Reflections on biocatalysis involving phosphorus.

Early studies on chemical synthesis of biological molecules can be seen to progress to preparation and biological evaluation of phosphonates as analogues of biological phosphates, with emphasis on their isosteric and isopolar character. Work with such mimics progressed into structural studies with a range of nucleotide-utilising enzymes. The arrival of metal fluorides as analogues of the phosphoryl group, PO(3)(-), for transition state (TS) analysis of enzyme reactions stimulated the symbiotic deployment of (19)F NMR and protein crystallography. Characteristics of enzyme transition state analogues are reviewed for a range of reactions. From the available MF(x) species, trifluoroberyllate gives tetrahedral mimics of ground states (GS) in which phosphate is linked to carboxylate and phosphate oxyanions. Tetrafluoroaluminate is widely employed as a TS mimic, but it necessarily imposes octahedral geometry on the assembled complexes, whereas phosphoryl transfer involves trigonal bipyramidal (tbp) geometry. Trifluoromagnesate (MgF(3)(-)) provides the near-ideal solution, delivering tbp geometry and correct anionic charge. Some of the forty reported tbp structures assigned as having AlF(3)(0) cores have been redefined as trifluoromagnesate complexes. Transition state analogues for a range of kinases, mutases, and phosphatases provide a detailed description of mechanism for phosphoryl group transfer, supporting the concept of charge balance in their TS and of concerted-associative pathways for biocatalysis. Above all, superposition of GS and TS structures reveals that in associative phosphoryl transfer, the phosphorus atom migrates through a triangle of three, near-stationary, equatorial oxygens. The extension of these studies to near attack conformers further illuminates enzyme catalysis of phosphoryl transfer.

Reversible conversion of monomeric human prion protein between native and fibrilogenic conformations.

Prion propagation involves the conversion of cellular prion protein (PrPC) into a disease-specific isomer, PrPSc, shifting from a predominantly alpha-helical to beta-sheet structure. Here, conditions were established in which recombinant human PrP could switch between the native alpha conformation, characteristic of PrPC, and a compact, highly soluble, monomeric form rich in beta structure. The soluble beta form (beta-PrP) exhibited partial resistance to proteinase K digestion, characteristic of PrPSc, and was a direct precursor of fibrillar structures closely similar to those isolated from diseased brains. The conversion of PrPC to beta-PrP in suitable cellular compartments, and its subsequent stabilization by intermolecular association, provide a molecular mechanism for prion propagation.

Determinants of DNA-binding specificity of ETS-domain transcription factors.

Several mechanisms are employed by members of transcription factor families to achieve sequence-specific DNA recognition. In this study, we have investigated how members of the ETS-domain transcription factor family achieve such specificity. We have used the ternary complex factor (TCF) subfamily as an example. ERK2 mitogen-activated protein kinase stimulates serum response factor-dependent and autonomous DNA binding by the TCFs Elk-1 and SAP-la. Phosphorylated Elk-1 and SAP-la exhibit specificities of DNA binding similar to those of their isolated ETS domains. The ETS domains of Elk-1 and SAP-la and SAP-2 exhibit related but distinct DNA-binding specificities. A single residue, D-69 (Elk-1) or V-68 (SAP-1), has been identified as the critical determinant for the differential binding specificities of Elk-1 and SAP-1a, and an additional residue, D-38 (Elk-1) or Q-37 (SAP-1), further modulates their DNA binding. Creation of mutations D38Q and D69V is sufficient to confer SAP-la DNA-binding specificity upon Elk-1 and thereby allow it to bind to a greater spectrum of sites. Molecular modelling indicates that these two residues (D-38 and D-69) are located away from the DNA-binding interface of Elk-1. Our data suggest a mechanism in which these residues modulate DNA binding by influencing the interaction of other residues with DNA.

Multiple folding pathways for heterologously expressed human prion protein.

Human PrP (residues 91-231) expressed in Escherichia coli can adopt several conformations in solution depending on pH, redox conditions and denaturant concentration. Oxidised PrP at neutral pH, with the disulphide bond intact, is a soluble monomer which contains 47% alpha-helix and corresponds to PrPC. Denaturation studies show that this structure has a relatively small, solvent-excluded core and unfolds to an unstructured state in a single, co-operative transition with a DeltaG for folding of -5.6 kcal mol-1. The unfolding behaviour is sensitive to pH and at 4.0 or below the molecule unfolds via a stable folding intermediate. This equilibrium intermediate has a reduced helical content and aggregates over several hours. When the disulphide bond is reduced the protein adopts different conformations depending upon pH. At neutral pH or above, the reduced protein has an alpha-helical fold, which is identical to that observed for the oxidised protein. At pH 4 or below, the conformation rearranges to a fold that contains a high proportion of beta-sheet structure. In the reduced state the alpha- and beta-forms are slowly inter-convertible whereas when oxidised the protein can only adopt an alpha-conformation in free solution. The data we present here shows that the human prion protein can exist in multiple conformations some of which are known to be capable of forming fibrils. The precise conformation that human PrP adopts and the pathways for unfolding are dependent upon solvent conditions. The conditions we examined are within the range that a protein may encounter in sub-cellular compartments and may have implications for the mechanism of conversion of PrPC to PrPSc in vivo. Since the conversion of PrPC to PrPSc is accompanied by a switch in secondary structure from alpha to beta, this system provides a useful model for studying major structural rearrangements in the prion protein.

Folding of peptide fragments comprising the complete sequence of proteins. Models for initiation of protein folding. I. Myohemerythrin.

In an attempt to delineate potential folding initiation sites for different protein structural motifs, we have synthesized series of peptides that span the entire length of the polypeptide chain of two proteins, and examined their conformational preferences in aqueous solution using proton nuclear magnetic resonance and circular dichroism spectroscopy. We describe here the behavior of peptides derived from a simple four-helix bundle protein, myohemerythrin. The peptides correspond to the sequences of the four long helices (the A, B, C and D helices), the N- and C-terminal loops and the connecting sequences between the helices. The peptides corresponding to the helices of the folded protein all exhibit preferences for helix-like conformations in solution. The conformational ensembles of the A- and D-helix peptides contain ordered helical forms, as shown by extensive series of medium-range nuclear Overhauser effect connectivities, while the B- and C-helix peptides exhibit conformational preferences for nascent helix. All four peptides adopt ordered helical conformations in mixtures of trifluoroethanol and water. The terminal and interconnecting loop peptides also appear to contain appreciable populations of conformers with backbone phi and psi angles in the alpha-region and include highly populated hydrophobic cluster and/or turn conformations in some cases. Trifluoroethanol is unable to drive these peptides towards helical conformations. Overall, the peptide fragments of myohemerythrin have a marked preference towards secondary structure formation in aqueous solution. In contrast, peptide fragments derived from the beta-sandwich protein plastocyanin are relatively devoid of secondary structure in aqueous solution (see accompanying paper). These results suggest that the two different protein structural motifs may require different propensities for formation of local elements of secondary structure to initiate folding, and that there is a prepartitioning of conformational space determined by the local amino acid sequence that is different for the helical and beta-sandwich structural motifs.

Characterisation of low free-energy excited states of folded proteins.

It is demonstrated that the identity of residues accessing excited conformational states that are of low free energy relative to the ground state in proteins can be obtained from amide proton NMR chemical shift temperature dependences displaying significant curvature. For the N-terminal domain of phosphoglycerate kinase, hen egg-white lysozyme and BPTI, conformational heterogeneity arises from a number of independent sources, including: structural instability resulting from deletion of part of the protein; a minor conformer generated through disulphide bond isomerisation; an alternative hydrogen bond network associated with buried water molecules; alternative hydrogen bonds involving backbone amides and surface-exposed side-chain hydrogen bond acceptors; and the disruption of loops, ends of secondary structural elements and chain termini. In many of these cases, the conformational heterogeneity at these sites has previously been identified by X-ray and/or NMR studies, but conformational heterogeneity of buried water molecules has hitherto received little attention. These multiple independent low free-energy excited states each involve a small number of residues and are shown to be within 2.5 kcal mol-1 of the ground state. Their relationship with the partially unfolded forms previously characterised using amide proton exchange studies is discussed.

The three-dimensional solution structure of human stefin A.

The three-dimensional solution structure of recombinant human stefin A has been determined by a simulated annealing protocol using a total of 1113 distance and angle constraints obtained from 1H and 15N HMR spectroscopy. The solution structure is represented by a family of 17 conformers with an average root-mean-square deviation relative to the mean structure of 0.44 A for backbone atoms and 0.94 A for all heavy atoms for the main body of the structure. The protein has a well-defined global fold consisting of five anti-parallel beta-strands wrapped around a central five-turn alpha-helix. There is considerable similarity between the structural features of free stefin A in solution and the X-ray structure of the homologous protein stefin B in its complex with papain, but there are also some important differences in the regions which are fundamental to proteinase binding. The differences consist primarily of two regions of high conformational heterogeneity in free stefin A which correspond in stefin B to two of the components of the tripartite wedge that docks into the active site of the target proteinase. These regions, which are shown to be mobile in solution, are the five N-terminal residues and the second binding loop. In the bound conformation of stefin B they form a turn and a short helix, respectively.

Mechanism-based inhibition of C5-cytosine DNA methyltransferases by 2-H pyrimidinone.

DNA duplexes in which the target cytosine base is replaced by 2-H pyrimidinone have previously been shown to bind with a significantly greater affinity to C5-cytosine DNA methyltransferases than unmodified DNA. Here, it is shown that 2-H pyrimidinone, when incorporated into DNA duplexes containing the recognition sites for M.HgaI-2 and M.MspI, elicits the formation of inhibitory covalent nucleoprotein complexes. We have found that although covalent complexes are formed between 2-H pyrimidinone-modified DNA and both M.HgaI-2 and M.MspI, the kinetics of complex formation are quite distinct in each case. Moreover, the formation of a covalent complex is still observed between 2-H pyrimidinone DNA and M.MspI in which the active-site cysteine residue is replaced by serine or threonine. Covalent complex formation between M.MspI and 2-H pyrimidinone occurs as a direct result of nucleophilic attack by the residue at the catalytic position, which is enhanced by the absence of the 4-amino function in the base. The substitution of the catalytic cysteine residue by tyrosine or chemical modification of the wild-type enzyme with N-ethylmaleimide, abolishes covalent interaction. Nevertheless the 2-H pyrimidinone-substituted duplex still binds to M.MspI with a greater affinity than a standard cognate duplex, since the 2-H pyrimidinone base is mis-paired with guanine.

Protein folding and intermediates.

With the exception of the discovery of the rate of formation of the earliest intermediates, there have been no major conceptual leaps in our understanding of protein folding reactions over the past two years. Rather, this period has seen an extension of two established techniques: first, mutational analysis combined with a kinetic definition of the energy landscape of the reaction; and second, the use of hydrogen/deuterium exchange of backbone amide groups combined with NMR. Owing to the application of these methods to a wider range of proteins, it is now possible to draw some general conclusions about the physical processes that direct a protein to its native fold.

Conformation of a T cell stimulating peptide in aqueous solution.

Using two-dimensional NMR spectroscopy and circular dichroism spectroscopy it is demonstrated that a T cell stimulating peptide corresponding to residues 132-153 of sperm whale myoglobin populates helical conformations in aqueous solution. This finding is in accordance with proposals that immunodominant sites in T cell stimulating peptides have a high conformational propensity. The observation of secondary structure in aqueous solutions of this and other immunogenic peptides has important implications for initiation of protein folding.

The natural design of vancomycin family antibiotics to bind their target peptides.

The vancomycin family of antibiotics provide a rare opportunity among natural systems to study a molecular recognition process in which both the 'receptor' and the 'ligand' are relatively small molecules. Unlike the vast majority of antibiotics, in the vancomycin family the antibiotic performs the role of the receptor. All members of the family are covalently cross-linked heptapeptides that contain a variety of glycosidic modifications. Their site of action in bacterial cell walls is modelled by simple dipeptides and tripeptides. NMR experiments have been used to characterize the binding of these species through the study of both the complex and the free components. In unbound antibiotics conformational freedom is observed in regions of the molecule not severely restricted by covalent linkages. On binding of the ligand much of this conformational freedom is lost and the hydrophobic side chains of the antibiotics reside close to the intermolecular hydrogen-bonding interactions, thus shielding these interactions from the solvent. The charged amino groups of the N-terminus and disaccharide region of vancomycin are orientated not to optimize intermolecular electrostatic interactions but rather to retain solvation. This causes further hydrophobic faces to be presented to the ligand. Removal of saccharide units from the antibiotics leads to small losses in binding energy but may have considerable influence on the selectivity of the antibiotics. Specific dimerization through the non-ligand-binding faces of ristocetin is observed at millimolar concentrations. The geometry of the dimeric complex enables a close approach of the ligand carboxylate anion and the charged amino group of the novel sugar, ristosamine.

Amyloid fibril formation by human stefin B: influence of the initial pH-induced intermediate state.

The amyloid fibril field is briefly described, with some stress put on differences between various proteins and possible role for domain swapping. In the main body of the text, first, a short review is given of the folding properties of both human stefins, alpha/beta-type globular proteins of 53% identity with a known three-dimensional fold. Second, in vitro study of amyloid fibril formation by human stefin B (type I cystatin) is described. Solvents of pH 4.8 and pH 3.3 with and without 2,2,2-trifluoroethanol (TFE) were probed, as it has been shown previously that stefin B forms acid intermediates, a native-like and molten globule intermediate, respectively. The kinetics of fibrillation were measured by thioflavin T fluorescence and CD. At pH 3.3, the protein is initially in the molten globule state. The fibrillation is faster than at pH 4.8; however, there is more aggregation observed. On adding TFE at each pH, the fibril formation is further accelerated.

Peptide models of protein folding initiation sites. 1. Secondary structure formation by peptides corresponding to the G- and H-helices of myoglobin.

Myoglobin has been extensively studied as a model system for protein folding in vitro. As part of an ongoing study of myoglobin folding, we have synthesized a series of peptide fragments corresponding to portions of the sequence of the sperm whale protein. The conformational preferences of these peptides have been investigated by circular dichroism and nuclear magnetic resonance spectroscopy in aqueous solution. In this paper we describe the folding propensities of two peptides (Mb-G and Mb-H), corresponding to the G- and H-helix segments of the myoglobin sequence. The Mb-G peptide shows evidence of a very small population of helical conformations in aqueous solution, both by CD and NMR. By contrast, the monomeric Mb-H peptide is found by CD to adopt a significant population (ca. 30%) of ordered helix and by NMR to populate helical conformations in rapid dynamic equilibrium with unfolded states. The Mb-H peptide undergoes a well-characterized, concentration-dependent monomer-tetramer equilibrium. At peptide concentrations greater than 1 mM there is an increase in the population of helix, to approximately 85% according to the CD spectrum, through self-association to form a tetramer. Both medium-range NOE connectivities and a CD spectrum characteristic of ordered helix are observed at low peptide concentrations, establishing that helical conformations are present in the monomeric state of Mb-H. The relative helicity at various sites throughout the Mb-H peptide has been estimated using a novel method for assessing the distribution of helical populations based on the relative magnitudes of medium-range d alpha beta (i,i+3) NOE connectivities. The population of ordered helix is seen to be highest in the center of the peptide sequence; the ends of the peptide show evidence of pronounced fraying.

Peptide models of protein folding initiation sites. 2. The G-H turn region of myoglobin acts as a helix stop signal.

A series of peptide fragments of sperm whale myoglobin, corresponding to segments of the region between the G- and H-helices of the protein, have been synthesized and their conformational preferences investigated using circular dichroism and nuclear magnetic resonance spectroscopy in aqueous solution and in solvent mixtures containing water and trifluoroethanol. The smallest fragment, Mb-GH5, a five-residue peptide with the sequence HPGDF corresponding to the connecting loop between the two helices in the folded protein, adopts highly populated turn conformations in aqueous solution. A 25-residue peptide, Mb-GH25, containing the same sequence flanked by contiguous segments of the G- and H-helix sequences, was also found to contain a high proportion of conformers with a turn in this region. No helix formation was observed in the flanking sequences in water solution, either in Mb-GH25 or in control 10-residue peptides (Mb-G10 and Mb-H10) with sequences corresponding to the G- and H-helix segments. No additional helicity above that of the sum of the components was observed for Mb-GH25, indicating that a helical hairpin structure is not formed in the monomeric peptide in aqueous solution. In the presence of TFE, ordered helix is formed in Mb-GH25 according to the CD spectrum, and NMR spectra indicate that this is localized in the N-terminal portion of the peptide. NOESY spectra clearly show that the turn conformation is retained under these conditions.(ABSTRACT TRUNCATED AT 250 WORDS)

Characterization of the Elk-1 ETS DNA-binding domain.

The ETS domain family of transcription factors is comprised of several important proteins that are involved in controlling key cellular events such as proliferation, differentiation, and development. One such protein, Elk-1, regulates the activity of the c-fos promoter in response to extracellular stimuli. Elk-1 is representative of a subgroup of ETS domain proteins that utilize a bipartite recognition mechanism that is mediated by both protein-DNA and protein-protein interactions. In this study, we have overexpressed, purified, and characterized the ETS DNA-binding domain of Elk-1 (Elk-93). Elk-93 was expressed in Escherichia coli as a fusion protein with glutathione S-transferase and purified to homogeneity from both the soluble and insoluble fractions using a two-column protocol. A combination of CD, NMR, and fluorescence spectroscopy demonstrates that Elk-93 represents an independently folded domain of mixed alpha/beta structure in which the three conserved tryptophans appear to contribute to the hydrophobic core of the protein. Moreover, DNA binding studies demonstrate that Elk-93 binds DNA with both high affinity (Kd approximately 0.85 x 10(-10)M) and specificity. Circular permutation analysis indicates that DNA binding by Elk-93 does not induce significant bending of the DNA. Our results are discussed with respect to predictive models for the structure of the ETS DNA-binding domain.

Forces in molecular recognition: comparison of experimental data and molecular mechanics calculations.

NMR studies of the rotation barrier of the disaccharide of the glycopeptide antibiotic vancomycin have been used to test the performance of computer simulation techniques using molecular mechanics. In the absence of any solvated water, no correlation could be found between experiment and calculation. By introducing solvent water molecules into the binding region of the antibiotic, the NMR results could be simulated both qualitatively and quantitatively within experimental error without using massive computational resources.

Calcium-induced structural changes and domain autonomy in calmodulin.

We have determined the solution structures of the apo and (Ca2+)2 forms of the carboxy-terminal domain of calmodulin using multidimensional heteronuclear nuclear magnetic resonance spectroscopy. The results show that both forms adopt well-defined structures with essentially equal secondary structure. A comparison of the structures of the two forms shows that Ca2+ binding causes major rearrangements of the secondary structure elements with changes in inter-residue distances of up to 15 A and exposure of the hydrophobic interior of the four-helix bundle. Comparisons with previously determined high-resolution X-ray structures and models of calmodulin indicate that this domain is structurally autonomous.

Differences in the effects of TFE on the folding pathways of human stefins A and B.

Trifluoroethanol (TFE) has been used to probe differences in the stability of the native state and in the folding pathways of the homologous cysteine protein inhibitors, human stefin A and B. After complete unfolding in 4.5 mol/L GuHCl, stefin A refolded in 11% (vol/vol) TFE, 0.75 mol/L GuHCl, at pH 6.0 and 20 degrees C, with almost identical first-order rate constants of 4.1 s-1 and 5.5 s-1 for acquisition of the CD signal at 230 and 280 nm, respectively, rates that were markedly greater than the value of 0.11 s-1 observed by the same two probes when TFE was absent. The acceleration of the rates of refolding, monitored by tyrosine fluorescence, was maximal at 10% (vol/vol) TFE. Similar rates of refolding (6.2s-1 and 7.2 s-1 for ellipticity at 230 and 280 nm, respectively) were observed for stefin A denatured in 66% (vol/vol) TFE, pH 3.3, when refolding to the same final conditions. After complete unfolding in 3.45 mol/L GuHCl, stefin B refolded in 7% (vol/vol) TFE, 0.57 mol/L GuHCl, at pH 6.0 and 20 degrees C, with a rate constant for the change in ellipticity at 280 nm of 32.8 s-1; this rate was only twice that observed when TFE was absent. As a major point of distinction from stefin A, the refolding of stefin B in the presence of TFE showed an overshoot in the ellipticity at 230 nm to a value 10% greater than that in the native protein; this signal relaxed slowly (0.01 s-1) to the final native value, with little concomitant change in the near-ultraviolet CD signal; the majority of this changes in two faster phases. After denaturation in 42% (vol/vol) TFE, pH 3.3, the kinetics of refolding to the same final conditions exhibited the same rate-limiting step (0.01 s-1) but were faster initially. The results show that similarly to stefin A, stefin B forms its hydrophobic core and predominant part of the tertiary structure faster in the presence of TFE. The results imply that the alpha-helical intermediate of stefin B is highly structured. Proteins 1999;36:205-216.