Assignment strategies in homonuclear three-dimensional 1H NMR spectra of proteins

Real-time nuclear magnetic resonance spectroscopy in the study of biomolecular kinetics and dynamics

The review describes the application of nuclear magnetic resonance (NMR) spectroscopy to study kinetics of folding, refolding and aggregation of proteins, RNA and DNA. Time-resolved NMR experiments can be conducted in a reversible or an irreversible manner. In particular, irreversible folding experiments pose large requirements for (i) signal-to-noise due to the time limitations and (ii) synchronising of the refolding steps. Thus, this contribution discusses the application of methods for signal-to-noise increases, including dynamic nuclear polarisation, hyperpolarisation and photo-CIDNP for the study of time-resolved NMR studies. Further, methods are reviewed ranging from pressure and temperature jump, light induction to rapid mixing to induce rapidly non-equilibrium conditions required to initiate folding.

Solution structure of the cytoplasmic linker between domain III-S6 and domain IV-S1 (III-IV linker) of the rat brain sodium channel in SDS micelles

The solution structure of the 36-mer peptide MP-5A in SDS micelles was investigated by CD and (1)H-NMR spectroscopies. The MP-5A was dissected from the cytoplasmic linker (K1482-A1517) connecting domain III-segment 6 (IIIS6) and domain IV-segment 1 (IVS1; III-IV linker) of the rat brain type IIA sodium channel. The molecular energy calculations including nuclear Overhauser effect and dihedral angle restraints gave a well-converged set of the structures of MP-5A for the region between I1488 and S1506. It was found that a large hydrophobic cluster is formed by I1488-F1489-M1490 (IFM motif), Y1497-Y1498, and M1501, which may be related to the fast inactivation process of the sodium channel. The solvent-accessible surface area of the IFM motif (195 A(2)), which is known to work essentially as an inactivation gate particle to occlude the ion permeation pore, gave the free energy (DeltaG) of stabilization of -3.9 kcal mol(-1) as a result of the hydrophobic interactions with its receptor. This value agreed well with the free energy of binding (inactivation) of -4.1 kcal mol(-1) calculated for the equilibrium between the open and the inactivated states of the sodium channels. It is concluded that the fast inactivation of the sodium channel is achieved by the environmental polarity-dependent conformational switching at the IFM motif, in response to the voltage-dependent activation and the movement of the S4 segments of the sodium channel.

The solution conformational features of two highly homologous antigenic peptides of foot-and-mouth disease virus serotype A, variant A and USA, correlate with their serological properties

The solution structure of a peptide corresponding to the VP1 region 141-160 of foot-and-mouth disease virus (FMDV) serotype A variant USA has been studied by NMR and computer calculations and compared with the results from a study on a highly homologous peptide deriving from serotype A, variant A. The two peptides differ in their serological behavior and contain the immunodominant epitope of the virus which partly overlaps with its receptor binding region. Distance constraints, derived both from 2D and 3D homonuclear NMR and 2D-heteronuclear NMR experiments, were combined with DG calculations to yield 50 structures. After refinement through EM and restrained molecular dynamics simulations the selected structures shared several general features. In particular the 151-158 region was a helix in all cases while a large loop similar to that found in peptide A but comprising less residues and stabilized by an H-bond between the side chains of D147 and S150 was found in the majority of structures. A further loop, common to all structures, was identified around the RGD sequence (145-147). This was different from that found in the corresponding region of peptide A as were the conformations of the individual residues within the RGDX sequence. The different structural features shown by the two peptides were rationalized in terms of the S148 (peptide A) to F148 (peptide USA) mutation. The second mutation, that at position 153 (L in A, P in USA) did not appear to affect the structure of the peptide significantly although the different dimensions of the loop in the central region and the type of H-bond stabilizing it could be potentially ascribed to this second mutation. All criteria used pointed to different structural features for the two peptides consistent with their serological behaviour.

1H and 13C NMR assignment and secondary structure of Chlorobium limicola f. thiosulfatophilum ferrocytochrome c555

The 1H resonances of the ferrocytochrome c555 from the anaerobic green sulfur bacterium Chlorobium limicola f thio-sulfatophilum (strain Tassajara) have been assigned. Identification of spin systems and sequential assignment of 1H was accomplished by automated assignment computer programs followed by manual verification. In addition, 13C resonances have been extensively assigned by HSQC experiments at natural abundance. As determined by short-range NOE connectivities, 13C alpha chemical shifts, and HN exchange experiments, the secondary structure consists of 3 helices ranging from residues 3-13, 43-53 and 70-86. Interestingly, the second helix is significantly longer than observed by X-ray crystallography [1977, Proc. Natl. Acad. Sci. USA 74, 5244-5247]. A topological model of the cytochrome c555 is presented based on a small number of long-range NOE contacts. The helices are shown to pack onto the heme according to the pattern common to all class I cytochromes c.

Solution structure of porcine pancreatic procolipase as determined from 1H homonuclear two-dimensional and three-dimensional NMR

Procolipase is the precursor of colipase, which acts as protein cofactor for the activity of pancreatic lipase. The solution structure of procolipase has been determined by 1H NMR using two- and three-dimensional measurements. The secondary structure determination identified two separate three-stranded beta-sheet regions with concomitant hydrogen bond patterns. The tertiary structure of the protein was determined using 863 non-trivial proton--proton distance constraints, 14 hydrogen bond distance constraints and 55 phi and 25 X1 dihedral constraints. The structure that was obtained from distance geometry and energy refinement contains three highly disordered loops as well as a disordered N- and C-terminal region. The remaining part of the structure is well defined with a root-mean-square deviation (rmsd) relative to the average of 0.09 +/- 0.02 nm for backbone atoms (residues 11-30, 37-50, 57-69, 83-89). The protein comprises two identical domains, each containing a three-strand beta-sheet and two disulfide bonds: a 15-residue region in each domain superimposes with 0.07 nm rmsd, measured on backbone atoms. The solution structure is nearly identical to the crystal structure. It is in agreement with previous NMR data and, in combination with these data, supports the current model of procolipase micelle interaction and the lipase activation by colipase.

Sequence analysis by NMR spectroscopy of the peptide lantibiotic epilancin K7 from Staphylococcus epidermidis K7

The amino acid sequence of the novel lantibiotic epilancin K7 from Staphylococcus epidermidis K7 was determined by NMR spectroscopy. NMR spectroscopy was used because sequencing by conventional Edman degradation techniques was prohibited by internal sequence blocks owing to the presence of modified residues. Epilancin K7 consists of 31 residues, including two alpha,beta-didehydroalanine (one-letter code U) and two alpha,beta-didehydrobutyrine (O) residues, one lanthionine (A-S-A), two beta-methyllanthionines (A*-S-A), and six lysines. Epilancin K7 has a molecular mass of 3032 +/- 1.5 Da. The amino acid sequence of epilancin K7 was derived from both through-space dipolar proton-proton interactions and through-bond scalar proton-carbon interactions as detected by two-dimensional 1H-NOESY, 1H-ROESY and three-dimensional 1H-TOCSY-NOESY, and by two-dimensional 1H,13C-heteronuclear multiple-bond correlation spectroscopy, respectively. The sequence is as follows: [sequence: see text] The N-terminal residue X partly resembles an alanine but its exact nature is unclear. The organization of the sulfide-bridge-containing (beta-methyl-)lanthionines was revealed by 1H-NMR and 1H,13C-NMR spectroscopy. Epilancin K7 has a linear structure and a high positive net charge, and therefore is classified as a type-A lantibiotic. NMR analysis of a degraded though still active form of epilancin K7 showed that two N-terminal residues of epilancin K7 were missing, owing to decomposition at the alpha,beta-didehydro alanine at position 3; it was called the epilancin K7-(3-31)-peptide (peptide fragment of epilancin K7 consisting of positions 3-31). The usefulness of three-dimensional 1H-TOCSY-NOESY, and two-dimensional 1H,13C-heteronuclear multiple-bond correlation spectroscopy at natural abundance for the study of (modified) polypeptides is demonstrated.

Evaluation of an algorithm for the automated sequential assignment of protein backbone resonances: a demonstration of the connectivity tracing assignment tools (CONTRAST) software package

The peptide sequential assignment algorithm presented here was implemented as a macro within the CONnectivity TRacing ASsignment Tools (CONTRAST) computer software package. The algorithm provides a semi- or fully automated global means of sequentially assigning the NMR backbone resonances of proteins. The program's performance is demonstrated here by its analysis of realistic computer-generated data for IIIGlc, a 168-residue signal-transducing protein of Escherichia coli [Pelton et al. (1991) Biochemistry, 30, 10043-10057]. Missing experimental data (19 resonances) were generated so that a complete assignment set could be tested. The algorithm produces sequential assignments from appropriate peak lists of nD NMR data. It quantifies the ambiguity of each assignment and provides ranked alternatives. A 'best first' approach, in which high-scoring local assignments are made before and in preference to lower scoring assignments, is shown to be superior (in terms of the current set of CONTRAST scoring routines) to approaches such as simulated annealing that seek to maximize the combined scores of the individual assignments. The robustness of the algorithm was tested by evaluating the effects of imposed frequency imprecision (scatter), added false signals (noise), missing peaks (incomplete data), and variation in user-defined tolerances on the performance of the algorithm.

Resonance assignment, cysteine-pairing elucidation and secondary-structure determination of capsicein, an alpha-elicitin, by three-dimensional 1H NMR

Difficulties encountered in the interpretation of two-dimensional NMR spectra of proteins exceeding roughly 100 amino acids, including resonance overlap and line broadening due to longer correlation times and/or aggregation phenomena, can be overcome by using three-dimensional 1H-NMR experiments. The improvement of spectral resolution using these experiments allows the size of molecules amenable to structure determination by NMR spectroscopy to be extended. A three-dimensional non-selective homonuclear Hartmann-Hahn/nuclear Overhauser effect spectroscopy experiment was performed on capsicein, a 10161-Da elicitin secreted by the Phytophthora capsici fungus. Sequential assignment and secondary structure determination is illustrated for beta-sheet, alpha-helix and loop structures by analysis of planar cross sections perpendicular to the omega 2 or omega 3 axis at the amide proton resonance frequencies. Cysteine pairing was established in the course of the investigation. The secondary structure topology of the molecule is composed of five helices and an antiparallel beta-sheet. Four of the helices compose two pairs running antiparallel while the last one is parallel to the beta-sheet.

1H resonance assignments and secondary structure of the carbon monoxide complex of soybean leghemoglobin determined by homonuclear two-dimensional and three-dimensional NMR spectroscopy

Homonuclear two-dimensional and three-dimensional 1H-NMR spectroscopy has been utilized to study the 15.9-kDa protein soybean leghemoglobin. NMR experiments were performed on the diamagnetic carbon monoxide complex at two temperatures and two pH values. Sequence-specific assignments have been made for 94% of the backbone and approximately 70% of the expected side-chain resonances. The secondary structure of leghemoglobin in solution has been determined on the basis of NOE connectivity patterns, hydrogen exchange and chemical-shift analyses. Leghemoglobin consists of seven helices and, unlike mammalian myoglobins, is missing the D helix. Instead an extended loop, the CE loop, is observed which might have importance for ligand entry into and exit from the protein interior. The hydrogen exchange behavior for the F helix and at the beginning of the A helix suggests different dynamic stability compared to other helical regions in leghemoglobin. Population of a second protein conformation, in which there is perturbation at the A-G-H helix interface, is observed at low pH.

Kinetics of amide proton exchange in parvalbumin studied by 1H 2-D NMR. A comparison of the calcium and magnesium loaded forms

The amide proton exchange rates have been measured for the pike parvalbumin loaded either with calcium (PaCa2) or with magnesium (PaMg2) by using 2-D total correlation spectroscopy experiments. The differences in the exchange rates observed between these two species were unexpected when compared with the small conformational changes induced in parvalbumin by the Ca/Mg exchange. With the calcium-loaded protein (PaCa2), a significant difference was observed for the amide proton exchange rates of residues located in the N-terminal domain AB in contrast to the slower exchange rates that were observed in the CD and EF domains. Such a difference does not exist for PaMg2, where faster exchange rates are observed over all the sequence. Since amide proton exchange rates are the signature of the solvent's accessibility in proteins, we interpreted our results in terms of difference of the equilibria between 'closed-states' and 'opened-states' for individual amide protons of the protein when calcium was replaced by magnesium. The CD and EF domains, and to a lesser extent the AB domain, would be more rigid when the protein was loaded with calcium ions. For the magnesium-loaded parvalbumin (PaMg2) the faster exchange rates we observed could be rationalized by a more flexible structure than in the case of the PaCa2.

Two- and three-dimensional proton NMR studies of apo-neocarzinostatin

Neocarzinostatin (NCS) is an antitumor protein from Streptomyces carzinostaticus that is identical in apo-protein sequence with mitomalcin (MMC) from Streptomyces malayensis. We describe the use of apo-NCS as a model system for applying combined two- and three-dimensional (2D and 3D) proton NMR spectroscopy to the structure determination of proteins (Mr greater than 10K) without isotope labeling. Strategies aimed at accurately assigning overlapped 2D cross-peaks by using semiautomated combined 2D and 3D data analysis are developed. Using this approach, we have assigned 99% of the protons, including those of the side chains, and identified about 1270 intra- and interresidue proton-proton interactions (fixed distances are not included) in apo-NCS. Comparing our results with those reported recently on 2D NMR studies of apo-NCS [Adjadj, E., Mispelter, J., Quiniou, E., Dimicoli, J.-L., Favadon, V., & Lhoste, J.-M. (1990) Eur. J. Biochem. 190, 263-271; Remerowski M. L., Glaser, S. J., Sieker, L., Samy, T. S. A., & Drobny, G. P. (1990) Biochemistry 29, 8401-8409] demonstrated advantages of proton 3D NMR spectroscopy in protein spectral assignments. We are able to obtain more complete proton resonance and secondary structural assignments and find several misassignments in the earlier report. Strategies utilized in this work should be useful for developing automation procedures for spectral assignments.

Three-dimensional correlated NMR study of Megasphaera elsdenii flavodoxin in the oxidized state

The value of a three-dimensional (3D) non-selective total correlation/nuclear Overhauser enhancement spectroscopy (TOCSY-NOESY) spectrum for making sequential resonance assignments in proteins is demonstrated using the relatively large Megasphaera elsdenii flavodoxin (molecular mass 15 kDa) in the oxidized state. An easy and concise method for the analysis of 3D-NMR spectra and a strategy for the resonance assignment of 3D-NMR protein spectra is introduced. In this context, non-selective TOCSY-NOESY is compared with selective TOCSY-NOESY and non-selective NOESY-TOCSY. Sequential assignments in various secondary structure elements of flavodoxin are made using the method of analysis introduced. NOEs not previously identified in 2D-NMR spectra due to resonance overlap are found in the 3D Clean-TOCSY-NOESY spectrum. Also additional side-chain assignments could be made.