Quantitative response relationships between nitrogen transformation rates and nitrogen functional genes in a tidal ?ow constructed wetland under C/N ratio constraints
Wei Zhi a ,b ,Guodong Ji a ,*
a
Key Laboratory of Water and Sediment Sciences,Ministry of Education,Department of Environmental Engineering,Peking University,Beijing 100871,China b
Department of Civil and Environmental Engineering,Virginia Tech,Blacksburg,VA 24061,United States
a r t i c l e i n f o
Article history:
Received 17February 2014Received in revised form 17June 2014
Accepted 25June 2014Available online 3July 2014Keywords:
Constructed wetland (CW)Tidal ?ow C/N ratio Functional gene
Nitrogen transformation
Quantitative response relationship
a b s t r a c t
The present study explored treatment performance and nitrogen removal mechanisms of a novel tidal ?ow constructed wetland (TF CW)under C/N ratios ranging from two to 12.High and stable COD (83e 95%),NH 4tàN (63e 80%),and TN (50e 82%)removal ef?ciency were simultaneously achieved in our single-stage TF CW without costly aeration.Results showed that a C/N ratio exceeding six was required to achieve complete denitri?cation without NO 2ààN and NO 3ààN accumulation in the system.Molecular biological analyses revealed aerobic ammonia oxidation was the dominant NH 4tàN removal pathway when the C/N ratio was less than or equal to six.However,when the C/N ratio was greater than six,anammox was notably enhanced,resulting in another primary NH 4tàN removal pathway,in addition to the aerobic ammonia oxidation.Quantitative response relation-ships between nitrogen transformation rates and nitrogen functional genes were estab-lished,and these relationships con?rmed that different nitrogen transformation processes were coupled at the molecular level (functional genes),and collaboratively contributed to nitrogen removal in the TF CW.Speci?cally,NH 4tàN transformation rates were collec-tively determined by amoA ,nxrA ,anammox,narG ,nirS ,nirK ,and nosZ ;and TN removal was in?uenced primarily by amoA and anammox.
?2014Elsevier Ltd.All rights reserved.
1.Introduction
Constructed wetlands (CWs)are engineered systems con-structed for wastewater treatment.The systems are designed to utilize and replicate many physical,chemical,and biolog-ical processes that occur in natural wetlands,but do so within a more controlled environment (Vymazal,2007).During the
last ?ve decades,CWs have evolved from empirical research into practical applications for treating various types of wastewater (Ji et al.,2007;Konnerup et al.,2009;Scholz and Hedmark,2010;Zhou et al.,1996).Given the economic and ecological bene?ts,CWs have become a widely applied tech-nique for economically underdeveloped areas and natural habitats,and a valuable complement to traditional sewage systems that dominate large cities (Zhi and Ji,2012).However,
*Corresponding author .Tel.:t86(010)62755914.
E-mail addresses:jiguodong@https://www.doczj.com/doc/ba14033690.html, ,luckysubmission@https://www.doczj.com/doc/ba14033690.html, (G.
Ji).Available online at https://www.doczj.com/doc/ba14033690.html,
ScienceDirect
journal homepage:
https://www.doczj.com/doc/ba14033690.html,/locate/watres
w a t e r r e s e a r c h 64(2014)32e 41
https://www.doczj.com/doc/ba14033690.html,/10.1016/j.watres.2014.06.035
0043-1354/?2014Elsevier Ltd.All rights reserved.
under varying organic loading rate,nitrogen removal in CWs exhibited substantial?uctuations and was often unsatisfac-tory.The North America database documented an average of 44%nitrogen removal ef?ciency in CWs(Kadlec et al.,2000).In Europe,the nitrogen removal ef?ciency of a typical CW was only35%,and could not exceed50%even with an optimized design(Verhoeven and Meuleman,1999).Therefore,how to improve nitrogen removal ef?ciency under varying organic loading rates have become an urgent issue and a research hotspot in the?eld of CWs.
Studies have demonstrated that nitrogen removal in CWs is often restricted by insuf?cient oxygen supply(EPA,1993;Ye et al.,2012).Speci?cally,microbial ammonia oxidation,which is the?rst and rate-limiting step for subsequent nitrogen transformation and removal,is impaired by limited oxygen supply to the CW substrate(Caffrey et al.,2007;Francis et al., 2005).Insuf?cient nitri?cation is the key factor for improving NH4tàN removal ef?ciency,and therefore an increasing effort has made to increase the oxygen content in CWs.Ver-tical?ow constructed wetlands(VF CWs),which increase oxygen transport capacity and therefore provide much better conditions for nitri?cation,have been extensively employed to enhance NH4tàN removal.However,total nitrogen(TN) removal is de?cient in VF CWs due to the limited denitri?ca-tion,agreeing with previous assessments that single-stage CWs cannot achieve high TN removal due to an inability to provide both aerobic and anaerobic conditions simulta-neously(Vymazal,2007).On the other hand,horizontal?ow constructed wetlands(HF CWs),which suffer from a lack of oxygen in beds,provide suitable denitri?cation conditions. Therefore,a growing interest in integrating speci?c advan-tages of different CW types to achieve better TN treatment performance has been observed.The most common con?gu-rations of hybrid systems are VF-HF and HF-VF CWs(Vymazal, 2011).Some other hybrid systems consisting of more than two different CW stages are also widely implemented(Brix et al., 2007;Vymazal and Kr€o pfelov a,2011;Ye and Li,2009).
Recently,tidal?ow constructed wetlands(TF CWs)have been proposed to enhance treatment performance(Leonard et al.,2003;Sun et al.,1999,2005).The“tidal?ow”refers to an operation strategy that repeatedly allows CWs to be?lled with wastewater,and then completely drained.During the ?lling phase,the CWs are gradually?ooded and air in bed is continuously squeezed out and consumed.In the draining phase,fresh air is drawn into CWs and the bed is replenished with oxygen.During the rhythmic cycle of“?ood/wet”and “drain/dry”phase,the wastewater acts as a passive pump to expel and draw air into the CWs,and consequently oxygen supply and consumption are substantially improved in the system.Hence,both nitri?cation and denitri?cation are intensi?ed in TF CWs,and high TN removal can be achieved. Cui et al.(2012)used a two-stage TF CW,and accomplished maximum NH4tàN and TN removal of72.7%and53.2%, respectively.In two separate studies of four-stage CWs with tidal?ow operation under different and varying chemical oxygen demand(COD)loading rates,Zhao et al.(2004,2011) reported NH4tàN(TN)removal ef?ciency ranged from49to 93%(11e78%)and30e91%.To date,only a few studies have been published on TF CWs.While these studies obtained some promising results in terms of improved NH4tàN and TN removal with multiple-stage TF CWs,no efforts have been made to achieve satisfactory NH4tàN and TN removal in one single-stage CW with tidal?ow operation.In addition,previ-ous studies have primarily focused on optimizing con?gura-tions and operational parameters to improve treatment performance under varying loading rates,and few attempts have been made to investigate nitrogen removal mechanisms at the molecular level in TF CWs.Therefore,research targeting nitrogen removal processes and pathways with functional genes under varying loading rates is much needed to optimize the design,operation,and application of TF CWs.Increased ef?cacy in treating wastewater will have notably positive environmental impacts,considering the ecological bene?ts these novel CWs can bring at such effective economic costs.
In one of the?rst attempts to achieve satisfactory nitro-gen removal in one single-stage CW with tidal?ow operation, the overall goal of this study was to understand nitrogen removal mechanisms at the molecular level under different C/N ratios,and the role of microbes in nitrogen trans-formation processes.The following four speci?c objectives were pursued:1)evaluate the treatment performance under different C/N ratios;2)quantify the absolute abundance of functional genes involved in nitrogen removal,and investi-gate ecological associations among these functional genes;3) determine quantitative response relationships between ni-trogen transformation processes and functional genes;and4) identify key functional genes that determine the treatment performance of nitrogen removal in TF CW.Establishment and analysis of the quantitative response relationships could also aid in our future efforts to:1)quantify the relative contribution of functional genes to nitrogen removal;2) regulate nitrogen removal processes at the molecular level with speci?c functional gene enrichment;and3)estimate the dynamics of nitrogen transformation rates using functional gene data.
2.Materials and methods
2.1.Tidal?ow constructed wetland
2.1.1.Experimental set up
One lab-scale vertical CW with40cm length?20cm width?120cm depth dimensions(working volume of40L) was built with PVC and organic glass(one facet for observa-tion).Iris pseudacorus was planted on surface of the CW at an initial density of22plants/m2.The CW bed consisted of three functional layers,the water distribution layer(0e20cm),the treatment layer(20e100cm),and the water collection layer (100e120cm).The treatment layer was?lled with lava rock with a particle diameter of8e10mm,and the water collection layer was?lled with gravel of10e20mm in diameter.
2.1.2.Start-up and operation strategy
C/N ratio was applied as an indicator to control organic loading rates in the in?uent.The experiment began on5 December2011,and involved the following six stages(total 245days):Start-up stage(C/N?2.0)from5December to21 February;Stage I(C/N?4.0)from22February to28March; Stage II(C/N?6.0)from29March to1May;Stage III(C/N?8.0)
w a t e r r e s e a r c h64(2014)32e4133
from 2May to 7June;Stage IV (C/N ?10.0)from 8June to 11July;and Stage V (C/N ?12.0)from 12July to 16August.The synthetic wastewater was prepared daily in a feeding tank,and pumped into the CW though ?umes in the distribution layer.The synthetic wastewater composition in each opera-tion stage is summarized in Table 1.Hydraulic and nutrient loading rates in all stages were maintained at 0.33m 3/m 2d and 8.23g-N/m 2d,respectively,whereas the organic loading rate varied between 20and 120g-COD/m 2d,depending on different operation stage (Table 1).The CW was placed in-doors,and in?uents and ef?uents ranged in temperature from 15.1to 27.0 C.
The CW was operated under tidal ?ow conditions,which were generated by a metering pump (ProMinent ?,China)and programmable timer.The ?ood-and-drain cycle was set to rhythmically occur every 36h,providing 24h of wastewater-bed contact in a “wet ”phase,and 12h of aeration in a “dry ”phase.The “wet ”phase began with a 10-min complete in?uent bed ?lling from a feeding tank,and the “dry ”phase began with a 10-min complete ef?uent bed drainage to an ef?uent tank.As the treated wastewater retreats during the “dry ”phase,air pulled into the porous bed from the top of the CW provides a considerable oxygen resource.The oxygen transfer from the atmosphere to the bed also occurred via natural diffusion,and contributed to re-oxygenation when the bed was drained for aeration.
2.2.Sample collection and determination
Water samples were collected from the inlet and outlet at least three times during each operation stage,and analyzed immediately at the Key Laboratory of Water and Sediment Sciences,Peking University.The following parameters were measured:chemical oxygen demand (COD)was determined with a HACH DR2800(HACH,USA);and ammonium-nitrogen (NH 4tàN),nitrite-nitrogen (NO 2ààN),nitrate-nitrogen (NO 3ààN),and total nitrogen (TN)were measured using a spectrophotometer UV-1800(SHIMADZU,Japan).All variables were analyzed according to standard analytical procedures (APHA,1995).
Microorganism samples were collected from pre-buried columns at the end of week 3,5,9,15,20,25,30,and 35.During each sampling,the pre-buried column was removed from the CW bed,placed in an ice incubator,and immediately sent to the Key Laboratory of Water and Sediment Sciences,Peking University for DNA extraction.
2.3.Quantitative Polymerase Chain Reaction (qPCR)2.3.1.
DNA extraction
Soil DNA kits D5625-01(Omega,USA)were used to extract and purify the total genomic DNA from samples.Extracted genomic DNA was detected by 1%agarose gel electrophoresis and stored at à20 C until use.
2.3.2.Primer design
Quantitative analysis was conducted on the 16S rRNA frag-ment of Bacteria (bacterial 16S rRNA),Archaea (archaeal 16S rRNA),and anammox bacteria (anammox bacterial 16S rRNA),and target fragments of the following functional genes:ammonia monooxygenase (amoA ),nitrite oxidoreductase (nxrA ),membrane-bound nitrate reductase (narG ),copper-containing nitrite reductase (nirK ),cd1-containing nitrite reductase (nirS ),and nitrous oxide reductase (nosZ ).All primers summarized in Supplementary Table S1were syn-thesized by Invitrogen Biotechnology Company (Shanghai,China),and diluted to a 10m mol/L concentration.
2.3.3.qPCR
qPCR was performed on a MyiQ2Real-Time PCR Detection System (Bio-Rad,USA)in ?nal 20m L volume reaction mixtures containing the following components:10m L SYBR Green I PCR master mix (Applied Biosystems,USA),1m L template DNA (sample DNA or plasmid DNA for standard curves),forward and reverse primers (Table S2),and sterile water (Millipore,USA).qPCR was performed in a three-step thermal cycling procedure,and the protocol and parameters for each target gene are presented in Supplementary Table S2.Each qPCR ampli?cation was performed in 40cycles and followed by a melting curve analysis.Sterile water was used as a negative control,and the data obtained from qPCR were normalized to copies per gram of lava rock.
2.3.4.Standard curves
The plasmids containing speci?c bacterial 16S rRNA,archaeal 16S rRNA,anammox bacterial 16S rRNA,and other functional genes,(i.e.amoA ,nxrA ,narG ,nirK ,nirS ,and nosZ )were man-ufactured by Majorbio BioPharm Technology Company (Shanghai,China).The standard samples were diluted to yield a series of 10-fold concentrations and subsequently used for qPCR standard curves.The R 2value for each standard curve exceeded 0.99,indicating linear relationships over the con-centration ranges used in this study.
w a t e r r e s e a r c h 64(2014)32e 41
34
2.4.Data analysis
In?uent and ef?uent concentrations and HRT(36h)were used to calculate removal ef?ciencies and transformation or accu-mulation rates of COD,NH4tàN,NO2ààN,NO3ààN,and TN.
Pearson correlation coef?cients were calculated to eval-uate ecological associations between nitrogen transformation genes using SPSS Statistics20(IBM,USA).
Stepwise regression models were built to determine the quantitative response relationships between nitrogen trans-formation rates and functional genes using SPSS Statistics20 (IBM,USA).
3.Results and discussion
3.1.Overall performance of TF CW
3.1.1.COD,NH4tàN,and TN removal ef?ciencies
COD removal ef?ciency was75%at the end of the Start-up stage,and increased to85e95%during the operation period (Stages I to V)(Fig.1).Various CW types have been adopted to remove COD from domestic wastewater(Vymazal,2011).In many typical HF CWs built in European countries(i.e.Czech Republic,Denmark,Germany,Poland,and Slovenia),the average COD removal ef?ciency achieved in treating domestic wastewaters(COD?200e287mg/L)ranged from64.3to82.0% (Vymazal,2009;Vymazal,2002;Vymazal and Kr€o pfelov a, 2008).VF CWs are also well known and widely used for COD removal from domestic wastewater.Zurita et al.(2009)built two VF CWs to remove COD(247.5±32.4mg/L)from domestic wastewater,and achieved a removal ef?ciency of77.2±2.5% and83.3±2.5%.Similarly,Sklarz et al.(2009)showed a270to 40mg/L(84%)reduction in COD from domestic wastewater using a VF https://www.doczj.com/doc/ba14033690.html,pared with these conventional systems, the TF CW in our study achieved a higher removal ef?ciency (85e95%)under a wide COD concentration range(120e360mg/ L)due to the enhanced oxygen transfer generated by the tidal ?ow operation.
NH4tàN removal ef?ciency was approximately63%dur-ing the Start-up stage,and reached73%at the end of Stage I, decreased to63%at the end of Stage II,remained at63%until the end of Stage III,increased to70%at the end of Stage IV,and reached an80%peak at the end of Stage V.TN removal ef?-ciency decreased from70%at week3to48%at week9during the Start-up period,then steadily increased from50%at the end of Stage I to66%at Stage II,67%at Stage III,73%at Stage IV,and82%at Stage V.Verhoeven and Meuleman(1999) indicated that even though CWs were effective and reliable in removing COD,treatment performance in eliminating inorganic nutrients from domestic wastewater was often limited and unsatisfactory.In particular,TN removal was often inadequate in most CWs,and some studies demon-strated a single-stage CW could not achieve satisfactory ni-trogen removal(Liu et al.,2013;Vymazal,2007).However,high TN removal ef?ciency was achieved within a range of50e82% under varying organic loading rates in our TF CW by inter-mittently providing aerobic conditions for nitri?cation,and anaerobic conditions for denitri?cation.
Therefore,the TF CW in our study combined the advan-tages of HF CW(more anaerobic conditions)and VF CW(more aerobic conditions),and successfully demonstrated increased COD and nitrogen removal ef?ciency could be achieved simultaneously in one single-stage CW.
3.1.2.Nitrogen transformation and removal
The synthetic wastewater used in this study was derived from Beijing groundwater,in which the NO3ààN concentration ?uctuated between1.2and5.3mg/L throughout our study. Results showed negligible nitri?cation during the?rst35d of the Start-up stage,i.e.NO3ààN in Fig.2(a)was from groundwater,not the nitri?cation product.Therefore,NH4tàN adsorption by porous lava rock was not only the dominant mechanism in NH4tàN removal from wastewater,but also the largest contributor to TN removal during the front half of the Start-up stage(Fig.2(a)).Chemoautotrophic nitrifying bacteria are slow growing organisms with optimal growing conditions above30 C(Loveless and Painter,1968),conse-quently it took more than35d for these microorganisms
to Fig.1e C/N effects on COD(a),NH4tàN(b),and TN removal(c).
w a t e r r e s e a r c h64(2014)32e4135
function in the CW,since we initiated our study in winter (5December 2011)under low temperature conditions (15.1 C).Signi?cant NO 2ààN was observed in the second half of the Start-up period (42,49,and 63d),suggesting nitri?cation began contributing to NH 4tàN removal in the system.How-ever,an increased TN ef?uent concentration during 36e 70d was observed from Fig.2(a),indicating NH 4tàN adsorption by lava rock decreased after the ?rst 35d and denitri?cation was absent and did not contribute to TN removal.
In Stage I (C/N ?4),adsorption and nitri?cation removed two-thirds of the NH 4tàN from the https://www.doczj.com/doc/ba14033690.html,pared to the previous stage,Stage I (C/N ?4)showed full nitri?cation in the system,as the NO 3ààN concentration (5.8e 9.9mg/L)was 1.2e 2.1-fold greater in ef?uent than the average in?uent (range of 1.2e 5.3mg/L;average 4.7mg/L).However,NO 2ààN and NO 3ààN accumulation in Stage I suggested denitri?ca-tion was de?cient and limited due to a lack of suf?cient organic carbon that could be utilized as electron donors in the denitri?cation process (Cervantes et al.,2001).When the C/N ratio was increased from six to 12during the remaining operation periods (Stages II to V),NO 2ààN and NO 3ààN accumulation essentially ceased in the system,and TN grad-ually decreased from 11.8to 6.2mg/L (66e 82%removal).Therefore,a C/N ratio greater than six was required to achieve
complete denitri?cation (without NO 2ààN and NO 3ààN accumulation),and a satisfactory TN removal (>66%).The highest TN removal achieved in our study was 82%under a C/N ratio of 12,consistent with the studies investigated by Munch et al.(1996)and Chiu et al.(2007),where results showed optimal nitrogen removal (in SBR system)was ach-ieved at respective C/N ratios of 11.2and 11.1.
The NH 4tàN,NO 3ààN,NO 2ààN,and TN trans-formation rates (or “accumulation rates ”if ef?uent N con-centration was increased)were calculated,and are shown in Fig.2(b).During the formal operation stages,the NH 4tàN transformation rate exhibited a ?uctuating increase from an average of 14.8g/m 3d in Stage I to an average of 15.1g/m 3d in Stage V.Signi?cant NO 3ààN transient accumulation was observed in Stage I with an average accumulation rate of 3.3g/m 3d,and steady NO 3ààN removal was achieved at an average transformation rate of 3.1g/m 3d during Stages II to V.The average NO 2ààN accumulation rate gradually decreased from 2.3g/m 3d in Stage I to 0.64g/m 3d in Stage V.TN transformation rate (denitri?cation rate)markedly increased from an average of 12.2g/m 3d in Stage I to 18.5g/m 3d in Stage V with an increased C/N ratio,congruent with other studies (Isaacs et al.,1994;Peng et al.,2007).These observations can be explained as follows:denitri?cation is primarily performed by heterotrophic bacteria (Vesilind,2003),which use organic carbon as an energy source,and require an appropriate C/N ratio to function properly in the system for TN removal (Chiu et al.,2007).Our results from Fig.2indicated that the increased C/N ratio had a signi?cant positive impact on TN removal (One-way ANOVA analysis,P <0.05).However,this positive effect became attenuated when insuf?cient nitri?-cation (ef?uent DO ranging from 0.49to 0.81mg/L)became the limiting factor for TN removal under a C/N ratio exceeding six.Hence,NH 4tàN and TN removal ef?ciency,to a certain extent,remained restricted by insuf?cient oxygen transfer under our operation strategy (24h ?ooded and 12h drained).Therefore,the maximum potential for NH 4tàN and TN removal must be further explored under an increased drain phase time for the TF CW.
3.2.Nitrogen removal pathway 3.2.1.
Absolute abundance
The absolute abundance of bacterial 16S rRNA,archaeal 16S rRNA,anammox bacterial 16S rRNA,amoA ,nxrA ,narG ,nirK ,nirS ,and nosZ were quanti?ed during the entire operation period to determine their dynamic population shifts.
Results from Fig.3(a)showed the absolute abundance of bacterial 16S rRNA markedly increased from 2.9?107copies/g on day 21to 1.5?109copies/g on day 245with the increased carbon source.Archaeal 16S rRNA increased from 1.2?102copies/g on day 105to 1.8?104copies/g on day 245,showing late emergence and low initial abundance.The standard curve of Archaeal 16S rRNA we used in this study ranged from 1.0?102to 1.0?106copies/g with R 2?0.995,leading to precise quanti?cation robust to measurement errors.Although Archaea were not dominant in the microbial community during the entire operation period,they may potentially play important roles in nitrogen transformation and removal in the system (Angnes et al.,2013
).
Fig.2e Nitrogen transformations (a)and nitrogen
transformation rates (b)of TN,NH 4tàN,NO 3ààN,and NO 2ààN under different C/N ratios.Note:The positive values of TN and NH 4tàN in the above ?gure indicated that they were transformed (reduced)in the system,while the negative values of NO 2ààN and NO 3ààN indicated that they were accumulated (increased)in the system.
w a t e r r e s e a r c h 64(2014)32e 41
36