ENVIRONMENTAL MICROBIOLOGY

Microbial Communities Involved in Biological AmmoniumRemoval from Coal Combustion Wastewaters

Tatiana A. Vishnivetskaya & L. Suzanne Fisher &

Greg A. Brodie & Tommy J. Phelps

Received: 10 August 2012 /Accepted: 6 December 2012 /Published online: 13 January 2013# Springer Science+Business Media New York (outside the USA) 2013

Abstract The efficiency of a novel integrated treatment sys-tem for biological removal of ammonium, nitrite, nitrate, andheavy metals from fossil power plant effluent was evaluated.Microbial communities were analyzed using bacterial andarchaeal 16S rRNA gene clone libraries (Sanger sequences)and 454 pyrosequencing technology. While seasonal changesin microbial community composition were observed, the sig-nificant (P=0.001) changes in bacterial and archaeal commu-nities were consistent with variations in ammoniumconcentration. Phylogenetic analysis of 16S rRNA genesequences revealed an increase of potential ammonium-oxidizing bacteria (AOB), Nitrosomonas, Nitrosococcus,Planctomycetes, and OD1, in samples with elevated ammoni-um concentration. Other bacteria, such as Nitrospira, Nitro-coccus, Nitrobacter, Thiobacillus, ε-Proteobacteria,Firmicutes, andAcidobacteria, which play roles in nitrificationand denitrification, were also detected. The AOB oxidized56% of the ammoniumwith the concomitant increase in nitriteand ultimately nitrate in the trickling filters at the beginning ofthe treatment system. Thermoprotei within the phylum Cren-archaeota thrived in the splitter box and especially in zero-valent iron extraction trenches, where an additional 25% of theammonium was removed. The potential ammonium-oxidizingArchaea (AOA) (Candidatus Nitrosocaldus) were detectedtowards the downstream end of the treatment system. The

design of an integrated treatment system consisting of tricklingfilters, zero-valent iron reaction cells, settling pond, and anaer-obic wetlands was efficient for the biological removal ofammonium and several other contaminants from wastewatergenerated at a coal burning power plant equipped with selec-tive catalytic reducers for nitrogen oxide removal.

Introduction

Excess ammonia (NH3) along with oxygen is fed into theselective catalytic reduction (SCR) systems of fossil-fueledelectric utility plants to achieve greater removal of nitrogenoxides (NOX) and increase the life of the catalyst [1]. Unreactedammonia typically results in a discharge stream that integrateswith the combustion byproduct (fly ash) handling system andultimately enters wastewater streams. Streams from SCR sys-tems combined with aqueous fly ash handling systems mayresult in localized transient ammonium (NH4

+) concentrationsapproaching 4 mg/L, often in conjunction with the trace levelsof other potential fly ash-related contaminants such as mercury(Hg), arsenic (As), selenium (Se), and boron (B) [2].

Ammonium, if discharged inappropriately, may negativelyimpact air and water quality, as well as fish and human health.The adverse impacts of NH4

+-containing wastewaters may bemitigated by facilitating the nitrification and denitrificationprocesses. An experimental onsite wastewater treatment tech-nology for removing SCR-derived ammonium, nitrite, andnitrate from wastewater has been developed by the TennesseeValley Authority (TVA) [3]. The treatment system, known asAquatic Toxicology Improvement and Control TreatmentTechnology (ATOXIC), also contains a separate parallel treat-ment train known as the Arsenic and Selenium ExtractionTrench (ASSET) designed to treat trace elements such as Hg,As, and Se. The ASSET portion is comprised of extractiontrench cells containing a mixture of zero-valent iron and lime-stone for trace contaminant removal and a settling/oxidation

T. A. Vishnivetskaya : T. J. Phelps (*)Biosciences Division, Oak Ridge National Laboratory,P. O. Box 2008, MS-6036, 1 Bethel Valley Rd.,Oak Ridge, TN 37831-6036, USAe-mail: phelpstj@ornl.gov

L. S. Fisher :G. A. BrodieTennessee Valley Authority, Knoxville, TN 37902, USA

Present Address:T. A. VishnivetskayaCenter for Environmental Biotechnology, University of Tennessee,Knoxville, TN 37932, USA

Microb Ecol (2013) 66:49–59DOI 10.1007/s00248-012-0152-5

basin for storage of iron hydroxide floc (Fig. 1). Other compo-nents of ATOXIC include trickling filters and constructedwetlands. The design of the system allows a comparativeevaluation of the parallel treatments for up to ~1 million Lday−1 of ammonium-enriched wastewater using only gravityflow. The systems operate under ambient conditions withannual average air temperature of 17.9 °C and annual averageprecipitation of 114 cm [3].

The aim of the current study was to evaluate the feasibil-ity of the ATOXIC/ASSET wastewater treatment system forbiological removal of ammonium with respect to ammoni-um concentration-dependent changes in the microbial com-munity composition. A combination of cultivation-independent molecular techniques and water chemistry datawas used to detect and identify putative AOB and AOAresponsible for ammonium oxidation and to ascertain theireffectiveness at ammonium removal.

Methods

Site Location and Description

The ATOXIC/ASSETwastewater treatment system is locatedat the Paradise Fossil Plant (PFP) near Drakesboro, KY, USA(36°58′N, 86°25′W). Flue gas desulfurization effluent isdiverted from the PFP via a diverter box (DB) to the ATOXICtreatment system (Fig. 1). The effluent first runs throughtrickling filters (TFs) and is deflected at a splitter box (SB)into two streams of equal volume of ~500,000 L/day. Oneeffluent stream flows to two 8×30 m zero-valent iron/lime-stone reaction cells (ZVI1 and ZVI2), and then through asettling pond (SP, 10×10 m and 3 m deep, with a 2:1 slopedberm) to one of two constructed wetlands (CW1). The second

effluent stream flows directly via gravity to another con-structed wetland (CW2). Each wetland is 65 m long×20 mwide and 2.5 m deep [3]. During construction of the treatmentsystem, the technologies that assured the development ofanaerobic environments, for example, anoxic zero-valentiron/limestone reaction cells and anaerobic constructed wet-lands, were employed [3]. The constructed cells were linedwith plastic film, and the sediments in the system becamesaturated with effluent. Monitoring of the systemwas initiatedin October 2005 and continued until May 2006. In May 2006,the wetlands were drained, refilled, and replanted with bul-rush, cattails, iris, pickerelweed, and arrowhead due to lowsurvival rates of the wetland flora. In September 2007, the TFswere inoculated with 20 L of water and “slime” from TVAexperimental wetlands in Muscle Shoals, AL, USA. Theinflow of the integrated treatment system was amended atthe injection site (Fig. 1) with ammonia beginning in May2007 and with As, Se, and Hg beginning inMay 2008 to bringthe concentrations to above 5 mgL−1 for NH4-N, 100 μgL−1

for As, 100 μgL−1 for Se, and 1 μgL−1 for Hg.

Water Analysis

Biweekly water samples were collected from sampling loca-tions (Fig. 1) during August–September 2008 and Septem-ber 2009. Temperature, specific conductance, dissolvedoxygen, and redox potential of water were measured usinga hand-held hydrolab. The total phospohorus (EPA Method365.1), Se (EPA Method 200.9), Hg (EPA Method 1634),As (EPA Method 200.9), and total ammonia (N) (EPAMethod 350.1) were analyzed as described in the EPAmethods indicated. The physico-chemical parameters mea-sured at different locations in the system are shown inTable 1.

Fig. 1 Diagram of theintegrated treatment systemillustrating water flow(indicated by arrows) throughATOXIC and ASSET paralleltreatments. TE trickling filterseffluent, TF1, TF2, TF3, TF4trickling filters, ZVI zero-valentiron reaction trenches, G gate[3]. Sampling sites are indicated

50 T.A. Vishnivetskaya et al.

Tab

le1

Phy

sico-chemical

parametersof

flue

gasdesulfurizationeffluent

measuredat

each

site

twicepermon

thdu

ring

Aug

ust–September20

08andSeptember20

09

Site

Ammon

ia,

mgL−1

Nitrateand

nitrite,mgL−1

Total

Pvalue,

mgL−1

Arsenic,mgL−1

Selenium,mgL−1

Mercury,

ngL−1

Tem

perature,°C

Con

ductivity,

μScm

−1

DO,mgL−1

DO,%

pHORP,mV

DB

0.9±0.1

3.9±0.9

0.1±0.0

0.1±0.0

0.2±0.0

93.8±115.5

24.0±2.6

4,80

8.0±28

.37.62

±0.04

95.5±5.1

7.3±0.1

331.0±111.7

TE

5.3±2.2

4.0±1.0

0.1±0.0

0.1±0.0

0.2±0.1

1324

.5±10

36.1

23.7±2.9

4,81

8.0±9.9

7.75

±0.21

96.3±2.7

7.6±0.4

365.0±16

2.6

TF4

2.2±1.2

7.3±1.2

0.1±0.0

0.1±0.0

0.2±0.1

723.8±47

1.4

23.0±3.3

4,81

6.0±15

.65.62

±0.81

69.4±14

.27.1±0.1

428.5±99

.7

TF3

2.4±1.4

6.9±1.0

0.1±0.0

0.1±0.0

0.2±0.1

733.5±39

7.7

22.9±3.2

4,81

7.0±0.0

5.64

±0.84

69.4±14

.47.1±0.1

421.5±10

2.5

TF2

2.5±1.4

6.9±1.2

0.1±0.0

0.1±0.0

0.2±0.1

853.2±52

5.6

23.0±3.0

4,80

7.0±1.4

5.51

±0.84

67.9±14

.07.1±0.1

399.5±95

.5

TF1

2.3±1.2

7.0±1.0

0.1±0.0

0.1±0.0

0.2±0.1

716.0±29

1.9

22.9±3.2

4,80

7.5±0.7

5.34

±1.04

65.8±16

.67.1±0.1

370.5±82

.7

SB

2.5±1.4

6.9±1.1

0.1±0.0

0.1±0.0

0.2±0.1

680.8±31

9.4

23.2±3.1

4,81

7.5±6.4

5.35

±0.98

66.2±16

.07.1±0.1

353.0±87

.7

ZVI1

1.2±0.8

7.9±1.2

0.1±0.0

0.1±0.0

0.2±0.1

270.7±26

2.2

21.5±3.4

4,78

4.0±62

.22.55

±1.68

30.9±21

.97.0±0.1

393.0±15

7.0

ZVI2

1.1±0.8

7.9±1.2

0.2±0.1

0.1±0.0

0.2±0.0

193.1±114.6

22.0±2.9

4,70

5.5±16

4.8

1.77

±1.58

21.8±20

.27.0±0.1

389.0±15

9.8

SP

1.0±0.6

8.1±0.8

0.1±0.0

0.03

±0.0

0.2±0.1

239.7±23

5.8

23.6±3.1

4,73

1.0±32

.57.23

±0.25

89.7±1.8

7.3±0.1

377.0±16

6.9

CW1

0.2±0.1

3.0±1.1

0.1±0.0

0.02

±0.0

0.1±0.1

20.9±6.3

23.4±1.4

4,60

0.5±115.3

0.22

±0.03

2.7±0.4

6.9±0.1

125.5±44

.6

CW2

0.1±0.05

3.9±1.2

0.1±0.0

0.01

±0.0

0.1±0.1

10.7±2.5

23.0±1.1

4,49

0.5±61

.50.26

±0.03

3.2±0.4

6.8±0.1

111.0±4.2

AP

0.2±0.1

0.9±0.7

0.1±0.0

0.01

±0.0

BDL

6.2±3.0

25.4±2.4

1,58

3.5±4.9

5.96

±2.02

75.3±22

.47.5±0.2

317.5±36

.1

DBdiverterbo

x,TEtricklingfilterseffluent,T

F1,TF2,TF3,andTF4tricklingfilters,SBsplitterbo

x,ZVIzero-valentironreactio

ncell,

SPsettlingpo

nd,C

Wconstructedwetland

,APashpo

nd,D

Odissolvedox

ygen

concentrationandpercentof

saturatio

n,ORPox

idation/redo

xpo

tential,BDLbelow

detectionlevel

Microbial Communities Involved in Biological Ammonium Removal 51

Collection of the Samples

The first sample collection occurred on March 27, 2007,1 month before ammonium was amended; the second set ofsamples was collected on August 28, 2007 after 3 months ofammonium addition; the third collection occurred on Sep-tember 29, 2008 after 16 months of continuous ammoniaamendment. Samples were collected during dry sunny daysat discrete locations (Fig. 1) by skimming the upper 2 cm ofwastewater stream sediments with a wide-mouthed contain-er and the material was transferred to sterile 2 L Nalgeneglass bottles. Samples were put on ice where they were keptduring the same day transportation to the Oak Ridge Na-tional Laboratory, Oak Ridge, TN, USA. On the next day inthe laboratory, sediments were collected by filtrationthrough Whatman Nuclepore polycarbonate membrane fil-ters or centrifugation at 4,000 rpm and 4 °C for 30 min(Sorvall Legend RT). Filters or sediment pellets were frozenat −80 °C prior to DNA extraction. The amount of sedi-ments varied from 0.3 to 2.2 gL−1.

DNA Extraction

Sediments were removed from filters by washing with anextraction buffer containing 0.1 M phosphate (NaH2PO4–Na2HPO4) solution, 0.1 EDTA, 0.1 M Tris–HCl, 1.5 M NaClwith pH 8.0. Total community genomic DNA (cgDNA) wasextracted following a described method [4]. The cgDNA fromsediments (0.5 g) was obtained using the PowerSoil™ DNAIsolation Kit (MoBio Laboratories, Inc., Carlsbad, CA, USA).

Cloning and Sanger Sequencing of Bacterial and Archaeal16S rRNA Genes

The samples collected in March 2007 (1 month beforeammonium amendment) and in August 2007 (3 months afterammonium addition) were analyzed by cloning approach.The cgDNA was amplified using Taq polymerase (Invitro-gen, La Jolla, CA, USA), bacteria-specific primers 8F (5′-AGA GTT TGA TCC TGG CTC AG-3′) and 1492R (5′-GGT TAC CTT TTA CGA CTT-3′) or Archaea-specificprimers 25Fa (5′ C(C/T)G GTT GAT CCT GCC (A/G)G3′), and 958Ra (5′ (C/T)CC GGC GTT GA(C/A) TCC AATT3′). The PCR products (~1.5 kb for bacteria and ~900 bpfor Archaea) were purified from UltraPureTM Agarose(Invitrogen) using QIAquick Gel Extraction kit (QuagenInc, Valencia, CA, USA). PCR products were ligated inpCR 2.1-TOPO vectors (Invitrogen, Carlsbad, CA, USA),transformed into One Shot TOP10 chemically competentEscherichia coli, and plated onto LB agar containing50 μgmL−1 kanamycin and 40 mgmL−1 X-gal. Transform-ants were incubated overnight at 37 °C, and 96 white colo-nies were selected and regrown separately in LB with 50 μg

mL−1 kanamycin at 37 °C and 200 rpm. Clones were thensequenced using the BigDye Terminator v3.1 Cycle Se-quencing kit and a reverse primer targeting positions 536–519 of E. coli 16S rRNA gene (5′-G(A/T)ATTA CCG CGGC(G/T)G CTG-3′) [5], in case of bacteria, or TAF primer(GCC GCC AGT GTG CTG GAA TT) from plasmid, incase of Archaea. Sequences were determined by resolvingthe sequence reactions on an Applied Biosystems 3730automated sequencer. Sanger sequences obtained in thisstudy were deposited in to the NCBI sequence databaseunder accession numbers from JX995370 to JX995497 forArchaea and from JX995498 to JX995906 for bacteria.

Pyrosequencing of Bacterial and Archaeal 16S rRNA Genes

The samples collected in September 2008 (after 16 months ofammonium spiking) were analyzed using 454 pyrosequencingthat provided about ten times more sequences than traditionalcloning and Sanger sequencing. The hypervariable V4 region(~290 or ~340 bp in case of bacteria or Archaea, respectively)of the 16S rRNA gene was amplified using bacterial primers(forward 5′-AYTGGGYDTAAAGNG-3′ and four reverse pri-mers 5′-TACCRGGGTHTCTAATCC-3′; 5′-TACCAGAGTATCTAATTC-3′; 5′-CTACDSRGGTMTCTAATC-3′; 5′-TACNVGGGTATCTAATCC-3′, mixed at ratio 6:1:2:12, re-spectively) or archaeal primers (forward 5′-CAGYMGCCRCGGKAAHACC-3′ and reverse 5′- GGACTACNSGGGTMTCTAAT-3′); the 6 bp unique tag sequences were as de-scribed at the Ribosomal Database Project’s (RDP) pyrose-quencing pipeline [6]. Conditions for PCR and purification ofamplicons were as described [7]. Sequencing reactions wereperformed on a 454 Life Sciences Genome Sequencer FLX(Roche Diagnostics, Indianapolis, IN, USA). Raw 454 FLXdata (~88 Mb) were initially processed through the RDP’spyrosequencing pipeline [8]. During this process, the sequen-ces were sorted by tag; the 16S primers, sequences shorter than200 bp, and sequences with any number of unknown nucleo-tides were removed. Each sample was obtained from ~6,000 to~10,000 bacterial sequences and from ~500 to 7,000 archaealsequences. Bacterial primers captured less than 0.2 % of thearchaeal sequences, whereas archaeal primers [9] capturedapproximately 60 % of the bacterial sequences, which wereremoved from analyses. Chimeric sequences were detected byusing the Chimera Slayer [10] and removed from the analyses.The 454 pyrosequences obtained in this study were depositedinto the Sequence Read Archive (SRA) of NCBI sequencedatabase under accession number SRA060259.

Phylogenetic and Statistical Analyses

Both bacterial and archaeal 16S rRNA sequences wereassigned to a set of hierarchical taxa using a Naïve BayesianrRNA Classifier v.2.0 with confidence threshold of 80 % for

52 T.A. Vishnivetskaya et al.

Sanger sequences and 50 % for 454 pyrosequences [11].Sanger sequences were aligned using ClustalWMultiple align-ment [12] within the program BioEdit v.7.0.5.3 [13]. To deter-mine variations between environments, the microbialcommunities were compared at the sequence level usingweighted UniFrac [14, 15]. For input to UniFrac, theneighbor-joining tree generated using MEGA4 software [16]was rooted with 16S rRNA gene sequence ofMethanosarcinamazei (EF452664) in case of bacteria and Arthrobacter globi-formis DSM20124 (NR_026187) in case of Archaea. Opera-tional taxonomic units (OTUs) were defined at 97 % identity,and one representative sequence was selected for major OTUs.The environmental input file for UniFrac contained the countof howmany times the representative sequence appeared in theclone library.P valueswere corrected formultiple comparisonsby multiplying the calculated P value by the number of com-parisons made (Bonferroni correction) [14, 15].

Results

Fate of Ammonium and Other Elements

Low concentrations of ammonium were detected in watersamples collected in the upstream DB site (average of0.9 mgL−1) and downstream AP site (average 0.2 mgL−1)(Fig. 1). After amendment with SCR-derived ammonium,the ammonium concentration increased to an average of5.3 mgL−1 with an insignificant change in pH (Table 1).Ammonium concentrations sampled at the end of the treat-ment systems were 0.12 mgL−1 in CW2 (ATOXIC) and0.18 mgL−1 CW1 (ASSET) resulting in greater than 97 %reduction. The higher ammonia and lower nitrate/nitriteconcentrations in CW1 compared to CW2 (Table 1) couldbe a result of nitrate reduction and conversion to ammoniumin ZVI extraction trenches (Fig. 1). Results indicated up to56 % of the ammonium oxidation with the concomitantincrease in nitrite and ultimately nitrate occurred in theTFs (Table 1), 25 % in the ZVI trenches or directly in thepassages of the ATOXIC system, and approximately 20 %in SP and CWs. The decreasing ammonium oxidationthrough the system is in agreement with the concentrationof dissolved oxygen which decreased to the end of thetreatment system (Table 1). After spiking, the nitrate andnitrite concentrations doubled in TF sites returning to thebackground levels in the CWs and afterwards were less than1 mgL−1 (Table 1). Arsenic and Se were lowered by a factorof less than 10 within the treatment systems. After adding1 μgL−1 Hg to the inflow to the treatment system, Hgconcentration increased 14-fold in the trickling filters efflu-ent and then decreased by 44 % in the TFs, 40 % in ZVIs orpassage through the ATOXIC system, 16 % in the CWs, andeven lower in AP (Table 1).

Microbial Community Size and Diversity

The yields of the cgDNA from water (0.001–0.083 μgmL−1) and sediments (0.2–26 μgg−1) were estimated. Thetotal bacterial/archaeal populations ranged from (0.03–2)×107 cells mL−1 to (0.2–6)×109 cells g−1 for water and wetsediments, respectively. This estimate was based on thepredicted effective genome size of 4.7 Mb for the soilbacterial/archaeal population as estimated from metage-nomics data [17, 18] assuming the weight of the ge-nome this size would be 4.05 fg [19]. Thougheukaryotic green algae were present and formed thickmats in the ZVI reaction cells, SP, and CWs, we did notinclude the eukaryotes in the population size estimate. Ifincluded, the eukaryotic component may reduce the sizeof the cell population by 25 % based on prediction ofthe effective genome size [17].

Microbial Community Composition in Background Samples

The ATOXIC/ASSET treatment systems represent an open,natural environment which could be effected by seasonalclimate changes. The samples collected in spring (March2007) prior to ammonium spiking and DB samples collectedin the upstream of the injection site represent a background.The DB samples could provide knowledge about seasonalchanges in bacterial and archaeal communities. Whereas theDB bacterial community showed seasonal variations inabundances of Proteobacteria, Actinobacteria, Bacteroi-detes, Verrucomicrobia , and Cyanobacteria , theammonium-oxidizers and the phyla Nitrospira, Planctomy-cetes, OD1, ε-Proteobacteria, Acidobacteria, Firmicuteswere detected at low (<1.5 %) levels (data not shown).The DB archaeal community was mostly (65.4–97.1 %) represented by phylum Euryarchaeota and un-classified Archaea, while Crenarchaeota (Thermoprotei)were minor components. The bacteria and Archaea pu-tatively involved in the nitrogen cycle were of lowabundance in background samples, suggesting that thisprocess may not be prevalent immediately in the up-stream of the treatment system.

In March 2007 prior to ammonium spiking, as waterflowed through the system toward the wetlands and dis-charge gates, the abundance of α-Proteobacteria, β-Proteo-bacteria, and Bacteroidetes decreased, whereas the oppositeeffect was observed for ε-Proteobacteria, Chloroflexi,OP11, and unclassified bacteria (not shown). Only δ-Proteobacteria showed differences in abundance anddiversity between treatment systems. The higher levelof δ-Proteobacteria in the ASSET system (up to27.7 % in CW1) in comparison to ATOXIC (4.0 % inCW2) system is likely to be a result of zero-valent ironextraction trenches.

Microbial Communities Involved in Biological Ammonium Removal 53

Changes in the Nitrifying/Denitrifying Community

The changes in the microbial community could be followedin two dimensions: (1) along the system maturation over thetime from March 2007 (no ammonium spiking) to Septem-ber 2008 (16 months of ammonium spiking) and (2) alongthe ammonium gradient at the same time point.

Temporal Shift

Bacterial Community In August 2007, after 3 months ofcontinuous ammonium injection, a significant (P=0.001)shift in the bacterial community structure was observed inCWs and treatment discharge gates (Fig. 2). According toprincipal component analysis (PCoA), 39.3 % of the differ-ences between bacterial communities could be explained bychanges in availability of nutrients, low dissolved oxygenconcentration, and slightly acidic conditions (Table 1) inorganic-rich wetlands. The variations along the second axis(18.2 %) could be a result of the presence of ammonium inAugust samples (Table 1). The continuous amendment withammonium led to the increased abundance of Nitrospira, α-Proteobacteria, β-Proteobacteria, ε-Proteobacteria, Planc-tomycetes, OD1, Firmicutes, Acidobacteria, Actinobacteria,and Cyanobacteria with concurrent decreases of δ-Proteo-bacteria and OP11 (Figs. 3 and 4). The bacterial phyla that

contain potential AOB and other members of nitrogen cy-cling showed up to 20-fold increases in their relativeabundance.

Archaeal Community Significant amounts of Archaea (22.7–76.2 %) in March 2007 and August 2007 samples studied bythe cloning approach were affiliated with uncultured archaeonor Crenarchaeote clones, whereas greater than 80 % of the454 pyrosequences from the September 2008 sampling wereclassified at class level (Figs. 5 and 6). The presence ofMethanomicrobia and Thermoplasmata both within the phy-lum Euryarchaeota and Thermoprotei (Crenarchaeota) wererevealed within the ASSET and ATOXIC systems regardlessof the date or sequencing procedure. Archaeal pyrosequencesfrom September 2008 samples affiliated with the Euryarch-aeota were detected in all samples with greater abundance inTFs (up to 88.9 %) and lower in ZVI (3.3 %). The long-termammonium spiking negatively affected the presence of Hal-obacteria (Euryarchaeota) from a maximum of 24.4 % inMarch 2007 before ammonium spiking to 1.4 % in August2007 by the cloning procedure and then to below detection bythe 454 pyrosequencing procedure from September 2008. Therelative abundance of Thermoprotei averaged at 27.6 % inMarch 2007 but reached 91.6 % of the sequences from thesplitter box and 96.1 % of the sequences from the ZVI bySeptember 2008 as estimated by 454 pyrosequencing (Fig. 6).

Fig. 2 PCoA analysis ofbacterial communities obtainedwith weighed and normalizedUniFrac using the 16S rRNAgene sequences isolated fromthe ASSET and ATOXICpurification systems in Marchand August of 2007. Neighbor-joining tree was rooted with anarchaeal outgroup. The ends ofthe alignment were trimmedprior to the analysis so that allof the aligned sequences werethe same length. The sampledescription is given in Fig. 1;the following after dash letterindicates months: M for March(black symbols) and A for Au-gust (gray symbols). Sample TFwas not collected in March be-cause trickling filters did notcontain any effluent at the timeof collection

54 T.A. Vishnivetskaya et al.

In comparison to up-gradient samples, the ZVI trenchesexhibited similar redox potentials characterized by a two-fold depletion of oxygen (Table 1). Regardless of the timeand sequencing method, Thermoprotei were a small fractionof the community in locations with higher concentrations ofdissolved oxygen. The sequences affiliated with CandidatusNitrosocaldus yellowstonii, ammonia-oxidizing Archaea [20],were detected in downstream environments.

Shift Along the Ammonium Gradient

Bacterial community Changes in the microbial communitycomposition and diversity along the ammonium gradientwere apparent at all times with more explicit differencesbetween the sites observed in August 2007 and September2008 (Figs. 3 and 4). At 16 months post-ammonium spik-ing, the potential AOB such as Nitrosomonas and Nitro-sococcus reached 2.5 % and 0.5 %, respectively, of the totalbacterial sequences in the beginning of the system whereNH4

+ concentration was higher in comparison to

background and downstream sites. Sites in the beginningof the system also showed elevated concentrations of nitrateand nitrite (Table 1), and community analyses detected therethe presence of 6.1 % Nitrospira, 3.6 % Nitrococcus, and0.1 % Nitrobacter sequences. The significant loss of ammo-nium and concurrent increase in nitrate and nitrite concen-tration evidence the ongoing nitrification process in thebeginning of the system. The concentration of nitrate andnitrite stayed at similar levels in TFs and SB, then increasedslightly in ZVI trenches, and decreased sharply in the con-structed wetlands, suggesting the denitrification towards thelow redox downstream end of the system. The wetlandswith pH 6.9 and low concentration of dissolved oxygenshowed increases in the relative abundance of ε-Proteobac-teria, δ-Proteobacteria, Planctomycetes, OD1, and anaero-bic denitrifying bacteria related to Denitratisoma from β-Proteobacteria (Figs. 3b and 4). Some bacteria weredetected only in certain parts of the system, for example,Nitrospira, nitrite-oxidizing bacteria, were present in siteswith high nitrate-N/nitrite-N concentrations; while bacteria,such as Planctomycetes and sulfur-oxidizing denitrifying bac-teria Sulfurimonas and Sulfurovum within ε-Proteobacteriawere detected only towards the end of the system regardless ofthe sequencing procedure employed (Figs. 3 and 4). Cyano-bacteria, even though they preferentially use NH4

+ over othernitrogen sources [21], were low in the TFs and SB and wereback to background level in CWs and discharge gates.

Archaeal community In August 2007 samples, the Metha-nococci (Euryarchaeota) were detected at high abundance(39.4 %) in TFs, with methanogens persisting at lowerabundances through ASSET and ATOXIC systems. In Sep-tember 2008, the significant changes in archaeal communitywere consistent with NH4

+ concentrations observed alongthe system resulting in increased relative abundance ofThermoprotei to 91.6 % in SB and 96.1 % in ZVI with theirsubsequent decrease along the remaining portion of thetreatment system to 37.6 % in CW and further to 7.3 % atthe gate (Fig. 6). Sequences affiliated with CandidatusNitrosocaldus, Methanobacteria, Thermoplasmata, Ther-mococci, and Archaeoglobi were detected only in samplestoward the end of the systems.

Discussion

Taxonomic evaluation of the microbial community in thenewly constructed ammonium treatment system by 454pyrosequencing and Sanger sequencing provided similarresults. Both approaches detected significant increases ofmicroorganisms essential to the nitrogen cycle consistentwith the system maturation. The prevalence of certainmicroorganisms in particular parts of the system was

Fig. 3 Changes in bacterial community structure in samples collectedin March (a) and August (b) 2007 analyzed by cloning and Sangersequencing. The sample description is given in Fig. 1. The bacteria ofthe phylum Chlamydiae were not detected in any of these samples.Group “Others” includes Gemmatimonadetes, Spirochaetes, Chlorobi,TM7, and OP11. Number of sequences analyzed from each sample isindicated at the top

Microbial Communities Involved in Biological Ammonium Removal 55

defined by the availability of oxygen and nutrients. Thus,slow-growing nitrifying bacteria Nitrococcus, Nitrosomo-nas, Nitrosococcus, Nitrobacter, and Nitrospira, which re-quire a steady supply of ammonia and oxygen [22], wereabundant in the oligotrophic (TF and SB) sites providing the

highest level of ammonium oxidation. Thermoproteales andDesulfurococcales comprised the majority of archaeal com-munity and dominated potential AOB in SB and ZVI sites.A number of cultured Crenarchaeota are sulfur-dependentthermophiles which can use nitrate, elemental sulfur, thiosul-fate, and sulfite as electron acceptors for growth andmolecularhydrogen as the electron donor with H2S,·H2O or NH4

+ as endproducts of their metabolism [23]. It was shown recently thatArchaea affiliated with the Crenarchaeota are ubiquitous andabundant microbial constituents of low temperature habitats[24]. Since these mesophilic Archaea were different fromhyperthermophilic Crenarchaeota, they were considered as athird archaeal phylum Thaumarchaeota [25]. The majority ofthese organisms have not been cultivated, and little is knownabout their physiological characteristics and ecological signif-icance. Recent studies described mesophilic archaeal culturescapable of chemolithoautotrophic oxidation of NH4

+ to NO2−

[26]. The first cultivated mesophilic thaumarchaeote Nitro-sopumilus maritimus can use HCO3

− and NH4+ as sole carbon

and energy sources [27, 28].The anaerobic ammonium oxidation component of the

nitrogen cycle was detected in anaerobic wetlands and wasrepresented by OD1, followed by Planctomycetes. Theseorganisms, despite their phylogenetic divergence, have beenshown to be involved in anaerobic ammonium oxidation usingnitrite produced by aerobic nitrifying bacteria and Archaea[29–31]. While 96.7 % of Planctomycetes sequences wererelated to uncultured representatives of the phylum, 15sequences were clustered at 91.3–94.2 % similarity with 16SrRNA gene sequences from anaerobic ammonium-oxidizingPlanctomycetes Candidatus Kuenenia stuttgartiensis(AF375995) and Candidatus Brocadia anammoxidans(AF375994). Seven sequences were 92.2 % identical to Rho-dopirellula baltica SH-1; other 12 and 11 sequences were100 % identical to Pirellula staleyi DSM6068 and uncultured

Fig. 4 Changes in bacterialcommunity structure in samplescollected in September 2008analyzed by 454pyrosequencing. The sampledescription is given in Fig. 1.Group “Others” includesDeinococcus–Thermus,Gemmatimonadetes,Spirochaetes, Chlorobi, TM7,OP11, BRC1, OP10, and WS3.Number of sequences analyzedfrom each sample is indicated atthe top

Fig. 5 Changes in archaeal community structure in samples collectedin March (a) and August (b) 2007 analyzed by cloning and Sangersequencing. The sample description is given in Fig. 1. Number ofsequences analyzed from each sample is indicated at the top

56 T.A. Vishnivetskaya et al.

Rhodopirellula sp., respectively. The whole genome sequencesof the anaerobic ammonium-oxidizer Candidatus Kueneniastuttgartiensis, and aerobic R. baltica SH-1 and P. staleyiDSM6068 showed the presence of similar pathways for am-monium transport via ammonium channel in these bacteria.

Archaea that dominated the CWs were methanogens,mostly Methanobacteria. A variety of methanogenic Ar-chaea are capable of ammonium assimilation [32] and ni-trogen fixation; however, as shown for Methanococcusmaripaludis, methanogens usually decrease nitrogen fixa-tion when cells encounter a nitrogen source more favorablethan N2, such as NH4

+ [33, 34]. Numerous microorganismsassimilate inorganic nitrogen during the formation of cellu-lar biomass, and they preferentially utilize ammonium overalternative inorganic nitrogen. A number of bacteria withinthe class γ-Proteobacteria such as Pseudomonas, Xantho-monadales, Chromatiales, Legionellales, and Aeromona-dales which are capable of ammonium assimilation weredetected in the treatment systems at high abundance.Sequences affiliated with Lactobacillus were also abundantin the treatment systems. This is important because Lacto-bacillus plantarum strain WCFS1 reduces nitrate with con-comitant formation of nitrite and ammonia [35]. Otherabundant representatives included, Lachnospiraceae, whichare known to consume ammonium as sole nitrogen source[36]. Acidobacteria were also abundant, and the presence ofnitrite reductase (NirA) in acidobacterial genome suggeststhat they are able to perform nitrate reduction, while noevidence of the ability to fix N2 (NifH), oxidize ammonium(AmoA), or denitrify nitrous oxide (NosZ) were found [37].

Other important autotrophic bacteria detected in the treat-ment system included thiobacilli, which may mediate Hgvolatilization and nitrate-dependent anoxic iron sulfide ox-idation in TF effluent and ZVI cells. The Hg-volatilizingbacterium Thiobacillus ferrooxidans has been shown to becapable of removing up to 62.9–75.1 % of Hg frommercury-contaminated sediments [38, 39]. A total of 4,895sequences related to thiobacilli were 97.5–99 % identical to

each other, and they had 97.1–99.5 % identity to Thioba-cillus denitrificans ATCC 25259 and T. denitrificans DSM739. As indicated by genome sequences, T. denitrificansATCC 25259 contains resistance to heavy metals [40] andis capable of nitrate-dependent oxidation of ferrous iron,iron sulfides, or sulfur [40, 41].

The potential sulfate-reducing and sulfur-oxidizing bac-teria of the phyla Firmicutes and classes of ε-, δ- and γ-Proteobacteria were also detected. The δ-Proteobacteriawere dominated by Desulfocapsa sp. and Desulfovibriosp., which are capable of sulfate reduction using nitrate asa terminal electron acceptor [42]. Even δ-Proteobacteriawere more diverse in the ASSET; both systems containedbacteria of the order Syntrophobacteraceae (Smithella). Inaquatic environments, Smithella usually exist in syntrophicassociation with H2-utilizing microorganisms and oxidizesubstrates incompletely to acetate [43]. A separate studyshowed that uncultured bacteria closely related to cul-tured members of the Syntrophobacteraceae (which donot use sulfate as an electron acceptor) carry dissimilato-ry sulfite reductase (dsrAB) gene and are able to reducesulfur [44]. The sulfur-reducing bacteria of the familyPeptococcaceae [45] were also detected. Other bacteriadetected were ε-Proteobacteria. Cultured representativesof the free-living ε-Proteobacteria are chemolithoauto-trophs or mixotrophs, capable of either oxidizing reducedsulfur compounds and hydrogen with oxygen and/or ni-trate or oxidizing hydrogen with elemental sulfur coupledto the fixation of inorganic carbon [46, 47]. Even thoughsulfide-containing wetland waters may represent excellenthabitats for sulfur-oxidizers, the bacteria Thiothrix sp.and Thiovirga sp., which obtain energy from the oxida-tion of sulfide, thiosulfate, or elemental sulfur and usenitrate and ammonia as sole nitrogen sources [48, 49],were minor community components. Along with biolog-ical processes, sulfide may facilitate precipitates withmetalloids such as As and Se, removing the pollutantsfrom discharged water [50, 51].

Fig. 6 Changes in archaealcommunity structure in samplescollected in September 2008analyzed by 454pyrosequencing. The sampledescription is given in Fig. 1.Number of sequences analyzedfrom each sample is indicated atthe top

Microbial Communities Involved in Biological Ammonium Removal 57

Both the ASSET and ATOXIC wastewater treatmentsystems showed effective biogeochemical treatment ofammonium, Hg, As, and Se from the flue gas desulfur-ization effluent. The microbial communities within TFand SB sites, which represent a part of both the ASSETand ATOXIC wastewater treatment systems, were themost efficient in removing ammonium and mercurycontaminations.

Acknowledgments This research was sponsored by the U. S. Depart-ment of Energy Office of Fossil Energy and Office of Science Biologicaland Environmental Research, Environmental Remediation Sciences Pro-gram and performed at Oak Ridge National Laboratory (ORNL). ORNL ismanaged by UT-Battelle, LLC, for the U. S. Department of Energy undercontract DE-AC05-00OR22725. We thank Zamin Yang and MarilynKerley for help with 454 FLX pyrosequencing and Sanger sequencing,respectively. We would also like to thank Alan Mays, David Lane, MarkWolfe, and Roy Quinn of TVA for help with sampling and maintaining theATOXIC/ASSET field sites.

References

1. TVA (2006) Operational improvements to optimize selective catalyt-ic reduction systems for nitrogen oxide control at allen fossil plantunits 1, 2, and 3. Tennessee Valley Authority, Shelby County, p 15

2. Feeley TJ, Pletcher S, Carney B,McNemar AT (2006) Department ofEnergy/National Energy Technology Laboratory’s Power Plant WaterR&D Program. Power-Gen International 2006, available via Googlesearch http://www.netl.doe.gov/technologies/coalpower/ewr/pubs/Power%20Gen%202006_Water%20RD.pdf. Accessed 26 Dec 2012

3. Yost TL, Pier PA, Brodie GA (2007) Fate of As, Se, and Hg in apassive integrated system for treatment of Fossil Plant wastewater.Final Report for DOE Project DE-FC26-03NT41910, pp 86

4. Zhou JZ, Bruns MA, Tiedje JM (1996) DNA recovery from soilsof diverse composition. Appl Environ Microbiol 62:316–322

5. Lane DJ (1991) 16S/23S rRNA sequencing. In: Stackebrandt E,Goodfellow M (eds) Nucleic acid techniques in bacterial system-atics, 1st edn. Chichester, Wileys, pp 15–176

6. Cole JR, Wang Q, Cardenas E, Fish J, Chai B, Farris RJ, Kulam-Syed-Mohideen AS, McGarrell DM, Marsh T, Garrity GM, TiedjeJM (2009) The Ribosomal Database Project: improved alignmentsand new tools for rRNA analysis. Nucleic Acids Res 37:D141–D145. doi:10.1093/nar/gkn879

7. Vishnivetskaya TA, Mosher JJ, Palumbo AV, Yang ZK, Podar M,Brown SD, Brooks SC, Gu B, Southworth GR, Drake MM, BrandtCC, Elias DA (2011) Mercury and Other Heavy Metals InfluenceBacterial Community Structure in Contaminated Tennessee Streams.Appl Environ Microbiol 77:302–311. doi:10.1128/aem.01715-10

8. Cole JR, Chai B, Farris RJ, Wang Q, Kulam SA, McGarrell DM,Garrity GM, Tiedje JM (2005) The Ribosomal Database Project(RDP-II): sequences and tools for high-throughput rRNA analysis.Nucleic Acids Res 33:D294–D296

9. Porat I, Vishnivetskaya TA, Mosher JJ, Brandt CC, Yang ZK,Brooks SC, Liang L, Drake MM, Podar M, Brown SD, PalumboAV (2009) Characterization of archaeal community in contaminat-ed and uncontaminated surface stream sediments. Microb Ecol60:784–795. doi:10.1007/s00248-010-9734-2

10. Haas BJ, Gevers D, Earl AM, Feldgarden M, Ward DV,Giannoukos G, Ciulla D, Tabbaa D, Highlander SK, SodergrenE, Methe B, DeSantis TZ, Petrosino JF, Knight R, Birren BW

(2011) Chimeric 16S rRNA sequence formation and detection inSanger and 454-pyrosequenced PCR amplicons. Genome Res21:494–504. doi:10.1101/gr.112730.110

11. Wang Q, Garrity GM, Tiedje JM, Cole JR (2007) NaiveBayesian classifier for rapid assignment of rRNA sequencesinto the new bacterial taxonomy. Appl Environ Microbiol73:5261–5267

12. Thompson JD, Higgins DG, Gibson TJ (1994) Clustal-W–improv-ing the sensitivity of progressive multiple sequence alignmentthrough sequence weighting, position-specific gap penalties andweight matrix choice. Nucleic Acids Res 22:4673–4680

13. Hall TA (1999) BioEdit: a user-friendly biological sequence align-ment editor and analysis program for Windows 95/98/NT. NucleicAcids Symp Ser 41:95–98

14. Lozupone C, HamadyM, Knight R (2006) UniFrac–an online tool forcomparing microbial community diversity in a phylogenetic context.BMC Bioinforma 7:371–385. doi:10.1186/1471-2105-7-371

15. Lozupone C, Knight R (2005) UniFrac: a new phylogenetic meth-od for comparing microbial communities. Appl Environ Microbiol71:8228–8235. doi:10.1128/aem.71.12.8228-8235.2005

16. Kumar S, Tamura K, Nei M (2004) MEGA3: Integrated softwarefor molecular evolutionary genetics analysis and sequence align-ment. Brief Bioinform 5:150–163

17. Raes J, Korbel JO, Lercher MJ, von Mering C, Bork P (2007)Prediction of effective genome size in metagenomic samples.Genome Biol 8:11. doi:10.1186/gb-2007-8-1-r10

18. Angly FE, Willner D, Prieto-Dava A, Edwards RA, Schmieder R,Vega-Thurber R, Antonopoulos DA, Barott K, Cottrell MT,Desnues C, Dinsdale EA, Furlan M, Haynes M, Henn MR, HuY, Kirchman DL, McDole T, McPherson JD, Meyer F, MillerRM, Mundt E, Naviaux RK, Rodriguez-Mueller B, Stevens R,Wegley L, Zhang L, Zhu B, Rohwer F (2009) The GAAS meta-genomic tool and its estimations of viral and microbial averagegenome size in four major biomes. PLoS Comput Biol 5:e1000593

19. Ellenbroek FM, Cappenberg TE (1991) DNA-synthesis andtritiated-thymidine incorporation by heterotrophic fresh-water bac-teria in continuous culture. Appl Environ Microbiol 57:1675–1682

20. de la Torre JR, Walker CB, Ingalls AE, Konneke M, Stahl DA(2008) Cultivation of a thermophilic ammonia oxidizing archaeonsynthesizing crenarchaeol. Environ Microbiol 10:810–818.doi:10.1111/j.1462-2920.2007.01506.x

21. Ohmori M, Ohmori K, Strotmann H (1977) Inhibition of nitrateuptake by ammonia in a blue-green alga, Anabaena cylindrica.Arch Microbiol 114:225–229. doi:10.1007/Bf00446866

22. Belser LW (1979) Population ecology of nitrifying bacteria. AnnRev Microbiol 33:309–333

23. Huber H, Stetter KO (2006) Desulfurococcales. In: Dworkin M,Falkow S, Rosenberg E, Schleifer K-H, Stackebrandt E (eds) Theprokaryotes archaea bacteria: Firmicutes. Actinomycetes Springer,New York, pp 52–68

24. Kemnitz D, Kolb S, Conrad R (2007) High abundance ofCrenarchaeota in a temperate acidic forest soil. FEMS MicrobiolEcol 60:442–448

25. Brochier-Armanet C, Boussau B, Gribaldo S, Forterre P (2008)Mesophilic Crenarchaeota: proposal for a third archaeal phylum,the Thaumarchaeota. Nat Rev Microbiol 6:245–252. doi:10.1038/Nrmicro1852

26. KonnekeM, Bernhard AE, de la Torre JR,Walker CB,Waterbury JB,Stahl DA (2005) Isolation of an autotrophic ammonia-oxidizingmarine archaeon. Nature 437:543–546. doi:10.1038/nature03911

27. You J, Das A, Dolan EM, Hu ZQ (2009) Ammonia-oxidizingarchaea involved in nitrogen removal. Water Res 43:1801–1809.doi:10.1016/j.watres.2009.01.016

28. Walker CB, de la Torre JR, Klotz MG, Urakawa H, Pinel N,Arp DJ, Brochier-Armanet C, Chain PSG, Chan PP, Gollabgir

58 T.A. Vishnivetskaya et al.

A, Hemp J, Hugler M, Karr EA, Konneke M, Shin M,Lawton TJ, Lowe T, Martens-Habbena W, Sayavedra-SotoLA, Lang D, Sievert SM, Rosenzweig AC, Manning G,Stahl DA (2010) Nitrosopumilus maritimus genome revealsunique mechanisms for nitrification and autotrophy in globallydistributed marine Crenarchaea. Proc Natl Acad Sci U S A107:8818–8823. doi:10.1073/pnas.0913533107

29. Jetten MSM, Sliekers O, Kuypers M, Dalsgaard T, van Niftrik L,Cirpus I, van de Pas-Schoonen K, Lavik G, Thamdrup B, Le PaslierD, Op den Camp HJM, Hulth S, Nielsen LP, Abma W, Third K,Engstrom P, Kuenen JG, Jorgensen BB, Canfield DE, Damste JSS,Revsbech NP, Fuerst J, Weissenbach J, Wagner M, Schmidt I,Schmid M, Strous M (2003) Anaerobic ammonium oxidationby marine and freshwater planctomycete-like bacteria. ApplMicrobiol Biotechnol 63:107–114. doi:10.1007/s00253-003-1422-4

30. Strous M, Fuerst JA, Kramer EHM, Logemann S, Muyzer G, vande Pas-Schoonen KT, Webb R, Kuenen JG, Jetten MSM (1999)Missing lithotroph identified as new planctomycete. Nature400:446–449

31. Kirkpatrick J, Oakley B, Fuchsman C, Srinivasan S, Staley JT,Murray JW (2006) Diversity and distribution of Planctomycetesand related bacteria in the suboxic zone of the Black Sea. ApplEnviron Microbiol 72:3079–3083

32. Kenealy WR, Thompson TE, Schubert KR, Zeikus JG (1982)Ammonia assimilation and synthesis of alanine, aspartate, andglutamate in Methanosarcina barkeri and Methanobacterium ther-moautotrophicum. J Bacteriol 150:1357–1365

33. Kessler PS, Leigh JA (1999) Genetics of nitrogen regulation inMethanococcus maripaludis. Genetics 152:1343–1351

34. Dworkin M, Falkow S, Rosenberg E, Schleifer K-H, StackebrandtE (2007) The prokaryotes: a hand book on the biology ofbacteria. Archaea. Bacteria: Firmicutes, Actinomycetes. Springer,New York

35. Brooijmans RJ, de Vos WM, Hugenholtz J (2009) Lactobacillusplantarum WCFS1 electron transport chains. Appl EnvironMicrobiol 75:3580–3585

36. Cotta M, Forster R (2006) The family Lachnospiraceae, includ-ing the genera Butyrivibrio, Lachnospira, and Roseburia. In:Dworkin M, Falkow S, Rosenberg E, Schleifer K-H,Stackebrandt E (eds) The prokaryotes. Springer, New York, pp1002–1021

37. Ward NL, Challacombe JF, Janssen PH, Henrissat B, CoutinhoPM, Wu M, Xie G, Haft DH, Sait M, Badger J, Barabote RD,Bradley B, Brettin TS, Brinkac LM, Bruce D, Creasy T,Daugherty SC, Davidsen TM, Deboy RT, Detter JC, DodsonRJ, Durkin AS, Ganapathy A, Gwinn-Giglio M, Han CS,Khouri H, Kiss H, Kothari SP, Madupu R, Nelson KE,Nelson WC, Paulsen I, Penn K, Ren QH, Rosovitz MJ,Selengut JD, Shrivastava S, Sullivan SA, Tapia R, ThompsonLS, Watkins KL, Yang Q, Yu CH, Zafar N, Zhou LW, KuskeCR (2009) Three genomes from the phylum Acidobacteriaprovide insight into the lifestyles of these microorganisms in

soils. Appl Environ Microbiol 75:2046–2056. doi:10.1128/aem.02294-08

38. Nakamura K, Hagimine M, Sakai M, Furukawa K (1999) Removalof mercury from mercury-contaminated sediments using a com-bined method of chemical leaching and volatilization of mercuryby bacteria. Biodegradation 10:443–447

39. Olson GJ, Porter FD, Rubinstein J, Silver S (1982) Mercuricreductase enzyme from a mercury-volatilizing strain ofThiobacillus ferrooxidans. J Bacteriol 151:1230–1236

40. Beller HR, Chain PS, Letain TE, Chakicherla A, Larimer FW,Richardson PM, Coleman MA, Wood AP, Kelly DP (2006) Thegenome sequence of the obligately chemolithoautotrophic, facul-tatively anaerobic bacterium Thiobacillus denitrificans. J Bacteriol188:1473–1488

41. Beller HR (2005) Anaerobic, nitrate-dependent oxidation of U(IV)oxide minerals by the chemolithoautotrophic bacteriumThiobacillus denitrificans. Appl Environ Microbiol 71:2170–2174. doi:10.1128/aem.71.4.2170-2174.2005

42. Lopez-Cortes A, Fardeau M-L, Fauque G, Joulian C, Ollivier B(2006) Reclassification of the sulfate- and nitrate-reducing bacte-rium Desulfovibrio vulgaris subsp. oxamicus as Desulfovibriooxamicus sp. nov., comb. nov. Int J Syst Evol Microbiol56:1495–1499. doi:10.1099/ijs.0.64074-0

43. Kuever J, Rainey FA, Widdel F (2005) Family II. Syntrophaceaefam. nov. In: Brenner DJ, Krieg NR, Staley JT, Garrity GM (eds)Bergey's Manual of Systematic Bacteriology, 2nd edn. Springer,New York, p 1033

44. Bahr M, Crump BC, Klepac-Ceraj V, Teske A, Sogin ML, HobbieJE (2005) Molecular characterization of sulfate-reducing bacteriain a New England salt marsh. Environ Microbiol 7:1175–1185

45. Isaksen MF, Bak F, Jorgensen BB (1994) Thermophilic sulfate-reducing bacteria in cold marine sediment. FEMS Microbiol Ecol14:1–8

46. Campbell BJ, Engel AS, Porter ML, Takai K (2006) The versatileepsilon-Proteobacteria: key players in sulphidic habitats. Nat RevMicrobiol 4:458–468

47. Sievert SM, Scott KA, Klotz MG, Chain PSG, Hauser LJ,Hemp J, Hugler M, Land M, Lapidus A, Larimer FW, LucasS, Malfatti SA, Meyer F, Paulsen IT, Ren Q, Simon J (2008)Genome of the epsilonproteobacterial chemolithoautotrophSulfurimonas denitrificans. Appl Environ Microbiol 74:1145–1156. doi:10.1128/aem.01844-07

48. Larkin JM, Shinabarger DL (1983) Characterization of Thiothrixnivea. Int J Syst Bacteriol 33:841–846

49. Ito T, Sugita K, Yumoto I, Nodasaka Y, Okabe S (2005) Thiovirgasulfuroxydans gen. nov., sp. nov., a chemolithoautotrophic sulfur-oxidizing bacterium isolated from a microaerobic waste-waterbiofilm. Int J Syst Evol Microbiol 55:1059–1064

50. Rittle KA, Drever JI, Colberg PJS (1995) Precipitation of arsenicduring bacterial sulfate reduction. Geomicrobiol J 13:1–11

51. Smedley PL, Kinniburgh DG (2002) A review of the source,behaviour and distribution of arsenic in natural waters. ApplGeochem 17:517–568

Microbial Communities Involved in Biological Ammonium Removal 59