Optimization of Arsenic Removal Water Treatment System through

Carollo Engineers, Sarasota, Florida 34232, United States. Environ. Sci. Technol. , 0, (),. DOI: 10.1021/es302145q@proofing. Copyright © American Che...
0 downloads 8 Views 2MB Size
Download PDF
Article pubs.acs.org/est

Optimization of Arsenic Removal Water Treatment System through Characterization of Terminal Electron Accepting Processes Giridhar Upadhyaya,†,‡ Tara M. Clancy,† Jess Brown,‡ Kim F. Hayes,† and Lutgarde Raskin†,* †

Department of Civil and Environmental Engineering, University of Michigan, Ann Arbor, Michigan 48109, United States Carollo Engineers, Sarasota, Florida 34232, United States

‡

S Supporting Information *

ABSTRACT: Terminal electron accepting process (TEAP) zones developed when a simulated groundwater containing dissolved oxygen (DO), nitrate, arsenate, and sulfate was treated in a fixed-bed bioreactor system consisting of two reactors (reactors A and B) in series. When the reactors were operated with an empty bed contact time (EBCT) of 20 min each, DO-, nitrate-, sulfate-, and arsenate-reducing TEAP zones were located within reactor A. As a consequence, sulfate reduction and subsequent arsenic removal through arsenic sulfide precipitation and/or arsenic adsorption on or coprecipitation with iron sulfides occurred in reactor A. This resulted in the removal of arsenic-laden solids during backwashing of reactor A. To minimize this by shifting the sulfate-reducing zone to reactor B, the EBCT of reactor A was sequentially lowered from 20 min to 15, 10, and 7 min. While 50 mg/L (0.81 mM) nitrate was completely removed at all EBCTs, more than 90% of 300 μg/L (4 μM) arsenic was removed with the total EBCT as low as 27 min. Sulfate- and arsenate-reducing bacteria were identified throughout the system through clone libraries and quantitative PCR targeting the 16S rRNA, dissimilatory (bi)sulfite reductase (dsrAB), and dissimilatory arsenate reductase (arrA) genes. Results of reverse transcriptase (RT) qPCR of partial dsrAB (i.e., dsrA) and arrA transcripts corresponded with system performance. The RT qPCR results indicated colocation of sulfate- and arsenate-reducing activities, in the presence of iron(II), suggesting their importance in arsenic removal.

■

INTRODUCTION A fixed-bed bioreactor contains a stationary bed consisting of sand, anthracite, plastic media, or granular activated carbon (GAC), which provides a large surface area for the attachment of microorganisms1,2 as biofilms.3 The formation of biofilms helps minimize washout of desired microorganisms, which is especially important when water treatment relies on slow growing microbes, such as sulfate-reducing bacteria (SRB). Microorganisms use electron acceptors present in the source water, such as dissolved oxygen (DO), nitrate, iron(III), and sulfate, based on their metabolic capabilities. This results in the generation of a sequence of terminal electron accepting process (TEAP) zones along the depth of the filter bed and within the biofilm.4 Consequently, a redox gradient develops along the filter bed, allowing for biotransformation and removal of a variety of organic and inorganic contaminants. The development of TEAP zones and subsequent removal of contaminants is governed by the time contaminants remain in contact with the filter bed in a fixed-bed bioreactor, which is generally expressed as the empty bed contact time (EBCT). The EBCT determines whether there is sufficient time for effective diffusion of contaminants into the biofilm and their subsequent utilization by microorganisms and, therefore, is a critical design and operational parameter of a fixed-bed © 2012 American Chemical Society

bioreactor. The minimum EBCT required to achieve sustained contaminant removal depends on many factors, including biotransformation kinetics, adsorption affinity of the contaminants to the filter material, and the target effluent concentrations of the contaminants. Increasing the EBCT generally leads to better performance by allowing more time for adsorption, biologically mediated precipitation of produced solids, and biodegradation.5−8 However, reactor size and cost increase with increasing EBCTs. Therefore, it is important to minimize the EBCT without compromising treatment performance. Associated with the optimization of EBCT is the need to establish effective treatment zones, especially when multiple TEAP zones are to be utilized for the treatment of cocontaminants within the same system. We previously demonstrated the simultaneous removal of nitrate and arsenic from groundwater using a fixed-bed bioreactor system consisting of two biologically active carbon (BAC) bioreactors (reactors A and B) in series.4 While nitrate was removed Received: Revised: Accepted: Published: 11702

May 30, 2012 September 26, 2012 October 2, 2012 October 2, 2012 dx.doi.org/10.1021/es302145q | Environ. Sci. Technol. 2012, 46, 11702−11709

Environmental Science & Technology

Article

were collected from the sampling ports along the depth of the reactors on days 301, 337, and 387, immediately before switching to a new EBCT. The samples were filtered through 0.22 μm filters, loaded in 2 mL glass vials and sealed immediately, and stored at 4 °C until analyzed, typically within 48 h. Samples for total arsenic and total iron were acidified to a final concentration of 0.02 N HCl before storage. Concentrations of DO (0.1 mg/L (3 μM)), acetate (0.2 mg/L (17 μM) as C), chloride (0.2 mg/L (6 μM)), nitrate (0.2 mg/L (3 μM)), sulfate (0.2 mg/L (2 μM)), total arsenic (2 μg/L (0.03 μM)), and total iron (0.1 mg/L (1.8 μM)) were determined as previously described4 (detection limits for each of the analytes are given in parentheses). Biomass Collection and Nucleic Acids Extraction. Biomass samples were collected on days 125, 227, 301, 337, and 387 by removing approximately 2 g of BAC particles from each of the sampling ports along the depth of the reactors. These samples were flash-frozen and stored at −80 °C. Genomic DNA was extracted from the samples using a phenolchloroform extraction protocol18 and quantified using a NanoDrop ND1000 (NanoDrop Technology, Wilmington, DE), and stored at −20 °C. Total RNA was isolated from samples collected on days 301, 337, and 387 following a low-pH, hot-phenol-chloroform extraction protocol.19 Contaminating DNA was digested using RNase-free Turbo DNase (Ambion Inc., Austin, TX) at 37 °C for 30 min. The purified RNA was quantified using a NanoDrop ND-1000 and stored at −80 °C. The effectiveness of DNase treatment was evaluated by PCR targeting the 16S rRNA gene with primers 8F and 1492R (below). PCR Amplification and Construction of Clone Libraries. Clone libraries of the 16S rRNA, dsrAB, and arrA genes were constructed from biomass samples collected on days 125, 227, and 301, respectively. The target genes were PCR amplified with a Mastercycler thermocycler (Eppendorf International, Hamburg, Germany) using DNA extracted separately from the profile biomass samples. Approximately 1.5 Kbp of the bacterial 16S rRNA gene was amplified using primers 8F (5′-AGAGTTTGATCCTGGCTCAG-3′) and 1492R (5′-GGYTACCTTGTTACGACTT-3′) as previously described20 except that Ex Taq polymerase (Takara Bio Inc., Shiga, Japan) was used. Using primer set DSR1Fmix and DSR4Rmix21 in a modified PCR reaction mix, approximately 1.9 Kbp of the dsrAB genes were amplified. Each 25 μL PCR reaction mix included 500 nM forward and reverse primers, 3 mM MgCl2, 0.4 μg/μL bovine serum albumin (Invitrogen Inc., Carlsbad, CA), 12.5 μL of HotStarTaq Mastermix (QIAGEN Inc., Valencia, CA), and 10 ng DNA template. PCR thermal conditions were adopted from Kjeldsen et al.21 By modifying the PCR reaction mixtures, approximately 628 bp of the arrA genes was amplified through nested PCR.17 Two separate initial PCR amplifications were performed; the first initial PCR used primers AS1F (5′-CGAAGTTCGTCCCGATHACNTGG-3′) and AS1R (5′-GGGGTGCGGTCYTTNARYTC-3′), while primers AS1F and AS2R (5′-ATANGCCCARTGNCCYTGNG-3′)17 were used for the second initial PCR. Each 25 μL initial PCR reaction mix included 400 nM forward and reverse primers, 1 mM MgCl2, 12.5 μL of HotStarTaq Mastermix (QIAGEN Inc., Valencia, CA), and 25 ng DNA template. PCR products (1 μL) from the initial PCR reactions were used as the template for the nested PCR with primers AS2F (5′-GTCCCNATBASNTGGGAN-

through denitrification, sulfate reduction generated sulfides, which resulted in the generation of iron sulfides in the presence of iron(II). Arsenic removal relied on precipitation of arsenic sulfides or adsorption or coprecipitation of reduced arsenic with iron sulfides.4 In this BAC system, reactors A and B required backwashing every 48 h and 3−4 months, respectively, to remove excess biomass.4,9 Given the need for frequent backwashing of reactor A, which contained high biomass generating and thermodynamically preferable TEAP zones (DO- and nitrate-reducing zones),10,11 large amounts of arsenic-containing solids were generated.4,9 The TEAP zones in which arsenic removal took place contained sulfate- and arsenate-reducing microbes,4 which produce comparatively low amounts of biomass.10−13 Therefore, it would be preferable to separate the high and low biomass generating TEAP zones. This idea was evaluated by optimizing the EBCT of the first bioreactor of the system consisting of two bioreactors in series. Optimizing the EBCT of the first bioreactor would not only help achieve the desired separation of TEAP zones, but would also minimize the frequency of handling hazardous solid waste. The objective of the current study was to assess the impact of varying the EBCT on reactor performance with the overall goal of confining the production of arsenic-laden solids primarily to reactor B without compromising treatment performance. Therefore, the abundance of sulfate- and arsenate-reducing bacteria and their activities were monitored at different locations in the system by quantitative PCR (qPCR) and reverse transcriptase (RT) qPCR targeting partial dissimilatory (bi)sulfite reductase (dsrAB)14,15 and the dissimilatory arsenate reductase (arrA)16,17 genes.

■

MATERIALS AND METHODS Reactor System and Operation. Two glass columns of 4.9 cm inner diameter and 26 cm height (Supporting Information (SI) Figure S1) containing BAC were operated in series.4 The BAC had been obtained from pilot- and benchscale perchlorate and nitrate removing bioreactors.29 The influent consisted of a synthetic groundwater with 300 μg/L (4 μM) arsenate (As(V)), 50 mg/L (0.81 mM) nitrate, and 22 mg/L (0.23 mM) sulfate along with other constituents (SI Table S1). Glacial acetic acid (final concentration of 35 mg/L (2.91 mM) as carbon) was injected into the influent through a syringe pump (Harvard Apparatus, Holliston, MA) along with 2 mg/L (35.8 μM) Fe(II). An additional 4 mg/L (71.6 μM) Fe(II) was added directly into the effluent of reactor A before entering reactor B starting on day 353. DO in the influent was maintained at less than 1 mg/L (31 μM).4 Reactor A was backwashed every 2 days with a mixed flow of N2-purged deionized (DI) water and N2 gas.9 Reactor B required backwashing every 3−4 months.4 Reactor B was backwashed 27 days prior to the start of the current study, but was not backwashed during this study. EBCT Experiment. The two reactors were initially operated with an EBCT of 20 min each, resulting in a total system EBCT of 40 min. To evaluate the possibility of shifting the sulfatereducing zone and arsenic removal into reactor B, the EBCT of reactor A was gradually lowered from 20 min to 15, 10, and 7 min on days 301, 337, and 387, respectively, by removing the appropriate amount of BAC. The EBCT of reactor B was kept constant at 20 min throughout the study. Liquid Sample Collection and Chemical Analyses. Influent (Inf), reactor A effluent (EA), and final effluent (EB) samples were obtained every day. In addition, liquid samples 11703

dx.doi.org/10.1021/es302145q | Environ. Sci. Technol. 2012, 46, 11702−11709

Environmental Science & Technology

Article

Figure 1. (A) Nitrate, (B) sulfate, and (C) total arsenic removed in reactor A and across the system versus time of operation. Influent concentrations of nitrate, sulfate, and arsenic are also shown. The empty bed contact time (EBCT) of reactor A was changed on days 301, 337, and 387 (marked by vertical lines). The EBCT of reactor B was maintained at 20 min throughout the experiment.

dissimilatory arsenate-reducing bacteria (DARB) were constructed using 214 and 219 amino acid positions, respectively, after aligning the sequences using ClustalW2.26 Quantitative PCR. The abundance of the dsrAB genes was assessed using qPCR with primers DSR1F+ (5′-ACSCACTGGAAGCACGCCGG-3′) and DSR-R (5′GTGGMRCCGTGCAKRTTGG-3′)27 in a modified PCR reaction mix targeting a fragment of the dsrAB gene (i.e., dsrA). Each 25 μL PCR reaction mix contained 12.5 μL 2× QuantiTect SYBR Green PCR Master Mix (QIAGEN Inc., Valencia, CA), 1 mM MgCl2, 400 nM forward and reverse primers, and DNA templates of known concentrations of standards or 50 ng DNA extracted from biomass samples. Purified E. coli plasmid DNA containing the dsrAB gene of Desulfovibrio vulgaris was used to generate a calibration curve. The gene copies determined by qPCR were divided by the amount of DNA (50 ng) used in each reaction. Two primer sets, ArrAF1 (5′-CCCGCTATCATCCAATCG3′) and ArrAR1 (5′-GGTCAGGAGCACATGAG-3′), and ArrAF2 (5′-CATCGCTTCTCGCTGTG-3′) and ArrAR2 (5′GAGGTAGTTGCAGTTTCG-3′), specific to the sequences related to clusters DARB-I and DARB-II (SI Figure S9) were designed. Primers ArrAF1 and ArrAR1 (target fragment 187 bp), and ArrAF2 and ArrAR2 (target fragment 201 bp) were used to determine the abundance of the arrA genes for clusters DARB-I and DARB-II, respectively. Each 20 μL PCR reaction mix contained 10 μL QuantiTect SYBR Green PCR Master

RARGCNMT-3′) and AS1R. The PCR reaction mixture remained the same as in the initial PCRs, except that 600 nM primers were used. PCR products of each target gene were purified using a MinElute Gel Extraction Kit (QIAGEN Inc., Valencia, CA), pooled together, and cloned into One Shot TOPO10 Chemically Competent Escherichia coli cells using the pCR4TOPO cloning kit (Invitrogen Inc., Carlsbad, CA). Glycerol stocks of randomly picked colonies were prepared in 96-well microplates for Sanger sequencing (Genomic Center, Washington University, Saint Louis, MO). The clone library of the 16S rRNA gene consisted of four 96-well plates, whereas one 96-well plate was used for each of the dsrAB and arrA gene clone libraries. Phylogenetic Analyses. The 16S rRNA gene sequences were processed using Mothur.22 Sequences identified as chimeras by Mothur and verified by Mallard23 were excluded from further analyses. The aligned sequences were clustered into operational taxonomic units (OTUs) based on a 97% sequence similarity.21 A phylogenetic tree of the identified Deltaproteobacteria-like sequences was constructed using 545 nucleotide positions in the 16S rRNA gene sequences starting from the 8F primer end with the software program MEGA4.24 The dsrAB and arrA gene sequences were analyzed and edited using BioEdit.25 After removing short and ambiguous dsrAB and arrA sequences, translated amino acid sequences were generated using MEGA4.24 Phylogenetic trees of SRB and 11704

dx.doi.org/10.1021/es302145q | Environ. Sci. Technol. 2012, 46, 11702−11709

Environmental Science & Technology

Article

reactor A consistently removed 10.6 ± 2.9 mg/L (0.11 ± 0.03 mM) sulfate and 251.8 ± 44.6 μg/L (3.4 ± 0.6 μM) arsenic, while sulfate and arsenic removals across the system were 18.5 ± 2.2 mg/L (0.19 ± 0.02 mM) and 284.6 ± 15.4 μg/L (3.8 ± 0.2 μM), respectively (Figures 1B and 1C). To avoid the frequent generation of arsenic-containing solids through backwashing reactor A every 48 h, the EBCT of reactor A was sequentially lowered to confine sulfate reduction, and thus arsenic removal, primarily to reactor B, which required less frequent backwashing. After lowering the EBCT of reactor A to 15 min on day 301 (total EBCT 35 min), complete nitrate removal was still achieved in reactor A (Figure 1A). As desired, sulfate reduction was shifted to reactor B with only 4.5 ± 2.2 mg/L (0.05 ± 0.02 mM) removal in reactor A (Figure 1B). Some of the arsenic removal also shifted to reactor B, but 130.9 ± 79.5 μg/L (1.75 ± 1.06 μM) arsenic was still removed in reactor A (Figure 1C). Additional sulfate reduction in reactor B resulted in a total removal of 16.2 ± 1.8 mg/L (0.17 ± 0.02 mM) sulfate and 250.7 ± 20.5 μg/L (3.3 ± 0.3 μM) arsenic across the system. These average values were calculated excluding days 315−318 and days 323−327 when the sulfate concentration in the influent was 0 and 14.2 ± 0.32 mg/L (0.15 mM), respectively, due to operational errors. The absence of sulfate also impacted arsenic removal during days 315−318 (Figure 1C). Further lowering the EBCT in reactor A to 10 min (total EBCT 30 min) on day 337 resulted in a further decrease in sulfate removal in reactor A. During days 338−387, while complete denitrification was still achieved in reactor A (Figure 1A), sulfate reduction and arsenic removal in reactor A averaged 2.7 ± 1.4 mg/L (0.03 ± 0.01 mM) and 112.4 ± 33.8 μg/L (1.5 ± 0.5 μM), respectively. Total sulfate and arsenic removals across the system were 17.9 ± 4.0 mg/L (0.19 ± 0.04 mM) and 252.0 ± 17.6 μg/L (3.4 ± 0.2 μM), respectively (Figures 1B and 1C). Overall reactor performance was comparable to that achieved with the total EBCT of 30 min even after the EBCT of reactor A was lowered to 7 min (total EBCT 27 min) on day 387 (Figure 1). To evaluate the impact of EBCT on TEAP zones, concentrations were monitored along the depth of the reactors. The cumulative removal profiles of nitrate, sulfate, and arsenic along the depth of the reactors illustrate how lowering the EBCT of reactor A shifted the TEAP zones in both reactors (Figure 2). A progressive shifting of nitrate- and sulfatereducing, and arsenic-removal zones toward the direction of flow was observed with the decrease in EBCT of reactor A. For example, nitrate was completely removed in the filter bed above port A7 in reactor A when the EBCT was 40 and 35 min. However, only 24.7 ± 0.1 mg/L (0.40 mM) nitrate was removed in the bed prior to port A7 when the EBCT was 30 min (Figure 2A). Similarly, sulfate reduction in reactor A declined from 56% to 17% of the total sulfate reduction when the total EBCT was lowered from 40 to 35 min (Figure 2B). Total sulfate reduction and arsenic removal across the system were lower on day 337 (EBCT 35 min), likely due to the low abundance and activity of SRB and DARB (discussed below). Arsenic removal relied on the generation of sulfides as suggested by the close match between the trends of sulfate-, arsenic-, and iron-removal (SI Figures S2, S3, and S4). Through X-ray absorption spectroscopy, we previously reported close association of arsenic with iron and sulfide, suggesting the precipitation of arsenic sulfides or adsorption of arsenic on or coprecipitation with iron sulfides as the possible mechanisms

Mix (QIAGEN Inc., Valencia, CA), 1 mM MgCl2, 300 nM forward and reverse primers, and DNA templates of known concentrations of standards or 50 ng DNA extracted from biomass samples. The amplification cycles included initial denaturation at 95 °C for 15 min followed by 35 cycles of denaturation at 95 °C for 30 s, annealing at 52 °C (cluster DARB-I) or 56 °C (cluster DARB-II) for 30 s, and extension at 72 °C for 1 min. Purified E. coli plasmids containing an approximately 628 bp fragment of the arrA genes from clone 62 (cluster DARB-I) and clone 34 (cluster DARB-II) were used to generate calibration curves (103−108 copies). The gene copies determined by qPCR were divided by the amount of DNA (50 ng) used. Reverse Transcriptase Quantitative PCR. Partial dsrAB gene transcripts were reverse transcribed from purified RNA extracts of biomass profile samples collected on days 301, 337, and 387 using a 2-Step RT-qPCR kit (ABgene House, UK). Each 20 μL RT reaction contained 1x cDNA synthesis buffer, 500 nM dNTP mix, 800 nM DSR-R primer, 1 μL RT enhancer, 1 μL Verso enzyme mix, and 5 μL RNA template. The reaction mixtures were incubated at 42 °C for 30 min and Verso enzyme was inactivated by heating at 95 °C for 5 min. A calibration curve was generated from known amounts of cDNA copies of the dsrAB gene prepared by modifying the protocol described by Smith et al.28 Briefly, partial dsrAB gene was amplified from DNA extract of Desulfovibrio vulgaris using primers DSR1F+ and DSR-R.27 The PCR product was purified using QIAquick PCR purification kit (QIAGEN Inc., Valencia, CA) and in vitro transcribed using MEGAscript T7 kit (Ambion Inc., Austin, TX). Contaminating DNA was removed through a treatment with Turbo DNase (Ambion Inc., Austin, TX). The transcripts were precipitated with ethanol and cDNA was synthesized using the 2-step RT-qPCR kit (Abgene House, UK). A standard series of known amplicon/μL was generated from the cDNA. The number of transcript copies determined by RT-qPCR was divided by the amount of total RNA (200 ng) utilized in each reaction. Similarly, partial arrA gene transcripts related to clusters DARB-I and DARB-II (SI Figure S9) were reverse transcribed from the RNA extracts of the biomass profile samples collected on day 301 as described above, except that reverse primers ArrAR1 or ArrAR2, respectively, replaced primer DSR-R. Due to an accidental loss of the RNA extracts from days 337 and 387, the activity of the arrA genes could not be evaluated for those days. Two standard series of known cDNA copies of the arrA genes were prepared from clones 62 (cluster DARB-I) and 34 (cluster DARB-II) using primer sets ArrAF1 and ArrAR1, and ArrAF2 and ArrAR2, respectively. qPCRs were performed using primer sets DSR1F+ and DSRR, ArrAF1 and ArrAF2, and ArrAF2 and ArrAR2, respectively, to determine the cDNA copies of the dsrA and arrA genes. DNA Sequence Accession Numbers. The 16S rRNA, dsrAB, and arrA gene sequences obtained in this study are available in Genbank under accession numbers JF808668JF808721, JF826136-JF826216, and JF827098-JF827148, respectively.

■

RESULTS AND DISCUSSION Reactor Performance. Initially, when each reactor was operated with an EBCT of 20 min, complete nitrate removal took place in reactor A, lowering the influent nitrate level of 47.1 ± 2.5 mg/L (0.76 ± 0.04 mM) to less than the detection limit of 0.2 mg/L (3 μM) in EA (Figure 1A). In addition, 11705

dx.doi.org/10.1021/es302145q | Environ. Sci. Technol. 2012, 46, 11702−11709

Environmental Science & Technology

Article

(i.e., Geobacter) exhibited a relative abundance of 12% and 7%, respectively. Geobacter-like clones were identified to be close relatives of Geobacter uraniireducens Rf4 (SI Table S3, and Figures S7, and S9), an arsenate-reducing species (discussed below). More detailed phylogenetic analyses of the Deltaproteobacteria-like 16S rRNA gene sequences revealed three clusters (SI Figure S7). Cluster I and II were closely related to the Desulfatirhabdium−Desulfococcus−Desulfosarcina assemblage and the Geobacter genus, respectively, and Cluster III represented members of the Desulfovibrio genus. The dsrAB gene clone library results were in agreement with 16S rRNA gene-based analyses; Cluster SRB-I (75% of the translated amino acid sequences) contained members closely related to the Desulfatirhabdium-Desulfococcus-Desulfosarcina assemblage, while Desulfovibrio-like clones were also present (SI Figure S8; Cluster SRB-II). The use of acetate as the electron donor explains the dominance of Cluster SRB-I since many members of the Desulfatirhabdium−Desulfococcus-Desulfosarcina assemblage can oxidize a variety of organic compounds, including acetate, completely to CO2.30,31 Desulfovibrio cannot utilize acetate as the electron donor for sulfate reduction, but can sustain growth through fermentation and the utilization of H2 (a fermentation product).30,31 Given that fermenting bacteria were relatively abundant (e.g., Cloacibacterium and Treponema, SI Table S3), suggesting that microbial products were used as their substrates, we speculate that Desulfovibriolike SRB were present due to their ability to ferment or utilize H2. Of the 58 gene sequences retrieved from the arrA gene clone library, 51 were translated into amino acid sequences for phylogenetic analysis, which revealed two distinct clusters (SI Figure S9). Cluster DARB-I, closely related to G. uraniireducens Rf4, dominated the DARB (37 sequences). This is in agreement with the 16S rRNA gene clone library (SI Table S1) and previous reports on the presence of Geobacter-related bacteria in arsenic-contaminated sites.32,33 Since G. uraniireducens shows sustained growth on arsenate32 and has putative genes for arsenate respiration in its genome, the predominance of G. uraniireducens-like DARB in the current system is not surprising. Cluster DARB-II contained 14 sequences distantly related (65−67% amino acid sequence similarity) to Alkalilimnicola ehrlichii strain MLHE-1 (SI Figure S9). This may be explained by the fact that one of the two homologues of putative respiratory arsenate reductase identified in the genome of A. eherlichii exhibits both arsenate reductase and arsenite oxidase activities.34 However, considering the comparatively low sequence identity of the clones in cluster DARB-II with A. ehrlichii, the possibility of the presence of novel uncultured arsenate respiring bacteria cannot be ruled out. Abundance and Activity of SRB and DARB. Biomass profile samples were collected to evaluate the impact of the changes in EBCT on sulfate- and arsenate-reducing populations. qPCR and RT-qPCR data require normalization to derive meaningful comparisons among samples. In the absence of a universally accepted method of normalization, the total mass of DNA and RNA, respectively, applied in the PCR reactions were used for normalization, even though we recognize the limitations with this approach. The relative abundance of dsrA genes indicated that SRB were roughly equally distributed throughout the BAC filter beds for a given EBCT, but their relative abundance levels varied for the three EBCTs evaluated (Figure 3). Sulfate reducing activity, expressed as dsrA

Figure 2. Cumulative removal of (A) nitrate, (B) sulfate, and (C) arsenic along the depth of the reactor beds on days 301, 337, and 387. A5−A8 and B1−B4 represent the sampling ports along the depth of reactors A and B, respectively. EA and EB represent concentrations in the effluents from reactors A and B, respectively. The mean values of three replicates are reported with the error bars representing one standard deviation. Pictures in insets show the filter bed depths at the empty bed contact times (EBCT) evaluated.

for arsenic removal.4 The poorer arsenic removal observed during days 315−318 due to the accidental absence of sulfate in the influent (Figure 1C) also suggests that the generation of fresh sulfides is important for arsenic removal. Despite the matching trends of sulfate reduction and arsenic removal profiles, the total arsenic removal did not track the shifts in sulfate reducing TEAP zones as almost 50% of arsenic was removed in reactor A regardless of the EBCT. This inability to shift arsenic removal completely to reactor B may be due to the occurrence of sufficient sulfate reduction in reactor A to facilitate arsenic removal (SI Figure S3). Indeed, even the low amount of sulfate reduction observed on day 337, that is, 2.2 ± 1.7 mg/L (23 ± 18 μM) provides excess sulfide relative to the total arsenic of 300 μg/L (4 μM) present in the influent. Microbial Community Structure. The 16S rRNA gene based clone library analysis (SI Figures S5 and S7) identified previously described nitrate-, sulfate-, iron-, and arsenatereducing bacteria (SI Table S3). Betaproteobacteria (36% of the clones) and Deltaproteobacteria (19%) were dominant (SI Figure S5). Zoogloea (13%), Azospira (12%), and Dechloromonas (6%) were the major genera within the Betaproteobacteria (SI Table S3). These clones likely represented the majority of nitrate-reducing bacteria in the system.29 Within the Deltaproteobacteria, clones closely related to SRB (i.e., Desulfatirhabdium and Desulfovibrio) and iron-reducing bacteria 11706

dx.doi.org/10.1021/es302145q | Environ. Sci. Technol. 2012, 46, 11702−11709

Environmental Science & Technology

Article

Figure 4. Abundance of two distinct phylogenetic clusters of dissimilatory arsenate reducing bacteria along the depth of the reactor beds on day 301 (A), day 337 (B), and day 387 (C). Abundance is expressed as the copies of arrA gene per ng of genomic DNA. A5−A8 and B1−B4 refer to the sampling ports along the depth of the reactor beds. The mean values of three replicates are presented with error bars representing one standard deviation.

Figure 3. Abundance and activity of sulfate reducing bacteria (SRB), and cumulative sulfate removal along the depth of the filter beds on day 301 (A), day 337 (B), and day 387 (C). Abundance is expressed as the dsrA gene copies per ng of genomic DNA, whereas SRB activity is expressed as the copies of dsrA transcripts per ng of total RNA. A5− A8 and B1−B4 refer to the sampling ports along the depth of the respective reactor beds. The mean values of three replicates are presented with error bars representing one standard deviation.

Even though the greatest abundance of DARB related to clusters I and II was observed in ports A6 and A5, respectively, on day 301 (Figure 4), arsenate reductase activity was not detected in these ports, but was observed in port A7 and the rest of the system where most of the arsenic removal took place (Figures 2 and 5). Since considerable sulfate reduction and iron removal occurred in ports A5 and A6 on day 301 (Figure 2 and SI Figure S4), the observed loss of arsenic at these ports despite the lack of arsenate reductase activity might have resulted from coprecipitation with or adsorption on iron sulfides. Previously

transcript numbers normalized to total mass of RNA, attained a maximum value in the middle of the system regardless of the EBCT (Figure 3) and correlated well with the cumulative sulfate removal profile. Specifically, the maximum dsrA activity was observed in the region with the steepest slope of the cumulative sulfate removal plot. In other words, the maximum SRB activity was observed at ports A8, B1, and B2 for EBCTs of 40 min, 35 min, and 30 min, respectively, suggesting that lowering the EBCT sequentially shifted the maximum SRB activity from reactor A to reactor B. Both dsrA abundance and activity levels were lowest on day 337. Although it is not clear why this occurred, this observation is consistent with the much lower cumulative sulfate removal detected on this day, as compared to the other two days for which profile data are available (Figure 3). DARB belonging to arrA cluster DARB-II were more abundant for all EBCTs evaluated (Figure 4). Similar to the abundance and activity levels of dsrA, arrA abundance was lower on day 337. Though a consistent trend of the abundance of arrA was not observed for the EBCTs evaluated, better arsenic removal was observed when arrA levels were higher toward the beginning of the system. For example, DARB related to both clusters were more abundant in ports A5 and A6 on day 301 when the total arsenic removal was 290 ± 12 μg/L (3.9 ± 0.2 μM), whereas their abundance in ports A7 and A8 was much lower on day 337 when only 255 ± 9 μg/L (3.4 ± 0.1 μM) was removed.

Figure 5. Activity of two distinct phylogenetic clusters of dissimilatory arsenate reducing bacteria along the depth of reactors A and B on day 301. Activity is presented as number of arrA transcripts normalized to total RNA. The mean values of three replicates are presented with error bars representing one standard deviation from the mean. ND represents “not detected”. 11707

dx.doi.org/10.1021/es302145q | Environ. Sci. Technol. 2012, 46, 11702−11709

Environmental Science & Technology

■

described DARB are not obligate arsenate respirers, except for strain MLMS-1,35 and can use other electron acceptors such as DO, nitrate, Fe(III), and sulfate.36 Therefore, it is possible that arrA genes detected in the nitrate-reducing TEAP zone (Figures 2 and 4), belonged to nitrate-reducing bacteria that can also utilize arsenate as an electron acceptor. In contrast to the DNA abundance data, the relative abundance levels of DARB-I and DARB-II arrA gene transcripts were similar in ports A7 and A8, but DARB-I exhibited greater activity than DARB-II in reactor B. While the reasons for this apparent contradiction in gene abundance and activity levels of the two clusters are not clear, DARB-I appears to be an important arsenate-reducing group in the system, despite its low abundance. In conclusion, this microbial activity-based EBCT analysis shows that changes in EBCT impact the activity of denitrifying bacteria, SRB, and DARB along the depth of BAC bioreactors. The presence of SRB and DARB in relatively high numbers and the colocation of sulfate- and arsenate-reducing activities in the presence of iron(II) were key for arsenic removal. Substantial arsenic removal was achieved down to a total EBCT of 27 min and most of the sulfate reduction could be shifted to reactor B by decreasing the EBCT of reactor A. Although it is conceivable that an even lower EBCT might shift the sulfate reducing zone entirely to reactor B, it is important to confine nitrate removal to reactor A. Otherwise, reactor B would require more frequent backwashing due to the higher biomass yield associated with denitrification and the benefit of minimizing the amount of arsenic-containing sludge generated may not be realized. However, a scaled-up evaluation of this process must consider the trade-off between total contact time, which affects reactor volume (i.e., capital costs), and arsenic-laden sludge generation, which impacts sludge handling and disposal costs.

■

ASSOCIATED CONTENT

The Supporting Information provides a schematic of the bioreactor system and the composition of the synthetic groundwater. Additional microbial data are also presented along with information on the primers designed in this study. This material is available free of charge via the Internet at http://pubs.acs.org.

AUTHOR INFORMATION

Corresponding Author

*Phone +1-(734)-647-6920; e-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

REFERENCES

(1) Weber, J.; Walter; Pirbazari, M.; Melson, G. Biological growth on activated carbon: An investigation by scanning electron microscopy. Environ. Sci. Technol. 1978, 12, 817−819. (2) Wilcox, D. P.; Chang, E.; Dickson, K. L.; Johansson, K. R. Microbial growth associated with granular activated carbon in a pilot water treatment facility. Appl. Environ. Microbiol. 1983, 46, 406−416. (3) Choi, Y. C.; Li, X.; Raskin, L.; Morgenroth, E. Effect of backwashing on perchlorate removal in fixed bed biofilm reactors. Water Res. 2007, 41, 1949−1959. (4) Upadhyaya, G.; Jackson, J.; Clancy, T. M.; Hyun, S. P.; Brown, J.; Hayes, K. F.; Raskin, L. Simultaneous removal of nitrate and arsenic from drinking water sources utilizing a fixed-bed bioreactor system. Water Res. 2010, 44, 4958−4969. (5) Brown, J. C.; Snoeyink, V. L.; Raskin, L.; Lin, R. The sensitivity of fixed-bed biological perchlorate removal to changes in operating conditions and water quality characteristics. Water Res. 2003, 37, 206− 214. (6) Liang, C.; Chiang, P.-C. Quantification elucidation of the effect of EBCT on adsorption and biodegradation of biological activated carbon filters. J. Chin. Inst. Chem. Eng. 2004, 35, 203−211. (7) Rhim, J. A. Characteristics of adsorption and biodegradation of dissolved organic carbon in biological activated carbon pilot plant. Korean J. Chem. Eng. 2006, 23, 38−42. (8) Wu, H.; Xie, Y. F. Effects of EBCT and water temperature on HAA removal using BAC. J. Am. Water Works Assoc. 2005, 97, 94− 101. (9) Upadhyaya, G.; Clancy, T. M.; Snyder, K. V.; Brown, J.; Hayes, K. F.; Raskin, L. Effect of air-assisted backwashing on the performance of an anaerobic fixed-bed bioreactor that simultaneously removes nitrate and arsenic from drinking water sources. Water Res. 2012, 46, 1309− 1317. (10) Lovley, D. R.; Phillips, E. J. Novel processes for anaerobic sulfate production from elemental sulfur by sulfate-reducing bacteria. Appl. Environ. Microbiol. 1994, 60, 2394−2399. (11) Rikken, G.; Kroon, A.; Van Ginkel, C. Transformation of (per) chlorate into chloride by a newly isolated bacterium: Reduction and dismutation. Appl. Microbiol. Biotechnol. 1996, 45, 420−426. (12) Cuthbertson, J.; Schumacher, M. Full scale implementation of sulfate enhanced biodegradation to remediate petroleum impacted groundwater. Proc. Annu. Int. Conf. Soils, Sed., Water Ecol. 2009, 14, 181−189. (13) Macy, J. M.; Nunan, K.; Hagen, K. D.; Dixon, D. R.; Harbour, P. J.; Cahill, M.; Sly, L. I. Chrysiogenes arsenatis gen. nov., sp. nov., a new arsenate-respiring bacterium isolated from gold mine wastewater. Int. J. Syst. Bacteriol. 1996, 46, 1153−1157. (14) Stahl, D. A.; Fishbain, S.; Klein, M.; Baker, B. J.; Wagner, M. Origins and diversification of sulfate-respiring microorganisms. Antonie van Leeuwenhoek 2002, 81, 189−195. (15) Wagner, M.; Loy, A.; Klein, M.; Lee, N.; Ramsing, N. B.; Stahl, D. A.; Friedrich, M. W. Functional marker genes for identification of sulfate-reducing prokaryotes. Methods Enzymol. 2005, 397, 469−489. (16) Malasarn, D.; Saltikov, C. W.; Campbell, K. M.; Santini, J. M.; Hering, J. G.; Newman, D. K. arrA is a reliable marker for As(V) respiration. Science 2004, 306, 455. (17) Song, B.; Chyun, E.; Jaffé, P. R.; Ward, B. B. Molecular methods to detect and monitor dissimilatory arsenate-respiring bacteria (DARB) in sediments. FEMS Microbiol. Ecol. 2009, 68, 108−117. (18) Pinto, A. J.; Xi, C.; Raskin, L. Bacterial community structure in the drinking water microbiome is governed by filtration processes. Environ. Sci. Technol. 2012, 46, 8851−9. (19) Berry, D.; Holder, D.; Xi, C.; Raskin, L. Comparative transcriptomics of the response of Escherichia coli to the disinfectant monochloramine and to growth conditions inducing monochloramine resistance. Water Res. 2010, 44, 4924−31. (20) Richardson, R. E.; Bhupathiraju, V. K.; Song, D. L.; Goulet, T. A.; Alvarez-Cohen, L. Phylogenetic characterization of microbial communities that reductively dechlorinate TCE based upon a

S Supporting Information *

■

Article

ACKNOWLEDGMENTS

We thank Jeff Jackson for laboratory assistance and helpful discussions. This study was funded by the Water Research Foundation (project 4293) and the U.S. National Science Foundation (project CBET 0967707). G.U. and T.M.C. were partially supported by Graham Environmental Sustainability Institute and Rackham Merit Fellowships from the University of Michigan. TMC was also supported by a U.S. National Science Foundation Graduate Research Fellowship. 11708

dx.doi.org/10.1021/es302145q | Environ. Sci. Technol. 2012, 46, 11702−11709

Environmental Science & Technology

Article

combination of molecular techniques. Environ. Sci. Technol. 2002, 36, 2652−62. (21) Kjeldsen, K. U.; Loy, A.; Jakobsen, T. F.; Thomsen, T. R.; Wagner, M.; Ingvorsen, K. Diversity of sulfate-reducing bacteria from an extreme hypersaline sediment, Great Salt Lake (Utah). FEMS Microbiol. Ecol. 2007, 60, 287−298. (22) Schloss, P. D.; Westcott, S. L.; Ryabin, T.; Hall, J. R.; Hartmann, M.; Hollister, E. B.; Lesniewski, R. a; Oakley, B. B.; Parks, D. H.; Robinson, C. J.; Sahl, J. W.; Stres, B.; Thallinger, G. G.; Van Horn, D. J.; Weber, C. F. Introducing mothur: Open-source, platformindependent, community-supported software for describing and comparing microbial communities. Appl. Environ. Microbiol. 2009, 75, 7537−7541. (23) Ashelford, K. E.; Chuzhanova, N. a; Fry, J. C.; Jones, A. J.; Weightman, A. J. New screening software shows that most recent large 16S rRNA gene clone libraries contain chimeras. Appl. Environ. Microbiol. 2006, 72, 5734−5741. (24) Tamura, K.; Dudley, J.; Nei, M.; Kumar, S. MEGA4: Molecular evolutionary genetics analysis (MEGA) software version 4.0. Mol. Biol. Evol. 2007, 24, 1596−1599. (25) Hall, T. A. BioEdit: A user-friendly biological sequence alignment editor and analysis program for Windows 96/98/NT. 1999, 41, 95−98. (26) Larkin, M. a; Blackshields, G.; Brown, N. P.; Chenna, R.; McGettigan, P. A.; McWilliam, H.; Valentin, F.; Wallace, I. M.; Wilm, A.; Lopez, R.; Thompson, J. D.; Gibson, T. J.; Higgins, D. G. Clustal W and Clustal X version 2.0. Bioinformatics 2007, 23, 2947−2948. (27) Kondo, R.; Nedwell, D.; Purdy, K.; Silva, S. de Q. Detection and enumeration of sulphate-reducing bacteria in estuarine sediments by competitive PCR. Geomicrobiol. J. 2004, 21, 145−157. (28) Smith, C. J.; Nedwell, D. B.; Dong, L. F.; Osborn, A. M. Evaluation of quantitative polymerase chain reaction-based approaches for determining gene copy and gene transcript numbers in environmental samples. Environ. Microbiol. 2006, 8, 804−815. (29) Li, X.; Upadhyaya, G.; Yuen, W.; Brown, J.; Morgenroth, E.; Raskin, L. Changes in the structure and function of microbial communities in drinking water treatment bioreactors upon addition of phosphorus. Appl. Environ. Microbiol. 2010, 76, 7473−7481. (30) Devereux, R.; Delaney, M.; Widdel, F.; Stahl, D. Natural relationships among sulfate-reducing eubacteria. J. Bacteriol. 1989, 171, 6689−6695. (31) Muyzer, G.; Stams, A. J. M. The ecology and biotechnology of sulphate-reducing bacteria. Nat. Rev. Microbiol. 2008, 6, 441−54. (32) Héry, M.; Gault, A. G.; Rowland, H. A. L.; Lear, G.; Polya, D. A.; Lloyd, J. R. Molecular and cultivation-dependent analysis of metalreducing bacteria implicated in arsenic mobilisation in south-east asian aquifers. Appl. Geochem. 2008, 23, 3215−3223. (33) Islam, F. S.; Gault, A. G.; Boothman, C.; Polya, D. A.; Charnock, J. M.; Chatterjee, D.; Lloyd, J. R. Role of metal-reducing bacteria in arsenic release from Bengal delta sediments. Nature 2004, 430, 68−71. (34) Richey, C.; Chovanec, P.; Hoeft, S. E.; Oremland, R. S.; Basu, P.; Stolz, J. F. Respiratory arsenate reductase as a bidirectional enzyme. Biochem. Biophys. Res. Commun. 2009, 382, 298−302. (35) Hoeft, S. E.; Kulp, T. R.; Stolz, J. F.; Hollibaugh, J. T.; Oremland, R. S. Dissimilatory arsenate reduction with sulfide as electron donor: Experiments with Mono Lake water and isolation of strain MLMS-1, a chemoautotrophic arsenate respirer. Appl. Environ. Microbiol. 2004, 70, 2741−2747. (36) Stolz, J. F.; Basu, P.; Santini, J. M.; Oremland, R. S. Arsenic and selenium in microbial metabolism. Annu. Rev. Microbiol. 2006, 60, 107−130.

11709

dx.doi.org/10.1021/es302145q | Environ. Sci. Technol. 2012, 46, 11702−11709