IBIS--a tool for automated sequential assignment of protein spectra from triple resonance experiments

We have developed a tool for computer-assisted assignments of protein NMR spectra from triple resonance data. The program is designed to resemble established manual assignment procedures as closely as possible. IBIS exports its results in XEASY format. Thus, using IBIS the operator has continuous visual and accounting control over the progress of the assignment procedure. IBIS achieves complete assignments for those residues that exhibit sequential triple resonance connectivities within a few hours or days.     

Sequential resonance assignment of medium-sized15 N/13C-labeled proteins with projected 4D triple resonance NMR experiments

We recently introduced a new line of reduced-dimensionality experiments making constructive use of axial peak magnetization, which has so far been suppressed as an undesirable artifact in multidimensional NMR spectra [Szyperski, T., Braun, D., Banecki, B. and Wüthrich, K. (1996) J. Am. Chem. Soc., 118, 8146–8147]. The peaks arising from the axial magnetization are located at the center of the doublets resulting from projection. Here we describe the use of such projected four-dimensional (4D) triple resonance experiments for the efficient sequential resonance assignment of ¹⁵N/¹³C-labeled proteins. A 3D ^{/ /}(CO)NHN experiment is recorded either in conjunction with 3D HNN< > or with the newly presented 3D HNN scheme. The first combination yields sequential assignments based on the measurement of¹³ C chemical shifts and provides a complete ¹H, ¹³C and ¹⁵N resonance assignment of polypeptide backbone and CH_n moieties. When employing the second combination, ¹³C=O chemical shifts are not measured, but the sequential assignment relies on both ¹³C and¹ H chemical shifts. The assignment is performed in a semi-automatic fashion using the program XEASY in conjunction with the newly implemented program SPSCAN. This program package offers routines for the facile mutual interconversion of single-quantum and zero/double-quantum frequencies detected in conventional and reduced-dimensionality spectra, respectively. In particular, SPSCAN comprises a peak picking routine tailored to cope with the distinct peak patterns of projected NMR experiments performed with simultaneous acquisition of central peaks. Data were acquired at 13 °C for the N-terminal 63-residue polypeptide fragment of the 434 repressor. Analysis of these spectra, which are representative for proteins of about 15 kDa when working at commonly used temperatures around 30 °C , demonstrates the efficiency of our approach for the assignment of medium-sized¹⁵ N/¹³C doubly labeled proteins.     

A 4D HCCH-TOCSY experiment for assigning the side chain1H and13C resonances of proteins

Summary A 4D HCCH-TOCSY experiment is described for correlating and assigning the¹H and¹³C resonances of protein amino acid side chains that has several advantages over 3D versions of the experiment. In many cases, both the¹H and¹³C chemical shifts can be obtained in the 4D experiment from a simple inspection of the¹³C(3),¹H(4) planes extracted at the¹H(1)/¹³C(2) chemical shifts. Together with the 3D and 4D triple resonance experiments, this allows sequence-specific assignments to be obtained. In addition, the increased resolution of the 4D experiment compared to its 3D counterpart allows. automation of resonance assignments.     

CO_H(N)CACB experiments for assigningbackbone resonances in 13C/15N-labeledproteins

A triple resonance NMR experiment, denoted CO_H(N)CACB, correlates¹H^N and ¹³CO spins with the¹³C and¹³C spins of adjacent amino acids. Thepulse sequence is an out-and-back design that starts with¹H^N magnetization and transfers coherence viathe ¹⁵N spin simultaneously to the ¹³CO and¹³C spins, followed by transfer to the¹³C spin. Two versions of the sequence arepresented: one in which the ¹³CO spins are frequency labeledduring an incremented t₁ evolution period prior to transfer ofmagnetization from the ¹³C to the¹³C resonances, and one in which the¹³CO spins are frequency labeled in a constant-time mannerduring the coherence transfer to and from the¹³C resonances. Because ¹³COand ¹⁵N chemical shifts are largely uncorrelated, thetechnique will be especially useful when degeneracy in the¹H^N-¹⁵N chemical shifts hindersresonance assignment. The CO_H(N)CACB experiment is demonstrated usinguniformly ¹³C/¹⁵N-labeled ubiquitin.     

Improved sensitivity and coherence selection for [15N,1H]-TROSY elements in triple resonance experiments

In experiments with proteins of molecular weights around 100 kDa the implementation of [¹⁵N,¹H]-TROSY-elements in [¹⁵N]-constant-time triple resonance experiments yields sensitivity enhancements of one to two orders of magnitude. An additional gain of 10 to 20% may be obtained with the use of sensitivity enhancement elements. This paper describes a novel sensitivity enhancement scheme which is based on concatenation of the ¹³ C ¹⁵N magnetization transfer with the ST2-PT element, and which enables proper TROSY selection of the ¹⁵N multiplet components.     

Improved resolution in three-dimensional constant-time triple resonance NMR spectroscopy of proteins

Summary Two new protocols for the three-dimensional, triple resonance, constant-time HCA(CO)N NMR experiment are presented that significantly increase the experimental resolution attainable in the C frequency dimension. Experimental verification of the new experiments is provided by spectra of the IIA domain of glucose permease fromBacillus subtilis.     

Structure of the Oct-3 POU-homeodomain in solution,as determined by triple resonance heteronuclear multidimensional NMR spectroscopy.

The POU-homeodomain (POUH) forms the bipartite DNA-binding POU domain in association with the POU-specific domain. The 1H, 15N, and 13C magnetic resonances of the 67-amino acid long POUH of mouse Oct-3 have almost completely been assigned, mainly through the combined use of three-dimensional triple resonance NMR methods. Based on the distance and dihedral angle constraints derived from the NMR data, the solution structure of the POUH domain has been calculated by the ab initio simulated annealing method. The average RMS deviation for all backbone heavy atoms of the 20 best calculated structures for residues 9-53 of the total 67 amino acid residues is 0.44 A. The POUH domain consists of three alpha-helices (helix-I, 10-20; helix-II, 28-38; and helix-III, 42-53), and helices-II and -III form a helix-turn-helix motif. In comparison with other classical homeodomains, the folding of the three helices is quite similar. However, the length of helix-III is fairly short. In the complex of the Oct-1 POU domain with an octamer site (Klemm JD, et al., 1994, Cell 77:21-32), the corresponding region is involved in helix-III. The structural difference between these two cases will be discussed.     

Assignments and structure determination of the catalytic domain of human fibroblast collagenase using 3D double and triple resonance NMR spectroscopy

We report here the backbone 1HN, 15N, 13C, 13CO, and 1H NMR assignmentsfor the catalytic domain of human fibroblast collagenase (HFC). Three independentassignment pathways (matching 1H, 13C, and 13CO resonances) were used to establishsequential connections. The connections using 13C resonances were obtained fromHNCOCA and HNCA experiments; 13CO connections were obtained from HNCO andHNCACO experiments. The sequential proton assignment pathway was established from a 3D(1H/15N) NOESY-HSQC experiment. Amino acid typing was accomplished using 13C and15N chemical shifts, specific labeling of 15N-Leu, and spin pattern recognition from DQF-COSY. The secondary structure was determined by analyzing the 3D (1H/15N) NOESY-HSQC. A preliminary NMR structure calculation of HFC was found to be in agreement withrecent X-ray structures of human fibroblast collagenase and human neutrophil collagenase aswell as similar to recent NMR structures of a highly homologous protein, stromelysin. Allthree helices were located; a five-stranded -sheet (four parallel strands, one antiparallelstrand) was also determined. -Sheet regions were identified by cross-stranddN and dNN connections and by strong intraresidue dN correlations, and were corroborated byobserving slow amide proton exchange. Chemical shift changes in a selectively 15N-labeledsample suggest that substantial structural changes occur in the active site cleft on the bindingof an inhibitor.     

Improved 3D triple resonance experiments, HNN and HN(C)N, for HN and 15N sequential correlations in (13C, 15N) labeled proteins: application to unfolded proteins

Two triple resonance experiments, HNN and HN(C)N, are presented which correlate H^N and ¹⁵N resonances sequentially along the polypeptide chain of a doubly (¹³C, ¹⁵N) labeled protein. These incorporate several improvements over the previously published sequences for a similar purpose and have several novel features. The spectral characteristics enable direct identification of certain triplets of residues, which provide many starting points for the sequential assignment procedure. The experiments are sensitive and their utility has been demonstrated with a 22 kDa protein under unfolding conditions where most of the standard triple resonance experiments such as HNCA, CBCANH etc. have limited success because of poor amide, C and C chemical shift dispersions.     

A tracked approach for automated NMR assignments in proteins (TATAPRO)

A novel automated approach for the sequence specific NMR assignments of ¹H^N, ¹³C, ¹³C, ¹³C/¹H and ¹⁵N spins in proteins, using triple resonance experimental data, is presented. The algorithm, TATAPRO (Tracked AuTomated Assignments in Proteins) utilizes the protein primary sequence and peak lists from a set of triple resonance spectra which correlate ¹H^N and ¹⁵N chemical shifts with those of ¹³C, ¹³C and ¹³C/¹H. The information derived from such correlations is used to create a `master_list' consisting of all possible sets of ¹H^N _i, ¹⁵N_i, ¹³C _i, ¹³C _i, ¹³C_i/¹H _i, ¹³C _i–1, ¹³C _i–1 and ¹³C_i–1/ ¹H _i–1 chemical shifts. On the basis of an extensive statistical analysis of ¹³C and ¹³C chemical shift data of proteins derived from the BioMagResBank (BMRB), it is shown that the 20 amino acid residues can be grouped into eight distinct categories, each of which is assigned a unique two-digit code. Such a code is used to tag individual sets of chemical shifts in the master_list and also to translate the protein primary sequence into an array called pps_array. The program then uses the master_list to search for neighbouring partners of a given amino acid residue along the polypeptide chain and sequentially assigns a maximum possible stretch of residues on either side. While doing so, each assigned residue is tracked in an array called assig_array, with the two-digit code assigned earlier. The assig_array is then mapped onto the pps_array for sequence specific resonance assignment. The program has been tested using experimental data on a calcium binding protein from Entamoeba histolytica (Eh-CaBP, 15 kDa) having substantial internal sequence homology and using published data on four other proteins in the molecular weight range of 18–42 kDa. In all the cases, nearly complete sequence specific resonance assignments (> 95%) are obtained. Furthermore, the reliability of the program has been tested by deleting sets of chemical shifts randomly from the master_list created for the test proteins.     

Spatially encoded strategies in the execution of biomolecular-oriented 3D NMR experiments

Three-dimensional nuclear magnetic resonance (3D NMR) provides one of the foremost analytical tools available for the elucidation of biomolecular structure, function and dynamics. Executing a 3D NMR experiment generally involves scanning a series of time-domain signals S(t ₃), as a function of two time variables (t ₁, t ₂) which need to undergo parametric incrementations throughout independent experiments. Recent years have witnessed extensive efforts towards the acceleration of this kind of experiments. Among the different approaches that have been proposed counts an “ultrafast” scheme, which distinguishes itself from other propositions by enabling—at least in principle—the acquisition of the complete multidimensional NMR data set within a single transient. 2D protein NMR implementations of this single-scan method have been demonstrated, yet its potential for 3D acquisitions has only been exemplified on model organic compounds. This publication discusses a number of strategies that could make these spatial encoding protocols compatible with 3D biomolecular NMR applications. These include a merging of 2D ultrafast NMR principles with temporal 2D encoding schemes, which can yield 3D HNCO spectra from peptides and proteins within ≈100 s timescales. New processing issues that facilitate the collection of 3D NMR spectra by relying fully on spatial encoding principles are also assessed, and shown capable of delivering HNCO spectra within 1 s timescales. Limitations and prospects of these various schemes are briefly addressed.     

Automated probabilistic method for assigning backbone resonances of (13C,15N)-labeled proteins

We present a computer algorithm for the automated assignment of polypeptide backbone and13C resonances of a protein of known primary sequence. Input to the algorithm consistsof cross peaks from several 3D NMR experiments: HNCA, HN(CA)CO, HN(CA)HA,HNCACB, COCAH, HCA(CO)N, HNCO, HN(CO)CA, HN(COCA)HA, and CBCA(CO)NH.Data from these experiments performed on glutamine-binding protein are analyzed statisticallyusing Bayes' theorem to yield objective probability scoring functions for matching chemicalshifts. Such scoring is used in the first stage of the algorithm to combine cross peaks fromthe first five experiments to form intraresidue segments of chemical shifts{Ni,HiN,Ci,Ci,Ci}, while the latter five are combined into interresiduesegments {Ci,Ci,Ci,Ni+1,HNi+1}. Given a tentative assignment of segments,the second stage of the procedure calculates probability scores based on the likelihood ofmatching the chemical shifts of each segment with (i) overlapping segments; and (ii) chemicalshift distributions of the underlying amino acid type (and secondary structure, if known). Thisjoint probability is maximized by rearranging segments using a simulated annealing program,optimized for efficiency. The automated assignment program was tested using CBCANH andCBCA(CO)NH cross peaks of the two previously assigned proteins, calmodulin and CheA.The agreement between the results of our method and the published assignments wasexcellent. Our algorithm was also applied to the observed cross peaks of glutamine-bindingprotein of Escherichia coli, yielding an assignment in excellent agreement with that obtainedby time-consuming, manual methods. The chemical shift assignment procedure described hereshould be most useful for NMR studies of large proteins, which are now feasible with the useof pulsed-field gradients and random partial deuteration of samples.     

Differences between the electronic environments of reduced and oxidized Escherichia coli DsbA inferred from heteronuclear magnetic resonance spectroscopy.

DsbA is the strongest protein disulfide oxidant yet known and is involved in catalyzing protein folding in the bacterial periplasm. Its strong oxidizing power has been attributed to the lowered pKa of its reactive active site cysteine and to the difference in thermodynamic stability between the oxidized and the reduced form. However, no structural data are available for the reduced state. Therefore, an NMR study of DsbA in its two redox states was undertaken. We report here the backbone 1HN, 15N, 13C(alpha) 13CO, 1H(alpha), and 13Cbeta NMR assignments for both oxidized and reduced Escherichia coli DsbA (189 residues). Ninety-nine percent of the frequencies were assigned using a combination of triple (1H-13C-15N) and double resonance (1H-15N or 1H-13C) experiments. Secondary structures were established using the CSI (Chemical Shift Index) method, NOE connectivity patterns, 3(J)H(N)H(alpha) and amide proton exchange data. Comparison of chemical shifts for both forms reveals four regions of the protein, which undergo some changes in the electronic environment. These regions are around the active site (residues 26 to 43), around His60 and Pro 151, and also around Gln97. Both the number and the amplitude of observed chemical shift variations are more substantial in DsbA than in E. coli thioredoxin. Large 13C(alpha) chemical shift variations for residues of the active site and residues Phe28, Tyr34, Phe36, Ile42, Ser43, and Lys98 suggest that the backbone conformation of these residues is affected upon reduction.     

(H)N(COCA)NH and HN(COCA)NH experiments for 1H-15N backbone assignments in 13C/15N-labeled proteins

Triple resonance HN(COCA)NH pulse sequences for correlating 1H(i), 15N(i),1H(i-1), and 15N(i-1) spins that utilize overlapping coherence transfer periods provide increased sensitivityrelative to pulse sequences that utilize sequential coherence transfer periods. Although theoverlapping sequence elements reduce the overall duration of the pulse sequences, theprincipal benefit derives from a reduction in the number of 180° pulses. Two versions of thetechnique are presented: a 3D (H)N(COCA)NH experiment that correlates 15N(i),1H(i-1), and 15N(i-1) spins, and a 3D HN(COCA)NH experiment that correlates 1H(i), 15N(i),1H(i-1), and 15N(i-1) spins by simultaneously encoding the 1H(i) and 15N(i) chemical shiftsduring the t1 evolution period. The methods are demonstrated on a 13C/15N-enriched sampleof the protein ubiquitin and are easily adapted for application to 2H/13C/15N-enrichedproteins.     

Backbone resonance assignment in large protonated proteins using a combination of new 3D TROSY-HN(CA)HA, 4D TROSY-HACANH and 13C-detected HACACO experiments

A new method for backbone resonance assignment suitable for large proteins with the natural ¹H isotope content is proposed based on a combination of the most sensitive TROSY-type triple-resonance experiments. These techniques include TROSY-HNCO, ¹³C-detected 3D multiple-quantum HACACO and the newly developed 3D TROSY multiple-quantum-HN(CA)HA and 4D TROSY multiple-quantum-HACANH experiments. The favorable relaxation properties of the multiple-quantum coherences, signal detection using the ¹³C antiphase coherences, and the use of TROSY optimize the performance of the proposed set of experiments for application to large protonated proteins. The method is demonstrated with the 44 kDa uniformly ¹⁵N,¹³C-labeled and fractionally (35%) deuterated trimeric B. Subtilis Chorismate Mutase and is suitable for proteins with large correlation times but a relatively small number of residues, such as membrane proteins embedded in micelles or oligomeric proteins.     

Chemical shift based editing of CH<Subscript>3</Subscript> groups in fractionally <Superscript>13</Superscript>C-labelled proteins using GFT (3, 2)D CT-H<Emphasis Type="Underline">CC</Emphasis>H-COSY: stereospecific assignments of CH<Subscript>3</Subscript> groups of Val and Leu residues

We propose a (3, 2)D CT-HCCH-COSY experiment to rapidly collect the data and provide significant dispersion in the spectral region containing (13)C-(1)H cross peaks of CH(3) groups belonging to Ala, Ile, Leu, Met, Thr and Val residues. This enables one to carry out chemical shift based editing and grouping of all the (13)C-(1)H cross peaks of CH(3) groups belonging to Ala, Ile, Leu, Met, Thr and Val residues in fractionally (10%) (13)C-labelled proteins, which in turn aids in the sequence-specific resonance assignments in general and side-chain resonance assignments in particular, in any given protein. Further, we demonstrate the utility of this experiment for stereospecific assignments of the pro-R and pro-S methyl groups belonging to the Leu and Val residues in fractionally (10%) (13)C-labelled proteins. The proposed experiment opens up a wide range of applications in resonance assignment strategies and structure determination of proteins.     

Solution NMR resonance assignment strategies for β‐barrel membrane proteins

Membrane proteins in detergent micelles are large and dynamic complexes that present challenges for solution NMR investigations such as spectral overlap and line broadening. In this study, multiple methods are introduced to facilitate resonance assignment of β‐barrel membrane proteins using Opa₆₀ from Neisseria gonorrhoeae as a model system. Opa₆₀ is an eight‐stranded β‐barrel with long extracellular loops (～63% of the protein) that engage host receptors and induce engulfment of the bacterium. The NMR spectra of Opa₆₀ in detergent micelles exhibits significant spectral overlap and resonances corresponding to the loop regions had variable line widths, which interfered with a complete assignment of the protein. To assign the β‐barrel residues, trypsin cleavage was used to remove much of the extracellular loops while preserving the detergent solubilized β‐barrel. The removal of the loop resonances significantly improved the assignment of the Opa₆₀ β‐barrel region (97% of the resonances corresponding to the β‐barrel and periplasmic turns were assigned). For the loop resonance assignments, two strategies were implemented; modulating temperature and synthetic peptides. Lowering the temperature broadened many peaks beyond detection and simplified the spectra to only the most dynamic regions of the loops facilitating 27 loop resonances to be assigned. To further assign functionally important and unstructured regions of the extracellular loops, a synthetic 20 amino acid peptide was synthesized and had nearly complete spectral overlap with the full‐length protein allowing 17 loop resonances to be assigned. Collectively, these strategies are effective tools that may accelerate solution NMR structure determination of β‐barrel membrane proteins.     

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

Summary 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 ²J_NC 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.     

A general Bayesian method for an automated signal class recognition in 2D NMR spectra combined with a multivariate discriminant analysis

Summary A generally applicable method for the automated classification of 2D NMR peaks has been developed, based on a Bayesian approach coupled to a multivariate linear discriminant analysis of the data. The method can separate true NMR signals from noise signals, solvent stripes and artefact signals. The analysis relies on the assumption that the different signal classes have different distributions of specific properties such as line shapes, line widths and intensities. As to be expected, the correlation network of the distributions of the selected properties affects the choice of the discriminant function and the final selection of signal properties. The classification rule for the signal classes was deduced from Bayes's theorem. The method was successfully tested on a NOESY spectrum of HPr protein from Staphylococcus aureus. The calculated probabilities for the different signal class memberships are realistic and reliable, with a high efficiency of discrimination between peaks that are true NOE signals and those that are not.     

Automated sequencing of amino acid spin systems in proteins using multidimensional HCC(CO)NH-TOCSY spectroscopy and constraint propagation methods from artificial intelligence

Summary We have developed an automated approach for determining the sequential order of amino acid spin systems in small proteins. A key step in this procedure is the analysis of multidimensional HCC(CO)NH-TOCSY spectra that provide connections from the aliphatic resonances of residue i to the amide resonances of residue i+1. These data, combined with information about the amino acid spin systems, provide sufficient constraints to assign most proton and nitrogen resonances of small proteins. Constraint propagation methods progressively narrow the set of possible assignments of amino acid spin systems to sequence-specific positions in the process of NMR data analysis. The constraint satisfaction paradigm provides a framework in which the necessary constraint-based reasoning can be expressed, while an object-oriented representation structures and facilitates the extensive list processing and indexing involved in matching. A prototype expert system, AUTOASSIGN, provides correct and nearly complete resonance assignments with one real and 31 simulated 3D NMR data sets for a 72-amino acid domain, derived from the Protein A of Staphylococcus aureus, and with 31 simulated NMR data sets for the 50-amino acid human type- transforming growth factor.