O2 availability impacts iron homeostasis in
Escherichia coli
Nicole A. Beauchenea, Erin L. Metterta, Laura J. Mooreb, Sündüz Keleşc,d, Emily R. Willeya, and Patricia J. Kileya,1
a
Department of Biomolecular Chemistry, University of Wisconsin–Madison, Madison, WI 53706; bDepartment of Chemistry and Biochemistry, Monmouth
College, Monmouth, IL 61462; cDepartment of Statistics, University of Wisconsin–Madison, Madison, WI 53706; and dDepartment of Biostatistics and
Medical Informatics, University of Wisconsin–Madison, Madison, WI 53706
Edited by Carol A. Gross, University of California, San Francisco, CA, and approved October 3, 2017 (received for review April 30, 2017)
The ferric-uptake regulator (Fur) is an Fe2+-responsive transcrip-
tion factor that coordinates iron homeostasis in many bacteria.
Recently, we reported that expression of the Escherichia coli Fur
regulon is also impacted by O2 tension. Here, we show that for
most of the Fur regulon, Fur binding and transcriptional repression
increase under anaerobic conditions, suggesting that Fur is con-
trolled by O2 availability. We found that the intracellular, labile
Fe2+ pool was higher under anaerobic conditions compared with
aerobic conditions, suggesting that higher Fe2+ availability drove
the formation of more Fe2+-Fur and, accordingly, more DNA bind-
ing. O2 regulation of Fur activity required the anaerobically in-
duced FeoABC Fe2+ uptake system, linking increased Fur activity
to ferrous import under iron-sufficient conditions. The increased
activity of Fur under anaerobic conditions led to a decrease in ex-
pression of ferric import systems. However, the combined positive
regulation of the feoABC operon by ArcA and FNR partially antag-
onized Fur-mediated repression of feoABC under anaerobic condi-
tions, allowing ferrous transport to increase even though Fur is
more active. This design feature promotes a switch from ferric im-
port to the more physiological relevant ferrous iron under anaerobic
conditions. Taken together, we propose that the influence of O2
availability on the levels of active Fur adds a previously undescribed
layer of regulation in maintaining cellular iron homeostasis.
ferric uptake regulator | iron homeostasis | anaerobiosis | labile iron pool |
protein metallation
I ron homeostasis requires coordination between iron uptake,
storage, and cofactor synthesis to supply iron for the iron-
containing proteome (1–3). In bacteria, this coordination is or-
chestrated by transcription factors that are predicted to provide a
direct readout of the cellular iron status (4). In particular, the
transcription factor ferric uptake regulator (Fur) plays a leading
role in many bacteria to maintain iron homeostasis by directly
sensing levels of ferrous iron (Fe2+) and altering gene transcrip-
tion accordingly (3, 5). Recently, we reported that transcriptional
repression of many genes of the Escherichia coli K12 Fur regulon
is enhanced under anaerobic growth conditions (6), suggesting
that O2 provides an additional input into Fur regulation.
The regulation of Fur activity by iron has provided the para-
digm for understanding how bacteria maintain appropriate iron
levels in vivo. When iron is available, each subunit of a Fur dimer
binds one molecule of Fe2+ ion (7, 8). In this Fe2+-bound form,
Fur exhibits increased affinity for specific DNA sequences [re-
ferred to a Fur boxes (9, 10)] and represses transcription of many
iron assimilation pathways. In contrast, when iron is limiting, Fur
shifts to an iron-free form, thereby inactivating Fur function. The
low micromolar binding affinity of Fe2+ for Fur (11–13) affords a
window into the physiologically relevant iron concentration that
Fur responds to in cells. Indeed, many studies suggest that under
aerobic growth conditions, the intracellular iron pool available for
Fur binding is in the low micromolar range (14, 15). However, our
recent finding that Fur-mediated repression is enhanced under
anaerobic conditions (6) raises the question of whether the in-
tracellular iron pool is altered during anaerobiosis.
www.pnas.org/cgi/doi/10.1073/pnas.1707189114
O2 tension influences the oxidation state of iron, and thereby
the means by which it is acquired by cells (1, 3, 16, 17). When O2
is absent, iron is primarily in its soluble, reduced Fe2+ form and
can be taken up directly by bacteria via dedicated pathways (e.g.,
the FeoABC system). In contrast, when O2 is present, iron is
oxidized to an insoluble ferric (Fe3+) state. The uptake of Fe3+
involves the synthesis and secretion of Fe3+-chelating molecules
called siderophores (e.g., enterobactin). Transport of the Fe3+-
bound siderophore into bacterial cells involves specific outer
membrane receptors, the energy-transducing TonB–ExbB–ExbD
complex, and ABC-type transporters that ultimately deliver the
Fe3+-siderophore to the cytoplasm, where the iron is reduced
and released. Given the impact of O2 on the oxidation state of
iron and the fact that Fur controls expression of both ferrous and
ferric iron assimilation pathways, an O2-dependent input into the
expression of the Fur regulon would make physiological sense.
In this study, we show that such an O2-dependent mechanism
exists by probing how O2 impacts Fur activity. We investigated
the correlation between increased anaerobic Fur DNA binding
across the genome and increased transcriptional repression
mediated by Fur. To dissect the mechanisms by which the
amount of active Fur is elevated during anaerobiosis, we mea-
sured the levels of intracellular chelatable Fe2+ and Fur protein
in aerobic and anaerobic cells, and we assayed Fur activity in
strains lacking proteins involved in iron uptake or storage. Fi-
nally, we measured the response of Fur to O2 following a shift
from anaerobic to aerobic conditions and addressed the influ-
ence of other O2-responsive transcription factors on expression
of genes in the direct Fur regulon. Together, our findings provide
insight on the means by which O2-dependent regulation of Fur
Significance
Our understanding of how cells regulate intracellular iron
pools has been largely shaped by studying cells grown under
aerobic conditions, in which the barrier to iron acquisition is
MICROBIOLOGY
dominated by O2-dependent insolubility. However, less is known
about how bacteria meet their iron demands in the O2-limiting or
anaerobic environments that are common to many ecosystems
and reflective of the ancient atmosphere of early Earth. Because
the transcription factor ferric-uptake regulator (Fur) plays a central
role in controlling iron homeostasis in many bacterial species, we
use Fur activity as a surrogate for understanding how anaerobi-
osis alters iron homeostasis. Our finding that levels of active Fur
are increased during anaerobiosis emphasizes fundamental dif-
ferences in bacterial iron requirements between aerobic and an-
aerobic growth conditions.
Author contributions: N.A.B. and P.J.K. designed research; N.A.B., E.L.M., L.J.M., and
E.R.W. performed research; S.K. contributed new reagents/analytic tools; N.A.B., E.L.M.,
and P.J.K. analyzed data; and N.A.B., E.L.M., and P.J.K. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
Published under the PNAS license.
1
To whom correspondence should be addressed. Email: pjkiley@wisc.edu.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
1073/pnas.1707189114/-/DCSupplemental.
PNAS | November 14, 2017 | vol. 114 | no. 46 | 12261–12266
occurs, which involves the up-regulation of Fe2+ uptake under
anaerobic conditions.
Results
Fur DNA Occupancy Increases During Anaerobiosis. A previous genome-
wide study of Fur–DNA interactions using ChIP-sequencing
(ChIP-seq) revealed that under anaerobic conditions, Fur bound
more regions in the E. coli K12 genome than under aerobic growth
conditions (6). This finding led us to hypothesize that during an-
aerobiosis, an increase in the amount of active Fe2+-Fur could
explain Fur binding to DNA sites that are more varied from the
consensus Fur binding motif (6) than those occupied by Fur in
both the presence and absence of O2. Here, we build upon these
results to address whether the O2-dependent changes in expres-
sion observed for the Fur regulon (6) can be attributed to changes
in Fur DNA occupancy genome-wide.
By quantifying the ratio of Fur enrichment for 74 of the
highest intensity iron-dependent ChIP-seq peaks (6) using the
statistics-based differential binding algorithm DBChIP (18), we
found that, overall, the ratio of anaerobic to aerobic Fur ChIP-
seq signal increased, ranging from ∼1.5- to 10-fold, with a me-
dian of threefold (Tables S1 and S2). The increase in the ratio of
anaerobic to aerobic Fur ChIP-seq signal at each site suggested
an increase in Fe2+-Fur binding at these sites.
The anaerobic increase in genome-scale Fur DNA binding
appears to be transcriptionally relevant since, for most of the Fur
regulon, the ratio of anaerobic to aerobic Chip-seq signal cor-
related positively with the fold increase in Fur-mediated re-
pression calculated from anaerobic and aerobic transcriptomic
data, respectively (Table S1). For example, Fur binding to the
promoter regions of exbB and tonB (both involved in Fe3+-
siderophore transport) and bfd (involved in iron storage/release
from bacterioferritin) increased 6.2-, 5.2-, and 3.9-fold under
anaerobic conditions, respectively (Fig. 1 and Table S1). This
corresponded to a 6.0-, 3.3-, and 7.0-fold respective increase in
Fur-mediated repression of these genes under anaerobic condi-
tions compared with aerobic conditions (Table S1). In instances
where the ChIP-seq ratio measured for Fur occupancy in an-
aerobic and aerobic cells was closer to 1, smaller O2-dependent
changes in gene expression were detected (e.g., entC and fhuE,
involved in Fe3+-siderophore transport; ryhB, encodes a regula-
tory small RNA) (Table S1). This latter group of promoters
represents some of the more strongly repressed Fur genes (6,
19), and the lower ratio likely stems from nearly comparable Fur
binding under both aerobic and anaerobic conditions. In sum-
mary, these data indicate that under anaerobic conditions, Fur
binding is increased at many promoters, which can lead to
comparable changes in Fur-mediated gene repression.
Fig. 1. Occupancy of Fur binding regions increases during anaerobiosis. Fur
ChIP-seq peak regions for exbB, tonB, and bfd from cells grown under an-
aerobic (blue) or aerobic (red) conditions in MOPS minimal glucose media
containing 10 μM FeSO4, were plotted from data reported by Beauchene
et al. (6). Shown are three replicates for each growth condition. The x axis
indicates the genomic position of the ChIP-seq peak, and the y axis indicates
the read count after each dataset was normalized to 20 million reads. The
ratio of the anaerobic ChIP signal to aerobic ChIP signal for all Fur ChIP-seq
peaks is presented in Tables S1 and S2.
12262 | www.pnas.org/cgi/doi/10.1073/pnas.1707189114
A Fur Binding Site Is Sufficient to Confer Regulation by O2. To test if
the presence of a Fur DNA binding site is sufficient to confer O2-
dependent regulation as suggested by our genome-wide analysis,
we engineered a synthetic Fur-dependent promoter (PfepBS), in
which the tac promoter (Ptac) contains the Fur binding site from
PfepB (which drives expression of FepB, an enterobactin binding
protein). This binding site is predicted to bind two Fur dimers
(20, 21) and was positioned in Ptac to maintain the same spacing
(−19 to +21 bp) with respect to the promoter −35 hexamer from
PfepB. In a strain containing PfepBS driving expression of a chro-
mosomal lacZ reporter gene, β-galactosidase activity was signif-
icantly decreased when Fur was present (Fig. S1), indicating that
Fur mediates repression of PfepBS in vivo. Furthermore, Fur-
dependent repression was increased threefold under anaerobic
growth conditions compared with aerobic growth conditions
(Fig. 2), demonstrating a direct effect of O2 availability on Fur-
dependent regulation of PfepBS. Taken together, the enhance-
ment in Fur occupancy at genomic sites, along with the increase
in Fur repression under anaerobic conditions, provides evidence
for a direct role of Fur in the global response to anaerobiosis.
Additional Transcription Factors Can Tailor O2-Mediated Expression of
Some Members of the Fur Regulon. For a few members of the Fur
regulon, we hypothesized from previous results that the impact
of O2 on their transcription might also require the O2-regulated
transcription factors FNR and ArcA (22, 23). Therefore, we
measured expression of lacZ fusions to PexbB, Pfiu, PcirA, or PfeoA
in strains lacking one or more of these transcription factors. In
the absence of any mutations, expression of PexbB, Pfiu, and PcirA
decreased under anaerobic conditions (Fig. 3 A–C), in agree-
ment with transcriptomic data (6). For PexbB, we found that de-
letion of fur increased expression to comparable levels under
both aerobic and anaerobic conditions, indicating Fur-mediated
repression was sufficient to explain its decreased expression
under anaerobic conditions (Fig. 3A). Deletion of fnr had no
effect on PexbB expression, despite a reported FNR binding site
(22). In contrast, for Pfiu and PcirA, we found that decreased
expression of these promoters under anaerobic conditions was
only partially dependent on Fur (Fig. 3 B and C). For both
promoters, we found that ArcA also contributes to anaerobic
repression (Fig. 3 B and C), in agreement with ArcA binding
these promoter regions in vivo (23). This additional role of ArcA
likely explains why the increase in Fur binding to Pfiu and PcirA
anaerobically did not correlate well with the fold increase in Fur-
mediated repression (Table S1).
Similar analysis of the promoter driving transcription of feoABC,
encoding a ferrous uptake system, revealed several distinctive fea-
tures (Fig. 3D). First, in contrast to the ferric uptake systems, PfeoA
is poorly expressed aerobically even in the absence of Fur. Second,
unlike the ferric uptake systems (e.g., cirA, exbB, fiu), expression of
PfeoA increased a small amount under anaerobic conditions com-
pared with aerobic conditions, despite increased anaerobic Fe2+-
Fur levels. Third, we found that this increased anaerobic expression
of PfeoA required both of the anaerobic transcription factors FNR
and ArcA, as suggested by genome-wide and other data (22–
24). This positive effect of both ArcA and FNR appears to limit
the negative effect of Fur under anaerobic conditions. Thus,
FNR and ArcA produce an expression pattern for PfeoA under
anaerobic conditions opposite to the expression pattern of the
ferric transport systems. Taken together, these data support the
notion that Fur plays a central role in reprogramming gene ex-
pression profiles in response to O2 because of changes in levels of
Fe2+-Fur available for DNA binding and by Fur acting in concert
with other transcription factors.
The Labile Fe2+ Pool Increases During Anaerobiosis. To determine
the basis for increased levels of Fe2+-Fur under anaerobic con-
ditions, we measured the levels of both Fur protein and iron.
Western blot analysis revealed that Fur protein levels are com-
parable under aerobic (22 μM) and anaerobic (17 μM) growth
conditions (Fig. S2). To test if increased intracellular Fe2+
Beauchene et al.

Fig. 2. Fur-dependent repression of a synthetic promoter containing a Fur
DNA binding site is influenced by O2 and the FeoABC ferrous uptake sys-
tem. Fur-mediated repression of a chromosomal Ptac-lacZ fusion bearing the
E. coli fepB Fur binding site (PfepBS) is shown in wild-type, Δfur, ΔfeoB, ΔfecC,
ΔentA, and ΔftnA Δbfr Δdps strains grown under aerobic (gray bars) or
anaerobic (white bars) conditions in MOPS minimal glucose media containing
1.0 μM FeSO4. Fold repression was calculated by dividing the average promoter
activity in the Δfur strain by the average promoter activity in each of the other
strains tested (at least two replicates of each; promoter activity was determined
using β-galactosidase assays, as shown in Fig. S1), and error bars represent the
propagation of errors calculated according to the method of Ku (45).
availability could explain the increase in active Fe2+-Fur protein
levels under anaerobic conditions, we used chelator-assisted elec-
tron paramagnetic resonance (EPR) spectroscopy (25) to measure
the intracellular chelatable Fe2+ pool. This labile pool is estimated
to be 1.0% of cellular iron (14) and is defined as iron not tightly
bound to cellular proteins, and possibly associated with metabolites,
such as glutathione, citrate, or phosphorylated sugar compounds
(25–27). We found that the amount of chelatable Fe2+ was ap-
proximately sevenfold higher in anaerobic cells compared with
aerobic cells (Fig. 4 A and B). When total cellular iron levels (which
encompass both Fe2+ and Fe3+ in proteins and the labile iron pool)
were measured by inductively coupled plasma mass spectrometry,
there was only a slight increase in the total iron content of anaer-
obically grown cells (Fig. 4C). These results indicate that while total
cellular iron in aerobically and anaerobically grown cells is com-
parable, the labile iron pool (that which is accessible for Fur
binding) is higher in the absence of O2.
To provide further support for an elevated labile Fe2+ pool in
anaerobic cells, we assayed the sensitivity of cells to hydrogen
peroxide (H2O2). H2O2 is reduced by Fe2+, resulting in the
composition (rich versus minimal) influences Fur-mediated re-
pression under aerobic or anaerobic conditions. Fur-mediated
repression was assayed using the synthetic PfebBS-lacZ fusion.
We did not find any changes in O2-dependent regulation of Fur
activity when the media were varied from our standard 3-(N-
morpholino)propanesulfonic acid (MOPS) minimal media (Fig.
S3). We also considered the possibility that decreased repression
by Fur under aerobic conditions could be explained if iron up-
take was less efficient due to the decreased solubility of iron in
the presence of O2. However, increasing external iron levels 10-
fold resulted in no change in the levels of Fur-mediated re-
pression under aerobic conditions, suggesting that iron avail-
ability was not limiting (Fig. S3). In contrast, under anaerobic
conditions, repression by Fur was increased even further, sug-
gesting that iron transport could be contributing to Fur regula-
tion under anaerobic conditions.
Since the FeoABC system has a major role in anaerobic Fe2+
uptake (17, 33, 34), we investigated whether anaerobic regulation
of Fur was disrupted in a strain lacking feoB. Indeed, we found
that elimination of the FeoB iron transporter caused a threefold
defect in PfepBS repression by Fur under anaerobic conditions (Fig.
2). Furthermore, O2-dependent regulation of Fur required FeoB
since Fur repressed PfepBS to a similar extent under aerobic and
anaerobic conditions in the ΔfeoB mutant (Fig. 2). The effect of
feoB on Fur activity did not appear to reflect a general defect in
iron-containing transcription factors since we did not find any
change in activity of two Fe-S cluster-containing transcription
factors (FNR and IscR) when similarly assayed under anaerobic
conditions (Fig. S4). These data suggest that FeoABC contributes
to the O2-dependent regulation of Fur by increasing the anaerobic
labile iron pool available for Fur binding.

MICROBIOLOGY
formation of hydroxyl radicals via the Fenton reaction (28).
Hydroxyl radicals are highly detrimental to the integrity of cel-
lular components, such as DNA (29, 30), and can lead to cell
death. We determined the sensitivity of aerobic and anaerobic
cells to H2O2 using a strain lacking RecA, which is defective in
DNA repair and thus more sensitive to H2O2 compared with a
wild-type strain (31). Anaerobically grown cells showed 10-fold
more sensitivity to killing by H2O2 than aerobically grown cells
(Fig. 4D), suggesting that Fe2+ is more readily available in an-
aerobic cells to promote hydroxyl radicals and DNA damage.
These results agree with previous findings that anaerobic cells
are highly sensitive to H2O2 (31, 32).
Taken together, these data show that increased levels of Fe2+-
Fur under anaerobic conditions are not due to elevated Fur
protein levels, but rather to an increase in the labile Fe2+ pool.
Consequently, these higher Fe2+ levels would allow for increased Fig. 3. Fur and other O2-regulated transcription factors tailor the expres-
formation of Fe2+-Fur, and thus explain increased Fur DNA sion of iron-uptake systems. Strains bearing lacZ fusions to PexbB (A), Pfiu (B),
binding and transcriptional repression during anaerobiosis. PcirA (C), and PfeoA (D) were assayed for β-galactosidase activity under aerobic
(gray) or anaerobic (white) growth conditions. Deletion of the relevant
Regulation of the Labile Iron Pool by O2. To determine how the transcription factor in the reporter strain is indicated by (−). Cultures were
labile iron pool is increased under anaerobic conditions, we grown in MOPS minimal glucose media containing 1.0 μM FeSO4. Error bars
asked whether the source of iron (Fe2+ or Fe3+) or media represent the SE from triplicate experiments.

Beauchene et al. PNAS | November 14, 2017 | vol. 114 | no. 46 | 12263

Fig. 4. Intracellular labile Fe2+ levels increase during anaerobiosis. EPR spectra
showing the signal at g = 4.3, normalized for OD600 from aerobic or anaerobic
cells grown in MOPS minimal glucose media containing 1.0 μM FeSO4. Aerobic
(red) or anaerobic (blue) cells were treated with desferoxamine mesylate (DFO)
or were left untreated (black). (A) Four to five replicates of each experiment
are shown. (B) Chelatable Fe levels in aerobic (gray) or anaerobic cells (white)
calculated from EPR spectra. (C) Inductively coupled plasma mass spectrometry was
used to measure whole-cell Fe levels in three biological replicates of aerobic (gray)
or anaerobic cells (white), which were grown in MOPS minimal glucose media
containing 1.0 μM FeSO4 and normalized for cell pellet (milligrams). (D) Viable cell
counts (assayed in triplicate) for aerobic (gray) or anaerobic cells (white) grown in
MOPS minimal glucose media containing 1.0 μM FeSO4, with and without treat-
ment of 2.5 mM H2O2 for 15 min. For B, C, and D, error bars represent the SE.
We also tested whether other iron transport systems affected
Fur activity (3). Enterobactin-mediated Fe3+ uptake was disrupted
by deletion of entA. The inability to synthesize the enterobactin
siderophore caused a twofold loss in the ability of Fur to repress
PfepBS under both aerobic and anaerobic conditions (Fig. 2).
However, the EntA− mutant retained O2-dependent regulation of
Fur, presumably because of the continued function and regulated
expression of FeoABC (Fig. 2). As a control, we showed that Fur-
mediated repression was not affected in a strain lacking fecC, in-
volved in transport of ferric citrate, which was not added to our
media (Fig. 2).
Finally, we tested whether iron storage proteins affected the
amount of iron available for Fur binding by measuring Fur-
mediated repression in iron storage-deficient strains. The only
known mechanisms of iron storage in E. coli require O2 or H2O2
(35), suggesting that less iron might be stored anaerobically, and
potentially contributing to the increase in the labile iron pool.
However, elimination of iron storage proteins by deletion of
ftnA, bfr, and dps did not alter the level of Fur-mediated re-
pression under either aerobic or anaerobic conditions (Fig. 2).
changes in environmental O2, conditions in which E. coli would
experience in its natural habitat.
Discussion
Our data provide insight into how O2 availability alters iron
homeostasis in the facultative bacterium E. coli. Overall, we
propose that an increase in the intracellular labile Fe2+ pool,
mediated by the FeoABC ferrous transport system, leads to an
increase in Fur-mediated repression under anaerobic conditions.
The rise in anaerobic Fe2+ availability increases the population
of Fur protein that is bound to Fe2+ and, accordingly, drives Fur
DNA binding and transcriptional repression. The global change
in expression of the Fur regulon under iron-sufficient anaerobic
conditions promotes a shift in expression from Fe3+ to Fe2+
transport systems, allowing E. coli to use the most physiologically
relevant form of iron.
The Labile Iron Pool Increases Under Anaerobic Conditions. It is
generally considered that the affinity of a metal sensor for its
metal, and the subsequent homeostatic control the regulator
exerts on its regulon, defines intracellular free metal concen-
trations (37–39). Thus, before this work, one would expect a
priori that the intracellular free (labile) Fe2+ pool would be
comparable in aerobically and anaerobically grown cells. Despite
this expectation, we found that intracellular chelatable iron lev-
els increased nearly sevenfold (∼26 μM in aerobic cells and
∼177 μM in anaerobic cells) under anaerobic conditions. Pre-
vious studies have shown that the FeoABC transporter, which is
widely distributed in bacteria (17), contributes to anaerobic iron
uptake (33, 34). Therefore, our finding that FeoABC was re-
quired for increased Fur activity under anaerobic conditions
provided a link between increased iron import under anaerobic
conditions and the anaerobic labile Fe2+ pool. The ability of
FeoABC to mediate an increase in the iron pool under anaerobic
conditions appears to be due, in part, to its unique mechanism of
transcriptional regulation (discussed below).
Fur Fe2+ Occupancy Varies with O2 Tension. Since the affinity of Fur
for Fe2+ is estimated at 1.2–55 μM (11–13), and the cellular Fur

Fur Activity Responds to a Shift in O2 Availability. Adapting to changes in

O2 tension is key for most organisms. Given the significant differ-
ences in the bioavailability of iron in the presence and absence of O2,
we wanted to test if the sudden introduction of O2 to anaerobic cells
altered Fur activity. To do this, we assayed β-galactosidase from lacZ
fusions to either our synthetic promoter (PfepBS) or the fepA promoter
(PfepA), which drives expression of the FepA Fe3+-enterobactin
transporter. Upon exposure to either O2 or α,α′-dipyridyl, a known
cell-permeable iron chelator, we observed an increase in PfepBS
and PfepA expression (Fig. 5 and Fig. S5). Although the rate of
increase in expression of these promoters after exposure to O2 was
slower than with α,α′-dipyridyl, de-repression of these promoters
occurred on a similar time scale.
To determine how other genes that are directly repressed by
Fur respond to O2 exposure, we analyzed time series tran-
Fig. 5. Anaerobic Fur activity decreases upon exposure to O2. A strain
scriptomic data from RNA isolated from E. coli at 0.5, 1, 2, 5,
containing PfepBS-lacZ was grown anaerobically and was then either shifted
and 10 min after a shift from anaerobic to aerobic conditions to aerobic growth conditions (●) or treated with an iron chelator (200 μM
[first reported by von Wulffen et al. (36)]. Similar to what we α,α′-dipyridyl; ▲) at time 0. Samples were removed and assayed for
observed for PfepBS and PfepA, expression of these genes showed β-galactosidase at various time points. Untreated cells (♦) are indicated.
an immediate response to the addition of O2, with almost all Cultures were grown in MOPS minimal glucose media containing 1.0 μM
RNA levels increasing by 2 min (Table S3). Thus, these data FeSO4. Error bars represent the SE from three experiments. (Inset) Expanded
indicate that expression of the Fur regulon responds to sudden view of O2- and α,α′-dipyridyl–treated cells from time 0–20 min.

12264 | www.pnas.org/cgi/doi/10.1073/pnas.1707189114 Beauchene et al.

concentration is comparable between aerobic (∼22 μM) and
anaerobic (∼17 μM) cells, our data suggest that the fraction of
Fur bound by Fe2+ is impacted by O2-dependent changes in Fe2+
levels. We propose that under aerobic conditions even when iron
levels are sufficient, Fur is not completely occupied with Fe2+,
whereas under anaerobic conditions, Fur-Fe2+ occupancy increases
due to the increase in the labile iron pool (Fig. 6). It was also
surprising that the decrease in iron available for Fur binding ob-
served with the ΔfeoB mutant, did not impact gene regulation
mediated by the transcription factors FNR and IscR that ligate Fe-S
clusters. This result could imply that the affinity of iron for Fe-S
cluster biogenesis machinery is much stronger than the binding
affinity to Fur, creating a hierarchy of iron usage in the cell.
The Strength of Fur DNA Recognition Sites May also Contribute to O2
Regulation of the Fur Regulon. The effect of O2 on Fur-mediated
repression varied for individual members of the Fur regulon.
Some genes exhibited only small differences in Fur-mediated
regulation between aerobic and anaerobic conditions, whereas
others showed a much stronger dependence on anaerobic con-
ditions to observe regulation by Fur. To explain these findings,
we propose that the increase in Fe2+-Fur levels under anaerobic
conditions allows binding to Fur sites of a wider range of affin-
ities. In contrast, under aerobic conditions, the reduced levels of
Fe2+-Fur bias binding to stronger affinity Fur sites. In support of
this model, bioinformatic analysis of sites bound by Fur only
under anaerobic conditions was much further from consensus
than that of sites bound by Fur both in the presence and absence
of O2 (6). Although, there is no resource to compare Fur af-
finities obtained under similar solution conditions for individual
sites across the regulon, information theory predictions of Fur
binding site strength suggest that many of the DNA sites, which
bound Fur well both in the presence and absence of O2 (e.g.,
ryhB, entC, cirA, yjjZ, fepA), should be high-affinity sites (21).
Thus, the strength of the Fur DNA binding site likely contributes
to how much of an effect O2 has on an individual regulon
member. For the Fur homolog Zur, the affinity of the transcrip-
tion factor for its DNA sites also correlates to the fold repression
(40). Further, it has been shown that stepwise metallation of the
Zur Zn binding sites fine-tunes the degree of Zur repression of its
target promoters (41, 42).
Physiological Significance of O2 Regulation of Fur. Our data reveal
that the O2-dependent changes in levels of active Fur play a role
in tailoring expression of iron uptake systems to the growth
Fig. 6. O2-dependent regulation of the Fur regulon. How O2 influences the
Fe2+ occupancy of Fur under aerobic and anaerobic growth conditions is
depicted. In the absence of O2, increased intracellular labile Fe2+ levels,
which are mediated by the FeoABC system, promote Fur Fe2+ occupancy, and
thus lead to increased repression of Fur-dependent promoters. In the pres-
ence of O2, cells shift to using Fe3+ uptake systems and intracellular labile
Fe2+ levels are lower compared with those in anaerobic cells. As a result,
levels of Fe2+-Fur are decreased, leading to partial occupancy and less re-
pression of promoters containing less conserved Fur binding sites. However,
promoters containing more conserved Fur binding sites remain repressed.
Beauchene et al.
conditions in which their substrate (Fe2+ or Fe3+) is most likely
to be found. In most cases, Fur-mediated repression was suffi-
cient to explain decreased expression of Fe3+ uptake systems
under anaerobic conditions. However, for a few Fe3+ uptake
systems (e.g., fiu, cirA), decreased expression under anaerobic
conditions required repression by both ArcA and Fur. In con-
trast, anaerobic expression of the FeoABC Fe2+ uptake system
is under positive control of ArcA and FNR. This regulatory
mechanism prevents additional Fur-mediated repression of PfeoA
(unlike that of the ferric uptake systems) and explains how ex-
pression increases under anaerobic conditions even when Fur is
more active. By wiring feoABC expression in this manner, Fe2+
can enter cells through a transport system whose expression is
not strictly controlled by Fur under anaerobic growth conditions.
In turn, anaerobic up-regulation of feoABC facilitates the in-
crease in the iron pool and recalibrates iron homeostasis.
It might be counterintuitive that even small changes in Fur-
mediated repression between aerobic and anaerobic conditions
could be physiologically significant. However, it is important to
note that under iron-sufficient conditions, the low expression
levels of Fur-repressed iron uptake systems are sufficient to drive
iron uptake. Iron uptake systems only become further induced
when external iron becomes scarce. Thus, it is logical to assume
that even small changes in levels of these transport systems will
affect iron uptake rates.
In summary, our data show that O2 has a direct effect on levels
of active Fur in vivo and that several factors contribute to the
expression profile of the Fur regulon in response to O2. The
expression output of each gene is a summation of Fur activity,
which depends on the extent of Fur Fe2+ occupancy, the Fur
DNA binding sequence, and coordinated regulation by other
transcription factors. Our data highlight the complexity of co-
ordinating a central cellular activity like iron homeostasis to
multiple environmental cues. This fine-tuning of transcriptional
regulation ensures that the necessary gene products are available
for cellular function. Finally, we predict that the differential ef-
fect of O2 on iron homeostasis observed in this study is likely to
be conserved in other Fur-containing facultative bacteria since it
reflects an adaptation to the intrinsic sensitivity of Fe2+ to oxi-
dation in aerobic environments.
Methods
Detailed materials and methods can be found in SI Methods.
Differential Fur DNA Binding Analyses. Differential binding of Fur between
aerobic and anaerobic conditions was assessed using the algorithm DBChIP
(18). This analysis was performed on 74 previously reported iron-dependent
Fur ChIP-seq peaks from anaerobically grown E. coli with read counts >4,000 at
the peak summit (6). Read counts of each dataset were normalized to 20 mil-
MICROBIOLOGY
lion reads. The fold change between aerobic and anaerobic Fur DNA occu-
pancy, reported here as the ChIP ratio, was calculated by dividing the average
ChIP signal from three anaerobic replicates by that of three aerobic replicates.
Fur was annotated as differentially bound if the adjusted P value of the fold
change was <0.05.
Growth Conditions. For the β-galactosidase assays (Figs. 2, 3, and 5 and Figs.
S1 and S3–S5) and H2O2 sensitivity assays (Fig. 4D), strains were grown at
37 °C in MOPS minimal 0.2% glucose media (43) containing either 1.0 μM or
10 μM FeSO4 under aerobic conditions (by shaking) or anaerobic conditions
(using filled screw-capped tubes) as described previously (44). In Fig. S3,
different media and iron sources are indicated in the figure legend.
β-galactosidase assays were performed as described (6).
For the in vivo O2 shift experiments (Fig. 5 and Fig. S5), strains bearing
PfepA-lacZ or PfepBS-lacZ were grown at 37 °C in MOPS minimal 0.2% glucose
media containing 1.0 μM FeSO4 by sparging with an anaerobic gas mix of 95%
N2 and 5% CO2. At time 0 min, which corresponded to an OD600 of 0.1, cultures
were either shifted to aerobic conditions by switching the gas mix to 70% N2,
5% CO2, and 25% O2, or were treated with α,α′-dipyridyl (final concentration of
200 μM). Samples were removed at various time points for measurement of
OD600 and β-galactosidase activity as described by Beauchene et al. (6).
For EPR spectroscopy and Western blot analysis, cultures were grown at
37 °C in MOPS minimal glucose media containing 1.0 μM FeSO4 to an OD600
PNAS | November 14, 2017 | vol. 114 | no. 46 | 12265
of ∼0.4 (Lambda 25 UV/Vis Spectrophotometer; PerkinElmer) by sparging
with either an aerobic (70% N2, 5% CO2, 25% O2) or anaerobic (95% N2, 5%
CO2) gas mix (44).
Chelator-Assisted EPR Spectroscopy. A MG1655 culture volume of 450 mL
(OD600 of ∼0.4) was pelleted, resuspended in 10 mL of fresh media con-
taining either 20 mM membrane-permeable iron chelator desferoxamine
mesylate (DFO; Sigma Aldrich) and 10 mM membrane-impermeable iron
chelator diethylenetriaminepentaacetic acid (DTPA; Sigma Aldrich) or just
10 mM DTPA (to remove contaminating iron) as described previously (25).
After a 15-min incubation at 37 °C, cells were washed twice with 5 mL of
20 mM Tris·HCl (pH 7.4) and resuspended in 350 μL of 20 mM Tris·HCl (pH 7.4)
and 30% glycerol. A portion of these cells (200 μL) was transferred to EPR
tubes (4-mm Thin Wall Quartz EPR sample tube with a length of 250 mm;
Wilmad Lab Glass) and frozen at −80 °C. The remaining cells were diluted to
measure the final OD600 (Lambda 25 UV/Vis Spectrophotometer; PerkinElmer).
All manipulations were completed in a Coy anaerobic chamber. Four to five
replicates of each sample were analyzed.
1. Braun V, Hantke K (2011) Recent insights into iron import by bacteria. Curr Opin
Chem Biol 15:328–334.
2. Frawley ER, et al. (2013) Iron and citrate export by a major facilitator superfamily
pump regulates metabolism and stress resistance in Salmonella typhimurium. Proc
Natl Acad Sci USA 110:12054–12059.
3. Andrews S, et al. (2013) Control of iron metabolism in bacteria. Met Ions Life Sci 12:
203–239.
4. Fleischhacker AS, Kiley PJ (2011) Iron-containing transcription factors and their roles
as sensors. Curr Opin Chem Biol 15:335–341.
5. McHugh JP, et al. (2003) Global iron-dependent gene regulation in Escherichia coli. A
new mechanism for iron homeostasis. J Biol Chem 278:29478–29486.
6. Beauchene NA, et al. (2015) Impact of anaerobiosis on expression of the iron-
responsive Fur and RyhB regulons. MBio 6:e01947–e15.
7. Lee JW, Helmann JD (2007) Functional specialization within the Fur family of met-
alloregulators. Biometals 20:485–499.
8. Fillat MF (2014) The FUR (ferric uptake regulator) superfamily: Diversity and versatility
of key transcriptional regulators. Arch Biochem Biophys 546:41–52.
9. Baichoo N, Helmann JD (2002) Recognition of DNA by Fur: A reinterpretation of the
Fur box consensus sequence. J Bacteriol 184:5826–5832.
10. Lavrrar JL, McIntosh MA (2003) Architecture of a fur binding site: A comparative
analysis. J Bacteriol 185:2194–2202.
11. Bagg A, Neilands JB (1987) Ferric uptake regulation protein acts as a repressor, em-
ploying iron (II) as a cofactor to bind the operator of an iron transport operon in
Escherichia coli. Biochemistry 26:5471–5477.
12. Hamed MY (1993) Binding of the ferric uptake regulation repressor protein (Fur) to
Mn(II), Fe(II), Co(II), and Cu(II) ions as co-repressors: Electronic absorption, equilibrium,
and 57Fe Mössbauer studies. J Inorg Biochem 50:193–210.
13. Mills SA, Marletta MA (2005) Metal binding characteristics and role of iron oxidation
in the ferric uptake regulator from Escherichia coli. Biochemistry 44:13553–13559.
14. Keyer K, Imlay JA (1996) Superoxide accelerates DNA damage by elevating free-iron
levels. Proc Natl Acad Sci USA 93:13635–13640.
15. Jacques JF, et al. (2006) RyhB small RNA modulates the free intracellular iron pool and
is essential for normal growth during iron limitation in Escherichia coli. Mol Microbiol
62:1181–1190.
16. Kosman DJ (2013) Iron metabolism in aerobes: Managing ferric iron hydrolysis and
ferrous iron autoxidation. Coord Chem Rev 257:210–217.
17. Lau CK, Krewulak KD, Vogel HJ (2016) Bacterial ferrous iron transport: The Feo sys-
tem. FEMS Microbiol Rev 40:273–298.
18. Liang K, Keles S (2012) Detecting differential binding of transcription factors with
ChIP-seq. Bioinformatics 28:121–122.
19. Seo SW, et al. (2014) Deciphering Fur transcriptional regulatory network highlights its
complex role beyond iron metabolism in Escherichia coli. Nat Commun 5:4910.
20. Brickman TJ, Ozenberger BA, McIntosh MA (1990) Regulation of divergent tran-
scription from the iron-responsive fepB-entC promoter-operator regions in Escherichia
coli. J Mol Biol 212:669–682.
21. Chen Z, et al. (2007) Discovery of Fur binding site clusters in Escherichia coli by in-
formation theory models. Nucleic Acids Res 35:6762–6777.
22. Myers KS, et al. (2013) Genome-scale analysis of Escherichia coli FNR reveals complex
features of transcription factor binding. PLoS Genet 9:e1003565.
23. Park DM, Akhtar MS, Ansari AZ, Landick R, Kiley PJ (2013) The bacterial response
regulator ArcA uses a diverse binding site architecture to regulate carbon oxidation
globally. PLoS Genet 9:e1003839.
24. Carpenter C, Payne SM (2014) Regulation of iron transport systems in Enter-
obacteriaceae in response to oxygen and iron availability. J Inorg Biochem 133:
110–117.
25. Woodmansee AN, Imlay JA (2002) Quantitation of intracellular free iron by electron
paramagnetic resonance spectroscopy. Methods Enzymol 349:3–9.
26. Hider RC, Kong X (2013) Iron speciation in the cytosol: An overview. Dalton Trans 42:
3220–3229.
12266 | www.pnas.org/cgi/doi/10.1073/pnas.1707189114
H2O2 Sensitivity Assays. Cells were treated with 2.5 mM H2O2 (equilibrated in
an anaerobic chamber for anaerobic cultures) for 15 min at 37 °C (31) and
then immediately diluted 1,000-fold into fresh MOPS glucose minimal media
containing 1.0 μM FeSO4. Treated and nontreated cultures were plated on
TYE agar plates to determine the number of viable cells. Anaerobic condi-
tions were maintained by use of a Coy anaerobic chamber and GasPak
anaerobic jar.
ACKNOWLEDGMENTS. We thank Martin Shafer (Wisconsin State Laboratory
of Hygiene) for assistance with whole-cell element analyses and Monika
Ivancic and Bob Shanks (Department of Chemistry EPR facilities, University of
Wisconsin) for their assistance with EPR instrumentation. We thank Jim Imlay
(University of Illinois) and Thomas Brunold (University of Wisconsin) for
helpful discussions on labile iron measurements. We also thank Dan Park for
the gift of Fur protein, and Kevin Myers for data analysis. This work was
funded by Grants GM045844 and GM115894 from the NIH (to P.J.K.). N.A.B.
was supported by University of Wisconsin–Madison NIH Chemistry Biology
Interface Training Grant T32GM008505.
27. Varghese S, Wu A, Park S, Imlay KR, Imlay JA (2007) Submicromolar hydrogen per-
oxide disrupts the ability of Fur protein to control free-iron levels in Escherichia coli.
Mol Microbiol 64:822–830.
28. Walling C (1975) Fenton’s reagent revisited. Acc Chem Res 8:125–131.
29. Imlay JA, Chin SM, Linn S (1988) Toxic DNA damage by hydrogen peroxide through
the Fenton reaction in vivo and in vitro. Science 240:640–642.
30. Imlay JA (2013) The molecular mechanisms and physiological consequences of oxi-
dative stress: Lessons from a model bacterium. Nat Rev Microbiol 11:443–454.
31. Imlay JA, Linn S (1986) Bimodal pattern of killing of DNA-repair-defective or anoxi-
cally grown Escherichia coli by hydrogen peroxide. J Bacteriol 166:519–527.
32. Keyer K, Gort AS, Imlay JA (1995) Superoxide and the production of oxidative DNA
damage. J Bacteriol 177:6782–6790.
33. Lodge JS, Emery T (1984) Anaerobic iron uptake by Escherichia coli. J Bacteriol 160:
801–804.
34. Kammler M, Schön C, Hantke K (1993) Characterization of the ferrous iron uptake
system of Escherichia coli. J Bacteriol 175:6212–6219.
35. Bradley JM, Moore GR, Le Brun NE (2014) Mechanisms of iron mineralization in fer-
ritins: One size does not fit all. J Biol Inorg Chem 19:775–785.
36. von Wulffen J, Sawodny O, Feuer R; RecogNice-Team (2016) Transition of an anaer-
obic Escherichia coli culture to aerobiosis: Balancing mRNA and protein levels in a
demand-directed dynamic flux balance analysis. PLoS One 11:e0158711.
37. Waldron KJ, Robinson NJ (2009) How do bacterial cells ensure that metalloproteins
get the correct metal? Nat Rev Microbiol 7:25–35.
38. Waldron KJ, Rutherford JC, Ford D, Robinson NJ (2009) Metalloproteins and metal
sensing. Nature 460:823–830.
39. Reyes-Caballero H, Campanello GC, Giedroc DP (2011) Metalloregulatory proteins:
Metal selectivity and allosteric switching. Biophys Chem 156:103–114.
40. Gilston BA, et al. (2014) Structural and mechanistic basis of zinc regulation across the
E. coli Zur regulon. PLoS Biol 12:e1001987.
41. Ma Z, Gabriel SE, Helmann JD (2011) Sequential binding and sensing of Zn(II) by
Bacillus subtilis Zur. Nucleic Acids Res 39:9130–9138.
42. Shin JH, et al. (2011) Graded expression of zinc-responsive genes through two reg-
ulatory zinc-binding sites in Zur. Proc Natl Acad Sci USA 108:5045–5050.
43. Blattner FR, et al. (1997) The complete genome sequence of Escherichia coli K-12.
Science 277:1453–1462.
44. Sutton VR, Kiley PJ (2003) Techniques for studying the oxygen-sensitive transcription
factor FNR from Escherichia coli. Methods Enzymol 370:300–312.
45. Ku HH (1966) Notes on the use of propagation of error formulas. J Res Natl Bur Stand
Sect C 70C:263–273.
46. Kang Y, Weber KD, Qiu Y, Kiley PJ, Blattner FR (2005) Genome-wide expression
analysis indicates that FNR of Escherichia coli K-12 regulates a large number of genes
of unknown function. J Bacteriol 187:1135–1160.
47. Baba T, et al. (2006) Construction of Escherichia coli K-12 in-frame, single-gene
knockout mutants: The Keio collection. Mol Syst Biol 2:2006.0008.
48. Datsenko KA, Wanner BL (2000) One-step inactivation of chromosomal genes in
Escherichia coli K-12 using PCR products. Proc Natl Acad Sci USA 97:6640–6645.
49. Sutton VR, Mettert EL, Beinert H, Kiley PJ (2004) Kinetic analysis of the oxidative
conversion of the [4Fe-4S]2+ cluster of FNR to a [2Fe-2S]2+ cluster. J Bacteriol 186:
8018–8025.
50. Imlay JA, Fridovich I (1991) Assay of metabolic superoxide production in Escherichia
coli. J Biol Chem 266:6957–6965.
51. Schwartz CJ, et al. (2001) IscR, an Fe-S cluster-containing transcription factor, re-
presses expression of Escherichia coli genes encoding Fe-S cluster assembly proteins.
Proc Natl Acad Sci USA 98:14895–14900.
52. Mettert EL, Outten FW, Wanta B, Kiley PJ (2008) The impact of O(2) on the Fe-S cluster
biogenesis requirements of Escherichia coli FNR. J Mol Biol 384:798–811.
53. Giel JL, et al. (2013) Regulation of iron-sulphur cluster homeostasis through tran-
scriptional control of the Isc pathway by [2Fe-2S]-IscR in Escherichia coli. Mol
Microbiol 87:478–492.
54. Keseler IM, et al. (2011) EcoCyc: A comprehensive database of Escherichia coli biology.
Nucleic Acids Res 39:D583–D590.
Beauchene et al.