respiration by Desulfovibrio vulgaris

0 downloads 0 Views 1MB Size Report
20 Feb 2018 - hydroxide and carbonate ligands are UO2CO3, UO2(CO3)2. 2,. UO2(CO3)3 ... complexes, the Log β values at T = 25 C, and zero ionic strength, are defined in ... equation (Tutem et al., 1991, VanBriesen & Rittmann, 1999) and ...... bacterial yield prediction: Modifications and comparative evaluations.
Received: 12 October 2017

|

Revised: 16 February 2018

|

Accepted: 20 February 2018

DOI: 10.1002/bit.26579

ARTICLE

pH-dependent speciation and hydrogen (H2) control U(VI) respiration by Desulfovibrio vulgaris Shilpi Kushwaha

|

Andrew K. Marcus | Bruce E. Rittmann

Biodesign Swette Center of Environmental Biotechnology, Arizona State University, Tempe, Arizon

Abstract In situ bioreduction of soluble hexavalent uranium U(VI) to insoluble U(IV) (as UO2) has

Correspondence Shilpi Kushwaha, Division of Organic Chemistry Division, CSIR-National Chemical Laboratory, Pune 411008, India. Email: [email protected]

been proposed as a means of preventing U migration in the groundwater. This work

Funding information Fulbright Foundation, Grant number: Fulbright Nehru postdoctoral fellowship 2013–14; DST-India, Grant number: SB/FT/ CS-072/2014

bicarbonate buffering. In the absence of sulfate, D. vulgaris reduced >90% of the total

focuses on the bioreduction of U(VI) and precipitation of U(IV). It uses anaerobic batch reactors with Desulfovibrio vulgaris, a well-known sulfate, iron, and U(VI) reducer, growing on lactate as the electron donor, in the absence of sulfate, and with a 30-mM soluble U(VI) (1 mM) to form U(IV) solids that were characterized by X-ray diffraction and confirmed to be nano-crystalline uraninite with crystallite size 2.8 ± 0.2 nm. pH values between 6 and 10 had minimal impact on bacterial growth and end-product distribution, supporting that the mono-nuclear, and poly-nuclear forms of U(VI) were equally bioavailable as electron acceptors. Electron balances support that H2 transiently accumulated, but was ultimately oxidized via U(VI) respiration. Thus, D. vulgaris utilized H2 as the electron carrier to drive respiration of U(VI). Rapid lactate utilization and biomass growth occurred only when U(VI) respiration began to draw down the sink of H2 and relieve thermodynamic inhibition of fermentation. KEYWORDS

bioavailable U(VI) species, Desulfovibrio vulgaris, hydrogen gas, UO2 nanoparticles, U(VI) reduction

1 | INTRODUCTION

(Reed, Deo, & Rittmann, 2010). U predominantly occurs in its oxidized (+VI) valence state, which is water soluble, and easily migrates to

A consequence of developing nuclear energy and weapons has been

threaten nearby water resources (Murphy & Shock, 1999). However,

the release of radionuclides, toxic heavy metals, and organic co-

its reduced (+IV) valence state is insoluble, and reduction of soluble

contaminants into the environment (Riley, Zachara, & Wobber, 1992;

U(VI)

Vogel, Borch, Mayes, Jardine, & Fendorf, 2005). Among the radio-

prevent U migration with groundwater (Murphy & Shock, 1999).

nuclides, uranium (U) is of particular concern because of its

Microbiological processes in the subsurface have a potential to

carcinogenicity, long half-life, widespread distribution, and mobility

immobilize U indirectly by altering local redox and pH conditions and

(Riley et al., 1992). Uranium is a redox-active actinide whose redox

directly by reducing U(VI) as an electron acceptor (Reed et al., 2010).

status and chemical speciation affect its mobility and exposure risk

Sulfate-reducing bacteria (SRB) can participate in anaerobic bio-

to

insoluble

U(IV)

(e.g.,

UO2(s),

or

uraninite)

can

reduction and bio-precipitation of U (Reed et al., 2010). Being Current Address of Shilpi Kushwaha is Organic Chemistry Division, CSIR-National Chemical Laboratory, Pune 411008, India.

Biotechnology and Bioengineering. 2018;1–10.

metabolically diverse, SRB can utilize many electron donors, including H2, formate, lactate, succinate, fatty acids, amino acids, and aromatic

wileyonlinelibrary.com/journal/bit

© 2018 Wiley Periodicals, Inc.

|

1

2

KUSHWAHA

|

ET AL.

compounds. Besides SO42−, SRB as a group are known to respire Fe(III),

respiratory electron acceptor for D. vulgaris (Elias et al., 2004;

U(VI), Cr(VI), Tc(VII), Mo(VI), and Pd(II) (Lloyd, 2003; Lloyd, Yong, &

Noguera, Brusseau, Rittmann, & Stahl, 1998). U(VI) respiration is

Macaskie, 1998). Bio-reduction can lead to precipitation of metal oxides,

carried out by periplasmic cytochrome c3, which accepts electrons

phosphates, and carbonates (Lloyd, Mabbett, Williams, & Macaskie,

from two catabolic processes: The inner-membrane protein ecp and

2001). Several SRB-based technologies for metal bioremediation have

periplasmic NiFe hydrogenase. They differ importantly in how they

been proposed, and some have been implemented at full scale (Barnes

receive electrons that originate in lactate.

et al., 1991; Benner, Blowes, Gould, Herbert, & Ptacek, 1999; Laspidou

When lactate is the electron donor, initial fermentation reactions

& Rittmann, 2002; Songkasiri, Reed, & Rittman, 2002; Webb, McGin-

are the same for both routes of electron transfer for U(VI) respiration:

ness, & Lappin-Scott, 1998; White, Sharman, & Gadd, 1998).

Lactate is first fermented into pyruvate and 2(H+ + e−), and subse-

Desulfovibrio vulgaris is among the best-studied SRB and one able to

quently pyruvate is fermented into acetate and 2(H+ + e−) in the

reduce U(VI) (Heidelberg et al., 2004; Lovley & Phillips, 1992a, b; Spear,

cytoplasm. In the pathway that involves ecp, pathway I in Figure 1, the

Figueroa, & Honeyman, 2000; Wall & Krumholz, 2006). However,

four electron equivalents are channelled directly to ecp, located on the

details on how D. vulgaris reduces U(VI) to U(IV) when U(VI) is the sole

inner membrane, before reducing the periplasmic cytochrome c3.

acceptor are scarce (Elias, Suflita, McInerney, & Krumholz, 2004). This is

When the NiFe hydrogenase is involved, pathway II, the four electron

due in large part to the complexity of U(VI)'s aqueous chemistry, as U(VI)

equivalents are first channelled to h-ev in the inner membrane (h-ev

can form monomeric and oligomeric complexes with carbonate,

stands for electrons being transported along the electron transport

hydroxide, and phosphate. Speciation of the ligands and, thus, their

chain in the membrane), where they then generate two H2 or two

ability to bind U(VI) depend on pH. Carbonate is a key ligand in natural

formate in the periplasm. H2 or formate is then oxidized by NiFe

environments, and it influences the speciation of uranium (Clark, Hobart,

hydrogenase, which donates the electron equivalents to cytochrome

& Neu, 1995; Langmuir, 1978; Stumm & Morgan, 1996), along with

c3 for dissimilatory U(VI) reduction (Elias et al., 2004). Should pathway

being an important pH buffer. Although U(VI) speciation and

II act, it is likely that H2 or formate has to build up to make the kinetics

bioavailability of U(VI) in its different complexes have been studied

of U(VI) respiration rapid. This should lead to a period when

for decades in isolation (Belli, DiChristina, Cappellen, & Taillefert, 2015;

fermentation is rapid, but respiration is minimal. Such a lag is unlikely

Brooks et al., 2003; Murphy & Shock, 1999; Stewart, Amos, Nico, &

with pathway I. Because respiration yields more energy per electron

Fendorf, 2011; Sheng & Fein, 2014; Ulrich, Veeramani, Latmani, &

equivalent than fermentation (Noguera et al., 1998; Niki et al., 1984;

Giammar, 2011), the connections between aqueous uranyl speciation,

Rittmann & McCarty, 2001), biomass synthesis also should exhibit a lag

and electron partitioning between respiration & fermentation during

when pathway I is utilized for U(VI) respiration.

U(VI) bioreduction were seldom studied. In sulfate-reducing conditions,

In aqueous systems, U(VI), the most common oxidation state, is

the availability of the electron acceptor affects how electrons from the

highly soluble as hydroxyl, and carbonyl complexes of UO22+. When

electron donor (lactate in our case) are distributed among fermentation

U(VI) is reduced to U(IV), it precipitates as UO2 (Heidelberg et al., 2004;

and respiration end products. Thus, pH ought to affect the bioavailability

Lovley, Phillips, Gorby, & Landa, 1991; Natrajan, Swinburne, Andrews,

and kinetics of U(VI) reduction, as well as the pattern of electron-donor

Randall, & Heath, 2014); in this way, dissimilatory U(VI)-reducing

utilization, and formation of respiratory and fermentation products.

bacteria accelerate the kinetics of reduction, and precipitation of

Here, we expand the mechanistic understanding of the role of

uranium. The distribution of U(VI) species in the presence and absence

U(VI) speciation by growing D. vulgaris in batch reactors with pH controlled between 5 and 10. The cultures were provided with 30 mM bicarbonate to accentuate the effects of uranium–carbonate complexes on the bioavailability of U(VI) and to stabilize the experimental pH. Sulfate was omitted to accentuate the phenomena related to U(VI) reduction. Integrating experiments with mathematical modelling, we explore pH speciation, reaction stoichiometry, and endproduct distribution for U(VI) reduction when U(VI) is the sole acceptor for D. vulgaris. From this, we characterize the temporal relationship between fermentation and respiration, as well as the effect of pHdependent speciation on the bioavailability of U(VI). We find that H2 accumulation is important during the transition from fermentation to respiration of U(VI) and growth of D. vulgaris.

2 | E L E C T R O N F L O W A N D SP E C I A T I O N Figure 1 presents the current understanding of electron flow when lactate is the exogenous electron donor and U(VI) is the only

FIGURE 1 Model for electron transfer and U(VI) reduction by D. vulgaris, based on (Elias et al., 2004, Noguera et al., 1998). Cytochrome c3 has a mid-point redox potential near −300 mV versus SHE (standard hydrogen electrode) at 25 °C (Niki et al., 1984), while the U(VI)-U(IV) couple has a mid-point potential of +334 mV versus SHE

KUSHWAHA

ET AL.

|

3

of dissolved inorganic carbon (DIC) and with a wide range of pH is of

Table 1 lists all the acid/base, dissociation, and complexation

paramount importance for understanding microbial reduction of U(VI),

reactions as formation reactions. The formation coefficients for the

since different U(VI) species could have different susceptibilities of

complexes, the Log β values at T = 25 C, and zero ionic strength, are

being reduced based on their bioavailability. A decrease in U(VI)

defined in Supplementary Information (SI), and the mass-action

reduction rate with increasing DIC concentration, along with

expressions are Supplementary Eqs. S1–S16 in Supplementary

sensitivity towards pH, has been well documented in the literature

Table S1. Activity coefficients were computed with the Davies

(Belli et al., 2015; Bernhard, Geipel, Brendler, & Nitsche, 1998; Brooks

equation (Tutem et al., 1991, VanBriesen & Rittmann, 1999) and

et al., 2003; Croteau, Fuller, Cain, Campbell, & Aiken, 2016; Drobot

used with the activity-corrected formation constants (cβ); they are

et al., 2015; Guillauont et al., 2003; Jones et al., 2015; Sheng & Fein,

shown in Supplementary Table S2 (Tutem et al., 1991). The ionic

2014; Ulrich et al., 2011).

strength of the medium was 0.172 for all experimental conditions.

The conventional thinking is that free UO22+ and its mono-nuclear

The total concentration for each component of the system was

complexes with hydroxides and carbonates are bioavailable to

determined as shown in Supplementary Eqs. S17 through S23 in

microorganisms (Tutem, Apak, Turgut, & Apak, 1991). Poly-nuclear

Supplementary Table S3. A special mass balance was needed for a

complexes are known to have lesser cell permeability and, hence, may

component defined as ACID. Called the proton condition, Supplemen-

have lower bioavailability (Bernhard et al., 1998; Croteau et al., 2016;

tary Eq. S24 balances acid and base equivalents generated by all

Drobot et al., 2015; Guillauont et al., 2003; Jones et al., 2015; Kulkarni,

reactions (Stumm & Morgan, 1996; VanBriesen & Rittmann, 1999).

Misra, Gupta, Ballal, & Apte, 2016). In addition, adsorption of U(VI) to

We combined all mass balances (Supplementary Eqs. S17–S24)

colloidal or particulate ligands should render it even less bioavailable

with the mass-action relationships (Supplementary Eqs. S1–S16) to

(Du et al., 2011).

form a set of coupled non-linear algebraic equations that were solved

Chemical components in the system we studied are lactate

for total concentration of all the complex species (VanBriesen &

(CH3CH(OH)COO−), acetate (CH3COO−), cells (C5H7O2N) (Rittmann

Rittmann, 1999). The Solver function in Microsoft Excel was used to

and McCarty, 2001; McCarty, 2007), the soluble uranyl(VI) cation

speciate the complexes to determine the equilibrium concentrations at

(UO22+), and insoluble uraninite(IV) solid (UO2). To represent specia-

a given pH.

tion in our microbiological setting, additional acid/base components are H2CO3, NH4+, and H2O. The mononuclear complexes of U(VI) with hydroxide and carbonate ligands are UO2CO3, UO2(CO3)22−, UO2(CO3)34−, UO2OH+, UO2(OH)42−. The poly-nuclear species are (UO2)3(CO3)66−,

(UO2)2(OH)22+,

(UO2)3(OH)5+,

(UO2)2(OH)3+,

(UO2)3(OH)42+, (UO2)3(OH)7−, and (UO2)4(OH)7+. Acid–base and

3 | MATERIALS AND METHO DS 3.1 | Media preparation All chemicals were high-purity analytical grade and obtained from

complexation reactions for the complexes with carbonate, ammonium,

Sigma–Aldrich (St. Louis, MO). D. vulgaris subsp. vulgaris Postgate, and

and water, along with their formation constants (Log β), are given in

Campbell (ATCC 25979) was procured from American Type Culture

Table 1 (Grenthe et al., 1991).

Collection (Rockville, Md.). The two growth media used for culture

TABLE 1

Acid-base and complexation reactions with their formation constants (log β)

Eq

Complex

Log β

Equation

1

UO2CO3

9.7

UO22+ + H2CO3 UO2CO30 + 2H+

2

UO2(CO3)22−

16.9

UO22+ + 2H2CO3 UO2(CO3)22- + 4H+

3

UO2(CO3)34−

21.6

UO22+ + 3H2CO3 UO2(CO3)34- + 6H+

4

UO2OH+

−5.2

UO22+ + H2O UO2OH+ + H+

5

(UO2)2(OH)22+

−5.6

2UO22+ + 2H2O (UO2)2(OH)22+ + 2H+

6

(UO2)3(OH)5+

−15.6

3UO22+ + 5H2O (UO2)3(OH)52+ + 5H+

7

UO2(OH)42−

−33.0

UO22+ + 4H2O UO2(OH)42- + 4H+

8

(UO2)2OH3+

−2.7

2UO22+ + H2O (UO2)2OH3+ + H+

9

(UO2)3(OH)42+

−11.9

3UO22+ + 4H2O (UO2)3(OH)42+ + 4H+

10

(UO2)3(OH)7−

−31.0

3UO22+ + 7H2O (UO2)3(OH)7− + 7H+

11

(UO2)4(OH)7+

−21.9

4UO22+ + 7H2O (UO2)4(OH)7+ + 7H+

12

(UO2)3(CO3)66−

54.0

3UO22+ + 6H2CO32− (UO2)3(CO3)66− + 12H+

13

NH3

−9.2

NH4+ NH3 + H+

14

CO32−

−10.3

H2CO3 <  > CO32− + 2H+

15

HCO3−

−6.4

H2CO3 HCO3- + H+

16

OH−

−14.0

H2O OH− + H+

n

4

KUSHWAHA

|

growth were modified versions of Baar's anaerobic medium for sulfate

ET AL.

3.4 | Analytical methods

reducers (ATCC 1249). Medium 1, used for culturing, was prepared using 18-mΩ deionized water adjusted to pH 7.5 with 1-N HCl, or 1-N NaOH and then sterilized for 30 min in an autoclave. A gas mixture of 80% N2 + 20% CO2 was purged through the hot solution to exclude oxygen. Prior to inoculation, 0.1 ml of filtered-sterilized 5% Fe(NH4)2(SO4)2 was added to 5.0 ml of medium. Medium 1 contained (in mM) CH3CHOHCOONa 31.2, Na-citrate 19.4, CaSO42− and MgSO42− 19.4, K2HPO4 2.9, and Fe(NH4)2(SO4)2 3.5. Preparation of medium one followed published methods (Zhou, Vannela, Hayes, & Rittmann, 2014; Zhou, Vannela, Hyun, Hayes, & Rittmann, 2014).

3.4.1 | Protein measurements Biomass was measured as protein using the Bicinchoninic Acid assay (Smith et al., 1985) with bovine serum albumin (BSA) the standard. The working reagent was prepared by mixing bicinchoninic acid and copper reagent. 0.1 ml of each sample was pipetted into a labelled test tube, and 2.0 ml of the working reagent was added to each tube and mixed well. Tubes were incubated at 37 C temperature for 30 min and assayed for absorbance at 562 nm. A series of dilutions of known concentration were prepared with BSA and assayed to create a calibration curve. Unknown protein concentrations were determined with reference to the calibration curve. Protein was converted to biomass with a ratio of 2 g biomass dry

3.2 | Growth and preparation of cell suspensions

weight per 1 g protein (Dobbernack, Schobert, & Sahm, 1988).

A headspace of 80% N2 + 20% CO2 was used throughout the experiments to maintain an anaerobic atmosphere and control

3.4.2 | High performance liquid chromatography

the pH. Cultures of D. vulgaris were grown in 160-ml glass serum

Concentrations of carboxylic acids (lactate, pyruvate, acetate, and

bottles with 100 ml Medium 1, which had 31 mM Na-lactate as the

formate) were measured using high performance liquid chromatography

electron donor, and 31 mM sulfate as the electron acceptor. The

(HPLC, Model LC-20AT, Shimadzu, Columbia, MD). Approximately one

bottles were sealed with rubber stoppers that were crimped with

milliliter samples, filtered through a 0.2-μm membrane filter (PTFE),

aluminum caps, and the contents were mixed in a rotary shaker at

were taken at fixed time intervals from each experiment. Separation was

200 rpm at 37 °C.

with an Aminex HPX- 87H (Bio-Rad, Hercules, CA) column operated

Once 0.49–0.58 mM of log-phase cells (assayed by protein) were

with 2.5 mM sulfuric acid as the eluent at a flow rate of 0.6 ml/min and at

produced, the serum bottles were allowed to stand (no mixing)

50 C (Parameswaran, Torres, Lee, Rittmann, & Krajmalnik-Brown, 2011).

overnight to separate FeS solids from the cells, which remained suspended. Supernatant from the culture bottles was dispensed into centrifuge tubes inside an anaerobic chamber, and the tubes were

3.4.3 | Gas chromatography

sealed and centrifuged (10,000g, 10 min, 25 °C). Cell-free centrate was

The gas percentages of H2 (if any) in the headspace of the batch bottles

removed carefully, and pellets were washed with a sterile, anoxic 30-

was monitored at the end of experiments (120 hr), by taking samples with

mM NaHCO3 buffer (Zhou, Vannela, Hayes, et al., 2014). Washing and

a gas-tight syringe (SGE 500 lL, Switzerland) using a gas chromatograph

centrifugation were repeated three times before the pellets were used

(GC 2010, Shimadzu) equipped with a thermal conductivity detector. N2

in U(VI)-reduction experiments.

was the carrier gas fed at a constant flow rate of 10 ml/min, and the temperature conditions for injection, column, and detector were 110, 140,

3.3 | U(VI)-reduction batch experiments

and 160 °C, respectively (Parameswaran et al., 2011). Analytical-grade H2 was used for standard calibration curves, analysis was done in duplicate,

Medium 2, used for U(VI)-reduction experiments, was prepared by

and the two results were averaged.

mixing (concentrations in mM) UO2Cl2 1, CH3CHOHCOONa 30, NaHCO3 30, KCl 1, NH4Cl 0.2, K2HPO4 0.01, CaCl2 0.002, FeCl2 0.01, NiCl2 0.01, plus 1 ml/L trace mineral solution (Chung,

3.4.4 | Inductive coupled plasma optical emission analysis of uranium concentration

Nerenberg, & Rittmann, 2006). For all batch experiments, 100 ml of medium 2 was used, test solutions were autoclaved for

Samples were prepared by filtering and diluting the samples 100-fold

sterilization, and transferred to an anaerobic glovebox, where the

before ICP-OES analysis. Standard solutions of uranyl chloride at 10, 20,

bottles were sealed with rubber stoppers and crimped with

50, 100, and 200 mg U L−1 (0.025, 0.05, 0.13, 0.25, and 0.51 mM U) were

aluminum caps. The 30-to-1 mole ratio between lactate and uranyl

used to obtain calibration curves. The standards and samples were

ensured that U(VI) was the limiting substrate for its reduction, since

analyzed by an inductively coupled plasma optical emission spectro-

the stoichiometry is 1.5 mol U(VI) per mol lactate. Uranyl was added

photometer (ICP-OES, Thermo iCAP6300, Waltham, MA) at wavelength

along with the inoculum into assay bottles, and then assay bottles

of 385.9 nm (Guine, 1998).

were moved out of the glovebox, and mixed in a shaker (200 rpm) at 37 °C. Liquid samples were taken over time with a syringe and

3.4.5 | X-ray diffraction analysis

filtered using membrane filters having 0.2 µm pore size (LC + PVDF membrane, Whatman Inc., Haverhill, MA). All analytical methods are

X-ray diffraction patterns of the separated solids were obtained by

detailed in Supporting Information section S1.

holding the solids in place on a quartz plate for exposure to CuKα

KUSHWAHA

ET AL.

|

5

radiation at a wavelength 1.5406 Å. The samples were analyzed at

eq/mol conversion factors. In batch experiments, we experimentally

room temperature over a range of 10–90 ° 2θ with sampling intervals

obtain the yield for each electron sink by closing mass balances on all

of 0.02 ° 2θ and a scanning rate of 0.75 °/min. Scherrer's equation

relevant compounds.

provided an estimate of the crystallite size:

Size of the particles ¼

4 | R E S U L T S AN D D I S CU S S I O N

kλ , B cos θ

ðS25Þ U(VI) bio-reduction experiments resulted in black nano-crystalline

where k = 0.9, λ is the wavelength in Å, B is full width at half maxima in

uraninite (2.8–3.2 nm particle size) that was verified by crystal-

radian based on instrumental correction, and θ is Bragg's angle of

diffraction patterns from the powder X-ray diffraction pattern (Bargar,

diffraction in degrees (Cullity & Stock, 2001).

Bernier-Latmani, Giammar, & Tebo, 2008), presented in Supplementary Figure S1.

3.5 | Electron-equivalent assessment

4.1 | Modeled pH speciation

Microbial synthesis involves partitioning the electron equivalents from the electron donor between cell synthesis and respiration (Rittmann & McCarty, 2001). The fraction of electron equivalents partitioned to cell 0

0

synthesis is termed fs , and the remaining fraction, fe , is goes to respiration. By definition,

0

fs + fe0

= 1. The true yield for biomass

synthesis is proportional to fs0: Y (in gVSS/ gCOD) = fs0 × (1 e−eq/ 8gCOD) × (113 gVSS/ 20 e−eq) when NH4+ is the nitrogen source. We used results from a batch growth experiments to estimate the net biomass yield:

Figure 2 shows the speciation for pH 1–10 with a total carbonate concentration of 30 mM (used in our experiments). All trends in Figure 2 agree with trends in previous speciation modeling of a similar system (Krestou & Panias, 2004). Free UO22+ ion dominates total U(VI) only below pH 3, and it declines steadily up to pH 7, where it is essentially absent. The hydroxide complexes of UO22+ begin forming around pH 3, and they reach a maximum in the pH range 3.5 to pH 7.5, with (UO2)3(OH)5+ being the prominent species. With a total carbonate concentration of 30 mM, the UO2-carbonate complexes become

Y net ¼ 

ΔXa , Δ½Lac

ð17Þ

important for pH above 7, although the predominant carbonate complexed remain UO2(CO3)22−, and UO2(CO3)34− (Supplementary Figure S2).

where ΔXa is the change in biomass concentration (mgVSS/L), VSS is

Figure 3 classifies U(VI) speciation based the level of uranium complexation: Free UO22+, the mono-nuclear complexes (UO2CO3,

volatile suspended solids, and COD is chemical oxygen demand. The metabolism of D. vulgaris is complex because it involves a

UO2(CO3)22−, UO2(CO3)34−, UO2OH+, UO2(OH)42−), and the poly-

number of fermentation and respiratory processes that operate in

nuclear complexes (UO2)3(CO3)36−, (UO2)2(OH)22+, (UO2)3(OH)5+,

parallel. This means that fe0 has to be partitioned among them. The

(UO2)2(OH)3+, (UO2)3(OH)42+, (UO2)3(OH)7−, and (UO2)4(OH)7+). If

partitioning of electrons depends on the synthesis yield, number and

the poly-nuclear species are not bioavailable (Bernhard et al., 1998;

amounts of fermentation products, and the electron equivalents

Croteau et al., 2016; Guillauont et al., 2003; Jones et al., 2015; Lovley

ultimately transferred to a respiratory acceptor. D. vulgaris subdivides

et al., 1991; McCarty, 2007; Natrajan et al., 2014), U(VI) respiration

its electron equivalents removed from lactate into H2, formate,

ought to be slow in the pH range of 5–8, where the poly-nuclear

acetate, pyruvate, U(IV), and biomass. Mathematically, the subdivision

species dominate U(VI) speciation. If poly-nuclear species are

is represented as the sum of all possible sinks for electrons from lactate: 1¼∑nj¼0 ea,i ,

ð18Þ

where ea,i is the electron equivalent fraction captured in respiration and fermentation. The stoichiometry coefficient for the generation of each product is related to fe0 and ea,i, and the corresponding product yield as Yj ¼ f 0e  ea,i ,

ð19Þ

where fe0 is the fraction of electron equivalents from lactate used for respiration, ea,i describes the subdivision of fe0 going to different electron sinks, and Yj is in units of e− eq product/e− eq lactate. The Yj value can be converted to mol product/mol lactate by applying e−

FIGURE 2 pH-dependent speciation of 1 mM total U(VI) in the presence of 30 mM carbonate

6

KUSHWAHA

|

ET AL.

and removal for the pH range 6–10. U(VI) removals were 94–99% at the end of the 5-day experiments, with the removal slightly greater for pH 7. Only pH 5 had significantly slower U(VI) removal, ∼83%. The fact that U(VI) reduction was fastest at pH values near seven supports that the poly-nuclear species were bioavailable; Figure 3 shows that they were the dominant forms of U(VI) at near-neutral pH. That the rate of U(VI) reduction continued to be high at pH > 7 means that the mono-nuclear species also were bioavailable. The lower removal rate for pH 5 is discussed below. The trends in bio-reduction experiments in Figure 4 can be classified into three stages: I (0–18 hr), II (18–48 hr), and III (48– 120 hr). Stage I had a moderate rate of lactate utilization for all pH values, but the key characteristic is the slow rate of U(VI) loss. In contrast, Stage II had a dramatically faster rate of U(VI) loss, and this FIGURE 3 Distribution of U(VI) species reported to be bioavailable or unavailable: Orange line, bioavailable Free UO2+2; Blue line, bioavailable mono-nuclear complexes (UO2CO3, UO2(CO3)22−, UO2(CO3)34−, UO2OH+, UO2(OH)42−); Green line, unavailable poly-nuclear complexes (UO2)3(CO3)36−, (UO2)2(OH)22+, (UO2)3(OH)5+, (UO2)2(OH)3+, (UO2)3(OH)42+, (UO2)3(OH)7−, and (UO2)4(OH)7+)

correlated to faster utilization of lactate, as well as production of acetate and pyruvate. In Stage III, U(VI) reduction already was nearly complete, and fermentation of lactate to acetate and pyruvate stopped. Fermentation of lactate can proceed only when the main products (acetate and H2 here) are consumed enough to make the fermentation reaction thermodynamically feasible. At the end of the experiment, the acetate concentration had accumulated, which means that the H2 concentration had to be very small to allow more

bioavailable, U(VI) reduction should be largely independent of pH as

fermentation. Thus, unmeasurable H2 could stop fermentation of

long as an extreme pH is not itself inhibitory (Wall, 2007).

lactate based on thermodynamics. The experiment at pH 5 was unique in that it had noticeably slower U(VI) loss (Figure 4a) and faster acetate accumulation (Figure 4b), even

4.2 | Effect of pH on bio-availability of uranium-complex species

though lactate loss (Figure 4d), and pyruvate accumulation (Figure 4c) were not affected. The inhibition of U(VI) respiration could have been

The effect of pH was studied by varying the initial pH from 5 to 10,

due to one or a combination of three possibilities: (i) free UO22+ was

presented in Supplementary Figure S3. The pH was measured at the

less bioavailable; (ii) free UO22+ was inhibitory to D. vulgaris respiration;

end of an experiment, and it decreased by less than 0.2 unit from the

and (iii) the low pH itself was an inhibitor of respiration (Wall, 2007).

initial

NaHCO3.

The most likely cause is that low pH was inhibitory. Wall (2007) studied

Figure 4a shows that pH had only a small impact of U(VI) reduction

D. vulgaris Hildenborough and found that low pH led to a prolonged lag

value

FIGURE 4

due

to

the

buffering

from

30 mM

Concentrations (mM) of (a) U(VI), (b) acetate, (c) pyruvate, and (d) lactate for batch tests at pH values

KUSHWAHA

ET AL.

|

7

FIGURE 5 Lactate, biomass, and product distributions in mM (panel a) and me- eq/L (panel b) during batch experiments for U(VI) bioreduction at pH 7. Each symbol is the average of six replicate experiments. The standard deviations are smaller than the size of the symbols and decreased yield, which were attributed to the accumulation of

electrons to lactate consumed, 1.02/4.94 = 0.21 me− eq missing per e−

acetate, which, by shuttling protons across the membrane, dissipates

eq lactate consumed, supports this view, as it is the ratio between H2 or

the proton gradient, and proton motive force.

formate electrons and lactate electrons in Figure 1; we discuss this ratio more below. The formate concentration remained below the method

4.3 | Bio-reduction kinetics and end-product distribution

detection limit (