Introduction: Science and connoisseurship in the European Enlightenment

A major theme of the European Enlightenment was the rationalization of value, the use of reason to determine the value of things, from diamonds to civilizations. This view of the Enlightenment is well-established in the human sciences. It is ripe for extension to the natural sciences, given the rich recent literature on affect, evaluation, and subjectivity in early modern science. Meanwhile, in art history, the new history of connoisseurship provides a model for the historical study of the evaluation of material things. Historians of natural history have already noted the connections between science, Enlightenment, and connoisseurship. The time has come to extend their insights to other areas of Enlightenment science. This means recognizing the breadth of connoisseurship - the social, linguistic, and disciplinary diversity of the practice - as understood in Europe in the eighteenth century and the latter part of the seventeenth century. An outline of the three papers in this special section gives an indication of how this historiographical project might be carried out.

The hand of the connoisseur: Gems and hardness in Enlightenment mineralogy

Historians of natural history have shown that the study of plants, animals, and minerals was a form of connoisseurship in the eighteenth century. Historians of early modern experiments have linked scientific knowledge to the manual skills of artisans. I combine these two insights, arguing that connoisseurship in the sciences meant learning to touch, not just learning to look. The focus is on gems and mineralogy in eighteenth-century France. I show, firstly, that the study of gems was linked to the connoisseurship ("connoissance") of paintings. Next, books on gems were closely related to the new mineralogical treatises that emerged in the middle of the eighteenth century. These treatises formalized a distinction between "Oriental" and "Occidental" gems that was also a distinction between hard and soft gems. The best judges of hardness were gem cutters, a group that participated in mineralogy through the culture of collecting. Finally, the knowledge of cutters contributed to the quantification of hardness in the form of the hardness scale and the scratch sclerometer.

Iatrochemistry and the Evaluation of Mineral Waters in France, 1600-1750

Existing literature on mineral springs in early modern France suggests that composition played a minor role in the evaluation of those springs. In fact it played a major role from at least the beginning of the seventeenth century. Composition was studied by a wide range of actors, from physicians in the provinces to chemists at the Paris Academy of Science, with a view to establishing the efficacy of particular springs against particular diseases. Iatrochemistry played a complex role in these evaluations. Followers of Paracelsus and van Helmont were among the first to perform chemical analyses on mineral waters. But there were physicians who studied composition without chemistry, or who used chemistry while opposing iatrochemistry. Conversely, there were iatrochemists who used chemistry to study mineral waters but not to evaluate them, and there were many chemists who gave at least as much weight to clinical experience as they did to composition.

The trials of theory: psychology and institutionalist economics, 1910-1931

The rise of the institutionalist school of economics, in the 1910s and 1920s, has recently been given the historical attention it deserves. However, historical studies of the school have left two questions unanswered. First, to what extent was the institutionalist's interest in academic psychology (frequently declared in their meta-economic writings) realized in their economic writings? Second, what evidence of a fruitful collaboration with institutional economics can be found in the work of psychologists? In this paper I consider the meta-economic statements of three key institutionalists, Wesley C. Mitchell, John M. Clark, and Walton H. Hamilton, and two key economic works by Mitchell and Clark. I contend that these works show little systematic engagement of academic psychology. A study of psychological literature of the period yields the same conclusion; in particular, industrial psychology did not come into fruitful contact with institutional economics, despite the parallel interests of the two fields and their parallel rise after World War I.

Regulation by phosphorylation of the relative affinities of the N-terminal transactivation domains of p53 for p300 domains and Mdm2

The transcriptional activity of the tumor suppressor p53 requires direct binding between its transactivation domain (TAD, 1-57) and the transcriptional coactivator p300. We systematically assessed the role of TAD phosphorylation on binding of the p300 domains CH3, Taz1, Kix and IBiD. Thr18 phosphorylation increased the affinity up to 7-fold for CH3 and Taz1, with smaller increases from phosphorylation of Ser20, Ser15, Ser37, Ser33, Ser46 and Thr55. Binding of Kix and IBiD was less sensitive to phosphorylation. Strikingly, hepta-phosphorylation of all Ser and Thr residues increased binding 40- and 80-fold with CH3 and Taz1, respectively, but not with Kix or Ibid. Substitution of all phospho-sites with aspartates partially mimicked the effects of hepta-phosphorylation. Mdm2, the main negative regulator of p53, competes with p300 for binding TAD. Binding of Mdm2 to TAD was reduced significantly only on phosphorylation of Thr18 (7-fold) or by hepta-phosphorylation (24-fold). The relative affinities of Mdm2 and p300 for p53 TAD can thus be changed by up to three orders of magnitude by phosphorylation. Accordingly, phosphorylation of Thr18 and hepta-phosphorylation dramatically shifts the balance to favouring binding of p300 with p53 and is thus likely to be an important factor in its regulation.

Application of thermophilic enzymes in commercial biotransformation processes

Biocatalysis is a useful tool in the provision of chiral technology and extremophilic enzymes are just one component in that toolbox. Their role is not always attributable to their extremophilic properties; as with any biocatalyst certain other criteria should be satisfied. Those requirements for a useful biocatalyst will be discussed including issues of selectivity, volume efficiency, security of supply, technology integration, intellectual property and regulatory compliance. Here we discuss the discovery and commercialization of an l-aminoacylase from Thermococcus litoralis, the product of a LINK project between Chirotech Technology and the University of Exeter. The enzyme was cloned into Escherichia coli to aid production via established mesophilic fermentation protocols. A simple downstream process was then developed to assist in the production of the enzyme as a genetically modified-organism-free reagent. The fermentation and downstream processes are operated at the 500 litre scale. Characterization of the enzyme demonstrated a substrate preference for N-benzoyl groups over N-acetyl groups. The operational parameters have been defined in part by substrate-concentration tolerances and also thermostabilty. Several examples of commercial biotransformations will be discussed including a process that is successful by virtue of the enzyme's thermotolerance.

The ligand-binding loops in the tunicate C-type lectin TC14 are rigid

C-type lectin-like domains are very common components of extracellular proteins in animals. They bind to a variety of ligands, including carbohydrates, proteins, ice, and CaCO3 crystals. Their structure is characterized by long surface loops in the area of the protein usually involved in ligand binding. The C-type lectin TC14 from Polyandrocarpa misakiensis specifically binds to D-galactose by coordination of the sugar to a bound calcium atom. We have studied the dynamic properties of TC14 by measuring 15N longitudinal and transverse relaxation rates as well as [1H-15N] heteronuclear NOEs. Relaxation rates and heteronuclear NOE data for holo-TC14 show minimal variations, indicating that there is no substantial difference in rigidity between the elements of regular secondary structure and the extended surface loops. Anisotropic tumbling of the elongated TC14 dimer can account for the main fluctuations in relaxation rates. Loss of the bound calcium does not significantly alter the internal dynamics, suggesting that the stability of the loop region is intrinsic and not dependent on the coordination of the calcium ion. Chemical shift differences between the holo and apo form show that main structural changes occur in the calcium-binding site, but smaller structural changes are propagated throughout the molecule without affecting the overall fold. The disappearance of two resonances for residues following the conserved cis-proline 87 (which is located in the calcium-binding site) in the apo form indicates conformational change on an NMR time scale between the cis and trans configurations of this peptide bond in the absence of calcium. Possible implications of these findings for the ligand binding in C-type lectin-like domains are discussed.

Crystallization and preliminary crystallographic studies of a ribosomal protein L30e from the hyperthermophilic archaeonThermococcus celer

Ribosomal protein L30e from the hyperthermophilic archaeon Thermococcus celer is a good model for the study of the thermostability of proteins. It has been overexpressed, purified and crystallized using the hanging-drop vapour-diffusion method using PEG 8000 as precipitant at 290 K. The crystal belongs to the hexagonal space group P6(1)/P6(5), with unit-cell parameters a = b = 48.32, c = 86.42 A. The asymmetric unit contains a single molecule of L30e, with a corresponding crystal volume per protein mass (V(M)) of 2.68 A(3) Da(-1) and a solvent content of 54%. A complete data set diffracting to 1.96 A resolution was collected from a single crystal at 100 K.

Structure of the C-terminal sterile alpha-motif (SAM) domain of human p73 alpha

p73 is a homologue of the tumour suppressor p53 and contains all three functional domains of p53. The alpha-splice variant of p73 (p73 alpha) contains near its C-terminus an additional structural domain known as the sterile alpha-motif (SAM) that is probably responsible for regulating p53-like functions of p73. Here, the 2.54 A resolution crystal structure of this protein domain is reported. The crystal structure and the published solution structure have the same five-helix bundle fold that is characteristic of all SAM-domain structures, with an overall r.m.s.d. of 1.5 A for main-chain atoms. The hydrophobic core residues are well conserved, yet some large local differences are observed. The crystal structure reveals a dimeric organization, with the interface residues forming a mini four-helix bundle. However, analysis of solvation free energies and the surface area buried upon dimer formation indicated that this arrangement is more likely to be an effect of crystal packing rather than reflecting a physiological state. This is consistent with the solution structure being a monomer. The p73 alpha SAM domain also contains several interesting structural features: a Cys-X-X-Cys motif, a 3(10)-helix and a loop that have elevated B factors, and short tight inter-helical loops including two beta-turns; these elements are probably important in the normal function of this domain.

The UBX domain: a widespread ubiquitin-like module

The UBX domain is an 80 amino acid residue module that is present typically at the carboxyl terminus of a variety of eukaryotic proteins. In an effort to elucidate the function of UBX domains, we solved the three-dimensional structure of the UBX domain of human Fas-associated factor-1 (FAF1) by NMR spectroscopy. The structure has a beta-Grasp fold characterised by a beta-beta-alpha-beta-beta-alpha-beta secondary-structure organisation. The five beta strands are arranged into a mixed sheet in the order 21534. The longer first helix packs across the first three strands of the sheet, and a second shorter 3(10) helix is located in an extended loop connecting strands 4 and 5. In the absence of significant sequence similarity, the UBX domain can be superimposed with ubiquitin with an r.m.s.d. of 1.9 A, suggesting that the two structures share the same superfold, and an evolutionary relationship. However, the absence of a carboxyl-terminal extension containing a double glycine motif and of suitably positioned lysine side-chains makes it highly unlikely that UBX domains are either conjugated to other proteins or part of mixed UBX-ubiquitin chains. Database searches revealed that most UBX domain-containing proteins belong to one of four evolutionarily conserved families represented by the human FAF1, p47, Y33K, and Rep8 proteins. A role of the UBX domain in ubiquitin-related processes is suggested.

Biophysical characterization of elongin C from Saccharomyces cerevisiae

Elongin C (ELC) is an essential component of the mammalian CBC(VHL) E3 ubiquitin ligase complex. As a step toward understanding the role of ELC in assembly and function of CBC-type ubiquitin ligases, we analyzed the quaternary structure and backbone dynamics of the highly homologous Elc1 protein from Saccharomyces cerevisiae. Analytical ultracentrifugation experiments in conjunction with size exclusion chromatography showed that Elc1 is a nonglobular monomer over a wide range of concentrations. Pronounced line broadening in (1)H,(15)N-HSQC NMR spectra and failure to assign peaks corresponding to the carboxy-terminal helix 4 of Elc1 indicated that helix 4 is conformationally labile. Measurement of (15)N NMR relaxation parameters including T(1), T(2), and the (1)H-(15)N nuclear Overhauser effect revealed (i) surprisingly high flexibility of residues 69-77 in loop 5, and (ii) chemical exchange contributions for a large number of residues throughout the protein. Addition of 2,2,2-trifluoroethanol (TFE) stabilized helix 4 and reduced chemical exchange contributions, suggesting that stabilization of helix 4 suppresses the tendency of Elc1 to undergo conformational exchange on a micro- to millisecond time scale. Binding of a peptide representing the major ELC binding site of the von Hippel-Lindau (VHL) tumor suppressor protein almost completely eliminated chemical exchange processes, but induced substantial conformational changes in Elc1 leading to pronounced rotational anisotropy. These results suggest that elongin C interacts with various target proteins including the VHL protein by an induced fit mechanism involving the conformationally flexible carboxy-terminal helix 4.

The structure of a LysM domain from E. coli membrane-bound lytic murein transglycosylase D (MltD)

The LysM domain is a widespread protein module. It was originally identified in enzymes that degrade bacterial cell walls but is also present in many other bacterial proteins. Several proteins that contain the domain, such as Staphylococcal IgG binding proteins and Escherichia coli intimin, are involved in bacterial pathogenesis. LysM domains are also found in some eukaryotic proteins, apparently as a result of horizontal gene transfer from bacteria. The available evidence suggests that the LysM domain is a general peptidoglycan-binding module. We have determined the structure of this domain from E. coli membrane-bound lytic murein transglycosylase D. The LysM domain has a betaalphaalphabeta secondary structure with the two helices packing onto the same side of an anti- parallel beta sheet. The structure shows no similarity to other bacterial cell surface domains. A potential binding site in a shallow groove on surface of the protein has been identified.

Crystallization and preliminary crystallographic studies of a SAM domain at the C-terminus of human p73alpha

p73 is a recently discovered homologue of the tumour suppressor p53 and contains all three functional domains of p53. The alpha-splice variant of p73 (p73alpha) contains an additional structural domain near its C--terminus that has sequence homology with the sterile alpha-motif (SAM) domain. This domain is considered to be responsible for mediating protein-protein interactions. Pyramidal crystals of human p73alpha SAM domain were obtained by the hanging-drop vapour-diffusion method with ammonium dihydrogen orthophosphate as the precipitant. The crystals diffract to 2.54 A resolution and belong to the tetragonal space group P4(1)2(1)2, with unit-cell parameters a = b = 32.02, c = 133.84 A. The structure was solved by molecular replacement using the NMR structure of the same protein as the search model.

Folding of a dimeric beta-barrel: residual structure in the urea denatured state of the human papillomavirus E2 DNA binding domain

The dimeric beta-barrel is a characteristic topology initially found in the transcriptional regulatory domain of the E2 DNA binding domain from papillomaviruses. We have previously described the kinetic folding mechanism of the human HPV-16 domain, and, as part of these studies, we present a structural characterization of the urea-denatured state of the protein. We have obtained a set of chemical shift assignments for the C-terminal domain in urea using heteronuclear NMR methods and found regions with persistent residual structure. Based on chemical shift deviations from random coil values, 3'J(NHN alpha) coupling constants, heteronuclear single quantum coherence peak intensities, and nuclear Overhauser effect data, we have determined clusters of residual structure in regions corresponding to the DNA binding helix and the second beta-strand in the folded conformation. Most of the structures found are of nonnative nature, including turn-like conformations. Urea denaturation at equilibrium displayed a loss in protein concentration dependence, in absolute parallel to a similar deviation observed in the folding rate constant from kinetic experiments. These results strongly suggest an alternative folding pathway in which a dimeric intermediate is formed and the rate-limiting step becomes first order at high protein concentrations. The structural elements found in the denatured state would collide to yield productive interactions, establishing an intermolecular folding nucleus at high protein concentrations. We discuss our results in terms of the folding mechanism of this particular topology in an attempt to contribute to a better understanding of the folding of dimers in general and intertwined dimeric proteins such as transcription factors in particular.

RNA recognition by a Staufen double-stranded RNA-binding domain

The double-stranded RNA-binding domain (dsRBD) is a common RNA-binding motif found in many proteins involved in RNA maturation and localization. To determine how this domain recognizes RNA, we have studied the third dsRBD from Drosophila Staufen. The domain binds optimally to RNA stem-loops containing 12 uninterrupted base pairs, and we have identified the amino acids required for this interaction. By mutating these residues in a staufen transgene, we show that the RNA-binding activity of dsRBD3 is required in vivo for Staufen-dependent localization of bicoid and oskar mRNAs. Using high-resolution NMR, we have determined the structure of the complex between dsRBD3 and an RNA stem-loop. The dsRBD recognizes the shape of A-form dsRNA through interactions between conserved residues within loop 2 and the minor groove, and between loop 4 and the phosphodiester backbone across the adjacent major groove. In addition, helix alpha1 interacts with the single-stranded loop that caps the RNA helix. Interactions between helix alpha1 and single-stranded RNA may be important determinants of the specificity of dsRBD proteins.

Efficient purification and kinetic characterization of a bimodular derivative of the erythromycin polyketide synthase

Modular polyketide synthases (PKSs), such as the 6-deoxyerythronolide B synthase (DEBS), are giant multienzymes that biosynthesize a number of clinically important natural products. The modular nature of PKSs suggests the possibility of a combinatorial approach to the synthesis of novel bioactive polyketides, but the efficacy of such a strategy depends critically on gaining fundamental insight into PKS structure and function, most directly through experiments with purified PKS proteins. Several recent investigations into important aspects of the activity of these enzymes have used only partially purified proteins (often 3-4% of total protein), reflecting how difficult it is to purify these multienzymes in amounts adequate for kinetic and structural analysis. We report here the steady-state kinetic analysis of a typical bimodular PKS, 6-deoxyerythronolide B synthase 1-thioesterase (DEBS 1-TE), purified from recombinant Saccharopolyspora erythraea JCB101 by a new, high-yielding procedure consisting of three steps: ammonium sulfate precipitation, hydrophobic interaction chromatography and size-exclusion chromatography. The method provides 13-fold purification with a recovery of 11% of the applied PKS activity. The essentially homogeneous synthase exhibits an intrinsic methylmalonyl-CoA hydrolase activity, which competes with polyketide chain extension. The most reliable value for the kcat for synthesis of (3S,5R)-dihydroxy-(2R,4R)-dimethyl-n-heptanoic acid-delta-lactone is 0.84 min-1, and the apparent Km for (2RS)-methylmalonyl-CoA is 17 microM. This kcat is approximately 10-fold lower than the value reported previously for a differently engineered version of the truncated PKS, DEBS 1+TE. The difference likely reflects the fact that the DEBS 1-TE contains a hybrid acyl carrier protein (ACP) domain in its second module, which lowers its catalytic efficiency.

Knowledge-based design of bimodular and trimodular polyketide synthases based on domain and module swaps: a route to simple statin analogues

BACKGROUND

Polyketides are structurally diverse natural products that have a range of medically useful activities. Nonaromatic bacterial polyketides are synthesised on modular polyketide synthase (PKS) multienzymes, in which each cycle of chain extension requires a different 'module' of enzymatic activities. Attempts to design and construct modular PKSs that synthesise specified novel polyketides provide a particularly stringent test of our understanding of PKS structure and function.

RESULTS

We have constructed bimodular and trimodular PKSs based on DEBS1-TE, a derivative of the erythromycin PKS that contains only modules 1 and 2 and a thioesterase (TE), by substituting multiple domains with appropriate counterparts derived from the rapamycin PKS. Hybrid PKSs were obtained that synthesised the predicted target triketide lactones, which are simple analogues of cholesterol-lowering statins. In constructing intermodular fusions, whether between modules in the same or in different proteins, it was found advantageous to preserve intact the acyl carrier protein-ketosynthase (ACP-KS) didomain that spans the junction between successive modules.

CONCLUSIONS

Relatively simple considerations govern the construction of functional hybrid PKSs. Fusion sites should be chosen either in the surface-accessible linker regions between enzymatic domains, as previously revealed, or just inside the conserved margins of domains. The interaction of an ACP domain with the adjacent KS domain, whether on the same polyketide or not, is of particular importance, both through conservation of appropriate protein-protein interactions, and through optimising molecular recognition of the altered polyketide chain in the key transfer of the acyl chain from the ACP of one module to the KS of the downstream module.

NMR structure of the N-terminal domain of Saccharomyces cerevisiae RNase HI reveals a fold with a strong resemblance to the N-terminal domain of ribosomal protein L9

In addition to the conserved and well-defined RNase H domain, eukaryotic RNases HI possess either one or two copies of a small N-terminal domain. The solution structure of one of the N-terminal domains from Saccharomyces cerevisiae RNase HI, determined using NMR spectroscopy, is presented. The 46 residue motif comprises a three-stranded antiparallel beta-sheet and two short alpha-helices which pack onto opposite faces of the beta-sheet. Conserved residues involved in packing the alpha-helices onto the beta-sheet form the hydrophobic core of the domain. Three highly conserved and solvent exposed residues are implicated in RNA binding, W22, K38 and K39. The beta-beta-alpha-beta-alpha topology of the domain differs from the structures of known RNA binding domains such as the double-stranded RNA binding domain (dsRBD), the hnRNP K homology (KH) domain and the RNP motif. However, structural similarities exist between this domain and the N-terminal domain of ribosomal protein L9 which binds to 23 S ribosomal RNA.

The structure of a tunicate C-type lectin from Polyandrocarpa misakiensis complexed with D -galactose

C-type lectins are calcium-dependent carbohydrate-recognising proteins. Isothermal titration calorimetry of the C-type Polyandrocarpa lectin (TC14) from the tunicate Polyandrocarpa misakiensis revealed the presence of a single calcium atom per monomer with a dissociation constant of 2.6 microM, and confirmed the specificity of TC14 for D -galactose and related monosaccharides. We have determined the 2.2 A X-ray crystal structure of Polyandrocarpa lectin complexed with D -galactose. Analytical ultracentrifugation revealed that TC14 behaves as a dimer in solution. This is reflected by the presence of two molecules in the asymmetric unit with the dimeric interface formed by antiparallel pairing of the two N-terminal beta-strands and hydrophobic interactions. TC14 adopts a typical C-type lectin fold with differences in structure from other C-type lectins mainly in the diverse loop regions and in the second alpha-helix, which is involved in the formation of the dimeric interface. The D -galactose is bound through coordination of the 3 and 4-hydroxyl oxygen atoms with a bound calcium atom. Additional hydrogen bonds are formed directly between serine, aspartate and glutamate side-chains of the protein and the sugar 3 and 4-hydroxyl groups. Comparison of the galactose binding by TC14 with the mannose binding by rat mannose-binding protein reveals how monosaccharide specificity is achieved in this lectin. A tryptophan side-chain close to the binding site and the distribution of hydrogen-bond acceptors and donors around the 3 and 4-hydroxyl groups of the sugar are essential determinants of specificity. These elements are, however, arranged in a very different way than in an engineered galactose-specific mutant of MBPA. Possible biological functions can more easily be understood from the fact that TC14 is a dimer under physiological conditions.

Hot-spot mutants of p53 core domain evince characteristic local structural changes

Most of the oncogenic mutations in the tumor suppressor p53 map to its DNA-binding (core) domain. It is thus a potential target in cancer therapy for rescue by drugs. To begin to understand how mutation inactivates p53 and hence to provide a structural basis for drug design, we have compared structures of wild-type and mutant p53 core domains in solution by NMR spectroscopy. Structural changes introduced by five hot-spot mutations (V143A, G245S, R248Q, R249S, and R273H) were monitored by chemical-shift changes. Only localized changes are observed for G245S, R248Q, R249S, and R273H, suggesting that the overall tertiary folds of these mutant proteins are similar to that of wild type. Structural changes in R273H are found mainly in the loop-sheet-helix motif and the loop L3 of the core domain. Mutations in L3 (G245S, R248Q, and R249S) introduce structural changes in the loop L2 and L3 as well as terminal residues of strands 4, 9, and 10. It is noteworthy that R248Q, which is often regarded as a contact mutant that affects only interactions with DNA, introduces structural changes as extensive as the other loop L3 mutations (G245S and R249S). These changes suggest that R248Q is also a structural mutant that perturbs the structure of loop L2-L3 regions of the p53 core domain. In contrast to other mutants, replacement of the core residue valine 143 to alanine causes chemical-shift changes in almost all residues in the beta-sandwich and the DNA-binding surface. Long-range effects of V143A mutation may affect the specificity of DNA binding.

Molecular basis of Celmer's rules: the role of two ketoreductase domains in the control of chirality by the erythromycin modular polyketide synthase

BACKGROUND

Polyketides are compounds that possess medically significant activities. The modular nature of the polyketide synthase (PKS) multienzymes has generated interest in bioengineering new PKSs. Rational design of novel PKSs, however, requires a greater understanding of the stereocontrol mechanisms that operate in natural PKS modules.

RESULTS

The N-acetyl cysteamine (NAC) thioester derivative of the natural beta-keto diketide intermediate was incubated with DEBS1-TE, a derivative of the erythromycin PKS that contains only modules 1 and 2. The reduction products of the two ketoreductase (KR) domains of DEBS1-TE were a mixture of the (2S, 3R) and (2R,3S) isomers of the corresponding beta-hydroxy diketide NAC thioesters. Repeating the incubation using a DEBS1-TE mutant that only contains KR1 produced only the (2S,3R) isomer.

CONCLUSIONS

In contrast with earlier results, KR1 selects only the (2S) isomer and reduces it stereospecifically to the (2S, 3R)-3-hydroxy-2-methyl acyl product. The KR domain of module 1 controls the stereochemical outcome at both methyl-and hydroxyl-bearing chiral centres in the hydroxy diketide intermediate. Earlier work showed that the normal enzyme-bound ketoester generated in module 2 is not epimerised, however. The stereochemistry at C-2 is therefore established by a condensation reaction that exclusively gives the (2R)-ketoester, and the stereo-chemistry at C-3 by reduction of the keto group. Two different mechanisms of stereochemical control, therefore, operate in modules 1 and 2 of the erythromycin PKS. These results should provide a more rational basis for designing hybrid PKSs to generate altered stereochemistry in polyketide products.

The structure of a PKD domain from polycystin-1: implications for polycystic kidney disease

Most cases of autosomal dominant polycystic kidney disease (ADPKD) are the result of mutations in the PKD1 gene. The PKD1 gene codes for a large cell-surface glycoprotein, polycystin-1, of unknown function, which, based on its predicted domain structure, may be involved in protein-protein and protein-carbohydrate interactions. Approximately 30% of polycystin-1 consists of 16 copies of a novel protein module called the PKD domain. Here we show that this domain has a beta-sandwich fold. Although this fold is common to a number of cell-surface modules, the PKD domain represents a distinct protein family. The tenth PKD domain of human and Fugu polycystin-1 show extensive conservation of surface residues suggesting that this region could be a ligand-binding site. This structure will allow the likely effects of missense mutations in a large part of the PKD1 gene to be determined.

Evaluating precursor-directed biosynthesis towards novel erythromycins through in vitro studies on a bimodular polyketide synthase

BACKGROUND

Modular polyketide synthases (PKSs) catalyse the biosynthesis of complex polyketides using a different set of enzymes for each successive cycle of chain extension. Directed biosynthesis starting from synthetic diketides is a potentially valuable route to novel polyketides. We have used a purified bimodular derivative of the erythromycin-producing polyketide synthase (DEBS 1-TE) to study chain extension starting from a variety of diketide analogues and, in some cases, from the alternative acyl-CoA thioester substrates.

RESULTS

Chain initiation in vitro by DEBS 1-TE module 2 using a synthetic diketide analogue as a substrate was tolerant of significant structural variation in the starter unit of the synthetic diketide, but other changes completely abolished activity. Interestingly, a racemic beta-keto diketide was found to be reduced in situ on the PKS and utilised in place of its more complex hydroxy analogue as a substrate for chain extension. The presence of a diketide analogue strongly inhibited chain initiation via the loading module. Significantly higher concentrations of diketide N-acetylcysteamine analogues than their corresponding acyl-CoA thioesters are required to achieve comparable yields of triketide lactones.

CONCLUSIONS

Although a broad range of variation in the starter residue is acceptable, the substrate specificity of module 2 of a typical modular PKS in vitro is relatively intolerant of changes at C-2 and C-3. This will restrict the usefulness of approaches to synthesise novel erythromycins using synthetic diketides in vivo. The use of synthetic beta-keto diketides in vivo deserves to be explored.

Origin of starter units for erythromycin biosynthesis

Modular polyketide synthases (PKSs), such as the 6-deoxyerythronolide B synthase (DEBS), are multifunctional proteins that govern the synthesis of a number of clinically important natural products. The modular arrangement of active sites within these enzymes suggests the possibility of a combinatorial approach to the synthesis of novel bioactive polyketides. The efficacy of combinatorial strategies toward altering the starter unit specificity of polyketide synthases critically depends on controlling the supply of competing endogenous starter acids. Using DEBS 1-TE, a bimodular derivative of DEBS, we aimed to determine whether the beta-ketosynthase (KS) domain responsible for condensation in the first module also has the ability to prime its own biosynthesis by catalyzing the decarboxylation of methylmalonyl-CoA to produce propionyl-CoA. In contrast to earlier reports with a closely similar mini-PKS DEBS 1+TE, we have found that rigorously purified DEBS 1-TE does not catalyze the decarboxylation of methylmalonyl-CoA.

Characterisation of urea-denatured states of an immunoglobulin superfamily domain by heteronuclear NMR

The structural and dynamic properties of an immunoglobulin superfamily domain (IgSF), Ig 18', have been characterised by NMR at 285 K, in the presence of 4.2 M and 6.0 M urea, respectively. Analysis of chemical shift deviations, 3JHNHalpha coupling constants, sequential NOE pattern, and 15N relaxation data reveals that although the two urea-denatured states are highly disordered, some local turn-like residual structures do exist. Moreover, some distinct differences between the properties of the two denatured states are observed. In 4.2 M urea-denatured Ig 18', regions 80-83 and 86-92 adopt turn-like conformations, furthermore, region 84-93 is involved in slow exchange processes that occur on a micro- to millisecond time-scale. In the 6.0 M urea-denatured state, these turn-like conformations are less occupied, and chemical exchange processes in region 84-93 are largely reduced. In contrast, region 32-36 has persistent turn-like structures in both urea-denatured states. Some correlation between the spectral density function at 0 frequency, Jeff(0), for the urea-denatured states and the secondary structure elements of the folded state have been observed. Except for the terminal regions, residues corresponding to beta-strands have higher Jeff(0) values compared to residues corresponding to loops. The characterisation and comparison of the two urea-denatured states highlight residues that possess properties that may be crucial for the initiation of folding of this domain.

Crystal structure of a calcium-phospholipid binding domain from cytosolic phospholipase A2

Cytosolic phospholipase A2 (cPLA2) is a calcium-sensitive 85-kDa enzyme that hydrolyzes arachidonic acid-containing membrane phospholipids to initiate the biosynthesis of eicosanoids and platelet-activating factor, potent inflammatory mediators. The calcium-dependent activation of the enzyme is mediated by an N-terminal C2 domain, which is responsible for calcium-dependent translocation of the enzyme to membranes and that enables the intact enzyme to hydrolyze membrane-resident substrates. The 2.4-A x-ray crystal structure of this C2 domain was solved by multiple isomorphous replacement and reveals a beta-sandwich with the same topology as the C2 domain from phosphoinositide-specific phospholipase C delta 1. Two clusters of exposed hydrophobic residues surround two adjacent calcium binding sites. This region, along with an adjoining strip of basic residues, appear to constitute the membrane binding motif. The structure provides a striking insight into the relative importance of hydrophobic and electrostatic components of membrane binding for cPLA2. Although hydrophobic interactions predominate for cPLA2, for other C2 domains such as in "conventional" protein kinase C and synaptotagmins, electrostatic forces prevail.

The molecular basis of Celmer's rules: the stereochemistry of the condensation step in chain extension on the erythromycin polyketide synthase

Modular polyketide synthases (PKSs), for example, the 6-deoxyerythronolide B synthase (DEBS) responsible for synthesis of the aglycone core of the macrolide antibiotic erythromycin, generate an impressive diversity of asymmetric centers in their polyketide products. However, as noted by Celmer, macrolides have the same absolute configuration at all comparable stereocenters. Understanding how the stereochemistry of chain extension in controlled is therefore crucial to determining the common mechanism of action of these enzymes. We aimed to elucidate the molecular basis of Celmer's rules through in vitro studies with DEBS 1-TE, a bimodular derivative of DEBS from Saccharopolyspora erythraea, which uses (2S)-methylmalonyl-coenzyme. A to produce both D- and L-methyl centers in its triketide lactone product. We show here that condensation of (2S)-methylmalonyl-CoA in module 2 proceeds with decarboxylative inversion without cleavage of the C-H bond adjacent to the methyl group; in contrast, in module 1 the chain extension process involves loss of the hydrogen attached to C-2 of the methylmalonyl-CoA precursor. The production of the D-methyl center in module 2 without loss of hydrogen from the asymmetric center of the (2S)-methylmalonyl-CoA establishes that condensation takes place with inversion of configuration as in fatty acid biosynthesis. The loss of the key hydrogen from the (2S)-methylmalonyl-CoA to produce the L-methyl center generated in module 1 implies that an additional obligatory epimerization step takes place in that module. The nature and timing of the epimerization remain to be established.

The solution structure of the S1 RNA binding domain: a member of an ancient nucleic acid-binding fold

The S1 domain, originally identified in ribosomal protein S1, is found in a large number of RNA-associated proteins. The structure of the S1 RNA-binding domain from the E. coli polynucleotide phosphorylase has been determined using NMR methods and consists of a five-stranded antiparallel beta barrel. Conserved residues on one face of the barrel and adjacent loops form the putative RNA-binding site. The structure of the S1 domain is very similar to that of cold shock protein, suggesting that they are both derived from an ancient nucleic acid-binding protein. Enhanced sequence searches reveal hitherto unidentified S1 domains in RNase E, RNase II, NusA, EMB-5, and other proteins.

Structure and stability of an immunoglobulin superfamily domain from twitchin, a muscle protein of the nematode Caenorhabditis elegans

The NMR solution structure of an immunoglobulin superfamily module of twitchin (Ig 18') has been determined and the kinetic and equilibrium folding behaviour characterised. Thirty molecular coordinates were calculated using a hybrid distance geometry-simulated annealing protocol based on 1207 distance and 48 dihedral restraints. The atomic rms distributions about the mean coordinate for the ensemble of structures is 0.55( +/- 0.09) A for backbone atoms and 1.10( +/- 0.08) A for all heavy atoms. The protein has a topology very similar to that of telokin and the titin Ig domains and thus it falls into the I set of the immunoglobulin superfamily. The close agreement between the predicted and observed structures of Ig 18' demonstrates clearly that the I set profile can be applied in the structure prediction of immunoglobulin-like domains of diverse modular proteins. Folding studies reveal that the protein has relatively low thermodynamic stability, deltaG(H2O)U-F = 4.0 kcal mol(-1) at physiological pH. Unfolding studies suggest that the protein has considerable kinetic stability, the half life of the unfolding is greater than 40 minutes in the absence of denaturant.

The dimeric DNA binding domain of the human papillomavirus E2 protein folds through a monomeric intermediate which cannot be native-like

The dimeric DNA binding domain of the human papillomavirus E2 protein displays a two-state concerted unfolding and dissociation, with no detectable monomeric intermediate species accumulated at equilibrium. We investigated the kinetic folding mechanism of the dimeric domain using stopped-flow spectroscopic techniques and observed a fast forming monomeric intermediate, followed by a slower bimolecular reaction. Both phases involve secondary structure rearrangements of similar magnitude. Our results support a folding pathway in which the formation of an early monomeric intermediate, with characteristics of hydrophobic collapse, is followed by a bimolecular step encompassing association and folding. The interwoven folding topology of this particular type of dimeric beta-barrel found in the E2 DNA binding domain strongly suggests that any monomeric species formed could not be native-like.

Equilibrium dissociation and unfolding of the dimeric human papillomavirus strain-16 E2 DNA-binding domain

The equilibrium unfolding reaction of the C-terminal 80-amino-acid dimeric DNA-binding domain of human papillomavirus (HPV) strain 16 E2 protein has been investigated using fluorescence, far-UV CD, and equilibrium sedimentation. The stability of the HPV-16 E2 DNA-binding domain is concentration-dependent, and the unfolding reaction is well described as a two-state transition from folded dimer to unfolded monomer. The conformational stability of the protein, delta GH2O, was found to be 9.8 kcal/mol at pH 5.6, with the corresponding equilibrium unfolding/dissociation constant, Ku, being 6.5 x 10(-8) M. Equilibrium sedimentation experiments give a Kd of 3.0 x 10(-8) M, showing an excellent agreement between the two different techniques. Denaturation by temperature followed by the change in ellipticity also shows a concomitant disappearance of secondary and tertiary structures. The Ku changes dramatically at physiologically relevant pH's: with a change in pH from 6.1 to 7.0, it goes from 5.5 x 10(-8) M to 4.4 x 10(10) M. Our results suggest that, at the very low concentration of protein where DNA binding is normally measured (e.g., 10(-11) M), the protein is predominantly monomeric and unfolded. They also stress the importance of the coupling between folding and DNA binding.

A comparison of the pH, urea, and temperature-denatured states of barnase by heteronuclear NMR: implications for the initiation of protein folding

The denatured states of barnase that are induced by urea, acid, and high temperature and acid have been assigned and characterised by high resolution heteronuclear NMR. The assignment was completed using a combination of triple-resonance and magnetisation-transfer methods. The latter was facilitated by selecting a suitable mutant of barnase (Ile-->Val51) which has an appropriate rate of interconversion between native and denatured states in urea. 3J NH-C alpha H coupling constants were determined for pH and urea-denatured barnase and intrinsic "random coil" coupling constants are shown to be different for different residue types. All the denatured states are highly unfolded. But, a consistent series of weak correlations in chemical shift, NOESY and coupling constant data provides evidence that the acid-denatured state has some residual structure in regions that form the first and second helices and the central strands of beta-sheet in the native protein. The acid/temperature-denatured states has less structure in these regions, and the urea-denatured state, less still. These observations may be combined with detailed analyses of the folding pathway of barnase from kinetic studies to illuminate the relevance of residual structure in the denatured states of proteins to the mechanism of protein folding. First, the folding of barnase is known to proceed in its later stages through structures in which the first helix and centre of the beta-sheet are extensively formed. Thus, embryonic initiation sites for these do exist in the denatured states and so could well develop into true nuclei. Second, it has been clearly established that the second helix is unfolded in these later states, and so residual structure in this region of the protein is non-productive. These data fit a model of protein folding in which local nucleation sites are latent in the denatured state and develop only when they make interactions elsewhere in the protein that stabilise them during the folding process. Thus, studies of the structure of denatured states pinpoint where nucleation sites may be, and the kinetic and protein engineering studies show which ones are productive.

NMR solution structure of a dsRNA binding domain from Drosophila staufen protein reveals homology to the N-terminal domain of ribosomal protein S5

The double-stranded RNA binding domain (dsRBD) is an approximately 65 amino acid motif that is found in a variety of proteins that interact with double-stranded (ds) RNA, such as Escherichia coli RNase III and the dsRNA-dependent kinase, PKR. Drosophila staufen protein contains five copies of this motif, and the third of these binds dsRNA in vitro. Using multinuclear/multidimensional NMR methods, we have determined that staufen dsRBD3 forms a compact protein domain with an alpha-beta-beta-beta-alpha structure in which the two alpha-helices lie on one face of a three-stranded anti-parallel beta-sheet. This structure is very similar to that of the N-terminal domain of a prokaryotic ribosomal protein S5. Furthermore, the consensus derived from all known S5p family sequences shares several conserved residues with the dsRBD consensus sequence, indicating that the two domains share a common evolutionary origin. Using in vitro mutagenesis, we have identified several surface residues which are important for the RNA binding of the dsRBD, and these all lie on the same side of the domain. Two residues that are essential for RNA binding, F32 and K50, are also conserved in the S5 protein family, suggesting that the two domains interact with RNA in a similar way.

Assignment of the backbone 1H,15N,13C NMR resonances and secondary structure of a double-stranded RNA binding domain from the Drosophila protein staufen

NMR spectroscopy has been used to determine the secondary structure of one of the double-stranded RNA binding domains from the Drosophila protein staufen. The domain has an alpha beta beta beta alpha arrangement of secondary structure, with the beta strands forming an antiparallel beta sheet. The secondary structure differs from that found in the RNP RNA binding domain.

Toward solving the folding pathway of barnase: the complete backbone 13C, 15N, and 1H NMR assignments of its pH-denatured state

The structures of the major folding intermediate, the transition state for folding, and the folded state of barnase have been previously characterized. We now add a further step toward a complete picture of the folding of barnase by reporting the backbone 15N, 13C, and 1H NMR assignments for barnase unfolded at pH 1.8 and 30 degrees C. These assignments, which were obtained from a combination of heteronuclear magnetization transfer and backbone triple-resonance NMR experiments, constitute the first stage in the structural characterization of this denatured state by NMR. Interresidue nuclear Overhauser effect contacts and deviations from 1H random-coil chemical shifts provide evidence for stable residual structure. The structured regions span residues in the native protein that contain its major alpha-helix and central strands of the beta-sheet. Earlier experiments have shown that these regions are predominantly intact in the major folding intermediate and that their docking is partly rate determining in folding.

Three-dimensional solution structure and 13C assignments of barstar using nuclear magnetic resonance spectroscopy

We present the high-resolution solution structure and 13C assignments of wild-type barstar, an 89 amino acid residue polypeptide inhibitor of barnase, derived from heteronuclear NMR techniques. These were obtained from measurements on unlabeled, uniformly 15N- and 13C/15N-labeled, and 10% 13C-labeled barstar samples that have both cysteines (at positions 40 and 82) fully reduced. In total, 30 structures were calculated by hybrid distance geometry-dynamical simulated annealing calculations. The atomic rms distribution about the mean coordinate positions is 0.42 A for all backbone atoms and 0.90 A for all atoms. The structure is composed of three parallel alpha-helices packed against a three-stranded parallel beta-sheet. A more poorly defined helix links the second beta-strand and the third major alpha-helix. The loop involved in binding barnase is extremely well defined and held rigidly by interactions from the main body of the protein to both ends and the middle of the loop. This structure will be used to aid protein engineering studies currently taking place on the free and bound states of barstar and barnase.

Refolding of barnase mutants and pro-barnase in the presence and absence of GroEL

Three mutants of barnase and a pro-barnase variant, which have a variety of different physical properties but the same overall protein structure, were analysed for their folding in the presence of the molecular chaperone GroEL. Mutants were chosen on the basis that changes in their refolding rate constants in solution are not correlated with the changes in their stability. All barnase variants fold considerably more slowly when bound to GroEL. However, barnase refolding on GroEL parallels that in solution: there is a linear relationship between the refolding rate constants, obtained for wild-type and all mutants of barnase, in the presence and absence of GroEL. Barnase is synthesized in vivo with a 13 amino acid pro-sequence attached to the N-terminus. The pro-sequence of pro-barnase is shown by NMR spectroscopy to be devoid of defined structure. The presence of this pro-sequence has no effect on the overall refolding rate constant or the activity of barnase. In the presence of GroEL, the refolding of pro-barnase is retarded relatively more strongly than that of wild-type and the mutant barnase proteins, suggesting that the pro-sequence provides additional binding sites for the chaperone.

Local breathing and global unfolding in hydrogen exchange of barnase and its relationship to protein folding pathways

We have measured the rate constants for exchange of amide protons in 15N-labeled wild-type barnase and a disulfide mutant that is more stable by 2 kcal.mol-1. The relative rate constants for exchange for wild type and mutant should reflect the changes in the equilibrium constants for local or global unfolding. The values for regions whose structure has been shown to be unaffected by the mutation fall into three subsets: those that are essentially unaffected by the mutation and so presumably exchange by local breathing; those where the energies change by close to 2 kcal.mol-1 and so presumably require global unfolding for exchange; and intermediate values that probably reflect a mixture of local and global unfolding in wild-type barnase. Amide protons that require the full change in unfolding energy are predominantly in the beta-sheet, which forms early in folding, but also include two that are involved in tertiary interactions that are known not to be formed until late in the folding pathway. Exchange in the major helix, which is known to form early, is largely unaffected by mutation and so exchanges by local breathing. There is thus no direct relationship between hydrogen-exchange behavior and the protein folding pathway. However, experiments on mutants of varying stability may provide further evidence on the sequence of events in folding.

Structure and dynamics of barnase complexed with 3'-GMP studied by NMR spectroscopy

The binding of 3'-GMP to the ribonuclease, barnase, has been studied using heteronuclear 2D and 3D NMR spectroscopy. The 1H and 15N NMR spectra of barnase complexed with 3'-GMP have been assigned. 2D and 3D NOESY spectra have been used to identify inter- and intramolecular NOEs, and a solution structure for the barnase-3'-GMP complex has been calculated. The position of the guanine ring of the ligand is reasonably well defined in the structures. The guanine ring forms hydrogen bonds with the NH protons of Ser57 and Arg59. These residues are located in a loop that is conserved among the microbial guanine-specific ribonucleases. The 2'-hydroxyl of 3'-GMP is close to His102 and Glu73, which have been shown to be involved in catalysis. The phosphate group of 3'-GMP is close to a number of positively charged residues that have also been shown to be important for activity. The position of the sugar moiety of 3'-GMP is less well defined in the structures. Structures calculated for the complex could not simultaneously satisfy all the observed intermolecular NOEs for the sugar protons, suggesting that the sugar samples several conformations when bound to barnase. The binding of 3'-GMP to barnase in solution is similar to that observed in the crystal structures of nucleotides bound to related ribonucleases. 3'-GMP binding causes no major conformational change in barnase. In contrast to the small structural changes that occur, there is a significant decrease in the rates of hydrogen/deuterium exchange and aromatic ring rotation in the active site of barnase upon ligand binding.

Assignment of the backbone 1H and 15N NMR resonances and secondary structure characterization of barstar

Barstar, a polypeptide inhibitor of ribonucleases, has been studied by 2D and 3D NMR techniques using uniformly 15N-labeled protein. Backbone (15NH-C alpha H-C beta H) resonances were assigned for all but 5 of the 89 residues. Dihedral angle and deuterium exchange studies were used in conjunction with medium range inter-proton NOEs to characterize the secondary structure of barstar. The protein is composed of four alpha-helices and three short stretches of extended strand. By further analysis of the NOE data three of the helices were found to be parallel to each other with the single disulphide bond linking the second and fourth helices at their C-terminal ends.

Identification of the barstar binding site of barnase by NMR spectroscopy and hydrogen-deuterium exchange

The extracellular ribonuclease from Bacillus amyloliquifaciens, barnase, forms a tightly-bound one-to-one complex with its intracellular inhibitor barstar. The barstar binding site on barnase was characterized by comparing the differences in the chemical shift and hydrogen-deuterium exchange rates between free and bound barnase. Chemical shift assignments of barnase in the complex with barstar were determined from 3D NOESY-HMQC and TOCSY-HMQC spectra of a complex that had been prepared with uniformly 15N-labelled barnase and unlabelled barstar. Hydrogen exchange rates were obtained from an analysis of a series of [15N]HMQC spectra of a sample prepared in the same manner exchanged into D2O. The largest changes in either chemical shift or hydrogen-deuterium exchange rate are observed for residues located in the active-site and substrate binding loops indicating that barstar inhibits barnase activity by sterically blocking the active site.

The folding of an enzyme. V. H/2H exchange-nuclear magnetic resonance studies on the folding pathway of barnase: complementarity to and agreement with protein engineering studies

Two major methods are currently being used to characterize transient intermediates during protein folding at the level of individual residues. Nuclear magnetic resonance (n.m.r.) measurements on the protection of peptide NH hydrogens against exchange with solvent during refolding can provide information about secondary structure formation. Protein engineering and kinetics can provide direct information about intramolecular interactions of protein side-chains and indirect evidence on secondary structure. These procedures have provided the most complete pictures so far about protein folding intermediates. Both methods have been applied to the characterization of an intermediate in the refolding of barnase. Although the two methods give complementary information, there are some regions of the protein where the methods overlap well. We show that, with one possible exception that is obscure, n.m.r. and protein engineering give identical results for those interactions that can be analysed by both methods. This suggests that these are valid approaches for the study of protein folding intermediates in the case of barnase and that the combination of the methods is a powerful analytical procedure. Information provided by n.m.r. data that is complementary to the protein engineering experiments is: (1) early formation of the C terminus of helix2; (2) early formation of helix3; (3) early formation of several beta-turns (46-49, 101-104 in loop5); and (5) partial formation of loop5. Confirmatory evidence of protein engineering data on the intermediate is: (1) helix1 is complete from residues 10 to 18; (2) the interactions between all beta-strands are present; (3) part of loop2 is not formed; (4) part of loop3 is formed; and (5) some specific tertiary interactions are not made. For some interactions the protein engineering and H/2H exchange methods overlap directly. The information obtained for direct overlap is self consistent.

Characterization of phosphate binding in the active site of barnase by site-directed mutagenesis and NMR

Phosphate is a competitive inhibitor of transesterification of GpC by the ribonuclease barnase. Barnase is significantly stabilized in the presence of phosphate against urea denaturation. The data are consistent with the existence of a single phosphate binding site in barnase with a dissociation constant, Kd, of 1.3 mM. The 2D 1H NMR spectrum of wild-type barnase with bound phosphate is assigned. Changes in chemical shifts and NOEs for wild type with bound phosphate compared with free wild type indicate that phosphate binds in the active site and that only small conformational changes occur on binding. Site-directed mutagenesis of the active site residues His-102, Lys-27, and Arg-87 to Ala increases the magnitude of Kd for phosphate by more than 20-fold. The 2D 1H NMR spectra of the mutants His-102----Ala, Lys-27----Ala, and Arg-87----Ala are assigned. Comparison with the spectra of wild-type barnase reveals that His-102----Ala and Lys-27----Ala have essentially the same structure as weild type, while some structural changes occur in Arg-87----Ala. It appears that phosphate binding by barnase is effected mainly by positively charge residues including His-102, Lys-27, and Arg-87. This may have applications for the design of phosphate binding sites in other proteins.

Determination of the three-dimensional solution structure of barnase using nuclear magnetic resonance spectroscopy

The solution conformation of the ribonuclease barnase has been determined by using 1H nuclear magnetic resonance (NMR) spectroscopy. The 20 structures were calculated by using 853 interproton distance restraints obtained from analyses of two-dimensional nuclear Overhauser spectra, 72 phi and 53 chi 1 torsion angle restraints, and 17 hydrogen-bond distance restraints. The calculated structures contain two alpha-helices (residues 6-18 and 26-34) and a five-stranded antiparallel beta-sheet (residues 50-55, 70-75, 85-91, 94-101, and 105-108). The core of the protein is formed by the packing of one of the alpha-helices (residues 6-18) onto the beta-sheet. The average RMS deviation between the calculated structures and the mean structure is 1.11 A for the backbone atoms and 1.75 A for all atoms. The protein is least well-defined in the N-terminal region and in three large loops. When these regions are excluded, the average RMS deviation between the calculated structures and the mean structure for residues 5-34, 50-56, 71-76, 85-109 is 0.62 A for the backbone atoms and 1.0 A for all atoms. The NMR-derived structure has been compared with the crystal structure of barnase [Mauguen et al. (1982) Nature (London) 297, 162-164].

Surface electrostatic interactions contribute little of stability of barnase

Electrostatic interactions are believed to play an important role in stabilizing the native structure of proteins. We have quantified the contribution to stability of an interaction between two oppositely charged side-chains on the surface of barnase. Using site-directed mutagenesis, glutamate 28 and lysine 32 were introduced onto the solvent-accessible side of the second alpha-helix in barnase. These two residues are separated by one turn of the helix, and so are ideally situated for their opposite charges to interact. Double mutant cycle analysis reveals that the interaction between Glu28 and Lys32 contributes only approximately 0.2 kcal/mol to stability of the protein. All other interactions between exposed charged side-chains in barnase examined so far also contribute little to stability. We explain this low value by their location on the surface, rather than in the interior, of the protein.

Pathway and stability of protein folding

We describe an experimental approach to the problem of protein folding and stability which measures interaction energies and maps structures of intermediates and transition states during the folding pathway. The strategy is based on two steps. First, protein engineering is used to remove interactions that stabilize defined positions in barnase, the RNAse from Bacillus amyloliquefaciens. The consequent changes in stability are measured from the changes in free energy of unfolding of the protein. Second, each mutation is used as a probe of the structure around the wild-type side chain during the folding process. Kinetic measurements are made on the folding and unfolding of wild-type and mutant proteins. The kinetic and thermodynamic data are combined and analysed to show the role of individual side chains in the stabilization of the folded, transition and intermediate states of the protein. The protein engineering experiments are corroborated by nuclear magnetic resonance studies of hydrogen exchange during the folding process. Folding is a multiphasic process in which alpha-helices and beta-sheet are formed relatively early. Formation of the hydrophobic core by docking helix and sheet is (partly) rate determining. The final steps involve the forming of loops and the capping of the N-termini of helices.

Aromatic-aromatic interactions and protein stability. Investigation by double-mutant cycles

The side-chains of phenylalanine and tyrosine residues in proteins are frequently found to be involved in pairwise interactions. These occur both within repeating elements of secondary structure and in tertiary and quaternary interactions. It has been suggested that they are important in protein folding and stability, and non-bonded potential energy calculations indicate that a typical aromatic-aromatic interaction has an energy of between -1 and -2 kcal/mol and contributes between -0.6 and -1.3 kcal/mol to protein stability. There is such an aromatic pair on the solvent-exposed face of the first alpha-helix of barnase, the small ribonuclease from Bacillus amyloliquefaciens. The edge of the aromatic ring of Tyr17 interacts with the face of that of Tyr13. The two residues have been mutated both singly and pairwise to alanine, and their free energies of unfolding determined by denaturation with urea. Application of the double-mutant cycle analysis gives an interaction energy of -1.3 kcal/mol for the aromatic pair in the folded protein relative to solvation by water in the unfolded protein. This value is similar to that calculated from the change in surface-accessible area between the rings on the formation of the pair. Analysis of a further double-mutant cycle in which the Tyr residues are mutated to Phe indicates that the aromatic-aromatic interactions of Tyr/Tyr and Phe/Phe make identical contributions to protein stability. However, Tyr is preferred to Phe by 0.3(+/- 0.04) kcal/mol at the solvent-exposed face of the alpha-helix.

Strength and co-operativity of contributions of surface salt bridges to protein stability

Many of the interactions that stabilize proteins are co-operative and cannot be reduced to a sum of pairwise interactions. Such interactions may be analysed by protein engineering methods using multiple thermodynamic cycles comprising wild-type protein and all combinations of mutants in the interacting residues. There is a triad of charged residues on the surface of barnase, comprising residues Asp8, Asp12 and Arg110, that interact by forming two exposed salt bridges. The three residues have been mutated to alanine to give all the single, double and triple mutants. The free energies of unfolding of wild-type and the seven mutant proteins have been determined and the results analysed to give the contributions of the residues in the two salt bridges to protein stability. It is possible to isolate the energies of forming the salt bridges relative to the solvation of the separated ions by water. In the intact triad, the apparent contribution to the stabilization energy of the protein of the salt bridge between Asp12 and Arg110 is -1.25 kcal mol-1, whereas that of the salt bridge between Asp8 with Arg110 is -0.98 kcal mol-1. The strengths of the two salt bridges are coupled: the energy of each is reduced by 0.77 kcal mol-1 when the other is absent. The salt-linked triad, relative to alanine residues at the same positions, does not contribute to the stability of the protein since the favourable interactions of the salt bridges are more than offset by other electrostatic and non-electrostatic energy terms. Salt-linked triads occur in other proteins, for example, haemoglobin, where the energy of only the salt-bridge term is important and so the coupling of salt bridges could be of general importance to the stability and function of proteins.