See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/12485407

Snoeyenbos-West OL, Nevin KP, Anderson RT, Lovley DR.. Enrichment of

Geobacter species in response to stimulation of Fe(III) reduction in sandy aquifer
sediments. Microb Ecol 39: 1...

Article  in  Microbial Ecology · March 2000

DOI: 10.1007/s002480000018 · Source: PubMed

CITATIONS READS

190 126

4 authors, including:

Oona Snoeyenbos-West
Natural History Museum Stockholm Sweden
58 PUBLICATIONS   1,780 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Direct Interspecies Electron Transfer (DIET) in multi-species syntrophic co-cultures. View project

Electroactive eukaryotes: electrogenic fungi and their industrial and ecological potential View project

All content following this page was uploaded by Oona Snoeyenbos-West on 18 September 2015.

The user has requested enhancement of the downloaded file.

 (QULFKPHQWRI*HREDFWHU6SHFLHVLQ5HVSRQVHWR6WLPXODWLRQRI)H,,,5HGXFWLRQLQ6DQG\
$TXLIHU6HGLPHQWV
$XWKRUV2/6QRH\HQERV:HVW.31HYLQ57$QGHUVRQDQG'5/RYOH\
6RXUFH0LFURELDO(FRORJ\9RO1R)HESS 
3XEOLVKHGE\6SULQJHU
6WDEOH85/KWWSZZZMVWRURUJVWDEOH
$FFHVVHG87&

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.

6SULQJHUis collaborating with JSTOR to digitize, preserve and extend access to 0LFURELDO(FRORJ\

http://www.jstor.org

This content downloaded from 128.119.49.10 on Fri, 18 Sep 2015 16:35:12 UTC
All use subject to JSTOR Terms and Conditions
 ECOLOGY
Microb Ecol (2000) 39:153-167
DOI: 10.1007/s002480000018
? 2000 Springer-VerlagNew York Inc.

Enrichment of GeobacterSpecies in Response to Stimulation of

Fe(III) Reduction in Sandy Aquifer Sediments

O.L. Snoeyenbos-West,K.P. Nevin, R.T. Anderson, D.R. Lovley

Department of Microbiology,University of Massachusetts,Amherst, MA 01003, USA

Received:19 October 1999; Accepted:28 December 1999; Online Publication:25 April 2000

|A B S T R A CT
Engineeredstimulation of Fe(III) has been proposed as a strategyto enhance the immobilizationof
radioactiveand toxic metals in metal-contaminatedsubsurfaceenvironments.Therefore,laboratory
and field studies were conducted to determine which microbial populations would respond to
stimulation of Fe(III) reduction in the sediments of sandy aquifers. In laboratory studies, the
addition of eithervarious organicelectron donors or electron shuttle compounds stimulatedFe(III)
reductionand resultedin Geobactersequencesbecoming importantconstituentsof the Bacterial16S
rDNA sequences that could be detected with PCR amplificationand denaturinggradientgel elec-
trophoresis (DGGE). Quantification of Geobacteraceaesequences with a PCR most-probable-
number technique indicatedthat the extent to which numbers of Geobacterincreasedwas relatedto
the degreeof stimulationof Fe(III) reduction. Geothrixspecieswere also enrichedin some instances,
but were orders of magnitude less numerous than Geobacterspecies. Shewanellaspecies were not
detected, even when organic compounds known to be electron donors for Shewanellaspecies were
used to stimulate Fe(III) reduction in the sediments. Geobacterspecies were also enriched in two
field experiments in which Fe(III) reduction was stimulated with the addition of benzoate or
aromatichydrocarbons.The apparentgrowth of Geobacterspeciesconcurrentwith increasedFe(III)
reduction suggests that Geobacterspecies were responsiblefor much of the Fe(1II)reduction in all
of the stimulation approachesevaluatedin three geographicallydistinct aquifers.Therefore,strat-
egies for subsurfaceremediationthat involve enhancingthe activityof indigenous Fe(III)-reducing
populations in aquifers should consider the physiologicalproperties of Geobacterspecies in their
treatment design.

Introduction als in subsurfaceenvironments[22]. Severalmetals and met-

Engineered stimulation of dissimilatory metal reduction
may be a useful method for immobilizing contaminantmet-
Correspondenceto: D.R. Lovley; Fax: (413) 545-1578; E-mail: dlovley@
microbio.umass.edu
alloids, including uranium, technetium, chromium, cobalt,
and selenate, are more soluble, and thus mobile, in the oxi-
dized state and tend to precipitatein the reduced state [22].
For example,many microorganismscapableof dissimilatory
metal reductioncan reducesoluble U(VI) to insolubleU(IV)
[25, 36], and stimulation of microbial U(VI) reduction un-

This content downloaded from 128.119.49.10 on Fri, 18 Sep 2015 16:35:12 UTC
All use subject to JSTOR Terms and Conditions
154
der laboratoryconditions has been shown to precipitateura-
nium from contaminated groundwaters[34].
Fe(III) is generallythe most abundant possible electron
acceptor for dissimilatory metal-reducing microorganisms
in soils and sediments [23, 25]. Most Fe(JII)-reducingmi-
croorganisms can also reduce Mn(IV), which is typically
present at concentrationsabout 10%of those of Fe(III) [23,
25]. Fe(III) and Mn(IV) are abundant in many subsurface
soils and are likely to be in higher concentrationsthan con-
taminant metals, even under conditions of severe contami-
nation. Therefore,stimulation of dissimilatorymetal reduc-
tion to immobilize metal contaminantsalso involves micro-
bial reduction of the Fe(III) and Mn(IV) in the soils. The
reduction of Fe(III) and Mn(IV) along with contaminant
metals may be beneficialbecause the Fe(II) and Mn(II) pro-
duced can serveas a redox bufferto help preventreoxidation
of the reducedcontaminantmetals. Furthermore,since most
of the dissimilatoryFe(III)- and Mn(IV)-reducingmicroor-
ganisms that have been evaluated are capable of reducing
contaminant metals [24, 25, 29], stimulation of Fe(III) and
Mn(IV) reduction may increase the population size of mi-
croorganismsthat can also reduce contaminantmetals. Such
stimulation may thus enhance rates of contaminant metal
reduction.
In order to understandhow to most effectivelyaccelerate
dissimilatorymetal reduction in subsurfaceenvironments,it
is important to know which microbialpopulations are pro-
moted when dissimilatory metal reduction is artificially
stimulated. A wide phylogenetic diversity of Bacteria [24]
and Archaea [53] are capableof dissimilatoryFe(III) reduc-
tion. It is conceivablethat the growth and activityof any of
these known Fe(III) reducers,or as yet undescribedspecies,
might be enhanced when Fe(III) reduction is stimulated in
contaminated subsurface environments. Which dissimila-
tory metal reducersbecome predominant might be affected
by such factorsas the physiologicalpropertiesof the Fe(III)-
reducing microorganisms,the geochemistryof the environ-
ment, the composition of the microbialcommunity prior to
stimulation, and the method used to stimulatemetal reduc-
tion.
Three mechanisms for stimulating Fe(III) reduction in
sedimentary environments have been identified. Previous
studies have demonstratedthat addition of known electron
donors for Fe(III) reducers can stimulate Fe(III) reduction
in aquaticsediments [33], and this is also a likely strategyfor
stimulatingFe(III)reduction in organic-poorsubsurfaceen-
vironments. The addition of electron shuttlingorganic com-
pounds such as humic acids or other extracellularquinones
This content downloaded from 128.119.49.10 on Fri, 18 Sep 2015 16:35:12 UTC
All use subject to JSTOR Terms and Conditions
O.L. Snoeyenbos-Westet al.
[30, 31] can also stimulate Fe(III) reduction in sediments
[46]. These compounds can greatlyaccelerateFe(III) reduc-
tion by alleviating the need for metal reducers to directly
contact insoluble metals or metals occluded within pore
spaces too tight for metal reducers to access [30, 31, 46].
Fe(I1I)reduction can also be stimulated in subsurfacesedi-
ments with the addition of Fe(III) chelators that solubilize
Fe(III) and make it more readilyavailablefor Fe(III) reduc-
tion [38, 39, 40]. However, this third option is not an ap-
propriatestrategyfor stimulatingdissimilatorymetal reduc-
tion in metal-contaminated subsurface environments be-
cause the chelators that solubilize Fe(III) may also chelate
contaminant metals and enhance their mobility.
Highly transmissivesandy aquifers are likely subsurface
environmentsin which contaminantmetal mobility may be
of concern. Such highly porous sediments are also among
the most amenable for in situ treatment strategiesthat in-
volve the addition of soluble stimulants. Therefore,a com-
bination of laboratoryand field studies was conducted in
order to determine which microbial populations are stimu-
lated when Fe(III)reductionis acceleratedin sandy aquifers.
The results from both field and laboratorystudies demon-
stratethat stimulation of dissimilatorymetal reductionwith
a variety of electron donors and/or electron shuttling com-
pounds results in the specific enrichment of Geobacterspe-
cies.
Materials and Methods
Incubations
Laboratory
Sediments for laboratory studies were collected from an aquifer
locatedin Bemidji,Minnesota.Portionsof the aquiferat this site
have been contaminatedwith crude oil, and the petroleum-
contaminatedportionof the aquifercontainsan extensivezone of
Fe(III)reduction [4, 5, 50]. For this study, sedimentswere collected
from a nearbyuncontaminated,aerobicportion of the aquiferwith
a truck-mounted drill rig and sterilizedcore barrels,as previously
described [5]. The cores were sectioned, immediately capped,
wrappedin plasticwrap, sealedwith duct tape, and placedin plastic
bags under an N2 atmosphere prior to shipment in coolers via
overnightcarrierto the laboratory.In the laboratory,sediments (40
g) were transferredinto serum bottles (5 ml) under N2 in a glove
bag. The bottles were sealed with thick butyl rubber stoppers and
removed from the glove bag, and the headspacewas flushed with
N2-CO2 (93:7). Acetate(10 mM), lactate(5 mM), formate (5 mM),
benzoate (2 mM), or glucose (3.3 mM) were added as electron
donors at the stated final concentrations from anaerobic stocks.
Triplicatebottles of each treatmentwere incubated in the dark at
20?C. The sediments were subsampled over time, and HCl-
extractableFe(II) and Fe(III) were determined as previously de-
SubsurfaceFe(III)-ReducingCommunity Structure
scribed [4]. Subsamples(2 g) of sediment were frozen at -80?C for
the molecular analysesdescribedbelow.
The effect of humics and anthraquinone 2,6-disulfonate
(AQDS) on the rate of Fe(III) reduction in the sedimentswere also
determinedin a relatedstudy [46], where all of the sedimentswere
amended with acetate (5 mM) as the electron donor and various
concentrationsof humics or AQDS. Subsamplesthat were frozen at
-800C for molecularanalysisduringthat study were analyzedin the
study reported here.
Columbus Field Site
A field study was conducted to determinethe effect of the addition
of petroleum hydrocarbonson the microbiologyand geochemistry
of a sand and gravel aquifer at the Columbus Air Force Base,
Columbus, Mississippi [6, 41]. This is the site of a groundwater
researchfacilityfor the U.S. Air Force.A syntheticjet fuel mixture
was placed within the aquifer in order to simulate the effect of a
hydrocarbonspill in an aquifer.The development of the hydrocar-
bon plume and its effects on the subsurfacemicrobialecology were
monitored over time. Sedimentswere cored and samplesprocessed
as described above. The predominant terminal electron-accepting
process in the sediments was determined with the [2-'4C]acetate
technique [26]. Samples from a portion of the Fe(III)-reducing
zone that developed in the aquiferas well as a nearby,uncontami-
nated site were frozen at -800C for molecular analysis.
Borden Field Site
Sedimentsfrom a field study in which benzoate was injected into a
sandy aquifer in order to promote anaerobic respiration were
kindly provided by B.J.Butler,Universityof Waterloo. Portions of
the Borden aquiferwere divided with sheet piling to fashion three
treatmentgates parallelto groundwaterflow. A contaminant solu-
tion containing perchloroethylene(6 MtM),carbon tetrachloride(6
riM), and toluene (12 1iM)was added to the groundwaterin the
gates. Gate 2 receivedthis contaminantsolution only and no other
additions. However, gate 3, in addition to the contaminants, also
received during the treatment period a solution of benzoate (2.1
mM), which was injected into the groundwaterat gate 3 approxi-
mately every 14 days. This was done in order to promote anaerobic
respirationin the gate. In this study, sediments of gate 2 and gate
3 were analyzed,as well as sedimentsfrom a nearbysite that did not
receive any additions. Sediment cores were collected from gates 2
and 3 and the backgroundsite over an intervalof 1.5 to 3 m below
ground surface on 4 July and 4 November 1997, after severalpe-
riodic injections of benzoate into the gate 3 site. Sediment cores
were collected in aluminum core barrels, capped and sealed with
duct tape, then returned to the lab where each 1.5-m core sample
was cut into 0.3-m sections. Each section was then placed into an
N2-filled anaerobicbag. About 1 cm of materialfrom the ends of
each section was discardedand the center portion of each section
was sampled, homogenized in sterilizedaluminum pans, and allo-
cated into sterile Whirlpakbags inside sterile canningjars contain-
This content downloaded from 128.119.49.10 on Fri, 18 Sep 2015 16:35:12 UTC
All use subject to JSTOR Terms and Conditions
155
ing an argonheadspace.Samplesfor molecularanalysiswere frozen
at -80?C.
Concentrationsof dissolved hydrogen at the gate 2 and gate 3
sampling sites were determined in October 1997 using the bubble
strip method as previously described [28]. HCl-extractableFe(III)
and Fe(II) on the same sediments used for molecularanalysiswere
determined as describedabove.
PCR-DGGE
Analysisof SedimentMicrobialCommunities
The phylogeneticdiversityof the microbialcommunity in the sedi-
ments was evaluatedwith PCR amplificationof 16S rDNA genes
and denaturinggradientgel electrophoresis(DGGE) [44, 45]. PCR-
DGGE methods employed were as previously described [50], but
included the following modifications: PCR was performed using
bacterialprimers 338F-GC (the complement of EUB338) [2] and
907R [20] with touchdown primer annealing from 65 to 55?C
(decreasing0.5?C per cycle), followed by 10 cycles at 55?C. This
prevents amplification of nonspecific PCR products. In addition,
archaealdiversitywas accessed in the sampled sediments with Ar-
chaea-specific primers ARC344 (5'-ACGGGGAGCAGCAGGC-
GCGA-3') [48] and 0915Ra Reverse (5'-GTGCTCCCCCGC-
CAATTCCT-3')[1]. 16S rDNA ampliconswere resolvedby DGGE
with a 40 to 80% denaturinggradient(where 100%is equivalentto
7 M urea and 40% formamide) [45] on a 7% acrylamidegel in lx
TAE (40mM Tris-acetate(pH 7.4), 20 mM sodium acetate, 1 mM
EDTA). Standardsof referenceDNA, extractedfrom various pure
cultures as previouslydescribed [50], were run along with the en-
vironmental samples for ease of comparison among the different
gels. DGGEgels were run for 15 h at 65 V and 60?C.PCRproducts
from excised DGGEbands were purified using the QlAquick PCR
Purificationkit (Qiagen Inc., Valencia, CA), and sequenced using
Dye Deoxy Terminator Cycle Sequencing (Perkin-ElmerCetus)
and an ABI 377 automated sequencer (Applied Biosystems,Foster
City, CA) at the University of MassachusettsSequencingFacility.
Phylogenetic
Analysisof SequenceData
The Ribosomal DatabaseProject (RDP) [42] CHECK_CHIMERA
programand secondarystructuredeterminationwere used to check
the partial16S rRNAgene sequencesfor potentialchimericartifacts
[18]. Sequenceswere analyzedusing BLAST(National Center for
Biotechnology Information) and SIMILARITY_RANK (RDP) in
order to find the most similar available database sequences. Se-
quences were then manuallyaligned with closely related 16S rRNA
sequencesfrom RDP and GenBank,using the graphicaluser inter-
face SeqLab (Wisconsin Package version 10; Genetics Computer
Group (GCG), Madison, WI). Only those sequence regions that
could be alignedwith confidencewere included in the analyses,and
gaps were treated as missing nucleotides. Phylogenetictrees were
inferredfrom unambiguouslyaligned sequence data, using the dis-
tance, maximum-likelihood, and maximum-parsimony tools of
PAUP* [51].
Distance and maximum likelihood analyseswere performedus-
ing heuristic tree searchingvia simple stepwise addition with tree
156
bisection reconnection rearrangement. Unweighted parsimony
analysis used the branch-and-bound algorithm. In addition,
the distance tools (neighbor joining of Kimura distances) of
TREECONfor Windows 1.3b [52] were employed. Three different
methods of phylogenetictree constructionwere used with the same
data set in order to test the robustness of the generatedtree topo-
logy. One thousand bootstrap replicates were performed on the
data set. Phylogenetictrees inferredusing the two differentsoftware
packagesdescribedabove showed the same topologies.
MPN-PCR
Analysisof Sediments
The relative numbers of Geobacteraceae, Geothrix,and Shewanella
species were determined in sediment samples using an MPN-PCR
technique [5, 47]. DNA was extracted from triplicate sediment
samplesas describedabove. Serial10-fold dilutions of the sediment
DNA were made in sterile water, and 3-[dlaliquots of the diluted
DNA were used as template in the PCR.PartialGeobacteraceae 16S
rDNA sequences were amplified with bacterialforwardprimer 8F
(5' -AGAGTTTGATCCTGGCTCAG-3') and Geobacteraceae-
specific reverse primer Geo825R (5'-TACCCGCRACACCTAGT-
3') in the first round of a semi-nested PCR protocol, followed by
338F [2] and Geo825R in the second round. The PCR conditions
were as previouslydescribed [5] except that 1 [l of the first-round
reaction product served as template in second-round reactions.
Geothrixspecies were also enumeratedwith a semi-nested MPN-
PCR protocol. The primersfor the first round of PCRwere 8F and
the Geothrix-specific reverse primer Gx.472R (5' -AGGT-
ACCGTCAAGTAACASS-3'). The Geothrix-specificforwardprimer
Gx. 182F (5' -AGACCTTCGGCTGGGATGCT-3')and Gx.472R
were the primers for the second round. Both the Geobacteraceae
and Geothrix-specificprimers were designed and tested for speci-
ficity in our lab prior to use in lab and field experiments (manu-
script in preparation). In addition, partial Shewanella16S rDNA
sequencesin sediment from the stimulationexperimentswere enu-
merated using the Shewanella-specificprimer set Sw.783-F (5'-
AAAGACTGACGCTCAKGCA-3')and Sw. 1245-R (5' -TTYG-
CAACCCTCTGTACT-3')(manuscriptin preparation).
Results
Effectof ElectronDonorAmendmentsto BemidjiAquifer
Sediments
Each of the electron donors evaluated produced an increase
in the rate of Fe(III) reduction in Bemidji sediment relative
to unamended controls (Fig. 1). The addition of acetate,
glucose, and lactate resulted in the fastest rates of Fe(III)
reduction; rates were intermediate with benzoate. Formate
additions resulted in the lowest rates of Fe(III) reduction.
Each sediment type was subsampled for analysis of 16S
rDNA sequences when approximately 60% of the Fe(LII)had
been reduced, with the exception of control sediments which
This content downloaded from 128.119.49.10 on Fri, 18 Sep 2015 16:35:12 UTC
All use subject to JSTOR Terms and Conditions
O.L. Snoeyenbos-Westet al.
80
70
60
50
uL 40
3 -0 0 Cont-ol
30 -U-- ~~~~~~~~Acetate
// Benzoate
20 -v- Formate
Glucose
-4-
10 Lactate
---~~~~~~~~~
10
0 10 20 30 40 50
Days
Fig. 1. Rates of Fe(IIL)reduction shown as an increase of Fe(II)
over time in sediments from the Bemidji aquifer incubated under
anaerobic conditions and amended with various electron donors.
The resultsare the means of triplicatedeterminations,and standard
deviations are shown on the graph as error bars.
were analyzedafter49 days of incubation.Samplesamplified
with primers designed to amplify the 16S rDNA from Ar-
chaea did not yield detectablePCRproduct. However,prim-
ers which amplify the 16S rDNA of most Bacteriadid yield
PCRproductswhich could be resolvedvia DGGE(Fig. 2). In
accordancewith previous analysesof the Bemidji sediments
[50], a low diversityof bandswas recoveredfrom the control
sediments. Sequenceanalysisof excised bands demonstrated
that the control sediments contained a 16S rDNA sequence
closely relatedto known Geobacterspecies. Other sequences
recoveredfrom the control sedimentswere closely relatedto
Thiobacillusaquaesulisand several unidentified ,B-Proteo-
bacteria. Thiobacillusaquaesulis is not known to reduce
Fe(III), but we do not rule out the possibility that some of
these as yet unidentified13-Proteobacteriamay be able to do
so. This question awaits further study.
DGGE bands with mobilities similar to that of the Geo-
bacterband observed in the control sediments were recov-
ered from sedimentsamendedwith the electrondonors (Fig.
2). The sequences of these bands were 99.8-100% similar
and could be assignedto the genus Geobacter(Fig. 3). Band
intensitywas muchgreaterin the sediments amended with
electron donors than' in the control. Sediments in which
Subsurface Fe(III)-Reducing Community Structure 157

Standard Standard
.. Fotn 3azaormate

Clostridium
_

Clostridium
redeubacterium
f1-Proteobacten

_ ~~~~ M 1i#t r .
_ ~~~~~~Formats Goo2
f3-Proteobacteri' Denftrobancterium
aq
Thiobaoillus

Fig. 2. DGGE profile of bacterial 16S rDNA fragmentsretrievedfrom Bemidji sediments. Individualbands that yielded useful sequence
data are labeled with the microorganismsto which the sequenceswere most similar. Other unlabeledbands were not separatedenough to
be cut from the gel, were chimeras, or did not yield good-quality sequence data. DGGE bands labeled in bold were sequenced, and the
phylogenetic placement of these sequences is shown in Fig. 3.

Fe(III) reductionwas greatlystimulated (acetateand glucose
amended) had the most intense bands (Fig. 2). The intensity
of an analogous band in the formate-amended sediments
was notably less, correspondingto the lower rates of Fe(III)
reduction (Fig. 1). Dominant and intense DGGEbands may
be consideredthe most abundantbacterialspecies [13], and
based on their intensity, the Geobacterbands can be consid-
ered to represent the predominant species in the amended
sediments. The formate- and lactate-amended sediments
also contained a second Geobacterband, Formate Geo-2
(Fig. 2), that differed in mobility, but was closely related to
the other Geobactersequences in the sediments (Fig. 3).
These resultssuggestedthat, regardlessof the electron donor
added, stimulation of Fe(III) reduction with the electron
donors enriched for Geobacterspecies that were very closely
related.
In order to further evaluate the effect of the added elec-
tron donors on the size of the Geobacterpopulation, num-
bers of Geobacteraceae16S rDNA sequences from donor-
amended sediments were enumerated with MPN-PCR [5,
47]. The sediments that initially had the highest rates of
Fe(III)reductionhad severalordersof magnitudemore Geo-
bacteraceaesequences than the control sediments or sedi-
ments amended with formate (Fig. 4). Formate-amended
sediments had slightly more Geobacteraceae sequences than
the control sediments. These results indicate that the stimu-
lation of Fe(III) reduction with electron donors resulted in
increasednumbers of Geobacteraceae. This is in accordance
with the qualitativeresults from the DGGE analysis.
Another notable result from the DGGE analysiswas the
detection of 16S rDNA sequences closely related to Clos-
tridiumspecies in the sediments amended with glucose, but
not in the other electron donor amendments (Fig. 2). This is
consistent with the concept that fermentative microorgan-
isms metabolize glucose in sediments in which Fe(III) re-
duction is the terminal electron-acceptingprocess with sub-
sequent metabolismof the fermentationproductsby Fe(III)-
reducing microorganisms [23, 35].
Sequences for other extensively studied Fe(III)-reducing
microorganismssuch as Shewanellaspecies or Geothrixfer-
mentanswere not detectedwith the DGGEanalysis.In order
to furtherevaluatethe potential role of these organisms,they
were enumeratedwith the MPN-PCRtechnique, with prim-
ers that were specific for these organisms. No Shewanella
sequences could be detected, even in sediments amended
with known electron donors for Shewanellaspecies, such as
lactate and formate.No Geothrixsequenceswere detected in
unamendedcontrol sediments,but Geothrixsequencescould
be enumeratedin sediments amended with electron donors
(Fig. 4). The highest number of Geothrixsequences were

This content downloaded from 128.119.49.10 on Fri, 18 Sep 2015 16:35:12 UTC
All use subject to JSTOR Terms and Conditions
158 O.L. Snoeyenbos-Westet al.

Borden Geo-1
_ Borden Geo-2
Borden Geo-3
70 Borden Geo-4
Borden Geo-5
Borden Geo-6
Borden Geo-7
62
Borden Geo-8
Borden Geo-9
Borden Geo-1 0
LBorden Geo-1I
Humics Geo-82
Humics Geo-24
ControlGeo-1
AQDS Geo-i
Humics Geo-1
Humics Geo-2 ma
94. Humics Geo-3 Z L
Formate Geo-2 co
Glucose Geo-1
Acetate Geo-1
Benzoate Geo-1
Formats Geo-1 0
Lactate Geo-1
BemidjiGeo-60
BemidjiGeo-83 O
Borden Geo-101 0 0
Bemidjibenzene enrichment(Benz-76) @ Fig. 3. Phylogenetic tree inferred
Columbus Geo-1 0 from 16S rRNA sequences showing
Columbus Geo-2 the phylogeneticplacementof Geo-
"Geobacterbemidjiensis"
Geobacter arculus bacter sequences amplified from
82 o
Geobactersulfureducens Q sediments in three different aqui-
Geobacter metallireducens 0 fers, and from environmental iso-
Geobacter hydrogenophilus lates. Phylogenetic relationships
L Pelobacterpropionicus shown here were inferredby using
Geobacter chapellei
GeobacterALA6 neighbor joining and Kimuratwo-
Geobacter humireducens parameter genetic distances in
GeobacterALA5 TREECON[52]. A total of 400 base
Geobacter akaganeitreducens positions were considered in the
84 100 Desulfuromrusa bakii
100 LDesulfuromusakysingii analysis.Bootstrapvalues above 60
Desulfuromnusasuccinoxidans are shown adjacentto branchnodes
Pelobacter carbinolicus and were calculatedfrom 1,000 re-
r Pelobacter acetylenicus sampled data sets using neighbor
Desulfuromonasacetoxidans joining. The scale bar shows ex-
Desulfuromonasacetexigens
pected nucleotide substitutionsper
Pelobacter venetanus
100 Humics Dsf-1 site per unit of branch length. Op-
100 FColumbus Dsfi1 erationalTaxonomic Units (OTUs)
BemidjiGeo-144 0 shown on the tree in bold are the
62 1 Borden Dsf-1o 0
same sequences obtained using
LI BemidjiGeo-47
uE DGGE from amended sediments
BemidjiGeo-52 0
100 BemidjiGeo-48 and fields sites in this study. Those
61 r 1Bemidji Geo-58 shown in plain text are from a pre-
c Ua
00 BemidjiGeo-43 q vious study of the Bemidji aquifer
BemidjiGeo-3 [50], and from GenBank and the
BemidjiGeo-2 RDP databases.A similar tree to-
Desulfovibriovulgaris
Escherichiacoli pology was generatedfor trees con-
structed using maximum-likeli-
0.1 substitutions/site
hood and maximum-parsimony
methods (data not shown).

This content downloaded from 128.119.49.10 on Fri, 18 Sep 2015 16:35:12 UTC
All use subject to JSTOR Terms and Conditions
SubsurfaceFe(III)-ReducingCommunity Structure
109
1 At8
-
lE Geobacter
Fig.4 Geothrix
106-
fee)n dn10 r
1- 104
E also recoveredfrom some of the humics-amendedsediments
z 0
" 100
recoveredfrom sediments amendedwith acetateand lactate.
This is in accordance with the physiological properties of
known Geothrix species, which oxidize acetate and lactate
with Fe(III) as the electron acceptor,but do not metabolize
glucose, benzoate, or formate [10].
In all of the sedimentsamendedwith electron donors, the
numbers of Geothrixsequenceswere severalorders of mag-
nitude lower than the number of Geobactersequences (Fig.
4). This is consistent with the predominance of Geobacter
sequences and lack of Geothrix sequences in the DGGE
analysis. These results further indicate that stimulation of
Fe(III) reduction by addition of electron donors primarily
resulted in an increase in Geobacterspecies.
Effectof Stimulationof Fe(lll)Reductionwith
Extracellular
Quinones
The addition of humics or the humics analog AQDS stimu-
lated Fe(III) reductionrinBemidji sediments [46]. Humics-
amended sediments from that study were analyzedafter 24
days of incubation,at which point the following percentages
of the HCl-extractableiron had been reduced: no humics
added, 3.7%;0.1 g humics added, 8.2%;0.5 g humics added,
54.4%; 1 g humics added, 58%; 2 g humics added, 82.1%
[46]. PCR,DGGE,and sequence analysisindicateda specific
enrichment of a Geobacterspecies in the humics-amended
sediments (Fig. 5A). The dominant Geobactersequence ob-
servedwas 99% similarto the Geobactersequencesthat pre-
dominated in the electron-donor amended sediments (Fig.
3). The enrichment of this Geobactersequence continued
through 82 days of incubation (Fig. SB) when most of the
This content downloaded from 128.119.49.10 on Fri, 18 Sep 2015 16:35:12 UTC
All use subject to JSTOR Terms and Conditions
159
Fe(III)had been reducedin all of the humics-amendedsedi-
ments [46]. The Geobacterband was also intensified in the
control sediments after 82 days of incubation,which is con-
sistent with significantFe(III) reduction in those sediments
by that time [46].
Other bands that comprised Geobactersequences were
V (Fig. 5A). These Geobactersequences were very closely re-
lated to the Geobactersequence that predominatedin all the
humics-amended sediments. However, one Geobacteraceae
sequence, which yielded a faint DGGE band (Dsf-1), fell
within the Desulfuromonascluster of the Geobacteraceae
[21]. A sequence closely relatedto the thermophilicFe(III)-
reducing microorganismBacillusinfernus[7] was recovered
from the sediments amendedwith the highest concentration
of humics, but not from the other sediments. Also, some
other unique bands emergedin the sedimentsamendedwith
high concentrations of humics (Fig. 5A and Fig. 5B), but
none of these are known humics- or Fe(III)-reducingmi-
croorganisms.
In results similar to those observed with the humics
amendments, the addition of 10, 50, or 100 iM AQDS
stimulatedFe(IIJ)reduction [46] and resultedin an enrich-
ment of a Geobacterspecies (Fig. 6) whose sequence was
99.6% similar to the sequences that predominated in the
humics- and donor-amended sediments (Fig. 3). This band
associatedwith Geobacterwas not observedin the sediments
amendedwith 500 [tMAQDS, which correspondsto the fact
that the addition of 500 [M AQDS did not stimulateFe(III)
reduction in these sediments [46].
MPN-PCR analysis indicated that the humics-amended
sediments contained up to three orders of magnitude more
Geobacteraceaethan the control sediment after 24 days of
incubation (Fig. 7). The number of Geobacteraceaese-
quences increasedin the control sediments between 24 and
82 days of incubation (Fig. 7), correspondingto increased
Fe(III) reduction in the control sediments over this period
[46]. However,the number of Geobactersequencesremained
significantlyhigher in the humics-amended sediments. The
AQDS-amendedsediments had a similar trend, but with an
even wider disparitybetween the number of sequencesin the
control sediments versus the sediments in which Fe(III) re-
duction was stimulated.Therewere five ordersof magnitude
more Geobactersequencesin the sediment amendedwith 10
and 100 MMAQDS than in the control (Fig. 8). The sedi-
ments that were amendedwith 500 [MM AQDS had numbers
of Geobactersequences similar to those in the control sedi-
160 O.L. Snoeyenbos-Westet al.

I
.e~~~~~~~~~~~~~~~~~1

G
Humics
Humics
Humics
Humics
Humics
GA *

ControlGo trII le

| ~~~~~~~~~~infemus
tLFS

A Increasing of humics(gramsperl_fer)
concentration
Stndr
Humics mct Standard2
DenitrffyingFe
(d)-oxidizing
aedum
Abateou
j~~~~~~~~~~ _

A (Grams concentration
Increasing of humics perler)
Standard Standard
-postbaduum
Bac-tq dirostasonisum
* -t!t w _ w -. + ~~~~Erysipelothnx

~~~~~~~~~~~~~~~~~~~~~~Unidentified e-e-

u_ Cysprotoobacter
:s:.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
....X_X
..........
s-s.s
..

B of humics(gramsperliter)um
Increasingconcentration
2 .^;.: ;rffl ^ ^ w --Cyst ystbacter
Fi. .Deaurn gadet 0.0
e eetroporsi (G 0.1 E)prfl of0.5 1.0 t1 2-o
>obacteria6rDAfgmnsetivdroh umic-mne eij

sediment after24 (A) and 82 (B) days of incubation. Individualbands that yielded useful sequence data are labeledwith the microorganisms
to which the sequenceswere most similar.DGGEbands labeled in bold were sequenced,and the phylogeneticplacement of these sequences
is shown in Fig. 3.

This content downloaded from 128.119.49.10 on Fri, 18 Sep 2015 16:35:12 UTC
All use subject to JSTOR Terms and Conditions
Subsurface Fe(III)-Reducing Community Structure 161

Stadard Standard

Dechiorimonas
aiau
PDsudnmonas

Pseudomonas

Go-
AQDS Itr_MIcans10-Protob
terIum
subgroup)

Increasingco raton of AQDS(pM)

Fig. 6. Denaturinggradientgel electrophoresis(DGGE) profile of bacterial16S rDNA fragmentsretrievedfrom AQDS-amendedBemidji

sediment after 82 days of incubation. DGGEbands labeled in bold were sequenced, and the phylogeneticplacement of these sequences is
shown in Fig. 3.

ments, which is consistent with the fact that rates of Fe(III)

in these sedimentswere similarto those in control sediments
[46]. c 107
Since Geothrixfermentansis known to oxidize acetateand E
.5 fl~~24
Days
reduce humics [10], the number of Geothrixsequences was '106
dyof incuban82.Days
also estimatedwith MPN-PCR.However,analysisof selected E
humics- and AQDS-amended sediments indicated that
numbers of Geothrixwere always at least four orders of
aL
CD) 104
magnitudelower than Geobacteraceae(unpublished data).
103
CO)
CD

In SituEnrichment
of Geobactersat a
CIS
Petroleum-Contaminated
Aquifer 0; 101
0) 10
In order to determine whether the stimulation of Fe(III)
reductionin the Columbus aquiferwith the addition of hy-
z
drocarbonsresultedin an enrichmentof Geobacteraceae, the 0 0.1 0.5 1 2
numberof Geobacteraceae sequences in the Fe(III)-reducing
Humics Added (grams/liter)
sedimentswere enumeratedwith MPN-PCR. Sedimentsad-
jacent to the hydrocarbonemplacement were anaerobic as
Fig. 7. MPN-PCRfor Geobacteraceaein Bemidjiaquifersediment
demonstratedby the accumulationof Fe(II) (44-394 iiM) in amended with humics. Sediments were sampled after 24 and 82
the groundwaterand the fact that 50% of the iron in the days of incubation.

This content downloaded from 128.119.49.10 on Fri, 18 Sep 2015 16:35:12 UTC
All use subject to JSTOR Terms and Conditions
162 O.L. Snoeyenbos-Westet al.

CD
MPN-PCRanalysisindicated that there were 3.7 x 1 +
E 7.8 x 102 (mean ? standard deviation) Geobacteraceaese-
a) 108.-
quences in these Fe(III)-reducingsediments.In comparison,
nearbybackgroundsediments that were not affectedby the
0 hydrocarbonemplacementcontained only 7.7 x 102 ? 2.7 x
107 -
CY 106 lI1Geobacteraceaesequences. Thus, the development of
L)
CL Fe(III)-reducingconditions was associatedwith an increase
105
U"
in the number of Geobacteraceae.
2' The Geobacteraceaepopulations in the hydrocarbon-
0102
impacted sediments were further evaluated by amplifying
a) 101
aI) Geobacteraceae16S rDNA from the sediments with primers
338F-GC and Geo 825R, separatingthe PCR products with
DGGE, and excising and sequencing the resultant bands.
This procedureyielded four distinctbands,which comprised
three different sequences. Two of these sequences (Colum-
bus Geo-1 and Columbus Geo-2) were closely related Geo-
bactersequences (Fig. 3). The other sequence (Columbus
Dsf-1), which came from a faint DGGE band, was in the
ment after 82 days. Desulfuromonascluster of the Geobacteraceae[21].

sediments extractablein 0.5 N HCl was in the form of Fe(II). In SituEnrichment

of Geobacter in a
When [2-'4C]acetatewas added to the sediments it was oxi- Benzoate-AmendedAquifer
co

dized
0 to0 '4C02 without the production of 14CH4,and '4C02 Fe(IIJ)reduction appearedto be the dominant TEAPin the
a)3
production
-0)
0 10 was not inhibited in the presence of molybdate sediments at all three sampling sites of the Borden aquifer.
(data not shown). These results indicated that Fe(III) was All three sediments contained Fe(JI), but there was also
the predominantTEAP in these sediments [26]. Fe(III) availablefor microbial reduction in that Fe(II) ac-
counted for the following percentagesof the HCl-extractable
iron in the sediments:backgroundcontrol site, 63%;gate 2,
63%;and gate 3, 66%. Dissolvedhydrogenconcentrationsat
6r
ate
men 1
2a3s4 the Gate 2 samplingsite were generallyin the concentration
range typicallyfound for sediments in which Fe(II) reduc-
tion is the predominant TEAP (Fig. 9). Hydrogen was not
measured in the control sediments because of a lack of ap-
Fig.e9 hydoge
Dissolvedat over four-
conentrationsedmeatsured
propriatewells. The periodic additions of benzoate to gate 3
prdepthsion ihbthed oaionsofeprsedimen cfollecionate
welsjutaosdto indicatedthat metabolismwas not at steady-stateconditions
the BoredemnaqufesTE.Pi hs eiet and thus the hydrogen concentrationscould not be consid-
2]
ered to be diagnostic of the TEAP at this gate [28, 32].
However, hydrogen concentrations in gate 3 were higher
theBodenaqifGaste3 than in gate 2, which suggested that anaerobicmetabolism
had been stimulated in gate 3.
The apparentstimulation of anaerobicmetabolism with
the addition of benzoate to gate 3 was associatedwith much
higher numbers of Geobacteraceae sequences in this zone in
comparisonto gate 2, which receivedthe low concentrations
of chlorinatedcontaminantsand toluene but no benzoate,or
the control site, which receivedno additions (Fig. 10). Num-
bers of Geobacteraceae sequences in sediments from Gate 2

This content downloaded from 128.119.49.10 on Fri, 18 Sep 2015 16:35:12 UTC
All use subject to JSTOR Terms and Conditions
Subsurface Fe(III)-Reducing Community Structure 163

108 to those that are readily available in pure culture. As dis-

l) I
0 1~07 July 1997 cussed earlier, the two methods that were used here to
stimulate Fe(III) reduction, the addition of electron donors
liX10_ 1 November 19 and the addition of electron shuttles, are the two most likely
CD~
OD 0 10?. -
101 strategies for stimulating dissimilatory metal reduction in
metal-contaminatedsubsurfaceenvironments.Thus, the re-
10
ICP
sults suggest that the design of strategiesfor the use of dis-
similatory metal reduction as a bioremediation tool should
102 consider that Geobacaer species are likely to be integralcom-
Z 101 ponents of the metal-reducingcommunity.

100

CONTAMINANTS BENZOATE& UNCONTAMINATED Effectof AddedElectronDonorson Fe(lIl)Reductionand

ONLY CONTAMINANTSBACKGROUND MicrobialCommunities
GATE2 GATE3
The addition of electrondonors effectivelystimulatedFe(III)
Fig. 10. Numbersof targetsequencesper gramof sedimentcor- reductionin both the laboratoryand field experiments.Elec-
respondingto membersof the Geobacteraceaedetectedby MPN-
tron donors used directly by a variety of microorganisms
PCRin Bordensedimentssampledat two differenttime points.
that conserve energy to support growth via metal reduction
include acetateand other short-chainfatty acids;long-chain
and the control site were similar,indicatingthat the addition fatty acids; monoaromatic compounds, and hydrogen [23,
of the low concentrationof chlorinatedsolvents and toluene 25]. Long-chain fatty acids are generallyinsoluble and thus
had little impact on this Fe(III)-reducingpopulation. can not be readilyadded to groundwater.Hydrogen is both
In order to determinewhich Geobacteraceae were present poorly soluble and explosive, and these characteristicspre-
in the Borden aquifersediments, 16S rDNA sequences were sent technical problems when introducing it into the sub-
amplified as described above, and the PCR products were surface. Although, as shown here, formate can serve as an
separatedwith DGGE. Although each sediment treatment electron donor for Fe(III) reduction, it did not stimulate
had a somewhat different complement of Geobacteraceae Fe(III) reduction as well as the other known electron donors
bands, no single band was found to be unique to the ben- that were evaluated.Formatehas additional potential draw-
zoate-treatedsediments in either position or intensity on the backs in that it is the salt of a corrosive acid and is also
gel (data not shown). Sequence analysis revealed that 11 rapidly disproportionatedto hydrogen and carbon dioxide
bands from the Borden gel form a monophyletic group in sediments. Therefore,formate is probably a poor choice
within the Geobactercluster of the Geobacteraceae, whereas as an electron donor for stimulation of Fe(III) reduction in
one sequence that was recovered from the control site falls the subsurface.
within the Desulfuromonascluster (Fig. 3). Glucose significantly stimulated Fe(III) reduction, but
this and other fermentable compounds have the potential
drawbackthat they may stimulate the excessive growth of
Discussion fermentative bacteria. The enrichment of microorganisms
with sequencessimilarto fermentativeClostridiain the sedi-
The results demonstrate that the stimulation of Fe(III) re- ments amendedwith glucose, but not in sediments amended
duction in sandy aquifer sediments was associated with a with other electron donors, suggests that the added glucose
specific enrichment of Geobacterspecies. This was true in did stimulate the growth of fermentativebacteria.Although
field as well as laboratorystudies, regardlessof the electron at least one Fe(III)-reducingShewanellaspecies can use glu-
donors or electron shuttles used to stimulate Fe(III) reduc- cose as an electron donor [10], it was apparentlynot com-
tion. This finding is of interest not only because of the petitive with the fermentative Clostridiaas no Shewanella
consistency of the enrichment for Geobacterspecies under sequenceswere detected. Fe(III) reducerscan be expected to
diverse conditions, but also because it represents a rare utilize the fermentation products that fermentativemicro-
instance in which microorganismsfound to be environmen- organismsproduce, as evidencedhere by the increasednum-
tally significantwith moleculartechniquesare closely related bers of Geobacterin the glucose-amendedsediments. How-

This content downloaded from 128.119.49.10 on Fri, 18 Sep 2015 16:35:12 UTC
All use subject to JSTOR Terms and Conditions
164
ever, stimulatingthe growthof fermentativemicroorganisms
with glucose or other more complex fermentablesubstrates
could increase the possibility of plugging the aquifer with
biomass that does not directly contribute to Fe(III) reduc-
tion.
In contrast to fermentable sugars, acetate, lactate, and
monoaromatic compounds can be directly oxidized to car-
bon dioxide by Fe(III)-reducingmicroorganisms[23, 24]. It
was previously speculated [23] that Fe(III) reducerswith a
metabolismlike that of the various Shewanellaspecies might
first partiallyoxidize lactateto acetateand carbon dioxide in
Fe(III)-reducing environments. However, it has subse-
quently been found that some Fe(III) reducers, including
Geobacterspecies [11], can completely oxidize lactateto car-
bon dioxide. The finding that a Geobacterpopulation be-
came dominant in the lactate-amendedsediments and that
there was no enrichmentof Shewanellasequencesin lactate-
amended sediments suggests that Geobacterspecies capable
of completely oxidizing lactate were most effective at scav-
enging the added lactate.
Acetate was very effective in stimulating Fe(III) reduc-
tion. Numerous members of the Geobacteraceae family can
oxidize acetate with the reduction of Fe(III) [24], as can
Geovibrioferrireducens[9], Ferribacteriumlimneticum [12],
and Geothrixfermentans[10]. However,acetateadditionsled
primarilyto the enrichment of a Geobacterpopulation. No
sequences related to the other known acetate-oxidizing
Fe(III)-reducing microorganisms were among the most
prominent sequences recoveredin the DGGEanalysisof the
sediments. Enumeration of Geothrix 16S rDNA sequences
with specific PCR primers indicated that the growth of this
bacterium was stimulated with the addition of acetate and
other electron donors. However, Geobactersequences were
typicallythree to five ordersof magnitudegreaterthan those
of Geothrix.
Benzoate additions stimulated Fe(III) reduction in the
laboratorystudies, and benzoate and aromatichydrocarbons
resulted in the development of Fe(III)-reducingconditions
in the field studies. This result might have been expected
based on previous studies,which have demonstratedthat the
releaseof monoaromatichydrocarbonsinto groundwateras
the result of hydrocarboncontamination greatly stimulates
Fe(III) reduction in aquifers [27, 26]. Geobacterspecies are
the only Fe(III)-reducingmicroorganismsthat are known to
have the ability to oxidize monoaromatic compounds with
the reduction of Fe(III) [24]. In previous studies of the pe-
troleum-contaminated portion of the Bemidji aquifer [5,
This content downloaded from 128.119.49.10 on Fri, 18 Sep 2015 16:35:12 UTC
All use subject to JSTOR Terms and Conditions
O.L. Snoeyenbos-Westet al.
50], it was found that there was a significantenrichment of
Geobacterin the zone where the most rapid oxidation of
aromatichydrocarbonssuch as benzene [5] and toluene [4]
was taking place. DGGE analysis indicated that Geobacter
species were major components of the community in those
aromatichydrocarbon-degradingsediments,and MPN-PCR
demonstratedthat the Geobacteraceae sequencesin the most
active Fe(III)-reducingsediments were four orders of mag-
nitude more numerous than in the uncontaminatedback-
ground sediments. This increase in Geobacterin the con-
taminatedsedimentswas attributedto Geobacterspecies oxi-
dizing the aromatic hydrocarbonsat that site. The findings
reported here that the growth of Geobacterspecies was en-
hanced with the addition of benzoate to Bemidji sediments,
with the addition of benzoate to the Borden aquifer,and in
the presence of petroleum hydrocarbonsat the Columbus
site demonstratethat Geobacterspecies are likely to be im-
portant components of the Fe(JII)-reducingcommunity in a
varietyof sandy aquifersin which Fe(III)reductionis stimu-
lated with monoaromatic compounds.
Stimulation
of Fe(lIl)ReductionwithExtracellular
Quinones
Recent studies have suggested that another strategy for
stimulatingFe(III) reduction in subsurfaceenvironmentsis
the addition of humics or other extracellularquinones [30,
31, 39, 46]. Fe(III)reducerscan reducethe quinones in these
organics to the hydroquinone state, and the hydroquinones
can then abioticallyreduceFe(III) [30, 31] as well as Mn(IV)
and contaminant metals such as uranium and technetium
(unpublished data). The Fe(III) reduction step oxidizes the
hydroquinone to the quinone, which can then undergo an-
other cycle of reduction and oxidation. This alleviatesthe
need for Fe(III)-reducingmicroorganismsto come into di-
rect physicalcontactwith insolubleFe(III)oxides in orderto
reduce them. In addition, hydroquinonesgeneratedvia qui-
none reductionby Fe(III)reducersmay be able to enter pore
spaces that are too small for Fe(III) reducers to enter and
permit reduction of Fe(III) and other metals that would
otherwise be inaccessible [31]. Although there has not yet
been an opportunityto conduct a field trial on the effect of
the addition of extracellularquinones on Fe(III) reduction
in sandy aquifers, the laboratory results suggest that the
addition of humics and other extracellularquinones will
stimulate the growth and activity of Geobacterspecies in
aquifer sediments.
SubsurfaceFe(III)-ReducingCommunity Structure
Implications
of the ApparentUbiquityof GeobacterSpeciesin
Fe(lIl)-Reducing
Sedimentsof SandyAquifers
The enrichment of Geobacterspecies in response to the
stimulation of Fe(IIL)reduction in the studies reportedhere
is consistent with the known physiologicalcharacteristicsof
Geobacterspecies that have been studied in pure culture.
However, it was not necessarilyexpected that stimulation of
Fe(III) reduction in sandy aquifersediments would result in
the growth of organismsso closely relatedto those alreadyin
pure culture. For example, it is commonly considered diffi-
cult to culture the microorganismsthat are the most envi-
ronmentally significant [3]. Yet, the Geobactersequences
that predominatein the aquifersediments are closely related
to Geobacterspecies in pure culture, and we have been able
to recover organisms in culture from the Bemidji aquifer
that are closely related to the sequences that predominate
there.
The consistent enrichment of similar Geobacterspecies,
regardlessof the method for stimulating Fe(JII) reduction,
also could not have been anticipated in advance. For ex-
ample, from the known physiological characteristicsof the
Fe(III)-reducingmicroorganismsin pure culture it seemed
possible that Shewanellaspecies that can oxidize lactate and
formate with the reduction of Fe(III) [8, 37] might have
been stimulated when these electron donors were added.
The finding with moleculartechniques that Shewanellaspe-
cies were either absentfrom the sediments or presentin very
low numbers is in accordancewith recent culturing studies,
which have indicated that even when lactate or formate are
used as the electron donor, Geobacterspecies rather than
Shewanellaare recoveredfrom soils (Coates, manuscriptin
preparation).
One of the most interesting results was the finding that
the Geobacterspecies that predominated in the Fe(III)-
reducing sediments from three geographicallydistinct aqui-
fers had 16S rDNA sequences that were very similar, desig-
nated here as the "SubsurfaceGeobacterCluster."The Geo-
bactersequences that predominatedin the Fe(III)-reducing
sediments from Columbus (Columbus Geo-1 and Geo-2)
and the Borden (Borden Geo-1 through Geo-11) aquifer
sediments were not only closely related to the Geobacter
sequences in the Bemidji sediments from the laboratory
studies reportedhere, but were also very similar to the pre-
viously reported [50] Geobactersequences that were abun-
dant in the Fe(III)-reducing zone of the Bemidji aquifer
(Geo-60 and Geo-83) and in a benzene-oxidizing, Fe(III)-
This content downloaded from 128.119.49.10 on Fri, 18 Sep 2015 16:35:12 UTC
All use subject to JSTOR Terms and Conditions
165
reducing enrichment culture (Benz-76) from this site (Fig.
3). An additionalGeobactersequence (BordenGeo-101) that
is 98% similar to Geo-83 was recovered from the Borden
sediments when the 16S rDNA was amplified with the PCR
primer set designed to amplify the 16S rDNA of most Bac-
teria (Fig. 3). Although Desulfuromonassequenceswere less
predominant than Geobactersequences, three partial 16S
rDNA sequences (Humics Dsf-1, Columbus Dsf-1, and Bor-
den Dsf- 1) that fall within the DesulfuromonasClusterof the
family Geobacteraceae[21] were retrievedfrom each of the
aquifer sites (Fig. 3). They are very similar to one another
and to those found in a previousstudy of the Bemidjiaquifer
[50] and are designated here as the "SubsurfaceDesulfu-
romonasCluster"(Fig. 3).
It is becoming increasinglyapparent[14, 15] that similar
microbial species may be ubiquitous in freshwater[43, 54],
marine [16, 17], and aerobicsoil environments [19, 49]. The
results presentedhere indicate that such microbial ubiquity
may also be extended to the anaerobiczones of sandy aqui-
fers.
Conclusions
The resultsof this study, as well as previousinvestigationsof
a petroleum-contaminatedaquifer [5, 50], demonstratethat
organisms with 16S rDNA sequences closely related to
known species of Geobacterspecies are significantlyenriched
under a variety of conditions that promote dissimilatory
Fe(III) reduction in sandy aquifers.The finding that it may
be possible to isolate the most important Fe(III)-reducing
microorganismsin pure culture provides a unique oppor-
tunity to combine environmentaland culture studies to si-
multaneously investigate the biochemical and geochemical
factorsinfluencingthe rate and extent of Fe(III)reductionin
the subsurface.
Acknowledgments
The NABIR Program of the Department of Energy (Grant
DEFG0297ER62475)supportedthis work. We thank G. De-
lin and W. Larsonfor Bemidjisediment samplecollection,T.
Staufferand R. Stapletonfor coordinatingsample collection
at the Columbus site, and B.J.Butlerfor sample collection at
Borden.
166
References
1. AchenbachLM,Woese C (1995) 16Sand 23S rRNA-likeprim-
ers. In: Robb FT, PlaceAR (eds) Archaea,a laboratorymanual,
Thermophiles. Cold Spring Harbor LaboratoryPress, New
York, pp 201-203
2. Amann RI, Binder BJ, Olson RJ, Chisholm SW, Devereux R,
Stahl DA (1990) Combination of 16S rRNA-targetedoligo-
nucleotide probes with flow cytometry for analyzing mixed
microbialpopulations.Appl EnvironMicrobiol 56:1919-1925
3. Amann RI, Ludwig W, Schleifer K-H (1995) Phylogenetic
identification and in situ detection of individual microbial
cells without cultivation. Microbiol Rev 59:143-169
4. Anderson RT, Lovley DR (1999) Naphthalene and benzene
degradationunder Fe(III)-reducingconditions in petroleum-
contaminated aquifers.BioremediationJ 3:121-135
5. Anderson RT, Rooney-VargaJ, Gaw CV, Lovley DR (1998)
Anaerobicbenzene oxidation in the Fe(III)-reductionzone of
petroleum-contaminated aquifers. Environ Sci Technol
32:1222-1229
6. Boggs JF,Young SC, Beard LM (1992) Field study of disper-
sion in a heterogeneousaquifer, 1. Overviewand site descrip-
tion. Wat Resour Res 28:3281-3291
7. Boone DR, Liu Y, Zhao Z-J, BalkwillDL, Drake GT, Stevens
TO, Aldrich HC (1995) Bacillus infernussp nov, an Fe(III)-
and Mn(IV)-reducinganaerobefrom the deep terrestrialsub-
surface. Int J Sys Bacteriol45:441-448
8. Caccavo Jr F, BlakemoreRP, Lovley DR (1992) A hydrogen-
oxidizing, Fe(III)-reducingmicroorganismfrom the GreatBay
Estuary,New Hampshire. Appl Environ Microbiol 58:3211-
3216
9. Caccavo Jr F, Coates JD, Rossello-Mora RA, Ludwig W,
SchleiferKH, LovleyDR, McInerneyMJ (1996) Geovibriofer-
rireducens,a phylogenetically distinct dissimilatory Fe(III)-
reducing bacterium.Arch Microbiol 165:370-376
10. CoatesJD, DJ, EllisDR, Lovley(1999) Geothrixfermentans gen
nov sp nov, an acetate-oxidizing Fe(III) reducer capable of
growth via fermentation. Internat J Sys Bacteriol 49:1615-
1622
11. Coates JD, LonerganDJ, JenterH, LovleyDR (1996) Isolation
of Geobacterspecies from diverse sedimentaryenvironments.
Appl Environ Microbiol 62:1531-1536
12. Cummings DE, CaccavoJrF, SpringS, RosenzweigRF (1999)
Ferribacterium limneticum, gen nov, sp nov, an Fe(III)-
reducing microorganism isolated from mining-impacted
freshwaterlake sediments. Arch Microbiol 17:183-188
13. Fantroussi SA, Verschuere 1, VerstraeteW, Top EA (1999)
Effectof phenylureaherbicideson soil microbialcommunities
estimatedby analysisof 16S rRNAgene fingerprintsand com-
munity-level physiological profiles. Appl Environ Microbiol
65:982-988
14. Fenchel T, Esteban GF, Finlay BJ (1997) Local versus global
diversity of microorganisms:Crypticdiversityof ciliated pro-
tozoa. Oikos 80:220-225
This content downloaded from 128.119.49.10 on Fri, 18 Sep 2015 16:35:12 UTC
All use subject to JSTOR Terms and Conditions
O.L. Snoeyenbos-Westet al.
15. FinlayBJ,ClarkeKL (1999) Ubiquitous dispersalof microbial
species. Nature 400:828
16. FuhrmanJA,McCallumK, Davis AA (1993) Phylogeneticdi-
versity of subsurfacemarine microbial communities from the
Atlanticand PacificOceans.Appl EnvironMicrobiol59:1294-
1302
17. Giovannoni SJ, BritschgiTB, Moyer C, Field KG (1990) Ge-
netic diversity in Sargasso Sea bacterioplankton. Nature
345:60-63
18. KopczynskiED, Bateson AM, Ward DM (1994) Recognition
of chimeric small-subunit ribosomal DNAs composed of
genes from uncultivated microorganisms.AppI Environ Mi-
crobiol 60:746-748
19. Kuske CR, Barns SM, Busch JD (1997) Diverse uncultivated
bacterial groups from soils of the arid southwesternUnited
States that are present in many geographicregions. Appl En-
viron Microbiol 63:3614-3621
20. Lane DL, Pace B, Olsen GJ, Stahl D, Sogin ML, Pace NR
(1985) Rapiddeterminationof 16S ribosomalRNA sequences
for phylogenetic analysis. Proc Nat Acad Sci USA 82:6955-
6959
21. LonerganDJ, Jenter H, Coates JD, Phillips EJP, Schmidt T,
Lovley DR (1996) Phylogenetic analysis of dissimilatory
Fe(III)-reducingbacteria.J Bacteriol 178:2402-2408
22. Lovley DR (1995) Bioremediationof organic and metal con-
taminants with dissimilatorymetal reduction. J Industr Mi-
crobiol 14:85-93
23. Lovley DR (1991) Dissimilatory Fe(I1I) and Mn(IV) reduc-
tion. Microbiol Rev 55:259-287
24. Lovley DR (2000) Fe(III) and Mn(IV) reducing prokaryotes.
In: Dworkin M, Falkow S, RosenbergE, and StackebrandtE
(eds) The Prokaryotes,Springer-Verlag,New York
25. Lovley DR (1995) Microbial reduction of iron, manganese,
and other metals. Adv Agron 54:175-231
26. Lovley DR (1997) Potential for anaerobicbioremediation of
BTEX in petroleum-contaminatedaquifers.J Industr Micro-
biol 18:75-81
27. Lovley DR, BaedeckerMJ, LonerganDJ, CozzarelliIM, Phil-
lips EJP,SiegelDI (1989) Oxidationof aromaticcontaminants
coupled to microbial iron reduction. Nature 339:297-299
28. Lovley DR, Chapelle FH, Woodward JC (1994) Use of dis-
solved H2 concentrationsto determinethe distributionof mi-
crobially catalyzed redox reactions in anoxic ground water.
Environ Sci Technol 28:1205-1210
29. Lovley DR, Coates JD (1997) Bioremediationof metal con-
tamination. Curr Opin Biotech 8:285-289
30. Lovley DR, Coates JD, Blunt-HarrisEL, Phillips EJP,Wood-
ward JC (1996) Humic substances as electron acceptors for
microbial respiration.Nature 382:445-448
31. LovleyDR, FragaJL,Blunt-HarrisEL,Hayes LA, PhillipsEJP,
Coates JD (1998) Humic substancesas a mediator for micro-
bially catalyzedmetal reduction. Acta Hydrochim Hydrobiol
26:152-157
32. LovleyDR, Goodwin 5 (1988) Hydrogenconcentrationsas an
indicator of the predominant terminal electron accepting re-
 SubsurfaceFe(III)-ReducingCommunity Structure
actions in aquatic sediments. Geochim Cosmochim Acta
52:2993-3003
33. LovleyDR, Phillips EJP(1986) Organicmatter mineralization
with reduction of ferric iron in anaerobic sediments. Appl
Environ Microbiol 51:683-689
34. Lovley DR, Phillips EJP (1992) Bioremediation of uranium
contamination with enzymatic uranium reduction. Environ
Sci Technol 26:2228-2234
35. Lovley DR, Phillips EJP (1989) Requirementfor a microbial
consortium to completely oxidize glucose in Fe(IIJ)-reducing
sediments. Appl Environ Microbiol 55:3234-3236
36. Lovley DR, Phillips EJP,Gorby YA, LandaER (1991) Micro-
bial reduction of uranium. Nature 350:413-416
37. Lovley DR, Phillips EJP,Lonergan DJ (1989) Hydrogen and
formate oxidation coupled to dissimilatoryreduction of iron
or manganeseby Alteromonasputrefaciens.Appl Environ Mi-
crobiol 55:700-706
38. Lovley DR, Woodward JC (1996) Mechanisms for chelator
stimulation of microbial Fe(IlI)-oxide reduction. Chem Geol
132:19-24
39. LovleyDR, WoodwardJC, ChapelleFH (1996) Rapidanaero-
bic benzene oxidation with a varietyof chelatedFe(III) forms.
Appl Environ Microbiol 62:288-291
40. Lovley DR, Woodward JC, Chapelle FH (1994) Stimulated
anoxic biodegradationof aromatichydrocarbonsusing Fe(III)
ligands. Nature 370:128-131
41. Maclntyre WG, Boggs M, Antworth CP, StaufferTB (1993)
Degradationkinetics of aromatic organics solutes introduced
into a heterogeneousaquifer.WaterResourRes 20:4045-4051
42. MaidakBL, Cole JR, CharlesT, ParkerJ, GarrityGM, Larsen
N, Li B, LilburnTG, McCaugheyMJ, Olsen GJ, OverbeekR,
PramanikS, SchmidtTM, TiedjeJM,Woese CR (1999) A new
version of the RDP (Ribosomal DatabaseProject). Nucl Acid
Res 27:171-173
43. Methe BA, Hiorns WD, Zehr JP (1998) Contrasts between
marine and freshwater bacterial community composition:
Analysesof communities in Lake George and six other lakes.
Limnol Oceanog 43:368-374
44. Muyzer G (1999) DGGE/TGGE,a method for identifying
This content downloaded from 128.119.49.10 on Fri, 18 Sep 2015 16:35:12 UTC
All use subject to JSTOR Terms and Conditions
View publication stats
167
genes from natural ecosystems. Curr Opin Microbiol 2:317-
322
45. MuyzerG, Wall ECD, UitterlindenG (1993) Profilingof com-
plex microbialpopulationsby denaturinggradientgel electro-
phoresisanalysisof polymerasechain reaction-amplifiedgenes
coding for 16S rRNA. Appl Environ Microbiol 59:695-700
46. Nevin KP, Lovley DR (2000) Potential for nonenzymatic re-
duction of Fe(IJJ)during microbialoxidation of organic mat-
ter coupled to Fe(III) reduction. Environ. Sci. Technol.
47. Picard C, Ponsonnet C, Paget E, Nesme X, Simonet P (1992)
Detection and enumerationof bacteriain soil by direct DNA
extraction and polymerasechain reaction. Appl Environ Mi-
crobiol 58:2717-2722
48. RaskinL, StromleyJM,RittmannBE,StahlDA (1994) Group-
specific 16S rRNA hybridization probes to describe natural
communities of methanogens. Appl and Environ Microbiol
60:1232-1240
49. RobertsMS, Cohan FM (1995) Recombinationand migration
rates in natural populations of Bacillus subtilis and Bacillus
mojavensis.Evolution 49:1081-1094
50. Rooney-VargaJN, AndersonRT, FragaJL,RingelbergD, Lov-
ley DR (1999) Microbialcommunities associatedwith anaero-
bic benzene mineralization in a petroleum-contaminated
aquifer.Appl Environ Microbiol 65:3056-3063
51. SwoffordDL (1998) PAUP*,PhylogeneticAnalysisUsing Par-
simony (* and other methods) Version 4. SinauerAssociates,
Sunderland,MA
52. Van de PeerY, De WachterY (1994) TREECONfor Windows:
A softwarepackagefor the construction and drawingof evo-
lutionary trees for the Microsoft Windows environment.
Comput Applic Biosci 10:569-570
53. VargasM, KashefiK, Blunt-HarrisEL,Lovley DR (1998) Mi-
crobiological evidence for Fe(IIJ) reduction on early Earth.
Nature 395:65-67
54. ZwartG, Hiorns WD, Methe BA, AgterveldMP, HuismansR,
Nold SC, Zehr JP, LaanbroekHJ (1998) Near-identical 16S
rRNA sequences recoveredfrom lakes in North America and
Europe indicate the existence of clades of freshwaterbacteria
with global distribution.Syst Appl Microbiol 21:546-556