Geomicrobiology Journal, 24:65?73, 2007
Copyright c? Taylor & Francis Group, LLC
ISSN: 0149-0451 print / 1521-0529 online
DOI: 10.1080/01490450601134309
Returning to Their Roots: Iron-Oxidizing Bacteria Enhance
Short-Term Plaque Formation in the Wetland-Plant
Rhizosphere
Scott C. Neubauer and Gloried E. Toledo-Dura?n
Smithsonian Environmental Research Center, Edgewater, MD 21037, USA
David Emerson
American Type Culture Collection, Manassas, VA 20110, USA
J. Patrick Megonigal
Smithsonian Environmental Research Center, Edgewater, MD 21037, USA
In the wetland rhizosphere, high densities of lithotrophic Fe(II)-
oxidizing bacteria (FeOB) and a favorable environment (i.e., high
Fe(II) availability and microaerobic conditions) suggest that these
organisms are actively contributing to the formation of Fe plaque
on plant roots. We manipulated the presence/absence of an Fe(II)-
oxidizing bacterium (Sideroxydans paludicola, strain BrT) in axenic
hydroponic microcosms containing the roots of intact Juncus ef-
fusus (soft rush) plants to determine if FeOB affected total rates of
rhizosphere Fe(II) oxidation and Fe plaque accumulation. Our ex-
perimental data highlight the importance of both FeOB and plants
in influencing short-term rates of rhizosphere Fe oxidation. Over
time scales ca. 1 wk, the FeOB increased Fe(II) oxidation rates by
1.3 to 1.7 times relative to FeOB-free microcosms. Across multiple
experimental trials, Fe(II) oxidation rates were significantly cor-
related with root biomass, reflecting the importance of radial O2
loss in supporting rhizosphere Fe(II) oxidation. Rates of root Fe(III)
plaque accumulation (time scales: 3 to 6 wk) were ?70 to 83% lower
than expected based on the short-term Fe(II) oxidation rates and
were unaffected by the presence/absence of FeOB. Decreasing rates
of Fe(II) oxidation and Fe(III) plaque accumulation with increasing
time scales indicate changes in rates of Fe(II) diffusion and radial
O2 loss, shifts in the location of Fe oxide accumulation, or tempo-
ral changes in the microbial community within the microcosms.
The microcosms used herein replicated many of the environmental
characteristics of wetland systems and allowed us to demonstrate
Received 5 June 2006; accepted 30 October 2006.
Neubauer?s Current address: Baruch Marine Field Laboratory, Uni-
versity of South Carolina, Georgetown, SC 29442, USA.
Current address for Toledo-Dura?n, University of Puerto Rico at
Mayagu?ez, Mayagu?ez, Puerto Rico 00681.
We thank Kim Givler and Amy Wolfe for laboratory assistance.
We also thank two anonymous reviewers for constructive comments
that improved the manuscript. This work was supported by NSF grant
DEB-9986981 to J.P.M and D.E.
Address correspondence to Scott C. Neubauer, Baruch Marine Field
Laboratory, University of South Carolina, PO Box 1630, Georgetown,
SC 29442, USA. E-mail: scott@belle.baruch.sc.edu
that FeOB can stimulate rates of Fe(II) oxidation in the wetland
rhizosphere, a finding that has implications for the biogeochemical
cycling of carbon, metals, and nutrients in wetland ecosystems.
Keywords Fe plaque, iron oxidation, iron-oxidizing bacteria (FeOB),
Juncus effusus, Rhizosphere processes, Wetland biogeo-
chemistry
INTRODUCTION
Microbially mediated reactions and chemical processes are
responsible for the formation of iron oxyhydroxide coatings (i.e.,
Fe plaque) on the surface of many wetland plant roots. This
plaque is a visual indication that subsurface oxidative processes
are occurring in otherwise anoxic wetland soils and sediments.
Oxygen that leaks out of plant roots (a processes known as ra-
dial O2 loss; Armstrong 1964) can react with ferrous iron [i.e.,
Fe(II)] to form iron oxides that are deposited on or near plant
roots. Iron oxides may also form in the absence of molecular
O2 if NO?3 or perchlorate are used as oxidants (Straub et al.
1996; Lack et al. 2002), although the significance of anaerobic
Fe(II) oxidation in the rhizosphere is currently unknown. Root
Fe deposits are biogeochemically significant and can sequester
significant amounts of PO3?4 and metals including As, Zn, Cu,
and Pb (Taylor and Crowder 1983; Peverly et al. 1995; Weis
and Weis 2004; Chen et al. 2005). Subsequent reduction of iron
plaque can suppress methanogenesis if Fe(III)-reducing bacte-
ria outcompete methanogens for electron donors (Roden and
Wetzel 1996; Neubauer et al. 2005). Because abiotic iron oxi-
dation can be rapid under circumneutral conditions (Stumm and
Morgan 1981), it has been largely assumed that plaque forma-
tion is primarily a chemically-driven process (Mendelssohn et al.
1995). Recent evidence, however, suggests that iron-oxidizing
bacteria (FeOB) may play a key role in mediating rhizosphere
Fe(II) oxidation and plaque formation.
65
66 S. C. NEUBAUER ET AL.
The wetland rhizosphere is an interface between aerobic and
anaerobic environments and contains a diverse community of
both aerobic and anaerobic microbes. Many microbial cells are
imbedded in Fe plaque (e.g. Trolldenier 1988; St-Cyr et al.
1993), but it was not until the late 1990s that the presence of
lithotrophic FeOB in the rhizosphere was conclusively demon-
strated (Emerson et al. 1999; Weiss et al. 2003). To date, at least
4 strains of FeOB have been enriched from the rhizosphere of
wetland plants (Emerson et al. 1999; Weiss et al. in review).
Despite some phylogenetic diversity among these FeOB, all are
lithotrophic and acquire energy via the oxidation of Fe(II) under
microaerobic conditions, an environment where rates of chem-
ical Fe(II) oxidation are depressed. A survey of mid-Atlantic
(USA) wetlands and aquatic habitats revealed that these FeOB
are widespread and may comprise up to 5% of the total micro-
bial community in the wetland rhizosphere (Weiss et al. 2003).
The favorable environment in the rhizosphere (i.e., high Fe(II)
availability and microaerobic conditions) and high FeOB densi-
ties are circumstantial evidence that these organisms are actively
contributing to the formation of Fe plaque.
Methodological difficulties associated with working in com-
plex microbial-plant-soil systems have limited our ability to
quantify how lithotrophic FeOB affect rates of rhizosphere Fe(II)
oxidation and plaque formation. In a series of batch culture lab-
oratory experiments, Neubauer et al. (2002) studied the Fe(II)
oxidation kinetics of a rhizosphere-isolated lithotrophic FeOB
and determined that this bacterium mediated up to 62% of the
total (biotic + abiotic) Fe(II) oxidation. In addition to success-
fully competing with chemical reactions for Fe(II) and O2, the
FeOB were able to accelerate total Fe(II) oxidation rates by
up to 18%. There was also evidence that the FeOB inhibited
rates of chemical Fe(II) oxidation, perhaps by temporarily bind-
ing Fe(II) within a matrix of bacterially-produced exopolymers
(Neubauer et al. 2002). In the present study, we used a series
of hydroponic microcosms containing the roots of intact Juncus
effusus (soft rush) plants to determine if FeOB have an effect on
total rhizosphere Fe(II) oxidation rates. Based on our previous
work (Neubauer et al. 2002), we hypothesized that microcosms
inoculated with FeOB would have higher rates of Fe(II) oxida-
tion and Fe(III) plaque accumulation than microcosms without
FeOB. The microcosms were designed to eliminate the microbial
complexity of a natural microbial-plant-soil system so the role of
FeOB could be specifically studied, while also having the FeOB
exposed to conditions representative of the wetland rhizosphere.
METHODS
Description of Iron-Oxidizing Bacteria. Strain BrT is a neu-
trophilic Fe(II)-oxidizing bacteria (FeOB) isolated from the rhi-
zosphere of Typha latifolia (broad-lead cattail) growing in a cre-
ated marsh in Maryland. This lithoautotrophic FeOB requires
Fe(II) and microaerobic conditions for growth. It cannot grow
using heterotrophic media, formate, Mn(II), H2, or reduced sul-
fur compounds as electron donors, nor can it use NO?3 as an
electron acceptor. Genotypically, BrT is a strain of Sideroxy-
dans paludicola and lies within the ?-proteobacteria (Weiss et al.
in review). Additional details about strain BrT and other similar
rhizosphere FeOB can be found elsewhere (Emerson et al. 1999;
Neubauer et al. 2002; Weiss et al. 2003; Weiss et al. in review).
All experiments described herein were conducted with BrT cells
maintained in gradient tubes with opposing gradients of O2 and
Fe(II).
Microcosm Setup. Approximately 3 months before the
April and July experiments, Juncus effusus seeds were planted
in a 50:50 mixture of moistened peat moss and potting soil. Fol-
lowing seed germination, the soil mix was consistently saturated
with water or flooded during plant growth because the O2 de-
mand of the substrate can influence aerenchyma development
and rates of radial O2 loss (Sorrell and Armstrong 1994; Sorrell
1999). For the microcosm experiments, the roots of J. effusus
plants were gently washed to remove any adhering soil particles,
carefully inserted through a hole in the lid of each microcosm,
and the space surrounding the base of the plant stem was filled
with wax to support the plant stems and seal the opening (Fig-
ure 1). The plant roots and the underside of the microcosm lids
were surface sterilized by agitation in a 0.2% sodium dodecyl
sulfate (SDS) solution for 5 minutes, followed by an overnight
soaking in an antibiotic solution containing streptomycin (0.1
g L?1), nalidixic acid (0.1 g L?1), ampicillin (0.05 g L?1), and
amphoterecin B (0.005 g L?1) (Calhoun and King 1997). Mi-
croscopic examination of roots treated with SDS and antibiotics,
and subsequently stained with a BacLight Live/Dead viability kit
(Molecular Probes, Eugene, Oregon), showed that few live cells
remained on the root surface (data not shown). Our goal was not
to completely sterilize the microcosms for the entire 3- to 6-week
FIG. 1. Schematic of microcosms used for Fe(II) oxidation experiments.
MICROBIAL IRON OXIDATION IN THE RHIZOSPHERE 67
duration of each experiment but rather to allow us to manipulate
the presence/absence of FeOB at the start of each experiment.
Following surface sterilization, the roots were rinsed 3 times in
sterile DI water and placed into a 0.94 L glass jar containing
800 ml of sterile 0.25X Hoagland?s solution, buffered with 5
or 10 mmol L?1 2-(N-Morpholino)ethanesulfonic acid (MES).
After the microcosm lid was sealed to the jar, the hydroponic
solution inside each microcosm was bubbled overnight with 0.2
?m filter-sterilized N2 to remove O2. Following bubbling, a
filter-sterilized and deoxygenated NaHCO3 solution containing
trace vitamins and minerals was added to adjust the pH to ?6
(final NaHCO3 concentration = 5.6 mmol L?1).
Three microcosm types were established for each experi-
ment: (i) plants inoculated with live cells of the Fe(II) oxidizer
BrT (?+FeOB? microcosms; n = 10 or 20), (ii) plants inoc-
ulated with killed (autoclaved) BrT cells (??FeOB;? n = 10
or 20), and (iii) experimental blanks containing no plants or
Fe(II) oxidizers (?blanks;? n = 5). Collectively, the +FeOB and
?FeOB microcosms are called the treatment microcosms. The
experimental blanks were used to determine rates of O2 leak-
age and Fe(II) oxidation in the absence of plants or microbes.
Following inoculation, sterile deoxygenated FeCl2?4H2O was
added to each microcosm to give a final Fe(II) concentration of
?1000 ?mol L?1. Additional FeCl2? 4H2O was added as needed
whenever Fe(II) concentrations fell below ?150 ?mol L?1. To
minimize O2 leakage through the sealed lids, each microcosm
was submerged until the top of the wax-filled cylinder holding
the plant was under water. Thus, the only source of O2 to sup-
port Fe(II) oxidation was radial O2 loss from the plant roots.
All microcosms were wrapped in foil to keep light out of the
hydroponic solution and were incubated in an environmental
growth chamber at 28?C under a light regime of 14 h light:10 h
dark.
Sampling the Microcosms. At 1- to 2-day intervals, aliquots
of the hydroponic solution were removed from each microcosm
using sterile needles and syringes. The concentration of solu-
ble Fe(II) (i.e. Fe2+) was determined using 0.1% ferrozine in
50 mmol L?1 HEPES buffer; samples were read on a spec-
trophotometer at 562 nm within minutes of sample collection.
FIG. 2. Examples of data used to calculate Fe(II) oxidation rates. The plotted data were selected to illustrate some of the variability in Fe(II) oxidation rates
between microcosms. (A) Fe(II) vs. time curves for treatment microcosms (e.g., A01, A14, A29) were well-described by exponential curves of the form Fe(II)
= a? e?b?time + c; Fe(II) concentrations decreased linearly in blank microcosms (e.g. A04). (B) The rate of Fe(II) oxidation varied linearly as a function of
[Fe(II)] in the treatment microcosms. For between-microcosm comparisons, Fe(II) oxidation rates were calculated at 750 ?M Fe(II).
At the beginning of each experiment and after ?3 weeks (April:
21 d; July: 18 d) and 6 weeks (41 d, April only), a subset of
the microcosms were destructively sampled so root Fe plaque
concentrations could be measured. After the microcosms were
opened, the pH of the hydroponic solution was measured.
The roots were removed from the solution and rinsed at least
5X in sterile DI water to remove any cells and Fe that were not
bound to the root surface (i.e., were not part of the Fe plaque).
This operational approach will provide a conservative estimate
of rhizosphere Fe(II) oxidation if the influence of the roots ex-
tends beyond the Fe plaque boundary. However, this protocol
provides an easy and reproducible means of defining and dis-
tinguishing root plaque from other Fe pools and is consistent
with our previous studies that have examined root-influenced
Fe cycling (e.g., Weiss et al. 2003, 2005). Approximately half
of the roots were extracted in 0.5 mol L?1 HCl for 30 minutes.
An aliquot of the HCl extract was then added to ferrozine and
analyzed for Fe(II) (as above) and total Fe (following reduction
of Fe(III) with 0.25 mol L?1 hydroxylamine hydrochloride in
0.25 mol L?1 HCl). The green (aboveground) biomass, roots
extracted for Fe plaque, and all remaining roots were separately
dried at 80?C and weighed.
Data Analysis and Calculations. For the treatment micro-
cosms, an exponential decay equation was used to describe tem-
poral changes in soluble [Fe(II)] during the experiments (e.g.,
Figure 2A).
[Fe(II)] = a ? e(?b?time) + c [1]
where [Fe(II)] is in ?mol L?1, time is h since a pulse of Fe(II)
was added to the microcosm, and a, b, and c are empirically
determined coefficients . At any given point along the curve, the
rate of Fe(II) disappearance is the first derivative of equation 1
with respect to time, or
d[Fe(II)]/dt = ?a ? e(?b?time) ? (?b) [2]
By convention, the rate of Fe(II) oxidation is calculated as
? d[Fe(II)]/dt. In other words, when Fe(II) concentrations
68 S. C. NEUBAUER ET AL.
decrease in the microcosm (i.e., d[Fe(II)]/dt < 0), there is a pos-
itive rate of Fe(II) oxidation. Because equation 1 is non-linear,
the rate of Fe(II) oxidation varies as a function of the ambient
Fe(II) concentration (which is, itself, a function of time) (Figure
2B). To allow meaningful comparisons between microcosms and
among treatments, we calculated the Fe(II) oxidation rate at a
standard Fe(II) concentration of 750 ?mol L?1, a value that was
within the range of Fe(II) values in the microcosms across both
experimental runs and is typical of Fe(II) concentrations in wet-
land soils. Thus, we can solve for t750 (that is, the time when the
target Fe(II) concentration, Fe(II)target, equals 750 ?mol L?1) as
t750 = ln[(Fe(II)target ? c)/a]/(?b) [3]
Replacing the variable ?time? in equation 2 with the output from
equation 3 allows us to calculate the Fe(II) oxidation rate for all
microcosms at a standardized Fe(II) concentration of 750 ?mol
L?1. Multiplying the results of equation 2 by the volume of
solution in the microcosms (0.8 L) gives the Fe(II) oxidation
rate in ?mol Fe(II) oxidized h?1.
For each experiment, the maximum rate of Fe(II) oxidation
(determined by linear regression) across all replicate blank mi-
crocosms was subtracted from Fe(II) oxidation in the treatment
microcosms to correct for any Fe(II) oxidation that was occur-
ring due to O2 leakage through the microcosm seals. Micro-
cosms that had Fe(II) oxidation rates lower than those in the
blank microcosms were excluded from all subsequent analyses
(see below for additional discussion). In the treatment micro-
cosms, rates of blank-corrected Fe(II) oxidation were expressed
on a per-microcosm basis and were also normalized to dry green
(i.e., aboveground) biomass and dry root biomass.
Statistics. Exponential or linear regressions were fit to the
Fe(II) versus time data, as described previously. Differences in
Fe(II) oxidation rates, root plaque concentrations, plant biomass,
and pH as a function of treatment (+FeOB vs. ?FeOB), month
(April vs. July), and time of sampling (0, 3, and 6 weeks) were
assessed using t-tests and standard least squares models, as ap-
propriate. Unless otherwise noted, statistical significance was
set at p ? 0.10 since preliminary experiments showed a high
level of variability between microcosms but it was logistically
unfeasible to increase the number of replicate microcosms in
each experiment. Thus, we are more likely incorrectly to reject
the null hypothesis of no difference between treatments (type I
error) than if a more conservative alpha level of 0.05 had been
used, but less likely to incorrectly accept the null hypothesis
(type II error). All curve fitting and statistical analyses were
conducted using JMP v.5.0 (SAS Institute, Cary, NC).
RESULTS
Temporal Patterns of Microcosm [Fe(II)] and Oxidation
Rates. For the microcosm experiment initiated in April, the
decrease in [Fe(II)] with time was well-described by an expo-
nential equation (equation 1 and Figure 2A), with r2 values for
treatment microcosms ranging from 0.96 to 1.0. In the blank
microcosms, Fe(II) concentrations decreased linearly with time,
with r2 values of 0.87 to 0.98. Fe(II) oxidation rates calculated
at a standard [Fe(II)] of 750 ?mol L?1 ranged from 0.8 to 5.3
?mol Fe(II) oxidized h?1 for the treatment microcosms (n = 20);
rates were significantly lower in the blank microcosms (0.17 to
0.31 ?mol h?1; n = 4). For microcosms where multiple pulses
of Fe(II) were added, reported Fe(II) oxidation rates are only
for the initial Fe(II) addition. There were no consistent trends
in Fe(II) oxidation rates across multiple Fe(II) pulses within a
single microcosm (data not shown).
In July, equation 1 was a good fit to the Fe(II) versus time
data for approximately half of the treatment microcosms (n =
24 of 40 total, average r2 = 0.90). For the remaining treatment
microcosms (n = 16) and the blanks (n = 5), linear regressions
were fit to the data (r2 = 0.69 to 0.96). At a Fe(II) concentration
of 750 ?mol L?1, the July Fe(II) oxidation rates in the treatment
microcosms (0.17 to 2.06 ?mol h?1) were significantly lower
than in April, whereas rates in the blank microcosms (0.46 to
0.81 ?mol h?1) were higher than in April. As a result, some of
the treatment microcosms from July were excluded from further
analysis because calculated oxidation rates were lower than in
the blank microcosms, presumably due to lower rates of radial
O2 loss as well as experimental difficulties in limiting O2 leakage
in the blank microcosms.
Effect of FeOB on Fe(II) Oxidation Rates. After correcting
for Fe(II) oxidation in the blank microcosms, Fe(II) oxidation
rates in the treatment microcosms were expressed as whole mi-
crocosm rates (?mol Fe(II) oxidized h?1) and as normalized to
dry green biomass and root biomass (?mol g?1 h?1). In April,
average Fe(II) oxidation rates tended to be greater in the +FeOB
relative to the ?FeOB microcosms when expressed as whole mi-
crocosm rates (2.78 ? 0.43 vs. 2.13 ? 0.43 ?mol h?1; averages
? 1 standard error), and when normalized to green biomass (6.95
? 0.91 vs. 5.33 ? 0.90 ?mol g?1 h?1) or root biomass (6.05 ?
0.81 vs. 5.14 ? 0.83 ?mol g?1 h?1) but these differences were
not statistically significant (p > 0.10; Figure 3). In July, whole
microcosm oxidation rates were 1.6 times greater in +FeOB mi-
crocosms (0.45 ? 0.10 vs. 0.28 ? 0.06 ?mol h?1; p = 0.08). As
in April, there was a trend for greater biomass-normalized oxi-
dation rates in the +FeOB microcosms (green-biomass normal-
ized: 1.45 ? 0.24 vs. 1.06 ? 0.29 ?mol g?1 h?1; root-biomass
normalized: 8.31 ? 3.42 vs. 3.63 ? 0.91 ?mol g?1 h?1).
The Fe(II) oxidation rates were also expressed relative to
average ?FeOB oxidation rates within that month (e.g., right
axes on Figure 3). When the data were pooled in this manner, we
found that Fe(II) oxidation rates in the+FeOB microcosms were
greater than in the ?FeOB microcosms, regardless of whether the
data were calculated as whole microcosm rates (1.5x greater, p
= 0.04) or normalized to green biomass (1.3x greater, p = 0.06)
or root biomass (1.7x greater, p = 0.08). The raw rate data could
not be similarly pooled because Fe(II) oxidation rates differed
substantially between experiments (note scales on April and July
y-axes in Figure 3).
MICROBIAL IRON OXIDATION IN THE RHIZOSPHERE 69
FIG. 3. Fe(II) oxidation rates expressed as: (A) whole microcosm rates, and as normalized to (B) dry green plant biomass and (C) dry root biomass. All rates
have been corrected for Fe(II) oxidation in blank microcosms. Fe(II) oxidation is expressed both as a raw rate (left axes) and relative to the average oxidation rate
in the ?FeOB microcosms (values provided on right axis). Note that the right axis applies for all graphs within a particular row, but the scale of the left axes within
a row varies between graphs. One-way t-tests were used to test the hypothesis that Fe(II) oxidation rates were greater in microcosms with FeOB than in ?FeOB
microcosms. n.s. = not significant (i.e., p > 0.10). Error bars are ? 1 standard error.
Fe Plaque Accumulation. In April, total root plaque Fe con-
centrations [Fe(III) + Fe(II)] increased significantly from 11.7
?mol g?1 at the beginning of the experiment to ?950 ?mol g?1
after 3 to 6 weeks (Table 1). There was no difference in total
Fe plaque accumulation between the 3 and 6 week sampling
points. In July, there was an order of magnitude increase in root
Fe plaque between 0 and 3 weeks, from 81.9 to 850.7 ?mol g?1.
Across the sampling dates, Fe(III) accounted for 82 to 91% of
the total plaque Fe, with the exception of the initial sampling
in July (?70% of plaque Fe was Fe(III); Table 1). Long-term
Fe(III) accumulation rates averaged ?40 ?mol g?1 d?1 for the
TABLE 1
Root Fe plaque concentrations and accumulation rates
Time sampled Plaque Fe(II) + Fe(III) % Fe(III) Fe(III) accumulation rate
Month (d since start) n (?mol g?1) (% of total) (?mol g?1 d?1)
April 0 10 11.7 ? 0.9a 84.4 ? 3.4a ?
21 10 943.8 ? 135.6c 88.4 ? 2.4a 41.0 ? 6.5a
41 10 955.1 ? 92.2c 91.4 ? 1.5a 22.6 ? 2.3b
July 0 10 81.9 ? 11.0b 67.6 ? 5.1b ?
18 21? 850.7 ? 98.7c 81.6 ? 1.9a 39.8 ? 5.0a
Values are means ? standard error. Within a column, values with the same superscript were statistically similar (standard least-squares model
with Tukey?s hsd, p < 0.05).
?n = 22 for % Fe(III).
samples analyzed after 3 weeks, and 22.6 ?mol g?1 d?1 for the
6 week microcosms sampled in April (Table 1). For each month
and sampling date, there were no differences in any root plaque
parameter (concentrations, %Fe(III), or long-term Fe(III) accu-
mulation rate) between the +FeOB and ?FeOB treatments (data
not shown).
General Microcosm Parameters. The biomass of above-
ground (?green?) vegetation averaged ?0.4 g in each month
(Table 2) and did not vary between months or as a function
of time of sampling (i.e., initial microcosms vs. those sampled
at end of experiments). Similarly, root biomass did not differ
70 S. C. NEUBAUER ET AL.
TABLE 2
Plant biomass
Time sampled (d since start) n Green biomass (g dry weight) Root biomass (g dry weight)
April 0 10 0.397 ? 0.038a 0.478 ? 0.038a
21 10 0.424 ? 0.035a 0.434 ? 0.043a
41 10 0.385 ? 0.044a 0.442 ? 0.042a
July 0 9 0.406 ? 0.098a 0.086 ? 0.029b
18 21 0.458 ? 0.062a 0.106 ? 0.013b
Values are means ? standard error. Within a column, values with the same superscript were statistically similar (standard least-squares model
with Tukey?s hsd, p < 0.05).
between the beginning of each experiment and in destructively
sampled microcosms after 3 or 6 weeks. However, average root
biomass was about 4 times greater in April than in July (Table
2). Within each month, there were no differences in biomass be-
tween +FeOB and ?FeOB microcosms (data not shown). The
pH of the hydroponic solution ranged from 5.8 to 6.0 and did
not vary between months, treatments, or time of sampling (data
not shown).
DISCUSSION
To date, all measurements of the activity of circumneutral,
lithotrophic, Fe(II)-oxidizing bacteria (FeOB) have been con-
ducted in experimental systems (e.g., diffusion gradient tubes
and bioreactors, Emerson and Revsbech 1994; Sobolev and Ro-
den 2001; Neubauer et al. 2002) that eliminate much of the
complexity of a natural wetland plant-microbe-soil system. It
is methodologically difficult to study these FeOB in situ be-
cause there are no known specific inhibitors for microbial Fe(II)
oxidation, and both stable and radioisotope techniques have
unresolved issues (Roden and Lovley 1993; Emerson 2000;
Bullen et al. 2001; Croal et al. 2004). In the FeOB-Juncus ef-
fusus microcosms discussed herein, we have replicated one el-
ement of the rhizosphere environment (namely, O2 provided
only via radial O2 loss from roots) within a system where it
is (relatively) easy to manipulate the presence or absence of
FeOB and measure rates of Fe(II) oxidation and Fe plaque
accumulation.
Factors Affecting Rates of Fe(II) Oxidation. Our experi-
ments showed that FeOB increased Fe(II) oxidation rates by 1.3
to 1.7 times relative to ?FeOB (i.e., FeOB-free) microcosms.
The differences between +FeOB and ?FeOB microcosms per-
sisted when Fe(II) oxidation rates were expressed on a whole
microcosm basis, or normalized to the aboveground or below-
ground biomass of the plants. Our finding that FeOB accelerated
rates of Fe(II) oxidation is consistent with that of Neubauer et al.
(2002) who found that FeOB strain BrT (the same FeOB used
herein) could accelerate total Fe(II) oxidation rates by up to 18%
relative to cell-free treatments. Similarly, Fe(II) oxidation rates
were higher in FeOB-containing microbial mat and groundwater
samples than in killed controls (Emerson and Revsbech 1994;
James and Ferris 2004). However, other studies have reported
that Fe(III) accumulation rates in diffusion-limited gradient cul-
tures did not differ between +FeOB cultures and abiotic con-
trols (Emerson and Moyer 1997; Sobolev and Roden 2004).
Regardless of the net effect of FeOB on Fe(II) oxidation rates,
these lithotrophic organisms are utilizing energy released by
Fe(II) oxidation to support cellular metabolism and growth. Our
data did not allow us to determine if the observed increases in
Fe(II) oxidation in the +FeOB microcosms were due solely to
microbially-mediated oxidation or if rates of abiotic oxidation
also increased.
For FeOB to accelerate Fe(II) oxidation rates, the microbes
must increase the availability of Fe(II) or O2, or affect the re-
action kinetics. If chemical Fe(II) oxidation is limited by Fe(II)
or O2 availability, FeOB could increase total rates by oxidizing
Fe(II) pools that cannot be accessed chemically, by oxidizing
Fe(II) at lower substrate concentrations than chemical oxida-
tion, or by increasing rates of Fe(II) and/or O2 diffusion to the
rhizosphere. If Fe(II) oxidation rates are not limited by substrate
availability, rates could increase if FeOB increase the kinetics of
the reaction. This could occur enzymatically or if the surfaces of
FeOB cells significantly increase chemical oxidation. Rates of
Fe(II) oxidation could have been higher in +FeOB microcosms
if FeOB were able to utilize Fe(II) that was otherwise bound
to organic matter (e.g., as hypothesized by Emerson and Moyer
1997; Neubauer et al. 2002) and was unavailable for chemical
oxidation in the ?FeOB microcosms. Previous work has docu-
mented that Fe(II)-organic matter interactions are complex and
can either inhibit, accelerate, or have no effect on rates of Fe(II)
oxidation, depending on the concentrations and nature of organic
matter, as well as concentrations of O2 and Fe(II) (Theis and
Singer 1974; Johnson-Green and Crowder 1991; Stone 1997;
Roth et al. 2000). Regardless of the mechanism, enhanced con-
sumption of both Fe(II) and O2 in the rhizosphere of the +FeOB
microcosms due to Fe(II) oxidation would lead to steeper con-
centration gradients that would further drive the delivery of both
Fe(II) and O2 to the root surface. Indeed, the presence of FeOB
in the rhizosphere does indicate that these microbes are suc-
cessfully competing for (and consuming) both Fe(II) and O2.
However, in a natural microbial-plant-soil system, however, it is
unlikely that FeOB would dramatically affect O2 concentration
MICROBIAL IRON OXIDATION IN THE RHIZOSPHERE 71
FIG. 4. Whole microcosm Fe(II) oxidation rates (?mol h?1) were not related
to dry green biomass. In contrast, dry root biomass was a significant predictor of
the Fe(II) oxidation rate across both experiments: Rate = 5.91 ? root biomass
? 0.18. For each panel, statistics are for all data points, regardless of month or
treatment.
gradients due to the presence of other aerobic microbes and O2-
consuming chemical reactions in the rhizosphere.
There were significant differences in rates of Fe(II) oxi-
dation between the April and July experiments, regardless of
treatment, that suggest that the presence/absence of FeOB is
not the only factor that affects Fe(II) oxidation rates. When
data from the two trials were combined, there was a strong
correlation between whole microcosm Fe(II) oxidation rates and
root biomass (Figure 4), indicating that between-month differ-
ences in root biomass (Table 2) explained a large part of the
between-month differences in Fe(II) oxidation rates. In con-
trast, there was no relationship between green biomass and
Fe(II) oxidation rates (Figure 4). Roots can affect Fe(II) oxi-
dation by controlling rates of radial O2 loss to the rhizosphere
(i.e., more root mass = more radial O2 loss), providing sur-
face area for FeOB to colonize, or affecting rates of autocat-
alytic Fe(II) oxidation onto existing Fe(III) oxides on the root
surface.
There is considerable between-plant variability in rates of ra-
dial O2 loss and the mechanisms that drive O2 transport through
aerenchyma (e.g., passive diffusion vs. through-flow of gases)
(Colmer 2003), but several field-based studies support a con-
nection between plant biomass and rates of Fe(II) oxidation.
For example, Sundby et al. (2003) found that seasonal cycles
of root growth and decay in a salt marsh affected the degree
of rhizosphere oxidation and porewater Fe2+ availability. Sim-
ilarly, Weiss et al. (2005) suggested that temporal variations
in porewater Fe2+ concentrations in a J. effusus wetland were
driven by plant-microbial-environmental interactions that affect
rates of rhizosphere Fe(II) oxidation and Fe(III) reduction. In a
tidal freshwater marsh, high rates of radial O2 loss and Fe(II)
oxidation coincident with peak aboveground biomass were hy-
pothesized to account for the high contribution of Fe(III) reduc-
tion to total anaerobic metabolism (Neubauer et al. 2005). As
in these field studies, our microcosm experiments suggest that
plant biomass plays a key role in driving rates of radial O2 loss
and rhizosphere Fe(II) oxidation.
In addition to plant activity (as proposed above), Fe(II) oxida-
tion rates can be affected by pH, temperature, the Fe(II) supply
rate, and concentrations of O2, Fe(II), and Fe(III) (Singer and
Stumm 1970; Theis and Singer 1974; Sung and Morgan 1980;
Stumm and Morgan 1981; Liang et al. 1993). However, these
factors were probably not major contributors to the between-
month differences in Fe(II) oxidation rates in this study. The
same experimental protocol was followed in the April and July
experiments, so the pH, temperature, composition of the hydro-
ponic solution, and concentrations of Fe(II) were comparable be-
tween months. Furthermore, the Fe(II) oxidation rates shown in
Figure 3 were all calculated at a Fe(II) concentration of 750?mol
L?1, eliminating another possible difference between months.
Fe(II) Oxidation versus Fe(III) Accumulation. Measuring
rates of Fe(II) disappearance provides a relatively short-term
measure of Fe(II) oxidation rates, whereas the accumulation of
plaque onto plant roots represents a longer, more-integrated in-
dicator of Fe(II) oxidation. In contrast to the above-described
Fe(II) oxidation rate data, there were no long-term differences
in Fe plaque accumulation between +FeOB and ?FeOB micro-
cosms. Furthermore, short term Fe(II) oxidation rates were 3 to 6
times higher than Fe(III) accumulation rates (e.g., compare Fig-
ure 3 with Table 1) which, in part, is because Fe(II) oxidation
rates were calculated at 750 ?mol Fe(II) L?1 whereas Fe(III)
plaque accumulation rates were integrated over the entire range
of Fe(II) concentrations (?100 to 1000 ?mol L?1) throughout
each incubation.
Differences between Fe(II) oxidation and Fe(III) accumula-
tion rates over time could result from temporal changes in the
availability of Fe(II) and or O2 in the rhizosphere, or the com-
position of the microbial community. As roots become coated
with Fe plaque (typical thickness of 10s of ?m, but up to 0.4
cm; e.g., Taylor et al. 1984; Vale et al. 1990) and the site of
oxidation moves farther from the root surface, increased resis-
tance to diffusion will theoretically decrease the rate at which
Fe(II) or O2 are made available for oxidation. In each experi-
ment, we observed visible Fe oxide accumulation on the plant
roots and in the hydroponic solution. If the surface chemistry of
the roots and Fe plaque changed sufficiently during the course
of the incubations, the focal site of Fe(III) accumulation could
have shifted from the root surface early in the incubations to
the bulk hydroponic solution later during the experiments. This
is highlighted in the April experiment where the ratio of ac-
cumulated root Fe plaque (?mol microcosm?1) to total Fe(II)
oxidized (?mol microcosm?1; based on Fe(II) disappearance
across all Fe pulses) averaged 0.49 ? 0.04 after 3 weeks and
0.31 ? 0.03 after 6 weeks (n = 10 microcosms per time point),
demonstrating that the fraction of oxidized iron associated with
the roots decreased throughout the duration of the experiment.
In part, the accumulation of Fe oxides in the hydroponic solution
is an artifact of using a hydroponic system. In a root/soil system,
the soil matrix itself would help keep oxides in close associa-
tion with the roots. Either of the above mechanisms would cause
measured root Fe(III) accumulation rates to decline over time.
72 S. C. NEUBAUER ET AL.
However, the lack of consistent patterns in Fe(II) oxidation rates
over the 3 to 6 weeks of each experiment indicates that a simple
unidirectional change in rates of radial O2 loss or diffusion rates
is unlikely to explain differences between Fe(II) oxidation and
Fe(III) accumulation rates.
In a study that measured rates of Fe(III) plaque accumulation
and reduction on the roots of Juncus effusus in the field, Weiss et
al. (2005) reported that rates of plaque accumulation averaged
0.08 to 0.24 ?mol Fe(III) g?1 h?1 (averaged over 5 months;
Fe(III) was ?79% of total Fe plaque). Average rates of Fe(III)
reduction were considerably higher and illustrate that Fe(III)
plaque accumulation in the field is the net result of the compet-
ing processes of Fe(II) oxidation and Fe(III) reduction. While
our sterilization techniques substantially reduced the numbers
of cells on root surfaces (as verified by microscopy), it is un-
likely the sterilization was 100% effective. Thus, some microbes
probably survived, either by antibiotic resistance or by ?hiding?
inside the roots, and subsequently recolonized the root surface
during the 3 to 6 week experiments. If Fe(III) reducers were
present in the community of surviving microbes, longer-term
Fe(III) accumulation rates may represent net Fe(II) oxidation
whereas the short-term Fe(II) oxidation rates may be closer to
gross oxidation rates. Population increases of other microaer-
obes within the microcosms would also have reduced the amount
of O2 available to support Fe(II) oxidation.
The change in Juncus effusus Fe oxidation and accumulation
rates over time was well-described by an exponential equation
that is similar in form to equation 1 (Figure 5), even though
the data set includes a combination of both laboratory (this
study; Snowden and Wheeler 1995) and field data (Weiss et
al. 2005). Together, these data indicate that changes in Fe oxida-
tion and accumulation rates over time are not simply a func-
tion of artifacts associated with our experimental set-up but
instead reflect fundamental changes in the underlying mecha-
nisms responsible for plaque formation (e.g., rates of Fe(II) dif-
fusion and radial O2 loss) or temporal changes in the microbial
community.
FIG. 5. Fe(II) oxidation rates on Juncus effusus roots decrease over time.
Data include the short-term Fe(II) oxidation rates shown in Figure 3, Fe(III)
accumulation rates reported in Table 1, laboratory Fe accumulation rates from
Snowden and Wheeler (1995), and Fe(III) accumulation rates calculated from
Weiss et al. (2005) based on field data. The regression line fit to the data is in the
form of eq. 1, where a = 6.87, b = 0.53, and c = 0.24; regression r2 = 0.94.
Symbols are: ? this study, April;  this study, July;  Snowden and Wheeler;
? Weiss et al. transplants; and 5 Weiss et al. root in-growth data.
Environmental Relevance. Circumneutral FeOB have been
found on the roots of Juncus spp. or in the soils of Juncus-
dominated wetlands in Virginia, West Virginia, and Alabama
(Emerson et al. 1999; Sobolev and Roden 2001; Weiss et al.
2003). The presence of FeOB on the roots of J. effusus may
be common since this plant can have moderate to high rates of
radial O2 loss (Sorrell 1999; Wie?ner et al. 2002) and a high
tolerance to dissolved Fe(II) (Snowden and Wheeler 1995).
The environmental conditions used in our microcosm exper-
iments approximate conditions in the rhizosphere and near-
surface soils and sediments. The pH used in the microcosms
(5.8 to 6.0) is within the range of porewater pH values re-
ported in a study that found circumneutral FeOB in 13 diverse
wetland and aquatic habitats (Weiss et al. 2003). Porewater
Fe(II) concentrations are highly variable between wetlands and
vary spatially (e.g., with depth or distance from a root) and
temporally.
By periodically replenishing Fe(II) within the microcosms,
we were able to prevent Fe(II) limitation of microbial Fe oxida-
tion and maintain Fe(II) concentrations between ?100 to 1000
?mol L?1, values that are typical of many wetland environ-
ments where FeOB have been observed or dynamic Fe cycles
have been proposed (e.g., Roden and Wetzel 1996; Weiss et al.
2003; Neubauer et al. 2005; Weiss et al. 2005).
Unlike other studies that have examined the effects of FeOB
in laboratory settings, this study mimicked the rhizosphere in
that radial O2 loss from plant roots was the only source of O2 to
support Fe(II) oxidation. To determine how FeOB affect Fe(II)
oxidation rates, it was necessary to simplify the microcosms to
exclude other microbes. Clearly, the rhizosphere in a natural
setting will have a diverse and complex mixture of both aer-
obic and anaerobic microbes that are competing for resources
including O2, Fe(II), and carbon. Determining rates of microbial
Fe(II) oxidation within a larger wetland microbial community
remains a daunting challenge that is not likely to be overcome
until the in situ activities of these organisms can be identified or
it becomes possible to selectively inhibit microbially-mediated
Fe(II) oxidation.
FeOB appear to be ubiquitous in circumneutral wetlands
(Weiss et al. 2003), but little is known about the factors that
control their distribution, abundance, activity, and ecological
importance. Is FeOB activity affected by plant biology (e.g.,
variations in root O2 loss rate), microbial biology (e.g., compe-
tition with other aerobic microbes), or physiochemical variables
(e.g., soil mineral content or pH)? Do the factors that increase
FeOB activity also positively affect Fe(III) reducing bacteria,
leading to an active rhizosphere Fe cycle, or are there condi-
tions under which Fe plaque can accumulate and increasingly
sequester metals and nutrients? Are there geochemical differ-
ences between Fe oxides produced via microbial or abiotic pro-
cesses that affect ecosystem-scale metal sequestration or rates
of Fe(III) reduction?
Given the importance of root Fe plaque in the biogeochem-
ical cycling of carbon, nutrients, and metals, our finding that
FeOB increase rates of Fe(II) oxidation in the rhizosphere raises
MICROBIAL IRON OXIDATION IN THE RHIZOSPHERE 73
important questions about the largely un-quantified role
that these organisms play in complex microbial-plant-soil-
environmental interactions.
REFERENCES
Armstrong W. 1964. Oxygen diffusion from the roots of some British bog plants.
Nature 204:801?802.
Bullen TD, White AF, Childs CW, Vivit DV, Schultz MS. 2001. Demonstration
of a significant iron isotope fractionation in nature. Geology 29:699?702.
Calhoun A, King GM. 1997. Regulation of root-associated methanotrophy by
oxygen availability in the rhizosphere of two aquatic macrophytes. Appl En-
viron Microbiol 63:3051?3058.
Chen Z, Zhu Y-G, Liu W-J, Meharg AA. 2005. Direct evidence showing the
effect of root surface iron plaque on arsenite and aresnate uptake in rice
(Oryza sativa) roots. New Phytol 165:91?97.
Colmer TD. 2003. Long-distance transport of gases in plants: a perspective
on internal aeration and radial oxygen loss from roots. Plant Cell Environ
26:17?36.
Croal LR, Johnson CM, Beard BL, Newman DK. 2004. Iron isotope fractionation
by Fe(II)-oxidizing photoautotrophic bacteria. Geochim Cosmochim Acta
68:1227?1242.
Emerson D. 2000. Microbial oxidation of Fe(II) and Mn(II) at circumneutral pH.
In: Lovley DR, editor. Environmental Microbe-Metal Interactions. Washing-
ton, DC: American Society for Microbiology. p 31?52.
Emerson D, Moyer C. 1997. Isolation and characterization of novel iron-
oxidizing bacteria that grow at circumneutral pH. Applied and Environmental
Microbiology 63:4784-4792.
Emerson D, Revsbech NP. 1994. Investigation of an iron-oxidizing microbial
mat community located near Aarhus, Denmark: Laboratory studies. Appl
Environ Microbiol 60:4032?4038.
Emerson D, Weiss JV, Megonigal JP. 1999. Iron-oxidizing bacteria are associated
with ferric hydroxide precipitates (Fe-plaque) on the roots of wetland plants.
Appl Environ Microbiol 65:2758?2761.
James RE, Ferris FG. 2004. Evidence for microbial-mediated iron oxidation at
a neutrophilic groundwater spring. Chem Geol 212:301?311.
Johnson-Green PC, Crowder AA. 1991. Iron oxide deposition on axenic and
non-axenic roots of rice seedlings (Oryza sativa L.). J Plant Nutr 14:375?
386.
Lack JG, Chaudhuri SK, Chakraborty R, Achenbach LA, Coates JD. 2002.
Anaerobic biooxodation of Fe(II) by Dechlorosoma suillum. Microb Ecol
43:424?431.
Liang L, McNabb JA, Paulk JM, Gu B, McCarthy JF. 1993. Kinetics of Fe(II)
oxygenation at low partial pressure of oxygen in the presence of natural
organic matter. Environ Sci Technol 27:1864?1870.
Mendelssohn IA, Kleiss BA, Wakely JS. 1995. Factors controlling the formation
of oxidized root channels: a review. Wetlands 15:37?46.
Neubauer SC, Emerson D, Megonigal JP. 2002. Life at the energetic edge:
Kinetics of circumneutral iron oxidation by lithotrophic iron-oxidizing bac-
teria isolated from the wetland-plant rhizosphere. Appl Environ Microbiol
68:3988?3995.
Neubauer SC, Givler K, Valentine S, Megonigal JP. 2005. Seasonal patterns
and plant-mediated controls of subsurface wetland biogeochemistry. Ecology
86:3334?3344.
Peverly JH, Surface JM, Wang T. 1995. Growth and trace metal absorption by
Phragmites australis in wetlands constructed for landfill leachate treatment.
Ecol Eng 5:21?35.
Roden EE, Lovley DR. 1993. Evaluation of 55Fe as a tracer of Fe(III) reduction
in aquatic sediments. Geomicrobiol J 11:49?56.
Roden EE, Wetzel RG. 1996. Organic carbon oxidation and suppression
of methane production by microbial Fe(III) oxide reduction in vegetated
and unvegetated freshwater wetland sediments. Limnol Oceanogr 41:1733?
1748.
Roth RI, Panter SS, Zegna AI, Levin J. 2000. Bacterial endotoxin (lipopolysac-
charide) stimulates the rate of iron oxidation. J Endotoxin Res 6:313?319.
Singer PC, Stumm W. 1970. Acidic mine drainage: The rate-determining step.
Science 167:1121?1123.
Snowden RED, Wheeler BD. 1995. Chemical changes in selected wetland plant
species with increasing Fe supply, with specific reference to root precipitates
and Fe tolerance. New Phytol 131:503?520.
Sobolev D, Roden EE. 2001. Suboxic deposition of ferric iron by bacteria in
opposing gradients of Fe(II) and oxygen at circumneutral pH. Appl Environ
Microbiol 67:1328?1334.
Sobolev D, Roden EE. 2004. Characterization of a neutrophilic, chemolithoau-
totrophic, Fe(II)-oxidizing ?-protobacterium from freshwater wetland sedi-
ments. Geomicrobiol J 21:1?10.
Sorrell BK. 1999. Effect of external oxygen demand on radial oxygen loss by
Juncus roots in titanium citrate solutions. Plant, Cell Environ 22:1587?1593.
Sorrell BK, Armstrong W. 1994. On the difficulties of measuring oxygen release
by root systems of wetland plants. J Ecol 82:177?183.
St-Cyr L, Fortin D, Campbell PGC. 1993. Microscopic observations of the iron
plaque of a submerged aquatic plant (Vallisneria americana Michx). Aquat
Bot 46:155?167.
Stone AT. 1997. Reactions of extracellular organic ligands with dissolved metal
ions and mineral surfaces. In: Banfield J, Nealson KH, editors. Geomicrobi-
ology: Interaction between Microbes and Minerals. Washington, DC: Miner-
alogical Society of America. p 309?344.
Straub KL, Benz M, Schink B, Widdel F. 1996. Anaerobic, nitrate-dependent
microbial oxidation of ferrous iron. Appl Environ Microbiol 62:1458?1460.
Stumm W, Morgan JJ. 1981. Aquatic Chemistry: An Introduction Emphasizing
Chemical Equilibria in Natural Waters. New York: Wiley-Interscience. 780 p.
Sundby B, Vale C, Caetano M, Luther III GW. 2003. Redox chemistry in the root
zone of a salt marsh sediment in the Tagus Estuary, Portugal. Aquat Geochem
9:257?271.
Sung W, Morgan JJ. 1980. Kinetics and product of ferrous iron oxygenation in
aqueous systems. Environ Sci Technol 14:561?568.
Taylor GJ, Crowder AA. 1983. Uptake and accumulation of heavy metals by
Typha latifolia L. in wetlands of the Sudbury, Ontario region. Can J Bot
61:63?73.
Taylor GJ, Crowder AA, Rodden R. 1984. Formation and morphology of an iron
plaque on the roots of Typha latifolia L. grown in solution culture. American
Journal of Botany 71:666-675.
Theis TL, Singer PC. 1974. Complexation of iron(II) by organic matter and its
effect on iron(II) oxygenation. Environ Sci Technol 8:569?573.
Trolldenier G. 1988. Visualization of oxidizing power of rice roots and of pos-
sible participation of bacteria in iron deposition. Zeitschrift Fur Pflanzen-
erna?hrung Und Bodenkunde 151:117?121.
Vale C, Caterino F, Cortesa?o C, Cac?ador I. 1990. Presence of metal-rich rhizo-
concretions on the roots of Spartina maritima from the salt marshes of the
Tagus Estuary, Portugal. Sci Total Environ 97/98:617?626.
Weis JS, Weis P. 2004. Metal uptake, transport and release by wetland plants:
Implications for phytoremediation and restoration. Environ Inter 30:685?
700.
Weiss JV, Emerson D, Backer SM, Megonigal JP. 2003. Enumeration of Fe(II)-
oxidizing and Fe(III)-reducing bacteria in the root zone of wetland plants:
Implications for a rhizosphere iron cycle. Biogeochemistry 64:77?96.
Weiss JV, Emerson D, Megonigal JP. 2005. Rhizosphere iron(III) deposition
and reduction in a Juncus effusus L.-dominated wetland. Soil Sci Soc Amer
J 69:1861?1870.
Weiss JV, Rentz JA, Plaia T, Neubauer SC, Floyd MM, Lilburn T, Bradburne C,
Megonigal JP, Emerson D. in review. Phenotypic and molecular characteriza-
tion of neutrophilic Fe(II)-oxidizing bacteria isolated from the wetland plant
rhizosphere and description of Ferritrophicum radicicola gen. nov. sp. nov,
and Sideroxydans paludicola sp. nov. Geomicrobiol J.
Wie?ner A, Kuschk P, Ka?stner M, Stottmeister U. 2002. Abilities of helophyte
species to release oxygen into rhizospheres with varying redox conditions in
laboratory-scale hydroponic systems. Inter J Phytoremed 4:1?15.