Blackwell Publishing LtdOxford, UKMMIMolecular Microbiology0950-382X 2006 The Authors; Journal compilation 2006 Blackwell Publishing Ltd 2006601209218Original ArticleIrr degradation in response to oxidative stressJ. Yang, H. R.

Panek and M. R. O Brian

Molecular Microbiology (2006) 60(1), 209 218 doi:10.1111/j.1365-2958.2006.05087.x

First published online 16 February 2006

Oxidative stress promotes degradation of the

Irr protein to regulate haem biosynthesis in

Bradyrhizobium japonicum

Jianhua Yang, Heather R. Panek and bacterial membranes to hydrogen peroxide (Seaver and

Imlay, 2001). Superoxide (O2 ) and hydrogen peroxide

Mark R. O Brian*

Department of Biochemistry and Witebsky Center for (H2O2), the initial partial reduction products of oxygen, can

Microbial Pathogenesis and Immunology, State University damage cell components but only to a limited extent.

However, O2 is capable of destroying exposed iron-

of New York at Buffalo, Buffalo, NY 14214 USA.

sulphur clusters with the release of free iron and the

released iron subsequently reacts with H2O2 via the Fen-

Summary

ton reaction generating extremely reactive hydroxyl radi-

cals (HO ). Hydroxyl radicals can directly attack most

The haem proteins catalase and peroxidase are stress

response proteins that detoxify reactive oxygen spe- macromolecules such as DNA, lipids and proteins, which

cies. In the bacterium Bradyrhizobium japonicum, is the basis of oxygen toxicity (reviewed in Imlay, 2003).

expression of the gene encoding the haem biosynthe- Bacteria have multiple defence strategies against oxi-

sis enzyme d-aminolevulinic acid dehydratase (ALAD) dative stress, including the direct detoxi cation of ROS by

is normally repressed by the Irr protein in iron-limited catalase, peroxidases and superoxide dismutase. Oxida-

cells. Irr degrades in the presence of iron, which tive stress responses require the activation of regulatory

requires haem binding to the protein. Here, we found proteins and the induction of genes under their control. In

that ALAD levels were elevated in iron-limited cells of many bacteria, the transcriptional regulator OxyR (Christ-

a catalase-de cient mutant, which corresponded with man et al., 1989; Tao et al., 1989) senses hydrogen per-

aberrantly low levels of Irr. Irr was undetectable in oxide (Zheng et al., 1998) and induces numerous genes

wild-type cells within 90 min after exposure to exog- whose products are involved in peroxide defence (Tartaglia

enous H2O2, but not in a haem-de cient mutant strain. et al., 1989; Altuvia et al., 1994), redox balance (Prieto-

In addition, Irr did not degrade in response to iron in Alamo et al., 2000; Ritz et al., 2000) and other factors

the absence of O2. The ndings indicate that reactive (Altuvia et al., 1997; Zheng et al., 1999). In Bacillus subtilis,

oxygen species promote Irr turnover mediated by PerR is the major peroxide regulator and represses a large

haem, and are involved in iron-dependent degrada- PerR regulon (Herbig and Helmann, 2001). The OhrR

tion. We demonstrated Irr oxidation in vitro, which family of antioxidant regulators is responsible for organic

required haem, O2 and a reductant. A truncated Irr hydroperoxide resistance (Mongkolsuk et al., 1998).

mutant unable to bind ferrous haem does not degrade Catalases and peroxidases are haem proteins that

in vivo, and was not oxidized in vitro. We suggest that detoxify H2O2 and peroxides respectively. Regulation of

Irr oxidation is a signal for its degradation, and that genes encoding haem biosynthesis enzymes in response

cells sense and respond to oxidative stress through to oxidative stress has been reported in several bacteria,

Irr to regulate haem biosynthesis. and this is presumably due to a need for haem for detox-

i cation. The B. subtilis PerR protein mediates the induc-

tion of hemAXCDBL operon encoding enzymes for the

Introduction

early steps of haem synthesis (Chen et al., 1995; Mon-

The generation of reactive oxygen species (ROS) is a gkolsuk and Helmann, 2002). In Escherichia coli and Sal-

natural consequence of aerobic respiration as a result of monella, the hemH gene encoding the haem biosynthetic

the partial reduction of molecular oxygen and the subse- enzyme ferrochelatase is induced in response to H2O2 in

quent reactions of those products with transition metals an OxyR-dependent manner (Zheng et al., 2001; Elgrably

and other compounds (Fridovich, 1995). The environment Weiss et al., 2002). In Salmonella, a mutant defective in

also can be a source of ROS given the permeability of the gene encoding the haem synthesis enzyme glutamyl-

tRNA reductase is hypersensitive to H2O2 (Elgrably Weiss

et al., 2002). However, this was attributed to an increase

Accepted 19 January, 2006. *For correspondence. E-mail

in reducing equivalents to drive the Fenton reaction rather

abph7a@r.postjobfree.com; Tel. +1-716-***-****; Fax +1-716-***-****.

than to a decrease in catalase activity.

Both authors contributed equally to this work.

2006 The Authors

Journal compilation 2006 Blackwell Publishing Ltd

210 J. Yang, H. R. Panek and M. R. O Brian

Our laboratory is interested in the regulation of haem

A katG

Wt

biosynthesis and its co-ordination with oxidative metabo-

M Fe

lism in the bacterium Bradyrhizobium japonicum. The 0 1 2 4 6 0 1 2 4 6

rhizobia are bacteria that exist as free-living soil organ- ALAD

isms or in symbiosis with leguminous plants. In symbiosis,

IRR

the bacteria convert atmospheric nitrogen to ammonia for

the nutritional nitrogen requirement of the plant. KatG

is the major functional catalase in cultured cells of

B katG

katG

Wt

B. japonicum (Panek and O Brian, 2004) and Rhizobium

M Fe

0 1 6 0 1 6

etli (Vargas Mdel et al., 2003). Sinorhizobium meliloti has

irr mRNA

three catalases which have different roles in free-living

growth and symbiosis (Herouart et al., 1996; Sigaud et al.,

Fig. 1. Iron-dependent expression of ALAD and Irr in parent strain

1999; Jamet et al., 2003). A haem-de cient mutant of

I110 and a catalase-de cient katG strain. The strains were grown in

B. japonicum is very sensitive to killing by H2O2 (Panek modi ed GSY with various concentrations of FeCl3 added to the

and O Brian, 2004). growth media as indicated. Samples were taken from each culture

for protein analysis and total RNA isolation.

Iron and oxidative stress are intertwined because iron

A. Steady state levels of ALAD and Irr proteins were detected by

is involved in oxygen chemistry (Touati, 2000). The hemB immunoblot analysis using polyclonal antibodies directed against the

gene encoding -aminolevulinic acid dehydratase (ALAD) respective proteins.

B. Steady state irr mRNA levels were analysed by RNase protection

is highly regulated by iron in B. japonicum. This control is

analyses as described in Experimental procedures.

mediated by the iron response regulator (Irr), which neg-

atively regulates hemB under iron limitation (Hamza et al.,

1998). Irr degrades in the presence of iron to derepress abundant Irr protein found in the parent strain grown under

low-iron conditions (0 2 M FeCl3), Irr levels were much

hemB. Irr is a conditionally stable protein that degrades

rapidly when cells are exposed to iron, allowing derepres- lower in the katG strain at those iron concentrations. The

sion of haem synthesis (Hamza et al., 1998). This iron- low Irr level in the katG strain is consistent with the higher

dependent turnover is mediated by haem, which binds ALAD accumulation because Irr negatively controls ALAD.

At 6 M Fe, Irr levels were very low in both the parent and

directly to Irr to promote degradation (Qi et al., 1999). Irr

interacts directly with the haem biosynthesis enzyme fer- mutant strains.

rochelatase, and thus responds to iron via the status of Previous work demonstrated that the irr gene is mod-

haem at the site of haem synthesis (Qi and O Brian, estly regulated by iron at the transcriptional level, but is

2002). strongly controlled by protein stability leading to degrada-

In the present study, we show that ROS promote Irr tion of Irr in the presence of iron (Hamza et al., 1998;

degradation in vivo, and that oxidation of Irr is involved in 2000; Qi et al., 1999). RNAse protection analysis revealed

iron-dependent turnover. essentially no difference between irr mRNA levels in the

parent and katG strains at 0, 1 or 6 M iron (Fig. 1B)

Results despite large differences in the protein level. In both cell

types, a reduction in the mRNA level was detected at 6 M

ALAD is upregulated in a catalase-de cient strain due to

iron. This indicates that the low level of Irr in the katG

a destabilized Irr protein

strain is post-transcriptional, and is likely due to

decreased stability of the protein. The similar levels of irr

The hemB gene encodes the enzyme ALAD, which

mRNA between the wild type and katG mutant at 0 and

catalyses the second step in haem biosynthesis in

1 M iron indicate that the intracellular iron levels were not

B. japonicum (Panek and O Brian, 2002) The expression

substantially different in the two strains. Haem promotes

of hemB is repressed under iron limitation and is upregu-

Irr degradation (Qi et al., 1999). However, total haem

lated in response to iron (Chauhan et al., 1997). Low

(protohaem) levels in the parent strain and the katG

ALAD protein levels are observed in wild-type cells grown

mutant were similar (0.16 versus 0.14 nmol haem per mg

under iron limitation (Chauhan et al., 1997) (Fig. 1A).

protein respectively). Thus, differences in haem unlikely

However, ALAD was easily detected in low-iron cells (0

2 M) of a katG mutant strain defective in the catalase contribute to lower Irr levels in the katG strain.

hydroperoxidase KatG (Fig. 1A). We found previously that

iron-dependent control of hemB is mediated by Irr, which

Evidence that the katG strain accumulates more H2O2

negatively regulates it under iron limitation (Hamza et al.,

than the parent strain

1998). Therefore, we compared Irr protein levels in cells

Although the B. japonicum genome contains several cat-

of the wild-type and katG strains grown in the presence

alase genes, we found that KatG is the primary scavenger

of various iron concentrations (Fig. 1A). In contrast to the

2006 The Authors

Journal compilation 2006 Blackwell Publishing Ltd, Molecular Microbiology, 60, 209 218

Irr degradation in response to oxidative stress 211

of endogenously generated H2O2 in aerobically grown Experimental procedures). H2O2 levels in the media from

cells, and is the only detectable catalase in cultured cells the katG mutant cultures increased to a greater extent

(Panek and O Brian, 2004). Thus, the katG strain may than for wild-type cultures over that time period. Although

contain more H2O2 than the parent strain. To address this, we cannot assume that the concentration of H2O2 in the

we measured H2O2 in the spent media of cultures of the media is equivalent to the cellular concentration, the data

wild-type and katG strains grown in iron-de cient media suggest that the katG mutant cells contain more H2O2 than

(Fig. 2A). H2O2 was produced by media alone when aer- the wild type. This raises the possibility that H2O2 or its

ated over the 70 h that the cells were grown, as was Fenton products contributes to the aberrant accumulation

reported previously (Seaver and Imlay, 2001). Thus, a of Irr in the katG strain.

control lacking cells was subtracted from the quantity Although a B. japonicum hemA strain contains reduced

measured in the spent media to estimate the H2O2 catalase activity, the remaining activity is suf cient to

excreted by cells. The katG mutant strain excreted about detoxify H2O2 produced by aerobic metabolism (Panek

sevenfold more H2O2 into the media in cultures grown and O Brian, 2004). Consistent with those ndings, the

under iron de ciency compared with the parent strain. We level of H2O2 in the spent media of cultures of the hemA

also measured the accumulation of H2O2 in culture media strain was low and similar to the wild type (Fig. 2A). This

over a 4 h time period (Fig. 2B). In these experiments, is likely the reason why Irr levels in the hemA strain are

cultures were washed and resuspended in fresh media, similar to the wild type when grown under iron limitation

and samples were taken at various time points (see (Qi et al., 1999).

Exogenous H2O2 promotes Irr degradation

A

To further address the possible effects of H2O2 on Irr, we

4

examined Irr levels in wild-type cells grown in media sup-

plemented with 0 or 2 M FeCl3 in response to the addi-

3

tion of 2 mM H2O2 (Fig. 3). We chose 2 M iron because

M H2O2

it is a low enough concentration that Irr remains detectable

2 in wild-type cells, but is suf ciently high to elicit a response

as judged by the studies with the katG strain (Fig. 1). A

relatively high concentration of H2O2 was used because

1

B. japonicum cells consumes it rapidly (Panek and

O Brian, 2004). In the absence of the iron supplement, the

0 Irr level was essentially constant in the presence or

Wt katG hemA

absence of added H2O2, with only slight reduction at

B 90 min in the presence of H2O2 (Fig. 3A). However, in the

presence of 2 M iron, addition of 2 mM H2O2 resulted in

0.7

a decrease in Irr levels, with little protein remaining after

0.6

90 min. The control protein GroEL did not respond to H2O2

0.5

M H2O2

katG

under either iron condition (Fig. 3A and B). The steady

0.4

state level of irr mRNA in the presence of 2 M iron was

0.3

not changed by H2O2 treatment (Fig. 3C), indicating that

0.2

the disappearance of Irr does not re ect a change in irr

Wt

0.1 mRNA, but rather is due to protein instability. Collectively,

0.0 the results indicate that H2O2 destabilizes Irr in the pres-

ence of iron.

-0.1

0 1 2 3 4

hours

Haem is required for Irr turnover in response to H2O2

Fig. 2. H2O2 accumulation in spent media of cultures of the wild-type

The evidence presented thus far shows that Irr degrades

and mutant strains.

A. H2O2 accumulated in media cultured with parent strain I110, katG in response to H2O2, but iron must also be present

or hemA cultures. Cells were grown in low-iron media to mid-log

(Fig. 3A and B). Previous work shows that iron acts indi-

phase and the amount of H2O2 in the media was assayed by the

rectly through haem, and that Irr remains stable in haem

Amplex Red/HRP procedure.

B. Time-course of H2O2 accumulation in media. Cells were grown to synthesis mutants irrespective of iron levels (Qi et al.,

mid-log phase in minimal media, washed and resuspended in minimal

1999). Thus, we asked whether H2O2-facilitated Irr degra-

media to OD540 = 0.1. Aliquots were taken at the indicated time and

dation is haem-dependent. The hemA gene encodes the

H2O2 concentrations were determined.

2006 The Authors

Journal compilation 2006 Blackwell Publishing Ltd, Molecular Microbiology, 60, 209 218

212 J. Yang, H. R. Panek and M. R. O Brian

Fig. 3. Effects of H2O2 on the Irr accumulation

B

A in wild-type strain I110 and hemA mutant

MLG1. Cultures were grown to mid-log phase

Wt ( 2 M Fe)

Wt ( 0 M Fe)

in media with 0 (A) or 2 M (B D) FeCl3. At time

zero, 0 or 2 mM H2O2 was added and cultures

- H2O2 2 mM H2O2

- H2O2 2 mM H2O2

were continually aerated at 29 C. Aliquots were

*-**-**-**-*-**-** 90 min

*-**-**-**-*-**-** 90 min harvested at the indicated time for protein

(A, B, D) or RNA (C) analysis. (E) The hemA

mutant MLG1 grown with 2 M FeCl3 was

IRR IRR

treated with 2 mM H2O2 as in panel D, except

that after 60 min, either buffer (control) or

GroEL GroEL

10 M haemin (ferric haem hydrochloride) was

added to cultures (indicated by the arrow ).

Irr was measured after an additional 60 min

D incubation.

C

probe + RNase

hemA ( 2 M Fe)

intact probe

Wt ( 2 M Fe) - H2O2 2 mM H2O2

2mM H2O2

*-**-**-**-*-**-** 90 min

0 90 min

IRR

irr mRNA

GroEL

E hemA ( 2 M Fe)

2 mM H2O2

0 60 120 min

control

haemin

haem biosynthetic enzyme -aminolevulinic acid syn- Cells were grown aerobically in low-iron media, which

allows Irr to accumulate. Iron (6 M FeCl3) was then

thase, and thus a hemA mutant is haem-de cient. The

status of Irr in response to H2O2 treatment in the presence added to cultures that were either continuously exposed

of 2 M iron was determined in the hemA strain MLG1 as to air (Fig. 4A) or ushed with N2 to remove O2 prior to

described above for the parent strain (Fig. 3D). Com- iron addition (Fig. 4B). As observed previously (Qi et al.,

pared with the wild type (Fig. 3B), Irr was stable in the 1999), Irr degraded rapidly in cultures exposed to air, with

presence of H2O2, with substantial protein remaining 60 little protein remaining 60 min after iron addition (Fig. 4A).

90 min after H2O2 exposure. When 10 M haemin (ferric By contrast, Irr was stable in the presence of iron under

haem hydrochloride) was added to cultures of the hemA anaerobic conditions (Fig. 4B). When the anoxic culture

strain after a 60 min incubation with H2O2, Irr disappeared was subsequently exposed to air after 60 min, Irr disap-

(Fig. 3E). Thus, Irr turnover in response to H2O2 is haem- peared rapidly (Fig. 4B). It is likely that the incubation

dependent. This suggests that the need for iron is for the with iron for 60 min prior exposure to air accounts for the

synthesis of haem as shown previously (Qi and O Brian, more rapid turnover than in the completely aerobic exper-

2002) rather than for direct chemistry between iron and iment in Fig. 4A. The ndings show that Irr turnover

H2O2. requires O2.

Oxygen is required for iron-dependent Irr degradation The Irr protein is subject to oxidation in vitro

Irr degrades rapidly in wild-type cells in response to addi- The oxygen requirement and implication for H2O2 for Irr

tion of 6 M iron to the growth media without the addition degradation suggests that protein oxidation by ROS may

of exogenous H2O2 (Hamza et al., 1998; Qi et al., 1999) be involved in that process. Haem is required for Irr turn-

(Fig. 4A). Nevertheless, the current studies raise the pos- over in vivo, and thus we established an in vitro protein

sibility that ROS are normally involved in iron-dependent oxidation system (modi ed from Aft and Mueller, 1984)

Irr turnover. ROS are derived from the metabolism of comprised of puri ed recombinant Irr, haemin (ferric

molecular oxygen, and thus we addressed whether O2 haem) and the reducing agent DTT. Protein carbonylation

was required for Irr turnover. is considered a hallmark in ROS-catalysed protein oxida-

2006 The Authors

Journal compilation 2006 Blackwell Publishing Ltd, Molecular Microbiology, 60, 209 218

Irr degradation in response to oxidative stress 213

To assess the role of each reagent in the assay, we

A Air carried out the carbonylation assay omitting one or more

component (Fig. 6A). There was essentially no oxidized

*-**-**-**-**-** 90 105 120 min

Irr in the absence of haem or DTT. These observations

are consistent with the requirement for haem in vivo, and

also the need for both redox states of haem (Yang et al.,

2005). In further support of this, we found that zinc-

B protoporphyrin, which is a metal porphyrin containing zinc

N2 Air

rather than iron, cannot substitute for haem (iron-

*-**-**-**-**-** 90 105 120 min

protoporphyrin) in promoting carbonylation of Irr (Fig. 6A).

Unlike iron, which readily undergoes oxidation and reduc-

tion, zinc is very redox stable. Finally, removal of O2 from

the in vitro reaction by ushing with N2 resulted in very

Fig. 4. Iron-dependent Irr degradation under anaerobic conditions.

Cultures of parent strain I110 were aerobically grown to mid-log low carbonylation. This is consistent with an O2 require-

phase in low-iron media. Cultures were split and either continued

ment for degradation in vivo (Fig. 4). This weak signal may

under aerobic conditions (A) or ushed with N2 (B). 6 M FeCl3 was

be due to incomplete removal of O2.

added, and aliquots were taken at various time points and analysed

for Irr by immunoblotting. At 60 min, the anaerobic culture was It was shown previously that the form of iron to which

exposed to air (indicated by the arrow ) for another 60 min and Irr

Irr responds for degradation in vivo is haem, and that Irr

monitoring was continued.

does not respond to iron in cells that cannot synthesize

haem from it (Qi et al., 1999). In agreement with that, we

tion (Climent et al., 1989; Stadtman, 1993), thus protein found that ferric chloride (FeCl3) only slightly promoted

oxidation was assessed by detection of carbonyl groups carbonylation of Irr (Fig. 6B) even though ferric iron is able

(aldehydes and ketones) introduced into amino acid side to reduce O2 in the presence of a reductant (Iwai et al.,

chains. A clear increase in Irr protein oxidation was 1998). This indicates that haem is speci c for Irr degrada-

observed over a 4 h time-course (Fig. 5). At later time tion, presumably due to its ability to directly bind the pro-

points, oxidized products both larger and smaller than the tein and oxidize Irr locally.

Irr monomer were often observed on SDS-PAGE. These We found that H2O2 destabilizes the Irr protein in vivo

minor products are likely due to oxidative cleavage of the in the presence of iron, and that haem is needed for the

peptide backbone and cross-linking between reactive oxi- action of H2O2 (Fig. 3). To determine the effect of H2O2 on

dized protein (Stadtman, 1993; Berlett and Stadtman, Irr oxidation, we assessed protein oxidation using H2O2 as

1997). We found a low level of oxidized protein at time the oxidant in place of oxygen. The reaction was carried

zero in some recombinant Irr preparations. We speculate out in the presence of 1 mM H2O2 under anaerobic

that this occurred in E. coli or during protein preparation, conditions. Anaerobic incubation diminished Irr oxidation

where Irr could be exposed to haem. The total Irr protein (Fig. 6C, lane 2), and the low observed amount was prob-

level was fairly constant throughout the time-course as ably due to O2 contamination. The addition of H2O2

discerned by Coomassie staining, but some disappear- partially restored oxidation compared with the aerobic

ance was observed at later time points (Fig. 5), which was reaction (Fig. 6C, lanes 1 and 3), and also required haem

probably due to peptide bond cleavage. and DTT (data not shown). Thus, H2O2 can serve as an

Fig. 5. Irr is subject to oxidation in vitro. Protein

0 5-10-20-30-60-120 180 240 min

oxidation was carried out in a 50 l reaction

45

containing 30 g of puri ed recombinant Irr pro-

tein, 10 M haemin and 10 mM DTT in phos-

phate buffered saline. The reaction mixture was

31 ushed with air for 45 s and then incubated at

29 C for 4 h. 5 l aliquots were taken at the

21 indicated time and immediately subject to DNP

derivatization as described in Experimental pro-

Oxidized Irr

cedures. The products were electrophoresed

on a 10 20% SDS-PAGE gradient gel. Deriva-

tized DNP groups were detected by immunoblot

14.4

analysis using anti-DNP antibody (top panel)

and total proteins were detected by Coomassie

staining (bottom panel). The approximate sizes

of pre-stained protein marker were shown in the

Total Irr top panel.

2006 The Authors

Journal compilation 2006 Blackwell Publishing Ltd, Molecular Microbiology, 60, 209 218

214 J. Yang, H. R. Panek and M. R. O Brian

A truncated Irr protein is not oxidized in vitro and does not

A degrade in vivo

Haemin + + +

If Irr degradation is linked to its oxidation, an Irr derivative

DTT + + + +

that is not readily oxidized should be affected in its deg-

O2 + + + + +

radation. Irr(1 116) is a truncated protein that lacks the

Zn-proto +

C-terminal 47 amino acids; it binds ferric haem but not

ferrous haem. We showed previously that a GST fusion of

45

Irr(1 116) is stabilized in vivo and does not degrade in

response to either iron or haem, whereas the GST fusion

31 of full-length Irr degrades normally (Yang et al., 2005)

(Fig. 7A). Unlike the full-length protein, Irr(1 116) was not

21

Oxidized Irr

oxidized in vitro (Fig. 7B). The truncated protein retains

ferric haem binding, showing that haem binding alone is

not suf cient for in vitro oxidation. These results establish

a strong link between haem binding, protein oxidation and

Total Irr

degradation of Irr.

Lane 1 2 3 4 5 6

A

B C Irr 1-116

- -

Haemin + Fe H Fe H

O2 N2 N2 + H2O2

Fe +

Irr-GST

45

Endogenous Irr

31

B

21

Oxidized Irr

Irr 1-116

0 4 0 4h

Total Irr

Fig. 6. Requirements for haem and O2 for in vitro Irr oxidation.

A. Protein oxidation was carried out in a 10 l reaction containing Oxidized Protein

1 g of Irr protein and various combinations of 10 M haemin, 10 mM

DTT, 10 M zinc protoporphyrin and O2 as indicated in the gure.

Oxidized (top panel) or total (bottom panel) Irr is shown.

B. The oxidation reactions were carried out with Irr protein, DTT, O2,

and either 10 M haem or 10 M FeCl3.

C. Protein oxidation in the presence of H2O2. Oxidation reactions were

Total Protein

carried out aerobically (lane 1), anaerobically by ushing with N2

(lane 2) or anaerobically with 1 mM H2O2 (lane 3).

Fig. 7. A truncated Irr protein is not oxidized in vitro and does not

oxidant, which is consistent with its ability to promote

degrade in vivo.

degradation in vivo. Exogenous H2O2 oxidized Irr less well A. Effects of iron and haem on Irr-GST fusion proteins. Fusion

than O2. Metal-catalysed protein oxidation is viewed as a proteins were expressed under the control of native irr promoter

on plasmid pLAFR3 in parent strain I110. Cells were grown in

caged processes such that ROS are not released into

medium containing either low iron, 6 M FeCl3 (Fe) or 15 M

solution but react directly with the protein (Stadtman, haemin (H), and Irr-GST levels were detected by immunoblot analysis

1993). It is possible that ROS produced in the vicinity of using anti-GST antibody. Endogenous Irr was detected with anti-Irr

antibody.

haem binding is more ef cient for oxidation than that pro-

B. In vitro protein oxidation assay of Irr and a truncated Irr derivative

vided exogenously. Collectively, the in vivo degradation Irr(1 116) was carried out as described in a 10 l reaction containing

1 g of Irr protein, 10 M haemin and 10 mM DTT in phosphate

and in vitro oxidation data provide strong evidence that

buffered saline at 29 C for 4 h. Samples from 4 h were subject to

haem-mediated protein oxidation is involved in the regu-

DNP derivatization and analysed by immunoblot analysis for DNP

lated degradation of Irr. Furthermore, the results indicate groups (top panel) and Coomassie staining for total proteins (bottom

that Irr senses oxidative stress. panel).

2006 The Authors

Journal compilation 2006 Blackwell Publishing Ltd, Molecular Microbiology, 60, 209 218

Irr degradation in response to oxidative stress 215

Discussion protein 2 (IRP2) (Iwai et al., 1998). The underlying mech-

anisms are not fully understood in most cases. IRP2 is

In the present study, we show that Irr degrades in oxidized by iron and subsequently ubiquitinated and

response to H2O2 added to the external medium, or gen- degraded by the proteasome (Iwai et al., 1998; Yamanaka

erated intracellularly. These ndings suggest that Irr can et al., 2003). Furthermore, a role for haem in IRP2 deg-

mediate a cellular response to oxidative stress to regulate radation was demonstrated, but it has not been estab-

haem biosynthesis. This is mediated by the OxyR protein lished that haem catalyses oxidation of the protein

in E. coli and Salmonella (Zheng et al., 2001; Elgrably (Goessling et al., 1998; Bourdon et al., 2003; Yamanaka

Weiss et al., 2002), and by the PerR protein in B. subtilis et al., 2003). Thus, Irr degradation may be similar to that

(Chen et al., 1995). B. japonicum has an oxyR gene for IRP2 in several respects.

homologue, but its role in oxidative stress response may

differ from other organisms (Panek and O Brian, 2004).

Experimental procedures

Like PerR, Irr belongs to the Fur superfamily of regulatory

proteins. However, the proteins appear to function differ- Bacterial strains, media and growth

ently. PerR activity is controlled by the oxidation state of

Bradyrhizobium japonicum I110 is the parent strain used in

histidine residues that co-ordinate Fe2+ to serve as a

the present work. MLG1 is a hemA deletion strain (Guerinot

regulatory switch (Lee and Helmann, 2006), whereas oxi-

and Chelm, 1986) and was grown in the presence of

dation of Irr leads to its destruction. In addition, Mn2+ or 50 g ml 1 kanamycin. The disruption of katG in strain I110

Fe2+ serve as co-repressors of PerR, thus metal is has been described previously (Panek and O Brian, 2004)

and the mutant cells were grown in the presence of 50 g

required for activity. In contrast, the status of Irr is depen-

ml 1 streptomycin and 100 g ml 1 spectinomycin. As the

dent on iron, but it is indirect and leads to degradation

katG strain cannot be aerobically cultured from low density,

rather than activation.

the mutant cells were routinely grown to an OD540 of 0.05 in

Degradation of Irr requires direct binding of haem to the

1% O2 and then grown aerobically to an OD540 of 0.3. For

protein (Qi et al., 1999; Qi and O Brian, 2002). Here, we

growth under restricted aeration, asks containing inoculated

show that haem catalyses the oxidation of Irr, which we media were sealed with rubber stoppers and ushed with N2;

suggest triggers its degradation. Iron-dependent turnover then, O2 was injected to the desired concentration. Strains

requires O2 in vivo (Fig. 4), and a truncated Irr protein I110(pIrr-GST) and I110(pIrr1-116-GST) contain plasmids

that express GST fusion proteins of Irr derivatives, in the

lacking its ferrous haem binding site did not degrade in

latter case a C-terminal truncation of Irr (Yang et al., 2005).

vivo, nor was it oxidized in vitro (Fig. 7). Thus, we propose

The plasmid-carrying strains were grown in the presence of

that ROS participate in Irr degradation not only as part of

50 g ml 1 tetracycline. B. japonicum strains were routinely

an oxidative stress response, but also in the normal deg-

grown at 29 C in GSY medium (Frustaci et al., 1991) with

radation in response to iron. Protein oxidation can result antibiotics. Modi ed GSY medium, containing 0.5 g l 1 yeast

in hydrolysis of peptide bonds (Berlett and Stadtman, extract and no exogenous iron source, was used for low-iron

1997) and thus, in principle, oxidation of Irr could be conditions. The actual iron concentration of the medium was

0.3 M as described previously (Hamza et al., 1998). The

suf cient for degradation. However, in vivo degradation of

media were supplemented with various concentrations of

Irr is rapid whereas carbonylation in vitro is slow. It is

FeCl3 for high-iron conditions. Cultures were harvested at an

probable that oxidized Irr is recognized by cellular pro-

optical density at 540 nm of 0.3 0.4.

teases as a damaged protein that is subsequently

degraded. We have not identi ed a candidate protease

thus far. Also, oxidation of Irr by H2O2 in vitro required a Immunoblot analysis

rather high concentration (Fig. 6C), and thus the in vitro

To examine the steady state levels of ALAD, Irr and GroEL

experiments may not completely recapitulate the cellular

by immunoblot analysis, cultures were harvested and washed

situation. Some H2O2 may get further reduced in solution

with 50 mM NaPO4 (pH 7.0). Total protein concentrations

that does not participate in protein oxidation. In addition, were determined by the BCA protein assay (Pierce, Rockford,

the physiological reductant needed may be more ef cient IL) and cells containing 30 g of total protein were boiled in

than DTT. H2O2 in cells may get effectively reduced to loading buffer and electrophoresed through 12% SDS-

more reactive forms, which react with Irr. polyacrylamide gels. Immunoblotting was carried out with

ALAD (Chauhan and O Brian, 1995), Irr (Hamza et al., 1998)

In addition to Irr, there are several other examples of

or GroEL (StressGen, Vancouver, Canada) polyclonal anti-

metal-induced protein degradation to regulate metal

bodies and HRP-conjugated goat anti-rabbit IgG (Jackson

homeostasis. These include the bacterial copper chaper-

Immunoresearch, West Grove, PA) with detection using the

one CopZ (Lu and Solioz, 2001), the yeast zinc trans- Western Lightning system (Perkin Elmer, Boston, MA). Poly-

porter Zrt1, the copper transporter Ctr1, and the clonal antibodies against GST (Zymed Laboratories, San

transcriptional factor Mac1 (Ooi et al., 1996; Gitan et al., Francisco, CA) were used in immunoblots to determine the

1998; Zhu et al., 1998) and the mammalian iron regulatory plasmid-borne GST fusion protein levels.

2006 The Authors

Journal compilation 2006 Blackwell Publishing Ltd, Molecular Microbiology, 60, 209 218

216 J. Yang, H. R. Panek and M. R. O Brian

Analysis of steady state levels of irr mRNA Overexpression and puri cation of recombinant Irr

proteins

Analysis of the irr mRNA levels was carried out by ribonu-

clease protection assay. Total RNA was prepared from cul- Recombinant proteins Irr (Qi et al., 1999) and the C-terminal

tured cells as described previously (Chauhan and O Brian, truncation Irr(1 116) (Yang et al., 2005) were overexpressed

1997). pRPA2 (Hamza et al., 1998) was linearized with EcoRI from plasmid pCYB1

Contact this candidate