Enumeration of Fe(II)-oxidizing and Fe(III)-reducing
bacteria in the root zone of wetland plants:
Implications for a rhizosphere iron cycle
JOHANNA V. WEISS1,2, DAVID EMERSON2, STEPHANIE M. BACKER1 and
J. PATRICK MEGONIGAL1,3,*
1Environmental Science and Policy Department, George Mason University, Fairfax, VA 22030, USA;
2American Type Culture Collection, Manassas, VA 22010, USA; 3Current address: Smithsonian Envi-
ronmental Research Center, 647 Contees Wharf Road, P.O. Box 28, Edgewater, Maryland 21037-0028,
USA; *Author for correspondence (e-mail: megonigal@serc.si.edu; phone: (443) 482-2346; fax: (443)
482-2380)
Received 16 January 2002; accepted in revised form 8 October 2002
Key words: Fe(II)-oxidizing bacteria, Fe(III)-reducing bacteria, Rhizosphere, Wetland
Abstract. Iron plaque occurs on the roots of most wetland and submersed aquatic plant species and is
a large pool of oxidized Fe(III) in some environments. Because plaque formation in wetlands with cir-
cumneutral pH has been largely assumed to be an abiotic process, no systematic effort has been made to
describe plaque-associated microbial communities or their role in plaque deposition. We hypothesized
that Fe(II)-oxidizing bacteria (FeOB) and Fe(III)-reducing bacteria (FeRB) are abundant in the rhizo-
sphere of wetland plants across a wide range of biogeochemical environments. In a survey of 13 wet-
land and aquatic habitats in the Mid-Atlantic region, FeOB were present in the rhizosphere of 92% of
the plant specimens collected (n = 37), representing 25 plant species. In a subsequent study at six of
these sites, bacterial abundances were determined in the rhizosphere and bulk soil using the most prob-
able number technique. The soil had significantly more total bacteria than the roots on a dry mass basis
(1.4 ? 109 cells/g soil vs. 8.6 ? 107 cells/g root; p < 0.05). The absolute abundance of aerobic,
lithotrophic FeOB was higher in the soil than in the rhizosphere (3.7 ? 106 /g soil vs. 5.9 ? 105 /g root;
p < 0.05), but there was no statistical difference between these habitats in terms of relative abundance
(
1% of the total cell number). In the rhizosphere, FeRB accounted for an average of 12% of all bac-
terial cells while in the soil they accounted for < 1% of the total bacteria. We concluded that FeOB are
ubiquitous and abundant in wetland ecosystems, and that FeRB are dominant members of the rhizo-
sphere microbial community. These observations provide a strong rationale for quantifying the contri-
bution of FeOB to rhizosphere Fe(II) oxidation rates, and investigating the combined role of FeOB and
FeRB in a rhizosphere iron cycle.
Abbreviations: DCB ? dithionite-citrate-bicarbonate, FeOB ? Fe(II)-oxidizing bacteria, FeRB ? Fe(III)-
reducing bacteria, ROL ? radial O2 loss, MPN ? most probable number, MWMM ? modified Wolfe?s
mineral media, PVC ? polyvinyl chloride, SAV ? submerged aquatic vegetation, SERC ? Smithsonian
Environmental Research Center
Biogeochemistry 64: 77?96, 2003.
? 2003 Kluwer Academic Publishers. Printed in the Netherlands.
Introduction
Wetland ecosystems are sites of rapid biogeochemical cycling due to interactions
between the aerobic soil surface and deeper anaerobic soils. The oxic-anoxic inter-
face is further extended by the presence of wetland vegetation that releases oxygen
from their root system in a process known as radial oxygen loss (ROL) (Armstrong
1964). Molecular oxygen (O2) and Fe(II) react to form a rust-colored precipitate on
the root surface referred to as iron plaque. Iron plaque is widely observed on the
roots of wetland and submersed aquatic plant species where it can account for a
large pool of Fe(III) (Chen et al. 1980; Mendelssohn et al. 1995) and influence the
mobility of trace metals and phosphorus (Greipsson and Crowder 1992; Christensen
and Sand-Jensen 1998; Hansel et al. 2001).
Iron plaque formation is known to be influenced by Fe(II) availability, ROL
rates, pH, soil texture, and redox potential (Mendelssohn et al. 1995). Although
bacteria have been found in iron plaque (Trolldenier 1988; St-Cyr et al. 1993), their
role in its formation is unclear. Our discovery of lithotrophic Fe(II)-oxidizing bac-
teria (FeOB) on the roots of wetland plants (Emerson et al. 1999) raises the possi-
bility that plaque-associated microbes may directly influence plaque formation. In
laboratory studies, FeOB mediate between 45 and 90% of Fe(II) oxidation (Emer-
son and Revsbech 1994; Sobolev and Roden 2001; Neubauer et al. 2002), suggest-
ing FeOB could play an important role in the formation of iron plaque.
O2 availability in the rhizosphere varies both spatially along a root system and
temporally (e.g. diurnal and seasonal), creating a mosaic of aerobic and anaerobic
microsites. Observations of both FeOB (Emerson et al. 1999) and Fe(III)-reducing
bacteria (FeRB) on the roots of wetland plants (King and Garey 1999) suggest that
Fe(II) oxidation and Fe(III) reduction are coupled within the rhizosphere, promot-
ing a localized iron cycle. A cycle in which poorly-crystalline Fe(III) is continually
regenerated through the action of FeOB has important biogeochemical implications,
such as a suppression of methane production (Roden and Wetzel 1996; van der Nat
and Middelburg 1998).
Although the existence of an iron cycle promoted by the presence of wetland
vegetation has been proposed (Giblin and Howarth 1984; Roden and Wetzel 1996),
the possibility that FeOB and FeRB may be involved in such a cycle has not been
investigated. In this study, we hypothesized that FeOB and FeRB are important
components of the wetland plant rhizosphere microbial community. Specifically,
our goals were to: (i) determine if FeOB are found in the rhizosphere of aquatic
plants across a variety of biogeochemically-distinct environments, (ii) compare the
abundance of FeOB, FeRB, and total bacteria in the rhizosphere and surrounding
soil, and (iii) scale the abundances of FeOB and FeRB to the ecosystem level. All
microbial investigations were limited to the study of aerobic, lithotrophic FeOB and
acetate-utilizing FeRB. Furthermore, we define the 
rhizosphere
 as a root?s zone
of influence on Fe(II) oxidation and use the presence of iron plaque to delineate
this zone.
78
Material and methods
Survey of Fe(II)-oxidizing bacteria
Thirteen wetland and aquatic habitats in Virginia, Maryland, and West Virginia were
sampled for rhizosphere FeOB between October of 1997 and November 1999 (Fig-
ure 1). The sites included tidal and non-tidal freshwater marshes, a salt marsh, ri-
parian forests, a bottomland hardwood forest, bogs, and submersed aquatic vegeta-
tion (SAV) (Table 1). At each site, two to three individuals of the dominant plant
species were removed in an intact soil block (approximately 15 ? 15 ? 25 cm) that
contained the plant and most of its root system. Water samples were collected from
the closest surface water at submerged sites, or from porewater that filled the soil
block cavity. Subsamples collected for measuring Fe(II) concentration were added
immediately to ferrozine, then subsequently read at 562 nm on a spectrophotom-
eter within 6 h of collection (Stookey 1970). The pH of the porewater and the soil
(1:1 slurry with DI water) were measured in situ with a portable Cole Parmer pH/
con 10 series meter.
The intact blocks were transported to the lab, stored at 4 ?C, then dissected
within 24 hours. The oxidized layer of the soil surface was discarded, then roots
and soil from the reduced zone were separated using sterile techniques. Roots were
washed in sterile deionized water and subsampled to determine the presence or ab-
sence of FeOB and quantify iron plaque concentrations. Soil samples were taken
from the root zone and dried at 110 ?C for two days for physical and chemical
analyses.
Figure 1. Location of sites sampled for FeOB associated with the roots of wetland and aquatic plants.
A sub-set of the sites (*) were also sampled to obtain abundances of FeOB and FeRB associated with
wetland plant roots and soil.
79
Ta
bl
e
1.
Ph
ys
ic
al
an
d
ch
em
ic
al
ch
ar
ac
te
ris
tic
s
o
fw
et
la
nd
s
an
d
aq
ua
tic
ha
bi
ta
ts
sa
m
pl
ed
fo
rF
e(I
I)-
ox
idi
zin
ga
n
d
Fe
(II
I)-
red
uc
ing
ba
ct
er
ia
.A
ll
m
ea
su
re
m
en
ts
ar
e
th
e
av
er
ag
e
o
ft
hr
ee
so
il
bl
oc
ks
?
1
st
an
da
rd
de
vi
at
io
n
Si
te
H
ab
ita
tT
yp
e
pH
So
il
O
rg
an
ic
C
(%
)
N
(%
)
So
il
Te
x
tu
re
(%
)
Te
x
tu
re
cl
as
sifi
ca
tio
n
So
il
W
at
er
Co
nt
en
t(
%)
Sa
nd
Cl
ay
D
og
ue
Cr
ee
k
R
ip
ar
ia
n
fo
re
st
5.
4
(0.
3)
6.
4
(0.
3)
5
(1)
2.
3
(0.
6)
0.
17
(0.
04
)
81
(3)
9
(5)
Lo
am
y
sa
n
d
H
un
tle
y
M
ea
do
w
s
Fr
es
hw
at
er
m
ar
sh
6.
5
(0.
1)
6.
5
(0.
1)
16
(0)
8.
7
(0.
2)
0.
50
(0.
04
)
48
(2)
29
(0)
Sa
nd
y
cl
ay
lo
am
Co
nt
ra
ry
Cr
ee
k
M
in
e-
im
pa
ct
ed
rip
ar
ia
n
st
rip
4.
2
(0.
5)
3.
9
(0.
2)
4
(1)
1.
5
(0.
5)
0.
10
(0.
03
)
79
(3)
11
(3)
Sa
nd
y
lo
am
SE
RC
Sa
lt
m
ar
sh
6.
1
(0.
5)
6.
5
(0.
2)
69
(5)
38
.1
(1.
6)
2.
22
(0.
04
)
84
(10
)
7
(3)
Lo
am
y
sa
n
d
B
ra
un
Co
ns
tru
ct
ed
he
rb
ac
eo
us
w
et
la
nd
6.
6
(0.
5)
5.
9
(0.
1)
2
(0)
0.
9
(0.
1)
0.
10
(0.
01
)
37
(11
)
45
(11
)
Cl
ay
Po
pl
ar
Tr
ee
Fr
es
hw
at
er
m
ar
sh
6.
8
(0.
3)
6.
6
(0.
2)
8
(3)
2.
5
(1.
3)
0.
20
(0.
11
)
39
(6)
43
(10
)
Cl
ay
G
re
at
M
ar
sh
Fr
es
hw
at
er
tid
al
m
ar
sh
5.
9
(0.
2)
6.
3
(0.
1)
38
(15
)
21
.6
(5.
3)
1.
36
(0.
41
)
7
(5)
47
(14
)
Si
lty
cl
ay
Li
nt
on
H
al
l
Fr
es
hw
at
er
m
ar
sh
6.
6
(0.
1)
7.
2
(0.
1)
7
(2)
3.
0
(1.
3)
0.
29
(0.
11
)
47
(4)
32
(2)
Sa
nd
y
cl
ay
lo
am
Le
e
D
riv
e
B
ot
to
m
la
nd
ha
rd
w
oo
d
4.
2
(0.
5)
5.
2
(0.
1)
13
(3)
7.
8
(1.
4)
0.
46
(0.
18
)
80
(8)
13
(9)
Sa
nd
y
lo
am
Ca
th
er
in
e?
s
Fu
rn
ac
e
R
ip
ar
ia
n
fo
re
st
5.
0
(0.
2)
5.
8
(0.
3)
4
(1)
1.
5
(0.
7)
0.
15
(0.
07
)
61
(3)
18
(4)
Sa
nd
y
lo
am
B
ig
R
un
B
og
H
ig
h
el
ev
at
io
n
bo
g
3.
7
(0.
1)
4.
4
(0.
1)
54
(30
)
29
.4
(15
)
1.
42
(0.
61
)
69
(5)
14
(2)
Sa
nd
y
lo
am
Tu
b
R
un
B
og
H
ig
h
el
ev
at
io
n
m
in
e-
im
pa
ct
ed
bo
g
3.
9
(0.
1)
4.
0
(0.
1)
24
(4)
13
.6
(0.
9)
0.
77
(0.
06
)
52
(15
)
24
(7)
Sa
nd
y
cl
ay
lo
am
Po
to
m
ac
R
iv
er
Su
bm
er
se
d
aq
ua
tic
v
eg
et
at
io
n
6.
0
(0.
1)
6.
1
(0.
0)
5
(0)
2.
7
(0.
3)
0.
18
(0.
01
)
54
(8)
28
(6)
Sa
nd
y
lo
am
80
Physical and chemical analyses
For the initial survey of 13 sites, we used dithionite-citrate-bicarbonate (DCB) to
extract iron on the roots (Taylor and Crowder 1983) and the soil (Darke and Wal-
bridge 1994), followed by analysis on an atomic absorption spectrometer (Perkin-
Elmer model 5100). The two methods differ slightly in the amount of sodium
dithionite that is recommended; we used 1 g for all samples taken for enumeration
of iron bacteria. Because we determined that DCB removes 99% of the iron plaque
(Weiss et al., in preparation), iron removed with DCB will be referred to as total
plaque iron. Due to a larger pool of non-DCB-extractable iron in the soil, the term

DCB-extractable soil iron
 will be used throughout. Soil organic matter content
was calculated by loss on ignition at 430 ?C for 16 hours (Nelson and Sommers
1982). Total soil C and N were determined using a CHN analyzer (Perkin Elmer
Series II 2400), and soil texture was measured with the hydrometer method (Day
1965).
Enrichment for Fe(II)-oxidizing bacteria
The technique used to enrich for aerobic, lithotrophic rhizosphere FeOB has been
previously described (Emerson and Moyer 1997; Emerson et al. 1999). Briefly, a
minimum of four 1-cm subsections of root from each plant specimen were indi-
vidually inoculated into gradient tubes containing an Fe(II)-sulfide plug overlaid
with a bicarbonate-buffered mineral salt solution containing 0.15% low-melting
agarose. All gradient tubes that were used for enrichment and enumeration of FeOB
had a pH of 6.5. However, each of our four rhizosphere FeOB isolates can tolerate
a wide range of pHs, from 4 to 7 or 8 (Weiss et al. in preparation).
Because air was present in the headspace, these tubes provided opposing diffu-
sion gradients of O2 and Fe(II). The tubes were considered positive if a discrete
band of Fe(III) oxides formed around the root surface (Emerson et al. 1999). The
presence of organisms such as sulfur-oxidizers in the initial enrichments does not
appear to affect the growth of FeOB. No differences in Fe(III) oxide band forma-
tion were observed between enrichments and pure cultures or when FeOB were
grown on another Fe(II) substrate such as ferrous carbonate. Subsets of samples
were examined microscopically to confirm the presence of bacteria within the
Fe(III) oxides.
Enumeration of total bacteria
Six of the initial 13 sites were sampled in the summer of 2000 to enumerate total
bacteria, FeOB and FeRB in the rhizosphere and surrounding soil. The sites were
chosen because of the presence of Typha spp., a common freshwater emergent
macrophyte with high rates of radial oxygen loss (Michaud and Richardson 1989).
Typha latifolia was collected at all sites with the exception of the Smithsonian En-
vironmental Research center (SERC) where Typha angustifolia occurs. At each site,
three Typha specimens were collected, stored, and dissected as described above.
81
Soil samples were collected solely from within the root zone of the plant, and gen-
erally within 5 cm of the root surface. In order to collect samples representative of
the full range in iron plaque concentrations, roots varying in color from white to
dark red were collected from many positions along the length of the root system.
Roots were washed by repeatedly vortexing with sterile anaerobic water until the
water became clear. No rhizomes were sampled. Aliquots of either washed root or
soil were preserved in 2% gluteraldehyde and stored at 4 ?C for up to 3 months
until enumeration of total bacteria.
Rhizosphere bacteria were enumerated by total direct counts after the extraction
and filtration of the iron plaque as described by Emerson et al. (1999). Acid am-
monium oxalate at pH 3 (Darke and Walbridge 1994) was substituted for 0.5 M
hydroxylamine/0.5 M HCl because it yielded higher total cell numbers while ex-
tracting all of the visible iron from the root surface (data not shown). Sterile ox-
alate solution (1 mL/cm root) was added to excised, washed roots and shaken in
the dark at 125 rpm until the roots appeared white (1?3 hrs). Ten 
L/mL of 0.1%
acridine orange was added to a portion of the extractant and allowed to sit four
minutes. The stained cells were filtered under low pressure onto 0.2 
M black
polycarbonate membrane filters then counted using epifluorescence microscopy at
1000 ? (Olympus BX 60 microscope). A minimum of 15 fields was counted for
each extracted root sample. A subsample of the extracted roots was stained and ex-
amined directly under the microscope to ensure that no cells were still associated
with the root surface. In order to further investigate the relationship between iron
plaque and rhizosphere bacteria (Emerson et al. 1999), oxalate-extractable iron
(equivalent to total iron in the root samples) was determined in each of the root
extracts. A portion of the oxalate extract was added to 0.5 M hydroxylamine/0.5 M
HCl to reduce all of the extracted Fe(III) to Fe(II) (Phillips and Lovley 1987). Af-
ter shaking the sample solution for 1 hour, Fe(II) was determined by the addition of
a subsample to ferrozine and measurement at 562 nm.
The total number of soil bacteria was determined by diluting 3 aliquots of ho-
mogenized, preserved samples by a factor of 200. After vortexing the diluted sam-
ple for 1 minute, two 10-
L spreads were made on agar-coated (1%), circle-im-
printed microscope slides and allowed to dry (Gold Seal Flourescent Antibody
Microslide). Ten 
L of 250 
M SYTO (Molecular Probes, Eugene, OR), a fluo-
rescent DNA-binding dye (Mason et al. 1998), was added to each circle. SYTO
reduced the background fluorescence observed with acridine orange and proved ex-
tremely effective at estimating bacterial cells associated with particles. After stain-
ing, the soil samples were enumerated in the same manner as the root samples.
Enumeration of Fe(II)-oxidizing and Fe(III)-reducing bacteria
Aerobic, lithotrophic FeOB and acetate-utilizing FeRB were enumerated using the
most probable number (MPN) method (Koch 1996). Root slurries were made by
adding 0.1 g of root to 1 mL of sterile water and gently grinding with a mortar and
pestle. Because this extract was enriched in solubilized carbon released from the
root, it was washed by vortexing for 30 seconds, then centrifuged at 16,000 ? g for
82
5 minutes to pellet the bacteria and remaining root tissue. The supernatant was dis-
carded and replaced by sterile anaerobic water at the beginning of each of three
successive washes. The root pellet was then resuspended in sterile water. Soil
samples were homogenized through vortexing and kept under anaerobic conditions
until dilution into MPN tubes.
The root slurry and homogenized soil were diluted in modified Wolfe?s mineral
media (MWMM) and inoculated into gradient tubes for the enumeration of FeOB.
This dilution series was used as a source to inoculate Fe-reducing media consisting
of sulfate-free MWMM, 10 mM sodium acetate, 50 mM sodium bicarbonate, 0.01%
yeast extract (w/v), 1 
L/mL Wolfe?s trace minerals, and 10% poorly crystalline
Fe(III) adjusted to pH 7.0 (final concentration about 10 mmol/L). Sulfate-free me-
dia was used to reduce the possibility of chemical reduction of iron by sulfides (Ja-
cobson 1994). A slightly modified enrichment was used for the salt marsh site in
which the MWMM in both FeOB and FeRB media was replaced with water from
the site. In these samples, 10 mM sodium molybdate was added to inhibit sulfate-
reducing bacteria from producing sulfides that are capable of chemically reducing
Fe(III) (Lovley et al. 1993).
A three-tube MPN was performed with dilution levels ranging from 10?3 to 10?8.
The development of a discrete band at the oxic-anoxic interface within the gradient
tubes indicated the presence of FeOB. Positive Fe(III)-reducing tubes were visually
assessed for consumption of the Fe(III) oxides and measured for the presence of
Fe(II) by color development with ferrozine. When questionable, the ferrozine sam-
ples were read on a spectrophotometer. A subset of tubes positive for FeOB and
FeRB was examined microscopically to confirm cell growth. Abundances of FeOB
and FeRB were calculated using MPN tables (Eaton and Franson 1995)
Calculation of root and soil bulk density
In order to express our data on a volume basis, intact soil cores were collected from
six Typha-dominated wetlands in February and March of 2001. At each site, six
cores centered on a single Typha shoot were taken to a depth of 20 cm using a
sharpened PVC pipe (15 cm diameter). The cores were taken as pairs; one core
was oven-dried at 90 ?C to determine bulk density and the other was washed to
determine root biomass. The cores for root biomass were washed through a 2 mm
sieve, then sorted into living and dead roots, rhizomes, and litter. All samples were
dried at 90 ?C until no weight change was observed. The sum of live and dead root
biomass was assumed to approximate live root biomass at the peak of the growing
season when samples for iron bacteria were taken. Rhizome biomass was not in-
cluded in the total because no rhizomes were evaluated for iron or bacterial densi-
ties in this study. Soil density was calculated as the difference between bulk density
and the sum of roots, rhizomes, and litter mass.
83
Statistical analyses
All statistics were performed in Microsoft Excel using the add-in statistical pack-
age XLSTAT version 5 (Kovach Computing Services, Wales, UK). Because much
of the data was not normally distributed, correlations were performed using spear-
man rank correlation analyses. Relative and absolute bacterial abundances were
compared with the Wilcoxon signed-rank tests. All standard errors were calculated
according to Sokal and Rohlf (Sokal and Rohlf 1995). Unless otherwise noted, re-
lationships with p-values < 0.05 were considered to be statistically significant.
Results
Soil characteristics and relationship to iron plaque
Our first objective was to conduct a survey for rhizosphere FeOB across a bio-
geochemically-diverse sample of wetland and aquatic habitats. The pH of the sites
ranged from 
4 at Contrary Creek and Tub Run, both mine-impacted sites, to 
7
(Table 1). Percent organic matter ranged from 2.5% at Braun, a constructed wet-
land, to 69% at the salt marsh. Sand generally dominated the mineral fraction of
the soil with notable exceptions at Braun and Poplar Tree where the clay fraction
was significant and Great Marsh where silt was abundant.
The habitats sampled for FeOB also varied with respect to iron in the porewater,
soil, and on the roots (Table 2). Porewater Fe(II) concentrations ranged from below
detection at the Potomac River site to > 12 mg L?1 at Big Run Bog, Tub Run Bog,
and Braun constructed wetland. DCB-extractable soil iron concentrations ranged
from < 10 mg Fe/g dry weight (gdw?1) of soil at Lee Drive and Big Run Bog to >
100 mg Fe gdw?1 in some samples collected at Contrary Creek and Great Marsh.
Iron plaque concentrations ranged from < 1 to > 225 mg Fe gdw?1 root.
No statistically significant relationships were found between iron plaque con-
centrations and porewater Fe(II), DCB-extractable soil Fe, pH, or soil texture. As
was observed previously for Typha latifolia (Macfie and Crowder 1987), we found
a weak negative correlation between iron plaque concentrations and soil organic
content (r2 = 0.40, n = 34, p = 0.01). Organic matter may bind iron, making it less
available for iron cycling (Perret et al. 2000). Correlations were better when Fe
plaque was expressed on a mass basis than a surface-area basis (Table 2).
Presence and abundance of FeOB and FeRB
A total of 37 plant specimens were sampled for rhizosphere FeOB. Aerobic,
lithotrophic FeOB were enriched from the rhizosphere of 92% of the plant speci-
mens, representing 23 of the 25 plant species collected (Table 2). The three plant
specimens lacking FeOB had moderate to high levels of iron plaque.
84
Table 2. The plant species collected and examined for Fe(II)-oxidizing bacteria and associated root, soil,
and water Fe(II) concentrations. Water samples are site averages because they could not be associated
with individual plant samples (n = 3). Soil DCB-extractable iron concentrations are reported as ? 1
standard error (n = 2). The presence of Fe(II)-oxidizing bacteria (FeOB) is denoted by 
+
 and the ab-
sence is denoted by a 
?

Site Plant Species Pore Water
Fe(II)
(mg/L)
DCB-ex-
tractable soil
Fe
DCB-extractable root Fe Presence of
root FeOB
(mg/g) (mg/g) (mg/cm2)1
Dogue Microstegium vimineum 4.1 14.6 (3.0) 3.3 0.03 +
Creek Lindera benzoin 26.0 (4.7) 125.7 1.04 +
Huntley Typha latifolia 3.8 (0.2) 61.6 (1.7) 76.7 0.40 +
Meadows Scirpus cyperinus 14.0 0.08 +
Saururus cernuus 90.6 0.69 +
Contrary Juncus effusus 4.3 (2.2) 49.5 (8.2) 68.7 0.57 +
Creek unknown sedge 49.1 (9.5) 224.9 0.71 +
Leersia oryzoides 130.5 (6.5) 43.1 0.23 +
SERC Spartina patens 0.8 (0.4) 4.2 (1.2) 0.3 0.00 +
Scirpus americanus 8.8 (0.5) 0.6 0.00 +
Typha angustifolia 15.3 (4.5) 9.5 0.04 +
Braun Typha latifolia 14.2 (3.3) 14.4 (1.1) 10.7 0.05 +
Scirpus americanus 18.2 (0.8) 21.7 0.06 +
Echinochloa colona 9.2 (1.5) 83.4 0.22 +
Poplar Tree Typha latifolia 6.1 (2.5) 69.9 (26.4) 24.0 0.06 +
Scirpus cyperinus 57.7 (10.4) 47.8 0.34 ?
Penthorum sedoides 27.6 (2.1) 46.1 0.16 +
Great Marsh Typha latifolia 10.4 (3.0) 79.1 (3.0) 7.8 0.02 +
Zizania aquatica 102.0 (75.9) 30.4 0.12 +
Rosa palustris 86.0 (9.6) 1.6 0.01 +
Linton Hall Typha latifolia 0.5 (0.4) 26.4 (8.3) 13.1 0.08 +
Scirpus cyperinus 46.7 (1.3) 5.6 0.03 +
Scirpus validus 9.6 (3.5) 30.4 0.14 +
Lee Drive Magnolia virginiana 3.4 (0.8) 4.0 (1.4) 25.3 0.26 +
Liquidambar styraciflua 5.8 (2.4) 0.5 0.01 +
Osmunda cinnamomea 6.4 (1.8) 3.5 0.04 ?
Catherine?s Cinna arundinacea 0.4 (0.2) 23.2 (4.2) 4.9 0.04 +
Furnace Scirpus sp. 23.0 (2.3) 8.5 0.08 +
Viburnum dentatum 21.0 (5.7) 35.3 0.47 +
Big Run Juncus acuminatus 12.6 (0.4) 5.2 (0.1) 42.4 0.33 +
Bog Rubus sp. 4.0 (0.4) 8.2 0.13 ?
unknown sedge 4.5 (0.6) 5.4 0.05 +
Tub Run Juncus effusus 15.9 (0.2) 6.1 (0.1) 63.6 0.85 +
Bog Hypericum densiflorum 13.6 (1.7) 13.5 0.02 +
Dulichium arundinaceum 6.6 (1.7) 6.9 0.10 +
Potomac Vallisneria americana < 0.1 15.7 (5.7) 91.0 0.72 +
River Hydrilla verticillata 12.6 (2.1) 97.1 1.29 +
1Root surface area was calculated from measurements of root length and diameter for each root in the
sample.
85
At the sites where Typha was abundant, total rhizosphere bacteria averaged 8.6
? 107 cells gdw?1 (Figure 2A). The highest FeOB abundances, 1.2 ? 106 FeOB
gdw?1 root, were observed at Linton Hall, a freshwater marsh with moderately high
iron plaque concentrations. The lowest abundance of rhizosphere FeOB was 3.4 ?
102 FeOB gdw?1 in the salt marsh (SERC), which also had the lowest levels of
iron on the roots and 80% soil organic matter content. Abundances at the other sites
ranged from 104 to 105 cells gdw?1. Excluding the SERC site, the average FeOB
abundance increased from 5.9 ? 105 FeOB gdw?1 to 6.7 ? 105 FeOB gdw?1.
Rhizosphere FeRB density was one to two orders of magnitude higher than
FeOB density at every site except Linton Hall, averaging 9.2 ? 106 FeRB gdw?1.
At SERC, there were almost 4,000-times more FeRB gdw?1 than FeOB. Microbial
Fe(III) reduction at this site was confirmed by transferring positive Fe(III)-reducing
tubes to media containing sodium molybdate to inhibit sulfate reduction. Nonethe-
less, some of the organisms enumerated at the SERC site may have been sulfate-
reducers capable of Fe(III) reduction that were not inhibited by sodium molybdate
and/or false positives due to chemical reduction of Fe(III) by sulfides (Lovley et al.
1993). A significant relationship was found between the amount of iron plaque and
the density of FeRB (r2 = 0.31, p = 0.01). There was evidence of a relationship
between FeOB densities and iron plaque concentrations. FeRB and FeOB abun-
dances were not significantly related to each other.
In soil, total microbial abundance averaged 1.4 ? 109 cells gdw?1 soil and was
significantly higher than on roots (p < 0.001, Figure 2B). FeOB abundance was
also significantly higher in the soil than on roots, but FeRB were more abundant on
roots than in the soil (p < 0.001). On average, FeOB were slightly more abundant
that FeRB in the soil. As was observed in the rhizosphere iron bacteria, soil FeOB
and FeRB abundances were not significantly correlated with each other. No signif-
icant relationships were observed between DCB-extractable soil Fe and the abun-
dance of FeOB or FeRB. DCB-extractable Fe concentrations were consistently
lower in the soil than on the roots on a dry weight basis.
Proportions of FeOB and FeRB on roots and in soil
The contributions of FeOB and FeRB to total microbial abundances varied widely
among sites and between the rhizosphere and soil (Table 3). The percentage of
FeOB in the rhizosphere ranged from 0.001% at SERC to 6.3% at Linton Hall. On
average, about 1.4% of the rhizosphere microbial community was aerobic,
lithotrophic FeOB. In contrast, FeRB averaged 12.5% of the rhizosphere microbial
community with sites such as Great Marsh and Braun yielding 20 to 30%. The per-
centage of iron bacteria in the soil microbial community was much lower, averag-
ing 0.5% for FeOB and 0.2% for FeRB. The percentage of soil FeRB and FeOB
were weakly correlated (r = 0.51, p = 0.02). The difference between root and soil
percentages of FeRB was highly significant (p < 0.001).
86
Root and soil density
On average, the minerals and soil organic matter of these wetland soils constituted
> 95% of the total mass while roots constituted 1.5% (Table 4). Root iron concen-
trations ranged from < 0.001 mg cm?3 at SERC to 0.09 mg cm?3 at Braun. In con-
trast, soil iron concentrations ranged from 0.002 mg cm?3 at SERC to 14 mg cm?3
Figure 2. A comparison of total bacteria, lithotrophic FeOB, acetate-utilizing FeRB and DCB-extract-
able iron in the rhizosphere of Typha spp. (A) and in the soil (B). With the exception of root bacterial
abundances at Linton Hall and Braun n = 2, all site data is a composite of three plant or soil samples.
Mean bacterial abundances are shown with ? 1 standard error.
87
at Poplar Tree. On a volume basis, the abundance of rhizosphere FeOB and FeRB
was two to three orders of magnitude lower than in the soil. On sites with low to
moderate levels of organic matter, < 50% of the cells and iron minerals were di-
rectly associated with the root surface.
Discussion
The role of microbes in plaque deposition
We recently showed that putatively lithotrophic FeOB exist in the rhizosphere of
plants growing in a mining-impacted riparian wetland with low porewater pH and
high levels of Fe(II) (Emerson et al. 1999). The present study greatly extends those
results by demonstrating that FeOB are ubiquitous in wetlands, occurring in the
rhizosphere of 25 plant species growing in 13 biogeochemically-diverse habitats.
In a more detailed analysis of the roots of Typha spp. at a subset of these sites,
recoverable FeOB had an average density of 105 FeOB gdw?1 root, and accounted
for 1% of the total microbial population. Although no significant correlation was
found between FeOB abundance and iron plaque concentration, the lowest num-
bers of FeOB were observed at the site with the lowest amount of iron plaque (i.e.
the SERC salt marsh). Furthermore, we found that FeRB were 12.5% of the rhizo-
sphere microbial community of the same plants. Although the MPN method does
not assess the activity of these organisms, high abundances suggest that Fe(II)-oxi-
dizing and Fe(III)-reducing bacteria may have a strong influence on the iron bio-
geochemistry of the rhizosphere.
The prevailing wisdom is that rhizosphere Fe(II) oxidation at circumneutral pH
is largely an abiotic process (Howeler and Bouldin 1971; Faulkner 1994; Kirby et
Table 3. The relative proportions (%) of FeOB and FeRB in the rhizosphere of Typha spp. and in bulk
soil. All values are presented as the average of three plant specimens or three soil samples ? 1 standard
error. Percentages were calculated by dividing the abundance of each group by the total bacterial counts.
The overall average is based on all the sampled plants or soils (n = 18). The asterisks (*) denote a
significant difference (p < 0.001) between the rhizosphere and bulk soil in the relative abundance of
FeRB; there was no significant difference between the rhizosphere and soil with respect to FeOB.
Site Rhizosphere Bulk Soil
FeOB FeRB FeOB FeRB
Linton Hall 6.3 (0.6) 5.4 (3.6) 0.04 (0.03) 0.06 (0.01)
Poplar Tree 1.8 (1.1) 12.2 (4.2) 0.08 (0.04) 0.10 (0.04)
SERC < 0.01 3.5 (1.6) < 0.01 0.14 (0.10)
Huntley Meadows 0.6 (0.3) 5.8 (4.0) 2.10 (0.86) 0.42 (0.08)
Great Marsh 0.8 (0.7) 20.1 (3.4) 0.60 (0.25) 0.12 (0.06)
Braun 0.1 (0.1) 32.1 (22.5) < 0.01 0.06 (0.01)
Average 1.4 (0.6) 12.5 (9.0)* 0.47 (0.22) 0.15 (0.04)*
88
Ta
bl
e
4.
R
oo
ta
n
d
so
il
bu
lk
de
ns
iti
es
,i
ro
n,
an
d
iro
n
ba
ct
er
ia
.V
o
lu
m
et
ric
Fe
co
n
ce
n
tr
at
io
ns
(m
gF
e/
cm
3 )
an
d
ab
un
da
nc
es
o
fi
ro
n
ba
ct
er
ia
(ce
lls
/cm
3 )
w
er
e
ca
lc
ul
at
ed
by
m
u
lti
pl
yi
ng
th
e
ro
o
ta
n
d
so
il
de
ns
ity
by
m
as
s-
ba
se
d
iro
n
co
n
ce
n
tr
at
io
ns
an
d
ab
un
da
nc
es
(F
igu
re
2).
Av
er
ag
es
re
pr
es
en
tt
he
co
m
po
sit
e
o
f3
di
ffe
re
nt
co
re
s
(de
ns
ity
)
o
r
sa
m
pl
es
(ir
on
an
d
ba
ct
er
ia
).
Av
er
ag
es
ar
e
?
1
st
an
da
rd
er
ro
r
Si
te
R
O
O
T
SO
IL
D
en
sit
y
(m
g/c
m3
)
D
CB
-e
x-
tr
ac
ta
bl
e
Fe
(m
g/c
m3
)
Fe
R
B
(ce
lls
/cm
3 )
Fe
O
B
(ce
lls
/cm
3 )
D
en
sit
y
(m
g/c
m3
)
D
CB
-e
x-
tr
ac
ta
bl
e
Fe
(m
g/c
m3
)
Fe
R
B
(ce
lls
/cm
3 )
Fe
O
B
(ce
lls
/cm
3 )
Li
nt
on
H
al
l
2.
29
(0.
56
)
0.
07
(0.
02
)
2.
5
?
10
3
(6.
1?
10
2 )
2.
7
?
10
3
(6.
8?
10
2 )
80
8
(23
)
12
.8
(0.
5)
8.
6
?
10
5
(2.
4?
10
4 )
6.
3
?
10
5
(1.
8?
10
4 )
Po
pl
ar
Tr
ee
0.
90
(0.
06
)
0.
04
(0.
01
)
6.
2
?
10
3
(4.
1?
10
2 )
9.
0
?
10
2
(6.
0?
10
1 )
50
0
(14
)
14
.2
(0.
7)
7.
5
?
10
5
(2.
1?
10
4 )
5.
3
?
10
5
(1.
5?
10
4 )
SE
RC
6.
06
(0.
85
)
0.
00
(0.
00
)
8.
9
?
10
3
(1.
7?
10
2 )
2.
4
(0.
05
)
94
(1)
0.
0
(0.
0)
1.
8
?
10
5
(2.
7?
10
3 )
5.
1
?
10
1
(0.
1)
H
un
tle
y
M
ea
do
w
s
1.
17
(0.
28
)
0.
02
(0.
01
)
6.
6
?
10
3
(2.
2?
10
3 )
1.
0
?
10
3
(3.
3?
10
2 )
32
1
(12
)
2.
7
(0.
2)
7.
5
?
10
5
(2.
6?
10
4 )
4.
1
?
10
6
(1.
4?
10
5 )
G
re
at
M
ar
sh
3.
94
(0.
79
)
0.
05
(0.
02
)
3.
0
?
10
4
(6.
5?
10
3 )
1.
0
?
10
3
(2.
2?
10
2 )
11
2
(16
)
0.
5
(0.
1)
1.
7
?
10
5
(3.
2?
10
4 )
8.
7
?
10
5
(1.
7?
10
5 )
B
ra
un
1.
22
(0.
25
)
0.
09
(0.
03
)
5.
0
?
10
4
(1.
0?
10
4 )
1.
3
?
10
2
(2.
6?
10
1 )
99
5
(43
)
12
.1
(0.
9)
1.
2
?
10
6
(5.
2?
10
4 )
7.
4
?
10
4
(3.
2?
10
3 )
89
al. 1999; van Bodegom et al. 2001). In one of the only studies to experimentally
examine the role of bacteria in iron plaque formation, microbial respiration in the
rhizosphere was hypothesized to limit the amount of O2 available for Fe(II) oxida-
tion, resulting in a negative correlation between microbial abundance and plaque
formation (Johnson-Green and Crowder 1991). However, because no precautions
were taken to mimic the microaerophilic environment of an in situ rhizosphere, it
is possible that relatively high O2 concentrations in the experimental microcosms
favored abiotic oxidation of Fe(II). In another set of studies using rice as a model
system, van Bodegom et al. (2001) determined that iron was the primary O2 sink in
the rhizosphere and that Fe(II) oxidation was entirely abiotic. The description of
their experimental set-up, which included periods of vigorous agitation, again sug-
gested conditions that would artificially favor abiotic Fe(II) oxidation (Neubauer et
al. 2002). Because iron plaque can potentially form rapidly due to autocatalytic
oxidation, elucidating the specific contribution of FeOB in plaque formation will
require experimental manipulations of FeOB abundance or the development of di-
rect techniques for distinguishing biotic Fe(II) oxidation using isotopic tracers or
specific inhibitors.
There are a number of reasons why we may have underestimated FeOB abun-
dance. Because cultivation-based methods commonly recover < 1% of the in situ
microbial population diversity, there may have been organisms capable of oxidiz-
ing Fe(II) that would not grow in the culture medium we used. For example, den-
sities of nitrate-reducing FeOB comparable to the O2-reducing FeOB values re-
ported in the current study have been found in sediments from the top 3?4 mm of
a rice paddy soil with substantial nitrate concentrations (Ratering and Schnell 2001)
and in the profundal sediments of a deep lake (Hauck et al. 2001). In contrast, a
study of European aquatic sediments found that nitrate-reducing FeOB comprised
 0.04% of the total bacterial community (Straub and Buchholz-Cleven 1998).
Generally, the use of NO3? as a terminal electron acceptor in wetland soils will be
limited due to assimilation of NO3? by plants and microorganisms (Bodelier et al.
1998). We also did not quantify the abundance of aerobic heterotrophic FeOB such
as members of the Siderocapsaceae, a group of organisms found in a wide range of
aquatic environments (Hanert 1992). Although these organisms have been hypoth-
esized to contribute to Fe(II) oxidation (Lunsdorf et al. 1997), there is little known
about the exact mechanism of heterotrophic Fe(II) oxidation (Emerson 2000) and,
to our knowledge, no studies have examined their natural abundances or potential
rates of iron oxidation. Further studies are needed to determine the contribution of
nitrate-reducing and aerobic heterotrophic FeOB to Fe(II) oxidation in wetlands.
Our observation that FeOB are a substantial proportion of the wetland rhizosphere
microbial community (1% average, 6% maximum) can be considered conservative.
The root as a site of Fe(III) reduction
Large numbers of acetate-utilizing FeRB were also associated with the roots of
Typha spp., contributing an average of 12% of the rhizosphere microbial commu-
nity compared to just 0.02% in the soil. One factor favoring the relatively high
90
densities of FeRB in the rhizosphere may be the high abundance of poorly-crystal-
line Fe(III) minerals that are superior substrates for FeRB (Thamdrup 2000). Poor-
ly-crystalline Fe(III) oxides have been observed on plants growing in natural and
mine-impacted wetlands (Taylor et al. 1984; Fisher and Stone 1991; Batty et al.
2000; Hansel et al. 2001). Conversely, much of the Fe(III) in aquatic sediments is
in a crystalline form not readily reduced by FeRB (Phillips et al. 1993; Thamdrup
2000). In a separate study, we found a significantly higher percentage of oxalate-
extractable iron on the roots (66%) versus the soil (23%) (p < 0.05), and a signif-
icant positive correlation between the percentage of oxalate-extractable iron and the
percentage of FeRB (r2 = 0.55, n = 5 sites, p = 0.01, Weiss et al. in preparation).
Because oxalate extracts primarily poorly-crystalline Fe(III), these results imply
that relatively large amounts of poorly-crystalline Fe(III) in the rhizosphere pro-
mote the high abundances of FeRB that were observed on the roots of wetland
plants.
Another factor contributing to the large difference in the percentage of FeRB in
the rhizosphere versus the soil may be the higher levels of labile carbon associated
with roots. High levels of acetate have been reported in the vicinity of the roots of
aquatic plants (Hines et al. 1994; Dannenberg and Conrad 1999), and concentra-
tions can increase dramatically under anaerobic conditions (Conrad and Klose
1999). Due to its abundance in the rhizosphere and common use as a carbon source
by FeRB such as Geobacter spp. (Lovley 2000), acetate was used in this study to
enrich for FeRB. Other carbon compounds associated with the rhizosphere that can
be used by Geobacter spp. include ethanol, propionate, and butyrate (Mendelssohn
et al. 1981; Conrad and Klose 1999). It is possible that non-acetate-utilizing FeRB
such as Shewanella spp. are also present in the rhizosphere, but a number of stud-
ies have indicated that members of the Geobacteracae dominate subsurface envi-
ronments (Lovley 2000). Humic-acids derived from roots may also be used as
electron shuttles by FeRB, further promoting high abundances of FeRB in the
rhizosphere (King and Garey 1999).
Because the observation of Fe(III) reduction in the rhizosphere is a novel one,
few studies have considered the role of this process in plaque formation and dis-
solution. King and Garey (1999) found that excised roots could reduce up to 28
mg Fe(III) gdw?1 day?1, illustrating that Fe(III) may be reduced from the root sur-
face quickly under anaerobic conditions. Our finding that high numbers of FeRB
are associated with roots and the dominance of poorly-crystalline iron in the rhizo-
sphere helps explain the high potential rates reported in that study. More studies
are needed to determine how quickly Fe(III) reduction occurs in situ where vari-
ability in Fe(III) and O2 concentrations might limit Fe(III) reduction potential. The
presence of iron plaque dominated by Fe(III) (Weiss, in preparation) under many
different environmental conditions indicates that rates of Fe(II) oxidation in the
rhizosphere are generally faster than Fe(III) reduction.
The lack of significant relationships between environmental parameters and iron
plaque concentrations was not surprising due to the dual control of plaque accu-
mulation by iron deposition and solubilization. Because both of these processes are
influenced by physical, chemical, and biological factors varying across spatial and
91
temporal scales, the relationship between iron plaque and any combination of fac-
tors is unlikely to be linear.
A rhizosphere iron cycle
Previous salt marsh studies have discussed the possibility of a geochemical iron
cycle promoted by the presence of wetland vegetation (Giblin and Howarth 1984;
Jacobson 1994), but did not specifically consider the role of FeOB and FeRB. Our
observation of high abundances of FeOB and FeRB on the same 1-cm subsection
of root suggests that oxidation and reduction are occurring simultaneously and both
processes are mediated, at least in part, by bacteria. In this cycle, Fe(II) is oxidized
by both autocatalytic and biotic mechanisms using O2 from roots as a terminal
electron acceptor. Under anaerobic conditions, FeRB use root-derived poorly crys-
talline Fe(III) and labile carbon for Fe(III) reduction. On a microscale, these two
processes are likely to be separated spatially and temporally. The edge of the plaque
may switch from microaerobic to anaerobic with diurnal variations in oxygenic
photosynthesis (e.g. Flessa (1994)). Older parts of the roots have been shown to
release less O2 than the root tip (Armstrong (1971, 1979); Brix and Schierup 1991;
Flessa 1994) and should support significant zones of Fe(III) reduction. Rhizosphere
Fe(II) oxidation may dominate during the summer when increases in plant biomass
stimulate radial O2 loss (Chen et al. 1980), while Fe(III) reduction dominates at
other times (Crowder and Macfie 1986).
The roots clearly have a smaller pool of iron bacteria and Fe(III) than the soil
on a volume basis (Table 4). However, we did not determine the microbial activity
of the FeOB and FeRB and several lines of evidence suggest the rhizosphere iron
pool is more dynamic than the larger soil pool. Over 12% of the root-associated
microbial community are FeRB as compared with less than 1% in the soil; differ-
ences in the availability of labile carbon and reducible Fe(III) may contribute to
this difference. The roots are also the primary source of O2 available to aerobic
FeOB living in otherwise anaerobic subsurface soils. In addition, the roots can have
a much larger aerobic surface area than the surface soil. A rough estimate of the
root surface area (calculated by converting gdw?1 to cm?2 using Table 2) is 16-
times higher than the surface area of the aerobic-anaerobic soil interface in our
cores. Higher availability of O2, labile carbon, and poorly crystalline Fe(III) may
all contribute to a more active iron cycle in the rhizosphere than in the bulk soil in
wetlands. More studies are needed to quantify to ecosystem-level rates of microbi-
ally-mediated iron cycling in the rhizosphere and the precise role that high abun-
dances of iron bacteria have on such a cycle.
Conclusions
Aerobic, lithotrophic FeOB were found on 25 different species sampled from a wide
range of biogeochemical conditions. In Typha-dominated wetlands, FeOB ac-
92
counted for an average of 1.4% of all rhizosphere bacteria. On the same 1-cm sec-
tions of root, FeRB accounted for an average of 12% of the rhizosphere microbial
community. In contrast, FeOB and FeRB abundance in non-rhizosphere soil aver-
aged 0.5% and 0.2% respectively. Relatively high proportions of FeOB, FeRB,
poorly-crystalline Fe(III), O2, and labile C suggest that the rhizosphere is a ?hot-
spot? of microbially-mediated iron cycling in wetlands. These observations provide
a strong rationale for quantifying the contribution of FeOB to rhizosphere Fe(II)
oxidation rates, and investigating the combined role of FeOB and FeRB in iron
cycling. Such a cycle has important biogeochemical implications including the
suppression of methane production.
Acknowledgements
This work was funded by grants from the National Science Foundation (MCB-
9723459 and DEB-9986981), the Jeffress Memorial Trust, and a NSF Dissertation
Improvement Grant. The work was also supported by an American Association of
University Women American Fellowship to JW. Thanks to Shiva Rajaram and Sha-
mus Goss for field and laboratory assistance. We thank Carol Litchfield, Scott Neu-
bauer, and two anonymous reviewers for insightful comments on an earlier draft of
this manuscript.
References
Armstrong W. 1964. Oxygen diffusion from the roots of some British bog plants. Nature 204: 801?802.
Armstrong W. 1971. Radial oxygen loss from intact rice roots as affected by distance from the apex,
respiration, and waterlogging. Physiol. Plantarum 25: 192?197.
Armstrong W. 1979. Aeration in higher plants. In: Woolhouse H.W.W. (ed.), Advances in Botanical Re-
search. Academic Press, London, pp. 226?332.
Batty L.C., Baker A.J.M., Wheeler B.D. and Curtis C.D. 2000. The effect of pH and plaque on the
uptake of Cu and Mn in Phragmites australis. Ann. Bot. 86: 647?653.
Bodelier P.L.E., Duyts H., Blom C.W.P.M. and Laanbroek H.J. 1998. Interactions between nitrifying
and denitrifying bacteria in gnotobiotic microcosms planted with the emergent macrophyte Glyceria
maxima. FEMS Microbiol. Ecol. 25: 63?78.
Brix H. and Schierup H. 1991. Soil oxygenation in constructed reed beds: the role of macrophyte and
soil-atmosphere interface oxygen transport. In: Cooper P.F. and Findlater B.C. (eds), Constructed
Wetlands in Water Pollution Control. Pergamon Press, Oxford, pp. 53?66.
Chen C.C., Dixon J.B. and Turner F.T. 1980. Iron coatings on rice roots ? mineralogy and quantity
influencing factors. Soil Sci. Soc. Am. J. 44: 635?639.
Chen C.C., Dixon J.B. and Turner F.T. 1980. Iron coatings on rice roots ? morphology and models of
development. Soil Sci. Soc. Am. J. 44: 1113?1119.
Christensen K.K. and Sand-Jensen K. 1998. Precipitated iron and manganese plaques restrict root up-
take of phosphorus in Lobelia dortmanna. Can. J. Bot. 76: 2158?2163.
Conrad R. and Klose M. 1999. Anaerobic conversion of carbon dioxide to methane, acetate and propi-
onate on washed rice roots. FEMS Microbiol. Ecol. 30: 147?155.
93
Crowder A.A. and Macfie S.M. 1986. Seasonal deposition of ferric hydroxide plaque on roots of wet-
land plants. Can. J. Bot. 64: 2120?2124.
Dannenberg S. and Conrad R. 1999. Effect of rice plants on methane production and rhizospheric meta-
bolism in paddy soil. Biogeochemistry 45: 53?71.
Darke A.K. and Walbridge M.R. 1994. Estimating non-crystalline and crystalline aluminum and iron by
selectable dissolution in a riparian forest soil. Commun. Soil Sci. Plant Anal. 25: 2089?2101.
Day P.R. 1965. Particle fractionation and particle size analysis. In: Black C.A. (ed.), Methods of Soil
Analysis. American Society of Agronomy, Madison, Wisconsin, USA, pp. 545?566.
Eaton A.D. and Franson M.A.H. 1995. Multiple-tube fermentation technique for members of the
coliform group. In: Standard Methods for the Examination of Water and Wastewater. American Pub-
lic Health Association, Washington, DC, USA, pp. 9?51.
Emerson D. 2000. Microbial oxidation of Fe(II) and Mn(II) at circumneutral pH. In: Lovley D.R. (ed.),
Environmental Microbe-Metal Interactions. ASM Press, Washington, DC, USA, pp. 31?52.
Emerson D. and Moyer C. 1997. Isolation and characterization of novel iron-oxidizing bacteria that grow
at circumneutral pH. Appl. Environ. Microb. 63: 4784?4792.
Emerson D. and Revsbech N.P. 1994. Investigation of an iron-oxidizing microbial mat community lo-
cated near Aarhus, Denmark ? laboratory studies. Appl. Environ. Microbiol. 60: 4032?4038.
Emerson D., Weiss J.V. and Megonigal J.P. 1999. Iron-oxidizing bacteria are associated with ferric hy-
droxide precipitates (Fe-plaque) on the roots of wetland plants. Appl. Environ. Microbiol. 65: 2758?
2761.
Faulkner S.P. 1994. Biogeochemistry of Iron and Manganese in Constructed Wetlands Receiving Coalm-
ine Drainage. Duke University, Durham, NC, USA.
Fisher H.M. and Stone E.L. 1991. Iron oxidation at the surfaces of slash pine roots from saturated soils.
Soil Sci. Soc. Am. J. 55: 1123?1129.
Flessa H. 1994. Plant-induced changes in the redox potential of the rhizospheres of the submerged vas-
cular macrophytes Myriophyllum verticillatum and Ranunculus circinatus. Aquat. Bot. 47: 119?129.
Giblin A.E. and Howarth R.W. 1984. Porewater evidence for a dynamic sedimentary iron cycle in salt
marshes. Limnol. Oceanogr. 29: 47?63.
Greipsson S. 1995. Effect of iron plaque on roots of rice on growth of plants in excess zinc and accu-
mulation of phosphorus in plants in excess copper or nickel. J. Plant Nutr. 18: 1659?1665.
Greipsson S. and Crowder A.A. 1992. Amelioration of copper and nickel toxicity by iron plaque on
roots of rice (Oryza sativa). Can. J. Bot. 70: 824?830.
Hanert H.H. 1992. The genus Siderocapsa (and other iron- and manganese-oxidizing eubacteria). In:
Tr?per H.G., Balows A., Dworkin M., Harder W. and Schleifer K.H. (eds), The Prokaryotes. 2nd
edn. Vol. 4. Springer-Verlag, New York, NY, USA, pp. 4102?4113.
Hansel C.M., Fendorf S., Sutton S. and Newville M. 2001. Characterization of iron plaque and associ-
ated metals on the roots of mine-waste impacted aquatic plants. Environ. Sci. Technol. 35: 3863?
3868.
Hauck S., Benz M., Brune A. and Schink B. 2001. Ferrous iron oxidation by denitrifying bacteria in
profundal sediments of a deep lake (Lake Constance). FEMS Microbiol. Ecol. 37: 127?134.
Hines M.E., Banta G.T., Giblin A.E., Hobbie J.E. and Tugel J.B. 1994. Acetate concentrations and oxi-
dation in salt-marsh sediments. Limnol. Oceanogr. 39: 140?148.
Howeler R.H. and Bouldin D.R. 1971. The diffusion and consumption of oxygen in submersed soils.
Soil Sci. Soc. Am. P. 35: 202?208.
Jacobson M.E. 1994. Chemical and biological mobilization of Fe(III) in marsh sediments. Biogeochem-
istry 25: 41?60.
Johnson-Green P.C. and Crowder A.A. 1991. Iron-oxide deposition on axenic and non-axenic roots of
rice seedlings (Oryza sativa). J. Plant Nutr. 14: 375?386.
King G.M. and Garey M.A. 1999. Ferric iron reduction by bacteria associated with the roots of fresh-
water and marine macrophytes. Appl. Environ. Microbiol. 65: 4393?4398.
Kirby C.S., Thomas H.M., Southam G. and Donald R. 1999. Relative contributions of abiotic and bio-
logical factors in Fe(II) oxidation in mine drainage. Appl. Geochem. 14: 511?530.
94
Koch A.L. 1996. Growth measurement. In: Gerhardt P., Murray R.G.E. and Wood W.A. (eds), Methods
for General and Molecular Bacteriology. ASM Press, Washington, DC, USA, pp. 257?260.
Lovley D.R. 2000. Fe(III) and Mn(IV) Reduction. In: Lovley D.R. (ed.), Environmental Microbe-Metal
Interactions. ASM Press, Washington, DC, USA, pp. 3?30.
Lovley D.R., Roden E.E., Phillips E.J.P. and Woodward J.C. 1993. Enzymatic iron and uranium reduc-
tion by sulfate-reducing bacteria. Mar. Geol. 113: 41?53.
Lunsdorf H., Brummer I., Timmis K.N. and Wagner-Dobler I. 1997. Metal selectivity of in situ micro-
colonies in biofilms of the Elbe River. J. Bacteriol. 179: 31?40.
Macfie S.M. and Crowder A.A. 1987. Soil factors influencing ferric hydroxide plaque-formation on roots
of Typha latifolia. Plant Soil 102: 177?184.
Mason D.J., Shamnuganathan S., Mortimer F.C. and Gant V.A. 1998. A fluorescent gram stain for flow
cytometry and epiflourescence microscopy. Appl. Environ. Microbiol. 64: 2681?2685.
Mendelssohn I.A., Kleiss B.A. and Wakeley J.S. 1995. Factors controlling the formation of oxidized
root channels ? a review. Wetlands 15: 37?46.
Mendelssohn I.A., Mckee K.L. and Patrick W.H. 1981. Oxygen deficiency in Spartina alterniflora roots
? metabolic adaptation to anoxia. Science 214: 439?441.
Michaud S.C. and Richardson C.J. 1989. Relative radial oxygen loss in five wetland plants. In: Hammer
D.A. (ed.), Constructed Wetlands for Wastewater Treatment. Lewis Publishers, Chelsea, Michigan,
USA, pp. 501?507.
Nelson D.W. and Sommers L.E. 1982. Total carbon, organic carbon, and organic matter. In: Page A.L.,
Miller R.H. and Kenny D.R. (eds), Methods of Soil Analysis: Part 2 ? Chemical and Microbiologi-
cal Properties. Soil Science Society of America, Inc, pp. 539?579.
Neubauer S.C., Emerson D. and Megonigal J.P. 2002. Life at the energetic edge: Kinetics of circum-
neutral iron oxidation by lithotrophic iron-oxidizing bacteria isolated from the wetland plant rhizo-
sphere. Appl. Environ. Microbiol. 68: 3988?3995.
Perret D., Gaillard J.F., Dominik J. and Atteia O. 2000. The diversity of natural hydrous iron oxides.
Environ. Sci. Technol. 34: 3540?3546.
Phillips E.J.P. and Lovley D.R. 1987. Determination of Fe(III) and Fe(II) in oxalate extracts of sedi-
ment. Soil Sci. Soc. Am. J. 51: 938?941.
Phillips E.J.P., Lovley D.R. and Roden E.E. 1993. Composition of non-microbially reducible Fe(III) in
aquatic sediments. Appl. Environ. Microbiol. 59: 2727?2729.
Ratering S. and Schnell S. 2001. Nitrate-dependent iron(II) oxidation in paddy soil. Environ. Microbiol.
3: 100?109.
Roden E.E. and Edmonds J.W. 1997. Phosphate mobilization in iron-rich anaerobic sediments: micro-
bial Fe(III) oxide reduction versus iron-sulfide formation. Arch. Hydrobiol. 139: 347?378.
Roden E.E. and Wetzel R.G. 1996. Organic carbon oxidation and suppression of methane production by
microbial Fe(III) oxide reduction in vegetated and unvegetated freshwater wetland sediments. Lim-
nol. Oceanogr. 41: 1733?1748.
Sobolev D. and Roden E.E. 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.
Sokal R.R. and Rohlf F.J. 1995. Biometry. The Principles and Practice of Statistics in Biological Re-
search. W.H. Freeman and Company, New York.
St-Cyr L., Fortin D. and Campbell P.G.C. 1993. Microscopic observations of the iron plaque of a sub-
merged aquatic plant (Vallisneria americana Michx). Aquat. Bot. 46: 155?167.
Stookey L.L. 1970. Ferrozine: A new spectrophotometric reagent for iron. Anal. Chem. 42: 779?781.
Straub K.L. and Buchholz-Cleven B.E.E. 1998. Enumeration and detection of anaerobic ferrous iron-
oxidizing, nitrate-reducing bacteria from diverse European sediments. Appl. Environ. Microbiol. 64:
4846?4856.
Taylor G.J. and Crowder A.A. 1983. Use of the DCB technique for extraction of hydrous iron-oxides
from roots of wetland plants. Am. J. Bot. 70: 1254?1257.
Taylor G.J., Crowder A.A. and Rodden R. 1984. Formation and morphology of an iron plaque on the
roots of Typha latifolia grown in solution culture. Am. J. Bot. 71: 666?675.
95
Thamdrup B. 2000. Bacterial manganese and iron reduction in aquatic sediments. In: Advances in Mi-
crobial Ecology. Kluwer Academic/Plenum Publishers, New York, pp. 41?84.
Trolldenier G. 1988. Visualization of oxidizing power of rice roots and of possible participation of bac-
teria in iron deposition. Z. Planz Bodenkunde 151: 117?121.
van Bodegom P., Goudriaan J. and Leffelaar P. 2001. A mechanistic model on methane oxidation in a
rice rhizosphere. Biogeochemistry 55: 145?177.
van Bodegom P., Stams F., Mollema L., Boeke S. and Leffelaar P. 2001. Methane oxidation and the
competition for oxygen in the rice rhizosphere. Appl. Environ. Microbiol. 67: 3586?3597.
van der Nat F.J.W.A. and Middelburg J.J. 1998. Seasonal variation in methane oxidation by the rhizo-
sphere of Phragmites australis and Scirpus lacustris. Aquat. Bot. 61: 95?110.
96