Three-dimensional solution structure of Acanthamoeba profilin-I.

We have determined a medium resolution three-dimensional solution structure of Acanthamoeba profilin-I by multidimensional nuclear magnetic resonance spectroscopy. This 13-kD actin binding protein consists of a five stranded antiparallel beta sheet flanked by NH2- and COOH-terminal helices on one face and by a third helix and a two stranded beta sheet on the other face. Data from actin-profilin cross-linking experiments and the localization of conserved residues between profilins in different phyla indicate that actin binding occurs on the molecular face occupied by the terminal helices. The other face of the molecule contains the residues that differ between Acanthamoeba profilins-I and II and may be important in determining the difference in polyphosphoinositide binding between these isoforms. This suggests that lipids and actin bind to different faces of the molecule.

Intracellular polymerization of sickle hemoglobin. Effects of cell heterogeneity.

To determine the extent to which the broad distribution in intracellular hemoglobin concentrations found in sickle erythrocytes affects the extent of intracellular polymerization of hemoglobin S, we have fractionated these cells by density using discontinuous Stractan gradients. The amount of polymer formed in the subpopulations was experimentally measured as a function of oxygen saturation using 13C nuclear magnetic resonance spectroscopy. The results for each subpopulation are in very good agreement with the theoretical predictions based on the current thermodynamic description for hemoglobin S gelation. We further demonstrate that the erythrocyte density profile for a single individual with sickle cell anemia can be used with the theory to predict the amount of polymer in unfractionated cells. We find that heterogeneity in intracellular hemoglobin concentration causes the critical oxygen saturation for formation of polymer to shift from 84 to greater than 90%; polymer is formed predominantly in the dense cells at the very high oxygen saturation values. The existence of polymer at arterial oxygen saturation values has significance for understanding the pathophysiology of sickle cell anemia. The utility of these techniques for assessing various therapeutic strategies is discussed.

Flap opening and dimer-interface flexibility in the free and inhibitor-bound HIV protease, and their implications for function.

BACKGROUND: (1)H and (15)N transverse relaxation measurements on perdeuterated proteins are ideally suited for detecting backbone conformational fluctuations on the millisecond-microsecond timescale. The identification of conformational exchange on this timescale by measuring the relaxation of both (1)H and (15)N holds great promise for the elucidation of functionally relevant conformational changes in proteins. RESULTS: We measured the transverse (1)H and (15)N relaxation rates of backbone amides of HIV-1 protease in its free and inhibitor-bound forms. An analysis of these rates, obtained as a function of the effective rotating frame field, provided information about the timescale of structural fluctuations in several regions of the protein. The flaps that cover the active site of the inhibitor-bound protein undergo significant changes of backbone (φ,psi) angles, on the 100 micros timescale, in the free protein. In addition, the intermonomer beta-sheet interface of the bound form, which from protease structure studies appears to be rigid, was found to fluctuate on the millisecond timescale. CONCLUSIONS: We present a working model of the flap-opening mechanism in free HIV-1 protease which involves a transition from a semi-open to an open conformation that is facilitated by interaction of the Phe53 ring with the substrate. We also identify a surprising fluctuation of the beta-sheet intermonomer interface that suggests a structural requirement for maturation of the protease. Thus, slow conformational fluctuations identified by (1)H and (15)N transverse relaxation measurements can be related to the biological functions of proteins.

Characterization of two hydrophobic methyl clusters in HIV-1 protease by NMR spin relaxation in solution.

Nearly 50 % of the amino acid residues of HIV-1 protease contain methyl side-chains, most of which appear to be organized into two clusters: the inner cluster that nearly surrounds the active site and the outer cluster that contains the hydrophobic core which stabilizes the inhibitor-free protease structure. NMR relaxation experiments sensitive to motions of methyl groups on the sub-nanosecond and the milli-microsecond time-scales revealed flexible methyl groups in residues that link the two clusters, the methyl groups of L10, L23, V75, and L76. We hypothesize that flexibility at the junctions of these clusters allows the protease to minimize conformational changes upon drug-binding. The two-methyl cluster motif appears to be a common structural feature among retroviral proteases and may play a similar role throughout this family of enzymes.

Refining the overall structure and subdomain orientation of ribosomal protein S4 delta41 with dipolar couplings measured by NMR in uniaxial liquid crystalline phases.

Prokaryotic protein S4 initiates assembly of the small ribosomal subunit by binding to 16 S rRNA. Residues 43-200 of S4 from Bacillus stearothermophilus (S4 Delta41) bind to both 16 S rRNA and to a mRNA pseudoknot. In order to obtain structure-based insights regarding RNA binding, we previously determined the solution structure of S4 Delta41 using NOE, hydrogen bond, and torsion angle restraints. S4 Delta41 is elongated, with two distinct subdomains, one all helical, the other including a beta-sheet. In contrast to the high resolution structures obtained for each individual subdomain, their relative orientation was not precisely defined because only 17 intersubdomain NOE restraints were determined. Compared to the 1.7 A crystal structure, when the sheet-containing subdomains are superimposed, the helical subdomain is twisted by almost 45 degrees about the long axis of the molecule in the solution structure. Because variations in subdomain orientation may explain how the protein recognizes multiple RNA targets, our current goal is to determine the orientation of the subdomains in solution with high precision. To this end, NOE assignments were re-examined. NOESY experiments on a specifically labeled sample revealed that one of the intersubdomain restraints had been misassigned. However, the revised set of NOE restraints produces solution structures that still have imprecisely defined subdomain orientations and that lie between the original NMR structure and the crystal structure. In contrast, augmenting the NOE restraints with N-H dipolar couplings, measured in uniaxial liquid crystalline phases, clearly establishes the relative orientation of the subdomains. Data obtained from two independent liquid crystalline milieux, DMPC/DHPC bicelles and the filamentous bacteriophage Pf1, show that the relative orientation of the subdomains in solution is quite similar to the subdomain orientation in the crystal structure. The solution structure, refined with dipolar data, is presented and its implications for S4's RNA binding activity are discussed.

The RNA binding domain of ribosomal protein L11: three-dimensional structure of the RNA-bound form of the protein and its interaction with 23 S rRNA.

The three-dimensional solution structure has been determined by NMR spectroscopy of the 75 residue C-terminal domain of ribosomal protein L11 (L11-C76) in its RNA-bound state. L11-C76 recognizes and binds tightly to a highly conserved 58 nucleotide domain of 23 S ribosomal RNA, whose secondary structure consists of three helical stems and a central junction loop. The NMR data reveal that the conserved structural core of the protein, which consists of a bundle of three alpha-helices and a two-stranded parallel beta-sheet four residues in length, is nearly the same as the solution structure determined for the non-liganded form of the protein. There are however, substantial chemical shift perturbations which accompany RNA binding, the largest of which map onto an extended loop which bridges the C-terminal end of alpha-helix 1 and the first strand of parallel beta-sheet. Substantial shift perturbations are also observed in the N-terminal end of alpha-helix 1, the intervening loop that bridges helices 2 and 3, and alpha-helix 3. The four contact regions identified by the shift perturbation data also displayed protein-RNA NOEs, as identified by isotope-filtered three-dimensional NOE spectroscopy. The shift perturbation and NOE data not only implicate helix 3 as playing an important role in RNA binding, but also indicate that regions flanking helix 3 are involved as well. Loop 1 is of particular interest as it was found to be flexible and disordered for L11-C76 free in solution, but not in the RNA-bound form of the protein, where it appears rigid and adopts a specific conformation as a result of its direct contact to RNA.

Solution structure and dynamics of linked cell attachment modules of mouse fibronectin containing the RGD and synergy regions: comparison with the human fibronectin crystal structure.

We report the three-dimensional solution structure of the mouse fibronectin cell attachment domain consisting of the linked ninth and tenth type III modules, mFnFn3(9,10). Because the tenth module contains the RGD cell attachment sequence while the ninth contains the synergy region, mFnFn3(9,10) has the cell attachment activity of intact fibronectin. Essentially complete signal assignments and approximately 1800 distance and angle restraints were derived from multidimensional heteronuclear NMR spectra. These restraints were used with a hybrid distance geometry/simulated annealing protocol to generate an ensemble of 20 NMR structures having no distance or angle violations greater than 0.3 A or 3 degrees. Although the beta-sheet core domains of the individual modules are well-ordered structures, having backbone atom rmsd values from the mean structure of 0.51(+/-0.12) and 0.40(+/-0.07) A, respectively, the rmsd of the core atom coordinates increases to 3.63(+/-1.41) A when the core domains of both modules are used to align the coordinates. The latter result is a consequence of the fact that the relative orientation of the two modules is not highly constrained by the NMR restraints. Hence, while structures of the beta-sheet core domains of the NMR structures are very similar to the core domains of the crystal structure of hFnFn3(9,10), the ensemble of NMR structures suggests that the two modules form a less extended and more flexible structure than the fully extended rod-like crystal structure. The radius of gyration, Rg, of mFnFn3(9,10) derived from small-angle neutron scattering measurements, 20.5(+/-0.5) A, agrees with the average Rg calculated for the NMR structures, 20.4 A, and is ca 1 A less than the value of Rg calculated for the X-ray structure. The values of the rotational anisotropy, D ||/D perpendicular, derived from an analysis of 15N relaxation data, range from 1.7 to 2.1, and are significantly less than the anisotropy of 2.67 predicted by hydrodynamic modeling of the crystal coordinates. In contrast, hydrodynamic modeling of the NMR coordinates yields anisotropies in the range of 1.9 to 2.7 (average 2.4(+/-0.2)), with NMR structures bent by more than 20 degrees relative the crystal structure having calculated anisotropies in best agreement with experiment. In addition, the relaxation parameters indicate that several loops in mFnFn3(9,10), including the RGD loop, are flexible on the nanosecond to picosecond time-scale. Taken together, our results suggest that, in solution, the limited set of interactions between the mFnFn3(9,10) modules position the RGD and synergy regions to interact specifically with cell surface integrins, and at the same time permit sufficient flexibility that allows mFnFn3(9,10) to adjust for some variation in integrin structure or environment.

Characterization of native and recombinant bone sialoprotein: delineation of the mineral-binding and cell adhesion domains and structural analysis of the RGD domain.

Bone sialoprotein is a small, sulfated, and phosphorylated integrin-binding glycoprotein apparently found only in tissues that eventually mineralize. Nondenatured bone sialoprotein (BSP) purified from rat osteosarcoma cell line (UMR 106-01 BSP) culture media is shown to have a hydroxyapatite Kd approximately 2.6 x 10(-9) M, perhaps the strongest affinity for this mineral of any of the matrix proteins. Both native BSP and a 47 kD fragment of UMR-BSP (Fragment 1 approximately 133A- approximately 265Y) are more potent inhibitors of seeded hydroxyapatite crystal growth than recombinant human BSP fragments lacking post-translational modifications. The recombinant proteins, however, do show reproducible inhibitory activity, suggesting that at least some of the strong mineral-binding properties are encoded directly within the protein sequence itself. BSP facilitates the adhesion of several cell types through its integrin binding (RGD) tripeptide sequence. Nuclear magnetic resonance (NMR) analysis of a 15N-enriched 59 amino acid recombinant domain containing the RGD tripeptide shows that the structure of this isolated domain is highly flexible with or without 5 mM calcium. Previous work has also shown that an endogenous fragment of UMR-BSP (Fragment 1) supports cell adhesion in the absence of the RGD sequence. In this report, non-RGD cell adhesion sites are localized within conserved amino- and carboxy-terminal tyrosine-rich domains of recombinant human BSP. Given the proximity of the latter non-RGD cell adhesion site to the RGD tripeptide, a model of BSP-receptor interactions is presented.

Tautomeric states of the active-site histidines of phosphorylated and unphosphorylated IIIGlc, a signal-transducing protein from Escherichia coli, using two-dimensional heteronuclear NMR techniques.

IIIGlc is an 18.1-kDa signal-transducing phosphocarrier protein of the phosphoenolpyruvate:glycose phosphotransferase system from Escherichia coli. The 1H, 15N, and 13C histidine ring NMR signals of both the phosphorylated and unphosphorylated forms of IIIGlc have been assigned using two-dimensional 1H-15N and 1H-13C heteronuclear multiple-quantum coherence (HMQC) experiments and a two-dimensional 13C-13C-1H correlation spectroscopy via JCC coupling experiment. The data were acquired on uniformly 15N-labeled and uniformly 15N/13C-labeled protein samples. The experiments rely on one-bond and two-bond J couplings that allowed for assignment of the signals without the need for the analysis of through-space (nuclear Overhauser effect spectroscopy) correlations. The 15N and 13C chemical shifts were used to determine that His-75 exists predominantly in the N epsilon 2-H tautomeric state in both the phosphorylated and unphosphorylated forms of IIIGlc, and that His-90 exists primarily in the N delta 1-H state in the unphosphorylated protein. Upon phosphorylation of the N epsilon 2 nitrogen of His-90, the N delta 1 nitrogen remains protonated, resulting in the formation of a charged phospho-His-90 moiety. The 1H, 15N, and 13C signals of the phosphorylated and unphosphorylated proteins showed only minor shifts in the pH range from 6.0 to 9.0. These data indicate that the pK alpha values for both His-75 and His-90 in IIIGlc and His-75 in phospho-IIIGlc are less than 5.0, and that the pK alpha value for phospho-His-90 is greater than 10. The results are presented in relation to previously obtained structural data on IIIGlc, and implications for proposed mechanisms of phosphoryl transfer are discussed.

Three-dimensional solution structure of the HIV-1 protease complexed with DMP323, a novel cyclic urea-type inhibitor, determined by nuclear magnetic resonance spectroscopy.

The three-dimensional solution structure of the HIV-1 protease homodimer, MW 22.2 kDa, complexed to a potent, cyclic urea-based inhibitor, DMP323, is reported. This is the first solution structure of an HIV protease/inhibitor complex that has been elucidated. Multidimensional heteronuclear NMR spectra were used to assemble more than 4,200 distance and angle constraints. Using the constraints, together with a hybrid distance geometry/simulated annealing protocol, an ensemble of 28 NMR structures was calculated having no distance or angle violations greater than 0.3 A or 5 degrees, respectively. Neglecting residues in disordered loops, the RMS deviation (RMSD) for backbone atoms in the family of structures was 0.60 A relative to the average structure. The individual NMR structures had excellent covalent geometry and stereochemistry, as did the restrained minimized average structure. The latter structure is similar to the 1.8-A X-ray structure of the protease/DMP323 complex (Chang CH et al., 1995, Protein Science, submitted); the pairwise backbone RMSD calculated for the two structures is 1.22 A. As expected, the mismatch between the structures is greatest in the loops that are disordered and/or flexible. The flexibility of residues 37-42 and 50-51 may be important in facilitating substrate binding and product release, because these residues make up the respective hinges and tips of the protease flaps. Flexibility of residues 4-8 may play a role in protease regulation by facilitating autolysis.

Anti-HIV and anti-tumor protein MAP30, a 30 kDa single-strand type-I RIP, shares similar secondary structure and beta-sheet topology with the A chain of ricin, a type-II RIP.

MAP30 is a 30 kDa single-stranded, type-I ribosome inactivating protein (RIP) possessing anti-tumor and anti-HIV activities. It binds both ribosomal RNA and the HIV-1 long-terminal repeat DNA. To understand the structural basis for MAP30 activities, we undertook the study of MAP30 by solution NMR spectroscopy. We report nearly complete 1H, 13C, and 15N chemical shift assignments of its 263 amino acids. Based upon an analysis of secondary 13C chemical shifts, 3J(HNHA) coupling constants, hydrogen exchange data, and nuclear Overhauser effect patterns, we find that the secondary structure and beta-sheet topology of MAP30 are very similar to those of the ricin A chain, a subunit of the well-known type-II RIP, even though two proteins display distinct activities. We therefore suggest that MAP30 and ricin A chain share a similar three-dimensional fold, and that the reported functional differences between two proteins arise primarily from differences in local three-dimensional structure and other structural properties such as surface electrostatic potentials.

An efficient NMR approach for obtaining sequence-specific resonance assignments of larger proteins based on multiple isotopic labeling.

By simultaneously incorporating in a protein 13C-carbonyl- and 15N-labeled amino acids with different levels of enrichment, characteristics asymmetric doublet-like patterns are observed for 15N nuclei that are directly adjacent to the 13C1-labeled residues, providing unambiguous identification of a large number of unique dipeptide fragments of the protein. Additional assignments and qualitative structural information can be obtained from such a selectively labeled protein by recording multiple bond correlation spectra. The procedure is demonstrated for the protein calmodulin, complexed with calcium.

Complete 1H and 13C assignment of Lys and Leu sidechains of staphylococcal nuclease using HCCH-COSY and HCCH-TOCSY 3D NMR spectroscopy.

Complete proton and carbon sidechain assignments are reported for 22 lysine and 11 leucine residues in staphylococcal nuclease, an enzyme with 149 residues. These assignments are readily obtained in a direct manner from the correlations observed in the 3D HCCH-COSY and HCCH-TOCSY spectra and the known protein backbone assignments. These assignments open the way to detailed studies of the sidechain structure and dynamics at the active site, in the hydrophobic core and on the surface of the protein.

Elucidation of the poly-L-proline binding site in Acanthamoeba profilin I by NMR spectroscopy.

The multifunctional protein profilin is one of the most abundant proteins in the cytoplasm and is thought to regulate actin assembly and the phosphoinositide signaling pathway. Profilin binds to several different ligands including actin, poly-L-proline, and the head groups of polyphosphoinositides. Knowledge of profilin/ligand interactions is important for understanding the physiology of profilin in the cell. As a first step in the characterization of profilin/ligand complexes, we have studied a profilin/poly-L-proline complex in solution using high resolution NMR spectroscopy. Analysis of profilin NOE's and chemical shift data indicates that the protein secondary structure is conserved upon binding to poly-L-proline and that the binding site is located between the N- and C-terminal helices in a region rich in highly conserved aromatic sidechains. This site is adjacent to the proposed binding site for actin. In addition, the rate constant for dissociation of the complex is found to be 1.6 +/- 0.2 x 10(4) s-1.

Use of the carbonyl chemical shift to relieve degeneracies in triple-resonance assignment experiments.

We illustrate an approach that uses the backbone carbonyl chemical shift to relieve resonance overlaps in triple-resonance assignment experiments conducted on protein samples. We apply this approach to two cases of simultaneous overlaps: those of ((1)H(N), (15)N) spin pairs and those of ((1)H(alpha), (13)C(alpha)) spin pairs in residues preceding prolines. For these cases we employed respectively CBCACO(N)H and H(CA)CON experiments, simple variants of the commonly used CBCA(CO)NH and HCA(CO)N experiments obtained by replacing one of the indirect dimensions with a carbonyl dimension. We present data collected on ribosomal protein S4 using these experiments, along with overlap statistics for four other polypeptides ranging in size from 76 to 263 residues. These data indicate that the CBCACO(N)H, in combination with the CBCA(CO)NH, can relieve >83% of the ((1)H(N), (15)N) and ((1)H(N), (13)C') overlaps for these proteins. The data also reveal how the H(CA)CON experiment successfully completed the assignment of triply and quadruply degenerate X-Pro spin systems in a mobile, proline-rich region of S4, even when X was a glycine. Finally, we discuss the relative sensitivities of these experiments compared to those of existing sequences, an analysis that reinforces the usefulness of these experiments in assigning extensively overlapped and/or proline-rich sequences in proteins.

Transverse 1H cross relaxation in 1H-15N correlated 1H CPMG experiments.

Transverse 1H cross relaxation was observed in Carr-Purcell-Meiboom-Gill (CPMG) experiments by recording 15N-1H correlated spectra of amides in HIV protease that was perdeuterated at nonexchangeable sites. Perdeuteration suppresses 1H-1H J coupling and improves spectral resolution and sensitivity. Measurements of cross-peak intensities, arising from cross relaxation, were made as a function of (i) Deltaf, the frequency difference between the spins, and (ii) tauCPMG, one-half of the duration between CPMG pi pulses. Cross peaks were observed when tauCPMG was less than 1/(2Deltaf), in agreement with theoretical calculations.

Enamel matrix: structural proteins.

Cell-free, fetal bovine enamel tissue was examined intact by high resolution. 13C Fourier transform, nuclear magnetic resonance spectroscopy. Two types of protein chains were observed under these conditions, one exhibiting rapid mobility and accounting for approximately two-thirds of the enamel matrix, while the other exhibited restricted or anisotropic segmental motion and accounted for the remaining third of the matrix. Sequential extraction of this fetal enamel under non-degradative conditions with dissociative solvents yielded two biochemically distinct populations of matrix protein. As expected, the bulk of the matrix consisted of proline-rich amelogenins, although the SDS-gel electrophoresis molecular weights for these proteins were somewhat higher than those reported using other extraction methods. Approximately fifteen percent of the total matrix consisted of much higher molecular weight phosphoproteins (46,000-72,000 daltons) whose amino acid composition closely resembled that reported for mature enamel protein. These high molecular weight proteins were tightly bound to the fetal enamel apatite crystallites.

A 13C magnetic resonance study of embryonic chick aorta.

13C nmr spectra of defatted aorta, obtained from chick embryos, ranging in age from 13 to 20 days, showed that linewidths were independent of the age of the embryo, this, in spite of the fact that crosslinking increases with age. As expected the dipolar decoupled spectrum of defatted aorta had a larger C-H signal intensity than the scalar decoupled spectrum, since collagen in the sample contributes signal intensity only in the former case. Significantly, the dipolar decoupled spectrum of autoclaved aorta also had a larger C-H signal intensity. This result indicates that ca. 30% of the carbons in the tissue, swollen by 0.15 M NaCl, has restricted motion at 37 degrees. Preliminary data indicate that tissue culture techniques can be used to enrich specific amino acids in chick aorta with 13C, thereby affording a means to study possible differences in molecular structure of the chemically distinct regions in elastin.