3D C(CC)H TOCSY experiment for assigning protons and carbons in uniformly 13C- and selectively 2H-labeled RNA.

We have prepared RNAs uniformly 13C-labeled in the ribose ring and with the H3',H4',H5'/H5" protons specifically replaced with deuterons (3',4',5'/5"-2H-13C, termed d4-13C ribose), which offers the advantage of spectral simplification without sacrifice of sensitivity for the remaining protons. A 3D C(CC)H TOCSY experiment that uses a combination of cross-polarization transfer and deuterium decoupling improves the transfer of magnetization around the ribose ring and facilitates assignment of all ribose carbons and H1' and H2' protons in a 30-nucleotide RNA from HIV-2. These and other combinations of labeling and pulse sequence methodology should prove invaluable for the study of large RNAs, RNA-ligand, and RNA-protein complexes.

Unusual lack of internal mobility and fast overall tumbling in oxidized flavodoxin from Anacystis nidulans.

Anacystis nidulans flavodoxin, an electron-transfer protein containing a flavin mononucleotide (FMN) molecule as its prosthetic group, has a redox potential for the oxidized/semiquinone equilibrium close to that of free flavin. Whereas the redox potential for the semiquinone/hydroquinone equilibrium is more negative. To gain an understanding of the contribution of mobility to redox potential modulation, we studied the backbone mobility of the oxidized A. nidulans flavodoxin at pH 6.6, 303 K by 15N NMR relaxation measurements. The spin-lattice relaxation rate constants (RN(Nz)=1/T1), spin-spin relaxation rate constants (RN(Nx,y)=1/T2) and 1H-15N nuclear Overhauser effects (NOE) were obtained for 143 of the 166 protonated backbone 15N nuclei and for the FMN N3 nucleus without ambiguity. The 15N T1, T2 and NOE data were analyzed by reduced spectral density mapping, and the so-called model-free approach. In contrast to most other proteins studied with 15N relaxation experiments, we found an almost complete absence of internal mobility. The overall correlation time of this 169-residue flavodoxin (>19 kDa) is significantly shorter (7.4 to 7.8 ns) than that of other proteins of this size, suggesting that the absence of internal mobility is correlated with faster overall rotational diffusion. The uniformity of the motional parameters along the backbone is in strong contrast to the crystallographic B-factors, which vary significantly along the sequence in this and other flavodoxins. The NMR relaxation parameters are primarily sensitive to rotational diffusive motions of the N-H bond vectors, while the crystallographic B-factors would be sensitive to translational internal motions as well. However, the large B-factors in this protein may originate from crystal packing and crystal lattice disorder. The relatively fast overall tumbling results in sharp NMR resonances. Hence, much larger proteins with such favorable dynamic behavior could be excellent candidates for studies by NMR.

Structure and mobility of the PUT3 dimer.

The solution structure and backbone dynamics of the transcriptional activator PUT3 (31-100) has been characterized using NMR spectroscopy. PUT3 (31-100) contains three distinct domains: a cysteine zinc cluster, linker, and dimerization domain. The cysteine zinc cluster of PUT3 closely resembles the solution structure of GAL4, while the dimerization domain forms a long coiled-coil similar to that observed in the crystal structures of GAL4 and PPR1. However, the residues at the N-terminal end of the coiled-coil behave very differently in each of these proteins. A comparison of the structural elements within this region provides a model for the DNA binding specificity of these proteins. Furthermore, we have characterized the dynamics of PUT3 to find that the zinc cluster and dimerization domains have very diverse dynamics in solution. The dimerization domain behaves as a large protein, while the peripheral cysteine zinc clusters have dynamic properties similar to small proteins.

The counterreceptor binding site of human CD2 exhibits an extended surface patch with multiple conformations fluctuating with millisecond to microsecond motions.

We have used 15N NMR relaxation experiments to probe, for the glycosylated human CD2 adhesion domain, the overall molecular motion, as well as very fast nanosecond-picosecond (ns-ps) and slow millisecond-microsecond (ms-microsecond) internal motions. Using a novel analysis method that considers all residues, we obtained a correlation time for the overall motion of 9.5 +/- 0.3 ns. Surprisingly, we found a large contiguous patch of residues in the counterreceptor (CD58) binding site of human CD2 exhibiting slow conformational exchange motions (ms-microsecond). On the other hand, almost none of the residues of the CD58 binding side display fast (ns-ps) internal motions of amplitudes larger than what is seen for well-ordered regions of the structure. Residues close to the N-glycosylation site, and the first N-acetylglucosamine of the high mannose glycan are as rigid as the protein core. Residues conserved in the immunoglobulin superfamily V-set domain are generally very rigid.

A robust method for estimating cross-relaxation rates from simultaneous fits to build-up and decay curves.

This paper introduces a novel computational method for estimating relaxation rates among pairs of spin orders. This method simultaneously estimates all the auto- and cross-relaxation rates from the same measurements, and avoids the ill-conditioning problems associated with multiexponential fits. The method models the relaxation dynamics by a system of linear differential equations, and assumes that measurements of the spin orders have been made at an equally spaced sequence of time points. It computes a nonlinear least-squares fit of the exponential of the rate matrix at the shortest time point to these measurements. Preliminary estimates of the exponential matrix and initial spin orders from which to start the computations are obtained by solving simpler linear-least-squares problems. The performance of the method on simulated 2 x 2 test problems indicates that when measurements at eight or more equally spaced times spanning the maximum and inflection points of the build-up curves are available, the relative errors in the rates are usually less than the relative errors in the measurements. The method is further demonstrated by applying it to the problem of determining the cross correlation-induced cross-relaxation rates between the in-phase and antiphase coherence of the amide groups in the 15N-labeled protein oxidized flavodoxin. Finally, the possibility of extending the method to other kinds of relaxation measurements and larger spin systems is discussed.

Internal mobility in the partially folded DNA binding and dimerization domains of GAL4: NMR analysis of the N-H spectral density functions.

The DNA binding domain (residues 1--65) of the yeast transcriptional activator GAL4 is only partially folded. While residues 10-41, the DNA recognition domain, form a well-defined structure in the free protein, the whole polypeptide folds up and dimerizes upon binding DNA. In order to describe the mobility of the protein, we have characterized the frequency spectrum of the motions of N-H bond vectors of GAL4(1-65) using a reduced spectral density mapping approach (an approximation of the full spectral density mapping technique) [Peng, J. W., & Wagner, G. (1992a) J. Magn. Reson. 98, 308-332; Peng. J. W., & Wagner, G. (1992b) Biochemistry 31, 8571-8586]. 15N spin-lattice relaxation [Rn(Nz)], spin-spin relaxation [Rn(Nx,y)], cross-relaxation [RN(Hz-->Nz)], two-spin order [RNH(2HzNz)], and antiphase [RNH(2HzNx,y)] rates were determined for 52 of the 65 backbone amide groups at 10 degrees C and ph 6.5 at 11.74 T. Calculations of the spectral density functions using a reduced set of RN(Nz),RN(Nx,y),RN(Hz-->Nz), and RNH(2HzNz) gave excellent agreement with those calculated using all six sets. The reduced method has the added advantage that the errant behavior seen at high field values is circumvented. A linear correlation was found between J(omega N) and J(0) with a limited and clearly defined range of J(0) values which defines the range of rates for internal motions in GAL4(1-65). It appears that all residues experience a combination of two movements: one of the overall tumbling (correlation time, 8.65 ns) and the other of fast internal fluctuations of the structure. The respective weights of these contributions vary with the primary sequence and faithfully mirror the secondary and tertiary elements of the protein. The position on the correlation line of J(omega N) versus J(0) indicates the amount of angular averaging relative to the overall motion of the protein. A spectral density function for internal motions can be described.

Local mobility within villin 14T probed via heteronuclear relaxation measurements and a reduced spectral density mapping.

Villin 14T, a representative domain from the actin severing and bundling protein villin, binds calcium ions and actin monomers. To begin to understand the contributions of mobility to the villin-calcium and villin-actin interactions, relaxation rates for magnetization involving the amide nitrogens and protons have been measured for 15N-labeled villin 14T in solution. Although we have measured the complete set of rates required for a full spectral density map, difficulties in the accurate measurement of relaxation rates for antiphase coherence and two-spin order led us to consider a reduced mapping formalism. From the reduced spectral density map, a model-free analysis, or directly from the measured Nx,y relaxation rates, local variations in mobility along the backbone of villin 14T have been revealed. Fast motions are evident not only at the amino and carboxyl termini but also in the turn between strands beta 4 and beta 5 of the central beta-sheet and in the turn between helix alpha 3 and strand beta 7. Slower motions are suggested for the turn between strands beta 2 and beta 3. Motions on the microsecond to millisecond time scale have been probed directly by examining the dependence of the proton transverse relaxation rate on the spin-locking field strength. Leu11 shows a strong dependence on field strength, implying conformational exchange with a time constant of 125 +/- 69 microseconds. The backbone at the actin-binding interface appears to be rather rigid.

Theory and practice of nuclear spin relaxation in proteins.

NMR relaxation experiments can provide information on overall and internal motions in proteins. This review consists of a concise report on the evolution of the theories for nuclear relaxation followed by an overview of mathematical models for internal motions in proteins. Next, the method of spectral density mapping with recent developments is reviewed. This is followed by a discussion of pulse sequences for relaxation experiments. Finally, we review recent studies correlating relaxation parameters, in particular spectral density functions, with structural features of proteins and with results of molecular dynamics simulations.