Triple-resonance NMR Techniques Research Articles

NMR “Crystallography” of Membrane Proteins Aligned in Native-Like Bilayers

Oriented-sample (OS) solid-state NMR can be used to determine structures of membrane proteins without crystallization and use of cryogenic temperatures. Here, the orientationally dependent dipolar couplings and chemical shifts provide direct input for structure calculations. However, so far only the 1H-15N dipolar couplings and 15N chemical shifts have been routinely assessed in 15N labeled protein samples. While highly efficient for determining the tilt angles of short α-helical peptides, these two measurements alone are likely insufficient for determining protein structures of arbitrary topology. Therefore, determination of tertiary structures of membrane proteins by OS NMR requires measurements of additional angular restraints. Here we have developed a new experimental triple-resonance NMR technique, which was applied to uniformly doubly (15N, 13C) labeled Pf1 coat protein reconstituted in magnetically aligned DMPC/DHPC bicelles. The previously inaccessible 1Hα-13Cα dipolar couplings, which represent powerful chiral angular restraints, have been measured for the resolved NMR resonances of the transmembrane domain of Pf1. Inclusion of the third restraint makes it possible to determine the torsion angles Φ and Ψ between the adjacent peptide planes without assuming α-helical structure a priori. Using our recently developed structure calculation algorithm [1], simultaneous fitting of three restraints per peptide plane has resulted in families of structures which have subsequently been filtered using various Rosetta scoring functions including the energetics within the membrane environment, hydrogen bonding, and potential steric clashes. Such structural validation protocol has led to a consensus α-helical transmembrane structure for Pf1 coat protein, albeit obtained in a de novo fashion, thus greatly enhancing utility of OS NMR as a “crystallography” in near-native membrane environments. [1]. Lapin, J. and Nevzorov A.A., Validation of protein backbone structures calculated from NMR angular restraints using Rosetta, J. Biomol. NMR, 2019; 73(5):229-244.

NMR “Crystallography” for Uniformly (13C, 15N)‐Labeled Oriented Membrane Proteins

In oriented-sample (OS) solid-state NMR of membrane proteins, the angular-dependent dipolar couplings and chemical shifts provide a direct input for structure calculations. However, so far only 1 H-15 N dipolar couplings and 15 N chemical shifts have been routinely assessed in oriented 15 N-labeled samples. The main obstacle for extending this technique to membrane proteins of arbitrary topology has remained in the lack of additional experimental restraints. We have developed a new experimental triple-resonance NMR technique, which was applied to uniformly doubly (15 N, 13 C)-labeled Pf1 coat protein in magnetically aligned DMPC/DHPC bicelles. The previously inaccessible 1 Hα -13 Cα dipolar couplings have been measured, which make it possible to determine the torsion angles between the peptide planes without assuming α-helical structure a priori. The fitting of three angular restraints per peptide plane and filtering by Rosetta scoring functions has yielded a consensus α-helical transmembrane structure for Pf1 protein.

In situ analysis of [8-13C-7-15N]-double-labelled theophylline by a triple resonance NMR technique

[8-13C-7-15N]-labelled theophylline was synthesized and applied for in situ analysis using a 1H–{13C–15N} triple resonance NMR technique. The triple resonance analysis enabled selective detection of labelled theophylline in a mixture containing excess amino acids or crude mouse liver extract, where conventional 1H NMR analysis suffers from difficulty because of background signals.

Turn-on Detection of Targeted Biochemical Reactions by Triple Resonance NMR Analysis Using Isotope-labeled Probe

Abstract We report on a strategy for the “turn-on” detection of target biochemical (metabolic) reactions using a triple resonance NMR technique with an isotope-labeled probe. Our NMR study clearly reveals that otherwise NMR-nonactive-13C/2H-labeled glucose actually turns “on” its 1H NMR signal by conversion to an important biomarker lactate as the end product of anaerobic glycolysis in cells and in injected mice with high selectivity.

Segmental Isotopic Labelling of a Multidomain Protein by Protein Ligation by Protein Trans‐Splicing

Segmental isotopic labelling is a powerful method for the incorporation of stable isotopes into particular regions within proteins for NMR detection, thereby reducing the complexity of NMR spectra and offering the potential to perform sequential assignments. Here we have demonstrated segmental isotopic labelling of a domain in a multidomain protein both in vivo and in vitro through protein ligation by protein trans-splicing. This robust protein trans-splicing approach could open possibilities for studying particular domains in intact proteins without dissection into smaller globular domains. Recent advances in optimization of transverse-relaxation in NMR spectroscopy have opened avenues for study of larger molecules (close to 1 MDa). However, sequential resonance assignments in large proteins remain time-consuming and challenging because of the increased number of signals and signal overlapping. Segmental isotopic labelling is one promising approach among numerous isotope-labelling techniques, because, unlike in the case of selective amino acid labelling, segmentally isotope-labelled samples can be directly analysed by tripleresonance NMR techniques developed for sequential resonance assignments. Segmentally isotope-labelled proteins have been prepared either by expressed protein ligation (EPL), which makes use of native chemical ligation (NCL), 6] or by protein trans-splicing (PTS), through the use of artificially split protein splicing domains (inteins). 7] EPL requires the preparation of an a-thioester group from a thiol reagent and an N-terminal cysteine residue by proteolysis in vitro (Scheme 1A), which demands considerable preparation efforts, 8] although an easier approach has recently been proposed. In protein splicing, an intein catalyses protein ligation of two polypeptide fragments fused to the Nand C-terminal ends of an intein. Protein splicing could take place in trans, when an intein is split into two fragments (Scheme 1B). 10] Segmental isotopic labelling through protein trans-splicing with artificially split inteins requires no additional thiol reagent nor cofactor, but denaturation and renaturation steps are necessary before protein-splicing activity can be restored. Unlike artificially split inteins, naturally split inteins do not require any denaturation and renaturation steps for protein splicing. Therefore, these have been suggested as potentially useful for segmental isotopic labelling of multidomain proteins. Protein trans-splicing with naturally split inteins has advantages over EPL because protein ligation can be performed not only in vitro but also in vivo, making it possible to achieve segmental isotopic labelling in vivo. Despite its many potential applications, it has never been used for segmental isotopic labelling of multidomain proteins except for a fusion tag for enhancing protein solubility. This is because the protein-splicing activity of the split inteins could be negatively affected even when naturally split inteins were fused with the Scheme 1. Protein ligation by A) expressed protein ligation and B) protein trans-splicing. C) Outline of the in vivo procedure for segmental isotopic labelling used in this article.

NMR assignment of the arenaviral protein Z from Lassa fever virus

The arenavirus protein Z from Lassa fever virus was recently found to inhibit mRNA translation through direct interaction with eIF4E. Here, we report the NMR assignment of this RING-containing protein that was determined by triple resonance NMR techniques.

Residue-Specific pKa Determination of Lysine and Arginine Side Chains by Indirect 15N and 13C NMR Spectroscopy: Application to apo Calmodulin

Electrostatic interactions in proteins can be probed experimentally through determination of residue-specific acidity constants. We describe here triple-resonance NMR techniques for direct determination of lysine and arginine side-chain protonation states in proteins. The experiments are based on detection of nonexchangeable protons over the full range of pH and temperature and therefore are well suited for pKa determination of individual amino acid side chains. The experiments follow the side-chain 15Nzeta (lysine) and 15Nepsilon or 13Czeta (arginine) chemical shift, which changes due to sizable changes in the heteronuclear electron distribution upon (de)protonation. Since heteronuclear chemical shifts are overwhelmed by the charge state of the amino acid side chain itself, these methods supersede 1H-based NMR in terms of accuracy, sensitivity, and selectivity. Moreover, the 15Nzeta and 15Nepsilon nuclei may be used to probe changes in the local electrostatic environment. Applications to three proteins are described: apo calmodulin, calbindin D9k, and FKBP12. For apo calmodulin, residue-specific pKa values of lysine side chains were determined to fall between 10.7 and 11.2 as a result of the high net negative charge on the protein surface. Ideal two-state titration behavior observed for all lysines indicates the absence of significant direct charge interactions between the basic residues. These results are compared with earlier studies based on chemical modification.

Sequence-specific 1H, 13C and 15N resonance assignments and secondary structure of a truncated protease from Simian Foamy Virus

The backbone and side chain assignments of the retroviral aspartate protease from Simian Foamy Virus from macaques (SFVmac) have been determined by triple resonance NMR techniques.

1H, 15N and 13C backbone and side chain chemical shifts of human ASC (apoptosis-associated speck-like protein containing a CARD domain)

The backbone and side chain resonance assignments of human ACS (approximately 22 KD), apoptosis-associated speck-like protein containing a caspase recruitment domain and a pyrin domain, have been determined by triple-resonance NMR techniques.

1H, 13C, and 15N Resonance Assignments for Escherichia coli ytfP, a Member of the Broadly Conserved UPF0131 Protein Domain Family

Protein ytfP from Escherichia coli (Swiss-Prot ID: YTFP-ECOLI; NESG target ID: ER111; Wunderlich et al., 2004) is a 113-residue member of the UPF0131 protein family (Pfam ID: PF03674) of unknown function. This domain family is found in organisms from all three kingdoms, archaea, eubacteria and eukaryotes. Using triple resonance NMR techniques, we have determined 97% of backbone and 91% of side chain 1H, 13C, and 15N resonance assignments. The chemical shift and 3J(HN?Ha) scalar coupling data reveal a mixed a/b topology,????????. BMRB deposit with Accession No. 6448. Reference: Wunderlich et al. (2004) Proteins, 56, 181?187.

1H, 13C, and 15N Resonance Assignments for the Protein Coded by Gene Locus BB0938 of Bordetella bronchiseptica

The product of gene locus BB0938 from Bordetella bronchiseptica (Swiss-Prot ID: Q7WNU7-BORBR; NESG target ID: BoR11; Wunderlich et al., 2004; Pfam ID: PF03476) is a 128-residue protein of unknown function. This broadly conserved protein family is found in eubacteria and eukaryotes. Using triple resonance NMR techniques, we have determined 98% of backbone and 94% of side chain 1H, 13C, and 15N resonance assignments. The chemical shift and 3J(HN?Ha) scalar coupling data reveal a b topology with a seven-residue helical insert, ??????????. BMRB deposit with accession number 6693. Reference: Wunderlich et al. (2004) Proteins, 56, 181?187.

Structural and dynamic characterization of a neuron-specific protein kinase C substrate, neurogranin.

Neurogranin/RC3 is a neuron-specific, Ca(2+)-sensitive calmodulin binding protein and a specific protein kinase C substrate. Neurogranin may function to regulate calmodulin levels at specific sites in neurons through phosphorylation at serine residue within its IQ motif, oxidation outside the IQ motif, or changes in local cellular Ca(2+) concentration. To gain insight into the functional role of neurogranin in the regulation of calmodulin-dependent activities, we investigated the structure and dynamics of a full-length rat neurogranin protein with 78 amino acids using triple resonance NMR techniques. In the absence of calmodulin or PKC, neurogranin exists in an unfolded form as evidenced by high backbone mobility and the absence of long-range nuclear Overhauser effect (NOE). Analyses of the chemical shifts (13)C(alpha), (13)C(beta), and (1)H(alpha) reveal the presence of a local alpha-helical structure for the region between residues G25-A42. Three-bond (1)H(N)-(1)H(alpha) coupling constants support the finding that the sequence between residues G25 and A42 populates a non-native helical structure in the unfolded neurogranin. Homonuclear NOE results are consistent with the conclusions drawn from chemical shifts and coupling constants. (15)N relaxation data indicate motional restrictions on a nanosecond time scale in the region from D15 to S48. Spectral densities and order parameters data further confirm that the unfolded neurogranin exists in conformation with residual secondary structures. The medium mobility of the nascent helical region may help to reduce the entropy loss when neurogranin binds to its targets, but the complex between neurogranin and calmodulin is not stable enough for structural determination by NMR. Calmodulin titration of neurogranin indicates that residues D15-G52 of neurogranin undergo significant structural changes upon binding to calmodulin.

An Empirical Correlation between Amide Deuterium Isotope Effects on 13Cα Chemical Shifts and Protein Backbone Conformation

Three-dimensional, triple resonance NMR techniques are described for measurement of two-bond (intraresidual) and three-bond (sequential) amide deuterium isotope effects on 13Cα chemical shifts. Measurements were carried out for uniformly 15N and 13C labeled human ubiquitin equilibrated in a 50% H2O/50% D2O mixed solvent. The three-bond isotope shift, 3ΔCα(ND), ranges from about 10−50 ppb, and, except for residues with positive φ angles and residues followed by Gly, its magnitude is described by 3ΔCα(ND) = 30.1 + 22.2 sin(ψ) ± 3.4 ppb. The two-bond isotope shift, 2ΔCα(ND), ranges from 70 to 116 ppb and is also dominated by the local backbone geometry at the Cα position: 2ΔCα(ND) = 93.1 + 10.1 sin(φ+62°) + 12.0 sin(ψ+42°) ± 4.1 ppb.

Application of H(N)CA,CO-E.COSY experiments for calibrating the φ angular dependences of vicinal couplings J(C′i−1,Hiα), J(C′i−1,Ciβ) and J(C′i−1,C′i) in proteins

A triple-resonance NMR technique suitable for the determination ofcarbonyl-related couplings in polypeptide systems is introduced. Theapplication of three novel pulse sequences to uniformly13C/15N-enriched proteins yields E.COSY-likemultiplet patterns exhibiting either one of the3J(C′i−1,Hiα), 3J(C′i−1,Ciβ) and3J(C′i−1,C′i)coupling constants in the indirectly detected 13C′dimension, depending on the passive spin selected. The experiments aredemonstrated with oxidized flavodoxin from Desulfovibrio vulgaris. On thebasis of the J-values measured and the backbone φ-angles derived from ahigh-resolution X-ray structure of the protein, the three associated Karplusequations were reparametrized. The root-mean-square differences between theexperimental coupling constants and those predicted by the optimized Karpluscurves are 0.41, 0.33 and 0.32 Hz for3J(C′i−1,Hiα),3J(C′i−1,Ciβ) and3J(C′i−1,C′i),respectively. The results are compared with the Karplus parameters previouslypublished for the same couplings.

Sequential assignment of 1H, 13C and 15N resonances of human carbonic anhydrase I by triple-resonance NMR techniques and extensive amino acid-specific 15N-labeling.

The backbone NMR resonances of human carbonic anhydrase I (HCA I) have been assigned. This protein is one of the largest monomeric proteins assigned so far. The assignment was enabled by a combination of 3D triple-resonance experiments and extensive use of amino acid-specific 15N-labeling. The obtained resonance assignment has been used to evaluate the secondary structure elements present in solution. The solution structure appears to be very similar to the crystal structure, although some differences can be observed. Proton-deuteron exchange experiments have shown that the assignments provide probes that can be used in future studies of HCA I.

HCN, A Triple-Resonance NMR Technique for Selective Observation of Histidine and Tryptophan Side Chains in13C/15N-Labeled Proteins

HCN, a new 3D NMR technique for stepwise coherence transfer from1H to13C to15N and reverse through direct spin couplings1JCHand1JCN, is presented as a method for detection and assignment of histidine and tryptophan side-chain1H,13C, and15N resonances in uniformly13C/15N-labeled proteins. Product-operator calculations of cross-peak volumes vs adjustable delay τ3were employed for determination of optimal τ3. For the phosphatidylinositol 3-kinase (PI3K SH3 domain, MW = 9.6 kD) at pH 6, H(C)N, the1H/15N projection, produced observable cross peaks within 20 min. and was completely selective for the single tryptophan and single histidine. The 3D HCN experiment yielded well-defined cross peaks in 20 h for the13C/15N-labeled origin-specific DNA binding domain from simian virus 40 T-antigen (T-ag-OBD131–259, MW = 15.4 kD) at pH 5.5. Resonances from all six histidines in T-ag-OBD were observed, and 11 of the 121H and13C chemical shifts and 10 of the 1215N chemical shifts were determined. The13C dimension proved essential in assignment of the multiply overlapping1H and15N resonances. From the spectra recorded at a single pH, three of the imidazoles were essentially neutral and the other three were partially protonated (22–37%). HCN yielded strong cross peaks after 18 h on a 2.0 mMsample of phenylmethanesulfonyl fluoride (PMSF)-inhibited α-lytic protease (MW = 19.8 kD) at pH 4.4. No spectra have been obtained, however, of native or boronic acid-inhibited α-lytic protease after 18 h at various temperatures ranging from 5 to 55°C, probably due to efficient relaxation of active-site imidazole1H and/or15N nuclei.

Complete 1H, 13C and 15N NMR assignments and secondary structure of the 269-residue serine protease PB92 from Bacillus alcalophilus

The 1H, 13C and 15N NMR resonances of serine protease PB92 have been assigned using 3D triple-resonance NMR techniques. With a molecular weight of 27 kDa (269 residues) this protein is one of the largest monomeric proteins assigned so far. The side-chain assignments were based mainly on 3D H(C)CH and 3D (H)CCH COSY and TOCSY experiments. The set of assignments encompasses all backbone carbonyl and CHn carbons, all amide (NH and NH2) nitrogens and 99.2% of the amide and CHn protons. The secondary structure and general topology appear to be identical to those found in the crystal structure of serine protease PB92 [Van der Laan et al. (1992) Protein Eng., 5, 405-411], as judged by chemical shift deviations from random coil values, NH exchange data and analysis of NOEs between backbone NH groups.

3D Triple-resonance NMR techniques for the sequential assignment of NH and 15N resonances in 15N- and 13C-labelled proteins

Two new 3D 1H-15N-13C triple-resonance experiments are presented which provide sequential cross peaks between the amide proton of one residue and the amide nitrogen of the preceding and succeeding residues or the amide proton of one residue and the amide proton of the preceding and succeeding residues, respectively. These experiments, which we term 3D-HN(CA)NNH and 3D-H(NCA)NNH, utilize an optimized magnetization transfer via the 2JNC alpha coupling to establish the sequential assignment of backbone NH and 15N resonances. In contrast to NH-NH connectivities observable in homonuclear NOESY spectra, the assignments from the 3D-H(NCA)NNH experiment are conformation independent to a first-order approximation. Thus the assignments obtained from these experiments can be used as either confirmation of assignments obtained from a conventional homonuclear approach or as an initial step in the analysis of backbone resonances according to Ikura et al. (1990) [Biochemistry, 29, 4659-4667]. Both techniques were applied to uniformly 15N- and 13C-labelled ribonuclease T1.

1H, 13C, and 15N NMR backbone assignments and secondary structure of human interferon-gamma.

1H, 13C, and 15N NMR assignments of the protein backbone of human interferon-gamma, a homodimer of 31.4 kDa, have been made using the recently introduced three-dimensional (3D) triple-resonance NMR techniques. It is shown that, despite the approximately 40-50-Hz 13C alpha and 1H alpha line widths of this high molecular weight dimer and the extensive overlap in the 1H alpha and 13C alpha spectral regions, unique sequential assignments can be made on the basis of combined use of the 3D HNCO, HNCA, HN(CO)CA, and HCACO constant-time experiments, the 15N-separated 3D NOESY-HMQC, and the 3D HOHAHA-HMQC experiments. Analysis of the 15N-separated 3D NOESY-HMQC and 13C/15N-separated four-dimensional (4D) NOESY-HMQC spectra together with the secondary C alpha and C beta chemical shifts yielded extensive secondary structure information. The NMR-derived secondary structure essentially confirms results of a recently published low-resolution crystal structure [Ealick et al. (1991) Science 252, 698-702], i.e., six helices in the monomer which are mostly alpha-helical in nature, no beta-sheets, a long flexible loop between helices A and B, and a very hydrophobic helix C. The functionally important carboxy terminus, which was not observed in the X-ray study, does not adopt a rigid conformation in solution. A high degree of internal mobility, starting at Pro-123, gives rise to significantly narrower resonance line widths for these carboxy-terminal residues compared to the rest of the protein.

Improved 3D triple-resonance NMR techniques applied to a 31 kDa protein

Recently proposed 3D triple-resonance techniques (I-5) make it possible to obtain sequential assignment of the backbone ‘H, 15N and 13C resonances in proteins that can be isotopically labeled with 13C and “N. Thi’s approach has been used successfully for a number of systems, including calmodulin ( 16.7 kDa) ( 1) , calmodulin complexed with a 26-residue peptide ( -20 kDa) (6), and the phospho-carrier protein IIIG’” ( 7) ( 18 kDa). All of these triple-resonance experiments rely on transfer of magnetization via heteronuclear one-bond Jcouplings and the sensitivity of the experiments depends strongly on the ratio of the size of the J coupling and the linewidths of the nuclei involved in the magnetization transfer process. Our original experiments were designed to minimize the number of RF pulses required for a particular pulse sequence, and thus to minimize the effect of pulse imperfections. Subsequent experience with these experiments has indicated that the effect of pulse imperfections is not as severe as originally expected, provided that pulses are properly calibrated. Hence, in order to optimize the experiments it is more important to minimize the relaxation of transverse magnetization by keeping the dephasing and rephasing delays as short as possible, especially for larger proteins. Here we demonstrate the applicability of three improved triple-resonance experiments to the study of interferon-y, a dimer with 134 residues per monomer and a total molecular weight of 3 1.4 kDa. Correlation of the amide proton and “N resonance with the intraresidue 13Ca can be obtained with an experiment known as HNCA (I, 2). Note that this experiment also provides connectivity to the Ca of the preceding residue, transferring coherence via the ’ JN-Ca coupling. Because ’ JN-Ca and ’ JNxn can be of comparable magnitude (8) it may not always be possible to distinguish intrafrom interresidue connectivities with this experiment. In contrast, the HN( CO)CA experiment (4) provides correlations exclusively to the Ca resonance of the preceding residue by relaying magnetization via the intervening carbonyl 13C spin. The HN( CO)CA experiment is closely related to the HNCO experiment, which correlates the amide ‘H and “N with the 13C of the preceding residue. In the present study of interferon-y, we employ modified versions of these experiments, using INEPT (9) instead of HMQC (10-12) type coherence transfers, and overlay the 13C15N dephasing period with a 15N evolution period