Observation of Slow Dynamic Exchange Processes in Ras Protein Crystals by 31P Solid State NMR Spectroscopy

doi:10.1016/S0022-2836(02)01010-0 available online at http://www.idealibrary.com on

w B

J. Mol. Biol. (2002) 323, 899–907

Observation of Slow Dynamic Exchange Processes in Ras Protein Crystals by 31P Solid State NMR Spectroscopy Michael Stumber1, Matthias Geyer1*, Robert Graf2 Hans Robert Kalbitzer3, Klaus Scheffzek4 and Ulrich Haeberlen1 1

Max-Planck-Institut fu¨r medizinische Forschung Jahnstrasse 29, 69120 Heidelberg, Germany 2 Max-Planck-Institut fu¨r Polymerforschung Ackermannweg 10, 55128 Mainz, Germany 3

Universita¨t Regensburg Institut fu¨r Biophysik und Physikalische Biochemie Postfach, 93040 Regensburg Germany 4

European Molecular Biology Laboratory, Structural and Computational Biology Programme, Meyerhofstraße 1 69117 Heidelberg, Germany

The folding, structure and biological function of many proteins are inherently dynamic properties of the protein molecule. Often, the respective molecular processes are preserved upon protein crystallization, leading, in X-ray diffraction experiments, to a blurring of the electron density map and reducing the resolution of the derived structure. Nuclear magnetic resonance (NMR) is known to be an alternative method to study molecular structure and dynamics. We designed and built a probe for phosphorus solid state NMR that allows for the first time to study static properties as well as dynamic processes in single-crystals of a protein by NMR spectroscopy. The sensitivity achieved is sufficient to detect the NMR signal from individual phosphorus sites in a 0.3 mm3 size singlecrystal of GTPase Ras bound to the nucleotide GppNHp, that is, the signal from approximately 1015 phosphorus nuclei. The NMR spectra obtained are discussed in terms of the conformational variability of the active center of the Ras – nucleotide complex. We conclude that, in the crystal, the protein complex exists in three different conformations. Magic angle spinning (MAS) NMR spectra of a powder sample of Ras – GppNHp show a splitting of one of the phosphate resonances and thus confirm this conclusion. The MAS spectra provide, furthermore, evidence of a slow, temperature-dependent dynamic exchange process in the Ras protein crystal. q 2002 Elsevier Science Ltd. All rights reserved

*Corresponding author

Keywords: solid state NMR; MAS NMR; protein dynamics; dynamics in crystals; Ras

Introduction Guanine nucleotide-binding proteins of the superfamily of small GTPases play a central role in cell signalling events. They function as molecular switches cycling between guanosine triphosphate (GTP)-bound “on” and guanosine diphosphate (GDP)-bound “off” states.1 In the GTP-bound state, they interact with effector proteins which trigger a variety of downstream events. Present address: M. Geyer, Max-Planck-Institut fu¨r molekulare Physiologie, Abteilung Physikalische Biochemie, Otto-Hahn-Str. 11, 44227 Dortmund, Germany. Abbreviations used: MAS NMR, magic angle spinning NMR. E-mail address of the corresponding author: [email protected]

The most prominent member of this family is the Ras protein, which relays signals from cell surface receptors to the nucleus to stimulate cell proliferation and differentiation.2 Specific mutations of Ras leading to inhibition of GTP hydrolysis or to acceleration of GDP release are involved in about 30% of human cancers.3 Mutation to oncogenic Ras is the most frequently occurring gain-offunction alteration detected in human tumors.4 The structure of Ras has been determined by X-ray crystallography5 – 7 and NMR spectroscopy8,9 in its diphosphate (GDP) and triphosphate (GppNHp and GppCH2p, two slowly hydrolizing GTP analogues, and also with GTP) bound forms. The structures obtained indicate that two regions, called switch 1 (effector loop, residues 30 –38) and switch 2 (loop L4, residues 60 –76), are crucial to the functioning of Ras as a signalling molecule. Both switch regions are close to the bound

0022-2836/02/$ - see front matter q 2002 Elsevier Science Ltd. All rights reserved

900

nucleotide, sensing the presence or absence of a g-phosphate group. The term “switch regions” reflects the capability of these regions to adopt different conformations depending on the type of the bound nucleotide.5 Besides the nucleotide-dependent conformational changes which have been observed elegantly by time-resolved X-ray crystallography,10 reversible dynamic exchange processes were detected in the GppNHp-bound state of Ras in solution by 31P NMR spectroscopy.11 Two distinct, almost equally populated conformational states were identified that exchange on the timescale of milliseconds with an activation enthalpy of 89 kJ mol21. A subsequent NMR structure determination of Ras – GppNHp indicated indeed the existence of multiple conformations in the switch regions9 and slow exchange processes were observed as well for Ras – GDP at low magnesium concentrations.12 These findings in solution are consistent with several conformations observed in different crystal forms of triphosphate-bound Ras,5 – 7,13 and also for other GTPases.14 The relevant observation is that the electron density maps show a blurring in the switch regions (resulting in high X-ray B-factors) which indicates that these regions are disordered. An example is Ras – GppCH2p, which contains four molecules in the asymmetric unit out of which three are found to be disordered in the switch 2 region, and, albeit to a lesser degree, also in the switch 1 region.5 However, there are also cases where this blurring is absent from the electron density map, that is, where the switch regions appear to be fixed for various reasons, e.g. special crystal packing forces or complexation with interacting proteins.14 A particular example for this situation is Ras –GppNHp, whose lattice is trigonal and has one molecule in the asymmetric unit.6 In this complex, low B-factors come with the switch 1 region, suggesting that this region is ordered. Indeed, a stabilizing crystal contact of residue 32 to the neighboring protein molecule has been identified.6 We thus wondered whether the multiple conformations and the exchange process in the active center of Ras observed in solution are absent or, in spite of the X-ray results, still present also in crystals of this protein– nucleotide complex. If they are present, a variety of distinct nucleotides should be observable, at least in the slow-exchange limit of the exchange process, that is, at low temperatures. Here we demonstrate that indeed we can observe multiple conformations and slow dynamic exchange processes in Ras protein crystals by solid state NMR.

Results and Discussion 31

P solid state NMR spectroscopy

We addressed the question about dynamic processes in crystals of the Ras protein by means of

Solid State NMR on Protein Single-crystals

31

P solid state NMR spectroscopy, which promises to establish a direct link between the previous findings of liquid state 31P NMR spectroscopy and the solid state. For Ras – nucleotide complexes, 31P NMR enables us to focus, without isotope labelling, selectively on the active center of the protein, since the phosphorus atoms of the nucleotide are the only ones present in the protein– nucleotide complex. For this purpose, two different solid state NMR approaches deserve consideration. First, NMR experiments on oriented single-crystals and, second, magic angle spinning (MAS) NMR on a sample of many, randomly oriented small crystallites.15 The first approach gives, in principle, detailed information about the orientation dependence of NMR quantities such as the chemical shift and the internuclear dipole – dipole interactions, which can be linked to the molecular structure. There are several experimental challenges connected with this technique.16 The most severe is the instrumental sensitivity that is necessary to record spectra of crystals as tiny and dilute in 31P as are the Ras –GppNHp crystals. Such crystals can be grown with a volume of up to 0.3 mm3. Since every Ras molecule is bound to only one nucleotide, the 31P atoms are highly diluted in the protein matrix. In addition, the total signal intensity is distributed to numerous resonance lines. For a general orientation of a trigonal Ras – GppNHp crystal (space group P3221) with all protein– nucleotide molecules in the same conformation, we expect six different 31P resonance lines for each of the Pa, Pb and Pg phosphorus atoms of the nucleotide, that is, a total of 18 lines, and even more if there are multiple conformations. This causes another problem, namely heavy crowding and overlapping of resonance lines. In the MAS approach the orientational information gets lost when averaging out the tensorial character of the chemical shift and dipole-dipole interactions. However, the spectra can be recorded with higher sensitivity and are similar to the solution spectra in their appearance and information content. The most advantageous aspect of this technique is that no “large” single-crystal is necessary but polycrystalline samples can be used. We decided to follow both approaches and show that in our case they complement one another in a fortunate way. Since no commercial probe with sufficient sensitivity was available for single-crystal 31P NMR, we developed a special probe able to detect the expected weak NMR signals from small crystals (Figure 1(a)). It operates at 190 MHz 31P Larmorfrequency and features an NMR coil with 7.5 turns, an inner diameter of 1.5 mm and a length of 1.7 mm. It is equipped with a goniometer that allows rotation of a sample crystal inside the coil in a controlled manner. We used double resonance 1 H ! 31P cross-polarization for 31P spin excitation and 1H decoupling during detection of the 31P NMR signal.

Solid State NMR on Protein Single-crystals

901

Figure 1. Single-crystal NMR, experimental setup and spectra. (a) View of the central part of the 31 P (190 MHz) and 1H (469 MHz) double resonance probe. Behind the double-tuned NMR coil is the worm-and-gear goniometer for rotating the sample crystal. Next to the coil are the 190 MHz (r.h.s.) and 469 MHz (l.h.s.) tuning capacitors. The “big” coils near the capacitors serve for matching the two circuits to the line impedance of 50 V. (b) Four proton decoupled 31P NMR spectra of a single-crystal of Ras(G12P) – GppNHp having a volume of 0.3 mm3, recorded at different rotation angels Y. 1 H ! 31P cross-polarization (contact time 2 ms) with 3% ramping in the 31 P channel was used and 42,000 scans were accumulated corresponding to one day of signal averaging. The arrows indicate resonances from individual 31P sites. In the Y ¼ 08 and Y ¼ 58 spectra, the open arrows demonstrate how rapidly the lines move in the spectra on rotating the sample crystal.

Inferences from single-crystal

31

P NMR spectra

Figure 1(b) shows a series of 31P NMR spectra of a single-crystal of Ras(G12P)–GppNHp with a volume of , 0.3 mm3 recorded at different orientations of the crystal with respect to the applied magnetic field B0. Experimental parameters are given in the Figure legend. The spectra demonstrate that the sensitivity of the probe is sufficient to record a meaningful solid state NMR signal from only about 1015 phosphorus nuclei. Note that the spectral resolution is not limited by instrumental shortcomings but results on the one hand from homonuclear dipolar couplings between the three phosphorus nuclei, and, on the other hand, from heteronuclear dipolar couplings of the 31P nuclei to the 14N nucleus of the NH group between the b and g-phosphate groups. While the homonuclear couplings could in principle, but hardly in practice, be removed by the application of linenarrowing multiple pulse sequences,17 no procedure is known to suppress the heteronuclear

coupling between the 31P and the quadrupolar 14N nuclei. Despite the limited spectral resolution some lines can be identified in the spectra that obviously arise from individual phosphorus sites in the crystal. Their wandering across the spectrum can be followed as the crystal is rotated with respect to B0. This is indicated by the colored sine curves in Figure 2, which shows a stacked plot of spectra obtained when incrementing the rotation angle of the crystal in steps of 58. Integration of those lines leads to intensities of the order of 1/40 of the total signal intensity. This is definitely much less than the ratio of 1/18 that is to be expected for Ras – GppNHp existing in one single conformation only. We made sure that instrumental effects and crystal twinning can be excluded as reasons for the discrepancy between the observed and expected intensity ratios. The inescapable conclusion then is that the Ras – GppNHp complex in the crystals studied exists in more than one (two, perhaps even three) different conformation. We point out that we have

902

Solid State NMR on Protein Single-crystals

Figure 2. Stacked plot of 31P spectra of a single-crystal of Ras(G12P) – GppNHp, recorded for different goniometer rotation angles, Y. The wandering of some individual resonances through the spectrum on rotating the crystal is indicated by colored sine curves. Note the loneliness of the line at d < 2 135 ppm in the spectrum (Y ¼ 458) highlighted by purple coloring. We think that the obvious splitting of this line on rotating the crystal clockwise or anticlockwise by only 158 is due to dipolar interactions. The fact that in all spectra a dominant part of the intensity appears between 2 30 ppm and þ30 ppm may be irritating but is not inconsistent from statistical considerations with the principal tensor components listed in Table 1.

also recorded 31P NMR spectra from crystals of the wild-type Ras – GppNHp. These spectra show the same essential characteristics as those from the G12P mutant: isolated resonances with an intensity of much less than 1/18 of the total intensity, obviously arising from individual 31P sites, and, for certain crystal orientations, no other resonances nearby in the spectrum. Magic angle spinning

31

P NMR

To check the conclusion just drawn we prepared a sample for the MAS technique. It consisted of about 50 crystallites of wild-type Ras – GppNHp. As the unit cell of Ras –GppNHp comprises six magnetically inequivalent sites, and many of the crystallites were partly forming multiple lattices or broken, our sample contains molecules in about 500 –600 different orientations, just sufficient to be considered as a randomly oriented powder. It turned out that the sample survives spinning rates up to 10 kHz without any damage (we did not try higher spinning rates) and it was stable for several months. A selection of 31P MAS spectra from this sample is shown in Figure 3. They have been recorded at the Max-Planck-Institute for Polymer Research in Mainz. Comparison with liquid-state spectra11 (boxed panel) leads to the assignment of the observed lines to the Pa, Pb and Pg nuclei of GppNHp as indicated in the Figure. The isotropic chemical shifts taken from the MAS spectra agree closely with those from Ras –GppNHp in solution (see Table 1). The spectra in Figure 3 show sharp

single lines for the Pa and Pg nuclei. In contrast, the resonance from the b-phosphate group is definitely split. The splitting is most pronounced in the n ¼ 2 1 sideband of the 6 8C and 20 8C spectra. This splitting confirms immediately the conclusion drawn above from the single-crystal spectra: in crystals of Ras – GppNHp, the asymmetric unit must contain multiple molecular conformations. The crystal packing forces inferred from the X-ray diffraction experiments obviously cannot prevent Ras – GppNHp from undergoing conformational changes. Note that the MAS NMR evidence for this conclusion is immune against instrumental problems and does not depend on the absence of crystal twinning. A closer inspection of the MAS spectra taken at different temperatures reveals, moreover, that a dynamic exchange process takes place on a very similar time scale as observed under the more physiological conditions of the liquid state NMR experiment.11 As in the liquid state, the Pb resonance shows the strongest temperature dependence, resulting, at higher temperatures, in narrower lines at averaged positions. On the other hand, Figure 3 contains evidence that at least one component of the Pb resonance, that is, some fraction of the Pb atoms, does not take part in this exchange process, and still is represented by a distinct line/shoulder (see, in particular, the n ¼ þ 1 sideband in the 31 8C MAS spectrum (open arrow)). Whether or not this fraction b(0) of the Pb atoms exchanges with the others on a time-scale of seconds or longer, remains an open question. The difference of the isotropic chemical shifts of

903

Solid State NMR on Protein Single-crystals

Figure 3. Selection of 202 MHz 31P MAS NMR spectra of a powder sample of Ras(wt) – GppNHp recorded at different temperatures. Sample weight, 4 mg; spinning speed, 10 kHz; number of scans, 24,000 – 36,000. The assignment of the lines is indicated in the center band n ¼ 0 in the 31 8C spectrum. Note the multiple lines in the resonances of the b-phosphate group, best visible in the n ¼ 2 1 spinning side band of the 6 8C spectrum, and the partial coalescence of these lines in the 31 8C spectrum. For comparision a 202 MHz liquid state 31P NMR spectrum of 0.6 mM Ras(wt) – GppNHp at 5 8C is boxed on top.

Table 1. 31P NMR data of Ras – GppNHp in the liquid and solid states A. Isotropic chemical shift d (ppm) a(1) Liquid statea 211.15 211.6 Solid stateb

a(2) 211.85

b(0) – 21.8

b(1) 22.69 22.8

b(2) 23.41 23.6

g(1) 20.41

g(2) 20.23 20.3

B. Principal values dii of the chemical shift tensors (ppm) Solid stateb dXX dYY dZZ a

a 60 36 2132

b 69 30 288

g 90 15 2105

From Geyer et al.11 Values are referenced to the liquid state by setting the chemical shift of the g-resonance in the 6 8C MAS-spectrum to 20.32 ppm, which is the mean value in the liquid state. b

904

Solid State NMR on Protein Single-crystals

must also be distinct even though the respective resonances cannot be resolved in the MAS spectrum. Implications of MAS results

Figure 4. Conformational variety around the nucleotide-binding site. (a) The structure of Ras – GppNHp according to Pai et al.6 (PDB accession code 5P21). (b) The structure of the complex Rap1A– GppNHp– Raf (1GUA).19 To display the two structures, projections are chosen such that the nucleotide is in essentially the same orientation. Note the different orientations for these projections of residues Tyr32 and Gln61 in (a) and (b). The Figure was generated using the program MOLSCRIPT.40

the b(1) and the b(2) lines is 0.7 ppm, which is, within the accuracy of the MAS experiment, equal to the value of 0.72 ppm found in the liquid state.11 The relative populations of the b(1) and b(2) states at 6 8C can be estimated to be roughly equal, which again matches the liquid state result. A Herzfeld –Berger analysis18 of the sideband patterns of the MAS spectra allows us to deduce the principal values dii of the chemical shift tensors of the 31P nuclei. These are given in Table 1. The numbers in this Table show that the Pa nuclei can experience, dependent on the crystal orientation, the largest downfield chemical shifts. In fact, any resonance line in the single-crystal spectra with a chemical shift d , 2 110 ppm must correspond to a Pa nucleus. Conversely, lines with shifts d . þ 75 ppm can safely be assigned to the g-phosphate group. The signal-to-noise ratio and the strong signal overlap does not allow us to extract from the MAS spectra separate principal shielding values for the b(0), b(1) and b(2) resonances on safe grounds. However, Herzfeld – Berger analyses of average versus separate b-resonances at different temperatures and spinning speeds give very similar results, suggesting that all Pb chemical shift tensors have similar principal values. It is reasonable to assume that the same is true of the Pa(0), Pa(1), Pa(2) and Pg(0), Pg(1), Pg(2) chemical shift tensors, which

The natural assumption formulated from the Herzfeld – Berger analyses has the following interesting implication. From the intensities of the resonances connected by the colored sine curves shown in Figure 2 we concluded that each of them arises from one of the phosphorus nuclei of a subset of nucleotides in one of the asymmetric units of the trigonal cell of the crystal. Now we see in addition that, in fact, they must arise from a subset of Pa nuclei because the maximum negative shielding is not compatible with the shielding tensors of Pb and Pg. The intriguing point is that, in particular, the pink sine curve shown in Figure 2 has no partner that follows it at a close distance as a function of the rotation angle Y. This means that the Pa nuclei of the different conformers in the asymmetric unit must have substantially differing chemical shift tensors. Because, as we have argued above, these tensors hardly differ with respect to their principal values, they must differ in their tensor orientations. This implies that the orientations of the whole molecular framework of the nucleotide in the various conformations of the protein– nucleotide complex must differ substantially. By calculating how much a Pa shielding tensor must differ in its orientation to give rise to the next nearest resonance line in the spectra, we may even quantify this statement: the result of this elementary calculation is that the angle through which the Pa shielding tensor, and thus the molecular framework of one of the nucleotide conformers must be rotated in order to bring it in coincidence with that of another conformer cannot be less that 208. Structural conclusions A rotation through (at least) 208 of the molecular framework of a nucleotide is a substantial structural change. How does our detection of such changes in Ras –GppNHp fit into the current picture of the structure and flexibility of Ras complexes? As an attempt to answer this question we compare in Figure 4(a) and (b) the structure, as determined by X-ray diffraction, of Ras – GppNHp alone6 with the structure of the complex Rap1A – GppNHp – Raf.19 Rap1A is a close homologue to Ras and its structure with Raf is believed to mimick the Ras –Raf complex formation.20 For clarity, only the neighborhood of the nucleotide is displayed. Both structures are supposed to depict Ras in its activated form; nevertheless, the orientations and displacements of the nucleotide relative to, in particular, the residues tyrosine 32 (in switch 1) and glutamine 61 (in switch 2) are remarkably different. Remembering the functional role of Raf, this suggests that the complexation

905

Solid State NMR on Protein Single-crystals

with Raf is necessary for Ras – GppNHp to assume a definite structure while without complexation with an effector protein there is considerable conformational variability and flexibility which is clearly seen by 31P NMR spectroscopy but could not be resolved in the analysis of the respective X-ray diffraction data. As pointed out in the Introduction, the data on Ras – GppNHp were refined in a trigonal cell with an asymmetric unit holding one protein– nucleotide complex. We think that the resolution of the X-ray experiment ˚ ),6 leaves enough room (actually excellent, 1.35 A to accomodate a bunch of nucleotide conformers with (estimated) 208 mutual rotations. On a speculative level it is well conceivable that one of the conformers (that corresponding to the b(0) resonance, which does not show chemical exchange in the MAS spectra) is stabilized by a crystal contact of tyrosine 32 to the neighboring molecule. In another conformer, possibly that corresponding to b(2), tyrosine 32 could be hydrogen-bonded to the g-phosphate group of the nucleotide while the third, b(1), might correspond to more disordered conformations of the effector loop of the protein. This would be in line with earlier findings that details of the arrangement of the effector loop vary in crystal structures of Ras5 – 7 and its related homologue Rap2A.21 Biological meaning The preservation of multiple conformations and the associated flexibility of Ras complexes even in the crystalline state leads, finally, to the question of the biological meaning of the conformational variability. Because many mechanisms of protein function require conformational changes, it has been proposed that conformational fluctuations must be related to protein function.22,23 For example, motions on the millisecond timescale have been shown to be important for optimal protein– protein interaction of the bacterial response regulator Spo0F.24 Another recent study showed that transitions on the millisecond timescale between a highly populated ground and an excited state of a cavity mutant of T4 lysozyme suggest that the excited state facilitates ligand entry although the ligand-binding site is sterically inaccessible according to X-ray studies.25 For Ras, transient kinetic methods showed that the interaction between Ras –GppNHp and the effector protein Raf is a two-step process, with an initial rapid equilibration step followed by a much slower isomerization reaction.26 A kind of “dynamic triggering” of Ras protein interactions is also indicated by the alternating selective stabilization of two states upon addition of Raf or the GTPase activating protein RasGAP, as observed by solution NMR.11 Indeed, analysis of the dynamic properties of the switch 1 region by mutants of Thr35 in Ras showed a two-step binding reaction for wild-type and (T35S)Ras with effector proteins. This interaction requires the existence of a rate-

limiting isomerization step, which is not observed for the partial loss-of-function mutant T35A.27 Similar slow-dynamic exchange processes in the active center of the molecule have been shown also for the nuclear import factor Ran in its triphosphate bound form28 and seem to emerge as a general feature of small GTP-binding proteins. Evidence thus accumulates that conformational changes and the associated activation barriers play an essential role for triggering the hydrolysis process of protein-bound GTP29,30 and the recognition of interacting proteins. Summary and outlook From single-crystal 31P NMR spectra and rotation patterns of such spectra we inferred a conformational variability of the nucleotide in crystals of Ras – GppNHp and quantified the orientational differences. That means, we have elucidated additional structural details of these crystals, which were not accessible to diffraction techniques. With this work we thus established single-crystal 31 P NMR as a new means to get structural information on protein –nucleotide complexes. Our 31P MAS NMR spectra from powder samples confirmed the structural conclusions inferred from the single-crystal work. The interpretation of MAS spectra is much less involved than that of singlecrystal spectra and rotation patterns: evidence for multiple conformations of the nucleotide in Ras – GppNHp is obtained by simply counting the number of resonance lines in a single MAS spectrum. The relative ease of obtaining MAS spectra allowed us to record such spectra for a range of different temperatures. Comparable to solution spectra, the solid state spectra recorded with Ras provided evidence for a dynamic exchange process occurring on a timescale of a few milliseconds between two of the nucleotide conformers while a third appeared to remain separate until at least T ¼ 31 8C. We thus introduced 31P MAS NMR as a means to get dynamical information on protein –nucleotide complexes in the solid state. Although the requirements for single-crystal NMR with regard to the size of the crystal to be studied are severe, phosphate groups, either as nucleotides or as phosphorylation sites, often are particularly interesting because they represent the active center of the molecule or the protein complex.31 These regions are inherently most crucial to protein dynamics and conformational changes. The use of specific spin labels, as is e.g. the nucleotide derivative GTPgF,32 may additionally reduce the requirements on the crystal size. In solution, NMR spectroscopy offers the possibility to correlate structural fluctuations of known magnitude and timescale with protein function.33 – 36 Here we present the feasibility to observe reversible exchange processes in protein crystals by NMR. The techniques developed have particular impact on the complementary aspects of solid state NMR spectroscopy and X-ray diffraction

906

experiments on protein crystals. The rapidly growing availability of superconducting magnets with higher and higher fields will soon ease the NMR sensitivity problem and make this technique more and more attractive.

Materials and Methods Protein crystallization Crystals of Ras – GppNHp were grown in the P3221 space group following the procedure of Scherer et al.,37 but including sitting drop methodology. To increase the crystal size, the drop volume was increased to 50 ml, using glass sitting drop rods (Hampton Research) in combination with regular hanging drop plates. This modification also resulted in slower nucleation and growth. We also used the non-transforming Ras(Gly12Pro) mutant for single-crystal NMR measurements, of which large single-crystals could be grown as in previous studies.38 Single-crystal NMR probe and spectrometer set-up To achieve the best possible signal-to-noise ratio in pulsed NMR experiments, it is necessary to adapt the size of the sample coil to the size of the sample. The inner diameter of our sample coil matched the outer diameter of the glass capillaries of 1.5 mm in which the crystals were mounted. After sealing a capillary, the protein crystal inside was shortly exposed to Cu Ka X-rays to determine, by diffraction, its orientation in the capillary. See also the legend to Figure 1 and Stumber.39 MAS NMR spectroscopy Sample preparation A collection of about 50 crystallites of minor quality (partly twinned, broken) was harvested, washed, and put in a glass capillary (1 mm diameter, 8 mm length), which was sealed on both ends to prevent the sample from drying out. This capillary was then glued into a cylinder of Vespelw which was subsequently machined down on a lathe to fit exactly into a 2.5 mm MAS rotor. Spectra The spectra were recorded on a Bruker ASX-500 spectrometer (24,000 – 36,000 accumulations, cross-polarization with a 2 ms contact time, 100 kHz continuous wave proton decoupling) with spinning speeds of 7 kHz and 10 kHz in a temperature range from 5 8C to 31 8C. Values are referenced to the liquid state by setting the chemical shift of the g-phosphorus resonance in the MAS-spectrum at 6 8C to 2 0.32 ppm, which is the mean value in the liquid state.

Acknowledgements We thank Anna Scherer for her patient efforts of growing “large” single-crystals of Ras – GppNHp, Frans Mulder for discussions and Kenneth C. Holmes for continuous support and encourage-

Solid State NMR on Protein Single-crystals

ment. M.S. & M.G. gratefully acknowledge support by the German-Israeli Science Foundation (GIF) and the Peter und Traudl Engelhorn Stiftung, respectively.

References 1. Bourne, H. R., Sanders, D. A. & McCormick, F. (1990). The GTPase superfamily: a conserved switch for diverse cell functions. Nature, 348, 125– 132. 2. Marshall, M. S. (1995). Ras target proteins in eukaryotic cells. FASEB J. 9, 1311 – 1318. 3. Barbacid, M. (1987). Ras genes. Annu. Rev. Biochem. 56, 779– 827. 4. Bos, J. L. (1989). Ras oncogenes in human cancer: a review. Cancer Res. 49, 4682– 4689. 5. Milburn, M. V., Tong, L., deVos, A. M., Bru¨nger, A., Yamaizumi, Z., Nishimura, S. & Kim, S. H. (1990). Molecular switch for signal transduction: structural differences between active and inactive forms of proto-oncogenic Ras proteins. Science, 247, 939– 945. 6. Pai, E. F., Krengel, U., Petsko, G. A., Goody, R. S., Kabsch, W. & Wittinghofer, A. (1990). Refined crystal structure of the triphosphate conformation of H-ras ˚ resolution: implications for the mechap21 at 1.35 A nism of GTP hydrolysis. EMBO J. 9, 2351– 2359. 7. Scheidig, A. J., Burmester, C. & Goody, R. S. (1999). The pre-hydrolysis state of p21(ras) in complex with GTP: new insights into the role of water molecules in the GTP hydrolysis reaction of ras-like proteins. Struct. Fold. Des. 7, 1311 – 1324. 8. Kraulis, P. J., Domaille, P. J., Campbell, S. L., van Aken, T. & Laue, E. D. (1994). Solution structure and dynamics of Ras p21·GDP determined by heteronuclear three and four-dimensional NMR spectroscopy. Biochemistry, 33, 3515– 3531. 9. Ito, Y., Yamasaki, K., Iwahara, J., Terada, T., Kamiya, A., Shirouzu, M. et al. (1997). Regional polysterism in the GTP-bound form of the human c-Ha-Ras protein. Biochemistry, 36, 9109– 9119. 10. Schlichting, I., Almo, S. C., Rapp, G., Wilson, K., Petratos, K., Lentfer, A. et al. (1990). Time resolved X-ray crystallographic study of the conformational change in Ha-Ras p21 protein on GTP hydrolysis. Nature, 345, 309– 315. 11. Geyer, M., Schweins, T., Herrmann, C., Prisner, T., Wittinghofer, A. & Kalbitzer, H. R. (1996). Conformational transitions in p21ras and in its complexes with the effector protein Raf-RBD and the GTPase activating protein GAP. Biochemistry, 35, 10308 –10320. 12. Ganguly, A. K., Wang, Y. S., Pramanick, B. N., Doll, R. J., Snow, M. E., Taveras, A. G. et al. (1998). Interaction of a novel GDP exchange inhibitor with the Ras protein. Biochemistry, 37, 15631 –15637. 13. Geyer, M. & Schlichting, I. (1999). Information contained in protein structures determined by NMR in solution or by X-ray diffraction in crystals. In Simplicity and Complexity in Proteins and Nucleic Acids (Frauenfelder, H., Deisenhofer, J. & Wolynes, P. G., eds), pp. 59 – 79, Dahlem University Press, Berlin. 14. Me´ne´trey, J. & Cherfils, J. (1999). Structure of the small G protein Rap2 in a non-catalytic complex with GTP. Proteins: Struct. Funct. Genet. 37, 465– 473. 15. Pines, A., Gibby, M. G. & Waugh, J. S. (1972). Protonenhanced nuclear induction spectroscopy. A method

Solid State NMR on Protein Single-crystals

16. 17. 18. 19.

20. 21.

22. 23. 24.

25.

26.

27.

for high resolution n.m.r. of dilute spins in solids. J. Chem. Phys. 56, 1776 –1777. Haeberlen, U. (1996). Optimizing sensitivity in solid state NMR. Bull. Ampere, 45, 13 – 30. Haeberlen, U. (1976). Advances in Magnetic Resonance (Waugh, J. S., ed.), Suppl. 1, Academic Press, New York. Herzfeld, J. & Berger, A. E. (1980). Sideband intensities in NMR spectra of samples spinning at the magic angle. J. Chem. Phys. 73, 6021– 6030. Nassar, N., Horn, G., Herrmann, C., Scherer, A., ˚ McCormick, F. & Wittinghofer, A. (1995). The 2.2 A crystal structure of the Ras binding domain of the serine/threonine kinase c-Raf1 in complex with Rap1A and a GTP analogue. Nature, 375, 554– 560. Wittinghofer, A. & Nassar, N. (1996). How Rasrelated proteins talk to their effectors. Trends Biochem. Sci. 21, 488– 491. Cherfils, J., Me´ne´trey, J., Le Bras, G., JanoueixLerosey, I., de Gunzburg, J., Garel, J. R. & Auzat, I. (1997). Crystal structure of the small G protein Rap2A in complex with its substrate GTP, with GDP and with GTPgS. EMBO J. 16, 5582– 5591. Frauenfelder, H., Parak, F. & Young, R. D. (1988). Conformational substates in proteins. Annu. Rev. Biophys. Chem. 17, 451– 479. Ringe, D. & Petsko, G. A. (1986). Study of protein dynamics by X-ray diffraction. Methods Enzymol. 131, 389– 433. Feher, V. A. & Cavanagh, J. (1999). Millisecond-timescale motions contribute to the function of the bacterial response regulator protein Spo0F. Nature, 400, 289– 293. Mulder, F. A., Mittermaier, A., Hon, B., Dahlquist, F. W. & Kay, L. E. (2001). Studying excited states of proteins by NMR spectroscopy. Nature Struct. Biol. 8, 932–935. Sydor, J. R., Engelhard, M., Wittinghofer, A., Goody, R. S. & Herrmann, C. (1998). Transient kinetic studies on the interaction of Ras and the Ras-binding domain of c-Raf-1 reveal rapid equilibration of the complex. Biochemistry, 37, 14292– 14299. Spoerner, M., Herrmann, C., Vetter, I. R., Kalbitzer, H. R. & Wittinghofer, A. (2001). Dynamic properties of the Ras switch I region and its importance for binding to effectors. Proc. Natl Acad. Sci. USA, 98, 4944–4949.

907

28. Geyer, M., Assheuer, R., Klebe, C., Kuhlmann, J., Becker, J., Wittinghofer, A. & Kalbitzer, H. R. (1999). Conformational states of the nuclear GTP-binding protein Ran and its complexes with the exchange factor RCC1 and the effector protein RanBP1. Biochemistry, 38, 11250– 11260. 29. Ma, J. & Karplus, M. (1997). Molecular switch in signal transduction: reaction paths of the conformational changes in ras p21. Proc. Natl Acad. Sci. USA, 94, 1905– 11910. 30. Soares, T. A., Miller, J. H. & Straatsma, T. P. (2001). Revisiting the structural flexibility of the complex p21(ras)-GTP: the catalytic conformation of the molecular switch II. Proteins, 45, 297– 312. 31. Westheimer, F. H. (1987). Why nature chose phosphates. Science, 235, 1173– 1178. 32. Stumber, M., Herrmann, C., Wohlgemuth, S., Kalbitzer, H. R., Jahn, W. & Geyer, M. (2002). Synthesis, characterization and application of two nucleoside triphosphate analogues, GTPgNH2 and GTPgF00. Eur. J. Biochem. 269, 3270– 3278. 33. Palmer, A. G., III (1997). Probing molecular motion by NMR. Curr. Opin. Struct. Biol. 7, 732– 737. 34. Kay, L. E. (1998). Protein dynamics from NMR. Nature Struct. Biol. 5, 145– 152. 35. Ishima, R. & Torchia, D. A. (2000). Protein dynamics from NMR. Nature Struct. Biol. 7, 740–743. 36. Wand, A. J. (2001). Dynamic activation of protein function: a view emerging from NMR spectroscopy. Nature Struct. Biol. 8, 926– 931. 37. Scherer, A., John, J., Linke, R., Goody, R. S., Wittinghofer, A., Pai, E. F. & Holmes, K. C. (1989). Crystallization and preliminary X-ray analysis of the human c-H-ras-oncogene product p21 complexed with GTP analogues. J. Mol. Biol. 206, 257– 259. 38. Franken, S. M., Scheidig, A. J., Krengel, U., Rensland, H., Lautwein, A., Geyer, M. et al. (1993). Threedimensional structures and properties of a transforming and a non-transforming glycine-12 mutant of p21H-ras. Biochemistry, 32, 8411– 8420. 39. Stumber, M. (2002). Dynamik in Kristallen des Proteinkomplexes p21ras:GppNHp untersucht mit 31 P-Festko¨rper-NMR und Aufbau des dazu erforderlichen Spektrometers. Dissertation, Universita¨t Heidelberg. 40. Kraulis, P. J. (1991). MOLSCRIPT: a program to produce both detailed and schematic plots of protein structures. J. Appl. Crystallog. 24, 946–950.

Edited by R. Huber (Received 12 June 2002; received in revised form 20 August 2002; accepted 27 August 2002)