This is an open access article published under a Creative Commons Attribution (CC-BY)

License, which permits unrestricted use, distribution and reproduction in any medium,
provided the author and source are cited.

Article

pubs.acs.org/est

Biostimulation by Glycerol Phosphate to Precipitate Recalcitrant

Uranium(IV) Phosphate
Laura Newsome,*, Katherine Morris, Divyesh Trivedi, Alastair Bewsher, and Jonathan R Lloyd,

Williamson Research Centre and Research Centre for Radwaste Disposal, School of Earth, Atmospheric and Environmental Sciences,
The University of Manchester, Williamson Building, Oxford Road, Manchester M13 9PL, U.K.

National Nuclear Laboratory, Chadwick House, Birchwood, Warrington WA3 6AE, U.K.
*
S Supporting Information

ABSTRACT: Stimulating the microbial reduction of aqueous

uranium(VI) to insoluble U(IV) via electron donor addition has
been proposed as a strategy to remediate uranium-contaminated
groundwater in situ. However, concerns have been raised regarding
the longevity of microbially precipitated U(IV) in the subsurface,
particularly given that it may become remobilized if the conditions
change to become oxidizing. An alternative mechanism is to
stimulate the precipitation of poorly soluble uranium phosphates
via the addition of an organophosphate and promote the
development of reducing conditions. Here, we selected a sediment
sample from a U.K. nuclear site and stimulated the microbial
community with glycerol phosphate under anaerobic conditions to
assess whether uranium phosphate precipitation was a viable bioremediation strategy. Results showed that U(VI) was rapidly
removed from solution and precipitated as a reduced crystalline U(IV) phosphate mineral similar to ningyoite. This mineral was
considerably more recalcitrant to oxidative remobilization than the products of microbial U(VI) reduction. Bacteria closely
related to Pelosinus species may have played a key role in uranium removal in these experiments. This work has implications for
the stewardship of uranium-contaminated groundwater, with the formation of U(IV) phosphates potentially oering a more
eective strategy for maintaining low concentrations of uranium in groundwater over long time periods.

INTRODUCTION
Uranium contamination in groundwater poses a signicant
environmental problem at current and former uranium mining
and nuclear facilities. Bacterially mediated precipitation of
uranium minerals can be used as a strategy to remove aqueous
uranium from groundwater in situ and therefore prevent its
uncontrolled migration and dispersal.14 Potential approaches
to stimulating microbial activity include the addition of an
electron donor to stimulate microbial U(VI) reduction, leading
to the precipitation of insoluble U(IV) minerals,5 or the use of
an organophosphate compound that bacteria break down to
release orthophosphate and, consequently, cause the precip-
itation of insoluble uranium phosphate minerals.6 To date,
most research has focused on microbial U(VI) reduction;1,2
however, biogenic U(IV) has been found to be susceptible to
oxidative remobilization after exposure to oxygen or nitrate710
and therefore may not be an ideal end-product for a long-term
in situ remediation strategy. The precipitation of uranium
phosphate minerals is a promising alternative because U(VI)
phosphates are largely insoluble and not sensitive to oxidative
changes, and their longevity has been demonstrated in natural
analogue sites.1114 Indeed, the presence of phosphate has been
found to limit the mobility of U(VI) in uranium-contaminated
systems.15,16 Most work on uranium phosphate biomineraliza-
tion has focused on aerobic systems that form U(VI)
phosphates. Interestingly, strategies to produce reduced
uranium(IV) phosphate minerals (for example, ningyoite
[CaU(PO4)2.2H2O]) have not been studied in detail, although
reduction coupled to phosphate biomineralization has signi-
cant merit when considering treatment for radionuclide-
contaminated groundwaters. Uranium(IV) phosphate would
be a desirable end product because of its very low solubility
even in comparison to that of U(VI) phosphates.17 Microbial
precipitation of U(IV) phosphate has only been observed in
one previous study,18 with putative identication of either
monomeric U(IV) complexed to phosphate19 or U(IV)
phosphate19,20 also reported. Indeed, it has been proposed
that some ningyoite ore deposits are of biogenic origin,
suggesting some precedent for microbially mediated ningyoite
formation in the natural environment.21
Stimulating the precipitation of U-containing phosphate
minerals in the subsurface is not straightforward because adding
inorganic phosphate can lead to metal phosphate precipitation
and localized clogging at injection wells.22 To prevent this,
phosphorus-containing compounds that are slowly hydrolyzed
Received: April 23, 2015
Revised: August 20, 2015
Accepted: August 20, 2015
Published: August 20, 2015

2015 American Chemical Society 11070 DOI: 10.1021/acs.est.5b02042

Environ. Sci. Technol. 2015, 49, 1107011078
Environmental Science & Technology
or biodegraded may be used, such as glycerol phosphate, long-
chain polyphosphates, or phytate from plant matter.6,23,24
Typically, glycerol phosphate has been used as both a carbon
and phosphate source for aerobic microorganisms.6,25 It should
also be noted that under phosphate-limited conditions, it is
possible that bacteria can mine uranyl phosphates, releasing
uranium to the solution,19,26,27 which may potentially aect
their longevity in the subsurface.
A number of bacterial species have been shown to precipitate
uranyl phosphates when supplied with glycerol phosphate,
including Serratia species,6,28 Caulobacter crescentus,29 environ-
mental isolates from the U.S. Department of Energy (DOE)
Oak Ridge site closely related to Bacillus and Rahnella
species25,30 and Aeromonas hydrophila31 as well as sediment
microbial consortia.32,33 The mechanism of glycerol phosphate
biodegradation involves the expression of a phosphatase
enzyme, which catalyzes the hydrolytic cleavage of the CP
bond and so releases inorganic phosphate to solution. The
released phosphate rapidly precipitates abiotically with the
U(VI) that is present in solution as uranyl complexes, forming
uranyl phosphate minerals.6,34 Phosphatase activity is a
characteristic common to many microorganisms because it
ensures that they are able to obtain phosphorus, an essential
nutrient, in its preferred form (orthophosphate). This suggests
that many more species may be capable of uranium phosphate
biomineralization via this mechanism. The composition of the
microbial community of uranium-contaminated sediments from
the U.S. DOE Oak Ridge site following stimulation with
glycerol phosphate under oxic conditions has been investigated
using the PhyloChip array.35 Over 2000 archaeal and bacterial
taxa were detected; the authors highlighted increases in the
relative abundances of hydrogen-dependent methanogens as
well as bacteria closely related to species known to denitrify,
exhibit phytase activity, or accumulate polyphosphate. This
suggests that glycerol phosphate amendment stimulated not
only microbes involved in phosphorus cycling but also many
other species, leading to complex changes in microbial
community dynamics.
Precipitation of uranium phosphates using extant microbial
communities has been investigated previously using sediments
from the U.S. DOE Oak Ridge site. Stimulation with glycerol
phosphate lead to the precipitation of uranyl phosphates over a
range of conditions, although uranium removal was dominated
by sorption eects.32,33 To further explore the interplay
between glycerol phosphate biostimulation and U phosphate
biomineralization, we conducted microcosm experiments using
sediment from a U.K. nuclear site under anaerobic conditions.
Throughout, our focus has been on uranium speciation and fate
coupled to sediment microbial community dynamics to gain
insight into the potential for targeted U(IV) phosphate
biomineralization in shallow subsurface environments.
MATERIALS AND METHODS
Solubility tests were carried out to assess whether uranium(VI)
remained in solution in articial groundwater containing the
organic phosphorus-containing compounds glycerol phosphate,
phytic acid, and sodium phytate. U(VI) as uranyl chloride at
0.05, 0.1, and 0.2 mM was added to a sterile articial
groundwater,10 containing the following in mM: K+, 0.089;
Na+, 3.37; Ca2+, 1.69; Mg2+, 0.795; Cl, 1.06; HCO3, 2.88;
NO3, 0.332; CO32, 1.69; SO42, 0.39, with 1 or 10 mM
glycerol phosphate or phytic acid. U(VI) in solution was
determined over a 12 day period by separating the supernatant
11071
Article
by centrifugation (16 200g for 5 min) then measuring aqueous
U(VI) via spectrophotometry.36 Because the addition of phytic
acid caused the pH to decrease from 7.8 to less than 3, a neutral
solution of 50 mM sodium phytate was prepared by adding
sodium hydroxide to phytic acid. Sodium phytate (1 or 5 mM)
was added to the articial groundwater with 0.05 or 0.5 mM
U(VI), and the solubility was monitored over 26 h.
To determine whether the microbial community present in
Sellaeld sediment could precipitate uranium phosphate
biominerals, microcosm experiments were set up containing
10 g of gravelly sand sediment (previously characterized as
RB2737), 100 mL of articial groundwater, 10 mM glycerol
phosphate (Sigma G6501), and 0.05 mM U(VI) in glass serum
bottles in triplicate. The headspace was degassed with argon,
and the microcosms were incubated in the dark at room
temperature with periodic shaking. Electron donor free controls
had no added glycerol phosphate. To investigate how the
presence of phosphate may aect U(VI) removal, an additional
set of microcosm experiments were set up containing 10 mM
glycerol as the electron donor. Sterile controls were created by
autoclaving and were sacricially sampled.
Aliquots from each biostimulation experiment were periodi-
cally withdrawn for biogeochemical monitoring. Sediment
slurry was added to 0.5 N HCl or 0.5 N hydroxylamineHCl
for digestion before analysis for Fe(II) and total bioavailable Fe
by the ferrozine assay.38 Supernatant was separated by
centrifugation (16 200g, 5 min) and analyzed for U(VI)36 and
nitrite39 by spectrophotometry and for Eh and pH. Surplus
supernatant was frozen (20 C) and stored before ion
chromatography analysis for phosphate and nitrate (Dionex
ICS 5000 with an AS18 2 mm ion exchange column at 0.25
mL/min) and for glycerol phosphate and volatile fatty acids by
ion chromatography (Dionex ICS 5000 with an AS11HC 0.4
mm high-capacity ion exchange column at 0.015 mL/min).
Sediment composition was analyzed using X-ray diraction
before and after biostimulation with glycerol phosphate. Scans
were performed using a Bruker D8 Advance tted with a Gobel
Mirror and a Lynxeye detector. Electron microscopy was
carried out on microbially reduced sediment using a Phillips
XL30 ESEM-FEG operated under low-vacuum conditions (0.5
Torr chamber pressure) at 20 kV accelerating voltage.
Elemental distributions were estimated using an EDAX Genesis
energy dispersive spectroscopy system.
An additional batch of microcosms was set up to analyze
uranium speciation and coordination via X-ray absorption
spectroscopy (XAS). These contained 0.2 mM U(VI) to
generate approximately 500 ppm U on sediments after reaction.
After conrming that U(VI) had been removed from the
solution, we stored the sediment at 80 C prior to analysis.
The sample was prepared under anaerobic conditions by
separating the sediment from the slurry by centrifugation
(5000g, 20 min) and then adding the wet paste to a cryovial
that was stored at 80 C under argon. XAS was carried out at
the Diamond Light Source, Harwell, U.K. on Beamline B18
using a liquid nitrogen cryostat; uranium LIII-edge spectra were
collected in uorescence mode using a 9 or 36 element Ge
detector.40 ATHENA41 was used to calibrate, background-
subtract, and normalize X-ray absorption near edge-structure
(XANES) spectra. ARTEMIS41 was used to t extended X-ray
absorption ne structure (EXAFS) spectra, and shells were
included in the nal t if they made a statistically signicant
change to the model as conrmed by the F-test.42
DOI: 10.1021/acs.est.5b02042
Environ. Sci. Technol. 2015, 49, 1107011078
Environmental Science & Technology
Figure 1. Geochemical changes observed following the biostimulation of Sellaeld sediments with glycerol phosphate (black lines) or glycerol (gray
lines). Controls contained no added donor (black dashed lines) or were sterilized by autoclaving (gray dashed lines).
The long-term stability of the precipitated uranium
phosphate was assessed via exposure to oxidants to represent
potential changes to the ambient subsurface conditions. Nitrate,
a common contaminant at nuclear sites, was added initially at 3
mM to represent the highest concentration reported in
Sellaeld groundwater,43 and the bottles were periodically
agitated throughout the course of the experiment. After 43
days, additional 30 mM nitrate was added to provide excess
oxidant to the system. To represent oxygen ingress, we
transferred sediment slurry into conical asks at a 1:3 slurry-to-
headspace ratio and aerated it periodically by gentle mixing
under ambient atmospheric conditions. Geochemical monitor-
ing of U, Fe, pH, and Eh was carried out throughout these
reoxidation experiments.
A pyrosequencing methodology was used to determine
changes in the microbial community during the glycerol
phosphate biostimulation experiment at selected time points.
Full details are documented previously.37 In brief, DNA was
extracted from sediment pellets collected on day 0, day 4, day
14, and day 92 using a PowerSoil DNA Isolation Kit (MO BIO
Laboratories Inc., Carlsbad, CA). The 16S23S rRNA
intergenic spacer region was amplied using primers ITSF
and ITSReub, and then electrophoresis (in Trisacetate
EDTA gel) was used to separate the polymerase chain reaction
(PCR) products. PCR products were then cleaned and
quantied before sequencing of the 16S rRNA gene (Roche
454 Life Sciences GS Junior system). Details of the sequence
processing are provided in Table S1. The 454 pyrosequencing
reads were analyzed using Qiime 1.6.0,44 taxonomic classi-
cation was performed using the Ribosomal Database Project
(http://rdp.cme.msu.edu), and the closest GenBank matches
were identied by Blastn nucleotide search (http://blast.ncbi.
nlm.nih.gov). A sample from day 92 of the glycerol-amended
experiment was also analyzed using the same methodology.
11072
Article
RESULTS AND DISCUSSION
Uranium Solubility. To help determine the appropriate
biostimulation method for the Sellaeld sediment, we added a
range of phosphate-delivery compounds to articial ground-
water containing uranium to assess its solubility. Uranium at 0.2
mM was soluble in articial groundwater in the presence of 1
mM and 10 mM glycerol phosphate. However, it was not
soluble in articial groundwater containing phytic acid or
sodium phytate at any of the concentrations tested. This rapid
interaction between U(VI) and phytic acid or phytate suggests
that in situ application of phytate is unlikely to be suitable for
bioremediation due to a lack of dispersivity into a contaminated
aquifer.
Biostimulation Geochemistry. On the basis of the initial
screening experiments, glycerol phosphate (and glycerol) were
added to sediment microcosm experiments to stimulate
microbial uranium precipitation. With glycerol phosphate,
U(VI) was removed rapidly from solution, with near-complete
removal by day 14 (Figure 1). Phosphate concentrations in
solution initially increased rapidly from 0 to 1.4 mM after 1 h of
incubation (Figure 1) but then decreased until day 4,
presumably when the rate of metal phosphate precipitation
exceeded the rate of biodegradation of glycerol phosphate. The
concentrations of U(VI) and phosphate both remained
constant between days 4 and 7, perhaps suggesting that in
this early stage, the rates of removal and production may be
controlled by phosphate geochemistry rather than the redox
geochemistry of iron or other related processes.
Additions of both glycerol phosphate and glycerol stimulated
a cascade of terminal electron accepting processes (Figure 1)
with very rapid nitrate reduction, followed by Fe(III) reduction
and sulfate removal. With glycerol phosphate, approximately
half of the U(VI) had already been removed from the solution
before Fe(III) reduction commenced, while with glycerol,
DOI: 10.1021/acs.est.5b02042
Environ. Sci. Technol. 2015, 49, 1107011078
Environmental Science & Technology
uranium removal occurred concomitantly with Fe(III)
reduction (day 7). These enhanced removal rates in the
glycerol phosphate system are presumably due to interactions
with the phosphate that was released following the microbial
metabolism of glycerol phosphate. Glycerol can be oxidized as
an electron donor or fermented and, consequently, may
generate a wide variety of breakdown products. The
concentrations of volatile fatty acids (VFAs) detected using
ion chromatography showed that with both electron donors,
acetate, propionate, and formate were generated (Figure S1),
although their proportions were dierent in each system.
Propionate was the dominant VFA breakdown product from
glycerol phosphate, while formate predominated following
glycerol stimulation. This highlights the complexity of glycerol
substrate biodegradation in these systems and suggests that the
presence of phosphate may stimulate dierent microbial
processes. Interestingly, glycerol phosphate (and glycerol)
addition stimulated the removal of sulfate from solution,
possibly due to sulfate reduction. To the best of our knowledge,
this is the rst time this has been observed for these electron-
donor systems and, for example, is in contrast to studies at the
U.S. DOE Oak Ridge site,33 although further research would be
required to conrm the precise mechanism of sulfate removal.
Comparison of the rates of uranium removal and Fe(III)
reduction between experiments containing the same sediment
and articial groundwater stimulated with glycerol phosphate,
glycerol, or an acetate/lactate mix37 revealed that U(VI)
removal was fastest with glycerol phosphate (Figure S2), while
rates with glycerol were remarkably similar to those with
acetate/lactate, perhaps suggesting the involvement of similar
metabolic pathways for these bioreduction-only stimulations.
Fe(III) reduction was most rapid in the presence of acetate/
lactate and slowest with glycerol (Figure S2), likely due to the
dierent substrates stimulating dierent microorganisms with
dierent metabolic pathways.
Clearly, adding glycerol phosphate stimulated microbial
activity, leading to the precipitation of uranium from solution.
Given that U(VI) and phosphate concentrations show similar
trends during the early stages of this experiment (Figure 1), this
suggests that microbial phosphatase activity may have played a
role in U(VI) removal, mediating the release of phosphate from
glycerol phosphate. Furthermore, 50% of U(VI) was removed
from solution before Fe(II) concentrations increased (Figure
1); this implies that a biogeochemical process other than
microbial U(VI) reduction by Fe(III)-reducing bacteria was
partly responsible for U(VI) removal in these experiments. This
theory is also supported by the faster rate of U(VI) removal,
with glycerol phosphate compared to that in experiments that
solely stimulated microbial U(VI) reduction, e.g., with glycerol
and acetate/lactate (Figure S2).
Mineralogy. XRD analysis found little dierence in the bulk
sediment pre- and post-biostimulation with glycerol phosphate.
Sediment mineralogy was composed mostly of silicate minerals
including quartz, feldspar, mica, and chlorite, with some calcite
also present. Pyrite made a minor contribution to the post-
biostimulation spectrum. Uranium concentrations were too low
to be detected with XRD, even in the high-uranium XAS
sediment, which contained up to 0.05% wt U.
Imaging of biostimulated sediment using ESEM (Figure S3)
revealed a heterogeneous substrate composed of predominantly
silicon, oxygen, calcium, aluminum, and carbon, with some
contribution from Na, Mg, P, S, and K (elements that are
common in silicate minerals). Numerous bright spots were
11073
Article
identied in backscatter mode, indicative of elements of higher
atomic weight. EDAX spectra conrmed that the majority of
these areas were rich in iron, although titanium and uranium
were also observed at selected sites. An area rich in uranium, of
approximately 1 m2 in size, was identied in backscatter mode
(Figure S4). Elemental mapping of this localized region
revealed the uranium hot-spot to be closely correlated with
phosphorus and located in an area generally high in iron and
titanium, although uranium did not appear to be correlated
with these elements.
X-ray Absorption Spectroscopy. To identify the
speciation of uranium in sediments stimulated with glycerol
phosphate, we collected XANES spectra and compared them to
those of U(IV) as uraninite and U(VI) as uranyl phosphate as
standards.28 Both the edge position of the sample and the shape
of the spectra were most similar to that of the U(IV) standard
(Figure 2). The position of the rst derivative, at 17170 eV, was
the same as that of uraninite and lower than that of uranyl
phosphate (17173 eV). Linear combination tting of XANES
spectra indicated that 100% of the uranium present in the
glycerol phosphate biostimulated sediment was present as
U(IV).
Figure 2. Uranium LIII-edge XANES spectra for sediment biostimu-
lated with glycerol phosphate (GP) compared to reference spectra for
uranium(IV) and uranium(VI) minerals.28
The local coordination environment of U(IV) in the glycerol
phosphate stimulated system was investigated by examining the
EXAFS spectra (Figure 3). Given the presence of U(IV) and
the likely presence of phosphate, we used the crystal structure
of the U(IV) phosphate mineral ningyoite to t the spectra
(Figure S5).45 An excellent t was obtained with the U(IV) in a
ningyoite-type environment, with the central U atom in 8-fold
coordination with oxygen (with four O atoms at 2.28 and
four O atoms at 2.44 ), and with two bidentate P atoms at
3.14 and four monodentate P atoms at 3.69 (Figure 3 and
Table 1). Given the presence of Ca in the ningyoite crystal
structure45 and the fact that calcite was detected in the XRD
analysis of this sediment,37 we attempted to include Ca in the
EXAFS t. Adding a shell of Ca atoms at 3.85 improved the
EXAFS tting parameters but not with statistical signicance,
and it also lead to relatively large errors on the DebyeWaller
factor (which is not surprising given the distance of Ca from
the central absorber and the fact that Ca is a relatively weak
scatterer); therefore, it was not included in the nal t. This
U(IV) phosphate coordination environment is very similar to
that of the ningyoite crystal structure, with only small variations
(0.010.06 ) compared to the published bond distances. In
contrast, tting to a model monomeric U(IV) phosphate phase
from our past work46 on this sediment (Figure S6) resulted in a
poor t with no resolution of key features in the Fourier
DOI: 10.1021/acs.est.5b02042
Environ. Sci. Technol. 2015, 49, 1107011078
Environmental Science & Technology Article

the reduction of aqueous U(VI) is integral to the formation of

U(IV) phosphate. If this were the case, it could explain the lack
of formation of U(IV) phosphate in the previous study, as the
U(VI) had rapidly sorbed to sediments well before phosphate
release to solution was measured (after 14 days) and Fe(III)-
reducing conditions developed (after 5 days).33 Clearly,
additional work is further required to elucidate the mechanism
of U(IV) phosphate formation in sediment systems. The
potential for stimulating microbial precipitation of recalcitrant
actinide phosphate minerals3 in situ has broader implications
for the management of radioactive legacy materials.

Figure 3. k3-weighted EXAFS data and non-phase shift corrected

Fourier transform of EXAFS data for glycerol phosphate stimulated
sediments.

transform between 2.5 and 3.5 (Figure S6). Additionally, our

spectrum did not match the published monomeric U(IV)
phosphate spectra,47 and overall, this suggested that the
biostimulation product was a U(IV) phosphate with a
ningyoite-like crystal structure. A comparison of our EXAFS
data with those for chemogenic crystalline U(IV) phosphate
that were previously published47 revealed remarkable similar-
ities; similar oscillations were observed in the EXAFS, and the
main peaks in the Fourier transforms for both data sets were
plotted at around 1.7, 2.7, 3.2, and 4.0 .
Our results dier from a previous study33 in which U(VI)
phosphates were formed in anaerobic sediment microcosms Figure 4. Comparison of U(VI) remobilization following the exposure
stimulated by glycerol phosphate. A key dierence is that of U(IV) phosphate and monomeric U(IV) to nitrate and air.
U(VI) rapidly sorbed to the sediments used in this earlier Reoxidation experiments were conducted under exactly the same
study, with just 0.7% (2 M) of the added U(VI) remaining in conditions with the same setup 90 days postbiostimulation with either
solution. This suggests that the observed formation of U(IV) glycerol phosphate to precipitate U(IV) phosphate or acetate/lactate
to stimulate monomeric U(IV) precipitation.46
phosphate in the current study could be linked to the
concomitant reductive precipitation of uranium as U(IV) and
the release of inorganic phosphate. Supporting evidence for this Recalcitrance of the U(IV) Phosphate Biomineral. The
theory may be found in the formation of a ningyoite-like long-term stability of the U(IV) phosphate biomineral was
mineral during U(VI)aq reduction experiments with Shewanella assessed in the context of reoxidation caused by nitrate and
putrefaciens CN32, in which the medium contained sodium oxygen ingress into the shallow subsurface at a nuclear site. The
orthophosphate.20 Another study observed the formation of addition of 3 mM nitrate did not cause signicant quantities of
ningyoite-like U(IV) from hydrogen uranyl phosphate (HUP) uranium to be released to the solution (Figure 4). Some Fe(II)
in the presence of Geobacter sulfurreducens PCA linked to the was initially oxidized to Fe(III) (data not shown), likely due to
dissolution of the HUP.19 This again supports the theory that the microbial community using the residual electron donor to

Table 1. Details of EXAFS Fit Parameters for the U(IV) Phosphate Biominerala
sample path coordination number atomic distance () DebyeWaller factor 2 (2) condence level of adding shell ()b
U(IV) phosphate biomineral O eq 4 2.28 (1) 0.004 (1) -
O eq 4 2.44 (1) 0.004 (1) 1.00c
P bidentate 2 3.14 (1) 0.006 (1) 0.99
P monodentate 4 3.69 (2) 0.013 (3) 0.96

a
Amplitude factor (S02) was xed at 1.0 for each sample. Numbers in parentheses are the SD on the last decimal place. Energy shift E0 from
calculated Fermi level (eV): 5.14 0.86. Reduced 2: 54.6. R goodness of t factor: 0.025. Number of variables: 9. Number of independent points:
20. bF-test results; > 0.95 statistically improves the t with 2 condence. cThis value is for splitting the shell of 8 equatorial oxygen atoms into
two shells, each containing 4 O atoms, after adding the P shells.

11074 DOI: 10.1021/acs.est.5b02042

Environ. Sci. Technol. 2015, 49, 1107011078
Environmental Science & Technology
reduce nitrate to nitrite, which could then oxidize Fe(II) to
Fe(III).48 This increase in Fe(III) was transient; by day 14, it
had been reduced back to Fe(II), probably as the added nitrate
had been metabolized by the microbial community as observed
previously.46 Consequently, to provide a stoichiometric excess
of oxidant, 30 mM nitrate was added on day 43. This reoxidized
all of the Fe(II) to Fe(III), but just 3% of the uranium was
released to the solution (Figure 4). A comparison of the results
of nitrate-mediated reoxidation of U(IV) phosphate and
monomeric U(IV)46 (Figure 4) clearly shows the recalcitrance
of U(IV) phosphate to remobilization under oxidizing
conditions. Other previous experiments where the products
of microbial U(VI) reduction were reoxidized during nitrate
additions showed variable results, from 3%8 to 97%,49 with the
presence of residual electron donor potentially being an
important factor in determining the levels of U(IV)
reoxidation.46
Exposure of the glycerol phosphate stimulated sediment to
air caused around 40% of the uranium to be remobilized after
90 days (Figure 4). Compared to previous data under similar
experimental conditions but stimulated with acetate/lactate and
containing monomeric U(IV)46 and where 100% was
remobilized after 60 days, the U(IV) phosphate biomineral
was considerably more resistant to oxidative remobilization.
Previous data published for the products of microbial U(IV)
reduction generally also showed U(IV) reoxidation following
exposure to oxygen or oxygenated groundwater.8,49,50 In this
study, analysis of the speciation and coordination of uranium
was possible by XAS in a parallel reoxidized sample, for which
the geochemical data indicated that 20% of the U(IV) had
been reoxidized (Figure 5). The spectra appeared remarkably
similar to that of the original U(IV) phosphate mineral (Figure
S7), and the best t was achieved using the same tting
parameters (Table S2), indicating the presence of a refractory
U(IV) phosphate phase.
Figure 5. k3-weighted EXAFS data and non-phase shift corrected
Fourier transform of EXAFS data for glycerol phosphate stimulated
sediments post-oxygen reoxidation. A good t was achieved for the
reoxidized sample using the same model as for the U(IV) phosphate
biomineral.
11075
Article
In summary, biogenic U(IV) phosphate was considerably
more recalcitrant to oxidative remobilization via exposure to 30
mM nitrate or oxygen in comparison to sediments when
bioreduction was stimulated with just an electron donor and
specically in similar sediment systems stimulated by acetate/
lactate where monomeric U(IV) had formed. These exper-
imental results suggest that targeted phosphate precipitation
may be a long-term treatment option for uranium-contami-
nated groundwaters.
Molecular Ecology. We performed 16S rRNA pyrose-
quencing to investigate the changes in the composition of the
microbial community during glycerol phosphate biostimulation
and to compare the changes to those following glycerol
biostimulation. At the phylum and class levels, the results were
dominated by Proteobacteria (, , and ) and Firmicutes
(Figure S8). At the start of the experiment (day 0, samples
taken 1 h after the experiment was set up), a relatively diverse
microbial community was present (Table S1 and Figure S9),
with Pseudomonas species making up four of the ve most
abundant operational taxonomic units (OTUs, Figure 6 and
Table S3).
After 4 days, species diversity had decreased (Table S1 and
Figure S9), and the microbial community was dominated by
those aliated with Pseudomonas species (Table S3); members
of the order Pseudomondales comprised 76% of the microbial
community (Figure 6). The most abundant OTU was closely
related to Pseudomonas mandelii (100% ID similarity) and
comprised 58% of all clones. P. mandelli is a facultative
anaerobe known to denitrify and has been previously used to
study the eects of denitrication on nitrate-rich radioactive
wastes in a geological disposal facility.51 Other bacteria in the
day 4 microbial community were closely related to Pseudomonas
migulae, known to be able to denitrify and x nitrogen, and a
Hydrogenophaga species that had been isolated from a uranium-
contaminated mine and was related to hydrogen-oxidizing and
nitrogen-metabolizing species (Table S3). These results are
similar to those of a previous study using sediments stimulated
with glycerol phosphate, which also identied increases in
species involved in hydrogen metabolism and denitrication.52
A signicant shift in the relative proportions of species had
occurred in the glycerol phosphate stimulated microbial
community between day 4 and day 14 (Figure 6), with large
increases in bacteria closely related to Pelosinus UFO1 (95% ID
similarity) comprising 33% of bacteria identied at the genus
level. Pelosinus species are fermentative bacteria and are known
to be able to reduce metals such as Cr, Fe, and U,53,54 and it is
noteworthy that this increase coincides with the start of Fe(III)
reduction in our experiments. The presence of Pelosinus UFO1
is particularly of interest because it was originally isolated from
pristine sediments at the U.S. DOE Oak Ridge site and has
been shown to be able to remove uranium from solution by
multiple mechanisms, including by reduction to U(IV) and via
precipitation as U(VI) phosphates.52 Additional support for its
potential role in uranium bioremediation comes from closely
related species also being present in uranium- and heavy-metal-
contaminated sediments and found in soils amended with
acetate to stimulate U(VI) bioremediation (Table S3).
By day 92, a more diverse microbial community was detected
following glycerol phosphate biostimulation (Table S1 and
Figure S9), with the most abundant OTUs including species
closely related to Rhizobium, Aztobacter, Magnetospirillum, and a
Bacteroidales species (Table S3). In comparison, 92 days after
glycerol biostimulation, the most abundant OTUs were
DOI: 10.1021/acs.est.5b02042
Environ. Sci. Technol. 2015, 49, 1107011078
Environmental Science & Technology
Figure 6. Changes in microbial ecology during glycerol phosphate biostimulation, plotted at the order level. The microbial community rapidly
became dominated by Pseudomonadales after biostimulation. Following this, Clostridales (Pelosinus) increased, and then after 92 days, a more
diverse community was detected. The glycerol-stimulated sediments (right column) appeared markedly dierent to the glycerol phosphate
stimulated sediments after 92 days, suggesting that the presence of phosphate may inuence the microbial community composition.
composed of species closely related to Hydrogenophaga, a
Bacteriodales species, and other uncultured species from
Bacteriodales and the family Gracilibacteraceae (Table S3).
The presence of phosphate in the glycerol phosphate
biostimulated sediments must account for these dierences in
the microbial community composition, given that no other
variables were changed. Following from this, a dierent
microbial community with dierent metabolic characteristics
may explain why propionate was the dominant VFA detected
following glycerol phosphate biostimulation and formate
predominated with glycerol (Figure S1).
These results highlight the dynamic changes that occur
following biostimulation that are not observed when just the
microbial community present at the end of an experiment is
examined. The large increase in bacteria closely related to
Pelosinus around day 14 could account for some uranium
removal. Questions remain about the role played by
Pseudomonas species in these anaerobic systems; although
they are known to denitrify, our results showed that nitrate
reduction was essentially complete after 24 h, and nitrite
reduction ended between day 1 and day 4. At day 14,
Pseudomonadales species comprised 50% of the microbial
community at the order level (Figure 6), so it is unclear how
and whether they were functioning under these anaerobic
conditions, and it would be of interest to investigate this
further.
In conclusion, stimulating sediments with glycerol phosphate
lead to the formation of U(IV) phosphate biominerals, which
were more recalcitrant to oxidative remobilization than the
products of microbial U(IV) reduction alone. Phosphate played
a key role in the formation of a genetically distinct microbial
community that generated dierent organic breakdown
products. This work has implications for the long-term
management of uranium-contaminated groundwater, where
targeted bioprecipitation of phosphate phases coupled with
bioreduction has the potential to treat a wide range of
radionuclides in the subsurface.
11076
Article
*
ASSOCIATED CONTENT
S Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.est.5b02042.
Additional results include tables of molecular ecology
sequence details, EXAFS t parameters, and phylogenetic
results; and gures showing volatile fatty acid generation
results, rate comparisons of U(VI) removal and Fe(III)
reduction, ESEM images, crystal structures, EXAFS data,
bacterial phylogenetic diversity, and sample diversity
rarefaction curves. (PDF)
AUTHOR INFORMATION
Corresponding Author
*Phone: +44 (0)161 275 0309; e-mail: laura.newsome@
manchester.ac.uk.
Notes
The authors declare no competing nancial interest.
ACKNOWLEDGMENTS
We thank Christopher Boothman and Athanasios Rizoulis
(University of Manchester) for assistance with sample
preparation and processing of pyrosequencing data and Jon
Fellowes (University of Manchester) for help with ESEM and
elemental mapping. Beamtime at beamline B18 was funded by
grants SP8941-2 and SP10163-2 from the Diamond Light
Source. We acknowledge nancial support from the Nuclear
Decommissioning Authority via a PhD student bursary,
managed by the National Nuclear Laboratory. J.R.L. acknowl-
edges the support of the Royal Society via an Industry
Fellowship. We also acknowledge nancial support from NERC
via the BIGRAD consortium (NE/H007768/1).
REFERENCES
(1) Williams, K. H.; Bargar, J. R.; Lloyd, J. R.; Lovley, D. R.
Bioremediation of uranium-contaminated groundwater: a systems
approach to subsurface biogeochemistry. Curr. Opin. Biotechnol. 2013,
24, 489497.
DOI: 10.1021/acs.est.5b02042
Environ. Sci. Technol. 2015, 49, 1107011078
Environmental Science & Technology
(2) Newsome, L.; Morris, K.; Lloyd, J. R. The biogeochemistry and
bioremediation of uranium and other priority radionuclides. Chem.
Geol. 2014, 363, 164184.
(3) Lloyd, J. R.; Macaskie, L. E. Bioremediation of radionuclide-
containing wastewaters. In Environmental Microbe-Metal Interactions;
Lovley, D. R., Ed.; ASM Press: Washington, DC, 2000; pp 277327.
(4) Lovley, D. R. Cleaning up with genomics: applying molecular
biology to bioremediation. Nat. Rev. Microbiol. 2003, 1, 3544.
(5) Lovley, D. R.; Phillips, E. J. P.; Gorby, Y. A.; Landa, E. R.
Microbial reduction of uranium. Nature 1991, 350, 413416.
(6) Macaskie, L. E.; Empson, R. M.; Cheetham, A. K.; Grey, C. P.;
Skarnulis, A. J. Uranium bioaccumulation by a Citrobacter sp. as a
result of enzymically mediated growth of polycrystalline HUO2PO4.
Science 1992, 257, 782784.
(7) Senko, J. M.; Istok, J. D.; Suflita, J. M.; Krumholz, L. R. In situ
evidence for uranium immobilization and remobilization. Environ. Sci.
Technol. 2002, 36, 14911496.
(8) Law, G. T. W.; Geissler, A.; Burke, I. T.; Livens, F. R.; Lloyd, J.
R.; McBeth, J. M.; Morris, K. Uranium redox cycling in sediment and
biomineral systems. Geomicrobiol. J. 2011, 28, 497506.
(9) Moon, H. S.; Komlos, J.; Jaffe, P. R. Uranium reoxidation in
previously bioreduced sediment by dissolved oxygen and nitrate.
Environ. Sci. Technol. 2007, 41, 45874592.
(10) Wilkins, M. J.; Livens, F. R.; Vaughan, D. J.; Beadle, I.; Lloyd, J.
R. The influence of microbial redox cycling on radionuclide mobility in
the subsurface at a low-level radioactive waste storage site. Geobiology
2007, 5, 293301.
(11) Finch, R.; Murakami, T.Systematics and paragenesis of uranium
minerals. In Reviews in Mineralogy, vol. 38; Mineralogical Society of
America: Golden, CO, 1999; pp. 91179
(12) Jensen, K. A.; Palenik, C. S.; Ewing, R. C. U6+ phases in the
weathering zone of the Bangombe U-deposit: observed and predicted
mineralogy. Radiochim. Acta 2002, 90, 761769.
(13) Jerden, J. L., Jr; Sinha, A. K. Phosphate based immobilization of
uranium in an oxidizing bedrock aquifer. Appl. Geochem. 2003, 18,
823843.
(14) Pinto, A. J.; Goncalves, M. A.; Prazeres, C.; Astilleros, J. M.;
Batista, M. J. Mineral replacement reactions in naturally occurring
hydrated uranyl phosphates from the Tarabau deposit: Examples in the
CuBa uranyl phosphate system. Chem. Geol. 2012, 312313, 1826.
(15) Mehta, V. S.; Maillot, F.; Wang, Z.; Catalano, J. G.; Giammar, D.
E. Transport of U(VI) through sediments amended with phosphate to
induce in situ uranium immobilization. Water Res. 2015, 69, 307317.
(16) Shi, Z. Q.; Liu, C. X.; Zachara, J. M.; Wang, Z. M.; Deng, B. L.
Inhibition effect of secondary phosphate mineral precipitation on
uranium release from contaminated sediments. Environ. Sci. Technol.
2009, 43, 83448349.
(17) Muto, T. Thermochemical stability of ningyoite. Mineral. J.
1965, 4, 245274.
(18) Khijniak, T. V.; Slobodkin, A. I.; Coker, V.; Renshaw, J. C.;
Livens, F. R.; Bonch-Osmolovskaya, E. A.; Birkeland, N. K.;
Medvedeva-Lyalikova, N. N.; Lloyd, J. R. Reduction of uranium(VI)
phosphate during growth of the thermophilic bacterium Thermoterra-
bacterium ferrireducens. Appl. Environ. Microbiol. 2005, 71, 64236426.
(19) Rui, X.; Kwon, M. J.; OLoughlin, E. J.; Dunham-Cheatham, S.;
Fein, J. B.; Bunker, B. A.; Kemner, K. M.; Boyanov, M. I. Bioreduction
of hydrogen uranyl phosphate: Mechanisms and U(IV) products.
Environ. Sci. Technol. 2013, 47, 56685678.
(20) Lee, S. Y.; Baik, M. H.; Choi, J. W. Biogenic formation and
growth of uraninite (UO2). Environ. Sci. Technol. 2010, 44, 8409
8414.
(21) Doinikova, O. A. Uranium deposits with a new phosphate type
of blacks. Geol. Ore Deposits 2007, 49, 8086.
(22) Wellman, D. M.; Icenhower, J. P.; Owen, A. T. Comparative
analysis of soluble phosphate amendments for the remediation of
heavy metal contaminants: Effect on sediment hydraulic conductivity.
Environ. Chem. 2006, 3, 219224.
11077
Article
(23) Paterson-Beedle, M.; Readman, J. E.; Hriljac, J. A.; Macaskie, L.
E. Biorecovery of uranium from aqueous solutions at the expense of
phytic acid. Hydrometallurgy 2010, 104, 524528.
(24) Wellman, D. M.; Pierce, E. M.; Bacon, D. H.; Oostrom, M.;
Gunderson, K. M.; Bovaird, C. C.; Cordova, E. A.; Clayton, E. T.;
Parker, K. E.; Ermi, R. M.; Baum, S. R.; Vermeul, V. R.; Fruchter, J. S.
300 Area treatability test: Laboratory development of polyphosphate
remediation technology for in situ treatment of uranium contamination in
the vadose zone and capillary fringe; Pacic Northwest National
Laboratory: Richland, WA, 2008; www.pnl.gov/main/publications/
external/technical_reports/PNNL-17818.pdf.
(25) Beazley, M. J.; Martinez, R. J.; Sobecky, P. A.; Webb, S. M.;
Taillefert, M. Uranium biomineralization as a result of bacterial
phosphatase activity: Insights from bacterial isolates from a
contaminated subsurface. Environ. Sci. Technol. 2007, 41, 57015707.
(26) Smeaton, C. M.; Weisener, C. G.; Burns, P. C.; Fryer, B. J.;
Fowle, D. A. Bacterially enhanced dissolution of meta-autunite. Am.
Mineral. 2008, 93, 18581864.
(27) Katsenovich, Y. P.; Carvajal, D. A.; Wellman, D. M.; Lagos, L. E.
Enhanced U(VI) release from autunite mineral by aerobic Arthrobacter
sp. in the presence of aqueous bicarbonate. Chem. Geol. 2012, 308
309, 19.
(28) Newsome, L.; Morris, K.; Lloyd, J. R. Uranium biominerals
precipitated by an environmental isolate of Serratia under anaerobic
conditions. PLoS One 2015, 10, e0132392.
(29) Yung, M. C.; Jiao, Y. Biomineralization of uranium by PhoY
phosphatase activity aids cell survival in Caulobacter crescentus. Appl.
Environ. Microbiol. 2014, 80, 47954804.
(30) Martinez, R. J.; Beazley, M. J.; Taillefert, M.; Arakaki, A. K.;
Skolnick, J.; Sobecky, P. A. Aerobic uranium (VI) bioprecipitation by
metal-resistant bacteria isolated from radionuclide- and metal-
contaminated subsurface soils. Environ. Microbiol. 2007, 9, 31223133.
(31) Shelobolina, E. S.; Konishi, H.; Xu, H. F.; Roden, E. E. U(VI)
sequestration in hydroxyapatite produced by microbial glycerol 3-
phosphate metabolism. Appl. Environ. Microbiol. 2009, 75, 57735778.
(32) Beazley, M. J.; Martinez, R. J.; Webb, S. M.; Sobecky, P. A.;
Taillefert, M. The effect of pH and natural microbial phosphatase
activity on the speciation of uranium in subsurface soils. Geochim.
Cosmochim. Acta 2011, 75, 56485663.
(33) Salome, K. R.; Green, S. J.; Beazley, M. J.; Webb, S. M.; Kostka,
J. E.; Taillefert, M. The role of anaerobic respiration in the
immobilization of uranium through biomineralization of phosphate
minerals. Geochim. Cosmochim. Acta 2013, 106, 344363.
(34) Macaskie, L. E.; Bonthrone, K. M.; Rouch, D. A. Phosphatase-
mediated heavy metal accumulation by a Citrobacter sp. and related
enterobacteria. FEMS Microbiol. Lett. 1994, 121, 141146.
(35) Martinez, R. J.; Wu, C. H.; Beazley, M. J.; Andersen, G. L.;
Conrad, M. E.; Hazen, T. C.; Taillefert, M.; Sobecky, P. A. Microbial
community responses to organophosphate substrate additions in
contaminated subsurface sediments. PLoS One 2014, 9, e100383.
(36) Johnson, D. A.; Florence, T. M. Spectrophotometric
determination of uranium(VI) with 2-(5-bromo-2-pyridylazo)-5-
diethylaminophenol. Anal. Chim. Acta 1971, 53, 7379.
(37) Newsome, L.; Morris, K.; Trivedi, D.; Atherton, N.; Lloyd, J. R.
Microbial reduction of uranium(VI) in sediments of different
lithologies collected from Sellafield. Appl. Geochem. 2014, 51, 5564.
(38) Lovley, D. R.; Phillips, E. J. P. Rapid assay for microbially
reducible ferric iron in aquatic sediments. Appl. Environ. Microbiol.
1987, 53, 15361540.
(39) Harris, S. J.; Mortimer, R. J. G. G. Determination of nitrate in
small water samples (100 L) by the cadmium-copper reduction
method: A manual technique with application to the interstitial waters
of marine sediments. Int. J. Environ. Anal. Chem. 2002, 82, 369376.
(40) Dent, A. J.; Cibin, G.; Ramos, S.; Smith, A. D.; Scott, S. M.;
Varandas, L.; Pearson, M. R.; Krumpa, N. A.; Jones, C. P.; Robbins, P.
E. B18: A core XAS spectroscopy beamline for Diamond. J. Phys. Conf.
Ser. 2009, 190, 012039.
DOI: 10.1021/acs.est.5b02042
Environ. Sci. Technol. 2015, 49, 1107011078
Environmental Science & Technology Article

(41) Ravel, B.; Newville, M. ATHENA, ARTEMIS, HEPHAESTUS:

data analysis for X-ray absorption spectroscopy using IFEFFIT. J.
Synchrotron Radiat. 2005, 12, 537541.
(42) Downward, L.; Booth, C. H.; Lukens, W. W.; Bridges, F. A
variation of the F-Test for determining statistical relevance of
particular parameters in EXAFS fits. AIP Conf. Proc. 2007, 882,
129131.
(43) McKenzie, H. M.; Coughlin, D.; Laws, F.; Stamper, A.
Groundwater Annual Report 2011; Sellaeld Ltd: Cumbria, U.K.,
2011; www.sellaeldsites.com/land/documents/
Annual%20Data%20Review%202011.pdf.
(44) Caporaso, J. G.; Kuczynski, J.; Stombaugh, J.; Bittinger, K.;
Bushman, F. D.; Costello, E. K.; Fierer, N.; Pena, A. G.; Goodrich, J.
K.; Gordon, J. I.; Huttley, G. A.; Kelley, S. T.; Knights, D.; Koenig, J.
E.; Ley, R. E.; Lozupone, C. A.; McDonald, D.; Muegge, B. D.;
Pirrung, M.; Reeder, J.; Sevinsky, J. R.; Turnbaugh, P. J.; Walters, W.
A.; Widmann, J.; Yatsunenko, T.; Zaneveld, J.; Knight, R. QIIME
allows analysis of high-throughput community sequencing data. Nat.
Methods 2010, 7, 335336.
(45) Dusausoy, Y.; Ghermani, N.-E.; Podor, R.; Cuney, M. Low-
temperature ordered phase of CaU(PO4)2: synthesis and crystal
structure. Eur. J. Mineral. 1996, 8, 667674.
(46) Newsome, L.; Morris, K.; Shaw, S.; Trivedi, D.; Lloyd, J. R. The
stability of microbially reduced U(IV); impact of residual electron
donor and sediment ageing. Chem. Geol. 2015, 409, 125135.
(47) Alessi, D. S.; Lezama-Pacheco, J. S.; Stubbs, J. E.; Janousch, M.;
Bargar, J. R.; Persson, P.; Bernier-Latmani, R. The product of microbial
uranium reduction includes multiple species with U(IV)phosphate
coordination. Geochim. Cosmochim. Acta 2014, 131, 115127.
(48) Weber, K. A.; Achenbach, L. A.; Coates, J. D. Microorganisms
pumping iron: anaerobic microbial iron oxidation and reduction. Nat.
Rev. Microbiol. 2006, 4, 752764.
(49) Moon, H. S.; Komlos, J.; Jaffe, P. R. Uranium reoxidation in
previously bioreduced sediment by dissolved oxygen and nitrate.
Environ. Sci. Technol. 2007, 41, 45874592.
(50) Begg, J. D. C.; Burke, I. T.; Lloyd, J. R.; Boothman, C.; Shaw, S.;
Charnock, J. M.; Morris, K. Bioreduction behavior of U(VI) sorbed to
sediments. Geomicrobiol. J. 2011, 28, 160171.
(51) Parmentier, M.; Ollivier, P.; Joulian, C.; Albrecht, A.; Hadi, J.;
Greneche, J.-M.; Pauwels, H. Enhanced heterotrophic denitrification in
clay media: The role of mineral electron donors. Chem. Geol. 2014,
390, 8799.
(52) Ray, A. E.; Bargar, J. R.; Sivaswamy, V.; Dohnalkova, A. C.;
Fujita, Y.; Peyton, B. M.; Magnuson, T. S. Evidence for multiple
modes of uranium immobilization by an anaerobic bacterium. Geochim.
Cosmochim. Acta 2011, 75, 26842695.
(53) Ray, A. E. Discovery and characterization of a novel anaerobe
with a potential role in bioremediation of metal-contaminated
subsurface environments. Ph.D. dissertation, Idaho State University,
Pocatello, ID, 2007.
(54) Beller, H. R.; Han, R.; Karaoz, U.; Lim, H.; Brodie, E. L.
Genomic and physiological characterization of the chromate-reducing,
aquifer-derived Firmicute Pelosinus sp. strain HCF1. Appl. Environ.
Microbiol. 2013, 79, 6373.

11078 DOI: 10.1021/acs.est.5b02042

Environ. Sci. Technol. 2015, 49, 1107011078