Sub-micromolar affinity of Escherichia coli NikR for Ni(II)w Rutger E. M. Diederix,* ab Caroline Fauquant, a Agne`s Rodrigue, c Marie-Andre´e Mandrand-Berthelot c and Isabelle Michaud-Soret* a Received (in Cambridge, UK) 20th December 2007, Accepted 21st January 2008 First published as an Advance Article on the web 18th February 2008 DOI: 10.1039/b719676h The dissociation constant of Ni(II) for Escherichia coli NikR was determined using three independent techniques, including binding kinetics, and shown to be in the sub-micromolar range. NikR is a Ni(II)-responsive transcription factor that regulates levels of Ni(II) in Escherichia coli and other bacteria such as Helicobacter pylori. 1 In E. coli, NikR is involved in nickel homeostasis through Ni(II)-dependent repression of the nik- ABCDE operon, thus suppressing expression of the ABC transporter specific for the import of Ni(II). E. coli NikR consists of 133 amino acids, of which 12 are histidines, and forms a tetrameric structure with several Ni(II)-binding sites. 2–5 One of these sites has square planar His 3 Cys co- ordination 3,6 and is buried at the interface between the two dimer pairs making up the tetramer. 2 A picomolar dissociation constant has previously been ascribed for this high affinity (HA) site, using competition assays. 7 This report presents a re- evaluation of the Ni(II) binding properties of E. coli NikR using direct measurements: UV-vis equilibrium titrations, filter binding assays, and rapid kinetics experiments. Ni(II) binding to the HA site activates NikR for operator binding (K D B30 nM). Ni(II) binds to several other (low affinity, LA) sites, which may number up to 7 per monomer. 8 These are less well characterized, and have a lower affinity (0.03–50 mM). 7 Ni(II) binding to (one of the) LA sites increases the affinity of NikR for DNA about 1000-fold (K D B10–20 pM). 7 This implies that NikR is sensitive or responsive to two greatly different Ni(II) concentrations, and thus acts at two distinct levels of regulatory control. This feature of NikR is, to our knowledge, unique for a metallo-regulatory protein. The cellular levels of Ni(II) in living E. coli cells have been estimated to lie between 10 8 and 10 6 M, 9 and the experi- mentally determined K D lies in the low mM range for many bacterial Ni(II)-binding proteins. 10 An apparent inconsistency thus exists between the pM affinity of NikR for Ni(II) on the one hand, and cellular nickel levels and affinities of Ni(II) binding proteins on the other hand. This clearly warrants further investigation, which is reported here. We have chosen for direct titration techniques to study Ni(II) binding by E. coli NikR, rather than assays based on chelator–protein competi- tion, which may lead to errors in interpretation (vide infra). We find that the HA site of NikR certainly does not have pM affinity for Ni(II), but rather an affinity four orders of magni- tude lower, consistent with NikR being responsive to sub- micromolar levels of nickel. Similarly, in a recent study on H. pylori NikR, 11 the affinity for Ni(II) as determined by a direct assay (isothermal titration calorimetry) was three orders of magnitude lower than previously determined by competi- tion assays suggesting consistent overestimation of NikR binding strength with indirect measurements using EGTA. 12 NikRz binds Ni(II) in the HA site with square planar geometry and with cysteine ligation, 2,6 as observed by UV- vis spectroscopy. 4,8 Increase of the Cys S g - Ni LMCT band at 302 nm (Fig. 1a) is linear up to one Ni(II) equivalent, after which mass aggregation is seen, which is complete at less than 1.5 equivalents of Ni(II). 8 Binding of Ni(II) is still possible by the aggregate however, between 4 and 7 Ni(II) ions per NikR monomer. 8 No aggregation at all is observed with 0.1 or less equivalents of Ni(II) using dynamic light scattering (see ESIw). In general, when the concentration of Ni(II) is much lower than that of NikR, Ni(II) binding is approximately linear. The proportion of bound Ni(II) is expected to decrease as the [NikR] of the experiment is lowered and nears the actual K D of the binding site. As justified in the ESI,w this leads to the Fig. 1 (a) UV-Vis difference spectrum of 215 mM NikR in the presence and absence of circa 20 mM NiSO 4 . (b) Dependence of the linear absorbance increase at 302 nm of Ni(II)-bound NikR as a function of Q2E NikR concentration (K). The data were fitted (line) to a hyperbolic function (eqn (1)). The plateau value corresponds to the extinction coefficient. For illustrative purposes, the expected dependence for binding with picomolar affinity (K D = 10 pM) is shown (dashed line). a CNRS, UMR 5249, 17 avenue des martyrs, Grenoble F-38054 cedex 9, CEA, DSV, iRTSV, Laboratoire de Chimie et Biologie des Me ´taux, Grenoble F-38054 cedex 9 Universite ´ Joseph Fourier, Grenoble F-38000, France. E-mail: imichaud@cea.fr; Fax: +33 4 38 78 54 87 b Departamento de Biofı´sica, Instituto de Quı´mica Fı´sica ‘‘Rocasolano’’, CSIC, Serrano 119, Madrid 28006, Spain. E-mail: diederix@iqfr.csic.es; Fax: +34 91 5642431 c Microbiologie, Adaptation et Pathoge ´nie, CNRS-UCB-INSA- BayerCropScience, Universite ´ Lyon 1, Villeurbanne, France w Electronic supplementary information (ESI) available: Additional experiments on the protein aggregation state, experimental details of binding assays and stopped-flow binding kinetics, and details on data fitting procedures and binding models used. See DOI: 10.1039/ b719676h This journal is  c The Royal Society of Chemistry 2008 Chem. Commun., 2008, 1813–1815 | 1813 COMMUNICATION www.rsc.org/chemcomm | ChemComm