FEMS Microbiology Letters 113 (1993) 219-222 © 1993 Federation of European Microbiological Societies 0378-1097/93/$06.00 Published by Elsevier

219

FEMSLE 05658

Activation of FNR-dependent transcription by iron: An in vitro switch for FNR

Je f f r ey G r e e n and J o h n R. G u e s t *

The Krebs Institute for Biomolecular Research, Department of Molecular Biology and Biotechnology, University of Sheffield, Western Bank, Sheffield SIO 2UH, UK

(Received 12 July 1993; accepted 10 August 1993)

Abstract: FNR is an iron-binding transcriptional regulator of Escherichia coli which controls the expression of target genes in response to anaerobiosis. The mechanism used by FNR to sense and respond to anaerobiosis is unknown but it is thought to involve iron. In vitro transcription analyses have shown that iron-deficient FNR is unable to activate transcription from the FF-melR promoter, but activity could be restored by preincubation with Fe 2+ and /3-mercaptoethanol. The reactivation of FNR was prevented and reversed by chelating agents, and this reactivation thus provides an artificial in vitro switch for FNR-dependent transcriptional activation.

Key words: FNR; Transcriptional regulation; Iron; Transcription, in vitro; Redox-sensing; Escherichia coli

Introduction

The FNR protein of Escherichia coli is an iron-binding transcriptional regulator which is structurally related to the cAMP receptor protein (CRP) or catabolite gene activator protein (CAP). FNR is predicted to contain all of the secondary structural elements of CRP. However, it differs in being monomeric rather than dimeric, and in possessing a cysteine rich N-terminal extension which is essential for the regulation of FNR-de- pendent genes in response to anaerobiosis (for reviews see [1,2]).

The mechanism for sensing and responding to anaerobiosis has not been defined but there is

* Corresponding author.

indirect evidence that iron is involved and that the N-terminal cysteine cluster contributes to an iron-binding redox-sensing domain. For example, isolated FNR can contain up to one iron atom per monomer [3] and the FNR-dependent activa- tion or repression of target genes is abolished by deleting residues 3-30 [4] and by site-directed substitiution of four of the five cysteine residues in FNR (C-20, C-23, C-29 and C-122, but not C-16) [4-6]. It has also been observed that chelat- ing agents mimic oxygen in preventing the anaer- obic activation or repression of FNR-regulated genes in vivo [7] and by increasing the rate of FNR cysteine carboxymethylation in perme- abilised cells [8]; effects which can be prevented by adding Fe E+ . The recent identification of an oxidised form of FNR (FNR27)which contains a disulphide bond linking C-122 to one of the N-

220

terminal cysteine residues, and the greater abun- dance of FNR27 in aerobic bacteria, further sup- port the existence of the putative cysteine-rich redox-sensing domain [9].

It has been shown that FNR responds to redox potential rather than to oxygen [10] and that FNR is reversibly inactivated and reactivated in vivo depending upon the iron and oxygen status of the environment [11]. It has also been shown that a high iron content is not essential for site- specific binding of FNR to target DNA although its presence increases the affinity up to two-fold [12]. In contrast, iron is essential for FNR-depen- dent open complex formation, both in vivo and in vitro, and it has thus been concluded that tran- scriptional activation by FNR is mediated by a ferrous iron cofaetor [13].

Here, studies with the FF-melR promoter have shown that iron is essential for FNR-dependent transcriptional activation in vitro. Transcription was not activated by iron-deficient FNR but in vitro activity could be restored by incubation with Fe z+ and 3-mercaptoethanol.

Materials and Methods

Plasmids, FNR and other materials The source of promoter DNA used in this

study was the plasmid pGS422, which contains the promoter region of an FNR-dependent melR-lacZ fusion, FF-melR (kindly provided by Dr. S. Busby, University of Birmingham), in a 343 bp EcoRI-HindIII fragment [12]. Standard proce- dures were used for the isolation and manipula- tion of DNA. Iron-deficient FNR protein (0.1 tool iron per mol of FNR monomer) was purified from aerobically grown E. coli JM101 (pGS330) in the absence of added iron, as described previ- ously [3]. RNA polymerase saturated with ~r 7° was obtained from Pharmacia, other enzymes from NBL, a-[35S]-dATP and a-[35S]-UTP from Amersham International, and the metal-ion bind- ing hexapeptide (KCTCCA) from Sigma.

Reactivation of iron-deficient FNR Samples of iron-deficient FNR (600 pmol in

200 txl of 50 mM Tris-HC1, pH 8.0) were preincu-

bated at 4°C for 6 h in the presence or absence of ferrous iron, 0.1 mM (NH4)2Fe(SO4)2, and dif- ferent reducing agents: /3-mercaptoethanol (200 raM); ascorbate (10 mM); dithiothreitol (100 raM); reduced glutathione (100 mM); prior to testing for the in vitro activation of FF-melR transcrip- tion. In some experiments, chelating agents were added either during the reactivation procedure, or after reactivation by treating for 2 h at 0°C, using ferrozine (5 mM), bipyridyl (2 raM), bathophenanthroline (2 raM), or KCTCCA (4 mM), prior to in vitro transcription analysis.

In vitro transcription studies The in vitro transcription analyses were per-

formed with the 343 bp EcoRI-HindIII FF-melR promoter fragment of pGS422 as the DNA tem- plate (0.1 pmol approx.) in reaction mixtures (20 /xl) containing: Tris-HC1 (40 mM) pH 7.5; MgCI 2 (6 mM); spermidine (2 mM); NaCl (10 raM) and dithiothreitol (2 mM). FNR (30 pmol) was added when required and, after preincubation for 5 min at 37°C, E. coli o-7°-saturated RNA polymerase (3.5 pmol) was added and the incubation contin- ued for a further 5 min. Transcription was initi- ated by the addition of ATP (0.5 mM), GTP (0.5 mM), CTP (0.5 mM), UTP (2.5/xM) and a-[35S] - UTP (0.5 ~1, > 1000 Ci/mmol), and allowed to proceed for 60 min at 37°C. After phenol extrac- tion, the nucleic acids were precipitated with ethanol, washed twice with 70% ethanol, resus- pended in 4 /xl loading buffer (95% by volume formamide, 20 mM EDTA, 0.05% bromophenol blue, 0.05% xylene cyanol FF) and the entire sample was analysed autoradiographically after fractionation in polyacrylamide-urea gel. Tran- scripts were sized by comparison with a DNA ladder generated by DMS-piperidine hydrolysis of end-labelled FF-melR promoter fragment [3,141.

Results and Discussion

Restoration of transcriptional activation activity to iron-deficient FNR

It has been shown previously that FNR with a high iron content can activate transcription from

221

[El---*

Isl---,

1 2 3 4 5 6

Fig. 1. In vitro transcription from the FF-melR promoter. Samples of iron-deficient FNR protein were preincubated under different conditions and then tested for in vitro tran- scriptional activation of the FF-melR promoter (see Materials and Methods). The preincubation reactions contained, no FNR (lane 1), iron-deficient FNR alone (lane 2), or iron-defi- cient FNR with: Fe 2+ and ascorbate (lane 3); /3- mercaptoethanol (lane 4); Fe z + and/3-mercaptoethanol (lane 5); Fe 2+, /3-mercaptoethanol and ferrozine (lane 6). The run-off transcripts, specifically initiated at the FF-melR pro- moter (S, 266 nt) or generated by RNA polymerase binding at

the end of the template (E, 343 nt), are arrowed.

the FF-melR promoter in vitro [12] and that iron is essential for the formation of the open complex at this promoter both in vivo and in vitro [13]. Further studies have now shown that iron-defi- cient F N R (0.1 tool iron per mol of FNR monomer) is almost incapable of activating tran- scription from the FF-melR promoter in vitro, but activity can be restored reproducibly by preincubating the iron-deficient protein with Fe 2 + (0.1 mM) in the presence of 200 mM /3- mercaptoethanol (Fig. 1). The requirement for /3-mercaptoethanol could not be satisfied by ascorbate or reduced glutathione, but dithiothre- itol was partially effective. It was also observed that Fe z+ (0.1 mM) could not be replaced by Mn 2+ (1 mM). Ferric iron (0.1 mM) was effective but this is probably because 50% of the Fe 3+ is converted to Fe z+ during the preincubation with /3-mercaptoethanol. Fur ther studies have shown that the reactivation of iron-deficient F N R could be prevented or reversed by the addition of chelating agents such as ferrozine, bipyridyl,

ba thophenan th ro l i ne , and the hexapep t ide KCTCCA, during or after the reactivation proce- dure (see Materials and Methods).

Quantitiative densitometric analyses with the autoradiographs indicated that reactivated FNR stimulates FF-melR transcription some six- to ten-fold relative to the level observed with iron- deficient FNR. This parallels the increase in pro- moter activity observed in vivo with an FF-melR- lacZ fusion during the aerobic to anaerobic tran- sition (seven- to nine-fold), and the increase in KMnOa-sensitivity of bases in the - 10 region of the FF-melR promoter observed with high-iron F N R relative to iron-deficient FNR (three- to six-fold) [13].

An interesting feature of the in vitro transcrip- tion studies is the severe reduction in the amount of end-transcript formed in the presence of FNR, even with protein that is incapable of activating transcription from the FF-melR promoter (Fig. 1). Since a high iron content is not essential for site-specific D N A binding by F N R [3,12], it is tempting to speculate that the F N R protein is targeting R N A polymerase to the FF-melR pro- moter and thus inhibiting formation of the end- transcript. It is conceivable that the inactive ternary complex containing FNR, R N A poly- merase and promoter DNA, could be activated by Fe e+ or the reductive incorporation of Fe 2+. A precedent for act ivator-dependent preinduc- tion binding of R N A polymerase exists for the Hg2+-regulated met promoter [15] and it could be relevant to the mode of action of FNR.

Although the nature of the switch mediating the in vivo act ivat ion/inact ivat ion of FNR-regu- lated transcription has yet to be defined, there is now good evidence that iron performs a central role [3,7,8,13]. The major system responsible for regulating iron-dependent processes is the iron- responsive repressor or ferric uptake regulator, Fur. However, recent in vivo studies on the ef- fects of iron limitation on the expression of aero- bic and anaerobic electron transport genes in E. coli have indicated that there may be another iron-responsive regulatory mechanism, which could involve F N R [16]. The present results would be consistent with a dual role for FNR in iron- and redox-responsive regulation.

222

Possible mechanisms for the redox-switching of FNR between inactive (aerobic) and active (anaerobic) forms could involve the reduction of a loosely-bound iron cofactor or the reversible binding of Fe 2+, with or without the participation of a thiol-dithiol transition. In either case, the affinity for Fe 3+ may be lower than for Fe 2÷, and this is consistent with the observed loss of iron during the purification of FNR under aerobic conditions. The present results do not discrimi- nate between these mechanisms but they do add significantly to our understanding of the role of iron in the mode of action of FNR. In particular, they provide the first report of the in vitro activa- tion and inactivation of FNR (with Fe 2÷ plus /3-mercaptoethanol and with chelating agents, re- spectively) which in essence represents an in vitro switch for FNR.

Acknowledgement

This work was supported by the Science and Engineering Research Council.

References

1 Spiro, S. and Guest, J.R. (1990) FNR and its role in oxygen-regulated gene expression in Escherichia coli. FEMS Microbiol. Rev. 75, 399-428.

2 Unden, G. and Trageser, M. (1991) Oxygen regulated gene expression in Escherichia coli: Control of anaerobic respi- ration by the FNR protein. Antonie van Leeuwenhoek 59, 65-76.

3 Green, J., Trageser, M., Six, S., Unden, G. and Guest, J.R. (1991) Characterization of the FNR protein of Escherichia coli, an iron-binding transcriptional regulator. Proc. Roy. Soc. B 244, 137-144.

4 Spiro, S. and Guest, J.R. (1988) Inactivation of the FNR protein of Escherichia coli by targeted mutagenesis in the N-terminal region. Mol. Microbiol. 2, 701-707.

5 Sharrocks, A.D., Green, J. and Guest, J.R. (1990) In vivo and in vitro mutants of FNR, the anaerobic transcriptional regulator of E. coli. FEBS Lett. 270, 119-122.

6 Melville, S.B. and Gunsalus, R.B. (1990) Mutations in fnr that alter anaerobic regulation of electron transport-asso- ciated genes in Escherichia coli. J. Biol. Chem. 265, 18733-18736.

7 Spiro, S., Roberts, R.E. and Guest, J.R. (1989) FNR-de- pendent repression of the ndh gene of Escherichia coli and metal ion requirement for FNR-regulated gene ex- pression. Mol. Microbiol. 3, 601-608.

8 Trageser, M. and Unden, G. (1989) Role of cysteine residues and of metal ions in the regulatory functioning of FNR, the transcriptional regulator of anaerobic respira- tion in Escherichia coli. Mol. Microbiol. 3, 593-599.

9 Green, J., Sharrocks, A.D., Green, B., Geisow, M. and Guest, J.R. (1993) Properties of FNR proteins substituted at each of the five cysteine residues. Mol. Microbiol. 8, 61-68.

10 Unden, G., Trageser, M. and Duchene, A. (1990) Effect of positive redox potentials ( > + 400 mV) on the expression of anaerobic respiratory enzymes of Escherichia coli. Mol. Microbiol. 4, 315-319

11 Engel, P., Trageser, M. and Unden, G. (1991) Reversible interconversion of the functional state of the gene regula- tor FNR from Escherichia coli in vivo by 0 2 and iron availability. Arch. Microbiol. 156, 463-470

12 Sharrocks, A.D., Green, J. and Guest, J.R. (1991) FNR activates and represses transcription in vitro. Proc. Roy. Soc. B 245, 219-226.

13 Green, J. and Guest, J.R. (1993) A role for iron in tran- scriptional activation by FNR. FEBS Lett. 329, 55-58.

14 Davies, R.W. (1988) DNA sequencing. In: Gel Elec- trophoresis of Nucleic Acids (Rickwood, D. and Hames, B.D., Eds.), pp. 164-165. IRL Press, Oxford.

15 Heltzel, A., Lee, I.W., Totis, P.A. and Summers A.O. (1990) Activator-dependent preinduction binding of ~70 RNA polymerase at the metal-regulated mer promoter. Biochem. 29, 9572-9584.

16 Cotter, P.A., Darie, S. and Gunsalus, R.P. (1992) The effect of iron limitation on the expression of the aerobic and anaerobic electron transport pathway genes in Es- cherichia coli. FEMS Microbiol. Lett. 100, 227-232.