Supporting Materials and Methods
MSG Samples. For mostof the NOE spectra used to generate distance restraints U-[15N,2H]- and U-[15N,2H], Iled 1-[13CH3], Leu,Val-[13CH3,12CD3]-labeled samples of malate synthase G (MSG) were used. U-[15N,2H]-labeled MSG was unfolded in H2O and refolded to protonate all amide positions buried in the protein core as described (1); the final sample was 0.9 mM in protein, 92% H2O/8% D2O. The U-[15N,2H], Iled 1-[13CH3], Leu,Val-[13CH3,12CD3]-labeled sample of MSG (0.9 mM in protein) was placed initially in 99.9% D2O for acquisition of methyl–methyl (1H-13C) NOE spectra (2). Subsequently, this sample was refolded in H2O to completely protonate amide sites (1) and dissolved in 92% H2O/8% D2O for the acquisition of HN–methyl NOE spectra. All NMR samples contained 25 mM sodium phosphate buffer (pH 7.1), 20 mM MgCl2, 5 mM DTT, and 0.05% NaN3.
Structure Calculations. Initially, a high-temperature torsion angle dynamics (TAD) phase was performed at 50,000 K, consisting of 1,000 molecular-dynamics steps, each of 15 fs. During this stage, all of the force constants were kept constant [Ktor = 100 kcal/rad2, KNOE = 150 kcal/Å2, KDIP = 0.03 kcal/(Hz)2, KCSA = 0.003 kcal/(ppb)2]. Subsequently, a cooling TAD phase composed of 180,000 steps, each of 5 fs, was used with the temperature decreasing from 50,000 to 2,000 K. KDIP and KCSA were scaled exponentially to their final values of 0.6 kcal/(Hz)2 and 0.06 kcal/(ppb)2, respectively, during this interval, with Ktor = 200 kcal/rad2, KNOE = 150 kcal/Å2. Last, a second cooling phase using Cartesian dynamics (10,000 steps of 2 fs each) was performed starting from 2,000–300 K, with all force constants at their final values with the exception of Ktor, which was doubled. The final cooling stage was followed by Powell minimization of 3,000 steps (Ktor = 1200 kcal/rad2, KNOE = 70 kcal/Å2, and other constants as before). NOE distance restraints involving methyl groups were averaged by means of an r–6 weighting. All structure calculations used a radius of gyration potential (3), with RG set at 26 Å and a force constant of 50 kcal/(A)2. The RG value was estimated on the basis of hydrodynamic radius (Rh) measurements (in the range of 33-38Å) calculated from dynamic light-scattering experiments performed on MSG samples with protein concentrations ranging from 0.2–0.4 mM and the same buffer composition as used for NMR data collection. A molecular radius of gyration (RG) can be readily estimated from these measurements using the relation RG = 0.774*Rh. We have used the lowest possible value of Rh (33 Å) because this takes into account its potential overestimation from slight deviations of the molecule from an ideal spherical shape. The calculation protocol is available from the authors upon request.
1. Tugarinov, V., Muhandiram, R., Ayed, A. & Kay, L. E. (2002) J. Am. Chem. Soc. 124, 10025–10035.
2. Tugarinov, V., Kay, L. E., Ibraghimov, I. & Orekhov, V. Y. (2004) J. Am. Chem. Soc., in press.
3. Kuszewski, J., Gronenborn, A. M. & Clore, G. M. (1999) J. Am. Chem. Soc. 121, 2337–2338.