioconversion of starch processing wasteto Phellinus linteus mycelium in solid-state cultivation. J. Ind. Microbiol.Biotechnol. 35, 859–865.

Lim, J., Do, H., Shin, S.G., Hwang, S., 2008. Primer and probe sets for group-specificquantification of the genera Nitrosomonas and Nitrosospira using real-time PCR.Biotechnol. Bioeng. 99, 1374–1383.

Limpiyakorn, T., Shinohara, Y., Kurisu, F., Yagi, O., 2005. Communities of ammonia-oxidizing bacteria in activated sludge of various sewage treatment plants inTokyo. FEMS Microbiol. Ecol. 54, 205–217.

S. Lee et al. / Bioresource Technology 102 (2011) 4196–4203 4203

Myers, R.H., Montagomery, D.C., 1995. Response Surface Methodology. John Wileyand Sons, Inc., New York, NY.

Park, S., Ely, R.L., 2008. Genome-wide transcriptional responses of Nitrosomonaseuropaea to zinc. Arch. Microbiol. 189, 541–548.

Price, P.B., Sowers, T., 2004. Temperature dependence of metabolic rates formicrobial growth, maintenance, and survival. Proc. Natl. Acad. Sci. USA 101,4631–4636.

Qin, Y.Y., Zhang, X.W., Ren, H.Q., Li, D.T., Yang, H., tul=0?>Qin et al., 2008. Populationdynamics of ammonia-oxidizing bacteria in an aerated submerged biofilm reactor formicropolluted raw water pretreatment. Appl. Microbiol. Biotechnol. 79, 135–145.

Radniecki, T.S., Ely, R.L., 2008. Znic chloride inhibition of Nitrosococcus mobilis.Biotechnol. Bioeng. 99, 1085–1095.

Radniecki, T.S., Semprini, L., Dolan, M.E., 2009. Expression of merA, amoA and hao incontinuously cultured Nitrosomonas europaea cells exposed to zinc chlorideadditions. Biotechnol. Bioeng. 102, 546–553.

Rittmann, B.E., McCarty, P.L., 2001. Environmental Biotechnology: Principles andApplications. McGraw-Hill, New York, NY.

Rotthauwe, J.-H., Witzel, K.-P., Liesack, W., 1997. The ammonia monooxygenasestructural gene amoA as a functional marker: molecular fine-scale analysis ofnatural ammonia-oxidizing population. Appl. Environ. Microbiol. 63, 4704–4712.

Sirianuntapiboon, S., Hongsrisuwan, T., 2007. Removal of Zn2+ and Cu2+ by asequencing batch reactor (SBR) system. Bioresour. Technol. 98, 808–818.

Tamura, K., Dudley, J., Nei, M., Kumar, S., 2007. MEGA4: molecular evolutionarygenetics analysis (MEGA) software version 4.0. Mol. Biol. Evol. 24, 1596–1599.

Vogeler, I., Vachey, A., Deurer, M., Bolan, N., 2008. Impact of plants on the microbialactivity in soils with high and low levels of copper. Eur. J. Soil Biol. 44, 92–100.

WSA, 2009. World Steel in Figures 2009. World Steel Association, Brussels, Belgium.Wyffels, S., Van Hulle, S.W.H., Boeckx, P., Volcke, E.I.P., Van Cleemput, O.,

Vanrolleghem, P.A., Verstraete, W., 2004. Modeling and simulation of oxygen-limited partial nitritation in a membrane-assisted bioreactor (MBR). Biotechnol.Bioeng. 86, 531–542.

Xue, Y., Yang, F., Liu, S., Fu, Z., 2009. The influence of controlling factors on the start-up and operation for partial nitrification in membrane bioreactor. Bioresour.Technol. 100, 1055–1060.

You, S.J., Tsai, Y.P., Huang, R.Y., 2009. Effect of heavy metals on nitrificationperformance in different activated sludge processes. J. Hazard. Mater. 165, 987–994.

Zhang, L.Q., Wei, C.H., Zhang, K.F., Zhang, C.S., Fang, Q., Li, S.G., 2009. Effects oftemperature on simultaneous nitrification and denitrification via nitrite in asequencing batch biofilm reactor. Bioprocess. Biosyst. Eng. 32, 175–182.