GORpnas08.pdf

DNA binding shifts the redox potential of the

transcription factor SoxR

Alon A. Gorodetsky

†

, Lars E. P. Dietrich

‡

, Paul E. Lee

†

, Bruce Demple

, Dianne K. Newman

‡¶

, and Jacqueline K. Barton

†

Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, CA 91125; Departments of

‡

Biology and

Earth and

Planetary Sciences, Massachusetts Institute of Technology, Cambridge, MA 02139; and

Department of Genetics and Complex Diseases, Harvard School of

Public Health, Boston, MA 02115

Contributed by Jacqueline K. Barton, January 4, 2008 (sent for review December 7, 2007)

Electrochemistry measurements on DNA-modified electrodes are

used to probe the effects of binding to DNA on the redox potential

of SoxR, a transcription factor that contains a [2Fe-2S] cluster and

is activated through oxidation. A DNA-bound potential of

200 mV

versus NHE (normal hydrogen electrode) is found for SoxR isolated

from

Escherichia coli

and

Pseudomonas aeruginosa

. This potential

value corresponds to a dramatic shift of

490 mV versus values

found in the absence of DNA. Using Redmond red as a covalently

bound redox reporter affixed above the SoxR binding site, we also

see, associated with SoxR binding, an attenuation in the Redmond

red signal compared with that for Redmond red attached below

the SoxR binding site. This observation is consistent with a SoxR-

binding-induced structural distortion in the DNA base stack that

inhibits DNA-mediated charge transport to the Redmond red

probe. The dramatic shift in potential for DNA-bound SoxR com-

pared with the free form is thus reconciled based on a high-energy

conformational change in the SoxR–DNA complex. The substantial

positive shift in potential for DNA-bound SoxR furthermore indi-

cates that, in the reducing intracellular environment, DNA-bound

SoxR is primarily in the reduced form; the activation of DNA-bound

SoxR would then be limited to strong oxidants, making SoxR an

effective sensor for oxidative stress. These results more generally

underscore the importance of using DNA electrochemistry to de-

termine DNA-bound potentials for redox-sensitive transcription

factors because such binding can dramatically affect this key

protein property.

DNA electrochemistry

iron-sulfur proteins

oxidative stress

oxR belongs to the MerR family of transcriptional regulators.

The members of this family are defined by an N-terminal

helix–turn–helix DNA binding motif, a coiled-coil dimerization

region, and a C-terminal sensory domain (1–3). Although the

DNA binding and dimerization regions are conserved among

MerR-type regulators, their sensory domains are diverse (2).

Typically, MerR type transcription factors occupy suboptimally

spaced 19

1-bp promoter elements in the inactivated state,

often inducing a slight bend of the promoter DNA. Upon

activation, these proteins are thought to undergo a conforma-

tional change that unwinds the promoter region, thereby allow-

ing RNA polymerase to initiate transcription (2).

SoxR regulates an oxidative stress response to superoxide in

the enterics

Escherichia coli

and

Salmonella enterica

(4, 5). This

unique transcription factor is a 17-kDa polypeptide that binds

DNA as a dimer and contains a [2Fe-2S] cluster in each

monomer (4). Loss of this cluster does not affect protein folding,

DNA binding, or promoter affinity (6–8), but oxidation of this

cluster by either oxygen or superoxide-generating agents (e.g.,

methyl viologen) triggers expression of the transcription factor

SoxS (8, 9). Subsequently, SoxS controls the expression of

100

genes in the SoxRS regulon that collectively act to repair or avoid

oxidative damage (10).

The role of SoxR appears to vary dramatically across organ-

isms. Whereas SoxR is conserved in both Gram-negative and

Gram-positive bacteria,

soxS

is exclusively found in enterics,

indicating that SoxR can be part of different regulatory networks

(11, 12). Indeed,

Pseudomonas putida

and

Pseudomonas aerugi-

nosa

do not rely on SoxR for an oxidative-stress response (13,

14). Instead,

P. aeruginosa

SoxR responds to phenazines, endog-

enous redox-active pigments, and activates transcription of two

probable efflux pumps and a putative monooxygenase (15) that

might aid in phenazine transport and modification. Considering

that SoxR shows functional diversity between pseudomonads

and enterics, it is surprising that the transcription factor is

biochemically conserved: (

) Expression of

P. putida

SoxR in

coli

can complement a

soxR

deletion mutant (14), and (

) the

redox potentials of soluble SoxR from

E. coli

and

P. aeruginosa

in vitro

are both approximately

290 mV (7, 8, 15).

That SoxR requires oxidation for its transcriptional activity

seems biologically reasonable but also leads to a conundrum.

Under normal physiological conditions, it is assumed that SoxR

is kept in its reduced, inactive state by the intracellular NADPH/

NADP

redox potential of approximately

340 mV versus NHE

(16, 17). Furthermore, it has been reported that NADPH-

dependent SoxR reduction is enzyme-mediated, allowing for a

rapid adjustment to changes in cellular conditions, although

direct enzymatic interaction with SoxR has not yet been dem-

onstrated (18, 19). The conundrum does not lie in the mecha-

nism of SoxR reduction but rather in the specificity of its

oxidation: At a low redox potential of

290 mV versus NHE (7,

8, 15), many cellular oxidants could react with SoxR, in particular

glutathione (20), and therefore SoxR would be primarily in an

oxidized form, even without imposing oxidative stress. Since this

is not the case, how is SoxR maintained in its reduced and

transcriptionally silent form?

The mechanism underlying the oxidation/activation of SoxR is

also not well understood. For

E. coli

SoxR, it was first suggested

that superoxide directly oxidizes the iron-sulfur cluster, but this

has not been established (21, 22). Alternatively, the redox state

of SoxR might be coupled to changes in the equilibrium of

biologically relevant redox couples, such as NADPH or gluta-

thione (16, 23). Recently, we have shown that the activation of

SoxR in

P. aeruginosa

can occur in an oxygen-independent

manner (14). Considering that both

E. coli

and

P. aeruginosa

SoxR can transfer electrons to the mediator safranin O, a

phenazine derivative (7, 8, 15), it seems reasonable that endog-

enous phenazines may oxidize SoxR in pseudomonads. Alter-

natively, given that pseudomonad phenazines can also modulate

the intracellular NADH/NAD

ratio, the possibility that phena-

zines activate SoxR indirectly must also be considered (24).

Author contributions: A.A.G. and L.E.P.D. contributed equally to this work; A.A.G., L.E.P.D.,

P.E.L., B.D., D.K.N., and J.K.B. designed research; A.A.G., L.E.P.D., and P.E.L. performed

research; A.A.G., L.E.P.D., P.E.L., B.D., D.K.N., and J.K.B. analyzed data; and A.A.G., L.E.P.D.,

P.E.L., B.D., D.K.N., and J.K.B. wrote the paper.

The authors declare no conflict of interest.

To whom correspondence may be addressed. E-mail: jkbarton@caltech.edu or dkn@

mit.edu.

3684 –3689

PNAS

March 11, 2008

vol. 105

no. 10

www.pnas.org

cgi

doi

10.1073

pnas.0800093105

One interesting possibility that has been suggested but never

addressed experimentally is the effect of DNA binding on the

redox potential of SoxR. The published redox potentials for

SoxR were measured in the absence of DNA (4, 7, 8). This is

particularly significant because SoxR activates transcription only

in its DNA-bound state, so determining the redox potential of

the DNA-bound form of SoxR becomes critical.

We have previously explored DNA-modified electrodes as

flexible platforms for the study of DNA-mediated charge trans-

port chemistry (25–27). Typically, self-assembled DNA mono-

layers on gold or graphite are interrogated electrochemically

with the efficiency of charge transfer to an electroactive probe

yielding information on the integrity of the intervening base pair

stack. In fact, duplexes that are covalently modified with redox-

active reporters at a fixed position provide particularly well

defined systems for study of DNA charge transport at electrode

surfaces, allowing for the electrochemical detection of even

small perturbations in the intervening base pair stack (28–31).

In addition, DNA-modified electrodes have proven useful for

probing redox centers within proteins bound to DNA (32–34).

We have used DNA monolayers to probe the redox potential of

MutY and Endonuclease III, base excision repair glycosylases

that contain a [4Fe-4S] cluster. Initial studies of these enzymes

had found no clear role for the clusters because, in the absence

of DNA, they did not display redox activity within a physiolog-

ically relevant range of potentials (35–37). We found, however,

that at DNA-modified Au surfaces, these repair enzymes display

reversible, DNA-mediated electrochemistry with redox poten-

tials of

90 mV (33). Moreover, experiments comparing directly

the electrochemistry of Endonuclease III on bare and DNA-

modified graphite demonstrated that binding to DNA shifts the

redox potential of the protein by

200 mV into a physiologically

relevant range, activating the cluster for oxidation (34). DNA

binding thus changes the redox properties of the enzymes from

being similar to ferredoxins to instead resembling high potential

iron proteins. Based on these data, we have proposed a redox

role for the [4Fe-4S] clusters in long-range DNA-mediated

signaling as a first step in detecting damaged sites that are to be

repaired in the genome (32–34, 38). Our ability to alter the redox

states of these proteins in a DNA-mediated manner further

suggests that DNA may be a medium through which oxidation/

reduction reactions occur. This mechanism may also be impor-

tant to consider in the context of SoxR.

Given the sensitivity of DNA-modified electrodes in probing

redox centers of proteins bound to DNA and the precedent that

DNA binding can alter redox potentials of the bound protein,

here we explore the redox properties of the DNA-bound form of

SoxR. Model studies have shown repeatedly the sensitivity of

redox potentials of iron-sulfur clusters to environmental pertur-

bations, which are expected to be significant for SoxR (39). Here,

using self-assembled DNA monolayers on highly oriented pyro-

lytic graphite (HOPG), we address the effect of DNA binding on

the redox potential of both

E. coli

and

P. aeruginosa

SoxR. The

DNA-bound potential provides convincing evidence for the

mechanism the cell uses to maintain SoxR in its reduced form

vivo

Results

Experimental Strategy Used to Probe SoxR Electrochemically.

Fig. 1

illustrates the experimental strategy used to investigate the

electrochemistry of DNA-bound SoxR from

E. coli

and

aeruginosa

PA14. DNA duplexes are prepared by hybridizing

pyrene-modified single-stranded DNA with its complement

(with or without covalently attached Redmond red). The du-

plexes are then self-assembled on HOPG in the absence of Mg

to form a loosely packed DNA monolayer, leaving room for

SoxR to bind (32–34). The surface is backfilled with octane or

decane to prevent direct charge transfer from the surface to the

protein (40, 41). The electrode is subsequently incubated with

protein, and electrochemical experiments are performed before

and after protein addition. The DNA binding sites for SoxR are

18-bp symmetrical sequences that are conserved across species

(15). For

P. aeruginosa

experiments, we chose the SoxR binding

site found upstream of an operon that encodes the efflux pump

MexGHI-OpmD in

P. aeruginosa

PA14.

SoxR Binding Is Reported Through the Redmond Red Electrochemical

Signal.

We can observe protein binding in electrochemistry

experiments by monitoring DNA-mediated transport between

the electrode and the redox active probe Redmond red that is

attached at either end of the DNA duplex (Fig. 2). The midpoint

potential of Redmond red is

160 mV versus NHE, and the

linearity of the plot of peak current as a function of scan rate

indicates that Redmond red behaves as a surface-bound species

(42). Although small potential shifts (

20 mV) in the Redmond

red signal are occasionally observed upon addition of

SoxR, Redmond red provides a convenient and reliable internal

standard.

It is expected that a redox-active probe located at the top of

the DNA monolayer will report on perturbations of the base pair

stack that intervene between the redox probe and the electrode,

whereas the same probe located at the bottom of the monolayer

near the electrode surface will not be affected by disruptions in

base stacking above the probe. Previously, we have reported

attenuation in charge accumulation by chronocoulometry for

daunomycin covalently attached near the top of a DNA film due

to perturbations in the intervening DNA structure by the

base-flipping methylase M.HhaI and TATA binding protein

(43). Here, as shown in Fig. 2, when Redmond red is incorpo-

rated above the SoxR binding site, a 16% decrease in the

integrated cathodic charge of Redmond red is observed upon

addition of SoxR. In contrast, when Redmond red is incorpo-

rated at the bottom of the DNA duplex below the SoxR binding

site, there is little detectable change in the Redmond red signal

in the presence of SoxR. Although the loss of signal observed

upon addition of SoxR is far smaller than that found for TATA

binding protein or M.HhaI, the decrease in electrochemical

signal when the binding site is positioned between the probe and

the electrode does provide evidence for SoxR binding.

Electrochemistry of

P. aeruginosa

SoxR.

As is evident in Fig. 2,

besides the Redmond red probe, we also observe a second

distinct and quasi-reversible electrochemical signal at

200 mV

Fig. 1.

Schematic illustration of the self-assembly/backfilling of a DNA monolayer followed by incubation with protein.

Gorodetsky

et al.

PNAS

March 11, 2008

vol. 105

no. 10

3685

BIOCHEMISTRY

CHEMISTRY

versus NHE upon addition of SoxR. The signal is observed only

after SoxR addition and is not affected by Redmond red because

it is also present in the absence of the probe. Note that no redox

signature is observed at

290 mV versus NHE, the potential

previously reported for SoxR in solution (Fig. 2). Incubation of

the DNA-modified surface with PA2274, a control protein that

lacks an iron-sulfur cluster, does not result in the appearance of

any redox signature. Furthermore, experiments with SoxR stocks

featuring low iron-sulfur content after purification do not lead to

appreciable cyclic voltammetric signals (data not shown). There-

fore, we can assign the new signal observed to the [2Fe-2S]

cluster of SoxR.

In a typical experiment, the observed SoxR signal increases

over a period of

15 min and is stable for a minimum of 18 scans

before slowly decaying (Fig. 3), although we have found a high

variability in electrode stability upon addition of protein. High

concentrations of protein (

M) are required for these

experiments, certainly concentrations higher than is required for

site-specific binding, and both the high protein concentrations

and long DNA sequences used make DNA/protein film forma-

tion difficult. It is important to note that the Redmond red signal

is highly stable and exhibits no noticeable degradation during

typical electrochemistry experiments.

As can be seen in Figs. 2 and 3, the cathodic and anodic waves

observed for SoxR are asymmetric: The oxidation wave is

pronounced and substantially less broad compared with the

reduction wave. In an ideal quasi-reversible system, the anodic

to cathodic peak current ratio is unity (42, 44), but this is

certainly not the case for SoxR. We find an anodic to cathodic

peak current ratio of 3.0 for SoxR in Fig. 2, strongly indicative

of a non-ideal and quasi-reversible electrochemical response. In

contrast, the anodic to cathodic peak current ratio is 1.3 for the

-Redmond red on the same film, far closer to the ideal value

for a fully reversible system. These data show that the electro-

chemistry of SoxR is complicated, hardly surprising given that

SoxR binds DNA as a dimer.

Interestingly, the asymmetries in the reduction and oxidation

waves of SoxR are qualitatively distinct from those previously

observed for the DNA repair enzymes MutY and Endo III; the

electrochemistry of those enzymes featured a reduction wave

that was somewhat more pronounced than the oxidation wave

(32). The better resolved anodic wave of SoxR integrates to very

low surface coverages of 0.5 pmol/cm

. This apparent low

coverage is comparable to that of 2 pmol/cm

previously found

for MutY at DNA monolayers on gold (33) and may reflect poor

coupling of the iron-sulfur cluster with the base pair stack.

However, the Redmond red probe at the bottom of the DNA

monolayer integrates to surface coverages of 1 pmol/cm

whereas the Redmond red probe at the top the film integrates

to coverages that are 3-fold lower (over sample sizes of at least

10 electrodes). These values are far less than the ideal DNA

surface coverage of 10 pmol/cm

expected for a loosely packed

DNA monolayer and indicate that the amount of DNA on the

surface is the main determinant of the size of the SoxR signal.

Despite the broad cathodic wave, we can calculate an upper

bound for the number of electrons transferred for the oxidation

of SoxR. For an ideal surface-bound species, the slope derived

from the plot of peak current as a function of scan rate divided

by the integrated charge

at any scan rate is equal to

where

is Faraday’s constant,

is the gas constant, and

is the

temperature (44); performing this operation for the Redmond

red probe at the bottom of the monolayer (

anodic/cathodic

nC at any scan rate) yields a value of

2, as expected for 2 e

transfer to the resorufin moiety. The integrated charge for the

anodic wave of SoxR on the same film varies from 3 to 9 nC,

indicating that SoxR receives at most half the number of

electrons transferred to the Redmond red. If we assume that all

of the DNA is bound and that the Redmond red signal at the

bottom of the monolayer corresponds perfectly to the number of

DNA molecules on the surface (highly likely for the sparse films

obtained), we can estimate that each DNA-bound SoxR dimer

undergoes at most a one electron oxidation/reduction. In fact, all

of these observations are consistent with titrations of free SoxR,

which deviate from ideal reductions, but appear also to yield

values of

1 (8, 45).

Fig. 2.

Cyclic voltammetry at 50 mV/s of electrodes modified with DNA

featuring Redmond red at the bottom (

), Redmond red above the binding

site (

), and no Redmond red (

). Voltammograms before addition of SoxR are

blue whereas those after addition of SoxR are red. The sequences used in the

course of these experiments are illustrated with the binding sequence for SoxR

highlighted, the location of Redmond red indicated by an ‘‘R,’’ and the

location of abasic sites underlined.

Fig. 3.

Binding of SoxR to the DNA-modified film. (

Left

) Background sub-

tracted cyclic voltammetry of

P. aeruginosa

SoxR at DNA-modified graphite

electrodes at a 50 mV/s scan rate immediately after addition of SoxR (light) and

20 min after addition of SoxR (dark) revealing the signal observed. (

Right

)

Integrated anodic charge for SoxR showing the growth of the signal as a

function of time.

3686

www.pnas.org

cgi

doi

10.1073

pnas.0800093105

Gorodetsky

et al.

Comparison of the Voltammetry of

E. coli

and

P. aeruginosa

SoxR.

expand our work to multiple organisms, we directly compared

the electrochemistry of

E. coli

and

P. aeruginosa

SoxR. Only

weak cyclic voltammetry for

E. coli

SoxR is obtained, irrespective

of the source, likely because of poor solubility. Therefore, the

comparison was made using square wave voltammetry, which is

a more sensitive technique and allows for better discrimination

of small signals. As can be seen in Fig. 4, the potentials,

referenced to the Redmond red internal standard, are nearly

identical for

P. aeruginosa

SoxR and

E. coli

SoxR. This obser-

vation is consistent with the

in vitro

redox titrations of SoxR in

the absence of DNA that found their potentials to differ from

one another by

10 mV (7, 8, 15). Here, we see that the

DNA-bound potentials are indistinguishable. The identical re-

dox potentials of the free and DNA-bound

E. coli

and

aeruginosa

SoxR should allow both to activate transcription upon

oxidation. In fact, previous work has shown that

P. putida

SoxR

is functional and can complement an

E. coli

SoxR deletion

mutant, lending credence to the

in vivo

importance of these

observations (13).

Discussion

The activity of SoxR, a transcription factor containing an Fe-S

cofactor, is regulated via a redox switch: SoxR triggers tran-

scription in its oxidized state (4, 5). However, the redox poten-

tials of free

E. coli

and

P. aeruginosa

SoxR have previously been

determined to be approximately

290 mV in solution (7, 8, 15).

Although the redox potential of SoxR can explain how it is

maintained in its reduced state by coupling it to the cellular

NADPH/NADP

pool (

40 mV), it was unclear how the

relatively low potential of

290 mV would allow for specificity

in vivo

. To understand the activation of SoxR at a mechanistic

level, it is crucial to determine its redox potential within an

appropriate context.

Here, using DNA-modified HOPG electrodes, we have dem-

onstrated that DNA association positively shifts the redox po-

tential of SoxR to 200 mV versus NHE. The

490-mV shift

between the free and DNA-bound states of SoxR is functionally

crucial because it keeps SoxR reduced across a range of intra-

cellular potential. For example, Fig. 5 shows standard and free

midpoint potentials for a variety of cellular redox pairs, and

where DNA-bound and free SoxR are positioned along this

series. Although numerous redox couples, ranging from gluta-

thione to FADH, are oxidants for soluble SoxR, they are

reductants to DNA-bound SoxR. In fact, the positive shift in

potential associated with DNA binding means that DNA-bound

SoxR is primarily in the reduced, transcriptionally silent form

vivo

. Oxidative stress serves to promote oxidation of DNA-

bound SoxR, activating the numerous genes required to protect

the organism. This provides a rationale for how DNA-bound

SoxR can serve as an effective sensor of oxidative stress in

E. coli

P. aeruginosa

, the paradigm for SoxR activation may be

different. Here, activation may be promoted by pyocyanin. When

we consider the standard potential of the phenazine pyocyanin

(

34 mV at pH 7 and

110 mV at pH 8), we predict

it would also act as a reductant for DNA-bound SoxR (46).

However, pyocyanin is an extracellular electron shuttle that

reacts readily with oxygen, as indicated by the bright blue color

P. aeruginosa

cultures. Uptake of oxidized pyocyanin increases

the intracellular ratio of the oxidized versus the reduced form

and thus favors the oxidation of SoxR. Considering a one-

electron transfer under physiological conditions (pH 7 and

37°C), to shift the redox equilibrium of DNA-bound SoxR

toward its oxidized state would require a ratio of oxidized to

reduced pyocyanin of at least 6,500:1. It remains to be deter-

mined whether this is of physiological relevance in

P. aeruginosa

The substantial shift in SoxR potential on DNA binding of

500 mV is striking but understandable. The significance of the

molecular environment for tuning the redox potentials of Fe-S

clusters is well documented (47–49): Each hydrogen-bonding

interaction with the cluster can cause a potential shift of

mV. Moreover, for [4Fe-4S] clusters in proteins, all with the

same ligating residues, cluster potentials vary from approxi-

mately

600 mV for ferredoxins to approximately

400 mV for

high potential iron proteins (50). We have previously observed

a negative shift of at least

200 mV for Endo III in the presence

of DNA (34). Because the structures of Endo III with and

without DNA were known and showed no significant distortion

in the protein (51–53), thermodynamically this shift was inter-

preted as a favorable shift in the binding affinity of the protein

in the oxidized form relative to the reduced form (34), perhaps

not so surprising on binding to the DNA polyanion.

By contrast, although the binding affinities of oxidized and

reduced SoxR are comparable (6, 7), SoxR and other MerR-type

transcriptional regulators have been shown to induce conforma-

tional changes of the promoter region (1, 54–59). In particular,

copper phenanthroline footprinting studies have provided strong

evidence that SoxR significantly distorts its promoter sequence

(60, 61). Although this experiment only reports on the reduced

form of SoxR, the observed loss of signal for Redmond red found

here strongly supports a DNA-distortion mechanism. We pro-

pose that the more positive reduction potential for DNA-bound

SoxR yields a higher energy complex, which may drive a con-

formational change in the protein/DNA complex. If this were the

case, it would constitute an effective means of allosteric regu-

lation

in vivo

It is important to note that the crystal structure of SoxR in any

form has not been reported (62). Therefore, the structural

difference between the free (low energy) and DNA-bound (high

energy) complexes is not clear, but a positive shift of the

magnitude we observe has been associated with bulk folding of

other metalloproteins from

P. aeruginosa

(63). These energetic

differences have been attributed to burying the cofactor in a

more hydrophobic environment. Consequently, we predict a

Fig. 4.

Square wave voltammetry of

P. aeruginosa

SoxR (

Left

) and

E. coli

(

Right

) at DNA-modified graphite electrodes at a frequency of 15 Hz showing

both the Redmond red and SoxR signals. The 5

Redmond red-modified

sequence was 5

-AGR GTA AAA CCT CAA GCA AAC TTG AGG TCA AGC CAA-3

plus pyrene-modified complement, where ‘‘R’’ indicates the position of the

probe.

Fig. 5.

Redox potentials of free and DNA-bound SoxR along with those of

cellular oxidants/reductants at pH 7.

Gorodetsky

et al.

PNAS

March 11, 2008

vol. 105

no. 10

3687

BIOCHEMISTRY

CHEMISTRY

large structural difference between the free and DNA-bound

SoxR to provide a rationale for the dramatic shift in potential

associated with binding.

Within a broader context, these data illustrate that it is critical

to take the effect of DNA binding into account when considering

the redox characteristics of DNA binding proteins. It is also likely

that it is the DNA-bound potential of these proteins that is most

relevant within the crowded environment of the cell, and this

potential may be altered even further upon recruitment of RNA

polymerase. Therefore, in many cases, as with SoxR, it is perhaps

the redox characteristics within a multiprotein/DNA complex

that must be considered. In fact, because several transcription

factors feature iron-sulfur clusters as sensor elements, a change

in the redox potential of these cofactors upon binding DNA may

generally be an important trait to consider.

Experimental Procedures

Materials.

All phosphoramidites and reagents for DNA synthesis were pur-

chased from Glen Research. 1,6-Diaminohexane was obtained from Acros

Organics. All organic solvents and other reagents were purchased from Al-

drich in the highest available purity.

Oligonucleotide Synthesis.

Oligonucleotides were prepared using standard

phosphoramidite chemistry on an ABI 394 DNA synthesizer. DNA was purified

by HPLC on a reverse-phase C18 column with acetonitrile and ammonium

acetate as eluents. The desired products were characterized by HPLC, UV-

visible spectroscopy, and MALDI-TOF mass spectrometry. For experiments on

HOPG, DNA was modified with pyrene at the 5

terminus by following the

procedure reported in ref. 27. In brief, oligonucleotides were prepared by

solid phase synthesis using standard reagents with an unprotected hydroxyl

group at the 5

terminus. The 5

-OH was treated with a 120 mg/ml solution of

carbonyldiimidazole in dioxane fo

r 2 h followed by an 80 mg/ml solution of

1,6-diaminohexane for 30 min. Subsequently, the free amine was treated with

1-pyrenebutyric acid,

-hydroxysuccinimide ester, resulting in the desired

pyrene moiety linked to the 5

terminus. The oligonucleotides were depro-

tected with concentrated NH

OH at 60°C for 8 h.

DNA modified with Redmond red at the 3

terminus or 3 bases in from the

terminus was prepared according to the ultra-mild protocols outlined on

the Glen Research web site (www.glenres.com). Pac-protected bases and ultra

mild reagents were used. The oligonucleotides were deprotected in 0.05 M

potassium carbonate in methanol at room temperature for 12–14 h to prevent

degradation of the Redmond red moiety under harsh conditions.

Expression and Purification.

E. coli

SoxR was prepared as described in ref. 64.

N-terminally histidine-tagged SoxR from

P. aeruginosa

PA14 were expressed

from plasmid pET16b in

E. coli

strain BL21 (DE3). Cells were grown in 1 liter of

LB medium with 100

g/ml ampicillin at 37°C. At an OD

600nm

of 0.3, protein

expression was induced by the addition of 1 mM IPTG and the cultures were

incubated for an additional 10 h at 16°C. All subsequent steps were performed

at 4°C. Cells were pelleted, resuspended in buffer A [50 mM NaH

(pH 8.0),

300 mM NaCl, 10% glycerol] containing 10 mM imidazole and PIC (protease

inhibitor mixture without EDTA; Roche), and lysed using a French Press. The

cell extract was centrifuged at 14,000

for 20 min. The supernatant was

incubated with Talon-beads (Clontech) for 30 min and then transferred to a

column. The beads were washed with buffer A containing 50 mM imidazole

and PIC. Histidine-tagged SoxR was eluted from the column with buffer A

containing 250 mM imidazole and PIC. Peak fractions and purity were deter-

mined by SDS/PAGE with Coomassie blue staining. Purified protein was dia-

lyzed against SoxR storage buffer [50 mM Pi (pH 8.0), 500 mM NaCl, 20%

glycerol].

To generate expression plasmid pET16b-soxR,

soxR

(PA14

35170) was PCR-

amplified from genomic DNA of

P. aeruginosa

PA14 using primers A (CGC

cat

atg

AAG AAT TCC TGC GCA TC) and B (GGC gga tcc CTA GCC GTC GTG CTC

G). Primer A contains an NdeI restriction site (small letters) and

soxR

’s start

codon (italicized). Primer B contains a BamHI site (small letters) and

soxR

’s stop

codon. The PCR fragment was ligated into NdeI/BamHI-digested pET16b.

Formation of DNA Monolayers and Electrochemical Measurements.

DNA films

were self-assembled on SPI-1 grade HOPG electrodes (SPI) with an estimated

surface area of 0.08 cm

defined by an

-ring. Duplex DNA was formed in (pH

8) 50 mM P

/500 mM NaCl/20% glycerol buffer (SoxR storage buffer) by

combining equimolar amounts of the pyrene-modified strand with its com-

plement. Loosely packed DNA monolayers were allowed to form over a period

of 24 – 48 h. The electrodes were then thoroughly rinsed with SoxR storage

buffer before being backfilled for 2– 4 h with 10% by volume octane or decane

solutions in SoxR storage buffer. The electrodes were then thoroughly rinsed

with SoxR storage buffer again and moved into a nitrogen atmosphere for

electrochemistry experiments.

Electrochemical data were collected with a Bioanalytical Systems CV-50W

potentiostat using the inverted drop cell configuration. All measurements

reported for the working electrode were taken versus a platinum (Pt) auxiliary

and a silver/silver chloride (Ag/AgCl) reference. The Ag/AgCl reference was

frequently standardized versus SCE, and all reported potentials have an

experimental uncertainty of

40 mV. Electrochemical experiments were per-

formed at ambient temperature and under an anaerobic atmosphere in SoxR

storage buffer. In a typical experiment, background electrochemical scans

were performed before SoxR was added to the storage buffer, resulting in an

15–35

M monomer concentration within the cell. Further scans were then

performed in the presence of SoxR, typically over a period of 30 – 45 min.

ACKNOWLEDGMENTS.

This work was supported by National Institutes of

Health Grant GM61077 (to J.K.B.), the Howard Hughes Medical Institute

(D.K.N.), and a European Molecular Biology Organization Long-Term Fellow-

ship (to L.E.P.D.).

1. Brown NL, Stoyanov JV, Kidd SP, Hobman JL (2003) The MerR family of transcriptional

regulators.

FEMS Microbiol Rev

27:145–163.

2. Hobman JL, Wilkie J, Brown NL (2005) A design for life: Prokaryotic metal-binding MerR

family regulators.

Biometals

18:429 – 436.

3. Giedroc DP, Arunkumar AI (2007) Metal sensor proteins: Nature’s metalloregulated

allosteric switches.

J Chem Soc Dalton Trans

, 3107–3120.

4. Pomposiello PJ, Demple B (2001) Redox-operated genetic switches: The SoxR and OxyR

transcription factors.

Trends Biotechnol

19:109 –114.

5. Demple B, Ding H, Jorgensen M (2002) Escherichia coli SoxR protein: Sensor/transducer

of oxidative stress and nitric oxide.

Methods Enzymol

348:355–364.

6. Hidalgo E, Demple B (1994) An iron-sulfur center essential for transcriptional activation

by the redox sensing SoxR protein.

EMBO J

13:138 –146.

7. Gaudu P, Weiss B (1996) SoxR, a [2Fe-2S] transcription factor, is active only in its

oxidized form.

Proc Natl Acad Sci USA

93:10094 –10098.

8. Ding H, Hidalgo E, Demple B (1996) The redox state of the [2Fe-2S] clusters in SoxR

protein regulates its activity as a transcription factor.

J Biol Chem

271:33173–

33175.

9. Ding H, Demple B (1997) In vivo kinetics of a redox-regulated transcriptional switch.

Proc Natl Acad Sci USA

94:8445– 8449.

10. Pomposiello J, Bennik MHJ, Demple B (2001) Genome-wide transcriptional profiling of

the Escherichia coli responses to superoxide stress and sodium salicylate.

J Bacteriol

183:3890 –3902.

11. Ha U, Jin S (1999) Expression of the soxR gene of Pseudomonas aeruginosa is inducible

during infection of burn wounds in mice and is required to cause efficient bacteremia.

Infect Immun

67:5324 –5331.

12. Stover CK,

etal.

(2000) Complete genome sequence of Pseudomonas aeruginosa PAO1,

an opportunistic pathogen.

Nature

406:959 –964.

13. Park W, Pena-Llopis S, Lee Y, Demple B (2006) Regulation of superoxide stress in

Pseudomonas putida KT2440 is different from the SoxR paradigm in Escherichia coli.

Biochem Biophys Res Commun

341:51–56.

14. Dietrich LEP, Price-Whelan A, Petersen A, Whiteley M, Newman DK (2006) The phena-

zine pyocyanin is a terminal signalling factor in the quorum sensing network of

Pseudomonas aeruginosa

. Mol Microbiol

61:1308 –1321.

15. Kobayashi K, Tagawa S (2004) Activation of SoxR-dependent transcription in Pseudo-

monas aeruginosa.

J Biochem

136:607– 615.

16. Zheng M, Storz G (2000) Redox sensing by prokaryotic transcription factors.

Biochem

Pharmacol

59:1– 6.

17. Liochev SI, Fridovich I (1992) Fumarase C, the stable fumarase of Escherichia coli, is

controlled by the soxRS regulon.

Proc Natl Acad Sci USA

89:5892–5896.

18. Kobayashi K, Tagawa S (1999) Isolation of reductase for SoxR that governs an oxidative

response regulon from Escherichia coli.

FEBS Lett

451:227–230.

19. Koo M-S,

et al

. (2003) A reducing system of the superoxide sensor SoxR in E. coli.

EMBO

22:2614 –2622.

20. Hwang C, Lodish HF, Sinskey AJ (1995) Measurement of glutathione redox state in

cytosol and secretory pathway of cultured cells.

Methods Enzymol

251:212–221.

21. Kobayashi K, Tagawa S (1997) A direct demonstration of reaction of superoxide anion

with SoxR as studied by pulse radiolysis.

J Inorg Biochem

67:258.

22. Liochev SI, Benov L, Touati D, Fridovich I (1999) Induction of the SoxRS regulon of

Escherichia coli by superoxide.

J Biol Chem

274:9479 –9481.

23. Ding H, Demple B (1998) Thiol-mediated disassembly and reassembly of [2Fe-2S]

clusters in the redox-regulated transcription factor SoxR.

Biochemistry

37:17280 –

17286.

24. Price-Whelan A, Dietrich LEP, Newman DK (2007) Pyocyanin alters redox homeostasis

and carbon flux through central metabolic pathways in Pseudomonas aeruginosa

PA14.

J Bacteriol

189:6372– 6381.

3688

www.pnas.org

cgi

doi

10.1073

pnas.0800093105

Gorodetsky

et al.

25. Kelley SO, Boon EM, Barton JK, Jackson NM, Hill MG (1999) Single-base mismatch

detection based on charge transduction through DNA.

Nucleic Acids Res

27:4830 –

4837.

26. Boon EM, Ceres DM, Drummond TG, Hill MG, Barton JK (2000) Mutation detection by

electrocatalysis at DNA-modified electrodes.

Nat Biotechnol

18:1096 –1100.

27. Gorodetsky AA, Barton JK (2006) Electrochemistry using self-assembled DNA mono-

layers on highly oriented pyrolytic graphite.

Langmuir

22:7917–7922.

28. Kelley SO, Jackson NM, Hill MG, Barton JK (1999) Long-range electron transfer through

DNA films.

Angew Chem Int Ed

38:941–945.

29. Inouye M, Ikeda R, Takase M, Tsuri T, Chiba J (2005) Single-nucleotide polymorphism

detection with ‘‘wire-like’’ DNA probes that display quasi ‘‘on– off’’ digital action.

Proc

Natl Acad Sci USA

102:11606 –11610.

30. Okamoto A, Kamei T, Saito I (2006) DNA hole transport on an electrode: Application

to effective photoelectrochemical SNP typing.

J Am Chem Soc

128:658 – 662.

31. Gorodetsky AA, Green O, Yavin E, Barton JK (2007) Coupling into the base pair stack

is necessary for DNA-mediated electrochemistry.

Bioconjugate Chem

18:1434 –1441.

32. Boon EM, Livingston AL, Chmiel NH, David SS, Barton JK (2003) DNA-mediated charge

transport for DNA repair.

Proc Natl Acad Sci USA

100:12543–12547.

33. Boal AK,

et al.

(2005) DNA-bound redox activity of DNA repair glycosylases containing

[4Fe-4S] clusters.

Biochemistry

44:8397– 8407.

34. Gorodetsky AA, Boal AK, Barton JK (2006) Direct electrochemistry of endonuclease III

in the presence and absence of DNA.

J Am Chem Soc

128:12082–12083.

35. Cunningham RP,

et al.

(1989) Endonuclease III is an iron-sulfur protein.

Biochemistry

28:4450 – 4455.

36. Asahara H, Wistort PM, Bank JF, Bakerian RH, Cunningham RP (1989) Purification and

characterization of Escherichia coli endonuclease III from the cloned nth gene.

Bio-

chemistry

28:4444 – 4449.

37. Fu W, O’Handley S, Cunningham RP, Johnson MK (1992) The role of the iron-sulfur

cluster in Escherichia coli endonuclease III. A resonance Raman study.

J Biol Chem

267:16135–16137.

38. Boal AK, Yavin E, Barton JK (2007) DNA repair glycosylases with a [4Fe-4S] cluster: A

redox cofactor for DNA-mediated charge transport?

J Inorg Biochem

101:1913–1921.

39. Stephens PJ, Jollie DR, Warshel A (1996) Protein control of redox potentials of iron-

sulfur proteins.

Chem Rev

96:2491–2513.

40. Kim Y, Long EC, Barton JK, Lieber CM (1992) Imaging of oligonucleotide-metal com-

plexes by scanning tunneling microscopy.

Langmuir

8:496 –500.

41. Giancarlo LC, Flynn GW (1998) Scanning tunneling and atomic force microscopy probes

of self-assembled, physisorbed monolayers: Peeking at the peaks.

Annu Rev Phys Chem

49:297–336.

42. Bard AJ, Faulkner LR (2001)

Electrochemical Methods

(Wiley, New York), 2nd Ed.

43. Boon EM, Salas JE, Barton JK (2002) An electrical probe of protein–DNA interactions on

DNA-modified surfaces.

Nat Biotechnol

20:282–286.

44. Udit AK, Hill MG, Bittner VG, Arnold FM, Gray HB (2004) Reduction of dioxygen

catalyzed by pyrene-wired heme domain cytochrome P450 BM3 electrodes.

J Am Chem

Soc

126:10218 –10219.

45. Hidalgo E, Ding H, Demple B (1997) Redox signal transduction via iron-sulfur clusters

in the SoxR transcription activator.

Trends Biochem Sci

22:207–210.

46. Price-Whelan A, Dietrich LEP, Newman DK (2006) Rethinking ‘‘secondary’’ metabolism:

Physiological roles for phenazine antibiotics.

Nat Chem Biol

2:71–78.

47. Rusling JF (1998) Enzyme bioelectrochemistry in cast biomembrane-like films.

Acc

Chem Res

31:363–369.

48. Carter, CW, Jr,

et al.

(1972) A comparison of Fe4S4 clusters in high-potential iron

protein and in ferredoxin.

Proc Natl Acad Sci USA

69:3526 –3529.

49. Heering HA, Bulsink YBM, Hagen WR, Meyer TE (1995) Influence of charge and polarity

on the redox potentials of high-potential iron-sulfur proteins: Evidence for the exis-

tence of two groups.

Biochemistry

34:14675–14686.

50. Beinert H (2000) Iron-sulfur proteins: Ancient structures, still full of surprises.

J Biol

Inorg Chem

5:2–15.

51. Kuo CF,

et al.

(1992) Atomic structure of the DNA repair [4Fe-4S] enzyme endonuclease

III.

Science

258:434 – 440.

52. Thayer MM, Ahern H, Xing D, Cunningham RP, Tainer JA (1995) Novel DNA-binding

motifs in the DNA repair enzyme endonuclease III crystal structure.

EMBO J

14:4108 –

4120.

53. Fromme JC, Verdine GL (2003) Structure of a trapped endonuclease III–DNA covalent

intermediate.

EMBO J

22:3461–3471.

54. Ansari AZ, Bradner JE, O’Halloran TV (1995) DNA-bend modulation in a repressor-to-

activator switching mechanism.

Nature

374:371–375.

55. Outten CE, Outten FW, O’Halloran TV (1999) DNA distortion mechanism for transcrip-

tional activation by ZntR, a Zn(II)-responsive MerR homologue in Escherichia coli.

J Biol

Chem

274:37517–37524.

56. Zheleznova-Heldwein EE, Brennan RG (2001) Crystal structure of the transcription

activator BmrR bound to DNA, a drug.

Nature

409:378 –382.

57. Godsey MH, Baranova NN, Neyfakh AA, Brennan RG (2001) Crystal structure of MtaN,

a global multidrug transporter gene activator.

J Biol Chem

276:47178 – 47184.

58. Changela A,

et al.

(2003) Molecular basis of metal-ion selectivity and zeptomolar

sensitivity by CueR.

Science

301:1383–1387.

59. Newberry KJ, Brennan RG (2004) The structural mechanism for transcription activation

by MerR family member multidrug transporter activation, N terminus.

J Biol Chem

279:20356 –20362.

60. Hidalgo E, Bollinger JM, Jr, Bradley TM, Walsh CT, Demple B (1995) Binuclear [2Fe-2S]

clusters in the Escherichia coli SoxR protein and role of the metal centers in transcrip-

tion.

J Biol Chem

270:28908 –28914.

61. Hidalgo E, Demple B (1997) Spacing of promoter elements regulates the basal expres-

sion of the soxS gene and converts SoxR from a transcriptional activator into a

repressor.

EMBO J

16:1056 –1065.

62. Watanabe S, Kita A, Kobayashi K, Takahashi Y, Miki K (2006) Crystallization and

preliminary x-ray crystallographic studies of the oxidative-stress sensor SoxR and its

complex with DNA.

Act Crystallogr F

62:1275–1277.

63. Wittung-Stafshede P,

et al.

(1998) Reduction potentials of blue and purple copper

proteins in their unfolded states: A closer look at rack-induced coordination.

J Biol

Inorg Chem

3:367–370.

64. Chander M, Demple B (2004) Functional analysis of SoxR residues affecting transduc-

tion of oxidative stress signals into gene expression.

J Biol Chem

279:41603– 41610

Gorodetsky

et al.

PNAS

March 11, 2008

vol. 105

no. 10

3689

BIOCHEMISTRY

CHEMISTRY