Nucleotide Spin Labeling for ESR Spectroscopy of ATP-Binding Proteins

Abstract

Site-directed spin labeling of proteins by chemical modification of engineered cysteine residues with the molecule MTSSL (1-Oxyl-2,2,5,5-tetramethylpyrroline-3-methyl methanethiosulfonate) has been an invaluable tool for conducting double electron electron resonance (DEER) spectroscopy experiments. However, this method is generally limited to recombinant proteins with a limited number of reactive Cys residues that when modified will not impair protein function. Here we present a method that allows for spin-labeling of protein nucleotide binding sites by adenosine diphosphate (ADP) modified with a nitroxide moiety on the β-phosphate (ADP-β-S-SL). The synthesis of this ADP analog is straightforward and isolation of pure product is readily achieved on a standard reverse-phase high-performance liquid chromatography (HPLC) system. Furthermore, analyses of isolated ADP-β-S-SL by LC-mass spectrometry confirm that the molecule is extremely stable under ambient conditions. The crystal structure of ADP-β-S-SL bound to the ATP pocket of the histidine kinase CheA reveals specific targeting of the probe, whose nitroxide moiety is mobile on the protein surface. Continuous wave and pulsed ESR measurements demonstrate the capability of ADP-β-S-SL to report on active site environment and provide reliable DEER distance constraints.

Fig. 1.

The reaction scheme for the synthesis of ADP-β-S-SL. ADP-β-S reacts with MTSSL in a 1:1 stoichiometric ratio to form an ADP analog with a nitroxide spin-label linked to the β-phosphate via a disulfide bond

Fig. 2.

Isolation and decomposition of ADP-β-S-SL analyzed by LC-MS. a A reaction mixture consisting of ADP-β-S-SL and MTSSL reveals the presence of the desired product, ADP-β-SSL. The precursors and product are well separated using a reverse-phase HPLC system. b Isolated ADP-β-S-SL was analyzed by LC-MS after a freeze-thaw cycle and incubation at 25°C for 24 hours. (Top) Three degradation products are identified; a reduced form of ADP-β-S-SL is the most abundant. (Bottom) Extracted ion chromatograms of each molecule present in the degraded sample. For LCMS experiments, reaction species were identified by either positive or negative ion detection modes

Fig. 3.

ADP-β-S-SL binds to the histidine kinase CheA from T. martima. a Competitive inhibition assays with ATP (25 μM) and ADP-β-S-SL demonstrate that ADP-β-S-SL binds to the nucleotide pocket of CheA with a Ki of 29 ± 6 μM. b cw-ESR of ADP-β-S-SL (100 μM) in the absence and presence of CheA (100 μM) after washing suggest that the nitroxide label is relatively mobile when bound to CheA. cw-ESR of ADP-β-S-SL bound to the N. crassa kinase CK1 and RNA helicase FRH, two other ATP binding proteins, reveal a more rigid nitroxides, in these cases. c The crystal structure of the CheA nucleotide-binding domain (P4) reconstituted with ADP-β-S-SL. The Fo-Fc omit map shown at σ = 1.3 (green contours) reveals density for the nucleotide and nitroxide moieties, although the nitroxide appears conformationally disordered to some degree. d The CheA[ADP-β-S-SL] structure predicts a ~35 Å distance between the nucleotide and residue 387.

Fig. 4.

DEER measurements with nitroxide-labeled CheA and ADP-β-S-SL. a Dimeric CheA labeled with MTSSL on each P4 subunit at position E387C produces a distance distribution with a broad peak centered at 45 Å. b When ADP-β-S-SL is reconstituted into the E387C-SL protein, the distribution become bimodal with a second peak centered at 35 Å. c Upon addition of excess ATP, ADP-β-S-SL is competed from the nucleotide pocket and the 35 Å peak is lost. d Time domain data for the three samples shown above after baseline correction. Inset shows DEER data before baseline subtraction to 2 μsec after the spectra were aligned for comparison purposes. Note that sample c differs from a in that the spin-label is free from the protein for c. Uncertainty in the SVD of the distance distributions are shown in red