Soil Biology & Biochemistry 43 (2011) 1220e1228

Contents lists available at ScienceDirect

Soil Biology & Biochemistry

journal homepage: www.elsevier.com/locate/soilbio

Microbial communities in rice rhizosphere altered by intermittent and continuous

flooding in fields with long-term arsenic application
Anil C. Somenahally a, *, Emily B. Hollister a, Richard H. Loeppert a, Wengui Yan b, Terry J. Gentry a
a
Department of Soil and Crop Sciences, Texas A&M University, 370 Olsen Blvd, College Station, TX 77843-2474, USA
b
USDA-ARS, Dale Bumpers National Rice Research Center, Stuttgart, AR 72160, USA

a r t i c l e i n f o a b s t r a c t

Article history: Rice cultivated on arsenic (As)-contaminated soils can, under some conditions, accumulate high
Received 16 September 2010 concentrations of As in grain, mostly as a result of the continuous flooding practices commonly used for
Received in revised form rice cultivation. Intermittent flooding, as opposed to continuous flooding, might reduce soluble As
10 February 2011
concentrations in the rice rhizosphere, but it might also alter soil microbial populations that may impact
Accepted 11 February 2011
Available online 26 February 2011
As chemistry. A field-scale study was conducted to analyze As concentrations and microbial populations
in the rice rhizosphere, in response to intermittent and continuous flooding in plots that were historically
amended with “As-containing” pesticide and unamended soil. Rhizosphere, pore-water and grain
Keywords:
Rice rhizosphere
As concentrations were quantified, and microbial populations in the rhizosphere were characterized
Pore-water As using community quantitative-PCR and 16S rRNA gene sequencing. Pore-water As concentrations
Rhizosphere-microbial communities decreased by 41e81% and grain As by 31e48% in the intermittently flooded plots relative to the
16S pyrosequencing continuously flooded plots. The relative abundance of bacteria increased over the course of the growing
qPCR season, while archaeal and fungal gene abundances decreased. Bacterial community structure and
Iron-reducing bacteria composition were significantly different between As amended and unamended plots, as well as between
the flooding treatments. Proteobacteria was the predominant phylum detected in most treatments with
relative abundance of 24e29%. The relative abundance of iron-reducing bacteria was higher with the
continuous flood compared to the intermittent-flood treatment, implying greater relative iron reduction
and possibly As release from the iron oxides under the continuously flooded conditions. These differ-
ences in rhizosphere-microbial communities may have contributed to the lower pore-water arsenic
concentrations in the intermittently flooded conditions.
Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction
Arsenic (As) is a naturally occurring metalloid that is toxic to most
forms of life, including most microorganisms. Natural As contami-
nation occurs throughout the world, resulting in more than 50-
million people being exposed to high As concentrations through
drinking water (BGS and DPHE, 2001; Duxbury et al., 2003). The
consumption of high As rice grain is another major exposure route for
a sizeable population in Southeast Asia (Mondal and Polya, 2008).
This potential source of exposure has resulted in increased concern
regarding the cultivation of rice on As contaminated soils, with
several recent studies reporting high As concentrations in rice grain
originating from different parts of the world, including the rice grown
in the south-central USA (Zavala et al., 2008; Meharg et al., 2009).
* Corresponding author. Tel.: þ1 979 845 5322; fax: þ1 979 845 0456.
E-mail address: asomenahally@ag.tamu.edu (A.C. Somenahally).
In 2008, a total of 1.19 million ha of rice was grown in the USA
with more than 80% being cultivated in the south-central states of
Arkansas (45.5%), Louisiana (16.3%), Mississippi (7.8%), Missouri
(6.8%) and Texas (5.4%) (USDA-ERS, 2009). Many of the rice fields in
this region were used historically for cotton production and
received repeated applications of As-based defoliants and pesti-
cides such as sodium hydroxy-methylarsinate, commonly known as
monosodium methyl-arsenate (MSMA) (Woolson, 1977). Although
As-based pesticides are no longer used in cotton or rice production
in the USA, soils with a history of amendment with arsenical
pesticides often contain considerable amounts of residual As
(Gilmour and Wells, 1980). Arsenic is also toxic to rice plants, and in
the production of rice on soils with high As concentrations it may
be responsible for the development of straighthead, a physiological
disorder characterized by sterility of the florets and significant
reduction in yield (Gilmour and Wells, 1980; Yan et al., 2005).
Arsenic can exist as a variety of chemical species, both inorganic
and organic, under typical soil conditions. The most commonly

0038-0717/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.soilbio.2011.02.011
 A.C. Somenahally et al. / Soil Biology & Biochemistry 43 (2011) 1220e1228
encountered inorganic species are arsenate (iAsV) and arsenite
(iAsIII), and the most commonly encountered organic species are
monomethylarsonic acid (MMAsV) and dimethylarsinic acid
(DMAsV), which are present either as the products of microbial
methylation (Cullen and Reimer, 1989) or from the application of
methylarsenical pesticides. Of these species, iAsIII and DMAV are
more soluble and generally considered to be more bioavailable than
others (Masscheleyn et al., 1991), and these species are also
commonly found in rice grain (Xu et al., 2008). Transformations
between iAsV and iAsIII and between organic and inorganic As occur
predominantly in microbially mediated As processes of oxidation,
reduction, methylation and demethylation (Oremland and Stolz,
2003). These transformations can result from a microbial detoxifi-
cation mechanism (Cullen and Reimer, 1989; Jackson et al., 2005),
or they can be linked to cellular metabolism and growth (Oremland
and Stolz, 2003). Other microorganisms, including iron- and
sulfate-reducing bacteria, can also impact As solubility and
bioavailability through the reductive dissolution of minerals
(e.g., iron oxides) that adsorb As (Horneman et al., 2004; Islam
et al., 2004).
Rice tends to accumulate larger amounts of As relative to other
cereals (Williams et al., 2007b), largely due to the continuous
flooding practices commonly used in rice production. Recent
studies have reported that when rice is grown under more aerobic
(i.e., non-flooded) conditions, As concentrations in both the
rhizosphere and rice grain are decreased (Xu et al., 2008; Li et al.,
2009a).
Although aerobic rice cultivation appears to have a potential for
reducing the accumulation of As in rice grain, there has been little
research to date targeting the impacts of different water-manage-
ment systems on rice rhizosphere-microbial communities in soils
with a long-term history of exposure to arsenic-based pesticides.
Both As concentration and redox conditions in soil can impact
microbial populations (Zhou et al., 2002; Edvantoro et al., 2003;
Xiong et al., 2010), and, as a result, microbial populations are
likely to be very different in aerobic compared to the continuously
flooded rhizosphere. Studies are needed to improve our under-
standing of the effects of water-management practices on rhizo-
sphere-microbial composition and specific microbial groups, such
as iron- and sulfate-reducing bacteria, that could potentially impact
As biogeochemistry and bioavailability. Therefore, we conducted
a field experiment to investigate changes in rice-rhizosphere As
concentration and microbial community composition, including
iron- and sulfate-reducing bacteria, during an entire rice growing
season under continuous and intermittent flooding and in separate
soils, amended and not amended with MSMA. We hypothesized
that lower relative abundances of iron reducing and sulfate-
reducing bacteria would occur in the intermittently flooded plots
because of wetedry cycles, and that divergent microbial commu-
nities would be found between MSMA amended and unamended
soils as a result of high As concentrations.
2. Material and methods
2.1. Field experiment
This field experiment was conducted in 2007 at Dale Bumpers
National Rice Research Center, US Department of Agriculture,
Agricultural Research Service, Stuttgart, Arkansas, USA. The soil
type at this location was a fine, montmorillonitic, thermic Typic
Albaqualf, with a soil texture of silt loam to loam (Supplementary
Table 1), and the chemical characterization of the soil samples
from the experimental plots is presented in Supplementary Table 2.
Our high As soil was located in a site where monosodium methyl-
arsenate (MSMA) has been applied in alternate years for more than
1221
twenty years (Yan et al., 2005). The MSMA solution was surface
sprayed before the soil was tilled for drill-planting at a rate of
6.7 kg ha1 (equivalent to 3.1 kg As ha1). In subsequent discussion,
this treatment will be referred to as ‘MSMA’. Our low As soil was
located in adjacent field that was not amended by MSMA and had
not received any As-containing products for at least the last 20
years (hereafter referred to as ‘unamended’).
The two flooding treatments included intermittent and contin-
uous flooding, which were superimposed on the MSMA and
unamended treatments. Under continuous flooding, plots were
flooded all the time with at least 10 cm of standing water, whereas,
under intermittent flooding, each time the plots were flooded,
allowed to dry until surface cracking was evident, and then re-
flooded. The treatment combinations used in this study were (1)
MSMA-continuous flood (M-CF), (2) MSMA-intermittent flood (M-
IF), (3) unamended-continuous flood (U-CF) and (4) unamended-
intermittent flood (U-IF).
Our study was part of a larger experiment to screen several rice
cultivars for As resistance, as well as evaluate the effects of inter-
mittent flooding on cultivar performance. Ten other cultivars were
evaluated along with ‘Wells’ that was used in our study. The
treatment plots were arranged using a splitesplit plot design with
soil-As as the main plot, water treatment as the sub-plot and
cultivar as the subesubplot. Forty subesubplot replicates (4 repli-
cates for each cultivar) were arranged in a completely randomized
design within each sub-plot. Each subesub plot replicate contained
six rows with 0.3 m spacing between rows and rows were 1.5 m
long. Each row was directly seeded with 3 g of paddy rice at 2 cm
deep. The cultivar used in our study was ‘Wells’, a popular cultivar,
commonly grown in Arkansas and also a variety that is susceptible
to As toxicity (Yan et al., 2005). The rice seeds were drill-planted in
the middle of April, and the first flood was introduced four weeks
after sowing, when the plants were about 30 cm tall. All other
management practices followed general procedure as outlined by
Yan et al. (2005).
2.2. Sampling and As analysis
Soil-core samples were collected at planting (1 d after sowing)
and rhizosphere samples were collected at 4 wk after first flooding
(60 d after sowing) and 3 months after first flooding (120 d after
sowing) from three of the four replicate plots. At 1 d, six soil cores
per plot, from 0 to 15 cm depth, were collected randomly from each
plot using a 4.5-cm diameter corer. To minimize the effect of soil
water content on rhizosphere and pore-water sampling in two
water treatments, sampling at 60 d and 120 d was done when the
rhizosphere of both the water treatments were completely satu-
rated. At 60 d and 120 d, three rice plants per plot were collected
along with the adhering bulk soil. Plants were shaken to drop loose
soil and the remaining non-rhizosphere clumps of soil were
removed manually. This procedure left only a few millimeters of
soil around the roots, which was then collected. The rhizosphere
samples were then split into two subsamples, one for chemical
analysis and another for microbial analysis. The samples for
microbial analysis were composited into one sample per plot. The
samples for chemical analysis were transported from the field to
the lab over ice and stored at 4  C, and the samples for microbial
analysis were immediately frozen on dry ice and then stored
at 80  C until further analysis. The samples for chemical analysis
were split into two additional subsamples, one for As analysis and
the other for fatty-acid methyl ester (FAME) analysis. The
subsample for As analysis was air-dried and ground to pass through
a 0.2 mm sieve. The total soil-As concentrations were determined
by inductively-coupled-plasma mass-spectrometry (ICP-MS), using
a DRC-ELAN II model (Perkin Elmer model, Waltham, MA, USA)
1222 A.C. Somenahally et al. / Soil Biology & Biochemistry 43 (2011) 1220e1228
following an open digestion US-EPA method with HNO3/H2O2
(US-EPA, 2007).
Rhizosphere pore-water samples were extracted from soil-core
samples that were collected from the rooting zone at 60 d and
120 d, but not at 1 d because the plots were not yet submerged and
were too dry to extract pore-water. Soil-core samples from 0 to
15 cm depth were collected between two adjacent plants using an
8.5-cm diameter corer with brass insert rings. The insert rings with
soil samples were capped with polypropylene end caps and stored
on ice during transport to the lab. The core samples were then
vacuum filtered through a 0.2 m pore size mixed cellulose ester
filters at a negative pressure of 138 kPa for 20 min to extract pore-
water samples, which were then acidified to pH 3 with HNO3,
stored at 4  C, for total-dissolved As analysis. The total As and Fe
concentrations in pore-water samples were determined using
a DRC-ELAN II model ICP-MS (Perkin Elmer).
Redox potential was measured in the field at each sampling time
between 8 and 10 AM using a platinum electrode and a Ag/AgCl
reference electrode. The platinum electrodes were inserted into the
rooting zone and also into the adjacent bulk soil (between two
rows), and were allowed to equilibrate for 24 h before measuring
the redox. The bulk soil redox measurements were taken for
comparison, as rhizosphere is expected to be more oxidized
compared to the bulk soil. The reference electrode was placed in the
flood water while measuring the redox potential. Detailed proce-
dures for measuring redox potential can be found in Fiedler et al.
(2007).
Grain samples were obtained during harvest from each treat-
ment plot. Approximately 50 g of grain was dehulled, milled,

ground to flour and stored at 4 C until further analysis. Total grain-
As concentration was determined following HNO3/H2O2 digestion
(Zavala and Duxbury, 2008). More information on grain analysis
can be found in Pillai et al. (2010).
2.3. Microbial FAME analysis
The subsamples for FAME analysis were air-dried for 3e6 h to
decrease water content and then extracted for FAMEs following the
procedure outlined by Olexa et al. (2000). Briefly, 3 g of soil from
each of the 3 sampling points were extracted using a FAME
procedure which involved cell lysis, initial saponification and
methylation of the fatty acids and subsequent extraction. Extracted
samples were analyzed at the Department of Plant and Soil
Sciences, University of Delaware using a model 6890 gas chro-
matograph with flame ionization detector (Agilent, Santa Clara, CA,
USA). Two micoliter of each sample were injected into a Hewlett
Packard (Agilent) ultra 2 column (0.33 mm, 5% cross-linked phenyl
methyl silicone, 25 m  0.20 mm) with a 100:1 split ratio and flow
rate of 0.6 mL min1 using hydrogen as the carrier gas. The injection
temperature was 250  C, and the detection temperature was
300  C. The initial oven temperature was 170  C which was ramped
at 5  C min1 to a final temperature of 300  C, for a total of 12.0 min
run time. Peaks were identified using the Sherlock Eukary program
(MIDI, Inc., Newark, DE, USA).
2.4. DNA extraction
Microbial community DNA from the rhizosphere was extracted
using MO BIO PowerMax DNA extraction kits (MO BIO Laboratories,
Carlsbad, CA, USA). The manufacturer’s protocol was modified to
include a lysozyme pre-incubation step in order to enhance gram-
positive bacterial DNA yields (Hollister et al., 2010). Approximately
10 g of soil was weighed into each tube. A 10 mL aliquot of lysozyme
solution (1 mg mL1 final concentration) was then added and the
tubes were incubated in a water bath for 1 h at 37  C with
occasional shaking. The manufacturer’s protocol from the bead-
beating step was continued from this step onward. After the final
elution step, the DNA samples were concentrated by ethanol
precipitation. The DNA samples were then purified using illustra
MicroSpinÔ G-25 Columns (GE Healthcare Biosciences, Pittsburgh,
PA, USA) and stored at 20  C for subsequent analysis.
2.5. Gene copy number quantification using qPCR targeting
The quantitative-PCR (qPCR) assays targeting total bacteria,
archaea, fungi, iron-reducing bacteria (FeRB) and sulfate-reducing
bacteria (SRB) were performed using the group specific primer sets
and qPCR conditions outlined in Supplementary Table 3. The FeRB
was determined by targeting Geobacteracaea and Shewanella spp.
and SRB by targeting the dsrA gene. The assays were performed in
a 10 mL reaction mixture containing 4.5 mL SYBR green real master
mix (5Prime, Gaithersburg, MD, USA), 0.5 mL of each primer
(concentration of 10 mM for bacteria, fungi and archaea, 200 nM for
dsrA and mcrA and 300 nM for Geobacteracaea and Shewanella spp.),
1 mL template (2.5 ng), 1 mL bovine serum albumin (10 mg mL1)
and 2.5 mL molecular grade water. Each analysis run included a set
of standards, positive and negative controls, and samples (each
with three analytical replicates) on a 96-well plate. The PCR reac-
tions with 40 amplification cycles were conducted at the temper-
atures listed in Supplementary Table 3. Melting curve analyses of
the PCR products were performed after each assay to confirm PCR-
amplification quality. The qPCR was performed using an Eppendorf
MastercyclerÒ ep realplex thermal cycler (Eppendorf, Hamburg,
Germany).
Standards for the qPCR assays were generated by PCR ampli-
fying each gene of interest from the genomic DNA of pure cultures
using the primers and details listed in Supplementary Table 3. The
PCR products were confirmed on an agarose gel, and then cloned
into a pGEMÒ-T Easy vector (Promega, Madison, WI, USA),
following the manufacturer’s instructions. Positive clones were
isolated and extracted for plasmid DNA using a Wizard SV Miniprep
kit (Promega, Madison, WI, USA). Plasmid DNA concentrations
ranging from 5.0  103 to 5.0  107 ng mL1 DNA were used to
generate the qPCR standard curves. Relative abundance was esti-
mated by calculating ratios of gene copy numbers for each micro-
bial population to the total community gene copy numbers (sum of
gene copy numbers for bacteria, archaea and fungi).
2.6. 16S rRNA gene sequencing
Equal quantities of community DNA from all 3 field replicates for
each treatment at 120 d were composited into one sample per
treatment. The composited DNA samples were then adjusted to
a concentration of approximately 100 ng mL1 and submitted to the
Research and Testing Laboratory (Lubbock, TX, USA) for tag-pyro-
sequencing of the 16S rRNA gene. Samples were amplified with
modified versions of primers 27F (GAGTTTGATCNTGGCTCAG)
and 519R (GTNTTACNGCGGCKGCTG), and the amplicons were
sequenced using Roche 454 Titanium chemistry (Acosta-Martinez
et al., 2008). Sequence data was submitted to the NCBI Sequence
Read Archive under project accession number SRA023542.
2.7. Statistical and sequence analyses
Nonmetric-multidimensional scaling (NMDS) analysis of FAMEs
was performed using the PC-ORD software version 5.0 (MjM Soft-
ware, Gleneden Beach, OR, USA). SigmaPlot version 11.0 (Systat
Software, San Jose, CA, USA) was used for calculating standard
error of mean for the qPCR data and for creating graphs. Fisher’s
least significant differences (LSD) were calculated for the data in
 A.C. Somenahally et al. / Soil Biology & Biochemistry 43 (2011) 1220e1228 1223

Table 1 using SAS software version 9.1.3. General linear model
ANOVA procedure in split-plot design was used to calculate LSDs
with soil-As as the main plot and water treatment as the sub-plot.
The 16S rRNA gene sequence data was analyzed using the
pyrosequencing pipeline tools available from the Ribosomal Data-
base Project (RDP) (Cole et al., 2009) and MOTHUR version 1.6.1
(Schloss et al., 2009). The 16S rRNA sequences were first trimmed
for primers and chimeras, and then sequences with <200 bases
were removed from the data sets. The total number of sequences in
each library ranged from 1200 to 1650 among the treatments. In
order to minimize any bias as a result of this divergence in the total
number of sequences, we randomly removed sequences from each
sequence library to retain a total of 1200 sequences for each
treatment. The individual sequence files were combined into one
single file using the BioEdit v7.0.5 software (Hall, 1999) and then
were submitted to the RDP aligner tool for multiple sequence
alignment. Pairwise distances between the aligned sequences were
calculated and used to assign sequences to operational taxonomic
units (OTUs). Both the distance matrix calculation and cluster
analyses were performed using the dist.seqs and cluster tools in
MOTHUR using default settings. The OTUs were defined at 97%
similarity cutoff for all the analyses. Both a and b-diversity
measures were estimated for the data sets using the summary
single and summary shared tools in MOTHUR. The a-diversity
measures included Shannon and Simpson measures. A dendrogram
(Fig. S2a) was constructed using the MEGA 4 software (Tamura
et al., 2007) after obtaining BrayeCurtis dissimilarity values (1 -
similarity) with the summary shared function in MOTHUR.
A phylip-formatted distance matrix file was constructed using
the dist.seqs function in MOTHUR and used to create a neighbor-
joining tree with neighbor tool available with the PHYLIP 3.68
package (Felsenstein, 2005). The PHYLIP generated tree was then
used as an input file to run the parsimony p test (Martin, 2002) using
MOTHUR. Parsimony analysis is commonly used to test whether
two or more communities harbor greater differences in phyloge-
netic structure than would be expected by chance; and a p value
of 0.01 is considered to be statistically significant (Martin, 2002).
Relative abundance of taxonomic phyla and families among the
treatment samples were determined with RDP Library Compare.
3. Results
3.1. As concentrations in rhizosphere soil, pore-water
and grain samples
The rhizosphere soil-As concentrations were significantly
higher in the M-CF and M-IF plots, with an average of 21.1
(2.0) mg kg1, than in the U-CF and U-IF plots with 6.4
(0.9) mg kg1 (Table 1). There were no apparent differences in
rhizosphere-As concentrations between the two flooding treat-
ments, nor among the sampling times. The rhizosphere of the M-CF
and U-CF plots were much more reduced, with a growing season
average redox potential of 216 (17) mv, than that of the M-IF and
U-IF plots with 143 (10) mv; and the plots were much more
reduced at 120 d than those at the 60 d sampling time.
Rhizosphere pore-water As concentrations at 60 d were signif-
icantly lower in the U-IF plots than in the M-IF plots, but only
slightly lower in U-CF than in the M-CF. Additionally, the M-IF and
U-IF plots had lower pore-water As concentrations (average of 21
(2.9) and 7 (1.5) mg L1, respectively) than those of the M-CF and
U-CF plots (average of 36 (3.8) and 34 (4.0) mg L1, respectively).
Similar trends were observed at 120 d with significantly lower
pore-water As concentrations in the U-CF and U-IF plots than in the
M-CF and M-IF plots, and also in M-IF and U-IF plots relative to the
M-CF and U-CF plots, respectively. The pore-water As concentra-
tions in the M-IF and U-IF plots were approximately 41 and 80%
lower at 60 d and 51 and 81% lower at 120 d relative to that of M-CF
and U-CF plots, respectively. The pore-water As concentrations also
varied temporally, with considerably higher concentrations
observed at 120 d compared to 60 d in all treatment plots. Inter-
estingly, pore-water As concentrations were higher in the U-CF
compared to the M-IF treatment, even though soil-As concentra-
tions were 4e5 fold higher in the M-IF plots (Table 1). Similar
trends were also observed for total Fe concentrations that were
significantly higher in M-CF than M-IF plots and in U-CF than in U-
IF plots, both at 60 d and 120 d. The total iron concentrations in the
pore-water with the continuous flood and intermittent flood
treatments ranged from 150 to 216 mg L1 and 101e131 mg L1,
respectively (Table 1).

Table 1
Total As concentrations in rhizosphere, pore-water and grain samples and total Fe concentrations in pore-water, and redox measurements in the rhizosphere under different
soil-As concentrations and flooding treatments at 1, 60, and 120 d after sowing.

Sampling time Treatments Rhizosphere As Redoxa (mv) Pore-watera Grain As

(mg kg1) (mg kg1)
Rooting-zone Bulk-soil As (mg L1) Fe (mg L1)
1 db M-CFc 20.5 (A,a)d
M-IF 21.8 (A,a)
U-CF 6.0 (B,a)
U-IF 6.8 (B,a)

60 d M-CF 20.5 (A,a) 179 (A,a) 216 (A,a) 36.2 (A,a) 150 (A,a)
M-IF 22.3 (A,a) 128 (A,b) 133 (A,b) 21.4 (A,b) 115 (A,b)
U-CF 6.3 (B,a) 181 (A,a) 207(A,a) 33.7 (A,a) 168 (A,a)
U-IF 6.7 (B,a) 131 (A,b) 154 (A,b) 6.8 (B,b) 101 (A,b)

120 d M-CF 19.5 (A,a) 246 (A,a) 281 (A,a) 59.4 (A,a) 188 (A,a) 816 (A,a)
M-IF 22.2 (A,a) 166 (A,b) 180 (A,b) 29.3 (A,b) 131 (A,b) 542 (A,b)
U-CF 6.2 (B,a) 258 (A,a) 273 (A,a) 47.9 (A,a) 216 (A,a) 304 (B,a)
U-IF 6.6 (B,a) 146 (A,b) 159 (A,b) 9.1 (B,b) 122 (A,b) 127 (B,b)
a
Pore-water samples and redox measurements were taken only at 60 d and 120 d, as the plots were not yet flooded at 1 d.
b
At 1 d soil samples were taken, but they are not considered to be rhizosphere samples.
c
M-CF ¼ MSMA amended under continuous flooding, M-IF ¼ MSMA amended under intermittent flooding, U-CF ¼ Unamended under continuous flooding,
U-IF ¼ Unamended under intermittent flooding.
d
Letters within parenthesis represent statistical class among the treatments at specific sampling time at p < 0.05 based on LSD mean difference. Capital letters represent
main plot (soil-As) comparisons i.e., M-CF v. U-CF and M-IF v. U-IF. Small letters represent sub-plot (water) comparisons within the same main plot i.e., M-CF v. M-IF and U-CF
v. U-IF.
1224 A.C. Somenahally et al. / Soil Biology & Biochemistry 43 (2011) 1220e1228

Total grain-As concentrations were significantly higher in the

MSMA-amended plots than the unamended plots (M-CF v. U-CF
and M-IF v. U-IF) (Table 1). Additionally, the M-IF and U-IF plots had
lower grain-As concentrations (average of 542 (42) and 127
(20) mg kg1, respectively) than those of the M-CF and U-CF plots
(average of 816 (73) and 304 (46) mg kg1, respectively). This
decrease in total grain-As concentrations in the IF plots compared
to the CF plots corresponded to 33% in the MSMA plots and 58% in
the No-MSMA plots.

3.2. Relative abundance of rhizosphere FAMEs

A nonmetric-multidimensional scaling (NMDS) plot for FAMEs

extracted from rhizosphere is presented in Fig. 1. The microbial
communities in the experimental plots varied temporally among
treatments. At 1 d, the samples were separated by soil-As treatment
(i.e., M versus U). Flooding treatment was not yet imposed as
a variable at 1 d. At 60 d, the samples were separated by both soil-
As concentrations and flooding treatments. At 120 d, the samples
were separated by flooding treatment and to a lesser extent than
those by soil-As treatment. Greater separation of the samples
occurred at 120 d between the flooding treatments, relative to that
at the 1 d and 60 d sampling times, indicating that the microbial
communities shifted and diverged over the course of the growing
season. The marker FAME analysis for fungi and bacteria indicated
that the fungal communities were present at a higher relative
abundance in the M-IF and U-IF plots compared to the M-CF and
U-CF plots, respectively, but there appeared to be no effect of MSMA
treatment on relative abundance of fungi or gram-positive bacteria
(Fig. S1). The fungi:bacteria marker FAME ratios were not different
between 1 d and 60 d; however, the ratios decreased slightly at
120 d, indicating that the fungal numbers decreased towards the
end of the growing season (Fig. S1).
3.3. qPCR assays for relative abundance of microbial populations
The relative abundances of bacteria, archaea and fungi
compared to the total-microbial community population (sum of
gene copy numbers for bacteria, archaea and fungi) are presented
in Fig. 2. At 1 d, the relative abundance of bacteria was significantly
greater than that of either archaea or fungi in all treatment plots. At
60 d, the relative abundance of bacteria decreased slightly
compared to that at 1 d, which corresponded to an increased
relative abundance of archaea in most treatments. Fungal relative
abundance also decreased in most treatments compared to that at
Fig. 2. Relative abundance of archaea, fungi and bacteria, determined using qPCR
assays, in the rice rhizosphere under different soil-As concentrations and flooding
treatments at 1, 60, and 120 d after sowing. The values on y-axis represent the
abundance of bacteria or archaea or fungi to total gene copy numbers of
bacteria þ archaea þ fungi. Error bars represent standard error of mean.
1 d. The relative abundances of bacteria and archaea were higher
than that of fungi in all treatments, except the M-CF treatment.
Archaeal abundance tended to be slightly higher in the U-CF
compared to the U-IF plots, but there was no apparent difference
between M-CF and M-IF. Fungal relative abundances were higher in
the U-IF plots compared to the U-CF plots; however the opposite
trend was obtained among the M-CF and M-IF plots. There were no
apparent differences in bacterial relative abundance among treat-
ments. At 120 d, bacteria dominated all treatments, with relative
abundances ranging from 63% in the M-CF plots to 75% in the M-IF
plots. Archaeal abundance was significantly higher in M-CF plots
relative to those under M-IF plots, but there was no apparent
difference between U-CF and U-IF plots.
The relative abundances of FeRB and SRB, as a proportion of the
total-microbial community (sum of gene copy number for bacteria,
fungi and archaea), are presented in Fig. 3. At 1 d, both FeRB and
SRB were present at relatively low levels, with FeRB comprising less
than 5% of the total community and SRB less than 2% in most
treatments. The relative abundance of FeRB was slightly higher in
the U-CF and U-IF plots compared to the M-CF and M-IF plots,
respectively, but there were no significant differences in relative
SRB abundance among the treatments. At 60 d, the relative abun-
dances of FeRB increased to 10e20% of the total community, with
slight (though not statistically significant) differences among the
treatments.
The relative abundances of SRB were also increased slightly at
60 d, relative to those at 1 d, comprising approximately 3e7% of the

4
M-CF*
M-IF*
2 1d
U-CF*
U-IF*
0 M-CF
Axis-2

M-IF
60d
U-CF
-2
U-IF
M-CF

-4 M-IF
120d
U-CF
U-IF
-6
-6 -4 -2 0 2 4
Axis-1
Fig. 3. Relative abundance of iron-reducing bacteria (FeRB) and sulfate-reducing
Fig. 1. Nonmetric-multidimensional scaling of FAMEs from the rice rhizosphere under bacteria (SRB) in the rice rhizosphere under different soil-As concentrations and
different soil-As concentrations and flooding treatments at 1, 60, and 120 d after flooding treatments at 1, 60, and 120 d after sowing. The values on y-axis represent the
sowing. Error bars represent standard error of mean. *The flooding treatments had not abundance of FeRB or SRB to total gene copy numbers of bacteria þ archaea þ fungi.
yet been imposed at 1 d. Error bars represent standard error of mean.
 A.C. Somenahally et al. / Soil Biology & Biochemistry 43 (2011) 1220e1228
total community; with higher relative abundance in the M-CF plots
relative to the M-IF plots. At 120 d, the FeRB relative abundances
were increased to around 16e26% in the U-CF and M-CF plots,
respectively, and 5e10% in the U-IF and M-IF plots, respectively. The
relative FeRB abundance was higher in the M-CF and U-CF plots
compared to the M-IF and U-IF plots, respectively. The relative
abundances of FeRB were also slightly higher in the MSMA-
amended plots (M-CF and M-IF) compared to the unamended plots
(U-CF and U-IF). The relative abundance of SRB was also higher in
the M-CF and U-CF plots compared to the M-IF and U-IF plots.
3.4. 16S rRNA sequence analyses
The bacterial communities in the rhizosphere were evaluated by
obtaining over 1200 partial 16S rRNA gene sequences per treatment
at 120 d (Table 2). There were no discernable differences among the
treatments based upon the number of OTUs, Shannon or Simpson’s
diversity indices. However, parsimony analysis and shared OTU
analysis for the pairwise comparisons revealed treatment differ-
ences with respect to community structure and composition
(Table 2). The parsimony analysis indicated that the communities
were significantly different from each other with p values <0.01 for
each of the comparisons (Table 2). The shared OTUs observed for
each pair indicated that M-CF and M-IF plots shared the greatest
number of OTUs, while M-CF and U-IF shared the least number of
OTUs (Table 2). The dendrogram (Fig. S2), based upon BrayeCurtis
values, also suggests that both water management and soil-As
concentration impacted the bacterial communities. These results
indicate that both MSMA and flooding treatments impacted the
bacterial communities and that community compositions were
significantly different between soil-As concentrations and flooding
treatments.
Proteobacteria was the predominant phylum detected in most
treatments, comprising 24e29% of each soil community (Fig. S2).
Other dominant phyla included Chloroflexi (20e34%), Acid-
obacteria (5e13%), Bacteroidetes (5e9%) and Firmicutes (3e5%).
With respect to treatment effects, Proteobacteria represented
a significantly higher proportion of the total bacteria detected in
M-CF and M-IF plots compared to U-CF and U-IF plots, respectively
(Fig. 4a). Among the flooding treatments, the M-IF and U-IF plots
harbored a significantly smaller fraction of Proteobacteria than the
M-CF and U-CF plots, respectively (Fig. 4b). Other phyla that
differed significantly among the treatments included Acidobacteria
and Firmicutes, each of which were detected with lower frequency
in M-IF and U-IF plots relative to M-CF and U-CF plots. The Chlor-
oflexi were more abundant in M-IF than M-CF plots but were less
abundant in the U-IF than the U-CF plots (Fig. 4b). The proportion of
Table 2
Diversity indices and pairwise shared operational taxonomic units (OTU)a for
bacterial communities in rice rhizosphere (at 120 d) under different soil-As
concentrations and flooding treatments.
Total No. of No. of Shannon Simpson M-CF M-IF U-CF
sequences OTUs (1/D)
Shared OTUs
M-CFb 1200 811 6.51 (0.05)c 811 (188)
M-IF 1200 862 6.60 (0.05) 1023 (302) 108*
U-CF 1200 819 6.51 (0.05) 741 (214) 34* 83*
U-IF 1200 760 6.45 (0.05) 927 (197) 15* 66*
*Indicates parsimony p test significantly different between the two treatment
communities with p value <0.01.
a
Operational taxonomic units (OTUs) were defined at 97% sequence identity.
b
M-CF ¼ MSMA amended under continuous flooding, M-IF ¼ MSMA amended
under intermittent flooding, U-CF ¼ Unamended under continuous flooding,
U-IF ¼ Unamended under intermittent flooding.
c
Values in parenthesis are 95% mean confidence intervals.
1225
sequences characterized as Acidobacteria were significantly higher
in M-CF and M-IF plots than U-CF and U-IF plots, while Chloroflexi
sequences were less abundant in M-CF and M-IF plots than U-CF
and U-IF plots (Fig. 4). Differences in frequency of occurrences (total
number of sequences) for major taxonomic families were analyzed
among the treatments, indicated that the Geobacteraceae and
Cystobacteraceae families, which include many FeRB, were lower in
the M-IF and U-IF plots than M-CF and U-CF plots, but the differ-
ences were only significant for Geobacteraceae in M-IF versus M-CF
(Fig. S3a) and for Cystobacteraceae in U-IF versus U-CF plots
(Fig. S4a).
4. Discussion
4.1. Impact of flooding treatment on rhizosphere As concentration
and relationships with microbial populations
Although water-management treatment did not impact total
soil-As concentration, it significantly altered the total pore-water
As concentration and thus potential As bioavailability to the rice
plants. Under flooded conditions, pore-water As has been reported
to be mostly in the form of inorganic AsIII (iAsIII) (Panaullah et al.,
2009; Xu et al., 2008); however, it has been established that rice
roots can absorb inorganic AsV (iAsV) and organic-As species, as
well as iAsIII (Li et al., 2009b). Therefore, all dissolved As is poten-
tially bioavailable, irrespective of its form. The lower pore-water As
concentrations in the IF plots than in the CF plots, as well as greater
pore-water As concentration in the U-CF plots relative to the M-IF
plots, imply that water management had a greater impact on total-
dissolved As concentration than did total soil-As concentration and
the reduced conditions favored As release to pore-water. It has been
demonstrated that As mobilization is mostly regulated by the
reduction and solubilization of iron oxides (Benner et al., 2002;
Rowland et al., 2007). At lower redox potentials, both increased
iron-oxide dissolution and the reduction of iAsV to iAsIII could have
resulted in higher dissolved-As concentrations in pore-water
(Masscheleyn et al., 1991; Xu et al., 2008). The reduction and
dissolution of iron oxides is linked to both biotic and abiotic
processes in the rice rhizosphere (Wang et al., 2009), implying that
biotic processes, abiotic processes, or a combination of both could
have contributed to the greater Fe(III) reduction and As release
under the CF conditions in the current study.
Significantly higher pore-water As concentrations in the rhizo-
sphere coincided with the higher abundances of in FeRB in M-CF
relative to the M-IF plots and in U-CF relative to the U-IF plots. The
increase in the relative abundance (up to 26% of the total-microbial
community) is suggestive of active dissimilatory iron reduction,
and probably As release. FeRB gain energy by coupling the oxida-
tion of organic compounds with the reduction of Fe(III) oxy-
hydroxides (Lovley et al., 2004). This process results in the
dissolution of soil iron oxides as well as the solubilization of As
previously adsorbed on the iron-oxide surfaces (Cummings et al.,
1999; Rowland et al., 2007), including the iron oxides precipi-
tated on the rice roots that can sequester significant amounts of As
(Wang et al., 2009). Microbially mediated Fe(III) reduction can be
quite significant in rice paddies, as studies have reported up to 24%
of total reduced Fe as a result of dissimilatory FeIII reduction
(Ratering and Schnell, 2001; Hori et al., 2009). However, it is still
49*
debated whether FeRB activity would actually result in the release
of As, since related studies have reported that microbial Fe(III)
reduction is also likely to result in the formation of secondary iron-
oxide phases that could also potentially adsorb As (Kocar et al.,
2006; Masue-Slowey et al., 2010; Tufano et al., 2008). In any case,
it is evident that FeRB activity does impact Fe cycling in the rice
rhizosphere.
1226 A.C. Somenahally et al. / Soil Biology & Biochemistry 43 (2011) 1220e1228
a b
Continuous f lood
Intermittent f lood
Fig. 4. Impacts of soil-As concentrations and flooding treatments on relative abundance of bacterial phyla in the rice rhizosphere at 120 d. (a) Differences in the relative abundance
of sequences in each phyla in the MSMA-amended plots relative to the unamended plots. (b) Differences in the relative abundance of sequences in each phyla in the intermittently
flooded (IF) plots relative to the continuously flooded (CF) plots. * Significant difference for the phylum between the treatments at 0.05 levels using RDP Library Compare.
The detection of SRB suggests that sulfate reduction could also
be actively occurring in the rice rhizosphere, which is not surprising
as several studies have reported SRB activity in rice paddy soils
(Liesack et al., 2000; Wind et al., 1999). If so, sulfides produced as
a result of sulfate reduction by SRB might react with AsIII to form
soluble and insoluble thioarsenic compounds (Newman et al., 1997;
Oremland et al., 2004). Sulfate reduction could also result in
incorporation of AsIII in FeII-sulfide minerals (Bostick and Fendorf,
2003), and these processes could potentially reduce As bioavail-
ability to the rice plant. We observed substantial sulfide-like black
coatings in the rice rhizosphere of the M-CF plots at 120 d, which
may be a result of dissimilatory sulfate reduction by SRB that were
detected at high relative abundance in M-CF plots at 120 d.
However, we do not have direct spectral evidence that precipitated
sulfides had accumulated in the experimental plots. The results of
our study provide substantial evidence that FeRB and SRB are
present in high abundances in the rice rhizosphere, and suggest
that these bacteria might be more active in continuously flooded
plots. Additional work beyond the scope of this study is needed to
conclusively link specific microbial populations to increased dis-
solved-As concentrations.
As expected, the total grain-As concentrations of rice from
MSMA plots were considerably higher than those from the
unamended plots were, and also higher than those reported in
several market-basket surveys (Williams et al., 2007a; Zavala and
Duxbury, 2008). The As concentrations from the unamended plots
were in similar range to the previously reported concentrations in
rice grains from an almost identical soil site in Arkansas, USA (Pillai
et al., 2010). The significant decrease in total grain-As concentra-
tions in IF plots compared to CF plots can be attributed primarily to
the lower pore-water As concentrations. Our results are similar to
the results of previous pot culture experiments, which also reported
lower grain-As concentrations in aerobically grown compared to
continuously flooded rice (Li et al., 2009a; Xu et al., 2008). These
trends in pore-water and grain As concentrations clearly demon-
strate that it is possible to reduce dissolved-As concentrations in the
rice rhizosphere by means of water management and to ultimately
reduce As uptake and accumulation in grain.
4.2. Response of rhizosphere-microbial communities
to soil-As concentrations and flooding treatments
As expected, higher relative abundances of FeRB and SRB in the
CF compared to the IF plots imply that the reduced conditions
favored anaerobic communities. The high abundance of FeRB
MSMA
Unamended
relative to SRB suggests that iron-reduction reactions dominated
these systems, which makes sense given the relatively high Fe
concentrations in the soils of these experimental plots. Iron- and
sulfate-reducing bacteria are commonly found in rice paddies (Hori
et al., 2009; Weiss et al., 2003), and studies have also reported FeRB
activity on the root surface (Wang et al., 2009). Relatively high
reactive Fe oxide, e.g., ferrihydrite and organic carbon concentra-
tions, in combination with anaerobic niches in the rhizosphere of
the CF plots, are likely to have favored FeRB growth. Under the
intermittent flooding scheme, we observed lower abundances of
FeRB and SRB, which may be due to higher oxygen intrusion into
the rhizosphere as a result of the wetedry cycles.
Long-term application of MSMA also impacted the soil microbial
communities, as indicated by the FAME data at 1 d. This trend was
expected, as As contamination has been reported to significantly
alter rhizosphere-microbial communities, with specific groups of
microbes becoming predominant over others (Turpeinen et al.,
2004; Xiong et al., 2010). Proteobacteria was the most abundant
phylum detected in our MSMA-amended plots, suggesting that
members of this phylum might include As-resistant bacteria. We
explored the NCBI database (accessed November, 2009) for As-
resistance genes, and our search results showed that 49% of all
arsenate-reductase genes (arsC) and 76% of all dissimilatory-arse-
nate respiration (arrA) and 38% of all arsenic-methylation (arsM)
gene-containing bacteria belonged to the phylum Proteobacteria. It
is possible, however, that these numbers may be biased since more
proteobacterial sequences have been submitted to the NCBI data-
base compared to that of any other phylum. We tried to isolate
dissimilatory arsenate-reductase (arrA gene specific primers) and
arsenate-reductase (arsC gene specific primers) genes from our
samples (data not shown), but the primers failed to detect any
arsenic-resistant bacteria among them. Since it is likely that
arsenic-resistant and arsenic-respiring populations were present in
the samples, we believe that the lack of detection of these pop-
ulations was at least partially due to limitations in the currently
available primers.
Iron- and sulfate-reducing bacteria, many of which also belong
to the phylum Proteobacteria, were also detected in greater relative
abundances in MSMA-amended plots than in the unamended plots
at 120 d. This phenomenon might be at least partially attributable
to a toxic effect of MSMA on root metabolism, a possible decrease
in radial oxygen loss, and a resulting lower redox potential at the
root/soil interface. This possibility would be more conducive for a
greater activity of FeRB and SRB in the rhizosphere and is supported
by the evidence of apparent black precipitates at the rice-root
 A.C. Somenahally et al. / Soil Biology & Biochemistry 43 (2011) 1220e1228
surface with the M-CF treatment. These relationships deserve
further attention.
Several FeRB are reported to also be capable of respiring iAsV
(Islam et al., 2005; Kocar et al., 2006), and some SRB are reported to
be able to reduce iAsV to iAsIII (Macy et al., 2000). In a study of
a metal-contaminated site with very high As concentrations, it was
observed that 78% of all bacterial sequences belonged to the
b-Proteobacteria, and the rest belonged to the g-Proteobacteria
(Rastogi et al., 2009). Lu et al. (2007) also reported that Proteo-
bacteria was the predominant phylum in the rice rhizosphere, with
a, g and b-Proteobacteria comprising its dominant bacterial classes.
Additional research is needed to explore these groups and under-
stand their possible role in As biogeochemistry in the rice
rhizosphere.
4.3. Conclusions
This field-scale study has demonstrated that water-manage-
ment impacts dissolved-As concentration in the rice rhizosphere,
with intermittent flooding resulting in significantly lower pore-
water and grain As concentrations than with the continuous
flooding treatment. Based on FAME, qPCR assays, and 16S rRNA
sequencing, this study has further demonstrated that microbial
community structure is impacted by soil-As concentration and
water management, with the microbial communities for the
various treatments diverging over the growing season. The
continuous flood treatment resulted in increased relative abun-
dances of FeRB and SRB relative to those of the intermittent-flood
treatment, which might have contributed toward increased pore-
water As concentrations in the rice rhizosphere. The results of this
study indicate that intermittent flooding has the potential to be
a viable management option to reduce dissolved-As concentrations
in the rhizosphere and reduce As accumulation in grain, when rice
is cultivated on soils with moderate to high As concentrations.
Acknowledgements
We would like to thank field research team at USDA-ARS,
Stuttgart for their help with the field experiment. We also would
like to thank Dr. Dennis James at the Department of Chemistry,
Texas A&M University for his help with the HPLC-ICP-MS, and three
anonymous reviewers whose comments helped to improve this
manuscript.
Appendix. Supplementary material
Supplementary material related to this article can be found
online at doi:10.1016/j.soilbio.2011.02.011.
References
Acosta-Martinez, V., Dowd, S., Sun, Y., Allen, V., 2008. Tag-encoded pyrosequencing
analysis of bacterial diversity in a single soil type as affected by management
and land use. Soil Biology and Biochemistry 40, 2762e2770.
Benner, S.G., Hansel, C.M., Wielinga, B.W., Barber, T.M., Fendorf, S., 2002. Reductive
dissolution and biomineralization of iron hydroxide under dynamic flow
conditions. Environmental Science and Technology 36, 1705e1711.
BGS and DPHE, B.A., 2001. Arsenic Contamination of Groundwater in Bangladesh.
British Geological Survey Report WC/00/19. British Geological Survey, Key-
worth, UK.
Bostick, B.C., Fendorf, S., 2003. Arsenite sorption on troilite (FeS) and pyrite (FeS2).
Geochimica Et Cosmochimica Acta 67, 909e921.
Cole, J.R., Wang, Q., Cardenas, E., Fish, J., Chai, B., Farris, R.J., Kulam-Syed-
Mohideen, A.S., McGarrell, D.M., Marsh, T., Garrity, G.M., Tiedje, J.M., 2009. The
ribosomal database project: improved alignments and new tools for rRNA
analysis. Nucleic Acids Research 37, D141eD145.
Cullen, W.R., Reimer, K.J., 1989. Arsenic speciation in the environment. Chemical
Reviews 89, 713e764.
1227
Cummings, D.E., Caccavo, F., Fendorf, S., Rosenzweig, R.F., 1999. Arsenic mobilization
by the dissimilatory Fe(III)-reducing bacterium Shewanella alga BrY. Environ-
mental Science and Technology 33, 723e729.
Duxbury, J.M., Mayer, A.B., Lauren, J.G., Hassan, N., 2003. Food chain aspects of
arsenic contamination in Bangladesh: effects on quality and productivity of rice.
Journal of Environmental Science and Health 38, 61e69.
Edvantoro, B.B., Naidu, R., Megharaj, M., Singleton, I., 2003. Changes in microbial
properties associated with long-term arsenic and DDT contaminated soils at
disused cattle dip sites. Ecotoxicology and Environmental Safety 55, 344e351.
Felsenstein, J., 2005. PHYLIP (Phylogeny Inference Package) Version 3.6. Distributed
by the Author. Department of Genome Sciences, University of Washington,
Seattle.
Fiedler, S., Vepraskas, M.J., Richardson, J.L., 2007. Soil redox potential: importance,
field measurements, and observations. Advances in Agronomy 94, 1e54.
Gilmour, J.T., Wells, B.R., 1980. Residual effects of MSMA on sterility in rice cultivars.
Agronomy Journal 72, 1066e1067.
Hall, T., 1999. BioEdit: a user-friendly biological sequence alignment editor and
analysis program for Windows 95/98/NT. Nucleic Acids Symposium Series 41,
95e98.
Hollister, E.B., Engledow, A.S., Hammett, A.J.M., Provin, T.L., Wilkinson, H.H.,
Gentry, T.J., 2010. Shifts in microbial community structure along an ecological
gradient of hypersaline soils and sediments. ISME Journal 4, 829e838.
Hori, T., Muller, A., Igarashi, Y., Conrad, R., Friedrich, M.W., 2009. Identification of
iron-reducing microorganisms in anoxic rice paddy soil by 13C-acetate probing.
ISME Journal 4, 267e278.
Horneman, A., Van Geen, A., Kent, D.V., Mathe, P.E., Zheng, Y., Dhar, R.K.,
O’Connell, S., Hoque, M.A., Aziz, Z., Shamsudduha, M., Seddique, A.A.,
Ahmed, K.M., 2004. Decoupling of As and Fe release to Bangladesh groundwater
under reducing conditions. Part 1: evidence from sediment profiles. Geo-
chimica Et Cosmochimica Acta 68, 3459e3473.
Islam, F.S., Boothman, C., Gault, A.G., Polya, D.A., Lloyd, J.R., 2005. Potential role of
the Fe(III)-reducing bacteria Geobacter and Geothrix in controlling arsenic
solubility in Bengal delta sediments. Mineralogical Magazine 69, 865e875.
Islam, F.S., Gault, A.G., Boothman, C., Polya, D.A., Charnock, J.M., Chatterjee, D.,
Lloyd, J.R., 2004. Role of metal-reducing bacteria in arsenic release from Bengal
delta sediments. Nature 430, 68e71.
Jackson, C.R., Dugas, S.L., Harrison, K.G., 2005. Enumeration and characterization of
arsenate-resistant bacteria in arsenic free soils. Soil Biology and Biochemistry
37, 2319e2322.
Kocar, B.D., Herbel, M.J., Tufano, K.J., Fendorf, S., 2006. Contrasting effects of
dissimilatory iron(III) and arsenic(V) reduction on arsenic retention and
transport. Environmental Science and Technology 40, 6715e6721.
Li, R.Y., Stroud, J.L., Ma, J.F., McGrath, S.P., Zhao, F.J., 2009a. Mitigation of arsenic
accumulation in rice with water management and silicon fertilization. Envi-
ronmental Science and Technology 43, 3778e3783.
Li, R.Y., Ago, Y., Liu, W.J., Mitani, N., Feldmann, J., McGrath, S.P., Ma, J.F., Zhao, F.J.,
2009b. The rice aquaporin Lsi1 mediates uptake of methylated arsenic species.
Plant Physiology 150, 2071e2080.
Liesack, W., Schnell, S., Revsbech, N.P., 2000. Microbiology of flooded rice paddies.
FEMS Microbiology Reviews 24, 625e645.
Lovley, D.R., Holmes, D.E., Nevin, K.P., 2004. Dissimilatory Fe(III) and Mn(IV)
reduction. Advances in Microbial Physiology 49, 219e286.
Lu, Y.H., Abraham, W.R., Conrad, R., 2007. Spatial variation of active microbiota in
the rice rhizosphere revealed by in situ stable isotope probing of phospholipid
fatty acids. Environmental Microbiology 9, 474e481.
Macy, J.M., Santini, J.M., Pauling, B.V., O’Neill, A.H., Sly, L.I., 2000. Two new arsenate/
sulfate-reducing bacteria: mechanisms of arsenate reduction. Archives of
Microbiology 173, 49e57.
Martin, A.P., 2002. Phylogenetic approaches for describing and comparing the
diversity of microbial communities. Applied and Environmental Microbiology
68, 3673e3682.
Masue-Slowey, Y., Loeppert, R.H., Fendorf, S., 2010. Alteration of ferrihydrite
reductive dissolution and transformation by adsorbed As and structural Al:
implications for As retention. Geochimica et Cosmochimica Acta. doi:10.1016/
j.gca.2010.11.016.
Masscheleyn, P.H., Delaune, R.D., Patrick, W.H., 1991. Arsenic and selenium chem-
istry as affected by sediment redox potential and pH. Journal of Environmental
Quality 20, 522e527.
Meharg, A.A., Williams, P.N., Adomako, E., Lawgali, Y.Y., Deacon, C., Villada, A.,
Cambell, R.C.J., Sun, G., Zhu, Y.G., Feldmann, J., Raab, A., Zhao, F.J., Islam, R., Hossain, S.,
Yanai, J., 2009. Geographical variation in total and inorganic arsenic content
of polished (white) rice. Environmental Science and Technology 43, 1612e1617.
Mondal, D., Polya, D.A., 2008. Rice is a major exposure route for arsenic in Chakdaha
block, Nadia district, West Bengal, India: a probabilistic risk assessment.
Applied Geochemistry 23, 2987e2998.
Newman, D.K., Beveridge, T.J., Morel, F.M.M., 1997. Precipitation of arsenic trisulfide
by Desulfotomaculum auripigmentum. Applied and Environmental Microbiology
63, 2022e2028.
Olexa, T.J., Gentry, T.J., Hartel, P.G., Wolf, D.C., Fuhrmann, J.J., Reynolds, C.M., 2000.
Mycorrhizal colonization and microbial community structure in the rhizo-
sphere of annual ryegrass grown in pyrene-amended soils. International Journal
of Phytoremediation 2, 213e231.
Oremland, R.S., Stolz, J.F., 2003. The ecology of arsenic. Science 300, 939e944.
Oremland, R.S., Stolz, J.F., Hollibaugh, J.T., 2004. The microbial arsenic cycle in Mono
Lake, California. FEMS Microbiology Ecology 48, 15e27.
1228 A.C. Somenahally et al. / Soil Biology & Biochemistry 43 (2011) 1220e1228
Panaullah, G.M., Alam, T., Hossain, M.B., Loeppert, R.H., Lauren, J.G., Meisner, C.A.,
Ahmed, Z.U., Duxbury, J.M., 2009. Arsenic toxicity to rice (Oryza sativa L.) in
Bangladesh. Plant and Soil 317, 31e39.
Pillai, T.R., Yan, W., Agrama, H.A., James, W.D., Ibrahim, A.M.H., McClung, A.M., Gentry, T.J.,
Loeppert, R.H., 2010. Total grain-arsenic and arsenic-species concentrations in
diverse rice cultivars under flooded conditions. Crop Science 50, 2065e2075.
Rastogi, G., Sani, R.K., Peyton, B.M., Moberly, J.G., Ginn, T.R., 2009. Molecular studies
on the microbial diversity associated with mining-impacted Coeur d’Alene river
sediments. Microbial Ecology 58, 129e139.
Ratering, S., Schnell, S., 2001. Nitrate-dependent iron(II) oxidation in paddy soil.
Environmental Microbiology 3, 100e109.
Rowland, H.A.L., Pederick, R.L., Polya, D.A., Pancost, R.D., Van Dongen, B.E.,
Gault, A.G., Vaughan, D.J., Bryant, C., Anderson, B., Lloyd, J.R., 2007. The control
of organic matter on microbially mediated iron reduction and arsenic release in
shallow alluvial aquafers in Cambodia. Geobiology 5, 281e292.
Schloss, P.D., Westcott, S.L., Ryabin, T., Hall, J.R., Hartmann, M., Hollister, E.B.,
Lesniewski, R.A., Oakley, B.B., Parks, D.H., Robinson, C.J., Sahl, J.W., Stres, B.,
Thallinger, G.G., Van Horn, D.J., Weber, C.F., 2009. Introducing MOTHUR: open
source, platform-independent, community-supported software for describing
and comparing microbial communities. Applied and Environmental Microbi-
ology 75, 7537e7541.
Tamura, K., Dudley, J., Nei, M., Kumar, S., 2007. MEGA4: molecular evolutionary
genetics analysis (MEGA) software version 4.0. Molecular Biology and Evolution
24, 1596e1599.
Tufano, K.J., Reyes, C., Saltikov, C.W., Fendorf, S., 2008. Reductive processes
controlling arsenic retention: revealing the relative importance of iron and
arsenic reduction. Environmental Science and Technology 42, 8283e8289.
Turpeinen, R., Kairesalo, T., Haggblom, M.M., 2004. Microbial community structure
and activity in arsenic-, chromium- and copper-contaminated soils. FEMS
Microbiology Ecology 47, 39e50.
Usda-Ers, (United States Department of Agriculture-Economic Research Service),
2009. Rice Situation and Outlook Yearbook. RCS 2008. Available at: http://usda.
mannlib.cornell.edu/usda/current/RCS-yearbook/RCS-yearbook-02-17-2009.pdf
(accessed February 2010).
US-EPA, 2007. Method 3050B: Acid Digestion of Sediments, Sludges and Soils.
Available at: http://www.epa.gov/osw/hazard/testmethods/sw846/pdfs/3050b.
pdf (accessed Sept 2007).
Wang, X.J., Chen, X.P., Yang, J., Wang, Z.S., Sun, G.X., 2009. Effect of microbial
mediated iron plaque reduction on arsenic mobility in paddy soil. Journal of
Environmental Sciences 21, 1562e1568.
Weiss, J.V., Emerson, D., Backer, S.M., Megonigal, J.P., 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, 77e96.
Williams, P.N., Raab, A., Feldmann, J., Meharg, A.A., 2007a. Market basket survey
shows elevated levels of As in South Central US processed rice compared to
California: consequences for human dietary exposure. Environmental Science
and Technology 41, 2178e2183.
Williams, P.N., Villada, A., Deacon, C., Raab, A., Figuerola, J., Green, A.J., Feldmann, J.,
Meharg, A.A., 2007b. Greatly enhanced arsenic shoot assimilation in rice leads
to elevated grain levels compared to wheat and barley. Environmental Science
and Technology 41, 6854e6859.
Wind, T., Stubner, S., Conrad, R., 1999. Sulfate-reducing bacteria in rice field soil and
on rice roots. Systematic and Applied Microbiology 22, 269e279.
Woolson, E.A., 1977. Fate of arsenicals in different environmental substrates. Envi-
ronmental Health Perspectives 19, 73e81.
Xiong, J., Wu, L., Tu, S., Van Nostrand, J.D., He, Z., Zhou, J., Wang, G., 2010. Microbial
communities and functional genes associated with soil arsenic contamination
and rhizosphere of the arsenic hyper-accumulating plant Pteris vittata L.
Applied and Environmental Microbiology, 00500e00510.
Xu, X.Y., McGrath, S.P., Meharg, A.A., Zhao, F.J., 2008. Growing rice aerobically
markedly decreases arsenic accumulation. Environmental Science and Tech-
nology 42, 5574e5579.
Yan, W.G., Dilday, R.H., Tai, T.H., Gibbons, J.W., McNew, R.W., Rutger, J.N., 2005.
Differential response of rice germplasm to straighthead induced by arsenic.
Crop Science 45, 1223e1228.
Zavala, Y.J., Duxbury, J.M., 2008. Arsenic in rice: I. Estimating normal levels of
total arsenic in rice grain. Environmental Science and Technology 42,
3856e3860.
Zavala, Y.J., Gerads, R., Gurleyuk, H., Duxbury, J.M., 2008. Arsenic in rice: II. Arsenic
speciation in USA grain and implications for human health. Environmental
Science and Technology 42, 3861e3866.
Zhou, J., Xia, B., Treves, D.S., Wu, L.Y., Marsh, T.L., O’Neill, R.V., Palumbo, A.V.,
Tiedje, J.M., 2002. Spatial and resource factors influencing high microbial
diversity in soil. Applied and Environmental Microbiology 68, 326e334.