Determination of Multiple φ-Torsion Angles in Proteins by Selective and Extensive C Labeling and Two-Dimensional Solid-State NMR

We describe an approach to efficiently determine the backbone conformation of solid proteins that utilizes selective and extensive ¹³C labeling in conjunction with two-dimensional magic-angle-spinning NMR. The selective ¹³C labeling approach aims to reduce line broadening and other multispin complications encountered in solid-state NMR of uniformly labeled proteins while still enhancing the sensitivity of NMR spectra. It is achieved by using specifically labeled glucose or glycerol as the sole carbon source in the protein expression medium. For amino acids synthesized in the linear part of the biosynthetic pathways, [1-¹³C]glucose preferentially labels the ends of the side chains, while [2-¹³C]glycerol labels the C^α of these residues. Amino acids produced from the citric-acid cycle are labeled in a more complex manner. Information on the secondary structure of such a labeled protein was obtained by measuring multiple backbone torsion angles φ simultaneously, using an isotropic–anisotropic 2D correlation technique, the HNCH experiment. Initial experiments for resonance assignment of a selectively ¹³C labeled protein were performed using ¹⁵N–¹³C 2D correlation spectroscopy. From the time dependence of the ¹⁵N–¹³C dipolar coherence transfer, both intraresidue and interresidue connectivities can be observed, thus yielding partial sequential assignment. We demonstrate the selective ¹³C labeling and these 2D NMR experiments on a 8.5-kDa model protein, ubiquitin. This isotope-edited NMR approach is expected to facilitate the structure determination of proteins in the solid state.     

Triple resonance experiments for aligned sample solid-state NMR of (13)C and (15)N labeled proteins

Initial steps in the development of a suite of triple-resonance (1)H/(13)C/(15)N solid-state NMR experiments applicable to aligned samples of (13)C and (15)N labeled proteins are described. The experiments take advantage of the opportunities for (13)C detection without the need for homonuclear (13)C/(13)C decoupling presented by samples with two different patterns of isotopic labeling. In one type of sample, the proteins are approximately 20% randomly labeled with (13)C in all backbone and side chain carbon sites and approximately 100% uniformly (15)N labeled in all nitrogen sites; in the second type of sample, the peptides and proteins are (13)C labeled at only the alpha-carbon and (15)N labeled at the amide nitrogen of a few residues. The requirement for homonuclear (13)C/(13)C decoupling while detecting (13)C signals is avoided in the first case because of the low probability of any two (13)C nuclei being bonded to each other; in the second case, the labeled (13)C(alpha) sites are separated by at least three bonds in the polypeptide chain. The experiments enable the measurement of the (13)C chemical shift and (1)H-(13)C and (15)N-(13)C heteronuclear dipolar coupling frequencies associated with the (13)C(alpha) and (13)C' backbone sites, which provide orientation constraints complementary to those derived from the (15)N labeled amide backbone sites. (13)C/(13)C spin-exchange experiments identify proximate carbon sites. The ability to measure (13)C-(15)N dipolar coupling frequencies and correlate (13)C and (15)N resonances provides a mechanism for making backbone resonance assignments. Three-dimensional combinations of these experiments ensure that the resolution, assignment, and measurement of orientationally dependent frequencies can be extended to larger proteins. Moreover, measurements of the (13)C chemical shift and (1)H-(13)C heteronuclear dipolar coupling frequencies for nearly all side chain sites enable the complete three-dimensional structures of proteins to be determined with this approach.     

Application of REDOR subtraction for filtered MAS observation of labeled backbone carbons of membrane-bound fusion peptides

Clean MAS observation of 13C-labeled carbons in membrane-bound HIV-1 and influenza fusion peptides was made by using a rotational-echo double-resonance spectroscopy (REDOR) filter of directly bonded 13C-15N pairs. The clean filtering achieved with the REDOR approach is superior to filtering done with sample difference spectroscopy. In one labeling approach, the peptide had labels at a single 13C carbonyl and its directly bonded 15N. The resulting chemical shift distribution of the filtered signal is used to assess the distribution of local secondary structures at the labeled carbonyl. For the influenza peptide, the Leu-2 carbonyl chemical shift distribution is shown to vary markedly with lipid and detergent composition, as well as peptide:lipid ratio, suggesting that the local peptide structure also has a strong dependence on these factors. Because most carboxylic- and amino-labeled amino acids are commercially available, this REDOR approach should have broad applicability to chemically synthesized peptides as well as bacterially synthesized proteins. In a second labeling approach, the HIV-1 fusion peptide had U-13C, 15N labeling over three sequential residues. When a 1.6 ms REDOR dephasing time is used, only backbone 13C signals are observed. The resulting spectra are used to determine spectral linewidths and to assess feasibility of assignment of uniformly labeled peptide.     

Optimal control based NCO and NCA experiments for spectral assignment in biological solid-state NMR spectroscopy

We present novel pulse sequences for magic-angle-spinning solid-state NMR structural studies of (13)C,(15)N-isotope labeled proteins. The pulse sequences have been designed numerically using optimal control procedures and demonstrate superior performance relative to previous methods with respect to sensitivity, robustness to instrumental errors, and band-selective excitation profiles for typical biological solid-state NMR applications. Our study addresses specifically (15)N to (13)C coherence transfers being important elements in spectral assignment protocols for solid-state NMR structural characterization of uniformly (13)C,(15)N-labeled proteins. The pulse sequences are analyzed in detail and their robustness towards spin system and external experimental parameters are illustrated numerically for typical (15)N-(13)C spin systems under high-field solid-state NMR conditions. Experimentally the methods are demonstrated by 1D (15)N-->(13)C coherence transfer experiments, as well as 2D and 3D (15)N,(13)C and (15)N,(13)C,(13)C chemical shift correlation experiments on uniformly (13)C,(15)N-labeled ubiquitin.     

Sample optimization and identification of signal patterns of amino acid side chains in 2D RFDR spectra of the alpha-spectrin SH3 domain

Future structural investigations of proteins by solid-state CPMAS NMR will rely on uniformly labeled protein samples showing spectra with an excellent resolution. NMR samples of the solid alpha-spectrin SH3 domain were generated in four different ways, and their (13)C CPMAS spectra were compared. The spectrum of a [u-(13)C, (15)N]-labeled sample generated by precipitation shows very narrow (13)C signals and resolved scalar carbon-carbon couplings. Linewidths of 16-19 Hz were found for the three alanine C(beta )signals of a selectively labeled [70% 3-(13)C]alanine-enriched SH3 sample. The signal pattern of the isoleucine, of all prolines, valines, alanines, and serines, and of three of the four threonines were identified in 2D (13)C-(13)C RFDR spectra of the [u-(13)C, (15)N]-labeled SH3 sample. A comparison of the (13)C chemical shifts of the found signal patterns with the (13)C assignment obtained in solution shows an intriguing match.     

Amino-acid selective experiments on uniformly 13C and 15N labeled proteins by MAS NMR: Filtering of lysines and arginines

Amino-acid selective magic-angle spinning (MAS) NMR experiments can aid the assignment of ambiguous cross-peaks in crowded spectra of solid proteins. In particular for larger proteins, data analysis can be hindered by severe resonance overlap. In such cases, filtering techniques may provide a good alternative to site-specific spin-labeling to obtain unambiguous assignments that can serve as starting points in the assignment procedure. In this paper we present a simple pulse sequence that allows selective excitation of arginine and lysine residues. To achieve this, we make use of a combination of specific cross-polarization for selective excitation [M. Baldus, A.T. Petkova, J. Herzfeld, R.G. Griffin, Cross polarization in the tilted frame: assignment and spectral simplification in heteronuclear spin systems, Mol. Phys. 95 (1998) 1197-1207.] and spin diffusion for transfer along the amino-acid side-chain. The selectivity of the filter is demonstrated with the excitation of lysine and arginine side-chain resonances in a uniformly 13C and 15N labeled protein preparation of the alpha-spectrin SH3 domain. It is shown that the filter can be applied as a building block in a 13C-13C lysine-only correlation experiment.     

Intraresidue HNCA and COHNCA experiments for protein backbone resonance assignment

Two novel experiments, intra-HNCA and intra-COHNCA, are presented for sequential backbone resonance assignment of (13)C, (15)N labeled proteins. The advantage with respect to conventional pulse schemes is the suppression of the sequential (15)N-->(13)C(alpha) coherence transfer pathway, which can be separately obtained from a HNCOCA correlation experiment. This results in a two-fold reduction of the number of detected correlation peaks. Spectral simplification is especially important for efficient automated assignment protocols as required in the context of high-throughput protein studies by NMR. The performance of the new experiments is demonstrated on an 18-kDa protein fragment of the E. coli sulfite reductase and compared to conventional techniques in terms of sensitivity and resolution.     

DEPT spectral editing in HCCONH-type experiments. Application to fast protein backbone and side chain assignment

2D DEPT-H(alpha,beta)C(alpha,beta)(CO)NH and 2D CT-DEPT-HC(CO)NH-TOCSY experiments are presented which allow fast resonance assignment of aliphatic protein side chains. In these 2D reduced-dimensionality experiments, two or three nuclei are frequency labeled in the indirect dimension. DEPT spectral editing reduces the number of correlation peaks detected in each 2D spectrum, and helps in amino-acid-type determination during sequential backbone resonance assignment. Applications are shown for a small 68-residue, and a highly deuterated 167-residue protein. The new experiments complement the set of 2D HNX correlation experiments, previously proposed for fast protein resonance assignment [J. Biomol. NMR, 27 (2003) 57].     

Assignment of congested NMR spectra: carbonyl backbone enrichment via the Entner-Doudoroff pathway

In NMR spectra of complex proteins, sparse isotope enrichment can be important, in that the removal of many (13)C-(13)C homonuclear J-couplings can narrow the lines and thereby facilitate the process of spectral assignment and structure elucidation. We present a simple scheme for selective yet extensive isotopic enrichment applicable for production of proteins in organisms utilizing the Entner-Doudoroff (ED) metabolic pathway. An enrichment scheme so derived is demonstrated in the context of a magic-angle spinning solid-state NMR (MAS SSNMR) study of Pf1 bacteriophage, the host of which is Pseudomonas aeruginosa, strain K (PAK), an organism that uses the ED pathway for glucose catabolism. The intact and infectious Pf1 phage in this study was produced by infected PAK cells grown on a minimal medium containing 1-(13)C d-glucose ((13)C in position 1) as the sole carbon source, as well as (15)NH(4)Cl as the only nitrogen source. The 37MDa Pf1 phage consists of about 93% major coat protein, 1% minor coat proteins, and 6% single-stranded, circular DNA. As a consequence of this composition and the enrichment scheme, the resonances in the MAS SSNMR spectra of the Pf1 sample were almost exclusively due to carbonyl carbons in the major coat protein. Moreover, 3D heteronuclear NCOCX correlation experiments also show that the amino acids leucine, serine, glycine, and tyrosine were not isotopically enriched in their carbonyl positions (although most other amino acids were), which is as expected based upon considerations of the ED metabolic pathway. 3D NCOCX NMR data and 2D (15)N-(15)N data provided strong verification of many previous assignments of (15)N amide and (13)C carbonyl shifts in this highly congested spectrum; both the semi-selective enrichment patterns and the narrowed linewidths allowed for greater certainty in the assignments as compared with use of uniformly enriched samples alone.     

Measurement of relaxation rates of N(H) and H(alpha) backbone protons in proteins with tailored initial conditions.

Several methods are presented for the selective determination of spin-lattice and spin-spin relaxation rates of backbone protons in labeled proteins. The relaxation rates of amide protons in (15)N labeled proteins can be measured by using two-way selective cross-polarization (SCP). The measurement of H(alpha) relaxation rates can be achieved by combining this method with homonuclear Hartmann-Hahn transfer using doubly selective irradiation. Various schemes for selective or nonselective inversion of the longitudinal proton magnetization lead to different initial recovery rates. The methods have been applied to lysine K6 in (15)N-labeled human ubiquitin and to leucine L5 in (15)N- and (13)C-labeled octapeptide YG*G*F*LRRI (GFL) in which the marked residues are (15)N- and (13)C-labeled.     

The DQ-HN[CACB] and DQ-HN(CO)[CACB] sequences with evolution of double quantum Calpha-Cbeta coherences

The new variant of known HNCACB and HN(CO)CACB techniques is proposed that employs excitation and evolution of double quantum Calpha-Cbeta coherences. The most important features of the new method are: increased signal dispersion, lack of splittings due to 1J(Calpha-Cbeta) spin-spin couplings, and absence of accidental cancellations of positive and negative signals. The acquisition of both DQ-HN[CACB] and DQ-HN(CO)[CACB] techniques enables sequential assignment of protein backbone, using only Calpha-Cbeta DQ-frequencies. The determination of all Calpha and Cbeta chemical shifts requires, however, a comparison with HN(CO)CA or HNCA spectra. Examples of applications of the DQ-HN[CACB] and DQ-HN(CO)[CACB] experiments are presented, employing the 2D Reduced Dimensionality approach for 13C, 15N-labeled ubiquitin, and the 3D acquisition for 13C, 15N-double labeled Ca2+ -binding bovine S100A1 protein in the apo state (21 kDa) with overall correlation time of 8.1 ns.     

Alanine check points in HNN and HN(C)N spectra

Rapid resonance assignment is a key requirement in structural genomics research by NMR. In this context we present here two new pulse sequences, namely, HNN-A and HN(C)N-A that have been developed by simple modification of the previously described pulse sequences, HNN and HN(C)N [S.C. Panchal, N.S. Bhavesh, R.V. Hosur, Improved 3D triple resonance experiments, HNN and HN(C)N, for H(N) and 15N sequential correlations in (13C, 15N) labeled proteins: application to unfolded proteins, J. Biomol. NMR, 20 (2001) 135-147]. These increase the number of start/check points in HNN and/or HN(C)N spectra and hence help in pacing up resonance assignment in proteins.     

Sample Optimization and Identification of Signal Patterns of Amino Acid Side Chains in 2D RFDR Spectra of the α-Spectrin SH3 Domain

Future structural investigations of proteins by solid-state CPMAS NMR will rely on uniformly labeled protein samples showing spectra with an excellent resolution. NMR samples of the solid α-spectrin SH3 domain were generated in four different ways, and their ¹³C CPMAS spectra were compared. The spectrum of a [u-¹³C, ¹⁵N]-labeled sample generated by precipitation shows very narrow ¹³C signals and resolved scalar carbon–carbon couplings. Linewidths of 16–19 Hz were found for the three alanine C^β signals of a selectively labeled [70% 3-¹³C]alanine-enriched SH3 sample. The signal pattern of the isoleucine, of all prolines, valines, alanines, and serines, and of three of the four threonines were identified in 2D ¹³C–¹³C RFDR spectra of the [u-¹³C,¹⁵N]-labeled SH3 sample. A comparison of the ¹³C chemical shifts of the found signal patterns with the ¹³C assignment obtained in solution shows an intriguing match.     

Rotational resonance in uniformly 13C-labeled solids: effects on high-resolution magic-angle spinning NMR spectra and applications in structural studies of biomolecular systems

We describe investigations of the effects of rotational resonance (R(2)) on solid state (13)C NMR spectra of uniformly (13)C-labeled samples obtained under magic-angle spinning (MAS), and of the utility of R(2) measurements as structural probes of peptides and proteins with multiple uniformly labeled residues. We report results for uniformly (13)C-labeled L-alanine and L-valine in polycrystalline form, and for amyloid fibrils formed by the 15-residue peptide A beta(11-25) with uniform labeling of a four-residue segment. The MAS NMR spectra reveal a novel J-decoupling effect at R(2) conditions that may be useful in spectral assignments for systems with sharp (13)C MAS NMR lines. Pronounced dependences of the apparent isotropic (13)C NMR chemical shifts on MAS frequency near R(2) conditions are also observed. We demonstrate the feasibility of quantitative (13)C-(13)C distance determinations in L-valine, and qualitative determinations of inter-residue (13)C-(13)C contacts in A beta(11-25) fibrils. Finally, we demonstrate a "relayed" R(2) technique that may be useful in structural measurements on systems with poorly resolved (13)C MAS NMR lines.     

A comparison of NCO and NCA transfer methods for biological solid-state NMR spectroscopy

Three different techniques (adiabatic passage Hartman-Hahn cross-polarization, optimal control designed pulses, and EXPORT) are compared for transferring (15)N magnetization to (13)C in solid-state NMR experiments under magic-angle-spinning conditions. We demonstrate that, in comparison to adiabatic passage Hartman-Hahn cross-polarization, optimal control transfer pulses achieve similar or better transfer efficiencies for uniformly-(13)C,(15)N labeled samples and are generally superior for samples with non-uniform labeling schemes (such as 1,3- and 2-(13)C glycerol labeling). In addition, the optimal control pulses typically use substantially lower average RF field strengths and are more robust with respect to experimental variation and RF inhomogeneity. Consequently, they are better suited for demanding samples.     

H,N,C-triple resonance NMR of very large systems at 900 MHz

We provide quantitative signal to noise data and feasibility study at 900 MHz for 1H-15N-13C triple resonance backbone assignment pulse sequences obtained from a medium sized 2H, 13C, 15N labeled protein slowed down in glycerol-water solution to mimic relaxation and spectroscopic properties of a much larger protein system with macromolecular tumbling correlation time of 52 and 80 ns, respectively, at 296 and 283 K (corresponding to molecular weights of 130 and 250 kDa). Comparisons of several different schemes for transferring magnetization from proton to nitrogen and back to proton confirms Yang and Kay's 1999 prediction that avoiding the unfavorable relaxation properties of 1H-15N multiple quantum coherence in the TROSY phase cycle of the final 15N-1H transfer before acquisition is crucial for maximal sensitivity from these very large molecular weight systems. We also show results which confirm some predictions regarding the superiority of TROSY at 900 MHz vs. 800 MHz especially as the molecular weights become very large.     

1H,15N,13C-triple resonance NMR of very large systems at 900 MHz

We provide quantitative signal to noise data and feasibility study at 900 MHz for 1H-15N-13C triple resonance backbone assignment pulse sequences obtained from a medium sized 2H, 13C, 15N labeled protein slowed down in glycerol-water solution to mimic relaxation and spectroscopic properties of a much larger protein system with macromolecular tumbling correlation time of 52 and 80 ns, respectively, at 296 and 283 K (corresponding to molecular weights of 130 and 250 kDa). Comparisons of several different schemes for transferring magnetization from proton to nitrogen and back to proton confirms Yang and Kay's 1999 prediction that avoiding the unfavorable relaxation properties of 1H-15N multiple quantum coherence in the TROSY phase cycle of the final 15N-1H transfer before acquisition is crucial for maximal sensitivity from these very large molecular weight systems. We also show results which confirm some predictions regarding the superiority of TROSY at 900 MHz vs. 800 MHz especially as the molecular weights become very large.     

AssignFit: a program for simultaneous assignment and structure refinement from solid-state NMR spectra

AssignFit is a computer program developed within the XPLOR-NIH package for the assignment of dipolar coupling (DC) and chemical shift anisotropy (CSA) restraints derived from the solid-state NMR spectra of protein samples with uniaxial order. The method is based on minimizing the difference between experimentally observed solid-state NMR spectra and the frequencies back calculated from a structural model. Starting with a structural model and a set of DC and CSA restraints grouped only by amino acid type, as would be obtained by selective isotopic labeling, AssignFit generates all of the possible assignment permutations and calculates the corresponding atomic coordinates oriented in the alignment frame, together with the associated set of NMR frequencies, which are then compared with the experimental data for best fit. Incorporation of AssignFit in a simulated annealing refinement cycle provides an approach for simultaneous assignment and structure refinement (SASR) of proteins from solid-state NMR orientation restraints. The methods are demonstrated with data from two integral membrane proteins, one α-helical and one β-barrel, embedded in phospholipid bilayer membranes.     

A suite of 3D NMR methods for characterizing complex hydrocarbon structure fragments

A suite of triple resonance 3D NMR experiments is presented for the complete connectivity assignment of the hydrocarbon network in complex macromolecular and supramolecular organic structures. These new 3D NMR methods rely only on the presence of a unique set of (13)C resonances (from (13)C(X)) which are separated from the rest of the (13)C NMR spectrum. These experiments take the advantage of region selective excitation and selective inversion by composite pulses to provide correlations among H(A), (13)C(A); H(B), (13)C(B) and neighboring (13)C(X) resonances along three frequency dimensions. These methods include: gHC(A)C(X), gHC(A)C(X)-HH-TOCSY and gHC(A)C(X)-CC-TOCSY experiments. The utility of this approach is illustrated with spectra of selected structure fragments in poly(ethylene-co-n-butyl acrylate-co-carbon monoxide) (polyEBC) prepared from 1,2,3-(13)C(3)-n-butyl acrylate.     

NMR assignment of protein side chains using residue-correlated labeling and NOE spectra

A new approach for the isotopic labeling of proteins is proposed that aims to facilitate side chain resonance assignments. Residue-correlated (RC) labeling is achieved by the expression of a protein on a medium containing a mixture of labeled, e.g., [U-13C,15N]amino acids, and NMR silent, [U-2H]amino acids. De novo synthesis of amino acids was suppressed by feedback inhibition by the amino acids in the growth medium and by the addition of beta-chloro-L-alanine, a transaminase inhibitor. Incorporation of these amino acids into synthesized proteins results in a relative diminution of inter-residue NOE interactions and a relative enhancement of intra-residue NOEs. Comparison of the resulting NOE spectra with those obtained from a uniformly labeled sample allows identification of intra-residue NOE peaks. Thus, this approach provides direct information for sidechain assignments in the NOE spectra, which are subsequently used for structural analysis. We have demonstrated the feasibility of this strategy for the 143 amino acid nuclease inhibitor NuiA, both at 35 degrees C, corresponding to a rotational correlation time of 9.5 ns, and at 5 degrees C, corresponding to a rotational correlation time of 22 ns.