An expectation/maximization nuclear vector replacement algorithm for automated NMR resonance assignments

We report an automated procedure for high-throughput NMR resonance assignment for a protein of known structure, or of an homologous structure. Our algorithm performs Nuclear Vector Replacement (NVR) by Expectation/Maximization (EM) to compute assignments. NVR correlates experimentally-measured NH residual dipolar couplings (RDCs) and chemical shifts to a given a priori whole-protein 3D structural model. The algorithm requires only uniform (15)N-labelling of the protein, and processes unassigned H(N)-(15)N HSQC spectra, H(N)-(15)N RDCs, and sparse H(N)-H(N) NOE's (d(NN)s). NVR runs in minutes and efficiently assigns the (H(N),(15)N) backbone resonances as well as the sparse d(NN)s from the 3D (15)N-NOESY spectrum, in O (n(3)) time. The algorithm is demonstrated on NMR data from a 76-residue protein, human ubiquitin, matched to four structures, including one mutant (homolog), determined either by X-ray crystallography or by different NMR experiments (without RDCs). NVR achieves an average assignment accuracy of over 99%. We further demonstrate the feasibility of our algorithm for different and larger proteins, using different combinations of real and simulated NMR data for hen lysozyme (129 residues) and streptococcal protein G (56 residues), matched to a variety of 3D structural models.     

A polynomial-time nuclear vector replacement algorithm for automated NMR resonance assignments.

High-throughput NMR structural biology can play an important role in structural genomics. We report an automated procedure for high-throughput NMR resonance assignment for a protein of known structure, or of a homologous structure. These assignments are a prerequisite for probing protein-protein interactions, protein-ligand binding, and dynamics by NMR. Assignments are also the starting point for structure determination and refinement. A new algorithm, called Nuclear Vector Replacement (NVR) is introduced to compute assignments that optimally correlate experimentally measured NH residual dipolar couplings (RDCs) to a given a priori whole-protein 3D structural model. The algorithm requires only uniform( 15)N-labeling of the protein and processes unassigned H(N)-(15)N HSQC spectra, H(N)-(15)N RDCs, and sparse H(N)-H(N) NOE's (d(NN)s), all of which can be acquired in a fraction of the time needed to record the traditional suite of experiments used to perform resonance assignments. NVR runs in minutes and efficiently assigns the (H(N),(15)N) backbone resonances as well as the d(NN)s of the 3D (15)N-NOESY spectrum, in O(n(3)) time. The algorithm is demonstrated on NMR data from a 76-residue protein, human ubiquitin, matched to four structures, including one mutant (homolog), determined either by x-ray crystallography or by different NMR experiments (without RDCs). NVR achieves an assignment accuracy of 92-100%. We further demonstrate the feasibility of our algorithm for different and larger proteins, using NMR data for hen lysozyme (129 residues, 97-100% accuracy) and streptococcal protein G (56 residues, 100% accuracy), matched to a variety of 3D structural models. Finally, we extend NVR to a second application, 3D structural homology detection, and demonstrate that NVR is able to identify structural homologies between proteins with remote amino acid sequences using a database of structural models.     

NMR restraint analysis of transforming growth factor alpha: a key component for NMR structure refinement.

Structures of the protein, transforming growth factor alpha (TGF-alpha), have been derived from NMR data using distance geometry and subsequent energy refinement. Analysis of the sequential NOE distance bounds using a template algorithm provides a check for consistency in the calculation of bounds, stereospecific assignment of prochiral centers, and secondary structure assignment. Application of the template algorithm to the long range NOEs found within the NMR data sets collected at pH 6.3 and pH 3.4 is used to assess the confidence levels for the accuracy of the structures obtained from modeling. The method also provides critical insight in differentiating regions of the structure that are well defined from those that are not. Use of the restraint analysis protocol is shown to be a powerful adjunct to currently used methods for the assignment of protein structures from NMR data.     

Backbone assignment of proteins with known structure using residual dipolar couplings

A prerequisite for NMR studies of protein-ligand interactions or protein dynamics is the assignment of backbone resonances. Here we demonstrate that protein assignment can significantly be enhanced when experimental dipolar couplings (RDCs) are matched to values back-calculated from a known three-dimensional structure. In case of small proteins, the program MARS allows assignment of more than 90% of backbone resonances without the need for sequential connectivity information. For bigger proteins, we show that the combination of sequential connectivity information with RDC-matching enables more residues to be assigned reliably and backbone assignment to be more robust against missing data. Structural or dynamic deviations from the employed 3D coordinates do not lead to an increased error rate in RDC-supported assignment. RDC-enhanced assignment is particularly useful when chemical shifts and sequential connectivity only provide a few reliable assignments.     

Mars - robust automatic backbone assignment of proteins

MARS a program for robust automatic backbone assignment of (13)C/(15)N labeled proteins is presented. MARS does not require tight thresholds for establishing sequential connectivity or detailed adjustment of these thresholds and it can work with a wide variety of NMR experiments. Using only (13)C(alpha)/(13)C(beta) connectivity information, MARS allows automatic, error-free assignment of 96% of the 370-residue maltose-binding protein. MARS can successfully be used when data are missing for a substantial portion of residues or for proteins with very high chemical shift degeneracy such as partially or fully unfolded proteins. Other sources of information, such as residue specific information or known assignments from a homologues protein, can be included into the assignment process. MARS exports its result in SPARKY format. This allows visual validation and integration of automated and manual assignment.     

Application of automated NOE assignment to three-dimensional structure refinement of a 28 kDa single-chain T cell receptor

An automated procedure for NOE assignment and three-dimensional structure refinement is presented. The input to the procedure consists of (1) an ensemble of preliminary protein NMR structures, (2) partial sequence-specific assignments for the protein and (3) the positions and volumes of unassigned NOESY cross peaks. Chemical shifts for unassigned side chain protons are predicted from the preliminary structures. The chemical shifts and unassigned NOESY cross peaks are input to an automated procedure for NOE assignment and structure calculation (ARIA) [Nilges et al. (1997) J. Mol. Biol., 269, 408–422]. ARIA is optimized for the task of structure refinement of larger proteins. Errors are filtered to ensure that sequence-specific assignments are reliable. The procedure is applied to the 27.8 kDa single-chain T cell receptor (scTCR). Preliminary NMR structures, nearly complete backbone assignments, partial assignments of side chain protons and more than 1300 unassigned NOESY cross peaks are input. Using the procedure, the resonant frequencies of more than 40 additional side chain protons are assigned. Over 400 new NOE cross peaks are assigned unambiguously. Distances derived from the automatically assigned NOEs improve the precision and quality of calculated scTCR structures. In the refined structures, a hydrophobic cluster of side chains on the scTCR surface that binds major histocompatibility complex (MHC)/antigen is revealed. It is composed of the side chains of residues from three loops and stabilizes the conformation of residues that interact with MHC.     

A complete algorithm to resolve ambiguity for intersubunit NOE assignment in structure determination of symmetric homo-oligomers

Assignment of nuclear Overhauser effect (NOE) data is a key bottleneck in structure determination by NMR. NOE assignment resolves the ambiguity as to which pair of protons generated the observed NOE peaks, and thus should be restrained in structure determination. In the case of intersubunit NOEs in symmetric homo-oligomers, the ambiguity includes both the identities of the protons within a subunit, and the identities of the subunits to which they belong. This paper develops an algorithm for simultaneous intersubunit NOE assignment and C(n) symmetric homo-oligomeric structure determinations, given the subunit structure. By using a configuration space framework, our algorithm guarantees completeness, in that it identifies structures representing, to within a user-defined similarity level, every structure consistent with the available data (ambiguous or not). However, while our approach is complete in considering all conformations and assignments, it avoids explicit enumeration of the exponential number of combinations of possible assignments. Our algorithm can draw two types of conclusions not possible under previous methods: (1) that different assignments for an NOE would lead to different structural classes, or (2) that it is not necessary to uniquely assign an NOE, since it would have little impact on structural precision. We demonstrate on two test proteins that our method reduces the average number of possible assignments per NOE by a factor of 2.6 for MinE and 4.2 for CCMP. It results in high structural precision, reducing the average variance in atomic positions by factors of 1.5 and 3.6, respectively.     

1H, 13C, and 15N NMR assignments and global folding pattern of the RNA-binding domain of the human hnRNP C proteins.

The hnRNP C1 and C2 proteins are abundant nuclear proteins that bind avidly to heterogeneous nuclear RNAs (hnRNAs) and appear to be involved with pre-mRNA processing. The RNA-binding activity of the hnRNP C proteins is contained in the amino-terminal 94 amino acid RNA-binding domain (RBD) that is identical for these two proteins. We have obtained the 1H, 13C, and 15N NMR assignments for the RBD of the human hnRNP C proteins. The assignment process was facilitated by extensive utilization of three- and four-dimensional heteronuclear-edited spectra. Sequential assignments of the backbone resonances were made using a combination of 15N-edited 3D NOESY-HMQC, 3D TOCSY-HMQC, and 3D TOCSY-NOESY-HSQC as well as 3D HNCA, HNCO, and HCACO spectra. Side-chain resonances were assigned using 3D HCCH-COSY and 3D HCH-TOCSY spectra. Four-dimensional 13C/13C-edited NOESY and 13C/15N-edited NOESY experiments were used to unambigously resolve NOEs. The overall global folding pattern was established by calculating a set of preliminary structures using constraints derived from the sequential NOEs and a small number of long-range NOEs. The beta alpha beta-beta alpha beta domain structure exhibits an antiparallel beta-sheet with the conserved RNP 1 and RNP 2 sequences [Dreyfuss et al. (1988) Trends Biochem. Sci. 13, 86-91] located adjacent to one another as the two inner strands of the beta-sheet.     

Determination of local conformational stability in fragment 96-133 of bovine growth hormone by high-resolution H NMR spectroscopy

The specific assignment of resonances in the 400-MHz nuclear magnetic resonance (NMR) spectrum of fragment 96-133 (AII) of bovine growth hormone (bSt) is described. Assignments have been made with homonuclear two-dimensional techniques, in particular that of sequential resonance assignment. Complete assignments were possible for the spin systems of 16 residues out of a total of 38 and partial assignments for another 5. Assignment of resonances to either residue type or a class of residue was possible for a number of other spin systems. Analysis of the type of nuclear Overhauser effect (NOE) indicates that segments 96-110 and 130-133 are nonregular stable structures and that the segment 111-127, which putatively spans the alpha-helix, is not sufficiently stable to generate NOEs.     

1H NMR assignments of apo calcyclin and comparative structural analysis with calbindin D9k and S100 beta.

The homodimeric S100 protein calcyclin has been studied in the apo state by two-dimensional 1H NMR spectroscopy. Using a combination of scalar correlation and NOE experiments, sequence-specific 1H NMR assignments were obtained for all but one backbone and > 90% of the side-chain resonances. To our knowledge, the 2 x 90 residue (20 kDa) calcyclin dimer is the largest protein system for which such complete assignments have been made by purely homonuclear methods. Sequential and medium-range NOEs and slowly exchanging backbone amide protons identified directly the four helices and the short antiparallel beta-type interaction between the two binding loops that comprise each subunit of the dimer. Further analysis of NOEs enabled the unambiguous assignment of 556 intrasubunit distance constraints, 24 intrasubunit hydrogen bonding constraints, and 2 x 26 intersubunit distance constraints. The conformation of the monomer subunit was refined by distance geometry and restrained molecular dynamics calculations using the intrasubunit constraints only. Calculation of the dimer structure starting from this conformational ensemble has been reported elsewhere. The extent of structural homology among the apo calcyclin subunit, the monomer subunit of apo S100 beta, and monomeric apo calbindin D9k has been examined in detail by comparing 1H NMR chemical shifts and secondary structures. This analysis was extended to a comprehensive comparison of the three-dimensional structures of the calcyclin monomer subunit and calbindin D9k, which revealed greater similarity in the packing of their hydrophobic cores than was anticipated previously. Together, these results support the hypothesis that all members of the S100 family have similar core structures and similar modes of dimerization. Analysis of the amphiphilicity of Helix IV is used to explain why calbindin D9k is monomeric, but full-length S100 proteins form homodimers.     

A multiple-template approach to protein threading

Most threading methods predict the structure of a protein using only a single template. Due to the increasing number of solved structures, a protein without solved structure is very likely to have more than one similar template structures. Therefore, a natural question to ask is if we can improve modeling accuracy using multiple templates. This article describes a new multiple-template threading method to answer this question. At the heart of this multiple-template threading method is a novel probabilistic-consistency algorithm that can accurately align a single protein sequence simultaneously to multiple templates. Experimental results indicate that our multiple-template method can improve pairwise sequence-template alignment accuracy and generate models with better quality than single-template models even if they are built from the best single templates (P-value <10(-6)) while many popular multiple sequence/structure alignment tools fail to do so. The underlying reason is that our probabilistic-consistency algorithm can generate accurate multiple sequence/template alignments. In another word, without an accurate multiple sequence/template alignment, the modeling accuracy cannot be improved by simply using multiple templates to increase alignment coverage. Blindly tested on the CASP9 targets with more than one good template structures, our method outperforms all other CASP9 servers except two (Zhang-Server and QUARK of the same group). Our probabilistic-consistency algorithm can possibly be extended to align multiple protein/RNA sequences and structures.     

Robust structure-based resonance assignment for functional protein studies by NMR

High-throughput functional protein NMR studies, like protein interactions or dynamics, require an automated approach for the assignment of the protein backbone. With the availability of a growing number of protein 3D structures, a new class of automated approaches, called structure-based assignment, has been developed quite recently. Structure-based approaches use primarily NMR input data that are not based on J-coupling and for which connections between residues are not limited by through bonds magnetization transfer efficiency. We present here a robust structure-based assignment approach using mainly H ^N –H ^N NOEs networks, as well as ¹ H–¹⁵ N residual dipolar couplings and chemical shifts. The NOEnet complete search algorithm is robust against assignment errors, even for sparse input data. Instead of a unique and partly erroneous assignment solution, an optimal assignment ensemble with an accuracy equal or near to 100% is given by NOEnet. We show that even low precision assignment ensembles give enough information for functional studies, like modeling of protein-complexes. Finally, the combination of NOEnet with a low number of ambiguous J-coupling sequential connectivities yields a high precision assignment ensemble. NOEnet will be available under: .     

RDC derived protein backbone resonance assignment using fragment assembly

Experimental residual dipolar couplings (RDCs) in combination with structural models have the potential for accelerating the protein backbone resonance assignment process because RDCs can be measured accurately and interpreted quantitatively. However, this application has been limited due to the need for very high-resolution structural templates. Here, we introduce a new approach to resonance assignment based on optimal agreement between the experimental and calculated RDCs from a structural template that contains all assignable residues. To overcome the inherent computational complexity of such a global search, we have adopted an efficient two-stage search algorithm and included connectivity data from conventional assignment experiments. In the first stage, a list of strings of resonances (CA-links) is generated via exhaustive searches for short segments of sequentially connected residues in a protein (local templates), and then ranked by the agreement of the experimental ¹³C_α chemical shifts and ¹⁵N-¹H RDCs to the predicted values for each local template. In the second stage, the top CA-links for different local templates in stage I are combinatorially connected to produce CA-links for all assignable residues. The resulting CA-links are ranked for resonance assignment according to their measured RDCs and predicted values from a tertiary structure. Since the final RDC ranking of CA-links includes all assignable residues and the assignment is derived from a “global minimum”, our approach is far less reliant on the quality of experimental data and structural templates. The present approach is validated with the assignments of several proteins, including a 42 kDa maltose binding protein (MBP) using RDCs and structural templates of varying quality. Since backbone resonance assignment is an essential first step for most of biomolecular NMR applications and is often a bottleneck for large systems, we expect that this new approach will improve the efficiency of the assignment process for small and medium size proteins and will extend the size limits assignable by current methods for proteins with structural models.     

Automatic methyl assignment in large proteins by the MAGIC algorithm

Selective methyl labeling is an extremely powerful approach to study the structure, dynamics and function of biomolecules by NMR. Despite spectacular progress in the field, such studies remain rather limited in number. One of the main obstacles remains the assignment of the methyl resonances, which is labor intensive and error prone. Typically, NOESY crosspeak patterns are manually correlated to the available crystal structure or an in silico template model of the protein. Here, we propose methyl assignment by graphing inference construct, an exhaustive search algorithm with no peak network definition requirement. In order to overcome the combinatorial problem, the exhaustive search is performed locally, i.e. for a small number of methyls connected through-space according to experimental 3D methyl NOESY data. The local network approach drastically reduces the search space. Only the best local assignments are combined to provide the final output. Assignments that match the data with comparable scores are made available to the user for cross-validation by additional experiments such as methyl-amide NOEs. Several NMR datasets for proteins in the 25–50 kDa range were used during development and for performance evaluation against the manually assigned data. We show that the algorithm is robust, reliable and greatly speeds up the methyl assignment task.     

Modified genetic algorithm resolves ambiguous NOE restraints and reduces unsightly NOE violations

In an ideal world, every NOE cross peak would have a unique assignment. However, the interpretation of NOE peaks is frequently complicated by overlapping resonances. In theory, ambiguous assignments could be resolved by performing separate structure calculations with each possible interpretation. Unfortunately, this would require an astronomical amount of computing time. A modified genetic algorithm has been developed that efficiently resolves hundreds of ambiguous restraints in parallel. Each NOE assignment becomes a gene that can be passed on to a new generation. New individuals are constructed by making a constraint lists from a subset of the genes. The constraint lists are then tested for self-consistency by using molecular dynamics to generate new structures for each list. To a first-degree approximation, there is enough information retained in each list to determine the global fold of the protein. Self-consistent constraint lists receive higher scores and their genes (or NOEs) stand a better chance of surviving into the next generation. The process selects NOEs that are consistent with the global fold. Under normal conditions, the program converges in 3 to 8 generations using 70 structures per generation. The final constraints are self-consistent and contain almost no residual NOE violations.     

Increased detection of structural templates using alignments of designed sequences

Protein structure prediction by comparative modeling benefits greatly from the use of multiple sequence alignment information to improve the accuracy of structural template identification and the alignment of target sequences to structural templates. Unfortunately, this benefit is limited to those protein sequences for which at least several natural sequence homologues exist. We show here that the use of large diverse alignments of computationally designed protein sequences confers many of the same benefits as natural sequences in identifying structural templates for comparative modeling targets. A large-scale massively parallelized application of an all-atom protein design algorithm, including a simple model of peptide backbone flexibility, has allowed us to generate 500 diverse, non-native, high-quality sequences for each of 264 protein structures in our test set. PSI-BLAST searches using the sequence profiles generated from the designed sequences ("reverse" BLAST searches) give near-perfect accuracy in identifying true structural homologues of the parent structure, with 54% coverage. In 41 of 49 genomes scanned using reverse BLAST searches, at least one novel structural template (not found by the standard method of PSI-BLAST against PDB) is identified. Further improvements in coverage, through optimizing the scoring function used to design sequences and continued application to new protein structures beyond the test set, will allow this method to mature into a useful strategy for identifying distantly related structural templates.     

Automated amino acid side-chain NMR assignment of proteins using <Superscript>13</Superscript>C- and <Superscript>15</Superscript>N-resolved 3D [<Superscript>1</Superscript>H,<Superscript>1</Superscript>H]-NOESY

ASCAN is a new algorithm for automatic sequence-specific NMR assignment of amino acid side-chains in proteins, which uses as input the primary structure of the protein, chemical shift lists of (1)H(N), (15)N, (13)C(alpha), (13)C(beta) and possibly (1)H(alpha) from the previous polypeptide backbone assignment, and one or several 3D (13)C- or (15)N-resolved [(1)H,(1)H]-NOESY spectra. ASCAN has also been laid out for the use of TOCSY-type data sets as supplementary input. The program assigns new resonances based on comparison of the NMR signals expected from the chemical structure with the experimentally observed NOESY peak patterns. The core parts of the algorithm are a procedure for generating expected peak positions, which is based on variable combinations of assigned and unassigned resonances that arise for the different amino acid types during the assignment procedure, and a corresponding set of acceptance criteria for assignments based on the NMR experiments used. Expected patterns of NOESY cross peaks involving unassigned resonances are generated using the list of previously assigned resonances, and tentative chemical shift values for the unassigned signals taken from the BMRB statistics for globular proteins. Use of this approach with the 101-amino acid residue protein FimD(25-125) resulted in 84% of the hydrogen atoms and their covalently bound heavy atoms being assigned with a correctness rate of 90%. Use of these side-chain assignments as input for automated NOE assignment and structure calculation with the ATNOS/CANDID/DYANA program suite yielded structure bundles of comparable quality, in terms of precision and accuracy of the atomic coordinates, as those of a reference structure determined with interactive assignment procedures. A rationale for the high quality of the ASCAN-based structure determination results from an analysis of the distribution of the assigned side chains, which revealed near-complete assignments in the core of the protein, with most of the incompletely assigned residues located at or near the protein surface.     

Sparsely-sampled,high-resolution 4-D omit spectra for detection and assignment of intermolecular NOEs of protein complexes

Unambiguous detection and assignment of intermolecular NOEs are essential for structure determination of protein complexes by NMR. Such information has traditionally been obtained with 3-D half-filtered experiments, where scalar coupling-based purging of intramolecular signals allows for selective detection of intermolecular NOEs. However, due to the large variation of ¹J_HC scalar couplings and limited chemical shift dispersion in the indirect proton dimension, it is difficult to obtain reliable and complete assignments of interfacial NOEs. Here, we demonstrate a strategy that combines selective labeling and high-resolution 4-D NOE spectroscopy with sparse sampling for reliable identification and assignment of intermolecular NOEs. Spectral subtraction of component-labeled complexes from a uniformly-labeled protein complex yields an “omit” spectrum containing positive intermolecular NOEs with little signal degeneracy. Such a strategy can be broadly applied to unbiased detection, assignment and presentation of intermolecular NOEs of protein complexes.     

MAP-XSII: an improved program for the automatic assignment of methyl resonances in large proteins

NMR studies of large proteins have gathered much interest in recent years, especially after methyl-transverse relaxation optimized spectroscopy was successfully applied to systems as large as ~1 MDa in molecular weight. However, to fully take advantage of these spectra, there is a need for convenient and robust methods for making resonance assignments rapidly. Here, we present an improved version of our program MAP-XS (methyl assignment prediction from X-ray structure) for the automatic assignment of methyl peaks, based on nuclear Overhauser effects (NOE) correlations and chemical shifts together with available structures. No manual analysis of the NOE data is needed in this new version, which helps to further accelerate the assignment process. A refined algorithm as well as more efficient sampling produces results from single runs of MAP-XSII using unanalyzed NOE data are comparable to those achieved by the old version using manually curated data with every NOE peak correctly attributed to the two related methyl peaks; in addition, checking the results from multiple parallel runs against each other provides an effective mechanism for getting rid of the wrong assignments while keeping the correct ones, which significantly improves the reliability of final assignments. The new program is tested against three different proteins and delivers ~95 % correct assignments; positive results are also achieved for tests using different cut-off distances for NOEs, structures of lower resolutions, and ambiguous residue types.     

Complete assignment of the deoxyribose 5'/5" proton resonances of the EcoRI DNA sequence using isotropic mixing

Using two-dimensional isotropic mixing spectroscopy all 5'/5" proton resonances of the EcoRI restriction site DNA dodecamer [d(CGCGAATTCGCG)]2 have been assigned. This completes the previous assignments of 1'H to 4'H resonances of the deoxyribose spin systems (Hare et al., 1983). With mixing times of up to 500 ms, many of these resonances showed connectivities of 5'/5" protons in the two-dimensional isotropic mixing spectrum. Relying only on through-bond connectivities makes these assignments independent of assumptions about the conformation of the DNA oligonucleotide. The assignment of the 5'H/5"H resonances will allow the interpretation of intra- and interresidue NOEs to these protons, providing information about the DNA backbone conformation.