Determination of protein structures from nuclear magnetic resonance data using a restrained molecular dynamics approach: the lac repressor DNA binding domain

Introduction to a special issue of Magnetic Resonance in honour of Robert Kaptein at the occasion of his 80th birthday

This publication, in honour of Robert Kaptein's 80th birthday, contains contributions from colleagues, many of whom have worked with him, and others who admire his work and have been stimulated by his research. The contributions show current research in biomolecular NMR, spin hyperpolarisation and spin chemistry, including CIDNP (chemically induced dynamic nuclear polarisation), topics to which he has contributed enormously. His proposal of the radical pair mechanism was the birth of the field of spin chemistry, and the laser CIDNP NMR experiment on a protein was a major breakthrough in hyperpolarisation research. He set milestones for biomolecular NMR by developing computational methods for protein structure determination, including restrained molecular dynamics and 3D NMR methodology. With a lac repressor headpiece, he determined one of the first protein structures determined by NMR. His studies of the lac repressor provided the first examples of detailed studies of protein nucleic acid complexes by NMR. This deepened our understanding of protein DNA recognition and led to a molecular model for protein sliding along the DNA. Furthermore, he played a leading role in establishing the cluster of NMR large-scale facilities in Europe. This editorial gives an introduction to the publication and is followed by a biography describing his contributions to magnetic resonance.

Molecular dynamics simulation of configurational ensembles compatible with experimental FRET efficiency data through a restraint on instantaneous FRET efficiencies

Förster resonance energy transfer (FRET) measurements are widely used to investigate (bio)molecular interactions or/and association. FRET efficiencies, the primary data obtained from this method, give, in combination with the common assumption of isotropic chromophore orientation, detailed insight into the lengthscale of molecular phenomena. This study illustrates the application of a FRET efficiency restraint during classical atomistic molecular dynamics simulations of a mutant mastoparan X peptide in either water or 7 M aqueous urea. The restraint forces acting on the donor and acceptor chromophores ensure that the sampled peptide configurational ensemble satisfies the experimental primary data by modifying interchromophore separation and chromophore transition dipole moment orientations. By means of a conformational cluster analysis, it is seen that indeed different configurational ensembles may be sampled without and with application of the restraint. In particular, while the FRET efficiency and interchromophore distances monitored in an unrestrained simulation may differ from the experimentally-determined values, they can be brought in agreement with experimental data through usage of the FRET efficiency restraining potential. Furthermore, the present results suggest that the assumption of isotropic chromophore orientation is not always justified. The FRET efficiency restraint allows the generation of configurational ensembles that may not be accessible with unrestrained simulations, and thereby supports a meaningful interpretation of experimental FRET results in terms of the underlying molecular degrees of freedom. Thus, it offers an additional tool to connect the realms of computer and wet-lab experimentation.

Biomolecular structure refinement using the GROMOS simulation software

For the understanding of cellular processes the molecular structure of biomolecules has to be accurately determined. Initial models can be significantly improved by structure refinement techniques. Here, we present the refinement methods and analysis techniques implemented in the GROMOS software for biomolecular simulation. The methodology and some implementation details of the computation of NMR NOE data, (3)J-couplings and residual dipolar couplings, X-ray scattering intensities from crystals and solutions and neutron scattering intensities used in GROMOS is described and refinement strategies and concepts are discussed using example applications. The GROMOS software allows structure refinement combining different types of experimental data with different types of restraining functions, while using a variety of methods to enhance conformational searching and sampling and the thermodynamically calibrated GROMOS force field for biomolecular simulation.

Glycine insertion in the hinge region of lactose repressor protein alters DNA binding

Amino acid alterations were designed at the C terminus of the hinge segment (amino acids approximately 51-59) that links two functional domains within lactose repressor protein (LacI). Gly was introduced between Gly(58) and Lys(59) to generate Gly(58+1); Gln(60) was changed to Gly or Pro, and up to three additional glycines were inserted following Gln(60) --> Gly. All mutant proteins exhibited purification behavior, CD spectra, assembly state, and inducer binding properties similar to wild-type LacI and only small differences in trypsin proteolysis patterns. In contrast, significant differences were observed in DNA binding properties. Gly(58+1) exhibited a decrease of approximately 100-fold in affinity for O(1) operator, and sequential Gly insertion C-terminal to Gln(60) --> Gly resulted in progressively decreased affinity for O(1) operator, approaching nonspecific levels for insertion of >/=2 glycines. Where sufficient affinity for O(1) operator existed, decreased binding to O(1) in the presence of inducer indicated no disruption in the allosteric response for these proteins. Collectively, these results indicate that flexibility and/or spacing between the core and N-terminal domains did not significantly affect folding or assembly, but these alterations in the hinge domain profoundly altered affinity of the lactose repressor protein for its wild-type target sequence.

Synthesis and conformational analysis by 1H NMR and restrained molecular dynamics simulations of the cyclic decapeptide [Ser-Tyr-Ser-Met-Glu-His-Phe-Arg-Trp-Gly]

The design of enzyme mimics with therapeutic and industrial applications has interested both experimental and computational chemists for several decades. Recent advances in the computational methodology of restrained molecular dynamics, used in conjunction with data obtained from two-dimensional 1H NMR spectroscopy, make it a promising method to study peptide and protein structure and function. Several issues, however, need to be addressed in order to assess the validity of this method for its explanatory and predictive value. Among the issues addressed in this study are: the accuracy and generizability of the GROMOS peptide molecular mechanics force field; the effect of inclusion of solvent on the simulations; and the effect of different types of restraining algorithms on the computational results. The decapeptide Ser-Tyr-Ser-Met-Glu-His-Phe-Arg-Trp-Gly, which corresponds to the sequence of ACTH1-10, has been synthesized, cyclized, and studied by two-dimensional 1H NMR spectroscopy. Restrained molecular dynamics (RMD) and time-averaged restrained molecular dynamics (TARMD) simulations were carried out on four different distance-geometry starting structures in order to determine and contrast the behavior of cyclic ACTH1-10 in vacuum and in solution. For the RMD simulations, the structures did not fit the NOE data well, even at high values of the restraining potential. The TARMD simulation method, however, was able to give structures that fit the NOE data at high values of the restraining potential. In both cases, inclusion of explicit solvent molecules in the simulation had little effect on the quality of the fit, although it was found to dampen the motion of the cyclic peptide. For both simulation techniques, the number and size of the NOE violations increased as the restraining potential approached zero. This is due, presumably, to inadequacies in the force field. Additional TARMD vacuum-phase simulations, run with a larger memory length or with a larger sampling size (16 additional distance-geometry structures), yielded no significantly different results. The computed data were then analyzed to help explain the sparse NOE data and poor chymotryptic activity of the cyclic peptide. Cyclic ACTH1-10, which contains the functional moieties of the catalytic triad of chymotrypsin, was evaluated as a potential mimic of chymotrypsin by measurement of the rate of hydrolysis of esters of L- and D-phenylalanine. The poor rate of hydrolysis is attributed to the flexibility of the decapeptide, the motion of the side chains, which result in the absence of long-range NOEs, the small size of the macrocycle relative to that of the substrate, and the inappropriate orientation of the Gly, His, and Ser residues. The results demonstrate the utility of this method in computer-aided molecular design of cyclic peptides and suggest structural modifications for future work based on a larger and more rigid peptide framework.

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.

Two-dimensional 1H, 15N NMR investigation of uniformly 15N-labeled lac repressor headpiece

15N uniformly labeled lac repressor and lac repressor headpiece were prepared. 15N NMR spectra of lac repressor were shown resolution inadequate for detailed study while the data showed that the 15N labeled N-terminal part of the protein is quite suitable for this type of study allowing future investigation of the specific interaction of the lac repressor headpiece with the lac operator. We report here the total assignment of proton 1H and nitrogen 15NH backbone resonances of this headpiece in the free state. Assignments of the 15N resonances of the protein were obtained in a sequential manner using heteronuclear multiple quantum coherence (HMQC), relayed HMQC nuclear Overhauser and relayed HMQC-HOHAHA spectroscopy. More than 80 per cent of residues were assigned by their 15NH(i)-N1H(i + 1) and 15NH(i)-N1H(i - 1) connectivities. Values of the 3JNH alpha splitting for 39 of the 51 residues of the headpiece were extracted from HMQC and HMQC-J. The observed 15NH(i)-C beta H cross peaks and the 3JNH alpha coupling constants values are in agreement with the three alpha-helices previously described [Zuiderweg, E.R.P., Scheek, R.M., Boelens, R., van Gunsteren, W.F. and Kaptein, R., Biochimie 67, 707 (1985)]. The 3JNH alpha coupling constants can be now used for a more confident determination of the lac repressor headpiece. From these values it is shown that the geometry of the ends of the second and third alpha-helices exhibit deviation from the canonical alpha-helix structure. On the basis of NOEs and 3JNH alpha values, the geometry of the turn of the helix-turn-helix motif is discussed.

31P NMR spectra of oligodeoxyribonucleotide duplex lac operator-repressor headpiece complexes: importance of phosphate ester backbone flexibility in protein-DNA recognition

The 31P NMR spectra of various 14-base-pair lac operators bound to both wild-type and mutant lac repressor headpiece proteins were analyzed to provide information on the backbone conformation in the complexes. The 31P NMR spectrum of a wild-type symmetrical operator, d(TGTGAGCGCTCACA)2, bound to the N-terminal 56-residue headpiece fragment of a Y7I mutant repressor was nearly identical to the spectrum of the same operator bound to the wild-type repressor headpiece. In contrast, the 31P NMR spectrum of the mutant operator, d(TATAGAGCGCTCATA)2, wild-type headpiece complex was significantly perturbed relative to the wild-type repressor-operator complex. The 31P chemical shifts of the phosphates of a second mutant operator, d(TGTGTGCGCACACA)2, showed small but specific changes upon complexation with either the wild-type or mutant headpiece. The 31P chemical shifts of the phosphates of a third mutant operator, d(TCTGAGCGCTCAGA)2, showed no perturbations upon addition of the wild-type headpiece. The 31P NMR results provide further evidence for predominant recognition of the 5'-strand of the 5'-TGTGA/3'-ACACT binding site in a 2:1 protein to headpiece complex. It is proposed that specific, strong-binding operator-protein complexes retain the inherent phosphate ester conformational flexibility of the operator itself, whereas the phosphate esters are conformationally restricted in the weak-binding operator-protein complexes. This retention of backbone torsional freedom in strong complexes is entropically favorable and provides a new (and speculative) mechanism for protein discrimination of different operator binding sites. It demonstrates the potential importance of phosphate geometry and flexibility on protein recognition and binding.

100ps molecular dynamic simulation of d(TATCACC)2

A heptanucleotide sequence d(TATCACC)2 from OR3 region of bacteriophage lambda is considered sufficient for the recognition of Cro protein. We present here results on molecular dynamic simulations on this sequence for 100 ps in 0.02 ps interval. The simulations are done using computer program GROMOS. The conformational results are averaged over each ps. The IUPAC torsional parameters for 100 conformations are illustrated using a wheal and a dial systems. Several other stereochemical parameters such as H-bonding lengths and angles, sugar puckers, helix twist and roll angles as also distances between opposite strand phosphorus are depicted graphically. We find that there is rupture of terminal H-bonds. The bases are tilted and shifted away from the helix axis giving rise to bifurcated H-bonds. H-bonds are seen even in between different base pairs. The role of these dynamic structural changes in the recognition of OR3 operator by Cro protein is discussed in the paper.

A model of the lac repressor-operator complex based on physical and genetic data

Computer graphics were used to build a molecular model of the complex of Lac repressor and lac operator. The model is based (a) on the NMR data of the Kaptein group [Boelens, R., Lamerichs, R. M. J. N., Rullmann, J. A. C., van Boom, J. H. & Kaptein, R. (1988) Protein Sequence Data Anal. 1, 487-498] and (b) on our genetic and biochemical data including specificity changes [Lehming, N., Sartorius, J., Kisters-Woike, B., von Wilcken-Bergmann, B. & Müller-Hill, B. (1990) EMBO J. 9, 615-621]. Effects of amino acid exchanges in the recognition helix could be predicted by the model and were subsequently tested and confirmed by genetic experiments. Comparison of the modelled lac complex with the known crystallographic structures of several helix-turn-helix DNA complexes reveals striking similarities and suggests rules which govern the recognition between particular amino acid side chains and particular base pairs in these systems.

Protein solution structure determination using distances from two-dimensional nuclear Overhauser effect experiments: effect of approximations on the accuracy of derived structures

Solution structures for many proteins have been determined to date utilizing interproton distance constraints estimated from two-dimensional nuclear Overhauser effect (2D NOE) spectra. Although the simple isolated spin pair approximation (ISPA) generally used can result in systematic errors in distances, the large number of constraints enables protein structure to be defined with reasonably high resolution. Effects of these systematic errors on the resulting protein structure are examined. Iterative relaxation matrix calculations, which account for dipolar interactions between all protons in a molecule, can accurately determine internuclear distances with little or no a priori knowledge of the molecular structure. The value of this additional complexity is also addressed. To assess these distance determination methods, hypothetical "experimental" data, including random noise and peak overlap, are calculated for an arbitrary "true" protein structure. Three methods of obtaining distance constraints from 2D NOE peak intensities are examined: one entails a conservative use of ISPA, one assumes the ISPA to be fairly accurate, and one utilizes an iterative relaxation matrix method called MARDIGRAS (matrix analysis of relaxation for discerning the geometry of an aqueous structure), developed in this laboratory. A distance geometry algorithm was used to generate a family of structures for each distance set. The quality of the average structure from each family was good. The root-mean-square deviation of that average structure from the true structure was improved about 2-5% using the more restrictive rather than the more conservative ISPA approach. Use of MARDIGRAS in a conservative fashion--i.e., with a poor initial model--resulted in improvement in the root-mean-square deviation by 8-15%. With a better initial model, MARDIGRAS obtained even more accurate distances. MARDIGRAS also permits analysis of 2D NOE data at longer mixing times, yielding additional distances. Use of more restrictive ISPA distances did, however, result in a few systematically incorrect structural features in local regions of the protein, producing distortions of 2-3 A. Comparison between experimental data and spectra calculated for the structures correlates with root-mean-square deviation, offering a method of structure evaluation. An R factor for evaluating fit between experimental and calculated 2D NOE intensities is proposed.

Two-dimensional 1H and 31P NMR spectra and restrained molecular dynamics structure of an oligodeoxyribonucleotide duplex refined via a hybrid relaxation matrix procedure

Assignment of the 1H and 31P resonances of a decamer DNA duplex, d(CGCTTAAGCG)2 was determined by two-dimensional COSY, NOESY and 1H-31P Pure Absorption phase Constant time (PAC) heteronuclear correlation spectroscopy. The solution structure of the decamer was calculated by an iterative hybrid relaxation matrix method combined with NOESY-distance restrained molecular dynamics. The distances from the 2D NOESY spectra were calculated from the relaxation rate matrix which were evaluated from a hybrid NOESY volume matrix comprising elements from the experiment and those calculated from an initial structure. The hybrid matrix-derived distances were then used in a restrained molecular dynamics procedure to obtain a new structure that better approximates the NOESY spectra. The resulting partially refined structure was then used to calculate an improved theoretical NOESY volume matrix which is once again merged with the experimental matrix until refinement is complete. JH3'-P coupling constants for each of the phosphates of the decamer were obtained from 1H-31P J-resolved selective proton flip 2D spectra. By using a modified Karplus relationship the C4'-C3'-O3'-P torsional angles (epsilon) were obtained. Comparison of the 31P chemical shifts and JH3'-P coupling constants of this sequence has allowed a greater insight into the various factors responsible for 31P chemical shift variations in oligonucleotides. It also provides an important probe of the sequence-dependent structural variation of the deoxyribose phosphate backbone of DNA in solution. These correlations are consistent with the hypothesis that changes in local helical structure perturb the deoxyribose phosphate backbone. The variation of the 31P chemical shift, and the degree of this variation from one base step to the next is proposed as a potential probe of local helical conformation within the DNA double helix. The pattern of calculated epsilon and zeta torsional angles from the restrained molecular dynamics refinement agrees quite well with the measured JH3'-P coupling constants. Thus, the local helical parameters determine the length of the phosphodiester backbone which in turn constrains the phosphate in various allowed conformations.

An NMR-based molecular dynamics simulation of the interaction of the lac repressor headpiece and its operator in aqueous solution

The results of a 125 psec molecular dynamics simulation of a lac headpiece-operator complex in aqueous solution are reported. The complex satisfies essentially all experimental distance information derived from two-dimensional nuclear magnetic resonance (2-D-NMR) studies. The interaction between lac repressor headpiece and its operator is based on many direct- and water-mediated hydrogen bonds and nonpolar contacts which allow the formation of a tight complex. No stable hydrogen bonds between side chains and bases are found, while specific contacts occur between both nonpolar groups and, to a lesser extent, through water-mediated hydrogen bonds. The simulated complex structure in water is intrinsically stable without application of nuclear Overhauser effect (NOE) distance restraints, while being compatible with most of the available biochemical, genetic, and chemically induced dynamic nuclear polarization (CIDNP) data.

Backbone folding of the polypeptide cardiac stimulant anthopleurin-A determined by nuclear magnetic resonance, distance geometry and molecular dynamics

The solution conformation of the cardiac stimulatory sea anemone polypeptide anthopleurin-A has been characterised using distance geometry and restrained molecular dynamics calculations. A set of 253 approximate interproton distance restraints and 14 peptide backbone torsion angle restraints derived from two-dimensional 1H-NMR spectra at 500 MHz were used as input for these calculations. 13 structures generated by either metric matrix or variable target function distance geometry calculations were refined using energy minimisation and restrained molecular dynamics. The resulting structures contain a region of twisted antiparellel beta-sheet to which two separate regions of unordered chain are linked by three disulphide bonds. Two loops, one including Pro-41 and the other encompassing residues 10-18, are poorly defined by the NOE data.

Photo-CIDNP study of the interaction between lac repressor headpiece and lac operator DNA

Lac repressor headpiece (HP) and intact lac repressor have been studied using the photo-CIDNP method. At neutral pH histidine 29, tyrosines 7, 12 and 17 and methionine 1 are polarised. His-29 polarizations are weaker and broader in HP59 than in HP51 indicating that the C-terminal octapeptide in HP59 adopts a conformation that allows an interaction with His-29. The photo-CIDNP spectra of intact lac repressor and HP51 are very similar, showing that the same residues are accessible to the photo-excited flavin. An equimolar mixture of HP51 and a 14 base pair lac operator fragment strongly suppresses the photo-CIDNP effect of tyrosines 7 and 17 and abolishes the His-29 polarizations. The results are compared with earlier photo-CIDNP measurements on a complex of headpiece with poly[d(AT)] and with a model derived from a 2D NMR study on a lac headpiece-operator complex.

Chemical modification of arginine residues in the lactose repressor

The lactose repressor protein was chemically modified with 2,3-butanedione and phenylglyoxal. Arginine reaction was quantitated by either amino acid analysis or incorporation of 14C-labeled phenylglyoxal. Inducer binding activity was unaffected by the modification of arginine residues, while both operator and nonspecific DNA binding activities were diminished, although to differing degrees. The correlation of the decrease in DNA binding activities with the modification of approximately 1-2 equiv of arginine per monomer suggests increased reactivity of a functionally essential residue(s). For both reagents, operator DNA binding activity was protected by the presence of calf thymus DNA, and the extent of reaction with phenylglyoxal was simultaneously diminished. This protection presumably results from steric restriction of reagent access to an arginine(s) that is (are) essential for DNA binding interactions. These experiments suggest that there is (are) an essential reactive arginine(s) critical for repressor binding to DNA.

Complex of lac repressor headpiece with a 14 base-pair lac operator fragment studied by two-dimensional nuclear magnetic resonance

Two-dimensional nuclear Overhauser enhancement spectra are presented of the complex of lac repressor headpiece with a 14 base-pair lac operator fragment. Analysis of nuclear Overhauser enhancements observed between protein and DNA shows that the second helix of the headpiece ("the recognition helix") binds in the major groove of DNA as has been suggested, but that the orientation of this helix is approximately 180 degrees different from the proposed models.