Relationship between thermal stability, degradation rate and expression yield of barnase variants in the periplasm of Escherichia coli.

An advantage of exporting a recombinant protein to the periplasm of Escherichia coli is decreased proteolysis in the periplasm compared with that in the cytoplasm. However, protein degradation in the periplasm also occurs. It has been widely accepted that the thermodynamic stability of a protein is an important factor for protein degradation in the cytoplasm of E.coli. To investigate the effect of the thermodynamic stability of an exported protein on the extent of proteolysis in the periplasm, barnase (an extracellular ribonuclease from Bacillus amyloliquefaciens) fused to alkaline phosphatase leader peptide was used as a model protein. A set of singly or doubly mutated barnase variants were constructed for export to the E.coli periplasm. It was found that the half-life of the barnase variants in vivo increased with their thermodynamic stability in vitro. A dominant factor for the final yield of exported barnase was not exportability but the turnover rate of the barnase variant. The yield of a stabilized mutant was up to 50% higher than that of the wild type. This suggests that exporting a protein to the periplasm and using protein engineering to enhance the stability can be combined as a strategy to optimize the production of recombinant proteins.

Folding and stability of endoglucanase III, a single-domain cellulase from Trichoderma reesei.

The reversible folding of an endoglucanase (EGIII) from the filamentous fungus Trichoderma reesei was investigated by activity, tryptophan fluorescence, and peptide CD measurements. Equilibrium stability was determined by urea denaturation at various pH and temperature values. Unfolding and refolding rates were measured over a range of urea concentrations. The data from the equilibrium and kinetic studies fit a simple two-state model, except at lower urea concentrations, where the folding kinetics indicate a transient intermediate. Unfolding is very slow, with a half-life of about 2 h in 8 M urea at pH 5.5 and 25 degrees C. Comparison of the urea dependence of the folding kinetics and equilibrium indicates the protein undergoes 93% of its total change in solvent exposure on going from the unfolded state to the transition state. Thus, the transition state is quite compact. The presence of dithiothreitol destabilized the protein by 7 kcal/mol, indicating the presence of an unusually strong disulfide linkage between the two cysteines in the molecule. Protein stability is dramatically reduced at alkaline pH values; this can be attributed to a titratable shift (pKa = 7.8) in the slope of the urea dependence of unfolding.

Inhibition of human aromatase by mammalian lignans and isoflavonoid phytoestrogens.

Isoflavonoid phytoestrogens and lignans in plants are known to be constituents of animal and human food and recently they have been found in human urine and other biological materials. These compounds have received increasing attention because of their interesting biological properties and possible role in human cancer and other diseases. The present study demonstrates that the main mammalian lignan enterolactone (trans-2,3-bis[(3-hydroxyphenyl)methyl]-butyrolactone) and some other diphenols are moderate or weak inhibitors of human estrogen synthetase (aromatase) and that this lignan binds to or near the substrate region of the active site of the P-450 enzyme. The inhibition is competitive with respect to testosterone and androstenedione, and the lignan affinity is 1/75-1/300 that of these natural substrates. It is suggested that the high concentration of lignans in vegetarians, by inhibiting aromatase in peripheral and/or cancer cells and lowering estrogen levels, may play a protective role as antipromotional compounds during growth of estrogen-dependent cancers.

The folding of an enzyme. II. Substructure of barnase and the contribution of different interactions to protein stability.

Barnase is described anatomically in terms of its substructures and their mode of packing. The surface area of hydrophobic residues buried on formation and packing of the structural elements has been calculated. Changes in stability have been measured for 64 mutations, 41 constructed in this study, strategically located over the protein. The purpose is to provide: (1) information on the magnitudes of changes in stabilization energy for mutations of residues that are important in maintaining the structure; and (2) probes for the folding pathway to be used in subsequent studies. The majority of mutations delete functional moieties of side-chains or make isosteric changes. The energetics of the interactions are variable and context-dependent. The following general conclusions may be drawn, however, from this study about the classes of interactions that stabilize the protein. (1) Truncation of buried hydrophobic side-chains has, in general, the greatest effect on stability. For fully buried residues, this averages at 1.5 kcal mol-1 per methylene group with a standard deviation of +/- 0.6 kcal mol-1. Truncation of partly exposed leucine, isoleucine or valine residues that are in the range of 50 to 80 A2 of solvent-accessible area (30 to 50% of the total solvent-accessible area on a Gly-X-Gly tripeptide, i.e. those packed against the surface) has a smaller, but relatively constant effect on stability, at 0.81 kcal mol-1 per methylene group with a statistical standard deviation of +/- 0.18 kcal mol-1. (2) There is a very poor correlation between hydrophobic surface area buried and the free energy change for an extensive data set of hydrophobic mutants. The best correlation is found to be between the free energy change and the number of methylene groups within a 6 A radius of the hydrophobic groups deleted. (3) Burial of the hydroxyl group of threonine in a pocket that is intended for a gamma-methyl group of valine costs 2.5 kcal mol-1, in the range expected for the loss of two hydrogen bonds.(ABSTRACT TRUNCATED AT 400 WORDS)

Protein stabilization by engineered metal chelation.

A ligand can shift a protein's folding/unfolding equilibrium by binding with higher affinity to the native state. A metal-chelating site consisting of two histidines separated by three residues (His-X3-His) engineered into an alpha-helix provides a general and easily-implemented means for protein stabilization by this mechanism. We have tested this approach with the iso-1-cytochrome c of Saccharomyces cerevisiae substituted with histidine at positions 4 and 8 in its N-terminal alpha-helix. One mM Cu(II) complexed to iminodiacetate stabilizes the cytochrome c variant by ca. 1 kcal/mol, as determined by guanidinium chloride-induced unfolding. The protein's folding/unfolding equilibrium is shifted by a free energy equal to that calculated from the metal ion's preferential binding to the native protein. Given the ubiquity of surface alpha-helices and the additional possibility of inserting di-histidine chelating sites into turns and beta-structures, we conclude that this is a useful method for protein stabilization.

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.

Inhibition of human placental aromatase by novel homologated 19-oxiranyl and 19-thiiranyl steroids.

Novel homologated 19-oxiranyl- and 9-thiiranylandrost-4-ene-3,17-diones (8a,b and 9a,b, respectively) have been synthesized. The configuration and conformation of compound 8a have been established by X-ray crystallographic analysis. All four compounds have been shown to be competitive inhibitors of human placental aromatase. The thiiranes were more potent inhibitors than the corresponding oxiranes, and the 2'S isomers (8b and 9b) were better inhibitors than the 2'R (8a and 9a) diastereomers in each series. Spectroscopic studies with purified human placental aromatase suggest that the oxiranyl oxygen and thiiranyl sulfur of 2'S compounds 8b and 9b coordinate to the enzyme's heme iron.

Transient folding intermediates characterized by protein engineering.

Kinetic experiments on engineered mutants of barnase detect an intermediate on the folding pathway and allow the mapping of the tertiary interactions of the side chains and their energetics. Many of the interactions present in the final folded state tend to be either fully formed or not formed at all in the intermediate or subsequent transition state for folding, but the hydrophobic core becomes progressively consolidated. These methods in combination with NMR provide extensive structural characterization of the folding intermediate and the sequence of events in the folding pathway.

Detection and characterization of a folding intermediate in barnase by NMR.

Protein engineering is being developed for mapping the energetics and pathway of protein folding. From kinetic studies on wild-type and mutant proteins, the sequence and energetics of formation of tertiary interactions of side chains can be mapped and the formation of secondary structure inferred. Here we cross-check and complement results from this approach by using a recently developed procedure that traps and characterizes secondary structure in intermediate states using 1H NMR. The refolding of barnase is shown to be a multiphasic process in which the secondary structure in alpha-helices and beta-sheets and some turns is formed more rapidly than is the overall folding.

6 alpha-fluorotestosterone: a nonaromatizable androgen inhibitor of aromatase cytochrome P450.

In an early survey of steroids which might serve as estrogen precursors, Gual et al. reported that 6 alpha-fluorotestosterone is not aromatized by human placental microsomes. Subsequently, 6 alpha-fluorotestosterone has been used to distinguish between androgen- and estrogen-mediated physiologic effects. We have reexamined the interaction of 6 alpha-fluorotestosterone with human placental and rat ovarian microsomes and with reconstituted purified aromatase cytochrome P450. Under conditions in which testosterone was readily aromatized, no aromatization of 6 alpha-fluorotestosterone was observed using either fluorescence detection of dansyl-estrogens separated by high-performance liquid chromatography or estrogen radioimmunoassay methods. The lack of aromatization is not due to failure of 6 alpha-fluorotestosterone to bind to P450arom, because 6 alpha-fluorotestosterone acts as a competitive inhibitor of the enzyme, and it exhibits a binding affinity similar to that of testosterone. Moreover, the addition of 6 alpha-fluorotestosterone to human placental microsomes elicits a spectral shift indicative of conversion of the heme from a low to a high spin state as observed for androgen substrates, consistent with its binding to the substrate site. The mechanism by which substitution of a fluorine at the 6 alpha-position interferes with the aromatization reaction remains to be determined, but the inhibitory action on estrogen formation may potentiate the androgenic properties of 6 alpha-fluorotestosterone in vivo due to a lowering of estrogen levels.

Mapping the transition state and pathway of protein folding by protein engineering.

In the transition state for unfolding of barnase, the hydrophobic core between the major alpha-helix and beta-sheet is somewhat weakened, the C terminus of the major helix is largely intact but its N terminus is exposed and a major loop has been invaded by solvent.

Energetics of complementary side-chain packing in a protein hydrophobic core.

The energetics of complementary packing of nonpolar side chains in the hydrophobic core of a protein were analyzed by protein engineering experiments. We have made the mutations Ile----Val, Ile----Ala, and Leu----Ala in a region of the small bacterial ribonuclease barnase where the major alpha-helix packs onto the central beta-sheet. The destabilization resulting from the creation of cavities was determined by measuring the decrease in free energy of folding from reversible denaturation induced by urea, guanidinium chloride, or heat. The different methods give consistent and reproducible results. The loss in free energy of folding for the mutant proteins is 1.0-1.6 kcal/mol per methylene group removed. This exceeds by severalfold the values obtained from model experiments of the partitioning of relevant side chains between aqueous and nonpolar solvents. Much of this discrepancy arises because two surfaces are buried when a protein folds--both the amino acid side chain in question and the portions of the protein into which it packs. These experiments directly demonstrate that the interior packing of a protein is crucial in stabilizing its three-dimensional structure: the conversion of leucine or isoleucine to alanine in the hydrophobic core loses half the net free energy of folding of barnase with a concomitant decrease in yield of the expressed recombinant protein.

Contribution of hydrophobic interactions to protein stability.

A major factor in the folding of proteins is the burying of hydrophobic side chains. A specific example is the packing of alpha-helices on beta-sheets by interdigitation of nonpolar side chains. The contributions of these interactions to the energetics of protein stability may be measured by simple protein engineering experiments. We have used site-directed mutagenesis to truncate hydrophobic side chains at an alpha-helix/beta-sheet interface in the small ribonuclease from Bacillus amyloliquefaciens (barnase). The decreases in stability of the mutant proteins were measured by their susceptibility to urea denaturation. Creation of a cavity the size of a -CH2-group destabilizes the enzyme by 1.1 kcal mol-1, and a cavity the size of three such groups by 4.0 kcal mol-1.

Aromatase cytochrome P-450. Purification and characterization of the enzyme from human placenta.

Aromatase cytochrome P-450 (P-450arom) was purified from human placental microsomes. Preparations exhibit a single major band of approximately 55 kDa on SDS-polyacrylamide gel electrophoresis and have a specific content of 11-13 nmol P-450/mg protein. The purified enzyme exhibits spectral properties typical of ferric and ferrous forms of cytochromes P-450. Full enzymatic activity can be reconstituted with rabbit liver P-450 reductase, and catalytic characteristics similar to aromatase in microsomes are observed. Rabbit antibodies to purified P-450arom were affinity purified and show high specificity and sensitivity on immunoblots.

The active site of aromatase cytochrome P-450. Differential effects of cyanide provide evidence for proximity of heme-iron and carbon-19 in the enzyme-substrate complex.

19-Norandrostenedione and androstenedione are shown to be metabolized by purified, reconstituted human placental aromatase cytochrome P-450. Kinetic evidence indicates that both steroids share a common catalytic site: 19-norandrostenedione is a competitive inhibitor of androstenedione aromatization, and the Ki value for its inhibition (120 nM) is similar to the Km value for its metabolism (132 nM). The two substrates differ, however, in their sensitivity to inhibition by the heme-iron ligand cyanide; 19-norandrostenedione is approximately 3-fold more sensitive to cyanide inhibition. Spectroscopic studies show that this differential inhibition by cyanide occurs because androstenedione competes with cyanide, whereas 19-norandrostenedione promotes cyanide binding to the heme-iron. It is proposed that these opposite effects on cyanide-iron coordination are due to the proximity of the heme-iron and C-19 of androstenedione in the enzyme-substrate complex, which results in steric exclusion of cyanide from the active site by the C-19 methyl group of androstenedione. Dioxygen is not excluded from binding to the heme-iron during catalysis, presumably because it bonds at an angle, in contrast to the linear bond of iron-cyanide complexes. A model for the active site of aromatase cytochrome P-450 is presented.

Purification and characterization of human placental aromatase cytochrome P-450.

Aromatase cytochrome P-450, which catalyzes the conversion of androgens to estrogens, was purified from human placental microsomes. The enzyme was extracted with sodium cholate, fractionated by ammonium sulfate precipitation, and subjected to column chromatography in the presence of its substrate, androstenedione, and the nonionic detergent, Nonidet P-40. The preparation exhibits a single major band when analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and has a specific content of 11.5 nmol of P-450/mg of protein. The purified enzyme displays spectroscopic properties typical of the ferric and ferrous forms of cytochrome P-450. Full enzymatic activity can be reconstituted with rabbit liver microsomal cytochrome P-450 reductase and Nonidet P-40. Purified aromatase cytochrome P-450 displays catalytic characteristics similar to the enzyme in intact microsomes in the aromatization of androstenedione, 19-hydroxyandrostenedione and 19-oxoandrostenedione. Testosterone and 16 alpha-hydroxytestosterone are aromatized at maximal rates similar to androstenedione, and all substrates exhibit relative affinities corresponding to those observed in microsomes. We have raised rabbit antibodies to the purified enzyme which show considerable specificity and sensitivity on immunoblots.

Inhibition of aromatase cytochrome P-450 by 10-oxirane and 10-thiirane substituted androgens. Implications for the structure of the active site.

The mechanism of inhibition of estrogen synthetase (P-450arom) by 19R- and 19S-isomers of 10-oxiranyl-and 10-thiiranyl-4-estrene-3,17-dione was investigated using human placental microsomes and purified enzyme preparations. The 19R-isomers were potent inhibitors and exhibited affinities 36-fold (10-oxirane) and 80-fold (10-thiirane) greater than the respective 19S-isomers. Kinetic experiments showed that inhibition by the 19R-isomers is competitive with respect to substrate; inhibition constants for the (19R)-10-oxirane (Ki = 10 nM) and the 19R-10-thiirane (Ki = 2 nM) indicate that each binds with greater affinity than the androgen substrates androstenedione and testosterone. Inhibition time courses and kinetic data were consistent with high affinity, reversible binding. Spectral titrations of microsomal preparations and purified P-450arom showed that binding of the 19R-isomers shifts the Soret maximum of the ferric enzyme to 411 nm (10-oxirane) or 425 nm (10-thiirane); addition of excess androstenedione reversed the spectral changes, producing the high spin form of the enzyme with a Soret peak at 393 nm. These spectral shifts suggest that the oxygen atom of the 10-oxirane and the sulfur atom of the 10-thiirane are bound to the heme iron in the inhibitor complexes. These results suggest that the high affinities of the inhibitors arise from their dual interaction with the androgen binding site and with the heme. Coordination of the C19 heteroatom to the heme indicates that C19 of androgen substrates may be positioned sufficiently close to the heme to allow direct attack by an iron-bound oxidant. Stereoselective binding of the 19R-isomers by P-450arom further suggests that the heme is likely to be positioned above C1 and C2 of the A ring.

Inhibition of aromatase cytochrome P-450 (estrogen synthetase) by derivatives of alpha-naphthoflavone.

alpha-Naphthoflavone (ANF; 7,8-benzoflavone) is a potent competitive inhibitor of human aromatase cytochrome P-450 [J. T. Kellis, Jr. and L. E. Vickery, Science 225, 1032 (1984)]. We have further investigated inhibition of aromatase by several derivatives of ANF. Using human placental microsomes and 40 nM androstenedione as substrate, the compounds tested and their I50 values were: ANF, 0.07 microM; 2-(2-naphthyl)-4H-naphtho[1,2b]pyran-4-one, 1.0 microM; 7,8-benzoisoflavone, approximately 100 microM; and 2-phenyl-4H-naphtho[1,2b]furan, greater than 100 microM. These findings show the necessity of the keto group of ANF in its binding to the enzyme and the importance of size and position of substitution of the exocyclic phenyl ring. Derivatives of ANF with hydroxyl substitution at positions 5, 6, 7, 8, 9, and 10 were also screened. 9-Hydroxy-ANF, a known metabolite of ANF in liver microsomes, was the most effective (I50 = 20 nM). Inhibition by 9-hydroxy-ANF was competitive, and its Ki value of 5 nM indicates a higher affinity for the enzyme than the natural steroid substrates--the Km values for androstenedione and testosterone under these conditions are 10 and 80 nM respectively. 9-Hydroxy-ANF also induced a change in the absorption spectrum of hte aromatase cytochrome P-450 indicative of substrate displacement. Based on these data we propose a model for the binding of 9-hydroxy-ANF in which the 7,8-benzochromone ring system of the ANF derivatives occupies the steroid ring binding site of the enzyme.

Inhibition of human estrogen synthetase (aromatase) by flavones.

Several naturally occurring and synthetic flavones were found to inhibit the aromatization of androstenedione and testosterone to estrogens catalyzed by human placental and ovarian microsomes. These flavones include (in order of decreasing potency) 7,8-benzoflavone, chrysin, apigenin, flavone, flavanone, and quercetin; 5,6-benzoflavone was not inhibitory. 7,8-Benzoflavone and chrysin were potent competitive inhibitors and induced spectral changes in the aromatase cytochrome P-450 indicative of substrate displacement. Flavones may thus compete with steroids in their interaction with certain monooxygenases and thereby alter steroid hormone metabolism.

Inhibition of estrogen synthetase (aromatase) by 4-cyclohexylaniline.

4- Cyclohexylaniline , a structurally simple analog of the drug aminoglutethimide [d,l-3-(4-aminophenyl)3-ethyl-2, 6- piperidinedione ], was found to be an effective inhibitor of the aromatization of testosterone and androstenedione. With human placental microsomes, 4- cyclohexylaniline was a more potent aromatase inhibitor than d-aminoglutethimide. For androstenedione and testosterone aromatization, competitive inhibition by 4- cyclohexylaniline was observed; a Ki value of 0.14 microM was found with both substrates. A Ki value for d-aminoglutethimide inhibition of androstenedione aromatization of 0.3 microM was obtained. Kinetic analysis of the simultaneous inhibition by 4- cyclohexylaniline and d-aminoglutethimide suggests that both compounds bind to the same site on the enzyme. 4- Cyclohexylaniline and d-aminoglutethimide were also tested for inhibition of androstenedione aromatization in human and rat ovarian microsomes. With the human aromatase, both inhibitors exhibited approximately the same effectiveness; with rat aromatase, however, d-aminoglutethimide was more potent. 4- Cyclohexylaniline and d-aminoglutethimide were also assayed for their inhibition of cytochrome P-450-catalyzed cholesterol side-chain cleavage. When the enzyme from human placenta was used, 4- cyclohexylaniline was 24-fold less effective than d-aminoglutethimide, and when the purified bovine adrenal enzyme was used, it was 16-fold less effective. Thus, 4- cyclohexylaniline exhibits inhibitory specificity toward aromatase. Difference spectral measurements using crude placental microsomes and cholate extracts of these microsomes show that binding of 4- cyclohexylaniline produces a type II spectral change; this is indicative of coordination of the arylamine to the heme-iron of the aromatase cytochrome P-450. Consistent with this spectral finding was the fact that blockade of the amino group of both 4- cyclohexylaniline and d-aminoglutethimide by acetylation resulted in essentially complete loss of inhibitory activity toward aromatase. These findings establish that the arylamine moieties of both 4- cyclohexylaniline and d-aminoglutethimide are essential for their inhibitory actions. The potent aromatase inhibition by 4- cyclohexylaniline indicates that neither the complex glutarimide ring nor the chiral 3-ethyl substituent of d-aminoglutethimide is required for high affinity binding of this class of inhibitor to aromatase P-450.