Folding pathway of guanidine-denatured disulfide-intact wild-type and mutant bovine pancreatic ribonuclease A

Minimization of cavity size ensures protein stability and folding: structures of Phe46-replaced bovine pancreatic RNase A

The Phe46 residue, located in the hydrophobic core of RNase A, was replaced with other hydrophobic residues, leucine, valine, or alanine, and their X-ray crystallographic structures were determined up to 1.50-1.80 A resolution in an attempt to examine the relationship between structural changes and conformational stability or folding kinetics. The backbone structure of F46L, F46V, and F46A was indistinguishable from that of the wild-type enzyme, retaining the correct active site structure. However, one water molecule was included in the hydrophobic core of F46A, forming two hydrogen bonds with the backbone peptide chain. The side chain of Met29 in F46V and F46A adopted two different conformations in an equal occupancy. A trapped water molecule and two conformations of Met29 represent changes that minimize the cavity volume. Nevertheless, the replacement of Phe46 with the above residues resulted in a marked decrease in both thermal stability and folding reaction. Thus, Phe46 ensures the thermal stability and the rapid and correct folding of RNase A by the role it plays in forming a highly packed, hydrophobic core.

Comparison of heat- and pressure-induced unfolding of ribonuclease a: the critical role of Phe46 which appears to belong to a new hydrophobic chain-folding initiation site

To clarify the structural role of Phe46 inside the hydrophobic core of bovine pancreatic ribonuclease A (RNase A), thermal and pressure unfolding of wild-type RNase A and three mutant forms (F46V, F46E, and F46K) were analyzed by fourth-derivative UV absorbance spectroscopy. All the mutants, as well as the wild type, exhibited a two-state transition during both thermal and pressure unfolding, and both T(m) and P(m) decreased markedly when Phe46 was replaced with valine, glutamic acid, or lysine. The strongest effect was on the F46K mutant and the weakest on F46V. Both unfolding processes produced identical blue shifts in the fourth-derivative spectra, indicating that the tyrosine residues are similarly exposed in the temperature- and pressure-induced unfolded states. A comparison of Gibbs free energies determined from the pressure and temperature unfoldings, however, gave DeltaG(p)/DeltaG(t) ratios (r) of 1.7 for the wild type and 0.92 +/- 0.03 for the mutants. Furthermore, the DeltaV value for each mutant was larger than that for the wild type. CD spectra and activity measurements showed no obvious major structural differences in the folded state, indicating that the structures of the Phe46 mutants and wild type differ in the unfolded state. We propose a model in which Phe46 stabilizes the hydrophobic core at the boundary between two structural domains. Mutation of Phe46 decreases protein stability by weakening the unfolding cooperativity between these domains. This essential function of Phe46 in RNase A stability indicates that it belongs to a chain-folding initiation site.

Acceleration of oxidative folding of bovine pancreatic ribonuclease A by anion-induced stabilization and formation of structured native-like intermediates

Phosphate anions accelerate the oxidative folding of reduced bovine pancreatic ribonuclease A with dithiothreitol at several temperatures and ionic strengths. The addition of 400 mM phosphate at pH 8.1 increased the regeneration rate of native protein 2.5-fold at 15 degrees C, 3.5-fold at 25 degrees C, and 20-fold at 37 degrees C, compared to the rate in the absence of phosphate. In addition, the effects of other ions on the oxidative folding of RNase A were examined. Fluoride was found to accelerate the formation of native protein under the same oxidizing conditions. In contrast, cations of high charge density or ions with low charge density appear to have an opposite effect on the folding of RNase A. The catalysis of oxidative folding results largely from an anion-dependent stabilization and formation of tertiary structure in productive disulfide intermediates (des-species). Phosphate and fluoride also accelerate the initial equilibration of unstructured disulfide ensembles, presumably due to non-specific electrostatic and hydrogen bonding effects on the protein and solvent.

Proline isomerization in bovine pancreatic ribonuclease A. 1. Unfolding conditions

The slow fluorescence unfolding phase of bovine pancreatic ribonuclease A is studied by stopped-flow kinetics and site-directed mutagenesis of tyrosines to phenylalanine and prolines to alanine. It is shown conclusively that this phase arises from two specific sources: Tyr92 reporting on the cis-trans isomerization of Pro93 and Tyr115 reporting on the cis-trans isomerization of Pro114. Previous studies have conjectured that the slow unfolding phase arises from only one source (Tyr92-Pro93 cis-trans isomerization) based primarily on studies of the homologous protein guinea pig ribonuclease A [Schmid, F. X., Grafl, R., Wrba, A., and Beintema, J. J. (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 872-876]; it is proposed here that Lys113 in the latter protein interferes with the isomerization of the Lys113-Pro114 peptide group. The site-directed mutations studied here enable the individual isomerizations of Pro93 and Pro114 to be monitored, providing an optical technique by which these well-defined molecular folding events can be studied, under both folding and unfolding conditions, and compared to molecular simulations. The time constants for Pro93 and Pro114 isomerization agree closely with those of our box model of proline isomerization under unfolding conditions, which had been derived from exhaustive statistical modeling of double-jump refolding data [Juminaga, D., Wedemeyer, W. J., Garduño-Júarez, R., McDonald, M. A., and Scheraga, H. A. (1997) Biochemistry 36, 10131-10145].

Tyrosyl interactions in the folding and unfolding of bovine pancreatic ribonuclease A: a study of tyrosine-to-phenylalanine mutants

Three tyrosine-to-phenylalanine mutants of ribonuclease A (Y25F, Y92F, and Y97F) are investigated for their enzymatic activities, molecular stabilities, and unfolding/refolding kinetics. These mutants exhibit 80, 90, and 80%, respectively, of the catalytic activity of the wild-type enzyme. Thermal, Gdn.HCl, and pH transition measurements indicate that Y25F and Y97F are less stable than the wild-type protein, whereas the bulk of the thermodynamic and kinetic evidence indicates that Y92F is as stable as the wild-type protein. Differences in molar extinction coefficients indicate that tyrosines 25, 92, and 97 contribute 38, 13, and 39%, respectively, to the absorption difference between the folded and unfolded states, in general agreement with previous studies but possibly indicating the contribution of a fourth tyrosine residue to account for the remaining 10%. Stopped-flow single- and double-jump kinetic experiments were carried out on the wild-type and three mutant proteins. At least one tyrosine residue besides tyrosine 92 contributes to the slow fluorescence-unfolding phase; the likely candidate for this residue is tyrosine 115 which monitors the cis-trans isomerization of the X-Pro114 peptide bond. Tyrosines 25 and 97 are involved in interactions that retard conformational unfolding and accelerate conformational refolding as well as the cis-trans proline isomerization of the slow-refolding phases, presumably by stabilizing the major beta-hairpin structure of RNase A. These interactions may contribute to the strong pH dependence of the folding and unfolding of ribonuclease A. In contrast, tyrosine 92 does not affect the folding and unfolding rates significantly. An improved "box" model of proline isomerization under unfolding conditions was derived from exhaustive fitting of all possible box models. The kinetic data support the identification of Pro93 as the proline whose isomerization distinguishes the slow-refolding species (USII and USI) from the other, faster-refolding species (Uvf, Uf, and Um), implying that Pro93 isomerizes in the slow-refolding reactions USI --> N and IN --> N. Similarly, Pro114 seems to distinguish between the very fast-refolding species Uvf and the fast-refolding species Uf. Lastly, Pro117 seems to distinguish the major slow-refolding species USII from the minor slow-refolding species USI and the medium-refolding species Um from the fast- and very fast-refolding species.

The role of a trans-proline in the folding mechanism of ribonuclease T1

Protein folding is often retarded by the cis reversible trans isomerizations of prolyl peptide bonds both in vitro and in vivo. An important role for the folding mechanism is well established for the prolyl peptide bonds that are cis in the native protein, but not for those that are trans. Here we investigated the role of trans-Pro73 for the folding of ribonuclease T1 (which additionally contains two cis-prolines) by comparing the wild-type protein with the Pro73-->Val variant. The Pro-->Val substitution led to a destabilization of the folded protein by 8.5 kJ/mol, which is explained by the strong, 25-fold increase in the rate of unfolding. In contrast, the rates and amplitudes of the fast and slow refolding reactions were virtually unchanged. trans-Proline residues remain largely trans after unfolding, and therefore their contributions to the observed folding kinetics should indeed be insignificant for proteins which also contain one or more cis prolines. The cis-proline residues dominate the kinetics of refolding, because almost all slow-folding molecules contain the respective incorrect (trans) isomers, and because trans-->cis isomerizations are slower than cis-->trans isomerizations. The inability to detect contributions from a trans-proline to the kinetics of folding does not imply that this proline is non-essential for folding in the sense that its cis reversible trans isomerization is energetically uncoupled from conformational folding.

Kinetic and thermodynamic studies of the folding/unfolding of a tryptophan-containing mutant of ribonuclease A

Tryptophan was substituted for Tyr92 to create a sensitive and unique optical probe in order to study the unfolding and refolding kinetics of disulfide-intact bovine pancreatic ribonuclease A by fluorescence-detected stopped-flow techniques. The stability of the Trp mutant was found to be similar to that of wild-type RNase A when denatured by heat or GdnHCl, and the mutant was found to have 85% of the activity of the wild-type protein. Single-jump unfolding experiments showed that the unfolding pathway of the Trp mutant contains a fast and a slow phase similar to those seen previously for the wild-type protein, indicating that the mutation did not alter the unfolding pathway significantly. The activation energy of the slow-unfolding phase suggested that proline isomerization is involved, with the Trp residue presumably reporting on changes in its local environment. Single-jump refolding experiments revealed the presence of GdnHCl-independent burst phase and a native-like intermediate, most likely IN, on the folding pathway. Single-jump refolding data at various final GdnHCl concentrations were fit to a kinetic folding model involving two pathways to the native state; one pathway involves the intermediate IN, and the other is a direct one to the native state. This model provides site-specific information, since Trp92 monitors the formation of local structure only in the neighborhood of that residue. Double-jump refolding experiments permitted the detection of a previously reported, hydrophobically collapsed intermediate, I phi. The refolding data support the hypothesis that the region around position 92 is a chain-folding initiation site in the folding pathway.

Nature of the unfolded state of ribonuclease A: effect of cis-trans X-Pro peptide bond isomerization

The equilibrium unfolded state of disulfide-intact bovine pancreatic ribonuclease A is a heterogeneous mixture of unfolded species. Previously, four unfolded species have been detected experimentally. They are Uvf, Uf, UsII, and UsI which have refolding time constants on the millisecond, millisecond to second, second to tens of seconds, and hundreds of seconds time scales, respectively. In the current study, the refolding pathway of the protein was investigated under favorable folding conditions of 0.58 M GdnHCl, pH 5.0, and 15 degrees C. In addition to the above four unfolded species, the presence of a fifth unfolded species was detected. It has a refolding time constant on the order of 2 s under the conditions employed. This new unfolded species is labeled Um, for medium-refolding species. Single-jump refolding experiments monitored by tyrosine burial and by cytidine 2'-monophosphate inhibitor binding indicate that the different unfolded species refold to the native state along independent refolding pathways. The buildup of the different unfolded species upon unfolding of the protein from the native state was monitored by absorbance using double-jump experiments. These experiments were carried out at 15 degrees C and consisted of an unfolding step at 4.2 M GdnHCl and pH 2.0, followed, after a variable delay time, by a refolding step at 0.58 M GdnHCl and pH 5.0. The results of these experiments support the conclusion that the different unfolded species arise from cis-trans isomerizations at the X-Pro peptide bonds of Pro 93, 114, and 117 in the unfolded state of the protein. The rates of these isomerizations were obtained for each of these three X-Pro peptide bonds at 15 degrees C.

Dihydrofolate reductase synthesis in the presence of immobilized methotrexate. An approach to a continuous cell-free protein synthesis system

Dihydrofolate reductase was synthesized in a batch system in the presence of the affinity ligand methotrexate, bound to various matrices. Two types of gel were used: commercial methotrexate-agarose with pores inaccessible for translation machinery and methotrexate-POROS with pores easily accessible for translation reaction mixture components. The transcription/translation reaction was not inhibited by either the immobilized methotrexate or the matrix. The enzyme was synthesized with a high yield and could simultaneously be removed from the reaction mixture by the affinity matrix during the synthesis. With methotrexate-POROS present the reaction probably proceeded mainly in the pores of the gel. Kinetic limitations to the reaction in the presence of the gel were not observed. Active dihydrofolate reductase was eluted from methotrexate-POROS. The activity recovered was higher than dihydrofolate reductase activity synthesized in free solution system. The influence of the presence of immobilized methotrexate on dihydrofolate reductase synthesis will be further studied in a novel type of a continuous protein synthesis system.

Contribution of a tyrosine side chain to ribonuclease A catalysis and stability

An intricate architecture of covalent bonds and noncovalent interactions appear to position the side chain of Lys 41 properly within the active site of bovine pancreatic ribonuclease A (RNase A). One of these interactions arises from Tyr 97, which is conserved in all 41 RNase A homologues of known sequence. Tyr 97 has a solvent-inaccessible side chain that donates a hydrogen bond to the main-chain oxygen of Lys 41. Here, the role of Tyr 97 was examined by replacing Tyr 97 with a phenylalanine, alanine, or glycine residue. All three mutant proteins have diminished catalytic activity, with the value of Kcat being perturbed more significantly than that of Km. The free energies with which Y97F, Y97A, and Y97G RNase A bind to the rate-limiting transition state during the cleavage of poly(cytidylic acid) are diminished by 0.74, 3.3, and 3.8 kcal/mol, respectively. These results show that even though Tyr 97 is remote from the active site, its side chain contributes to catalysis. The role of Tyr 97 in the thermal stability of RNase A is large. The conformational free energies of native Y97F, Y97A, and Y97G RNase A are decreased by 3.54, 12.0, and 11.7 kcal/mol, respectively. The unusually large decrease in stability caused by the Tyr-->Phe mutation could result from a decrease in the barrier to isomerization of the Lys 41-Pro 42 peptide bond.

Autocatalytic folding of the folding catalyst FKBP12

Prolyl isomerases are folding enzymes and thus have the potential to catalyze their own folding. We show here that the folding of cytosolic FKBP12 (FK 506 binding protein) is an autocatalytic process both for the mature protein and for a fusion protein with an amino-terminal extension of 16 residues. Native FKBP contains seven trans-prolyl peptide bonds, and the cis-to-trans isomerizations of some or all of them constitute the slow, rate-limiting events in folding. The rate of an autocatalytic reaction increases with reactant concentration, because the product catalyzes its own formation. Accordingly, the folding of the fusion protein was more than 10-fold accelerated when the protein concentration was increased from 0.05 microM to 10 microM. At high concentrations of both forms of FKBP12 autocatalysis was very efficient, and the observed folding rate seemed to approach the rate of the fast direct folding reaction of the protein molecules with the correct (all trans) peptidyl-prolyl bond conformation.

Folding and unfolding kinetics of the proline-to-alanine mutants of bovine pancreatic ribonuclease A

Four single mutants (P42A, P93A, P114A, and P117A) of bovine pancreatic ribonuclease A (RNase A) in which each mutant has one of the four prolines of RNase A changed to alanine were prepared. The physical properties of these four mutants indicate that their native structure is essentially identical to that of wild-type RNase A. The disulfide-intact forms of these proteins were denatured in guanidine hydrochloride (Gdn.HCl) and then refolded by dilution of the Gdn.HCl. Single-jump folding, single-jump unfolding, and double-jump unfolding/folding stopped-flow experiments were carried out on wild-type and the four proline mutants of RNase A using absorption detection to follow the folding kinetics. The single-jump folding experiments carried out at six different final Gdn.HCl concentrations indicate that the folding rate constants of individual steps for the mutants are similar to those of wild-type RNase A. The Tyr92-Pro93 peptide bond has a cis conformation in native wild-type RNase A, and the results from our double-jump stopped-flow experiments indicate that the Tyr92-Ala93 peptide bond in the P93A mutant of RNase A is also cis in the native state. The existence of two cis peptide bonds (preceding Pro93 and Pro114) in wild-type RNase A is probably due to (as-yet-unidentified) long-range interactions, and such interactions are presumably the origin of a cis peptide bond even when alanine is substituted for Pro93. The data from the double-jump stopped-flow experiments are interpreted in terms of a folding/unfolding model. This model specifies the cis/trans isomerization state of the unfolded species (Uvf, Uf, Um, and Us) at each X-Pro peptide bond. Also, this model confirms the existence of several previously postulated chain-folding initiation sites.