A code governing specific binding of regulatory proteins to DNA and structure of stereospecific sites of regulatory proteins]

A model is proposed for the structure of stereospecific sites in regulatory proteins. On its basis a possible code is suggested that governs the binding of regulatory proteins at specific control sites on DNA. Stereospecific sites of regulatory proteins are assumed to contain pairs of antiparallel polypeptide chain segments which form a right-hand twisted antiparallel beta-sheet, with single-stranded regions at the ends of the beta-structure. The model predicts that binding reaction between a regulatory protein and double-helical DNA is a cooperative phenomenon and is accompanied by significant structural alteration at the stereospecific site of the protein. Half of hydrogen bonds normally existing in beta-structure are broken upon complex formation with DNA and a new set of hydrogen bonds is formed between polypeptide amide groups and DNA base pairs. In a stereospecific site, one chain (t-chain) is attached through hydrogen bonds to the carbonyl oxygens of pyramides and N3 adenines lying in one DNA strand, while the second polypeptide chain (g chain) is hydrogen bonded to the 2-amino groups of guanine residues lying in the opposite DNA strand. The amide groups serve as specific reaction sites being hydrogen bond acceptors in g-chain and hydrogen bond donors in t-chain. The single-stranded portions of t- and g-chains lying in neighbouring subunits of regulatory protein interact with each other forming deformed beta-sheets. The recognition of regulatory sequences by proteins is based on the structural complementarity between stereospecific sites of regulatory proteins and base pairs sequences at the control sites. An essential feature of these sequences is the asymmetrical distribution of guanine residues between the two DNA strands. The code predicts that there are six fundamental amino acid residues (serine, threonine, asparagine, histidine, glutamine and cysteine) whose sequence in stereospecific site determines the base pair sequence to which a given regulatory protein would bind preferentially. The code states a correspondence between four amino acid residues at the stereospecific site of regulatory protein with the two residues being in t- and g-segments, respectively, and AT(GC) base pair at the control site. It is thus possible to determine which amino acid residues in the repressor and which base pairs in the operator DNA are involved in specific interactions with each other, as exemplified by lac repressor binding to lac operator.     

The 2 A resolution structure of the sulfate-binding protein involved in active transport in Salmonella typhimurium

The crystal structure of the liganded form of the sulfate-binding protein, an initial receptor for active transport of sulfate in Salmonella typhimurium, has been solved and refined at 2.0 A resolution (1 A = 0.1 nm). The final model, which consists of 2422 non-hydrogen atoms, one sulfate substrate and 143 water molecules, yields a crystallographic R-factor of 14.0% for 16,959 reflections between 8 and 2 A. The structure deviates from ideal bond lengths and angle distances by 0.015 A and 0.037 A, respectively. The protein is ellipsoid with overall dimensions of 35 A x 35 A x 65 A and consists of two similar globular domains. The two domains are linked by three distinct peptide segments, which though widely separated in the amino acid sequence, are in close proximity in the tertiary structure. As these connecting segments are located near the periphery of the molecule, they further serve as the base or a "boundary" of the deep cleft formed between the two domains. Despite the unusual interdomain connectivity, both domains have similar supersecondary structure consisting of a central five-stranded beta-pleated sheet sandwiched by alpha-helices on either side. The arrangement of the two domains gives rise to the ellipsoidal shape and to the cleft between the two domains wherein the sulfate substrate is found and completely engulfed. A discovery of considerable importance is that the sulfate substrate is tightly held in place primarily by seven hydrogen bonds, five of which are donated by main-chain peptide NH groups, another by a serine hydroxyl and the last by the indole NH moiety of a tryptophan side-chain; there are no positively charged residues, nor cations, nor water molecules within van der Waals' distance to the sulfate dianion. All the main-chain peptide units associated with the sulfate are in turn linked (via the peptide CO group) to arrays of hydrogen bonds. Three of these arrays are composed of alternating peptide units and hydrogen bonds within the solvent-exposed part of three alpha-helices and two are linked to a histidine and an arginine residue. The sulfate-binding protein bears strong similarity to the structures of four other periplasmic binding proteins solved in our laboratory which are specific for L-arabinose, D-galactose/D-glucose, leucine/isoleucine/valine and leucine. The similarity includes the ellipsoidal shape and the two globular domain structures, each domain consisting of a central beta-pleated sheet flanked by alpha-helices.(ABSTRACT TRUNCATED AT 400 WORDS)     

Protein anatomy: structure and function of peptide fragments corresponding to the secondary structure units of barnase

Globular proteins can be decomposed into several modules or secondary structure units. It is useful to investigate the functions of such structural units in order to understand the folding units of proteins. In our previous work, barnase was divided into six peptide fragments corresponding to modules, and some of them were shown to have RNA-binding and RNase activity [Yanagawa, et al. (1993) J. Biol. Chem. 268, 5861-5865]. Barnase mutant proteins obtained by permutation of the structural units also had RNase activity [Tsuji, T. et al. (1999) J. Mol. Biol. 286, 1581-1596]. Here we investigated the structure and function of peptide fragments corresponding to secondary structure units of barnase. The results of circular dichroism spectroscopy indicated that some of the peptide fragments form helical structures in aqueous solutions containing over 30% 2,2,2-trifluoroethanol, and the S6 (94-110) peptide fragment is induced to form a beta-sheet structure in the presence of RNA. The S6 peptide fragment forms aggregate complexes with RNA. Electron microscopic analysis showed that the aggregate complexes were comprised of filaments. These results indicate that not only modules but also secondary structure units dissected from a globular protein have functional and structure-forming capabilities.     

Stereochemical Criteria for Polypeptide and Protein Chain Conformations: III. Helical and Hydrogen-Bonded Polypeptide Chains

The previous study, for a pair of peptide units, of the conformations which are allowed on the basis of stereochemical criteria of van der Waals contacts has been extended to the analysis of possible conformations of helical polypeptide chains. Computer methods have been developed which select conformations on the basis of both satisfactory interatomic contacts as well as the formation of good intramolecular hydrogen bonds. Such programs have been used to map the allowed dihedral angle pairs (varphi, psi) for helical polypeptide chains. This survey has been made for values of the N-C(a)-C' angle (tau) of 105 degrees , 110 degrees , and 115 degrees , from which the significant influence of this angle in determining allowed helical conformations can be demonstrated. Calculations have also been carried out using potential energy functions for the interaction between nonbonded atoms. The potential energy contour maps obtained in this manner are basically similar to the conformational maps calculated by the first method.     

Correlation between prebiotic amino acid compositions and contemporary proteins.

A method of energy minimization for conformational energy calculations using the least-squares technique has been reported. The method has been tested in the simple case of a pair of alanyl peptide units using the non-bonded potential energy. Starting from any point in the (φ, ψ) energy map (not necessarily near a minimum), convergence to one of the energy minima is achieved rapidly, within a few cycles of iteration.     

Theoretical Investigation of the Molecular Structure of the πκ DNA Base Pair

Abstract

The structure of the nonclassical πκ base pair (7–methyl-oxoformycin … 2,4-diaminopyrimidine) was studied at the ab initio Hartree-Fock (HF) and MP2 levels using the 6–31G* and 6–31G** basis sets. The πκ base pair is bound by three parallel hydrogen bonds with the donor-acceptor-donor recognition pattern. Recently, these bases were proposed as an extension of the genetic alphabet from four to six letters (Piccirilli et al. Nature 343, 33(1990)). By the HF/6- 31G* method with full geometry optimization we calculated the 12 degree propeller twist for the minimum energy structure of this complex. The linearity of hydrogen bonds is preserved in the twisted structure by virtue of the pyramidal arrangement of the κ-base amino groups. The rings of both the π and κ molecules remain nearly planar. This nonplanar structure of the πκ base pair is only 0.1 kcal/mol more stable than the planar (Cs) conformation. The HF/6- 31G* level gas-phase interaction energy of πκ (—13.5 kcal/mol) calculated by us turned out to be nearly the same as the interaction energy obtained previously for the adenine-thymine base pair (—13.4 kcal/mol) at the same computational level. The inclusion of p-polarization functions on hydrogens, electron correlation effects (MP2/6–31G** level), and the correction for the basis set superposition error (BSSE) increase this energy to -14.0 kcal/mol.     

A code controlling specific binding of regulatory proteins to DNA

A possible code is suggested that describes a correspondence between amino acid sequences in stereospecific sites of regulatory proteins and nucleotide sequences at the control sites on DNA. Stereospecific sites of regulatory proteins are assumed to contain pairs of antiparallel polypeptide chain segments which form a right-hand twisted antiparallel -sheet with single-stranded regions at the ends of the -structure. The binding reaction between regulatory protein and double-helical DNA is accompanied by significant structural alterations at stereospecific sites of the protein and DNA. Half of the hydrogen bonds normally existing in -structure are broken upon complex formation with DNA and a new set of hydrogen bonds is formed between polypeptide amide groups and DNA base pairs. The code states a correspondence between four amino acid residues at a stereospecific site of the regulatory protein and an AT (GC) base pair at the control site. It predicts that there are six fundamental amino acid residues (serine, threonine, histidine, asparagine, glutamine and cysteine) whose arrangement in the stereospecific site determines the base pair sequence to which a given regulatory protein would bind preferentially.     

Synthesis of delta(E)Phe-containing tripeptide via photoisomerization and its conformation in solution

A new synthetic route to (E)-beta-phenyl-alpha,beta-dehydroalanine (delta(E)Phe)-containing peptide was presented via photochemical isomerization of the corresponding (Z)-beta-phenyl-alpha,beta-dehydroalanine (delta(Z)Phe)-containing peptide. By applying this method to Boc-Ala-delta(Z)Phe-Val-OMe (Z-I: Boc, t-butoxycarbonyl; OMe, methoxy), Boc-Ala-delta(E)Phe-Val-OMe (E-I) was obtained. The identification of peptide E-I was evidenced by 1H-nmr, 13C-nmr, and uv absorption spectroscopy, elemental analysis, and hydrogenation. The conformation of peptide E-I in CDCl3 was investigated by 1H-nmr spectroscopy (solvent dependence of NH chemical shift and difference nuclear Overhauser effect). Interestingly, peptide E-I differed from peptide Z-I in the hydrogen-bonding mode. Namely, for peptide Z-I, only Val NH participates in intramolecular hydrogen bonding, which leads to a type II beta-turn conformation supported by hydrogen bonding between CO(Boc) and NH(Val). On the other hand, for peptide E-I, two NHs, delta(E)Phe NH and Val NH, participate in intramolecular hydrogen bonding. In both peptides, a remarkable NOE (approximately 11-13%) was observed for Ala C(alpha) H-deltaPhe NH pair. Based on the nmr data and conformational energy calculation, it should be concluded that peptide E-I takes two consecutive gamma-turn conformations supported by hydrogen bonding between CO(Boc) and NH(delta(E)Phe), and between CO(Ala) and NH(Val) as its plausible conformation.     

CNDO/2 quantum-mechanical calculations of the conformational flexibility of the diketopiperazine skeleton

Eight conformers typical of diketopiperazine (DKP) ring folding were chosen for analysis. Conformational energy calculations were carried out using the semiempirical quantummechanical CNDO/2 method. The results obtained confirm considerable flexibility of the DKP skeleton. As the degree of folding increases, twisted boat conformations with the nonplanar peptide bonds tend to be more stable, while more rigid regular boat conformations with planar peptide bonds appear to be less stable than a flat one. The CNDO/2 method was found to be reliable enough for conformational studies of cyclic peptide skeletons with cis-peptide bonds.     

Structure of Horse-muscle Phosphoglycerate Kinase at 6 Å Resolution

The single peptide chain of 3-phosphoglycerate kinase is folded into two distinct globular units, only one of which seems to be involved in substrate binding.     

Influence of solvent molecules on the stereochemical code of glycyl residues in proteins

The Ramachandran steric map and energy diagrams of the glycyl residue are symmetric. A plot of (phi,psi) angles of glycyl residues in 250 nonhomologous and high-resolution protein structures is also largely symmetric. However, there is a clear aberration in the symmetry. Although there is a cluster of points corresponding to the right-handed alpha-helical region, the "equivalent" cluster is clearly shifted to in and around the (phi,psi) values of (90 degrees, 0 degrees ) instead of being centered at the left-handed alpha-helical region of (60 degrees, 40 degrees ). This lack of symmetry exists even in the (phi,psi) distribution of residues from non-alpha-helical regions in proteins. Here we provide an explanation for this observation. An analysis of glycyl conformations in small peptide structures and in "coil" proteins, which are largely devoid of helical and sheet regions, shows that glycyl residues prefer to adopt conformations around (+/-90 degrees, 0 degrees ) instead of right- and left-handed alpha-helical regions. By using theoretical calculations, such conformations are shown to have highest solvent accessibility in a system of two-linked peptide units with glycyl residue at the central C(alpha) atom. This finding is consistent with the observations from 250 nonhomologous protein structures where glycyl residues with conformations close to (+/-90 degrees, 0 degrees ) are seen to have high solvent accessibility. Analysis of a subset of nonhomologous structures with very high resolution (1.5 A or better) shows that water molecules are indeed present at distances suitable for hydrogen bond interaction with glycyl residues possessing conformations close to (+/-90 degrees, 0 degrees ). It is suggested that water molecules play a key role in determining and stabilizing these conformations of glycyl residues and explain the aberration in the symmetry of glycyl conformations in proteins.     

Molecular orbital study on the effect of the binding of water to base pair

The effects of binding water to base pairs was studied by means of the CNDO/2 molecular orbital method. The solvation energy is largest when water is bound as a proton donor and is smallest when it is stacked parallel to the plane of the base pair. The effects of two water molecules are nearly additive. The binding of one water molecule to the adenineuracil pair makes one of the two hydrogen bonds stronger and the other weaker. The change in the hydrogen bonding force is explained in terms of electrostatic and charge transfer energies. By the comparison with the adenine-cytosine pair, it is revealed that the binding of water to adenine serves to yield larger solvation energy for the complementary A-U pair than for the non-complementary A-C pair. It was also observed that the solvation energy due to the binding of water to pyrimidine was larger for A-C than for A-U.     

Molecular dynamics simulation provides a possible structure for substance P-like peptides in aqueous solution.

A hypothetical conformation of the undecapeptide Substance P in aqueous solution is generated by molecular dynamics simulation for 284 ps. The conformation takes explicit solvent interactions into account as well as entropic effects to the extent that phase space is sampled in simulation. The initial conformation is taken from energy minimization studies and modified. In spite of fluctuations through 180 degrees in some backbone dihedral angles, the peptide settles with all backbone dihedrals within +/- 60 degrees from the initial values. In 130 ps, the radius of gyration decreases from 6.2 A to 5.5 A, whereas only fluctuation (+/- .2 A) is observed during the last 150 ps. The root-mean-square deviation at optimal superposition for a pair of conformations from the last 150 ps is 0.6 A, based on backbone atoms. The final structure is close-knit, nearly globular, and stabilized by several long-lived hydrogen bonds. The simulation conformation agrees with the scarce experimental data including a large number of structure-activity relationships. Thus, the simulation conformation is a likely candidate for one of the several conformations, the existence of which has been deduced from nuclear magnetic resonance data. Simulation results and experimental modification studies suggest that Phe 8 and Leu 10 are involved in the primary binding of SP to its receptors.     

Mechanism-based protein design: attempted "nucleation-condensation" approach to a possible minimal helix-bundle protein

In an intended mechanism-based de novo approach, a 22-mer peptide was so designed as to make it both a stereochemically nucleatable and hydrophobically condensable minimal globular protein. Framework-like nucleation of a triple-helix bundle was targeted by employing as folding nucleators composite beta-turns that could both nucleate helices and place them in close juxtaposition for possible interhelical interaction. To promote the targeted triple-helix bundle to condense as a globular protein, an amphipathic sequence pattern was adopted for possible hydrophobic interhelical interaction. A predominantly helicogenic 22-mer amphipathic peptide was thus designed, punctuating it with composite type II'-III and type II-Asx type beta-turns as the helix nucleators cum chain reversal elements. The peptide made by solid-phase synthesis was shown by NMR and CD to be a nascent and distorted triple-helix bundle in a trifluoroethanol (TFE)-water mixture, but more or less a random coil in water. A fold nucleation effect is evident in the TFE-water mixture, but apparently the hydrophobic effect cannot sustain the peptide conformational order in water. A lack of synergy between folding nucleation and hydrophobic condensation of the peptide is possible. Indeed, a mismatch between the sequential H,P pattern of the peptide and its nascent-type globular fold in a TFE-water mixture is evident based on a simulated annealing study guided by NMR.     

Molecular dynamics simulation of the nanofibrils formed by amyloid-based peptide amphiphiles

Abstract

Atomistic molecular dynamics simulations have been performed on the peptide amphiphiles (PAs) with four amyloid beta peptide fragments as head groups. The stable structures were monitored by the root mean square deviation with respect to the energy minimised initial structures. Random coil and β-sheet structures with hydrogen bonds along and perpendicular to the long axis of the nanofibre were obtained due to the different nature of the head groups. Influences of pH and capping ends on the nanofibre structures were investigated through variation of the protonation states of the ionic amino acids in the peptides. The peptides with opposite charges on both sides were found to have the fewest β-sheet structures, and the charges on the outer terminal tended to destruct the β-sheets while those at the inner side did not. The isolated charge in the centre of peptides was found to be able to promote the formation of regular β-sheets, while multiple charged residues could not support ordered β-sheet structures. When charge neutralisation occurred between adjacent residues, regular β-sheet laminates might also occur for systems with charges at the outer terminal. With the increase of β-sheet structures formed, the original twisted structures found for random coil structures of the PAs could be diminished by the hydrogen bonds.     

Effects of net charge on the hydrogen-deuterium exchange parameters of lysozyme.

Possible effects of changes in net charge on protein hydrogen exchange rates were investigated by desalting hen egg-white lysozyme, which allowed its net charge to increase with decreasing pH in the acid region. Chloride ion-binding ratios, expressed as ratios of free to total Cl^?, were measured with a chloride-specific electrode at pH 5 on a 2.4% solution of a five-time-desalted product. This ratio was used to show a 97% reduction of the 11% Cl^? present in a commercial lysozyme preparation upon three passes of the enzyme through a column of ion-retardation resin. Net charges on the purified product were assigned from a combination of electrophoretic mobility and proton titration data gathered under minimal ionic strength conditions. The net charge on the desalted product increased by 1.64 units between pH 5.0 and 3.0. Hydrogendeuterium exchange studies on the purified lysozyme in D₂O were obtained using the near-infrared region of a Cary 14R spectrophotometer. The rate-pD profile for k₂, the rate constant for the intermediate class of exchanging hydrogens, showed a decrease in the apparent pD of minimum exchange rate of 0.3 units, when compared to that obtained earlier in 0.2 m added NaCl. However, the rate of exchange at pD minimum and the number of hydrogens in the class remained largely unaffected. A similar shift was observed for the rate-pD profile of the class 1 hydrogens. Thus, the effect of an increase in net positive charge is to shift the rate-pD profile to a lower pD. Moreover, the effect extended to the interior peptide hydrogens of this globular protein. Consequently, the exchange rates of all the observable hydrogens are altered by the net charge changes, and the effect appeared uniform. The shift can be accounted for quantitatively by applying electrostatic interaction terms to the acid and base catalytic constants characterizing the exchange process. The calculated electrostatic interaction factors in minimal salt and 0.2 m added NaCl were found to be 29 and 18% lower, respectively, than those obtained theoretically. Therefore, under conditions where changes in net charge may occur for a globular protein, the effect on hydrogen exchange rates can be estimated fairly well theoretically, especially at moderate ionic strengths.     

The pseudomolecule method and the structure of globular proteins. II. The example of ribonuclease F1 and T1

The pseudomolecule approach to the structure of globular proteins in which a small number of water molecules are incorporated into the "molecule" is tested again by comparing the ribbon of hydrogen bonds in two proteins, ribonuclease F1 and T1. These two molecules are 59% homologous and have the same backbone conformation both globally and locally. The two ribbons of hydrogen bonds that cover the whole of the backbone are conserved with an accuracy of some 95% providing that allowance is made for the intrusion into one of the pair of such extra factors as the presence of adducts or metal ions, the insertions and the absence of a few water molecules from one of the x-ray data sets. Without these corrections, the conservation of the ribbon is some 85%. There are 35 conserved hydrogen-bonding residues, nearly all of which show many unions to the backbone or interactions with the active site. There are 36 point mutations that involve one or two hydrogen-bonding side chains and nearly all of these have either none or one hydrogen bond to the backbone. These are minor contributors to the ribbon of hydrogen bonds. Of the 71 residues involved in these two categories, all but six fit into the pseudomolecular picture of the structure of globular proteins. The remaining 30 residues almost all contain conserved hydrocarbon side chains that may have a second order effect on the structure through their space filling effects.     

Localization of hydrogen-bonds within modules in barnase

Proteins in eukaryotes are composed of structural units, each encoded by discrete exons. The protein module is one such structural unit; it has been defined as the least extended or the most compact contiguous segment in a globular domain. To elucidate roles of modules in protein evolution and folding, we examined roles of hydrogen bonds and hydrophobic cores, as related to the stability of these modules. For this purpose we studied barnase, a bacterial Rnase from Bacillus amylolique-faciens. Barnase is decomposed into at least six modules, M1–M6; the module boundaries are identified at amino acid residues 24, 52, 73, 88, and 98. Hydrogen bonds are localized mainly within each of the modules, with only a few between them, thereby indicating that their locations are designed to primarily stabilize each individual module. To obtain support for this notion, an analysis was made of hypothetical modules defined as segments starting at a center of one module and ending at the center of the following one. We found that the hydrogen bonds did not localize in each hypothetical module and that many formed between the hypothetical modules. The native conformations of modules of barnase may be specified predominantly by interactions within the modules. © 1993 Wiley-Liss, Inc.     

Electrostatic interactions in proteins: Calculations of the electrostatic term of free energy and the electrostatic potential field

The pH-dependence of the electrostatic energy of interactions between titratable groups is calculated for some well studied globular proteins: basic pancreatic trypsin inhibitor, sperm whale myoglobin and tuna cytochrome c. The calculations are carried out using a semi-empirical appraach in terms of the macroscopic model based on the Kirkwood-Tanford theory. The results are discussed in the light of their physicochemical and biological properties. It was found that the pH-dependence of the electrostatic energy correlates with the III–IV transition of cytochrome c. The electrostatic field of the cysteine proteinase inhibitor, cystatin, was calculated in two ways. In the first one, the electrostatic field created by the pH dependent charges of the ionizable groups and peptide dipoles was calculated using the approach proposed. In the second one, the finite-difference method was used. The results obtained by the two methods are in overall agreement. The calculated field was discussed in terms of the binding of cystatin to papain.     

Interaction of monoclonal anti-peptide antibodies with lysozyme

The interaction of monoclonal anti-peptide antibodies with the free peptide and its protein counterpart has been evaluated for hen egg white lysozyme and the peptide constituting residues 38 to 45. Fluorescence methodology has been developed for the measurement of association constants based on resonance energy transfer between the excited tryptophan of antibody and bound peptide ligand conjugated to a fluorescent probe. Five antibodies, four IgM and one IgG, have been assayed by ELISA, and have demonstrated binding to the adsorbed peptide alone, to the adsorbed lysozyme alone, or to both. Multivalent interaction with the adsorbed ligand is a key factor in the efficacy of binding. Measurement of binding constants in homogeneous solution, by equilibrium dialysis and energy transfer, demonstrated that lysozyme was bound to an IgG antipeptide antibody with an association constant (4 X 10(2) M-1) 200-fold less than that for the free peptide (8 X 10(4) M-1). It was also inferred for IgM that an association constant of the order of 10(2) M-1 was sufficient to effect selective interaction in a system providing multivalent interaction. The shared conformations between protein and peptide, implied by the specific reactivity of the anti-peptide antibody with the protein, points to structural fluctuations of the surface regions and residues of globular proteins.