Uranium removal and microbial community in a H2-based membrane biofilm reactor

w a t e r r e s e a r c h 6 4 ( 2 0 1 4 ) 2 5 5 e2 6 4

Available online at www.sciencedirect.com

ScienceDirect journal homepage: www.elsevier.com/locate/watres

Uranium removal and microbial community in a H2-based membrane biofilm reactor Chen Zhou a, Aura Ontiveros-Valencia a,*, Louis Cornette de Saint Cyr a,b, Alexander S. Zevin a, Sara E. Carey a, Rosa Krajmalnik-Brown a, Bruce E. Rittmann a a b

Swette Center for Environmental Biotechnology, Biodesign Institute, Arizona State University, USA Institut Sup'Biotech de Paris, France

article info

abstract

Article history:

We evaluated a hydrogen-based membrane biofilm reactor (MBfR) for its capacity to reduce

Received 18 May 2014

and remove hexavalent uranium [U(VI)] from water. After a startup period that allowed

Received in revised form

slow-growing U(VI) reducers to form biofilms, the MBfR successfully achieved and main-

6 July 2014

tained 94e95% U(VI) removal over 8 months when the U surface loading was 6e11 e
mEq/

Accepted 7 July 2014

m2-day. The MBfR biofilm was capable of self-recovery after a disturbance due to oxygen

Available online 17 July 2014

exposure. Nanocrystalline UO2 aggregates and amorphous U precipitates were associated with vegetative cells and apparently mature spores that accumulated in the biofilm matrix.

Keywords:

Despite inoculation with a concentrated suspension of Desulfovibrio vulgaris, this bacterium

MBfR

was not present in the U(VI)-reducing biofilm. Instead, the most abundant group in the

U removal

biofilm community contained U(VI) reducers in the Rhodocyclaceae family when U(VI) was

Uraninite

the only electron acceptor. When sulfate was present, the community dramatically shifted

Sulfate

to the Clostridiaceae family, which included spores that were potentially involved in U(VI)

Spore

reduction.

Bicarbonate

1.

Introduction

The hydrogen (H2)-based Membrane Biofilm Reactor (MBfR) is a novel technology to reduce oxidized contaminants in water and wastewater treatment. H2 is a promising electron donor, as it is nontoxic, provides low biomass yields, and can be utilized by diverse bacteria that can reduce many electron acceptors (Martin and Nerenberg, 2012). H2 gas is delivered into the reactor by diffusion through the walls of bubbleless gas-transfer membranes, and the outside surface of the membrane accumulates H2-oxidizing autotrophic bacteria as

© 2014 Elsevier Ltd. All rights reserved.

a biofilm (Lee and Rittmann, 2000, 2002). Oxidation of H2 allows reduction of oxidized contaminants in the bulk solution. This demand for H2 to reduce electron acceptors is the driving force to pull H2 through the membrane, and this eliminates practical problems of over- or under-dosing the electron donor, a shortcoming inherent to heterotrophic reactors (Chung et al., 2007a; Rittmann, 2006). Successful removal of a broad spectrum of oxidized contaminants has been reported for the MBfR: e.g., nitrate, nitrite, perchlorate, selenate, arsenate, chromate, tetrachloroethene, trichloroethene, and N-nitrosodimethylamine (Chung et al., 2008a, b; 2007a; Chung and Rittmann, 2007, 2008; Chung

* Corresponding author. Present address: Swette Center for Environmental Biotechnology, Biodesign Institute, Arizona State University, Tempe, AZ 85207-5701, USA. Tel.: þ1 480 217 3742. E-mail address: [email protected] (A. Ontiveros-Valencia). http://dx.doi.org/10.1016/j.watres.2014.07.013 0043-1354/© 2014 Elsevier Ltd. All rights reserved.

256

w a t e r r e s e a r c h 6 4 ( 2 0 1 4 ) 2 5 5 e2 6 4

Fig. 1 e Schematic of the bench-scale MBfR system used to investigate U(VI) reduction. The major components include: ① mixed gas tank with 80% N2 and 20% CO2 to feed the headspace of the medium bottle; ② pure H2 gas tank to feed the fiber bundles of the MBfR; ③ gas pressure regulators; ④ medium bottle with a stir bar; ⑤ feeding pump; ⑥ MBfR configuration (the red arrow indicates the flow direction; ⑦ main bundle with 49 fibers; ⑧ biofilm sampling bundle with 10 coupon fibers; ⑨ recirculation pump; and ⑩ small serum bottle for gas sampling. The black solid arrows indicate the liquid flow, and the black dashed arrows indicate the gas flow. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

et al., 2007b; Karatas‚ et al., 2014; Martin and Nerenberg, 2012; Rittmann et al., 2005; Zhao et al., 2011; Ziv-El et al., 2012). However, the application of MBfR for reducing and removing hexavalent uranium [U(VI)] from contaminated groundwater has not been reported. U contamination is widely present in aquifers and soils at over 120 U.S. Department of Energy (DOE) sites, and it threatens human health due to its strong toxicity (Arfsten et al., 2001; Bleise et al., 2003). In general, U remediation involves immobilizing the soluble U(VI) (uranyl ions; UO2 2þ ) in the groundwater through adsorption, complexation, and reductive precipitation (Baird and Cann, 2005; Bru¨mmer et al., 1986; Gavrilescu et al., 2009; Langmuir, 1978). Microorganisms are able to catalyze U(VI) reduction, and many studies have established that a variety of prokaryotes can enzymatically reduce U(VI) to U(IV) (Wall and Krumholz, 2006). Those prokaryotes include iron reducers (Jeon et al., 2004), sulfate reducers (Lovley and Phillips, 1992a), denitrifiers (Martins et al., 2010c), dehalogenators (Sanford et al., 2007), and sporeforming fermenters (Gao and Francis, 2008). The reductive reactions of the dominant uranyl-carbonate species in groundwater, UO2 ðCO3 Þ3 4
and UO2 ðCO3 Þ2 2
at neutral pH, produce solid-phase uraninite [UO2(s)] (Wan et al., 2005): UO2 ðCO3 Þ2 2
þ 2Hþ þ 2e
/UO2ðSÞ þ 2HCO3


(1)

UO2 ðCO3 Þ3 4
þ 3Hþ þ 2e
/UO2ðSÞ þ 3HCO3


(2)

Previous studies discovered that biogenic uraninite was mainly nanocrystalline, with physical properties quite

different from bulk uraninite (Bargar et al., 2008; Burgos et al., 2008; Lee et al., 2010): smaller particle size and large specific surface area. The surface structure and reactivity of biogenic UO2 also was affected by the presence of surface-associated organic matter (Fernandez-Garcia et al., 2004; Gilbert et al., 2004; Singer et al., 2009). Biological reduction and precipitation have been recognized as the most economical, efficient, and sustainable practice for in situ remediation of U-contaminated groundwater and soil (Wall and Krumholz, 2006). However, complete understanding the biochemistry in the process is still lacking. Furthermore, ex situ remediation has been studied only rarely. In this study, we evaluated a bench-scale MBfR for treating a synthetic groundwater containing abundant U(VI). We examined U reduction and removal, biofilm formation, and solids precipitation. We also studied how environmental conditions, including O2 disturbance, varied U(VI) loading, and an additional electron acceptor, affected U-removal performance. Our results document the mechanism underlying microbial U(VI) removal in a H2-based MBfR and also provide the foundation for optimizing the process in practical application to water treatment in the future.

2.

Material and Methods

2.1.

MBfR setup

A schematic of the MBfR used in this study is shown in Fig. 1. The MBfR system consisted of two glass tubes connected with Norprene tubing, polycarbonate stopcocks, and plastic barbed fittings. The total working volume of the MBfR was 60 mL. One glass tube contained a main bundle of 49 hollow-fiber membranes (nonporous polypropylene fiber, 200 mm OD, 100e110 mm ID, wall thickness 50e55 mm; Teijin, Ltd., Japan), each 25 cm long. Biofilm samples are collected by cutting short lengths of a separate fiber from the “coupon” bundle of 10 fibers, located in the second glass tube. This allowed sample collection without disturbing the main bundle of fibers and without causing a significant change in total biofilm surface area in the reactor. The experimental permeability of the polypropylene fiber was 1.8  10
7 m3 H2$m membrane thickness/m2 hollow fiber surface area$d$bar at standard temperature and pressure (Tang et al., 2012). For the main fiber bundle, both ends were glued into an H2supply manifold. For the “coupon” bundle, the top was glued into an H2-supply manifold, but the bottom was sealed and not fixed to the end of the tube. The MBfR was completely mixed using a high recirculation rate of 150 mL/min achieved with a peristaltic pump (Master Flex, model 7520-40, ColeParmer Instrument Company, U.S.A); therefore, the concentration in the MBfR was equal to the effluent concentration. A peristaltic pump (Rainin Dynamax Peristaltic pump, model RP-1) and PVC tubing (Rainin Silicone pump tube, Yellow Blue) provided an influent feed rate within the range of 0.03e3.00 mL/min. The H2 pressure supplied to the MBfR was 23 psi (1.6 atm) throughout all experiments. A top-sealed 15mL serum bottle was connected after the effluent line in order to create inside a ca. 7-mL headspace for collecting gas samples (Fig. 1, ⑩).

w a t e r r e s e a r c h 6 4 ( 2 0 1 4 ) 2 5 5 e2 6 4

2.2.

Strain, culturing, and biomass preparation

The MBfR was inoculated with a culture of Desulfovibrio vulgaris subsp. vulgaris Postgate and Campbell (ATCC 29579) originally purchased from the American Type Culture Collection (Rockville, Md.). We chose D. vulgaris to test this application because D. vulgaris is known to achieve U(VI) reduction coupled with oxidation of H2 gas as the only electron donor (Elias et al., 2004; Lovley et al., 1993a, b), and it was abundant in the biofilm of an MBfR used for nitrate and sulfate reductions (Ontiveros-Valencia et al., 2013). We followed the same the procedure of biomass preparation as Zhou et al. (2014).

2.3.

Feeding medium

We prepared and stored the feeding medium in a 5-L glass bottle (VWR, Radnor, PA). We maintained an air-tight seal of the bottle with a rubber stopper containing one outlet from the liquid phase and one inlet to the headspace. The basic medium consisted of (in mM) UO2Cl2 0.5, CH3COONa 0.05, NaHCO3 30, KCl 1, NH4Cl 0.2, CaCl2 0.002, MgCl2 0.01, FeCl2 0.01, NiCl2 0.01, and 5 mL trace metal solution. Acetate was added to facilitate the growth of Desulfovibrio species. Desulfovibrio species derive approximately 80% of their carbon from acetate and 20% from carbon dioxide when utilizing H2 as the electron donor (Badziong et al., 1978; Noguera et al., 1998). The acetate concentration was set to be twice that required by stoichiometry for complete reduction of 0.5 mM U(VI) (Noguera et al., 1998). The composition of the trace metal solution was described by Chung et al. (2006). All the components except UO2Cl2, FeCl2, and NiCl2 were added at the beginning. Immediately after being autoclaved for deoxygenation and sterilization, the medium bottle was transferred to an anaerobic glove box. UO2Cl2 powder and FeCl2 þ NiCl2 concentrate solution were added in the glove box after the medium cooled to room temperature. Once outside the glove box, the medium bottle was sparged by 80% N2/20% CO2 premixed gas for 2 h to make it ready for feeding. The 80% N2/20% CO2 gas was supplied subsequently to the headspace at <2 psi. The small positive pressures in the sealed medium bottle maintained the soluble carbonate level and minimized oxygen intrusion from outside. In consequence, influent pH was maintained at 7.5 ± 0.3 until we reduced the bicarbonate concentration in the medium in the last month, and measured influent D.O. levels were minimal (<10 ppb) during the entire operation.

2.4.

Inoculation, startup, and operation

We inoculated the MBfR with 10 mL (~17% of the reactor volume) of a freshly prepared D. vulgaris biomass suspension after filling the reactor with medium. The initial suspending biomass in the bulk liquid was 120e150 mg/L as protein. Note that we did not strictly keep the whole system sterile throughout operation, and thus expected intruding microorganisms other than D. vulgaris. Most denitrifying MBfRs start continuous feeding after short-term batch mode, typically up to 24 h (Chung et al., 2007a; Ontiveros-Valencia et al., 2012; Zhao et al., 2011; Ziv-

257

El and Rittmann, 2009). However, U(VI) reducers, as well as other metal reducers, grow more slowly than denitrifiers and may take longer to form a biofilm on the membranes. Furthermore, our preliminary results using the typical start up method (not shown) failed to produce enough biofilm accumulation to allow noticeable U(VI) reduction. Thus, we adapted the startup procedure used successfully for a TCEreducing MBfR featuring Dehalococcoides, a group of slowgrowing dehalogenators (Ziv-El et al., 2012). In Stage 1, after inoculating the MBfR with the freshly prepared biomass suspension, we maintained it in an extended batch mode for 9 days. This was followed by semi-continuous operation: in each cycle of semicontinuous operation, we fed the reactor at a flow rate of 0.20 mL/min for 24 h, and then we stopped feeding for the next three days. Upon observing stable U removal after four cycles of semi-continuous mode, we initiated continuous operation at a flow rate of 0.04 mL/min. This stage lasted 118 days, and we increased the flow rate to 0.08 mL/min on day 119. In Stage 2, we added into the medium 0.125 mM sulfate (in Na2SO4) as the additional electron acceptor at the same electron equivalents as uranyl (1 mEq e
/L). The flow rate was kept at 0.04 ± 0.003 mL/min through this stage.

2.5.

Sampling and analyses

For routine analyses, we collected liquid samples from the MBfR with 6-mL syringes and filtered part of them immediately through 0.2-mm membrane filters (LC þ PVDF membrane, Whatman Inc., Haverhill, MA). We measure the unfiltered and filtered samples for total and soluble U concentrations, respectively, using inductively coupled plasma optical emission spectrometer (ICP-OES, Thermo iCAP6300) at the wavelength of 385.9 nm. The methods for measuring concentrations of soluble U, sulfate, acetate and other volatile fatty acids are described in Zhou et al. (2014). For pH and dissolved oxygen (D.O.) assays, we took 8-mL liquid samples with 10-mL gas-tight syringes (Hamilton Company, Reno, NV), injected them into sealed anaerobic 10mL serum bottles, and then put the serum bottles into the glove box for analysis. We measured pH values of the unfiltered samples with an Epoxy Semi-Micro Combination pH Electrode (Beckman Coulter BKA57187) and a pH Meter (Beckman Coulter BKA58734). We measured D.O. of the unfiltered samples by the Rhodazine D™ Method test kits (CHEMetrics K-7501, K-7599 and K-7512).

2.6.

Flux calculation

We calculated the U-removal fluxes (e
mEq/m2-day) according to Eqn. 3 J¼

ðS0
SÞQ A

(3)

where S0 and S are the influent and effluent U concentration (e
mEq/L), Q is the influent flow rate to the MBfR system (L/d), and A is the membrane surface area (m2). The maximum H2 delivery capacity (e
mEq/m2-day) was calculated according to Tang et al. (2012).

258

w a t e r r e s e a r c h 6 4 ( 2 0 1 4 ) 2 5 5 e2 6 4

compositions at selected areas were identified using an EDX detector (EDAX Inc., Mahwah, NJ, USA). For XRD analysis, we washed out most of the precipitates on the remaining fibers by vigorously rinsing the bundle in a 200-mL beaker containing freshly-prepared sterile anoxic 15 mM NaHCO3 solution. After overnight precipitation, we removed most supernatant and transferred the remaining solids into 50-mL Falcon tubes. The subsequent rinse and drying steps were same as described in Zhou et al. (2014). After freeze drying, we transferred the grounded solids in 10-mL serum bottles sealed with rubber stoppers in the glove box and stored the bottles in the
20  C freezer. The detailed information of XRD analysis is described in Zhou et al. (2014).

Fig. 2 e Performance parameters in the MBfR during Stage 1. Top: measured flow rate; Middle: measured concentrations of total U in influent (black bars), total U in effluent (open dots), and soluble U in effluent (solid dots); Bottom: calculated U-removal percentage (the filled area), U surface loading (the solid line), and U-removal flux (the dashed line).

2.7.

DNA extraction and microbial community analyses

We sampled the biofilms in Stage 1 and Stage 2. We followed the procedures of biofilm separation and DNA extraction described by Ontiveros-Valencia et al. (2012). The only exception is that, at the end of Stage 2, we stopped the H2 gas supply and placed the whole reactor module into the glove box before opening it up. We amplified and analyzed 16S rRNA genes in all DNA samples for subsequent terminal restriction fragment length polymorphism (T-RFLP) (Liu et al., 1997) and pyrosequencing. The procedures were modified from Sheng et al. (2011) and Ontiveros-Valencia et al. (2013), respectively, and are described in the Supplementary Material.

2.8.

3.

Results and discussion

3.1.

Stage 1 e U(VI) as the sole electron acceptor

Fig. 2 shows the measured flow rate and U concentrations in Stage 1, as well as calculated U-removal percentages, U surface loadings, and U-removal fluxes. During the first 6 days in batch mode, 98% of the influent U was removed, and Fig. 3 shows that black-colored solids were observable around the fiber bundles, verifying successful biofilm formation and U precipitation. After two cycles of semi-continuous operation, the U-removal ratio increased to 99.6%, with the effluent U concentration dropping to 0.002 mM. When continuous mode began, the U removal slightly decreased to 95%, but rebounded and reached steady state at over 97%, with an average effluent U concentration of 0.01 mM and an average U surface loading of 5.9 e
mEq/m2-day. The average U surface loading was lower than the nitrate loading in denitrifying MBfRs of similar design (20e200 e
mEq/m2-day) (Lee and Rittmann, 2002; Ontiveros-Valencia et al., 2012; Zhao et al., 2011) due to the lower flow rate. After 19 days of continuous operation, we sampled the fiber biofilm. This process exposed the reactor to the air for

Solid separation and characterization

After sampling the biofilm in the end of Stage 2, we took out the whole coupon fiber bundle and cut off a few fiber pieces 0.5e1 cm in length. These fiber samples were chemically fixed and then thin-sectioned at 70 nm thickness for analysis on transmission electron microscopy coupled with energydispersive X-ray spectroscopy (TEM-EDX). The details are described in the Supplementary Material. For TEM-EDX analysis, the section samples were loaded on a Lacey carbon 300-mesh copper TEM grid (Ted-Pella, Inc., Redding, CA, USA). TEM images were captured using a Philips CM200-FEG high resolution TEM/STEM (FEI Corp., Eindhoven, The Netherlands) operated at 200 kV, and elemental

Fig. 3 e Photograph of a part of the main fiber bundles in the MBfR, with associated black-colored solids attributable to U immobilization, taken after 9 days of continuous operation in Stage 1.

(3) (4) (3) (1) (1) 105 102 182 87 106 5.7 5.4 10.1 4.6 5.6

(0.3) (0.4) (0.1) (0.1) (<0.1)

97 95 94 81 95

(1) (1) (<1) (<1) (<1)

mg/L/day % e mEq/m -day

5.9 (0.3) 5.7 (0.4) 10.7 (0.1) 5.5 (0.1) 5.7 (<0.1) (0.9) (1.5) (0.7) (0.8) (0.1) 3.2 5.2 6.9 9.5 2.5 1833 (29) 1849 (54) 1851 (42) 1815 (76) 777 (28)

10

e

d

c

HRT is short for hydraulic retention time. ALK is short for alkalinity. Open-up indicates the event we opened up the reactor for biofilm sampling. Numbers in the parentheses are the standard deviations of the average parameter values. Q stands for the flow rate. b

2

Before open-up After open-up Doubled Qe 30 mM HCO3 15 mM HCO3 1

a

1.1 (0.1) 1.1 (0.1) 0.6 (0.1) 1.1 (0.1) 1.1 (0.1) (0.1) (0.1) (0.4) (0.1) (0.1) 32e45 73e112 119e154 48e69 76e94

10

3.9 3.9 7.4 3.9 3.9

mL/min

Day

7.5 7.5 7.6 7.7 6.6

(0.2) (0.1) (0.2) (0.1) (0.1)

c.a.

8.2 8.3 8.3 8.8 8.8

(0.4) (0.4) (0.3) (0.1) (0.1)

mg/L CaCO3

Fig. 4 e Performance parameters in the MBfR during Stage 2. Top: measured concentrations of soluble U and sulfate in influent and effluent. Bottom: reduction fluxes (bars) and average surface loadings (dotted lines) of soluble U (dark gray color) and sulfate (light gray color).

1856 (40) 1870 (32) 1861 (42) 1860 (46) 783 (26)

mM



2
2

d

Ratio

259

2

c

ALKinfb pHeff pHinf HRTa Flow rate Days Steady state condition Stage

Table 1 e The average performance parameters at five steady states in Stages 1 and 2.

ALKeff

[U]eff

U surface loading

Flux

U removal

Rate

w a t e r r e s e a r c h 6 4 ( 2 0 1 4 ) 2 5 5 e2 6 4

5 min, and this exposure seemed to affect greatly the reactor's performance for up to 28 days. The temporary O2 intrusion seemed to inhibit the activity of the anaerobic bacteria, perhaps by competing with U(VI) as the electron acceptor for H2 oxidation. Moreover, it is highly possible that some of the precipitated UO2 was oxidized and dissolved. As a result, the U concentration in the effluent increased to 100 mg/L, and net U removal dropped to only 12% for 14 days. Subsequently, the U concentration in the effluent began to decrease, and the U removal returned to 95% after another 14 days. From that point, the system performed stably for the next 39 days. Thus, O2 exposure had far-reaching impacts on microbial activities and UO2 dissolution, but the biofilm community also had the capacity for self-recovery. On the 112th day of this stage, we doubled the U surface loading by increasing the flow rate two-fold. U removal dropped to 91% within 3 days, but rapidly rebounded to ~94% after another 3 days, after which the system performed stably for 35 days. The average steady-state performance parameters for the two flow rates e including hydraulic retention time (HRT), U surface loading, U removal flux, effluent U concentration, U removal ratio, and U removal rate – are summarized in Table 1. The fact that the U-removals for the two flow rates were almost the same indicates that a U(VI)-surface loading as high as 11 e
mEq/m2-day in MBfR did not approach a loading threshold based on the maximum H2 delivery capacity (113 e
mEq/m2-day at 23 psig in the nonporous polypropylene fiber), U(VI)-reduction kinetics, U(VI) mass transport, or build up of UO2 in the biofilm.

260

w a t e r r e s e a r c h 6 4 ( 2 0 1 4 ) 2 5 5 e2 6 4

Fig. 5 e TEM images of various locations in the fixed cross section of a single fiber with biofilm. (A) the overall cross section of the fiber sample; (B) a layer of crystallite UO2 aggregates were attached directly on the fiber surface and sandwiched biofilms with another inner layer; (C) the magnified image of a selected area in B, showing the crystallite lattice fringes; (D) amorphous U precipitates associated with a microbial cell; (E) amorphous U precipitates associated with a spore-shaped structure; and (F) the magnified image of a selected area in E, showing the U precipitates were associated on the surface of the spore-shaped structure.

3.2. Stage 2 e U(VI) and sulfate as concomitant electron acceptors In Stage 2, we supplemented the medium with sulfate as a second electron acceptor. Fig. 4 shows the concentrations and fluxes of U and sulfate in Stage 2. Unexpectedly, sulfate reduction was insignificant throughout the 94 days of operation. However, the presence of sulfate affected microbial U(VI) reduction. U removal reached 96% within 3 days, but slowly decreased during the following 45 days, reaching a steady state at 81 ± 1% for the next 21 days. On the 69th day, we reduced the bicarbonate concentration in the medium from 30 mM to 15 mM. U removal subsequently returned to 95% within 7 days and remained there until we opened up the reactor for sampling 18 days later. The average performance parameters of these two steadystate conditions with different bicarbonate dosages are summarized in Table 1. One factor possibly affecting U-removal is pH. The primary U(VI) species at neutral pH are ðUO2 Þ2 ðCO3 Þ2 2
and ðUO2 Þ2 ðCO3 Þ3 4
, but (ðUO2 Þ2 ðCO3 Þ3 4
is predominant over ðUO2 Þ2 ðCO3 Þ2 2
when pH is higher than 8 (Bernhard et al., 1996, 2001; Gorman-Lewis et al., 2005; Kalmykov and Choppin, 2000; Merroun and Selenska-Pobell, 2008). According to Eqns. (1) and (2), U(VI) reduction coupled with H2 oxidation by biofilms releases free bicarbonate and also consumes protons. These reactions caused a notable pH increase in the reactor e from about 7.5 to 8.8 for 30 mM

bicarbonate in influent e even though the measured alkalinity changed little. Halving the bicarbonate addition caused a noticeable pH decrease in the influent medium. However, the reduced buffer capacity, due to less bicarbonate, along with increased bicarbonate release and proton consumption from enhanced U(VI) reduction, caused a sharper pH increase, from 6.6 to 8.8, as high as for 30 mM bicarbonate. Since the pH was the same for both bicarbonate concentrations, the decrease of U removal for 30 mM bicarbonate probably was due to the direct effect of U bioreduction itself (Abdelouas et al., 2000; Phillips et al., 1995). In particular, earlier research (Wu et al., 2007) discovered that reducing the bicarbonate concentration from 40 mM to 10 mM stimulated sporulation, and the combination of our pyrosequencing and TEM imaging data (shown in sections 3.3 and 3.4) suggest the presence of Clostridiaceae spores in the biofilm in Stage 2. During the entire course of operation in both stages, sulfide, methane and acetate in the effluent were not detected, indicating the insignificance of sulfate reduction, methanogenesis and acetogenesis.

3.3.

Solid characterization of biofilm samples

Fig. S1, which shows the XRD pattern of the solids separated from the whole coupon fiber bundle, reveals the presence of crystallite UO2 with a thickness of 4.0 ± 0.2 nm after over 8 months operation. Fig. 5 and S2eS4 present the TEM images

w a t e r r e s e a r c h 6 4 ( 2 0 1 4 ) 2 5 5 e2 6 4

and EDX spectra of selected locations in the fixed fiber sample. In some places, biofilms look detached and torn from the fiber wall (Fig. 5A); this occurred because the epoxy resin for TEM did not completely penetrate into the biofilm matrix due to the abundant presence of the U precipitates; thus, the fixation was not complete. We avoided these areas when capturing magnified images (Fig. 5BeF and S2eS4). The image of the whole fiber cross-section (Fig. 5A) reveals that a thin layer of biofilm that included precipitates was formed around the outside wall of the hollow fiber. Magnified images of selected locations further reveal that most UO2 crystallites formed mmsized aggregates throughout the biofilm matrix; in particular, a portion of the aggregates attached directly on the fiber surface in a relatively organized arrangement, while microorganisms grew outside the aggregates and were scarcely able to reach the fiber surface. In other places, two layers of U aggregates sandwiched microbial cells (Fig. 5B and S2). This is confirmed by the lattice fringes as pointed out by the arrows in Fig. 5C; the fringes showed an estimated average (111) dspacing of 3.10 ± 0.1 Å, which corresponds with the structure of biogenic UO2 (Singer et al., 2009). U precipitates also were strongly associated with microbial cells (Fig 5D), and Figure S3 reveals that precipitates were located on the cell membrane, inside the cell, or associated with extracellular polymers in-between cells (indicated by higher C-to-U signals in the EDX spectra). In particular, some cells were encrusted by U and probably had undergone lysis as a consequence, a common phenomenon previously observed during metal immobilizations by biofilms (Brown et al., 1998; Gillan and DeRidder, 1997). Fig. 5C and E show plentiful round-shaped structures that are similar in size and morphology to mature spores produced by Clostridium spp. (Hofstetter et al., 2012; Permpoonpattana et al., 2011; Yang et al., 2009). U precipitates also appear in close proximity to the exosporium-like surface, the outermost shell surrounding the mature spore (Fig. 5F and S4); this proximity is consistent with what Dalla Vecchia et al. (2010) observed in batch tests of U(VI) reduction by spores of Clostridium acetobutylicum. In contrast to crystalline phase of the layered UO2 aggregates, the U precipitates associated with cells and putative spores were mostly in amorphous gel-like phases without observable lattice fringes; this suggests the presence of amorphous UO2 and/or U(VI) immobilized by adsorption. Additional analytical techniques (e.g., XANES) are necessary to allow us to differentiate elemental oxidation states and chemical bonds at certain locations of a heterogeneous microenvironment. In addition, neither Fe nor S signals was found in the EDX spectra (Figs S2eS4). This suggests the absence of iron sulfide precipitates and further confirms the insignificant sulfate reduction.

3.4.

261

day continuous operation in Stage 1. Instead, members of the Rhodocyclaceae family and the Veillonellaceae family (all Thermosinus spp.; Fig S5) became predominant in the biofilm community, with abundances of 69% and 14%, respectively. The Rhodocyclaceae family belongs to Rhodocyclales, an order predominant in nitrate- and perchlorate-reducing MBfRs (Ontiveros-Valencia et al., 2013; 2014; Zhao et al., 2013). Members in this family have been reported to be responsible for U(VI) bioreduction coupled with organic electron donors (Martins et al., 2010c; Williams et al., 2013). Thus, our results suggest that Rhodocyclaceae family members grew lithotrophically by respiring U(VI) coupled with H2 oxidation. No previous research has reported that Thermosinus spp. has U(VI)-reducing capacity. However, the Veillonellaceae family belongs to the Clostridiales order, which is believed to play a role in breaking down biomass or other organic matter during in situ U(VI) bioreduction (Bargar et al., 2013; Boonchayaanant et al., 2009). Thus, we interpret that Thermosinus spp. did not contribute directly to U(VI) reduction, but facilitated U(VI) reducers by degrading biomass and EPS through fermentation (Sokolova et al., 2004). The major shift of the community away from D. vulgaris during Stage 1 is similar to what Martins et al. (2010b) observed in batch tests of U(VI) reduction: a drastic change from SRB (including D. vulgaris) to Clostridium genus and Rhodocyclaceae families after U(VI) exposure. We interpret that D. vulgaris was not able to conserve energy and form a good biofilm in the H2-based system with U(VI) reduction alone. This caused D. vulgaris to be outcompeted by Rhodocyclaceae. Regardless of the near absence of sulfate reduction in Stage 2, the microbial community underwent a drastic shift: the Rhodocyclaceae family decreased to 2%, the Veillonellaceae family became negligible, and two other families in the Clostridiales order e Clostridiaceae and an unknown Clostridiales e became predominant, with abundances of 67% and 17%, respectively. Clostridium is the only genus in the Clostridiaceae family known to be capable of U(VI) reduction: heterotrophically by vegetative cells or lithotrophically by spores. The spores, in spite of being in a dormant state, catalyzed redox reactions via hydrogenases and reductases at the surface of exosporium

Microbial community structure of the biofilm

Fig. 6 shows relative abundances of the most abundant microbial phylotypes at the family level for the biofilm samples for Stage 1 (U alone) and Stage 2 (U and sulfate). Perhaps surprisingly, the inoculated D. vulgaris almost disappeared (a negligible abundance of 0.2%) after the 26-day startup and 19-

Fig. 6 e Relative abundances of the most abundant microbial phylotypes at the family level for the biofilm samples in Stages 1 and 2. Note that Desulfovibrionaceae were present at such a low level (<0.2%) that they can hardly be seen by the bar graph.

262

w a t e r r e s e a r c h 6 4 ( 2 0 1 4 ) 2 5 5 e2 6 4

(Dalla Vecchia et al., 2010). TEM images in Fig. 5C and E add weight to the existence of spores. The appearance of Clostridiaceae after SO4 2
addition is similar to the Cr(VI)-reducing tests reported by Martins et al. (2010a): the presence of SO4 2
stimulated sporulation of Clostridium spp., leading to U(VI) reduction in the exosporium. The Clostridiaceae became predominant over the Rhodocyclaceae probably due to greater tolerance of their spores to alkaline conditions in long term (Wiegel et al., 2006). The results of T-RFLP analysis performed with the MseI restriction enzyme (Fig S6) were consistent with the pyrosequencing data. The single T-RF of 551 base pairs confirmed the relative purity of the D. vulgaris cells for initial inoculation of the reactor. In stage 1, the absence of the T-RF of 551 indicates the exclusion of D. vulgaris in the microbial community; instead, two detectable T-RFs of 201 and 595 with the fluorescence intensities of 79.9% and 5.2% probably represent the Veillonellaceae and the Rhodocyclaceae families detected with pyrosequencing, respectively. The microbial community shifted again from Stage 1 to Stage 2, with completely new bacteria present at T-RFs of 542 and 593 with the fluorescence intensities of 85.8% and 5.3%, respectively; these base pairs may represent the Clostridiaceae and Chlostridiales species, respectively.

biofilm by XRD and TEM-EDX. Nanocrystalline UO2 aggregates and amorphous U precipitates were associated with vegetative cells and mature spores that accumulated in the biofilm matrix. Sulfate was not reduced, and D. vulgaris and other known SRB were not present in the biofilm; rather, the biofilm community mainly contained members of Rhodocyclaceae family and Thermosinus spp. when U(VI) was the only electron acceptor, and it underwent a dramatic shift to two other groups of the Clostridiales order when sulfate was added. The presence of the Clostridiaceae family can be linked to the observed spores that may have been responsible for U(VI) reduction in the biofilm in Stage 2.

Acknowledgment We gratefully acknowledge the use of facilities supervised by David Lowry in the School of Life Science, Thomas Groy in the Department of Chemistry and biochemistry and by Karl Weiss in the LeRoy Eyring Center for Solid State Science, all at Arizona State University.

Appendix A. Supplementary data 3.5.

Environmental implications

Uranium in contaminated groundwater is a threat to the health of people living in nearby areas as it enters into the drinking water distribution system (Lytle et al., 2014). Microbial U removal has a number of advantages over conventional chemical methods; however, although widely studied for in situ remediation, microbial U removal has very limited application in drinking water treatment (Katsoyiannis and Zouboulis, 2013; Lovley and Phillips, 1992b). Our study reports successful removal of most U from water during 8 months of continuous operation. The system reached a maximum removal rate of 180 mg-U/L/day, which was 5e10 times higher than reported for other bench-scale biofilm systems (Beyenal et al., 2004). This demonstrates that the MBfR technology may be a practical treatment for cleaning up the U-contaminated groundwater. In addition, our study confirms that a particularly high loading of U enabled formation and growth of biofilms through U(VI) respiration alone; this phenomenon has not been reported before and supports practical applications of the MBfR technology for treating uranium-mill wastewater, which has a high concentration of soluble U, but no other oxidized contaminants.

4.

Conclusion

Slow-growing U(VI) reducers formed biofilms on the fibers of the H2-based MBfR after an extended startup. With an average U surface loading of 5.7 e- mEq/m2-day and an average flow rate of 0.039 mL/min, the MBfR successfully achieved and maintained 94e95% U(VI) removal at steady state with or without sulfate during over 8 months of operation, and the biofilm was capable of self-recovery after a disturbance due to O2 exposure. Reduced U solids were clearly identifiable in the

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.watres.2014.07.013.

references

Abdelouas, A., Lutze, W., Gong, W.L., Nuttall, E.H., Strietelmeier, B.A., Travis, B.J., 2000. Biological reduction of uranium in groundwater and subsurface soil. Sci. Total Environ. 250 (1e3), 21e35. Arfsten, D.P., Still, K.R., Ritchie, G.D., 2001. A review of the effects of uranium and depleted uranium exposure on reproduction and fetal developmentt. Toxicol. Ind. Health 17 (5e10), 180e191. Badziong, W., Thauer, R.K., Zeikus, J.G., 1978. Isolation and characterization of Desulfovibrio growing on hydrogen plus sulfate as the sole energy source. Arch. Microbiol. 116 (1), 41e49. Baird, C., Cann, M., 2005. Environmental Chemistry. University of Western Ontario, London, Ontario, Canada. Bargar, J.R., Bernier-Latmani, R., Giammar, D.E., Tebo, B.M., 2008. Biogenic uraninite nanoparticles and their importance for uranium remediation. Elements 4 (6), 407e412. Bargar, J.R., Williams, K.H., Campbell, K.M., Long, P.E., Stubbs, J.E., Suvorova, E.I., Lezama-Pacheco, J.S., Alessi, D.S., Stylo, M., Webb, S.M., Davis, J.A., Giammar, D.E., Blue, L.Y., BernierLatmani, R., 2013. Uranium redox transition pathways in acetate-amended sediments. Proc. Natl. Acad. Sci. U. S. A. 110 (12), 4506e4511. Bernhard, G., Geipel, G., Brendler, V., Nitsche, H., 1996. Speciation of uranium in seepage waters of a mine tailing pile studied by time-resolved laser-induced fluorescence spectroscopy (TRLFS). Radiochim. Acta 74, 87e91. Bernhard, G., Geipel, G., Reich, T., Brendler, V., Amayri, S., Nitsche, H., 2001. Uranyl(VI) carbonate complex formation: validation of the Ca2UO2(CO3)3(aq.) species. Radiochim. Acta 89 (8), 511e518.

w a t e r r e s e a r c h 6 4 ( 2 0 1 4 ) 2 5 5 e2 6 4

Beyenal, H., Sani, R.K., Peyton, B.M., Dohnalkova, A.C., Amonette, J.E., Lewandowski, Z., 2004. Uranium immobilization by sulfate-reducing biofilms. Environ. Sci. Technol. 38 (7), 2067e2074. Bleise, A., Danesi, P.R., Burkart, W., 2003. Properties, use and health effects of depleted uranium (DU): a general overview. J. Environ. Radioact. 64 (2e3), 93e112. Boonchayaanant, B., Nayak, D., Du, X., Criddle, C.S., 2009. Uranium reduction and resistance to reoxidation under ironreducing and sulfate-reducing conditions. Water Res. 43 (18), 4652e4664. Brown, D.A., Beveridge, T.J., Keevil, C.W., Sheriff, B.L., 1998. Evaluation of microscopic techniques to observe iron precipitation in a natural microbial biofilm. Fems Microbiol. Ecol. 26 (4), 297e310. Bru¨mmer, G., Gerth, J., Herms, U., 1986. Heavy metal species, mobility and availability in soils. Pflanzenernaehr. Bodenkd 149, 382e398. Burgos, W.D., McDonough, J.T., Senko, J.M., Zhang, G.X., Dohnalkova, A.C., Kelly, S.D., Gorby, Y., Kemner, K.M., 2008. Characterization of uraninite nanoparticles produced by Shewanella oneidensis MR-1. Geochim. Et. Cosmochim. Acta 72 (20), 4901e4915. Chung, J., Rittmann, B.E., 2007. Bio-reductive dechlorination of 1,1,1-trichloroethane and chloroform using a hydrogen-based membrane biofilm reactor. Biotechnol. Bioeng. 97 (1), 52e60. Chung, J., Rittmann, B.E., 2008. Simultaneous bio-reduction of trichloroethene, trichloroethane, and chloroform using a hydrogen-based membrane biofilm reactor. Water Sci. Technol. 58 (3), 495e501. Chung, J., Nerenberg, R., Rittmann, B.E., 2006. Bioreduction of selenate using a hydrogen-based membrane biofilm reactor. Environ. Sci. Technol. 40 (5), 1664e1671. Chung, J., Nerenberg, R., Rittmann, B.E., 2007a. Evaluation for biological reduction of nitrate and perchlorate in brine water using the hydrogen-based membrane biofilm reactor. J. Environ. Eng. ASCE 133 (2), 157e164. Chung, J., Rittmann, B.E., Wright, W.F., Bowman, R.H., 2007b. Simultaneous bio-reduction of nitrate, perchlorate, selenate, chromate, arsenate, and dibromochloropropane using a hydrogen-based membrane biofilm reactor. Biodegradation 18 (2), 199e209. Chung, J., Ahn, C.H., Chen, Z., Rittmann, B.E., 2008a. Bio-reduction of N-nitrosodimethylamine (NDMA) using a hydrogen-based membrane biofilm reactor. Chemosphere 70 (3), 516e520. Chung, J., Krajmalnik-Brown, R., Rittmann, B.E., 2008b. Bioreduction of trichloroethene using a hydrogen-based membrane biofilm reactor. Environ. Sci. Technol. 42 (2), 477e483. Dalla Vecchia, E.C., Veeramani, H., Suvorova, E.I., Wigginton, N.S., Bargar, J.R., Bernier-Latmani, R., 2010. U(VI) reduction by spores of Clostridium acetobutylicum. Res. Microbiol. 161 (9), 765e771. Elias, D.A., Suflita, J.M., McInerney, M.J., Krumholz, L.R., 2004. Periplasmic cytochrome c3 of Desulfovibrio vulgaris is directly involved in H2-mediated metal but not sulfate reduction. Appl. Environ. Microbiol. 70 (1), 413e420. Fernandez-Garcia, M., Martinez-Arias, A., Hanson, J.C., Rodriguez, J.A., 2004. Nanostructured oxides in chemistry: characterization and properties. Chem. Rev. 104 (9), 4063e4104. Gao, W.M., Francis, A.J., 2008. Reduction of uranium(VI) to uranium(IV) by Clostridia. Appl. Environ. Microbiol. 74 (14), 4580e4584. Gavrilescu, M., Pavel, L.V., Cretescu, I., 2009. Characterization and remediation of soils contaminated with uranium. J. Hazard. Mater. 163, 475e510.

263

Gilbert, B., Huang, F., Zhang, H.Z., Waychunas, G.A., Banfield, J.F., 2004. Nanoparticles: strained and stiff. Science 305 (5684), 651e654. Gillan, D.C., DeRidder, C., 1997. Morphology of a ferric ironencrusted biofilm forming on the shell of a burrowing bivalve (Mollusca). Aquat. Microb. Ecol. 12 (1), 1e10. Gorman-Lewis, D., Elias, P.E., Fein, J.B., 2005. Adsorption of aqueous uranyl complexes onto Bacillus subtilis cells. Environ. Sci. Technol. 39 (13), 4906e4912. Hofstetter, S., Denter, C., Winter, R., McMullen, L.M., Ganzle, M.G., 2012. Use of the fluorescent probe LAURDAN to label and measure inner membrane fluidity of endospores of Clostridium spp. J. Microbiol. Methods 91 (1), 93e100. Jeon, B.H., Kelly, S.D., Kemner, K.M., Barnett, M.O., Burgos, W.D., Dempsey, B.A., Roden, E.E., 2004. Microbial reduction of U(VI) at the solid-water interface. Environ. Sci. Technol. 38 (21), 5649e5655. 2Kalmykov, S.N., Choppin, G.R., 2000. Mixed Ca2þ/UO2þ 2 /CO3 complex formation at different ionic strengths. Radiochim. Acta 88 (9e11), 603e606. € Karatas‚, S., Hasar, H., Tas‚kan, E., Ozkaya, B., S‚ahinkaya, E., 2014. Bio-reduction of tetrachloroethen using a H2-based membrane biofilm reactor and community fingerprinting. Water Res. 58, 21e28. Katsoyiannis, I., Zouboulis, A., 2013. Removal of uranium from contaminated drinking water: a mini review of available treatment methods. Desalination Water Treat. 51 (13e15), 2915e2925. Langmuir, D., 1978. Uranium solution-mineral equilibria at lowtemperatures with applications to sedimentary ore-deposits. Geochim. Et. Cosmochim. Acta 42 (6), 547e569. Lee, K.C., Rittmann, B.E., 2000. A novel hollow-fibre membrane biofilm reactor for autohydrogenotrophic denitrification of drinking water. Water Sci. Technol. 41 (4e5), 219e226. Lee, K.C., Rittmann, B.E., 2002. Applying a novel autohydrogenotrophic hollow-fiber membrane biofilm reactor for denitrification of drinking water. Water Res. 36 (8), 2040e2052. Lee, S.Y., Baik, M.H., Choi, J.W., 2010. Biogenic formation and growth of uraninite (UO2). Environ. Sci. Technol. 44 (22), 8409e8414. Liu, W.T., Marsh, T.L., Cheng, H., Forney, L.J., 1997. Characterization of microbial diversity by determining terminal restriction fragment length polymorphisms of genes encoding 16S rRNA. Appl. Environ. Microbiol. 63 (11), 4516e4522. Lovley, D.R., Phillips, E.J., 1992a. Reduction of uranium by Desulfovibrio desulfuricans. Appl. Environ. Microbiol. 58 (3), 850e856. Lovley, D.R., Phillips, E.J.P., 1992b. Bioremediation of uranium contamination with enzymatic uranium reduction. Environ. Sci. Technol. 26 (11), 2228e2234. Lovley, D.R., Roden, E.E., Phillips, E.J.P., Woodward, J.C., 1993a. Enzymatic iron and uranium reduction by sulfate-reducing bacteria. Mar. Geol. 113 (1e2), 41e53. Lovley, D.R., Widman, P.K., Woodward, J.C., Phillips, E.J.P., 1993b. Reduction of uranium by cytochrome c3 of Desulfovibrio vulgaris. Appl. Environ. Microbiol. 59 (11), 3572e3576. Lytle, D.A., Sorg, T., Wang, L., Chen, A., 2014. The accumulation of radioactive contaminants in drinking water distribution systems. Water Res. 50, 396e407. Martin, K.J., Nerenberg, R., 2012. The membrane biofilm reactor (MBfR) for water and wastewater treatment: principles, applications, and recent developments. Bioresour. Technol. 122, 83e94. Martins, M., Faleiro, M.L., Chaves, S., Tenreiro, R., Costa, M.C., 2010a. Effect of uranium (VI) on two sulphate-reducing

264

w a t e r r e s e a r c h 6 4 ( 2 0 1 4 ) 2 5 5 e2 6 4

bacteria cultures from a uranium mine site. Sci. Total Environ. 408 (12), 2621e2628. Martins, M., Faleiro, M.L., Chaves, S., Tenreiro, R., Santos, E., Costa, M.C., 2010b. Anaerobic bio-removal of uranium (VI) and chromium (VI): comparison of microbial community structure. J. Hazard. Mater. 176 (1e3), 1065e1072. Martins, M., Faleiro, M.L., da Costa, A.M.R., Chaves, S., Tenreiro, R., Matos, A.P., Costa, M.C., 2010c. Mechanism of uranium (VI) removal by two anaerobic bacterial communities. J. Hazard. Mater. 184 (1e3), 89e96. Merroun, M.L., Selenska-Pobell, S., 2008. Bacterial interactions with uranium: an environmental perspective. J. Contam. Hydrol. 102 (3e4), 285e295. Noguera, D.R., Brusseau, G.A., Rittmann, B.E., Stahl, D.A., 1998. A unified model describing the role of hydrogen in the growth of Desulfovibrio vulgaris under different environmental conditions. Biotechnol. Bioeng. 59 (6), 732e746. Ontiveros-Valencia, A., Ziv-El, M., Zhao, H.P., Feng, L., Rittmann, B.E., Krajmalnik-Brown, R., 2012. Interactions between nitrate-reducing and sulfate-reducing bacteria coexisting in a hydrogen-fed biofilm. Environ. Sci. Technol. 46 (20), 11289e11298. Ontiveros-Valencia, A., Ilhan, Z.E., Kang, D.W., Rittmann, B., Krajmalnik-Brown, R., 2013. Phylogenetic analysis of nitrateand sulfate-reducing bacteria in a hydrogen-fed biofilm. FEMS Microbiol. Ecol. 85 (1), 158e167. Ontiveros-Valencia, A., Tang, Y., Krajmalnik-Brown, R., Rittmann, B.E., 2014. Managing the interactions between sulfate-and perchlorate-reducing bacteria when using hydrogen-fed biofilms to treat a groundwater with a high perchlorate concentration. Water Res. 55, 215e224. Permpoonpattana, P., Tolls, E.H., Nadem, R., Tan, S., Brisson, A., Cutting, S.M., 2011. Surface layers of Clostridium difficile endospores. J. Bacteriol. 193 (23), 6461e6470. Phillips, E.J.P., Landa, E.R., Lovley, D.R., 1995. Remediation of uranium contaminated soils with bicarbonate extraction and microbial U(Vi) reduction. J. Ind. Microbiol. 14 (3e4), 203e207. Rittmann, B.E., 2006. The membrane biofilm reactor: the natural partnership of membranes and biofilm. Water Sci. Technol. 53 (3), 219e225. Rittmann, B.E., Chung, J.W., Wright, W.F., Bowman, R.H., 2005. Remediation of perchlorate and other co-contaminants using the hydrogen-based membrane biofilm reactor. Abstr. Pap. Am. Chem. Soc. 230, U1548eU1549. Sanford, R.A., Wu, Q., Sung, Y., Thomas, S.H., Amos, B.K., € ffler, F.E., 2007. Hexavalent uranium supports Prince, E.K., Lo growth of Anaeromyxobacter dehalogenans and Geobacter spp. with lower than predicted biomass yields. Environ. Microbiol. 9 (11), 2885e2893. Sheng, J., Kim, H.W., Badalamenti, J.P., Zhou, C., Sridharakrishnan, S., Krajmalnik-Brown, R., Rittmann, B.E., Vannela, R., 2011. Effects of temperature shifts on growth rate and lipid characteristics of Synechocystis sp PCC6803 in a bench-top photobioreactor. Bioresour. Technol. 102 (24), 11218e11225. Singer, D.M., Farges, F., Brown, G.E., 2009. Biogenic nanoparticulate UO2: synthesis, characterization, and factors affecting surface reactivity. Geochim. Et. Cosmochim. Acta 73 (12), 3593e3611.

Sokolova, T.G., Gonzalez, J.M., Kostrikina, N.A., Chernyh, N.A., Slepova, T.V., Bonch-Osmolovskaya, E.A., Robb, F.T., 2004. Thermosinus carboxydivorans gen. nov., sp nov., a new anaerobic, thermophilic, carbon-monoxide-oxidizing, hydrogenogenic bacterium from a hot pool of Yellowstone National Park. Int. J. Syst. Evol. Microbiol. 54, 2353e2359. Tang, Zhou, C., Van Ginkel, S.W., Ontiveros-Valencia, A., Shin, J., Rittmann, B.E., 2012. Hydrogen permeability of the hollow fibers used in H2-based membrane biofilm reactors. J. Membr. Sci. 407, 176e183. Wall, J.D., Krumholz, L.R., 2006. Uranium reduction. Annu. Rev. Microbiol. 60, 149e166. Wan, J.M., Tokunaga, T.K., Brodie, E., Wang, Z.M., Zheng, Z.P., Herman, D., Hazen, T.C., Firestone, M.K., Sutton, S.R., 2005. Reoxidation of bioreduced uranium under reducing conditions. Environ. Sci. Technol. 39 (16), 6162e6169. Wiegel, J., Tanner, R., Rainey, F.A., 2006. The prokaryotes. In: Dworkin, M.e. (Ed.), Proteobacteria: Gamma Subclass, vol. 6. Springer, USA, pp. 654e678. Williams, K.H., Bargar, J.R., Lloyd, J.R., Lovley, D.R., 2013. Bioremediation of uranium-contaminated groundwater: a systems approach to subsurface biogeochemistry. Curr. Opin. Biotechnol. 24 (3), 489e497. Wu, C.G., Zhang, G.S., Liu, X.L., Dong, X.Z., 2007. Bicarbonate is a stimulus in the inter-species induced sporulation of strict anaerobic Syntrophomonas erecta subsp sporosyntropha. Extremophiles 11 (6), 827e832. Yang, W.W., Crow-Willard, E.N., Ponce, A., 2009. Production and characterization of pure Clostridium spore suspensions. J. Appl. Microbiol. 106 (1), 27e33. Zhao, H.P., Van Ginkel, S., Tang, Y.N., Kang, D.W., Rittmann, B., Krajmalnik-Brown, R., 2011. Interactions between perchlorate and nitrate reductions in the biofilm of a hydrogen-based membrane biofilm reactor. Environ. Sci. Technol. 45 (23), 10155e10162. Zhao, H.P., Ontiveros-Valencia, A., Tang, Y.N., Kim, B.O., Ilhan, Z.E., Krajmalnik-Brown, R., Rittrnann, B., 2013. Using a two-stage hydrogen-based membrane biofilm reactor (MBfR) to achieve complete perchlorate reduction in the presence of nitrate and sulfate. Environ. Sci. Technol. 47 (3), 1565e1572. Zhou, C., Vannela, R., Hayes, K.F., Rittmann, B.E., 2014. Effect of growth conditions on microbial activity and iron-sulfide production by Desulfovibrio vulgaris. J. Hazard. Mater. 272, 28e35. Ziv-El, M.C., Rittmann, B.E., 2009. Systematic evaluation of nitrate and perchlorate bioreduction kinetics in groundwater using a hydrogen-based membrane biofilm reactor. Water Res. 43 (1), 173e181. Ziv-El, M., Delgado, A.G., Yao, Y., Kang, D.W., Nelson, K.G., Halden, R.U., Krajmalnik-Brown, R., 2012. Development and characterization of DehaloR^2, a novel anaerobic microbial consortium performing rapid dechlorination of TCE to ethene. Appl. Microbiol. Biotechnol. 95 (1), 273e274. Ziv-El, M., Popat, S.C., Cai, K., Halden, R.U., Krajmalnik-Brown, R., Rittmann, B.E., 2012. Managing methanogens and homoacetogens to promote reductive dechlorination of trichloroethene with direct delivery of H2 in a membrane biofilm reactor. Biotechnol. Bioeng. 109 (9), 2200e2210.