the genome-wide transcriptional and physiological...
TRANSCRIPT
[1]
The Genome-Wide Transcriptional and Physiological Responses of 1
Bradyrhizobium japonicum to Paraquat-Mediated Oxidative Stress 2
3
4
Andrew J. Donati1, Jeong-Min Jeon1,2, Dipen Sangurdekar3, Jae-Seong So2 and Woo-Suk 5
Chang1* 6
7
1 Department of Biology, University of Texas, Arlington, Texas 76019; 2 Department of Marine 8
Science and Biotechnology BK21 Program, Inha University, Incheon 402-751, South Korea 9
3 Lewis-Sigler Institute for Integrative Genomics, Princeton University, Princeton, New Jersey 10
08544 11
12
13
14
Running title: Paraquat-mediated oxidative stress in B. japonicum 15
16
17
*Corresponding author. 18
Mailing address: Department of Biology, University of Texas at Arlington, Life Science 19
Building 216, Arlington, TX 76019. 20
Phone: (817) 272-3280 21
Fax: (817) 272-2855 22
E-mail address: [email protected] 23
Copyright © 2011, American Society for Microbiology and/or the Listed Authors/Institutions. All Rights Reserved.Appl. Environ. Microbiol. doi:10.1128/AEM.00047-11 AEM Accepts, published online ahead of print on 15 April 2011
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[2]
ABSTRACT 24
The rhizobial bacterium Bradyrhizobium japonicum functions as a nitrogen-fixing symbiont 25
of the soybean plant (Glycine max). Plants are capable of producing an oxidative burst, a rapid 26
proliferation of reactive oxygen species (ROS), as a defense mechanism against pathogenic and 27
symbiotic bacteria. Therefore, B. japonicum must be able to resist such a defense mechanism to 28
initiate nodulation. In this study, paraquat, a known superoxide radical-inducing agent, was used 29
to investigate this response. Genome wide transcriptional profiles were created for both 30
prolonged exposure (PE) and fulminant shock (FS) conditions. These profiles revealed that 190 31
and 86 genes were up and down-regulated for the former condition, and that 299 and 105 genes 32
were up and down-regulated for the latter condition, respectively (> 2.0-fold, P < 0.05). Many 33
genes within putative operons for F0F1-ATP synthase, chemotaxis, transport, and ribosomal 34
proteins were up-regulated during PE. The transcriptional profile for the FS condition strangely 35
resembled that of a bacteroid condition, including the FixK2 transcription factor and most of its 36
response elements. However, genes encoding canonical ROS scavenging enzymes such as 37
superoxide dismutase and catalase were not detected, suggesting constitutive expression of those 38
genes by endogenous ROS. Various physiological tests including exopolysaccharide (EPS), 39
cellular protein, and motility characterization were performed to corroborate the gene expression 40
data. The results suggest that B. japonicum responds to tolerable oxidative stress during PE 41
through enhanced motility, increased translational activity, and EPS production in addition to the 42
expression of genes involved in global stress responses such as chaperones and sigma factors. 43
44
45
46
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[3]
INTRODUCTION 47
Bradyrhizobium japonicum, which belongs to the family Bradyrhizobiaceae of the order 48
Rhizobiales, is an inhabitant of the soil which can exist in either a free-living state or an 49
endosymbiont state in association with host plants. In the latter state, the host specificity of B. 50
japonicum allows it to establish a symbiotic relationship with the soybean plant (Glycine max) by 51
engaging in specific signaling and communication which involves plant-borne flavonoids and 52
bacterial lipo-chito-oligosaccharide Nod factors (11, 40). In addition, B. japonicum is capable 53
of diversity in its metabolism and energy production in response to environmental cues (24). 54
One variable that changes as B. japonicum transforms from a free-living state to a symbiotic state 55
is the concentration of oxygen. It has been demonstrated that a concentration difference on the 56
magnitude of five orders of ten (34) exists between the two states. Reactive oxygen species 57
(ROS) are created within cellular systems as a by-product of aerobic respiration because of the 58
incomplete reduction of molecular oxygen at the terminus of the electron transport chain. These 59
reactive oxygen molecules are capable of modifying cellular components such as lipids, proteins 60
and nucleic acids due to their contribution to the propagation of free radicals within a cell. 61
A key aspect of a plant-microbe interaction involving either pathogenic or non-pathogenic 62
but invasive bacteria is an induction of the plant’s innate defense mechanism which has been 63
termed the hypersensitive disease resistance response (35). This non-specific localized response 64
is mediated by ROS such as hydrogen peroxide and the superoxide anion radical (3, 4, 31, 36, 65
51, 53). Symbiotic bacteria of necessity must possess compensatory mechanisms for the plant 66
innate defense mechanisms to initiate symbiosis. These mechanisms include activation of 67
detoxification enzymes, exopolysaccharide (EPS) production, and biofilm formation, all of 68
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[4]
which have been suggested to play a role in tolerance to a variety of stresses such as ROS (33), 69
desiccation (16), and cytotoxic host immune responses (18, 19). 70
Paraquat (1,1’-dimethyl-4,4’-dipyridylium dichloride or methyl viologen) is the 71
ROS-inducing agent and indirectly contributes to oxidative stress in the cytosol by its interaction 72
with membrane-bound cytochrome c. The paraquat radical is capable of diffusion across the 73
cellular membrane of bacteria (27). The overall mechanism for its toxicity involves a net 74
production of superoxide anion radicals in the intracellular compartment and a moderate degree 75
of extracellular production of superoxide anion radicals and hydrogen peroxide (13). The toxic 76
effects of paraquat on a prokaryotic system have been demonstrated to be augmented by the 77
presence of oxygen, and also depend on nutritional factors such as substrate availability and 78
medium composition (27, 29). The transcriptional response to paraquat in prokaryotic systems, 79
namely Escherichia coli (28, 29), has been shown to involve the canonical mechanism of 80
detoxification by manganese-superoxide dismutase and downstream enzymes such as catalases 81
and peroxidases. 82
Many studies on oxidative stress have been performed using shock experiments in which the 83
bacteria are exposed to a high concentration of the ROS for a short time period (5, 39, 50). 84
However, little is known about how bacteria respond to tolerable, low concentrations of ROS for 85
a long time. It is not difficult to conceive of the notion that bacteria inhabiting the soil and the 86
rhizosphere experience low concentrations of ROS in addition to the high concentrations that 87
have been conventionally used in most transcriptional analyses. It can also be reasonably 88
hypothesized that cells subjected to a high concentration rarely experienced in nature may not 89
respond in a coherent and realistic manner because of the excessive levels of stress. A state of 90
prolonged exposure to a moderate concentration allows a cell to gradually acclimatize to the 91
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[5]
experimental variable and may provide a more comprehensive overview of its metabolic 92
adjustment to ROS. Moreover, the concentration of ROS within the rhizosphere is not static, but 93
is characterized by a fluctuation of the relative concentration. Therefore, the proceeding 94
experimentation has been designed to consist of two categories. The first method is referred to 95
as the fulminant shock (FS) treatment of cells which uses a high, sub-lethal concentration of 96
paraquat. The other method has been termed the prolonged exposure (PE) condition in which 97
low concentrations are used to provide a balanced approach to the transcriptional profile of B. 98
japonicum. 99
In this study, we demonstrate the transcriptional and physiological responses of B. 100
japonicum to oxidative stress induced by paraquat. The elucidation of possible stress tolerance 101
mechanisms will provide insight into the plant-microbe interaction between B. japonicum and 102
Glycine max as well as other rhizobia and their respective symbiotic partners. 103
104
MATERIALS AND METHODS 105
Bacterial strain and growth conditions. The wild-type strain, B. japonicum USDA 110, 106
was routinely cultured in arabinose-gluconate medium (AG) (48) which contained 125 mg 107
NaHPO4, 250 mg Na2SO4, 320 mg NH4Cl, 180 mg MgSO4 x 7H2O, 10 mg CaCl2, 4 mg FeCl3, 108
1.3 g HEPES, 1.1 g MES, 1.0 g yeast extract, 1.0g L-arabinose, and 1.0 g D-gluconic acid sodium 109
salt per liter with pH adjusted to 6.8. All cultures were maintained aerobically at 30 °C with 110
shaking (200 rpm). 111
Prolonged exposure (PE) experimental conditions. Growth curves were produced for B. 112
japonicum during PE to paraquat (Aldrich) at various concentrations (10, 50, 100 μM, and 113
1 mM) in order to ascertain which concentration had an intermediate effect on generation time 114
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[6]
and the final optical density. Measurements of turbidity (O.D.600) were taken every 8-12 hours 115
with a UV-Vis spectrophotometer (Genesys 5, Spectronic Instruments) and plotted against time 116
on the abscissa. Paraquat was added to AG medium at the time of inoculation to create final 117
concentrations of 10, 50, 100 μM and 1 mM. 118
Fulminant shock (FS) experimental conditions. For induction of a FS stress condition, 119
paraquat was added after cell cultures had achieved O.D.600 of 0.8. Cells were treated with 120
various concentrations of paraquat (100, 250, 500, 750 μM, and 1, 5, 10, and 20 mM) for 10, 20, 121
and 30 min. After being subjected to paraquat, cells were plated on AG medium plates. Then, 122
colony forming units (CFUs) were counted after 4 days of incubation at 30 °C. The 123
experimental condition was chosen from the concentration and time variables which had 124
intermediate and sub-lethal effects on the viable cell counts. 125
Isolation of total RNA. Cells were harvested from 250-ml cultures during mid-log phase 126
(O.D.600 ~0.6-0.8) after PE to paraquat and after FS treatment with paraquat by 10% stop solution 127
(5% H2O-phenol, pH 4.3, in 100% ethanol). Cells were recovered by centrifugation at 8,000 x g 128
for 20 min, flash frozen in liquid N2, and stored at -80 °C until use. Total RNA was isolated 129
from cell pellets using a modified hot-phenol method as previously described (7). DNase 130
treatment and purification of total RNA were performed by using RNase-free DNase set 131
(Qiagen) and the RNeasy® mini kit (Qiagen) according to the manufacturer’s protocols. RNA 132
quality was analyzed on 0.8% agarose gels and RNA quantity was measured by the NanoDrop 133
ND-1000 spectrophotometer (Thermo Scientific). 134
Global gene expression analysis by microarray hybridization. Genome-wide 135
transcriptional profiles were created from the hybridization of cDNA samples labeled with 136
Amersham Cy3 and Cy5 monoreactive dyes (GE Healthcare) to microarray chips containing 137
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[7]
70-mer oligonucleotides which were complementary to each of the 8,453 annotated open reading 138
frames (ORFs) of B. japonicum (15). Thirty micrograms of total RNA were used for cDNA 139
synthesis and 5 μg of cDNA from both control and experimental conditions were used for 140
labeling and hybridization. The detailed protocols for cDNA synthesis, cDNA labeling, 141
hybridization, and washing have been described previously (15). A total of three independent 142
biological replicates were prepared for each condition, which included a dye-swap for each 143
replicate, resulting in a total of 6 slides for each experimental condition. 144
Statistical analysis of microarray data. The slides were scanned with the Axon GenePix 145
4200 scanner and GenePix Pro 6.0 software was used to measure intensity values at each spot. 146
The signal intensities were normalized for slide and spot abnormalities using the Lowess 147
algorithm and subsequently analyzed by mixed-effect microarray ANOVA (MAANOVA) (32). 148
Values obtained from this round of analysis were input into a significance analysis of microarray 149
(SAM) statistical package (57) to create a list of differentially expressed genes with a fold-150
induction threshold of 1.5 or 2.0, and a false-discovery rate (FDR) of 5% or less (q value ≤ 0.05). 151
Quantitative reverse transcription-PCR (qRT-PCR) analysis. Representative genes 152
which were differentially expressed from the microarray data were chosen for qRT-PCR analysis 153
to confirm the microarray data. Primers were designed with Primer 3 software, available at 154
http://frodo.wi.mit.edu/primer3/, to amplify 80-250 bp regions of the chosen genes (Table 1). 155
Two micrograms of the same RNA which was used in the microarray experiment was employed 156
to synthesize cDNA. The process was performed according to a previously described protocol 157
(20). Relative expression values for 3 biological replicates were normalized to the expression 158
values of a housekeeping gene (bll0631, parA), which encodes a chromosome partitioning 159
protein. Fold-induction values were calculated in accordance with the method of Pfaffl (42). 160
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[8]
Construction of ΔcheA mutant. The locus blr2343 (cheA) was chosen for mutagenesis 161
based on the transcriptomics data. The mutant strain (ΔcheA) was created by amplification of a 162
3,663 bp fragment containing the 1,911-bp cheA gene and additional 892-bp 5’ upstream and 163
860-bp 3’ downstream portions with the primers TGATTCGCGTCAAACG TGTATCGG and 164
TCATCTTGGCGTATGTCACAGCGA, forward and reverse respectively. The 3,663 bp 165
fragment was cloned into the pKnockout Ω suicide vector (59). A kanamycin cassette from 166
pHP45Ω-Km (22) was inserted inside the cloned gene to disrupt it. The resulting construct was 167
transferred from E. coli DH5α to B. japonicum USDA110 by triparental mating with the helper 168
strain containing pRK2073. Transconjugants were selected based on double homologous 169
recombination, which confers kanamycin resistance and streptomycin sensitivity. The mutant 170
strain was confirmed by Southern blot analysis and colony PCR (data not shown). 171
Measurement of ROS in vitro. Fifty micro-liters of a 20 μM stock solution of 172
dihydroethidium (Molecular Probes) in dimethyl sulfoxide was combined with 50 μL of the 173
samples to create a final concentration of 10 μM and incubated for 1 h at 37 °C in the dark. 174
Absorbance was measured at the λmax for fluorescence excitation (300 nm) of dihydroethidium 175
using a 96-well microtiter plate reader (Synergy 2, BioTek). Nine replicates were performed at 176
the harvesting time for the growth of B. japonicum when subjected to 100 μM paraquat (for the 177
PE condition) and at 10 min following treatment with 5 mM paraquat (for the FS treatment). 178
The medium itself and medium with paraquat were also used as internal controls. 179
Capillary assay for chemotaxis. The effect of paraquat on chemotaxis was investigated 180
using a previously described capillary assay (1). Briefly, a mid-log phase culture was harvested 181
by centrifugation at 8,000 x g and resuspended into AG media to adjust cell numbers to ca. 109 182
cells per milliliter. The sample was divided into three and media was added to reconstitute ca. 183
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[9]
3.3 x 108 cells for each sample in 1 milliliter. The capillary test was performed using the capillary 184
tube apparatus as described previously (1). CFUs were counted after incubation for 3 to 4 days. 185
Therefore the approximate influx of motile cells into the capillary tube was measured. A ratio of 186
the treated cells over the control cells was calculated to infer the effects of 100 μM paraquat on 187
chemotaxis. A ratio greater than one indicates that the test substance is an attractant and a ratio 188
less than 1 indicates that the substance is a repellant. A ratio close to one is indicative of no net 189
effect on chemotaxis. The chemotaxis of the mutant strain (ΔcheA) was also assayed to compare 190
it with that of the wild type. 191
EPS isolation and quantification. B. japonicum cultures treated with 0, 100, 250, and 500 192
µM paraquat were grown for the late log phase and the extracellular materials were recovered by 193
centrifugation at 16,000 x g for 30 min at 4 °C. The supernatant was filtered with sterilized 0.45-194
µm filters and treated with DNase I and proteinase K as described previously (6, 46). The EPSs 195
were precipitated with three volumes of 100% ethanol at -20 °C overnight. The ethanolic 196
supernatant was subject to the second ethanol precipitation to ensure adequate EPS recovery (10, 197
16). The EPSs were dried at room temperature before being resuspended in deionized water. 198
Total carbohydrate content was measured with the phenol-sulfuric acid method (21) in which 199
glucose was the standard, and the acidic carbohydrate content was determined by the m-200
phenylphenol method (8) with D-glucuronic acid as the standard. The carbohydrate content was 201
normalized to the total protein. 202
Isolation and quantification of cellular protein. Cellular protein was isolated from the 203
same cell pellets from the cultures used for EPS quantification. The cell pellets were 204
resuspended in 20 ml 1N NaOH and incubated at 80 °C for 30 min to lyse the cells and recover 205
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[10]
the cellular proteins in the supernatant. The cellular protein content was determined by the 206
method of Bradford (9) in which bovine serum albumin served as a reference. 207
Microarray data accession numbers. The microarray data from this study are compiled in 208
the NCBI Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo/) and 209
are accessible through the GEO series accession numbers GSE26252 and GSE26236 for PE and 210
FS treatments, respectively. 211
212
RESULTS 213
Bacterial growth characteristics during prolonged exposure (PE) to paraquat. The 214
growth characteristics of B. japonicum when subjected to various concentrations of paraquat 215
from the time of inoculation were examined and an optimal concentration for the PE experiment 216
was determined to create a genome-wide transcriptional profile. As shown in Fig. 1, the high 217
concentration (1 mM) of paraquat affected both the lag phase and the generation time. B. 218
japonicum cultures supplemented with 10, 50, and 100 μM paraquat also experienced a 219
lengthened acclimation period, but had a growth rate comparable to that of the control, although 220
the maximum density of cells treated with 100 μM paraquat is lower than that of the control 221
(Fig. 1). Given that the oxidative stress condition of 100 μM paraquat had intermediate effects on 222
cellular processes and physiology without excessive lethality, this condition was chosen for PE 223
experiments and the transcriptomics study. 224
Bacterial survival after fulminant shock (FS) treatment with paraquat. The effects of a 225
concentration spectrum of paraquat on the survival of B. japonicum were investigated to 226
determine an optimal condition for FS experiments. After a 10 min treatment period, 5, 10, and 227
20 mM paraquat caused a diminution of the viable cell count by 22%, 24 %, and 32% 228
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[11]
respectively (Fig. 2). However, by 30 min there was a steady-state in CFUs with a little decline, 229
approximately 40% for the three concentrations. Due to the intermediate consequences on cell 230
viability, we chose the treatment parameters of a 5 mM concentration and 10 min for the creation 231
of genome-wide transcriptional profiles. 232
Genome-wide transcriptional analyses of cells exposed to paraquat-mediated oxidative 233
stress. Genome-wide transcriptional profiles were generated from both cells subjected to the low 234
concentration (100 μM) of paraquat throughout their entire growth phase and cultures treated 235
with the high concentration (5 mM) of paraquat for a short time period (10 min). Both 236
transcriptional profiles reveal that the distribution of affected cellular functional categories 237
between the two conditions is different but does contain some overlap (Fig. 3). At the time of 238
cell harvest, we also checked the ROS level since the concentration of paraquat used, specifically 239
in the PE treatment, may have been detoxified earlier. To insure that superoxide was still present 240
at the time of the transcriptional analysis, we conducted an assay for superoxide radicals. 241
Relative superoxide levels were inferred from the intercalation of DNA with 2-hydroxyethidium 242
which is a product of the oxidation of dihydroethidium by the superoxide anion radical. 243
Endogenous superoxide has been shown to selectively oxidize dihydroethidium (47, 60, 61). As 244
shown in Fig. 4, the difference in superoxide levels is displayed at the time of cell harvest for 245
transcriptional profiling for both conditions. A statistically significant amount of intracellular 246
superoxide is still present in the experimental cultures. This result eliminates the possibility of 247
the paraquat-mediated production of superoxide radicals having been quenched before the 248
transcriptional analysis was performed. 249
A total of 190 genes were up-regulated and 86 genes were down-regulated (2.0-fold cutoff 250
with P < 0.05) during PE to paraquat with diversity in the functional classification of the 251
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[12]
differentially-expressed genes (Table S1). A substantial difference was observed within the 252
functional classifications of cellular processes, cell envelope, energy metabolism, translation, and 253
transport and binding proteins (Fig. 5A). Most differentially-expressed genes within these 254
categories were up-regulated. However, many of the genes within the regulation category were 255
down-regulated by PE to paraquat. The most interesting results from this experiment were the 256
comprehensive induction of genes involved in chemotaxis, translation factors, and the two 257
putative ATP synthase operons within the B. japonicum genome (Table 2). Specifically, out of a 258
total of 31/44 (2.0/1.5 fold-induction) differentially-expressed chemotaxis genes, all of them 259
were up regulated (Table S1). In addition, genes belonging to the sub-classifications of 260
molecular chaperones, nucleic acid modification and repair proteins, and oxidative stress 261
mechanisms were expected to be augmented in their relative expression levels, and such was the 262
case. The most unexpected result was the failure to detect the induction of canonical 263
detoxification enzymes such as superoxide dismutase and catalase. The two superoxide-264
dismutase enzymes (bll7559 and bll8073) which depend on a manganese co-factor were not 265
detected in B. japonicum, although the manganese co-factor enzyme has been demonstrated to be 266
the primary detoxification response to superoxide in E. coli (25, 26, 28). 267
A total of 299 genes were up-regulated and 105 were down regulated (2.0-fold cutoff with 268
P < 0.05) for the FS condition (Table S2). A majority of the differentially expressed genes are 269
listed under the classification of hypothetical or other categories. Most genes within the 270
categories of central intermediary metabolism, of which most function in nitrogen fixation or 271
nitrogen metabolism, and regulatory functions had augmented expression levels relative to 272
control. The FS treatment also induced the expression of the FixK2 transcription factor and most 273
of its target genes (Table 3) which are necessary for microaerobic, anaerobic, and symbiotic 274
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[13]
growth conditions (38). These targets include the fixNOQP operon which encodes proteins 275
related to an alternative, high-affinity cytochrome c oxidase complex (44, 45). Other targets 276
include the fixGHIS operon, rpoN1 sigma factor, heme biosynthesis genes, and possibly genes for 277
nitrate respiration all of which are induced when oxygen availability is limited (23, 37, 38, 45). 278
In a comparison of the transcriptional profiles of bacteroids isolated from soybean root 279
nodules and aerobically grown free-living cells, Pessi et al. report two statistically significant 280
overrepresented functional categories in the bacteroid cells which are central intermediary 281
metabolism and transport and binding proteins (41). A similar pattern was observed in cells 282
exposed to a FS treatment of paraquat (Fig. 5B and Table 3). In general, the results of the 283
transcriptional profiles of B. japonicum when subjected to paraquat-mediated oxidative stress are 284
strangely similar to a transcriptional profile of a bacteroid state. Genes within the sub-categories 285
of molecular chaperones, respiration, and transport and binding function were also up-regulated 286
(Table 3). Within the context of this experimental condition, the transcript levels for superoxide 287
dismutase and catalase were not changed as in the case of the PE condition. A complete list of 288
all genes for both conditions can be found in Tables S1 and S2 in the supplementary material. 289
Validation of microarray expression by qRT-PCR. In order to confirm the microarray 290
expression data, fold-induction values for representative genes from each experimental condition 291
were determined by qRT-PCR. The genes for qRT-PCR analysis were chosen based on a range 292
of fold-induction values and by their diversity in function. All of the chosen genes from both PE 293
and FS conditions yielded consistent results with little deviation (R2 = 0.92), indicating that the 294
microarray data have been validated by the qRT-PCR method (Fig. 6). 295
Differential activation of oxidative stress response mechanisms between PE and FS. PE 296
to paraquat induced 17 oxidative stress-responsive genes of which most were only moderately 297
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[14]
up-regulated, while the FS treatment induced 5 oxidative stress-responsive genes (Table 4). 298
Organic hydroperoxide resistance protein (bll0735) was the only common gene induced in both 299
PE and FS conditions. In addition, 4 sigma factor related genes were up-regulated in the FS 300
treatment (Table 3), indicating that these sigma factors function in the amelioration of fulminant 301
shock. Interestingly, the failure to detect the induction of catalase (kat) and superoxide dismutase 302
(sod) genes commonly known to be induced by oxidative stresses led us to employ qRT-PCR to 303
confirm the negative result. Genes encoding a Fe/Mn superoxide dismutase (bll7559), alkyl 304
hydroperoxide reductase (bll1777), and catalase (blr0778) also failed to meet 2.0 threshold 305
values by qRT-PCR analysis (data not shown), which suggests constitutive expression of these 306
proteins and dependence upon alternative methods for oxidative stress tolerance. 307
Translational activity is increased during PE, but decreased by FS. One interesting 308
finding in the gene expression data is related to translational machinery. Twenty one/fifty six 309
(2.0/1.5-fold cutoff) translation-related genes were expressed higher in the PE condition (Table 310
S1), while no genes encoding ribosomal proteins were expressed in the FS treatment (Tables 3 311
and S2). In general, there is an increased demand for translational machinery for metabolically 312
active cells. For example, more translation related genes were up-regulated in the faster growing 313
cells (15, 56). However, our growth data do not support an increased demand for translational 314
machinery due to higher metabolic activity, since the growth rate in 100 µM paraquat is 315
comparable to that of the control (Fig. 1). This discrepancy led us to measure total protein 316
content in response to paraquat-mediated oxidative stress during PE. The total cellular protein for 317
cultures supplemented with 100 μM paraquat was 15.61 ± 0.63 mg/ml while the protein content 318
of control cultures was 9.50 ± 0.47 mg/ml, which means a 61% increase in total cellular protein 319
was evident for treatment with paraquat. Given that all cells were harvested at the same O.D. 320
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[15]
(approximately 0.6 at 600 nm), our protein data are consistent with the transcriptomic data. In the 321
FS condition, there was no difference of total protein content between the control and 5 mM 322
paraquat for 10 min (data not shown), which is also in agreement with gene expression data. 323
However, we prefer to interpret the latter result in that 10 min was not sufficient enough to 324
compare levels of total proteins between before and after the treatment. 325
Paraquat-mediated oxidative stress modulates motility characteristics. Another 326
interesting finding was induction of motility related gene clusters in the PE treatment. Thirty 327
one/forty four (2.0/1.5-fold cutoff) chemotaxis genes were up-regulated (Table S1). The FS 328
treatment, to the contrary, caused a repression of most chemotaxis genes that were differentially 329
expressed at 2-fold cutoff (Table S2). The high expression of chemotaxis genes may indicate that 330
the chemical added to the culture is an attractant, and therefore bacteria move towards it. 331
However, our capillary assay has determined that paraquat has a negative chemotactic effect, 332
indicating that it is a repellent (Fig. 7). The ∆cheA (blr2343) mutant strain was constructed and 333
used as a negative control in the assay (Fig. 7). This gene was chosen because it encodes a key 334
two-component sensor histidine kinase and is part of a more complete gene cluster. Moreover, 335
the contiguous genes within the same putative operon were concomitantly induced during PE. 336
The result suggests that B. japonicum activates chemotaxis functions to stay away from a 337
potential toxicity caused by paraquat. This finding is of particular importance because expression 338
of chemotaxis genes can be induced by either an attractant or a repellant. The latter case is often 339
neglected in bacterial chemotaxis studies. Alternatively, we cannot rule out the possibility that 340
apparent gene expression may be modified at the post-transcriptional level. 341
Exopolysaccharide (EPS) production is enhanced in the presence of ROS. In this 342
transcriptional profiling, we expected that more EPS related genes would be expressed because 343
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[16]
of the protective nature of EPS against ROS (33). In a pre-determined hypothesis, one of the 344
physiological responses of B. japonicum to oxidative stress was an increase in the net production 345
of EPS as a possible compensatory mechanism for the innate host defense mechanisms (12, 33). 346
Nonetheless, there was no transcriptomic evidence for enhanced EPS production except for 347
blr7573 (exoU), which does not exclude the possibility of post-transcriptional regulation as the 348
cause of excessive EPS. Thus, we measured the total amount of EPSs during PE to paraquat. It is 349
nearly two-fold more than in the case of the control, although there is approximately 50% 350
increase in the 500 µM paraquat treatment (Fig. 8A), suggesting that EPS production is induced 351
by paraquat-mediated oxidative stress. However, there is no direct correlation between 352
increasing concentration and EPS production, which may suggest a refractory effect upon 353
achieving a threshold concentration. We also quantified the sub-class of acidic carbohydrate 354
content as expressed in glucuronic acid equivalents (Fig. 8B). Uronic acids such as alginate, 355
composed of mannuronic and guluronic acids, have been known to be involved in tolerance to 356
desiccation and associated oxidative stresses in Pseudomonas putida, a typical soil bacterium (16, 357
17). Acidic sugars were not detected in control cultures, while relatively similar amounts of 358
uronic acids were present across various concentrations of paraquat (Fig. 8B). EPS 359
quantification analysis for both the total and acidic sugars indicates that more EPS production is 360
one of the ROS-responsive mechanisms in B. japonicum. Regulation of EPS production in 361
response to paraquat-mediated oxidative stress may also be controlled at the post-transcriptional 362
level. 363
364
DISCUSSION 365
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[17]
In prolonged exposure (PE), B. japonicum requires longer acclimation time, compared with 366
the control (no paraquat), to adapt to superoxide-mediated oxidative stress presumably through 367
detoxification mechanisms as well as through other indirect mechanisms before exponential 368
growth is achieved. In fulminant shock (FS), the net effect of adding 5 mM paraquat to a 369
proliferating culture was a swift reduction in viable cell counts within a short time period (10 370
min) as was expected. A shock experimental condition within a short time period exceeds a 371
capacity of cells to respond and return to homeostasis which is evidenced by a rapid decrease in 372
viable cells or cell survival. This high concentration was chosen to generate a gene expression 373
profile of B. japonicum during a systemic and desperate response to oxidative stress whose 374
magnitude is rarely experienced for a lengthy period under normal environmental conditions. 375
A more consummate picture of the transcriptional activities of B. japonicum during oxidative 376
stress is made possible by comparing these two distinct parameters of PE and FS. Chemotaxis, 377
chaperones, respiration, translation, and transport are the primary functional categories affected 378
during PE to paraquat. Chemotaxis is worth addressing because most of the differentially 379
expressed genes in this category are up-regulated. However, in the FS treatment most of the 380
chemotaxis genes did not change or some were even down-regulated, which is consistent with 381
other genome-wide transcriptional studies of Staphylococcus aureus (14), Pseudomonas 382
aeruginosa (50), and E. coli (58) where shock concentrations of hydrogen peroxide were applied 383
to the cultures. Another study on E. coli which used paraquat for treatment of early-log phase 384
cultures did not show induction of chemotaxis related genes (43). The observed increase in 385
chemotaxis can be explained as an evasive mechanism. When exposed to low levels of ROS, 386
bacteria can respond by increasing motility to limit the damage of the superoxide radicals. 387
However, in the high level of ROS they may not have enough time to adjust and activate their 388
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[18]
motility machinery. Instead, cells may utilize much quicker mechanisms such as chaperones and 389
sigma factors to ameliorate the toxicity of ROS. 390
Aerobic respiration also appears to be regulated in both PE and FS conditions. The putative 391
ATP synthase operons and some proteins associated with electron transport immediately 392
upstream in the aerobic respiratory chain to ATP synthase, such as the Rieske Fe-S 393
protein/cytochrome bc1 complex (blr2485) and the cytochrome aa3 terminal oxidase (blr1171), 394
were moderately up-regulated during PE to paraquat. Tamarit et al. (54) report that the β-subunit 395
of the F0F1-ATPase is a specific protein damaged by superoxide stress. The induction of 396
cytochrome c oxidase genes has also been observed in global transcriptome analysis of E. coli 397
exposed to paraquat-mediated oxidative stress (43). Hassan et al. have demonstrated that toxicity 398
of paraquat is oxygen- and nutrient-dependent, and also augments what they have termed as a 399
cyanide-insensitive respiration which applies to the terminal cytochrome c oxidase complex (28, 400
29). They also postulate that paraquat is able to subvert electron flow through the cytochrome c 401
oxidase complex directly to molecular oxygen, thereby increasing endogenous superoxide. The 402
transcriptional profiles of B. japonicum appear to be consistent with these conjectures. This 403
phenomenon could confer a metabolic advantage to the bacteria as they adapt from a free-living 404
state into a symbiotic state which invariably involves the plant’s hypersensitive disease 405
resistance response. Further experiments would be required to corroborate this hypothesis. 406
Various genes encoding proteins from the tricarboxylic acid (TCA) cycle were up-regulated 407
in both conditions such as aconitase (bll0466), enolase (bll4794), succinate dehydrogenase 408
(blr0513), and most notably the pyruvate dehydrogenase complex (blr6333 and blr6334) which 409
links the TCA cycle and the glycolytic pathway. Genome-wide transcriptional profiles of P. 410
aeruginosa and Sinorhizobium meliloti indicate a moderate increase in expression levels for 411
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[19]
proteins within the pyruvate dehydrogenase complex in response to oxidative stress (5, 49). In 412
the case of P. aeruginosa, genes encoding proteins that function in the TCA cycle and the 413
terminal aerobic respiratory chain complex are up-regulated, which is in agreement with B. 414
japonicum gene expression. 415
The transcriptional profile for B. japonicum during FS oxidative stress was strangely 416
identical to a symbiotic state which included activation of the FixLJ-FixK2 regulatory pathway. 417
However, Mesa et al. reported a repression of these target genes of the FixLJ-FixK2 pathway in 418
response to oxidative stress (37), which is incongruous with our findings. The rationale for the 419
contradictory results would be the differing background conditions. In this study, all 420
transcriptional analyses were performed in aerobic conditions, while a micro-oxic condition was 421
used in the previous study, which may not allow for a direct comparison. Furthermore, Western 422
blot analysis conducted in vivo across a variety of conditions (i.e., oxic, micro-oxic, and anoxic) 423
reveals a somewhat constitutive expression of FixK2 transcripts, which hints at a possible control 424
mechanism at the post-translational level. Alternatively, it can be explained by the toxicity 425
mechanism of the paraquat radical which reacts with molecular oxygen, resulting in a superoxide 426
anion radical. A fulminant increase in the concentration of paraquat would cause a rapid 427
conversion of free oxygen into superoxide, thereby creating a transient micro-oxic condition in 428
which the induced FixK2 positively regulates the fixNOQP and fixGHIS operons (38). Hence, the 429
bacteria would adjust their transcriptional profiles as if they were in a bacteroid state whose 430
environment is typically micro-oxic. 431
Only one gene (blr7573, exoU) encoding UDP-hexose transferase associated with EPS 432
production was moderately up-regulated. Nevertheless, we performed EPS quantification 433
because an a priori hypothesis was that B. japonicum would produce more EPS as a defense 434
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[20]
mechanism against oxidative stress. In addition, a number of previous studies investigating 435
various stresses such as desiccation (16, 20, 55) and cytotoxic immune responses (12, 18, 19) 436
have reported elevated EPS production against those stresses. Indeed, our EPS quantification 437
results are consistent with that of previous findings, and therefore support our original hypothesis 438
regardless of the transcriptomic profiles, suggesting possible regulation of EPS production at the 439
post-transcriptional level. In a study on the virulence of Streptococcus mutans in response to 440
oxygen availability, the authors demonstrate that the EPS machinery is modulated post-441
transcriptionally in the regulation of the biofilm microhabitat (2). The subset of acidic, anionic 442
carbohydrate content of the EPS was also included in addition to total EPS content because the 443
acidic sugars have been demonstrated to play a crucial role in biofilm and EPS integrity in 444
bacteria within the Pseudomonas genus (30, 50). 445
Only a few genes encoding proteins involved in oxidative stress were highly expressed in 446
both PE and FS. The expression of the canonical detoxification enzymes superoxide dismutase 447
and catalase was not detected in either condition. However, moderate induction (between 1.5- 448
and 2.0-fold) was evident of organic hydroperoxide resistance proteins, glutathione related 449
genes, and some oxidoreductases (Table 4). Similarly, Salunke et al. did not detect up-regulation 450
of catalase (katA), superoxide dismutase (sodB) or alkyl hydroperoxide reductases (ahpB and 451
ahpC) in their expression analysis of three strains of P. aeruginosa in response to superoxide 452
stress (49). Moreover, evidence has been provided that these genes are already constitutively 453
expressed at such a high level, and thus any exposure to oxidative stress would not be effective 454
in further induction of these genes. This certainly appears to be the case for B. japonicum. The 455
superoxide assay in this study not only revealed that excessive superoxide was present at the 456
time of cell harvest, but that a basal level, presumably from endogenous ROS, was present in 457
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[21]
control cultures (data not shown) which would induce transcription of these detoxification 458
enzymes. In addition, FS caused significant up-regulation of a gene (hemN2) involved in heme 459
biosynthesis which has been demonstrated to be functional during anaerobic and symbiotic 460
growth (23). This is also evidence for a transcriptional profile that resembles that of a symbiotic 461
condition. The most pertinent conclusion from these enumerated findings is that B. japonicum is 462
capable of diverse physiological responses to tolerable oxidative stress levels which include 463
enhanced motility, translational activity, respiration, and EPS production. 464
465
ACKNOWLEDGEMENTS 466
We thank Raymond Jones in the Genomics Core Facility at the University of Texas-467
Arlington for providing technical support for the qRT-PCR analysis and Hae-In Lee for his 468
assistance in the chemotaxis mutant strain construction. This research was supported in part by 469
Basic Science Research Program through the National Research Foundation of Korea (NRF) 470
funded by the Ministry of Education, Science and Technology (2010-0016797), Inha University 471
Research grant, and a Research Enhancement Program (REP) grant from the University of 472
Texas-Arlington. 473
474
REFERENCES 475
1. Adler, J. 1973. A method for measuring chemotaxis and use of the method to determine 476 optimum conditions for chemotaxis by Escherichia coli. J. Gen. Microbiol. 74:77-91. 477 2. Ahn, S.J., Z. T. Wen, and R. A. Burne. 2007. Effects of oxygen on virulence traits of 478 Streptococcus mutans. J. Bacteriol. 189:8519-8527. 479 3. Baker, C. J, E. W. Orlandi. 1995. Active oxygen in plant pathogenesis. Annu. Rev. 480 Phytopathol. 33:299-321. 481 4. Baker, C. J., N. Mock, J. Glazener, and E. Orlandi. 1993. Recognition responses in 482 pathogen/non-host and race/cultivar interactions involving soybean (Glycine max) and 483 Pseudomonas syringae pathovars. Physiol. Mol. Plant Pathol. 43:81-94. 484
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[22]
5. Barloy-Hubler, F., A. Cheron, A. Hellegouarch, and F. Galibert. 2004. Smc01944, a 485 secreted peroxidase induced by oxidative stresses in Sinorhizobium meliloti 1021. 486 Microbiology. 150:657-664. 487 6. Bergmaier, D., C. Lacroix, M. G. Macedo, and C.P. Champagne. 2001. New method 488 for exopolysaccharide determination in culture broth using stirred ultrafiltration cells. 489 Appl. Microbiol. Biotechnol. 57:401-406. 490 7. Bittner, M., R. Butow, J. DeRisi, M. Diehn, J. Eberwine, C. B. Epstein, R. Glynne, 491 S. Grimmond, T. Ideker, J. E. Kacharmina, S. Katsabanis, J. Khan, J. Lee, C. L. 492 Liu, P. Marciano, F. M. Marincola, T. McIntosh, D. Monte, J. R. Pollack, V. 493 Rhodius, S. Somerville, E. Tom, E. Wang, J. S. Wei, D. Willhite, and S. Ybarra. 494 2003. Expression analysis of RNA, p. 101–288. In D. Bowtell and J. Sambrook (ed.), 495 DNA microarrays: a molecular cloning manual. Cold Spring Harbor Laboratory Press,496 Cold Spring Harbor, NY. 497 8. Blumenkrantz, N., and G. Asboe-Hansen. 1973. New method for quantitative 498 determination of uronic acids. Anal. Biochem. 54:484-489. 499 9. Bradford, M. M. 1976. A rapid and sensitive method for the quantitation of microgram 500 quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 501 72:248-254. 502 10. Breedveld, M. W., L. P. T. M. Zevenhuizen, and A. J. B. Zehnder. 1991.Osmotically- 503 regulated trehalose accumulation and cyclic -(1,2)-glucan excretion by Rhizobium 504 leguminosarum biovar trifolii TA-1. Arch. Microbiol. 156:501-506. 505 11. Broughton, W. J., S. Jabbouri, and X. Perret. 2000. Keys to symbiotic harmony. 506 J. Bacteriol. 182:5641-5652. 507 12. Bylund, J., L. A. Burgess, P. Cescutti, R. K. Ernst, and D. P. Speert. 2006. 508 Exopolysaccharides from Burkholderia cenocepacia inhibit neutrophil chemotaxis and 509 scavenge reactive oxygen species. J. Biol. Chem. 281:2526-2532. 510 13. Carr, R. J. G., R. F. Bilton, and T. Atkinson. 1986. Toxicity of paraquat to 511 microorganisms. Appl. Environ. Microbiol. 52:1112-1116. 512 14. Chang, W., D. A. Small, F. Toghrol, and W. E. Bentley. 2006. Global transcriptome 513 analysis of Staphylococcus aureus response to hydrogen peroxide. J. Bacteriol. 514 188:1648-1659. 515 15. Chang, W.-S., W. L. Franck, E. Cytryn, S. Jeong, T. Joshi, D.W. Emerich, M. 516 Sadowsky, D. Xu, and G. Stacey. 2007. An oligonucleotide microarray resource for 517 transcriptional profiling of Bradyrhizobium japonicum. Mol. Plant-Microbe Interact. 518 20:1298-1307. 519 16. Chang, W.-S., M. van de Mortel, L. Nielson, G. N. de Guzman, X. Li, and L. J. 520 Halverson. 2007. Alginate production by Pseuodomonas putida creates a hydrated 521 microenvironment and contributes to biofilm architecture and stress tolerance under 522 water-limiting conditions. J. Bacteriol. 189:8290-8299. 523 17. Chang, W.-S., X. Li, and L. J. Halverson. 2009 Reduced water availability 524 influences endogenous oxidative stress and cell death within unsaturated Pseudomonas 525 putida biofilms. Environ. Microbiol. 11: 1482-1492. 526 18. Costerton, J. W., K. J. Cheng, G. G. Geesy, T. I. Ladd, J. C. Nickel, M. Dasgupta, 527 and T. J. Marrie. 1987. Bacterial biofilms in nature and disease. Annu. Rev. Microbiol. 528 41:435-464. 529
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[23]
19. Costerton, J. W., P. S. Stewart, and E. P. Greenberg. 1999. Bacterial biofilms: a 530 common cause of persistent infections. Science. 284:1318-1322. 531 20. Cytryn, E. J., D. P. Sangurdekar, J. G. Streeter, W. L. Franck, W.-S. Chang, G. 532 Stacey, D. W. Emerich, T. Joshi, D. Xu, and M. J. Sadowsky. 2007. Transcriptional 533 and physiological responses of Bradyrhizobium japonicum to desiccation-induced stress. 534 J. Bacteriol. 189:6751-6762. 535 21. Dubois, M., K. A. Giles, J. K. Hamilton, P. A. Rebers, and F. Smith. 1956. 536 Colorimetric method for determination of sugars and related substances. Anal. Chem. 537 28:350-356. 538 22. Fellay, R., J. Frey, and H. Krisch. 1987. Interposon mutagenesis of soil and 539 water bacteria: a family of DNA fragments designed for in vitro insertional mutagenesis 540 of gram-negative bacteria. Gene 52:147-154. 541 23. Fischer, H. M., L. Velasco, M. J. Delgado, E. J. Bedmar, S. Scharen, D. Zingg, M. 542 Gottfert, and H. Hennecke. 2001. One of two hemN genes in Bradyrhizobium 543 japonicum is functional during anaerobic growth and in symbiosis. J. Bacteriol. 544 183:1300-1311. 545 24. Franck, W. L., W.-S. Chang, J. Qiu, M. Sugawara, M. J. Sadowsky, S. A. Smith, 546 and G. Stacey. 2008. Whole-genome transcriptional profiling of Bradyrhizobium 547 japonicum during chemoautotrophic growth. J. Bacteriol. 190:6697-6705. 548 25. Gregory, E. M., F. J. Yost, Jr., and I. Fridovich. 1973. Superoxide dismutases of 549 Escherichia coli: intacellular localization and functions. J. Bacteriol. 115:987-991. 550 26. Hassan, H. M., and I. Fridovich. 1977. Enzymatic defenses against the toxicity of 551 oxygen and of streptonigrin in Escherichia coli. J. Bacteriol. 129:1574-1583. 552 27. Hassan, H. M., and I. Fridovich. 1979. Paraquat and Escherichia coli: mechanism and 553 production of extracellular superoxide radical. J. Biol. Chem. 254:10846-10852. 554 28. Hassan, H. M., and I. Fridovich. 1977. Regulation of the synthesis of superoxide 555 dismutase in Escherichia coli. J. Biol. Chem. 252:7667-7672. 556 29. Hassan, H. M., and I. Fridovich. 1978. Superoxide radical and oxygen enhancement of 557 the toxicity of paraquat to Escherichia coli. J. Biol. Chem. 253:8143-8148. 558 30. Kachlany, S.C., S. B. Levery, J. S. Kim, B. L. Reuhs, L. W. Lion, and W. C. 559 Ghiorse.2001. Structure and carbohydrate analysis of the exopolysaccharide capsule of 560 Pseudomonas putida G7. Environ. Microbiol. 3:774-784. 561 31. Keppler, L. D., C. J. Baker, and M. M. Atkinson. 1989. Active oxygen production 562 during a bacteria-induced hypersensitive reaction in tobacco suspension cells. 563 Phytopathology. 79:974-978. 564 32. Kerr, M. K., M. Martin, and G. A. Churchill. 2000. Analysis of variance for gene 565 expression microarray data. J. Comput. Biol. 7:819-837. 566 33. Kiraly, Z., H. M. El-Zahaby, and Z. Klement. 1997. Role of extracellular 567 polysaccharide (EPS) slime of plant pathogenic bacteria in protecting cells to reactive 568 oxygen species. J. Phytopathol. 145:59-68. 569 34. Kuzma M. M., S. Hunt, and D. B. Layzell. 1993. Role of oxygen in the limitation and 570 inhibition of nitrogenase activity and respiration rate in individual soybean nodules. Plant 571 Physiol. 101:161-169. 572 35. Levine, A., R. Tenhaken, R. Dixon, and C. Lamb. 1994. H2O2 from the oxidative burst 573 orchestrates the plant hypersensitive disease resistance response. Cell. 79:583-593. 574
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[24]
36. Mehdy, M. C. 1994. Active oxygen species in plant defense against pathogens. Plant 575 Physiol. 105:467-472. 576 37. Mesa, S., Z. Ucurum, H. Hennecke, and H. M. Fischer. 2005. Transcription activation 577 in vitro by the Bradyrhizobium japonicum regulatory protein FixK2. J. Bacteriol. 578 187:3329-3338. 579 38. Nellen-Anthamatten, D., P. Rossi, O. Preisig, I. Kullik, M. Babst, H. M. Fischer, and 580 H. Henneke. 1998. Bradyrhizobium japonicum FixK2, a crucial distributor in the FixLJ-581 dependent regulatory cascade for control of genes inducible by low oxygen levels. J. 582 Bacteriol. 180:5251-5255. 583 39. Panek, H. R., and M. R. O’Brian. 2004. KatG is the primary detoxifier of hydrogen 584 peroxide produced by aerobic metabolism in Bradyrhizobium japonicum. J. Bacteriol. 585 186:7874-7880. 586 40. Parniske, M., and J. A. Downie. 2003. Locks, keys, and symbioses. Nature. 425:569-587 570. 588 41. Pessi, G., C. H. Ahrens, H. Rehrauer, A. Lindemann, F. Hauser, H. M. Fischer, and 589 H. Henneke. 2007. Genome-wide transcript analysis of Bradyrhizobium japonicum 590 bacteroids in soybean root nodules. Mol. Plant-Microbe Interact. 20:1353-1363. 591 42. Pfaffl, M. W. 2001. A new mathematical model for relative quantification in real-time 592 RT-PCR. Nucleic Acids Res. 29:2002-2007. 593 43. Pomposiello, P. J., M. H. J. Bennik, and B. Demple. 2001. Genome-wide 594 transcriptional profiling of the Escherichia coli responses to superoxide stress and 595 sodium salicylate. J. Bacteriol. 183:3890-3902. 596 44. Preisig, O., D. Anthamatten, and H. Hennecke. 1993. Genes for a microaerobically 597 induced oxidase complex in Bradyrhizobium japonicum are essential for a nitrogen-fixing 598 endosymbiosis. Proc. Natl. Acad. Sci. USA. 90:3309-3313. 599 45. Preisig, O., R. Zufferey, L. Thony-Meyer, C. A. Appleby, and H. Hennecke. 1996. A 600 high-affinity cbb3-type cytochrome oxidase terminates the symbiosis-specific respiratory 601 chain of Bradyrhizobium japonicum. J. Bacteriol. 178:1532-1538. 602 46. Read, R. R., and J. W. Costerton. 1987. Purification and characterization of adhesive 603 exopolysaccharides from Pseudomonas putida and Pseudomonas fluorescens. Can. J. 604 Microbiol. 33:1080-1090. 605 47. Robinson, K. M., M. S. James, M. Pehar, J. S. Monette, M. F. Ross, T. M. Hagen, 606 M. P. Murphy, and J. S. Beckman. 2006. Selective fluorescent imaging of superoxide 607 in vivo using ethidium-based probes. Proc. Natl. Acad. Sci. USA. 103:15038-15043. 608 48. Sadowski, M. J., R. E. Tully, P. B. Cregan, and H. H. Keyser. 1987. Genetic diversity 609 in Bradyrhizobium japonicum serogroup 123 and its relation to genotype-specific 610 nodulation of soybean. Appl. Environ. Microbiol. 53:2624-2630. 611 49. Salunkhe, P., F. von Gotz, L. Wiehlmann, J. Lauber, J. Buer, and B. Tummler. 612 2002. Genechip expression analysis of the response of Pseudomonas aeruginosa to 613 paraquat-induced superoxide stress. Genome Lett. 4:165-174. 614 50. Salunkhe, P., T. Topfer, J. Buer, and B. Tummler. 2005. Genome-wide transcriptional 615 profiling of the steady-state response of Pseudomonas aeruginosa to hydrogen peroxide. 616 J. Bacteriol. 187:2565-2572. 617 51. Santos R., D. Herouart, S. Siguad, D. Touati, and A. Puppo. 2001. Oxidative burst in 618 alfalfa-Sinorhizobium meliloti symbiotic interaction. Mol. Plant-Microbe Interact. 14:86-619 89. 620
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[25]
52. Sutherland, I. 2001. Biofilm exopolysaccharides: a strong and sticky framework. 621 Microbiology. 147:3-9. 622 53. Sutherland, M. W. 1991. The generation of oxygen radicals during host plant responses 623 to infection. Physiol. Mol. Plant Pathol. 39:79-94. 624 54. Tamarit, J., E. Cabiscol, and J. Ros. 1998. Identification of the major oxidatively 625 damaged proteins in Escherichia coli cells exposed to oxidative stress. J. Biol. Chem. 626 273:3027-3032. 627 55. Tamaru, Y., Y. Takani, T. Yoshida, and T. Sakamoto. 2005. Crucial role of 628 extracellular polysaccharides in desiccation and freezing tolerance in the terrestrial 629 cyanobacterium Nostoc commune. Appl. Environ. Microbiol. 71:7327-7333. 630 56. Tao, H., C. Bausch., C. Richmond, F. R. Blattner, and T. Conway. 1999. Functional 631 Genomics: expression analysis of Escherichia coli growing on minimal and rich media. 632 J. Bacteriol. 181:6425-6440. 633 57. Tusher, V. G., R. Tibshirani, and G. Chu. 2001. Significance analysis of microarrays 634 applied to the ionizing radiation response. Proc. Natl. Acad. Sci. USA. 98:5116-5121. 635 58. Wang, S., K. Deng, S. Zaremba, X. Deng, C. Lin, Q. Wang, M. L., Tortorello, and 636 W. Zhang. 2009. Transcriptomic response of Escherichia coli O157:H7 to oxidative 637 stress. Appl. Environ. Microbiol. 75:6110-6123. 638 59. Windgassen, M., A. Urban, and K. E. Jaeger. 2000. Rapid gene inactivation in 639 Pseudomonas aeruginosa. FEMS Microbiol. Lett. 193:201-205. 640 60. Zhao, H., J. Joseph, H. M. Fales, E A. Sokoloski, R. L. Levine, J. Vasquez-Vivar, 641 and B. Kalyanaraman. 2005. Detection and characterization of the product of 642 hydroethidine and intracellular superoxide by HPLC and limitations of fluorescence. 643 Proc. Natl. Acad. Sci. USA. 102:5727-5732. 644 61. Zhao, H., S. Kalivendi, H. Zhang, J. Joseph, K. Nithipatikom, J. Vasquez-Vivar, and 645 B. Kalyanaraman. 2003. Superoxide reacts with hydroethidine but forms a fluorescent 646 product that is distinctly different from ethidium: potential implications in intracellular 647 fluorescence detection of superoxide. Free Radic. Biol. Med. 34:1359-1368. 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[26]
Table 1. Gene-specific primers used for qRT-PCR. 667 Primer Sequence Gene Description
bll6866F GGTATCGATCGCTCCAAGATCCA flagellin bll6866R TCAGCCAGTTGATGCCGTTGAT
blr6333F ATCAATCATCCCGAGGTCGCCA pyruvate dehydrogenase complex E2 blr6333R AGCTGCAGGAGAGGTTCATCAT
blr1602F AGCTTCGCTTGTTTCCTGTCGT ABC transporter permease protein blr1602R TGAATCCGCTCAAGACCGCTT
bll5412F AAAGAAGCGGTCGAACAGCTCA 50S ribosomal Protein L10 blr5412R GACGGTGAGGCCGGAATATTGA
bll6879F AGTCTGCATCTGCATCGGAAGAG probable flagellar motor switch protein bll6879R TCGCATCGACATTGGCGAACTT
bsr4258F TGAAGTGGACGCGGTAAACCT hypothetical protein bsr4258 bsr4258R GGCCGAACTTGAACGTGGATGAA
bll0256F GACATTGCCGATGCGAAATGGA transcriptional regulatory protein bll0256R GCGACATGGTCCTTGGCATTT
blr2581F TCGAATCGCCGGCATTGGGAT putative D-fructose-1,6-bisphosphatase protein blr2581R ACCCAGCGCATGTTGAAATCGT
bll1028F AGGCGACTTCGGACGATATGCTT RNA polymerase sigma factor (carQ) bll1028R TTGTGCCGGCAATAGAGGATGT
blr2764F ATCGTCGGCATCCTCTTGGTGAT cytochrome-c oxidase (fixO) blr2764R TCTCGATCGTGCTCTTGAGGTAGA
blr4635F TCGAGGGAATGAAATTCGAC chaperonin GroEL (groEL5) blr4635R CGCAATTGATTGACGATGAG
blr3815F GATGTTGCCGGAAGAATTTG putative cation-transporting ATPase blr3815R TCTGGTTCTGCGTCAGTGTC
bll4012F TGCCAAGCTTCTCTTTACCG organic hydroperoxide resistance protein bll4012R TCGGATTGTTGAGGTCGAT
bsl2574F AACGCGTTTGGAAAGTGAAC hypothetical protein bsl2574 bsl2574R CTACTTCCGCAGCAATTCG
blr4701F TGGAACAACGTGCTCTGGTA putative outer-membrane immunogenic protein blr4701R CCGAGGAAGGTGAAAGTGTC precursor bll5813F TCTCAGAGAACATCGCGAAC flagellar basal-body rod protein (flgC) bll5813R CGGATTGTTCGGTTCGTATT
blr2221F CTGCTGATCTTCGACGAGGT adenosylmethionine-8-amino-7-oxononanoate blr2221R ACAGCTCCATCTGGTTCTCG aminotransferase (bioA) blr5517F TGGAAATGTACAGCCACTACGTGC two-component response regulator blr5517R TGATGATCGGCGCGTAAGGATT
bll0631F TCAACCTTCTGACGGTGAACGC chromosome partitioning protein A (parA) bll0631R TGCAGCAATTGCGACAGACCTT for normalization
668 669
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[27]
Table 2. A subset of B. japonicum genes and their clusters significantly induced during prolonged 670 exposure (PE) to paraquat a. 671 Locus (gene ID) Gene descriptionb Fold induction Chemotaxis
bll1439 (ctpD) pilus assembly protein 2.3 bll1440 (ctpC) pilus assembly protein 2.1 bll6851 (flhA) flagellar biosynthesis protein 2.1 bll6853 (flgD) hook formation protein 7.4 bll6854 (flbT) flagellin synthesis repressor protein 3.2 bll6855 (fliC) probable flagellar protein 2.9 bll6856 (flgL) probable flagellar hook-associated protein 5.9 bll6857 (flgK) hook associated protein I homolog 7.5 bll6858 (flgE) flagellar hook protein 7.8 bll6861 (motC) probable chemotaxis protein precursor 2.7 bll6862 (motB) probable flagellar motor protein 5.3 bll6865 (fla) flagellin 18.8 bll6866 (fla) flagellin 19.5 bll6869 (flgH) flagellar L-ring protein precursor 2.9 bll6871 (flgI) flagellar P-ring protein precursor 4.1 bll6873 (flgG) flagellar basal-body rod protein 4.5 bll6874 (fliE) flagellar hook-basal body complex protein 3.7 bll6875 (flgC) flagellar basal-body rod protein 4.6 bll6876 (flgB) flagellar basal-body rod protein 6.0 bll6877 (flhB) flagellar biosynthetic protein 2.9 bll6878 (fliG) probable flagellar motor switch protein 2.9 bll6879 (fliN) probable flagellar motor switch protein 2.6
Chaperones bll2059 (groEL3) GroEL3 chaperonin 2.0
bll5834 (dksA) dnaK deletion suppressor protein 2.2 blr5625 (groES) 10 KD chaperonin 8.4 blr5626 (groEL) 60 KDA chaperonin 8.3 blr6978 (groES2) chaperonin 3.7 blr7533 (groEL) 60 KDA chaperonin 3.9 bsr7532 (groES) 10 KD chaperonin (protein CPN10) 5.1
Respiration bll0439 (atpC) ATP synthase epsilon chain 2.0
bll0441 (atpG) ATP synthase gamma chain 2.0 bll0442 (atpA) ATP synthase alpha chain 2.3 bll0443 (atpH) ATP synthase delta chain 2.1 bll1185 (atpB) FoF1 ATP synthase B chain 3.7 bll1186 (atpB) FoF1 ATP synthase B' chain 2.8
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[28]
bsl1187 (atpC) FoF1 ATP synthase C chain 2.3 bll1188 (atpA) FoF1 ATP synthase A chain 2.7 blr6333 (bkdB) dehydrogenase complex E2 7.0 blr6334 (lpdA) dihydrolipoamide dehydrogenase 6.0
Transcription and Translation bll4076 50S ribosomal protein L9 2.0
bll4079 30S ribosomal protein S6 2.1 bll5399 50S ribosomal protein L4 2.0 bll5401 30S ribosomal protein S10 2.2 bll5402 elongation factor TU 2.3 bll5403 translation elongation factor G 2.2 bll5404 30S ribosomal protein S7 2.4 bll5405 30S ribosomal protein S12 2.1 bll5411 50S ribosomal protein L7/L12 2.3 bll5412 50S ribosomal Protein L10 3.2 bll5414 50S ribosomal protein L1 2.1 bll7440 peptidyl-tRNA hydrolase 2.0 bll7441 50S ribosomal protein L25 2.4
Transport bll6453 ABC transporter ATP-binding protein 2.2
bll6455 ABC transporter substrate-binding protein 3.0 bll6832 probable ABC transporter permease protein 2.4 bll6833 probable ABC transporter permease protein 3.6 bll6834 probable ABC transporter substrate-binding protein 5.6 blr1601 ABC transporter substrate-binding protein 4.4 blr1602 ABC transporter permease protein 5.3 blr1604 ABC transporter ATP-binding protein 2.4 blr3917 ABC transporter ATP-binding protein 2.1 blr3920 ABC transporter permease protein 2.6
a Differentially expressed genes from each functional category were selected based on a 2-fold 672 cut-off with q value ≤ 0.05. 673 b Gene description (annotations) represents the third level from the three-tiered functional level 674 system of B. japonicum (www.kazusa.or.jp/rhizobase/Bradyrhizobium/cgi-675 bin/category_brady.cgi). 676 677 678 679 680 681 682
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[29]
Table 3. A subset of B. japonicum genes and their clusters significantly induced during paraquat 683 fulminant shock (FS) treatmenta. 684
Locus (Gene ID) Descriptionb Fold induction Chaperones bll0729 small heat shock protein 2.0 blr4635 (groEL) chaperonin GroEL 6.4 blr4637 probable HspC2 heat shock protein 3.4 blr4653 molecular chaperone DnaJ family 7.5 blr7740 small heat shock protein 2.1 blr7961 probable HspC2 heat shock protein 2.5
Respiration bll3998 probable succinate-semialdehyde dehydrogenase 8.1 blr1309 acetyl-coenzyme A synthetase 2.7 blr2584 (cbbA) putative fructose-1,6-bisphosphate aldolase protein 2.0 blr4655 Phosphoenolpyruvate synthase 2.8 blr4657 beta-glucosidase 4.4 blr4658 probable Glucokinase (EC 2.7.1.2) 3.1 blr4955 putative cytochrome B561 2.0 blr6062 putative cytochrome C6 precursor 2.0 blr6128 cytochrome c552 5.6 blr6333 (bkdB) dehydrogenase complex E2 2.2 blr6636 ATP synthase subunit 2.2 blr7040 (napC) cytochrome C-type protein 5.1
Nitrogen fixation and metabolism bll2757 (fixK2) transcriptional regulatory protein 3.4 blr2763 (fixN) cytochrome-c oxidase 6.7 blr2764 (fixO) cytochrome-c oxidase 8.0 blr2766 (fixP) cbb3 oxidase subunit III 4.5 blr2767 (fixG) iron-sulfur cluster-binding protein 5.9 blr2768 (fixH) FixH protein 3.0 blr3125 (cycH) cytochrome C-type biogenesis protein 2.0 blr3126 (cycJ) cytochrome C-type biogenesis protein 2.0 blr3127 (cycK) cytochrome C-type biogenesis protein 2.2 blr3128 (cycL) cytochrome C-type biogenesis protein 2.3 blr5778 (fixG) nitrogen fixation protein 2.0 blr7089 respiratory nitrite reductase 3.2
Lipid metabolism bll1200 (hemA) 5-aminolevulinic acid synthase 5.0 blr1288 probable long-chain-fatty-acid-CoA ligase 2.6
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[30]
blr2989 putative acyl-CoA dehydrogenase 2.0
Transcription and Translation bll1028 (carQ) RNA polymerase sigma factor 16.4 bll1447 (rhlE) dead-box ATP-dependent RNA helicase 2.2 blr1883 (rpoN1) RNA polymerase sigma-54 subunit 2.0 blr3042 putative RNA polymerase sigma factor 3.9 blr7337 (rpoH2) sigma32-like transcription factor 2.3 blr1404 (clpB) ATP-dependent protease ATP-binding subunit 2.5
Transport bll6063 ABC transporter substrate-binding protein 3.3 bll6064 ABC transporter ATP-binding protein 3.3 bll7988 probable ATP-binding protein 4.4 blr0316 (nosD) periplasmic copper-binding precursor 3.0 blr1093 (pstA) phosphate ABC transporter permease protein 2.0 blr1094 (pstB) phosphate ABC transporter ATP-binding protein 2.3 blr1233 putative sulfonate binding protein 3.6 blr3815 putative cation-transporting ATPase 4.9 blr4112 Probale cation efflux system protein 4.5 blr4932 putative cation efflux system protein 2.0 blr6523 putative iron transport protein 3.0 blr7037 (napD) periplasmic nitrate reductase 9.2 blr7038 (napA) periplasmic nitrate reductase large subunit precursor 5.9
blr7039 (napB) periplasmic nitrate reductase small subunit precursor 4.4
blr7090 probable periplasmic nitrate reductase 4.0 blr7873 ABC transporter ATP-binding protein 3.2 blr7874 ABC transporter permease protein 2.0 bsr4636 putative cation transport regulator 5.6 bsr7036 (napE) periplasmic nitrate reductase protein 3.3
Amino acid biosynthesis blr0314 (nosR) nitrous oxide reductase expression regulator 3.1 blr0315 (nosZ) nitrous oxide reductase 4.1 blr3212 (norE) nitric oxide reductase subunit E 2.3 blr3214 (norC) nitric oxide reductase subunit C 3.2 blr3215 (norB) nitric oxide reductase subunit B 2.7 blr3216 (norA) NorQ protein 2.7 blr5774 probable sulfide-quinone reductase 3.1
685
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[31]
a Differentially expressed genes from each functional category were selected based on a 2-fold 686 cut-off with q value ≤ 0.05. 687 b Gene description (annotations) represents the third level from the three-tiered functional level 688 system of B. japonicum (www.kazusa.or.jp/rhizobase/Bradyrhizobium/cgi-689 bin/category_brady.cgi). 690 691 692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[32]
Table 4. Oxidative stress-responsive genes differentially expressed during paraquat-mediated oxidative 714 stressa. 715
Locus (Gene ID) Descriptionb Fold inductionc
Prolonged Exposure (PE) blr0528 probable oxidoreductase 1.5 blr0594 thioredoxin 1.5 bll0668 glutathione synthetase 1.7 bll0735 organic hydroperoxide resistance protein 1.8 bll0773 hypothetical glutathione S-transferase like protein 1.7 bll1162 glutathione S-transferase 1.6 bll2508 hypothetical glutathione S-transferase like protein 1.9 blr2609 oxidoreductase -2.8 bll2737 oxidoreductase with iron-sulfur subunit -2.0 bll2856 probable cytochrome P450 1.6 blr2939 glutathione S-transferase 1.7 bll3914 oxidoreductase 2.8 bll4059 NAD/NADP dependent oxidoreductase 1.7 bll5338 probable oxidoreductase 1.9 blr5618 putative oxidoreductase protein 1.7 blr6218 putative oxidoreductase protein -1.7 bll7007 putative oxidoreductase 1.9 bll7404 hypothetical glutathione S-transferase like protein 1.6 blr7566 oxidoreductase 1.7 bll7606 oxidoreductase 1.5
Fulmimant Shock (FS) bll0088 oxidoreductase -2.0 bll0735 organic hydroperoxide resistance protein 1.5 bll1244 flavin dependant oxidoreductase 1.8 bll4012 organic hydroperoxide resistance protein 1.8 bll5081 putative multidrug resistance protein 3.6 bll7086 (hemN) anaerobic coproporphyrinogen III oxidase 8.2
a Differentially expressed genes were selected based on a 1.5-fold cut-off with q value ≤ 0.05. 716 b Gene description (annotations) represents the third level from the three-tiered functional level 717 system of B. japonicum (www.kazusa.or.jp/rhizobase/Bradyrhizobium/cgi-718 bin/category_brady.cgi). 719 c Positive and negative values indicate up and down regulation in response to paraquat compared 720 to the control without paraquat. 721
722
723
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[33]
Figure legends 724
725
Figure 1. The effect of paraquat on the growth of B. japonicum USDA 110 in AG media. The 726
raw absorbance data have been log10 transformed and plotted as a function of time in hours. 727
Control cultures were treated with sterilized ddH2O. Symbols are as follows: control (); 10 μM 728
(); 50 μM (); 100 μM (∆); and 1 mM (). Each point is the mean of three replicates and the 729
error bars represent the standard error of the mean. 730
731
Figure 2. The effect of fulminant shock (FS) treatment with paraquat on the survival of B. 732
japonicum USDA 110. Control cultures were treated with ddH2O. Symbols are as follows: 733
control (); 100 μM (); 250 μM (); 500 μM (∆); 750 μM (); 1 mM (); 5 mM (♦); 10 mM 734
(◊); and 20 mM (). Values are means ± standard error of the mean for three replications. 735
736
Figure 3. A comparison of the total number of differentially expressed genes between prolonged 737
exposure (PE) and fulminant shock (FS) conditions. Values shown represent more than 2-fold 738
(in bold) and 1.5-fold (in parentheses) differential expression with P < 0.05. 739
740
Figure 4. Measurement of ROS in B. japonicum cultures. Absorbance at 300 nm of 2-741
hydroxyethidium was calculated for each condition at time of cell harvest. The contribution of 742
paraquat and the media to the absorbance values were accounted for by including them as 743
controls and subtracting those values from the measured absorbance values, resulting in relative 744
absorbance. Black bars represent control without treatments and grey bars represent paraquat 745
treatment conditions (100 μM for prolonged exposure (PE), and 5 mM for fulminant shock (FS)). 746
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[34]
The differences between control and treatment for both treatments are statistically significant (P 747
< 0.05). Error bars represent the standard error of the mean for nine replicates. 748
749
Figure 5. Functional classification of differentially expressed genes with > 2.0 fold changes. (A) 750
Comparison of differential expression in the prolonged exposure (PE) condition. (B) Comparison 751
of differential expression in the fulminant shock (FS) condition. Black bars represent positive 752
fold-induction values and grey bars represent negative fold-induction values. Functional 753
classifications were derived from B. japonicum genome annotations available through Rhizobase 754
(http://bacteria.kazusa.or.jp/rhizobase/). 755
756
Figure 6. Comparison of log2 transformed qRT-PCR data and microarray data of 18 757
representative genes from both conditions. These genes were selected based on fold-induction 758
and functional categories. The corresponding gene names for the locus IDs are as follows: 759
bll1028 (carQ), blr2764 (fixO), blr4635 (groEL), bll5813 (flgC), blr2221 (bioA), bll6866 (fla), 760
blr6333 (bkdB), bll5412 (rplJ), bll6879 (fliN), and blr2581 (cbbF). blr1602, blr4258, bll0256, 761
blr3815, bll4012, bsl2574, blr4701, blr5517, and bll0631 lack assigned names. Open triangles, 762
prolonged exposure (PE); closed triangles, fulminant shock (FS). 763
764
Figure 7. The ratio of treatment (100 µM paraquat) to control of relative CFUs in capillaries 765
plotted against time. Error bars are the standard error of the mean for three replicates. If the ratio 766
> 1, chemotaxis is positive and the test substance is an attractant. If the ratio < 1, chemotaxis is 767
negative and the test substance is a repellant. A ratio close to 1 is indicative of no net effect on 768
chemotaxis. Symbols are as follows: , wild-type; ∇, ∆cheA mutant. 769
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[35]
770
Figure 8. Quantification of the total EPS (A) and acidic sugar content (B) isolated from the 771
supernatant of B. japonicum cultures treated with 0, 100, 250, and 500 µM of paraquat. All 772
values in the paraquat treatments are statistically significant from control (P < 0.05). Error bars 773
represent the standard error of the mean for three replicates. 774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[36]
Figure 1: 793
Time (hours)
0 20 40 60 80
Abso
rban
ce (6
00nm
)
0.001
0.01
0.1
1
10
794
795
796
797
798
799
800
801
802
803
804
805
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[37]
Figure 2: 806
Time (min)
0 10 20 30
Surv
ival
(%)
0
20
40
60
80
100
120
807
808
809
810
811
812
813
814
815
816
817
818
819
820
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[38]
Figure 3: 821
822 823
Up Up 824 PE FS 825 826 827 828 178 5 287 829 (443) (35) (475) 830 831 832 7 7 833 (40) (13) 834 835 836 98 0 79 837 (271) (5) (107) 838 839 840 841 842 Down Down 843 FS PE 844 845
846
847
848
849
850
851
852
853
854
855
856
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[39]
Figure 4: 857
Rel
ativ
e Ab
sorb
ance
(300
nm
)
0.0
0.5
1.0
1.5
2.0
2.5
3.0
PE FS 858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[40]
Figure 5: 874
Number of genes
0 10 20 30 40 50 60 70
Func
tiona
l cat
egor
ies
Amino acid biosynthesis
Biosynthesis of cofactors, prosthetic groups, and carriers
Cell envelope
Cellular processes
Central intermediary metabolism
DNA replication, recombination, and repair
Energy metabolism
Fatty acid, phospholipid and sterol metabolism
Purines, pyrimidines, nucleosides, and nucleotides
Regulatory functions
Transcription
Translation
Transport and binding proteins
Other categories
Hypothetical protein
0 50 150
A B
PE FS
200
875
876
877
878
879
880
881
882
883
884
885
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[41]
Fig. 6: 886
Microarray data (log2 fold change)
qRT-
PCR
dat
a (lo
g 2 fo
ld c
hang
e)
-4 -2 0 2 4
-4
-2
0
2
4
y = 0.93x - 0.30
R2 = 0.92bll5813blr4701
blr2221bsl2574
blr5517
bsr4258
bll0256blr2581
bll5412bll6879
bll4012blr1602
blr6333blr2764
bll1028bll6866
blr3815blr4635
887
888
889
890
891
892
893
894
895
896
897
898
899
900
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[42]
Figure 7: 901
Time (min)
Rel
ativ
e ra
tios
of C
FUs
in c
apill
ary
(Tre
atm
ent/C
ontro
l)
0.4
0.6
0.8
1.0
1.2
20 30 40
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from
[43]
Figure 8: 917
µ g G
lc e
q's
/ mg
prot
ein
0
20
40
60
80
100
120
A
Paraquat (µM)
µ g G
lcU
A eq
's /
mg
prot
ein
0
2
4
6
8
10
0 100 250 500
B
918
on Septem
ber 6, 2018 by guesthttp://aem
.asm.org/
Dow
nloaded from