Dissimilarity in the Reductive Unfolding Pathways of Two Ribonuclease Homologues

doi:10.1016/j.jmb.2004.03.014

J. Mol. Biol. (2004) 338, 795–809

Dissimilarity in the Reductive Unfolding Pathways of Two Ribonuclease Homologues Mahesh Narayan1, Guoqiang Xu1, Daniel R. Ripoll2, Huili Zhai1 Kathrin Breuker3, Celestine Wanjalla1, Howard J. Leung1 Amiel Navon4, Ervin Welker1, Fred W. McLafferty1 and Harold A. Scheraga1* 1 Baker Laboratory of Chemistry and Chemical Biology, Cornell University, Ithaca, NY 14853-1301, USA 2

Computational Biology Service Unit, Cornell Theory Center Cornell University, Ithaca, NY 14853-3801, USA 3 Institute of Organic Chemistry University of Innsbruck Innrain 52a, 6020 Innsbruck Austria 4

Department of Cell Biology Harvard Medical School, 240 Longwood Avenue, Boston, MA 02115, USA

Using DTTred as the reducing agent, the kinetics of the reductive unfolding of onconase, a frog ribonuclease, has been examined. An intermediate containing three disulfides, Ir, that is formed rapidly in the reductive pathway, is more resistant to further reduction than the parent molecule, indicating that the remaining disulfides in onconase are less accessible to DTTred. Disulfide-bond mapping of Ir indicated that it is a single species lacking the (30 –75) disulfide bond. The reductive unfolding pattern of onconase is consistent with an analysis of the exposed surface area of the cysteine sulfur atoms in the (30 –75) disulfide bond, which reveals that these atoms are about four- and sevenfold, respectively, more exposed than those in the next two maximally exposed disulfides. By contrast, in the reductive unfolding of the homologue, RNase A, there are two intermediates, arising from the reduction of the (40 –95) and (65 –72) disulfide bonds, which takes place in parallel, and on a much longer time-scale, compared to the initial reduction of onconase; this behavior is consistent with the almost equally exposed surface areas of the cysteine sulfur atoms that form the (40 – 95) and (65 – 72) disulfide bonds in RNase A and the fourfold more exposed cysteine sulfur atoms of the (30 – 75) disulfide bond in onconase. Analysis and in silico mutation of the residues around the (40 – 95) disulfide bond in RNase A, which is analogous to the (30 –75) disulfide bond of onconase, reveal that the side-chain of tyrosine 92 of RNase A, a highly conserved residue among mammalian pancreatic ribonucleases, lies atop the (40 –95) disulfide bond, resulting in a shielding of the corresponding sulfur atoms from the solvent; such burial of the (30 –75) sulfur atoms is absent from onconase, due to the replacement of Tyr92 by Arg73, which is situated away from the (30 – 75) disulfide bond and into the solvent, resulting in the large exposed surface-area of the cysteine sulfur atoms forming this bond. Removal of Tyr92 from RNase A resulted in the relatively rapid reduction of the mutant to form a single intermediate (des [40 –95] Y92A), i.e. it resulted in an onconase-like reductive unfolding behavior. The reduction of the P93A mutant of RNase A proceeds through a single intermediate, the des [40 – 95] P93A species, as in onconase. Although mutation of Pro93 to Ala does not increase the exposed surface area of the (40 – 95) cysteine sulfur atoms, structural analysis of the mutant reveals that there is greater flexibility in the (40 – 95) disulfide bond compared to the (65 – 72) disulfide bond that may make the (40 – 95) disulfide bond much easier to expose, consistent

Abbreviations used: ONC, frog (Rana pipiens) onconase; RNase A, bovine pancreatic ribonuclease A; BPTI, bovine pancreatic trypsin inhibitor; AEMTS, 2-aminoethylmethylthiosulfonate; DTTred, reduced dithiothreitol; SCX, strong cation-exchange; nS, an ensemble of disulfide-containing intermediates each having n disulfide bonds; des species, a structured intermediate containing all but one native disulfide bond; R, reduced ONC; WT, wild-type; ESI/FTMS, electrospray ionization/Fourier-transformed mass spectrometry; IRMPD, infrared multiphoton dissociation; MALDITOF, matrix-assisted laser desorption/ionization time-of-flight; PDB, Protein Data Bank. E-mail address of the corresponding author: [email protected] 0022-2836/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.

796

Reductive Unfolding Pathways of Two Homologues

with the reductive unfolding pathway and kinetics of P93A. Mutation of Tyr92 to Phe92 in RNase A has no effect on its reductive unfolding pathway, suggesting that the hydrogen bond between the hydroxyl group of Tyr92 and the carbonyl group of Lys37 has no impact on the local unfolding free energy required to expose the (40 – 95) disulfide bond. Thus, these data shed light on the differences between the reductive unfolding pathways of the two homologous proteins and provide a structural basis for the origin of this difference. q 2004 Elsevier Ltd. All rights reserved. *Corresponding author

Keywords: reductive unfolding; exposed surface area; local unfolding; global unfolding; in silico mutation

Introduction Ongoing research in our laboratory is focused on identifying the intramolecular interactions that govern the oxidative folding processes of proteins containing many disulfides.1 – 4 Since oxidative folding and reductive unfolding studies of such proteins are key for understanding the intramolecular interactions that govern their folding/ unfolding processes, we have used bovine pancreatic ribonuclease A (RNase A), which contains four disulfides, as our model system for this purpose.5 Reductive unfolding is a useful procedure for probing the nature of the transition state during protein unfolding and can provide information complementary to data gathered from studies of oxidative regeneration.6 – 17 Reductive unfolding is a probe for structural analysis of a protein. By varying the reducing conditions, it is possible to compute the free energy required for reduction of several of the disulfides of the protein, and thereby identify the nature of the unfolding process (global or local) that might be required to expose those disulfide bonds.13 By comparing the rates of reduction of protein disulfide bonds with those of model peptides, reductive unfolding can shed light on the effect of environment and degree of exposure of disulfide bonds on their relative reactivity.18 Finally, reductive unfolding studies of proteins containing multi-disulfides can sometimes be used to gather information about the contribution of some of the disulfide bonds toward maintaining a stable protein structure. Stable intermediates that may be formed during the reductive unfolding process indicate that the missing disulfide bond in those intermediates is not critical for maintaining the native-like structure under the given conditions, even though they may provide additional stability to the whole protein. In order to investigate the response of structurally homologous multi-disulfide proteins to the reductive unfolding process, and to assess the contributions of their disulfide bonds for maintaining their respective three-dimensional structures, and, in the absence of X-ray or NMR data, to obtain information about inter-residue interactions within their intermediates, we have

undertaken a study of the reductive unfolding of onconasew (ONC; registered trademark of Alfacell Corp., Bloomfield, NJ, USA), which contains four disulfides and is a homologue of RNase A (the defining member of pancreatic ribonucleases). ONC, a cytotoxin19,20 that is found in the oocytes and early embryos of the Northern leopard frog Rana pipiens,21 can degrade tRNAs selectively22 and is currently in phase III clinical trials for the treatment of asbestos-related lung cancer†. The amino acid sequences of ONC and RNase A share , 30% identity,21,23 – 25 and the positions of three of their four disulfides are conserved ((19 – 68), (30 –75), and (48 –90) in ONC; (26 – 84), (40 – 95), and (58 –110) in RNase A). While the disulfide analogous to the (65–72) bond in RNase A is missing from ONC, the frog variant has a synapomorphic disulfide at position (87 –104) that links its C terminus with a central b-strand, an arrangement that is unique to the amphibian variants of RNase A.24 The (65 – 72) disulfide bond of RNase A is of special importance in the oxidative folding and reductive unfolding of this protein.13,26 – 29 Disulfide mapping studies of the unstructured 1S and 2S ensembles of wild-type (WT) RNase A and of its three-disulfide mutants [C40A, C95A] and [C65A, C72A] reveal that the (65 – 72) bond is the predominant disulfide bond in unstructured intermediates (for example, it is present to the extent of 40% of the 1S ensemble of WT RNase A).26 – 29 The dominance of this disulfide bond in the unstructured intermediates is reflected in the predominant formation of a structured intermediate, des [40 – 95] (containing the (65 – 72) disulfide bond and constituting about 80% of the structured intermediates that form the native protein, the remaining 20% being des [65 – 72]) in the rate-determining steps of the oxidative folding of RNase A.30,31 Similar to oxidative folding, the reductive unfolding of RNase A also takes place through two parallel pathways with des [40 – 95] and des [65 – 72] being the early-detected intermediates, in its reductive unfolding pathway.13 ONC23 and RNase A32 have similar threedimensional structures but, because ONC lacks

† http://www.alfacell.com/clinical.htm

Reductive Unfolding Pathways of Two Homologues

797

the (65 –72) disulfide bond of RNase A, the folding/unfolding pathways probably differ. Here, therefore, we have examined the reductive unfolding pathway of ONC, and compared it with those of RNase A, of its P93A mutant (whose threedimensional structure is known33,34) and of its Y92F35 and Y92A mutants. This study provides information about important inter-residue interactions in the two variants, and their impact on the respective reductive unfolding pathways of the two proteins.

Results Reductive unfolding of ONC Figure 1A shows the elution profile of ONC on a cation-exchange column immediately after the start of the reduction process (10 mM DTTred). Only one peak corresponding to native ONC is visible, indicating that the starting material was essentially pure (verified by mass spectrometry; experimentally determined mass 11,820 Da; calculated mass,21 11,819 Da). Figure 1B shows the progress of the reduction of ONC by DTTred (10 mM), five minutes after the start of the reaction. A species that elutes after native ONC, which we denote as Ir (I, for intermediate; r for reductive unfolding pathway), is clearly evident in the chromatogram. The sharpness of the HPLC peak, when taking into account the relatively shallow gradient (1 mM NaCl change in 1.5 minutes), suggests that Ir is a single species. The fact that Ir elutes after ONC on a cation-exchange column indicates that it is more positively charged than the parent molecule; this is the expected behavior of an 2-aminoethylmethylthiosulfonate (AEMTS)blocked intermediate lacking at least one disulfide bond, compared to the parent molecule.5 Figure 2 shows the kinetic profiles for the partial reduction of ONC (B) by DTTred (10 mM) and the concomitant formation of Ir (X). No reductive intermediates of Ir or fully reduced ONC were observed under these conditions, indicating that the reduction of this intermediate is much slower than the initial reduction of the parent molecule. This implies that the remaining disulfides in Ir are less accessible to DTTred than the particular disulfide that was reduced in ONC to produce Ir. While buried disulfides can be exposed periodically by unfolding (breathing) of the protein, the concentration of DTTred used here does not appear to be sufficient to reduce the protein effectively during the short time interval when its buried disulfides are exposed. For comparison, we studied the reductive unfolding of RNase A under similar conditions (Figure 2) (w, native RNase A). From the plot, it is clear that ONC can be (partially) reduced very easily compared to RNase A. In fact, under the given reducing conditions (10 mM DTTred), no intermediate of RNase A could be detected within

Figure 1. HPLC chromatograms showing the elution of ONC and its intermediate, Ir, which is populated during the reduction of native ONC by 10 mM DTTred (20 mM Tris – HCl (pH 8), 0.2 mM EDTA, 15 8C). A, Native ONC immediately upon starting the reduction process. B, Native ONC and AEMTS-blocked Ir, five minutes after the start of the reduction process.

the time of the experiment. The kinetic data, therefore, suggest the presence of a relatively more easily exposable disulfide bond in ONC compared to the disulfides in RNase A. Since we could not detect fully reduced ONC in the previous experiment (using 10 mM DDTred), we studied the reductive unfolding of ONC at a higher concentration of DTTred (100 mM). Figure 3 is a plot of the consumption of native ONC upon addition of DTTred (100 mM). Within the time of the experiment (300 minutes), only one peak, corresponding to a disulfide-containing intermediate (i.e. Ir), was found on the HPLC, and a very small amount (9%) of fully reduced ONC was detected (data not shown). No disulfide-containing reductive intermediate of Ir was detected.

Figure 2. Kinetics of partial reduction of ONC (0.5 mg/ml) obtained by using 10 mM DTTred, 20 mM Tris – HCl (pH 8), 0.2 mM EDTA, 15 8C. (B) Consumption of ONC; (X) formation of Ir; and (w) reduction of RNase A.

798

Figure 3. Kinetics of consumption of native ONC upon addition of 100 mM DTTred (20 mM Tris – HCl (pH 8), 0.2 mM EDTA, 15 8C).

Identity of Ir ESI/FTMS mass spectrometry36 was used to determine the relative molecular mass ðMr Þ of AEMTS-blocked Ir. Figure 4 shows the spectrum of Ir at the 9 þ charge state. The experimentally obtained Mr of 11,970.8-7 (calculated 11,970.8-7) corresponds to a mass increase of 151.8 Da over native ONC, which is the mass of two adducts (– SCH2CH2NH2) derived from the AEMTS blocking.5 Since one AEMTS molecule reacts with one free thiol group, this indicates that Ir has two free thiol groups, i.e. it is a species containing three disulfides. Since Ir is able to survive mildly reducing conditions (10 mM DTTred), this indicates that its disulfides are protected from the reducing agent.2,3 Under the conditions of the experiment, disulfide bonds in unstructured species are reduced rapidly;30 therefore, Ir appears to be a stable, structured intermediate of ONC. Identification of the missing disulfide bond in Ir Two methods (proteolytic digestion/matrix-

Figure 4. ESI/FTMS spectrum of Ir blocked by two AEMTS groups. Dots represent the theoretical abundance distribution of the isotopic peaks. The mass difference (0.7 Da) between the most abundant peak and the monoisotopic peak is indicated as 27.

Reductive Unfolding Pathways of Two Homologues

assisted laser desorption/ionization time-of-flight (MALDI-TOF)26 and electrospray ionization/ Fourier-transformed mass spectrometry (ESI/FTMS)36) were used to determine the position of the missing disulfide bond in Ir. For both of these methods, the free cysteine residues of Ir were first carboxymethylated, and then the remaining disulfides in Ir were reduced and blocked with AEMTS. Thus, by determining the cysteine residues that are carboxymethylated, the initially reduced disulfide bond that results in the formation of Ir can be identified. The molecular mass of carboxymethylated Ir and carboxymethylated-then-AEMTS-blocked Ir was determined by MALDI-TOF (data not shown). The molecular mass of carboxymethylated Ir indicated that both cysteine residues of Ir were blocked, consistent with previous mass spectroscopic results obtained from AEMTS-blocked Ir (see the previous section). The mass of the carboxymethylated then AEMTS-blocked sample showed that six cysteine residues of Ir were blocked with AEMTS in addition to the two that had previously been carboxymethylated (data not shown). Method 1 In method 1, these blocked proteins were digested and the peptide fragments were isolated and analyzed by MALDI-TOF as described in Material and Methods. The theoretical and experimental masses of digested and fractionated fragments of AEMTS-blocked carboxymethylated Ir, which are sufficient for determining the missing disulfide bond, are listed in Table 1. The assignments of the residues in native ONC that correspond to the experimental and theoretical molecular mass values are listed in Table 1. The subscripts C and X in the fragment sequences represent the blocking groups, – CH2CONH2 and –SCH2CH2NH2, respectively, obtained from the reaction of free cysteine residues with iodoacetamide and AEMTS, respectively. The superscripted numerals indicate cysteine positions in the amino acid sequence and are determined as described below. The experimental masses are in good agreement with the calculated masses (Table 1). Analysis of the fragment sequences and masses indicates that, in fragment 3, either Cys68 or Cys75 is carboxymethylated. Since carboxymethylation of the two free cysteine residues in Ir was carried out prior to reduction of the remaining three disulfide bonds and AEMTS-blocking of the resulting six free cysteine residues, the mass number of fragment 3 indicates that either the (19 – 68) or the (30 – 75) disulfide bond in Ir is broken. From the sequence and mass of fragment 1, it can be seen that Cys19 was blocked with AEMTS; therefore, the (30 – 75) disulfide bond, but not the (19 –68) disulfide bond, is the broken one, and Ir is des [30 – 75]; des [30 – 75] refers to a species containing all native disulfide bonds, except (30 – 75). Fragments 1 and

799

Reductive Unfolding Pathways of Two Homologues

Table 1. MALDI-TOF mass spectroscopy results and assignments of the fragments obtained from chymotrypsin digestion of reduced AEMTS-blocked carboxymethylated Ir No. 1 2 3 4

Fragmentsa 19

QKKHITNTRDVD CXDNIMSTNLF IYSRPEPVKAI48CXKGIIASKNVLTTSEF YLSD68CXNVTSRP75CCKY 87 CXVT90CXENQAPVHF

Expt. mass (Da)

Calc. mass (Da)

Assignment

2669.6 3041.7 1781.6 1499.8

2669.9 3041.5 1781.9 1499.6

7–28 37–63 64–77 87–98

a Subscripts C and X represent the blocking groups, –CH2CONH2 and –SCH2CH2NH2, obtained from the reaction of free cysteine residues with iodoacetamide and AEMTS, respectively. The superscripts to the cysteine residues indicate the location of the particular cysteine residue in the amino acid sequence. Deducing the type of the blocking group on each cysteine residue, and the assignment of the two free cysteine residues in Ir is described in detail in Results under the heading Identification of the missing disulfide bond in Ir (Method 1).

3 are the unique minimal markers required to deduce the identity of the first-reduced disulfide bond in Ir in the current mapping experiment. The other two fragments (2 and 4) confirm this conclusion by eliminating the possibility that the other two disulfide bonds are the first-reduced disulfide bond during the reductive unfolding of ONC because Cys48, Cys87, and Cys90 were blocked with AEMTS in the mass spectroscopic results. In this method, we assume that there is no thiol – disulfide reshuffling within the structured intermediate(s) during reductive unfolding, which would otherwise result in the formation of at least one unstructured species. However, under the strongly reducing conditions, the unstructured species would be reduced immediately to R (fully reduced protein) before it can reshuffle to form another structured species. Therefore, it is reasonable to assume that all the disulfides in the reductive unfolding intermediate(s) are native disulfides. Method 2 In method 2, a top-down mass spectrometric approach36 was used to identify the disulfide distribution in Ir without involving its proteolytic digestion. The ESI/FTMS mass spectrum (not shown) of the carboxymethylated then AEMTSblocked sample showed dominant (90%) molecular ions of mass 12,391.9-7 Da; the calculated mass for this species is 12,391.9-7 Da. The only other observable molecular ion (with an almost negligible signal-to-noise ratio) has a mass of 12,448.9-7 Da; this corresponds to the non-thiolspecific substitution of a third – CH2CONH2 group (58 Da) for H (1 Da) in the treatment with ICH2CONH2. After ejection of all but the 12,391.9 Da ions from the FTMS ion cell, infrared multiphoton dissociation (IRMPD), which cleaves the molecule in only one position, resulted in 43 fragment ions that contain either the N-terminal (b ions) or the C-terminal (y ions) fragment of the protein. These fragments (not shown) showed alkylation at only Cys30 and Cys75, with no evidence for AEMTS derivatization of these cysteine residues, indicating

that Ir lacked the (30 –75) disulfide bond. For example, the b22 fragment ion (the ion containing the N-terminal 22 amino acid residues) had calculated masses of 2730.41 Da and 2712.33 Da, depending on whether Cys19 is derivatized by AEMTS or by alkylation, respectively. The IRMPD spectrum (not shown) has only a peak of 2730.39 Da, whose intensity is . 15 times the noise level at 2712.33 Da, showing that AEMTS derivatization is dominant at Cys19. In a similar manner, the b32 ion would have a mass of 3983.04 Da or 3964.96 Da for derivatization of Cys30 by AEMTS or by alkylation, respectively; the intensity of the IRMPD 3964.99 Da peak is about six times the noise level of the 3964.96 Da peak, consistent with dominant alkylation of Cys30. In a similar manner, the data for the b73 and b81 peaks (not shown) demonstrate that alkylation at Cys75 is eight times and six times, respectively, greater than the noise level for AEMTS derivatization. Thus, the (30 –75) disulfide bond is identified in the detection limits of the experiments exclusively as the one lost in forming the Ir intermediate. Exposed surface areas of cysteine sulfur atoms in ONC and RNase A In order to understand the structural basis of the dissimilar unfolding of the two homologous proteins, we examined the contribution to the molecular surface areas of their cysteine sulfur atoms using the Connolly algorithm,37 and we examined the region surrounding the (40 –95) and (30 – 75) disulfide bonds in their X-ray structures obtained from the RCSB Protein Data Bank (PDB).38,39 Furthermore, since previous studies40 have shown that the reductive unfolding of the P93A mutant of RNase A is similar to that of ONC, in that it unfolds reductively through only one pathway, i.e. through des [40 – 95] P93A, since its (40 –95) disulfide bond is reduced very rapidly, we included the P93A RNase A mutant in our considerations. The cysteine sulfur atoms of disulfide bonds (30 – 75), (87 – 104), and (19 – 68) in ONC, and (40 – 95) and (65 – 72) in RNase A have non-zero exposed surface areas and the calculated values for these exposed areas are listed in Table 2. In

800

Reductive Unfolding Pathways of Two Homologues

Table 2. Calculated contribution to the exposed surface areas of cysteine sulfur atoms forming the four individual disulfide bonds of RNase A and ONC using ˚ probe radius a 1.4 A RNase A

ONC

Disulfide

˚ 2) Area (A

Disulfide

˚ 2) Area (A

(26–84) (40–95) (58–110) (65–72)

0 5.8 0 5.2

(19– 68) (30– 75) (48– 90) (87– 104)

3.3 23.8 0 5.5

The exposed surface area of the cysteine sulfur atom in Gly˚ 2, and corresponds to approximately half of Cys-Gly is 26.7 A that for an exposed disulfide bond.

ONC, the sulfur atoms of cysteine residues 30 and 75, corresponding to the (30 –75) disulfide bond, ˚ 2 of exposed surface area, whereas the have 23.8 A cysteine sulfur atoms of the (87 – 104) and (19 – 68) ˚ 2 exposed ˚ 2 and 3.3 A disulfide bonds have 5.5 A surface areas, respectively. These numbers indicate that the cysteine residues in the (30 –75) disulfide

bond have about fourfold and sevenfold more exposed surface area compared to the other two exposed cysteine sulfur atoms, (87 –104) and (19 – 68), respectively (disulfide bond (48 –90) is completely buried). In RNase A, the cysteine sulfur atoms forming the (40 – 95) and (65 –72) disulfide bonds are ˚ 2) ˚ 2 and 5.2 A approximately equally exposed (5.8 A and are about fourfold less exposed than the cysteine sulfur atoms forming disulfide bond (30 – 75) in ONC (Table 2). Furthermore, examination of the residues occluding the (40 – 95) cysteine sulfur atoms from solvent in RNase A revealed that the side-chain of Tyr buries the cysteine sulfur atoms of (40 – 95) (Figure 5A) from the solvent. Such a protection is absent from ONC (Figure 5B), since Tyr92 is replaced by Arg73, whose side-chain lies away from the protein, and into the solvent. The magnitude of the burial provided by Tyr92 and other side-chains in the vicinity of the (40 – 95) disulfide bond of RNase A is assessed by comparing the exposed surface areas of cysteine sulfur

Figure 5. A, Illustration of how the ring of Tyr92 in RNase A (in blue) is oriented to cover a large portion of the cysteine 40 and cysteine 95 sulfur atoms (in yellow), since it hydrogen bonds with the backbone carbonyl group of Lys37.32,38,39 B, However, in ONC, Tyr92 is replaced by Arg73, whose side-chain (color-coded according to the temperature factor) is oriented away from the (30 – 75) disulfide bond, resulting in a large contribution to the exposed surface area of this disulfide. For clarity, the atomic details of the molecules have been omitted and only the secondary-structure elements are shown (b-sheets as light blue ribbons, a-helices as dark blue cylinders, and turns and coil regions as light green worms), with the N termini of the polypeptide chains indicated by N. The disulfide bonds in these structures are shown as yellow sticks, while the side-chain atoms of Lys37 and Tyr92 in RNase A, and Arg73 in ONC, are shown as sticks and are colored according to the values of the temperature factors (tf),41 using the following color code: 0 , tf , 5.0, blue; 5.0 , tf , 25.0, light blue; 25.0 , tf , 30.0, white; 30.0 , tf , 40.0, pink; tf . 40.0, red.

801

Reductive Unfolding Pathways of Two Homologues

Table 3. Exposed surface areas of the sulfur atoms of the (40 –95) and (30 – 75) disulfide bonds in RNase A and ONC and in their in silico mutants

Protein WT RNase A ONC Y92G RNase A Y92A RNase A T36G RNase A Y92G/T36G RNase A R39G RNase A R41G RNase A N94G RNase A N96G RNase A S24G ONC P74G ONC

Exposed surface area for the (40–95) or (30– 75) disulfide ˚ 2) bond cysteine sulfurs (A

Change from WT RNase A ˚ 2) (A

5.8 23.8 15.5 10.5 10.7 20.6 5.6 5.6 5.6 5.6 26.8 23.5

– 18 9.7 4.7 4.9 14.8 ,0 ,0 ,0 ,0 21 (3 w.r.t. WT ONC) ,18 (0 w.r.t. WT ONC)

The difference between the solvent exposure of the cysteine sulfur atoms in each in silico mutant and the cysteine sulfur atoms in the (40–95) disulfide bond of RNase A is shown.

atoms 40 and 95 of in silico Tyr to Gly mutants of each of those residues (in the vicinity of the (40 – 95) disulfide bond) with the exposure of the cysteine sulfur atoms forming the (40 – 95) disulfide bond in the WT protein. The effect of a few mutations in the vicinity of the (30 –75) disulfide bond of ONC is provided for comparison. The results are listed in Table 3 and considered in the Discussion. Figure 6 shows the effects of the Tyr92-to-Gly mutation on the exposed surface areas of the sulfur atoms of Cys40 and Cys95; Figure 6(a) corresponds to the WT protein, and Figure 6(b) corresponds to the Y92G mutant. Finally, we analyzed the changes in flexibility of the residues in the neighborhood of the (40 – 95) disulfide bond in the P93A mutant (which displays “ONC-like” reductive unfolding behavior;40 see Discussion) and compared it with the change in flexibility of its (65 – 72) disulfide bond as inferred from the temperature factors of the X-ray structures.41 A summary of our findings is provided in the Discussion (and see Figure 7). Reductive unfolding of Y92F and Y92A RNase A The reductive unfolding kinetics of the Y92F35 and Y92A mutants of RNase A were followed using DTTred. The reductive unfolding of Y92A, to form the sole intermediate des [40 – 95] Y92A (Figure 8), was found to be 63-fold faster than that of WT protein. Comparison of the reductive unfolding rate and reductive intermediates of Y92F with those of WT RNase A revealed no significant differences (data not shown).

Discussion The reductive unfolding of ONC The reductive unfolding of ONC was examined by using DTTred as a reducing agent (Figure 1).

Under mildly reducing conditions (10 mM DTTred), the partial reduction of ONC and the formation of an intermediate, Ir, could be followed easily (Figure 2). Mass spectrometry of AEMTS-blocked Ir revealed that it has two free cysteine residues. Thus, we can conclude that Ir is formed by the reduction of one disulfide bond in the native molecule to yield a stable, structured des species (containing three relatively buried disulfide bonds). The rapid partial reduction of ONC to form an intermediate containing three disulfides indicated that the (30 – 75) disulfide bond is highly exposed (or easily exposable by a local unfolding event). Furthermore, no intermediate containing two disulfides is observed in the further reduction of des [30 – 75] (Ir). The absence of an intermediate containing two disulfides could be due either to lack of stability of the two disulfide-containing intermediate(s) that is formed after the reduction of a single disulfide bond in Ir (as in RNase A), or alternatively that an intermediate containing two disulfides is stable, but the rate of formation of such an intermediate (by reduction of Ir) is much lower than the rate of its reduction to 1S and finally R. Ir is a structured des species because it withstands the strongly reducing condition of the experiments. ESI/FTMS mass spectrometry confirmed the results from the proteolytic digestion/ MALDI-TOF experiments, which indicated that Ir is a single-disulfide species, lacking the (30 –75) disulfide bond. Differences in degree of exposed surface areas of the cysteine sulfur atoms in ONC and in RNase A The rates of reduction of disulfide bonds in structured proteins depend upon the individual free energies that are required to expose those disulfide bonds for the reduction process. It is plausible to suppose that disulfides that need global unfolding (relatively larger free energy) to be exposed tend to be buried inside the protein,

802

Reductive Unfolding Pathways of Two Homologues

Figure 6. Stereo view of WT RNase A and its in silico Y92G mutant (a) WT RNase A oriented to show Tyr92 (in red) and the 40 – 95 disulfide bond in yellow (b) The Y92G mutant (Gly in red) showing a larger exposed surface area for the cysteine 40 and cysteine 95 sulfur atom compared to WT protein.

while those disulfides that can be exposed by local unfolding (relatively smaller free energy, which results in a more rapid reduction) tend to be on the surface. Therefore, there may be some degree of correlation between exposed surface-areas of the cysteine sulfur atoms and the free energy required to expose them. In order to assess this assumption, we carried out a molecular surface-area analysis of ONC and RNase A. It showed that the cysteine sulfur atoms of the (30 – 75) disulfide bond in ONC are substantially more exposed compared to the cysteine sulfur atoms in the three other disulfide

bonds (Table 2). The large difference between the contributions to the molecular surface area of the sulfur atoms in the (30 – 75) disulfide bond ˚ 2) and in the next most-exposed disulfide (23.8 A ˚ 2), are in good agreement bond, (87 –104) (5.5 A with the much faster reduction rate of the (30 – 75) disulfide bond in native ONC. A comparison of the exposed surface areas of the sulfur atoms in ONC and RNase A revealed that the cysteine sulfur atoms forming the (30 – 75) disulfide bond in ONC have about fourfold more exposed cysteine sulfur atoms than those forming the (65–72) and (40–95) disulfide bonds in RNase A

Reductive Unfolding Pathways of Two Homologues

803

Figure 7. A, Illustration of how the ring of Tyr92 in the P93A mutant of RNase A is oriented in a manner different from that in the WT protein (Figure 6(b)) (according to the PDB coordinates, no hydrogen bond is present between Tyr92 and Lys37 in this mutant). The temperature factors are colored coded as described in the legend to Figure 5. B, Plot of residue number versus averaged temperature factor (tf) for the full residue for P93A (open bars) and WT RNase A (filled bars).

(Table 2). Such a difference in the exposed surfaceareas might translate into a lower free energy requirement for exposing the (30 – 75) disulfide bond of ONC compared to the free energy requirement for exposing the (65 – 72) and (40 – 95) disulfide bonds of RNase A. In turn, this would account for the faster rate of partial reduction of ONC compared to the partial reduction of RNase A. Structural constraints around the (40– 95) disulfide bond of RNase A In order to understand the structural basis for the fourfold greater degree of exposure of cysteine sulfur atoms in the (30 – 75) disulfide of ONC compared to those in the (40–95) disulfide of RNase A, we compared the residues around cysteine residues 40 and 95 of RNase A with the corresponding residues of the (30 – 75) disulfide bond in ONC. Figure 5A shows that the ring of Tyr92 in

Figure 8. HPLC chromatogram showing the reductive unfolding of Y92A RNase A, 100 minutes after initiation of reduction (100 mM DTTred, 20 mM Tris – HCl (pH 8), 0.2 mM EDTA, 15 8C).

RNase A is oriented to cover a large portion of its (40 – 95) disulfide bond. However, in ONC, Tyr92 is replaced by Arg73, whose side-chain is oriented away from the (30 – 75) disulfide bond, resulting in a large exposed surface-area for the (30 – 75) disulfide (Figure 5B and Table 2). Tyr92 is conserved in 36 of the 41 mammalian pancreatic ribonucleases previously examined,42 with Phe replacing Tyr in four species and Trp substituted for Tyr in one of the species. However, Tyr92 is replaced by an Arg residue in the frog ribonucleases (R. pipiens, Rana catesbeiana and Rana japonica).43 We examined the magnitude of the burial of the 40 and 95 cysteine sulfur atoms by the side-chain of Tyr92 by an in silico conversion of Tyr to Gly. Table 3 shows the effect of such a mutation. Removal of the Tyr92 side-chain causes a very significant change in the exposed surface area of the sulfur atoms of Cys40 and Cys95 (from ˚ 2 in the Y92G ˚ 2 in the WT protein to 15.5 A 5.8 A ˚ 2; see Figure 6(a) and (b)). mutant, a gain of 9.7 A We also mutated (in silico) other residues in the vicinity of the cysteine sulfur atoms of the (40 – 95) disulfide bond to Gly and calculated the exposure of the sulfur atoms of Cys40 and Cys95 in the in silico mutants. These results are summarized in Table 3, which shows the corresponding value for WT ONC and its relevant in silico mutants for comparison. From a comparison of WT RNase A and its mutants, it is clear that Tyr92 makes the largest contribution to the burial of the cysteine sulfur atoms of the (40 – 95) disulfide bond. Interestingly, Thr36, which is conserved in all 41 mammalian pancreatic ribonucleases studied,42 contributes to a ˚2 gain in the burial of the sulfur atoms of 4.9 A (this residue, while absent from ONC, is spatially replaced by Ser24 in the frog variant). A double

804

mutant of RNase A, Y92G/T36G, causes a gain of ˚ 2 in the exposed surface area of the (40 – 95) 14.8 A cysteine sulfur atoms, and the resulting exposed surface area for the cysteine sulfur atoms of the ˚ 2) is (40 – 95) disulfide bond of RNase A (20.6 A very close to the exposed surface area for the corresponding cysteine sulfur atoms of the (30 – 75) ˚ 2). bond of ONC (23.8 A It must be noted that, although there are several other residues in the vicinity of the cysteine sulfur atoms of the (30 –75) bond in ONC, any gain in exposed surface area by mutating such residues ˚ 2 gain in exposed surface area (such as the 3 A obtained by mutating Ser24 to Gly) is minimal and, more importantly, would have no influence on the reductive unfolding pathway, given that the (30 – 75) disulfide bond is already the sole disulfide bond to be reduced rapidly to form a single intermediate containing three disulfides (des [30 –75]). We tested whether the solvent occlusion of the (40 – 95) disulfide bond in RNase A by Tyr92 is responsible for the dissimilarity in the reductive unfolding behavior of the two homologues. Such a test was accomplished by following the reductive unfolding of Y92A RNase A and comparing the data with WT protein. Upon elimination of the bulky aromatic group, the reductive unfolding of the mutant protein was found to be 63-fold faster than that of the WT protein. Moreover, only one intermediate, des [40 –95] Y92A (Figure 8), was found in the reductive unfolding pathway in the mutant (as opposed to two intermediates, des [40 – 95] and des [65 –72], in the WT protein). The increase in solvent accessibility of the (40 – 95) cysteine sulfur atoms in Y92A (Table 3) appears sufficient to result in the sole reduction of only one disulfide bond (i.e. the (40 – 95) disulfide bond) during the first step in its reductive unfolding pathway. In a similar vein, other mutants are being prepared in our laboratory, and reductive unfolding experiments will be carried out on these mutants for a further check of our in silico observations. A mutant of RNase A, in which Pro93 was replaced by Ala (P93A RNase A),33,34 demonstrated a reductive unfolding behavior40 that is very similar to that of ONC, i.e. the reductive unfolding of P93A RNase A is dominated by the presence of only one, very rapidly formed intermediate, des [40 – 95] P93A (formed 120 times faster than the formation of des [40 –95] WT RNase A).13,40 In order to examine the origin of the “ONC-like” reductive unfolding behavior of P93A RNase A, we considered this mutant in our studies. The exposure of the cysteine sulfur atoms forming the (40 – 95) ˚ 2), while less than disulfide bond in P93A (2.5 A that for the (40 –95) disulfide bond in the WT ˚ 2), is consistent with the presence of protein (5.8 A Tyr92. However, while the ring of Tyr92 is still oriented above the (40 – 95) disulfide bond in its X-ray structure,33 the manner of orientation is different from that in the WT protein32 (Figure 7A).

Reductive Unfolding Pathways of Two Homologues

To explain the rapid reduction of the (40 – 95) disulfide in P93A (despite the small exposed surface area of its cysteine sulfur atoms), we examined the changes in flexibility of the (40 – 95) and (65 – 72) disulfide bonds in the mutant, using the same parameter, i.e. the temperature factor, in the WT protein as a reference (Figure 7B). The effects of the mutation are strongest in the region of the (40 – 95) disulfide bond of the mutant (Figure 7B, open bars). While the effects of the mutation are felt in other regions of the protein, they are relatively more diminished compared to the region around the mutation. For example, while there is an increased change in flexibility in the (65 – 72) disulfide bond of P93A (Figure 7B, open bars), the magnitude is less than that in the (40 – 95) disulfide bond. The temperature factors for the WT protein are shown for reference purposes (Figure 7B, filled bars). The dramatic increase in the flexibility of the (40 – 95) disulfide in P93A may be explained as follows. It is known from the X-ray structure32 of WT RNase A that Pro93 is involved in ring– ring stacking interactions with Y92, and constrains the flexibility of this residue,44 and the loop containing these residues. When replaced by Ala, both the ring –ring stacking interaction and the rigidity of the loop (containing these residues) are lost, and hence there is a large change in flexibility of the (40 – 95) disulfide bond. The preferential lack of rigidity of the (40 – 95) disulfide bond compared to the (65 – 72) disulfide bond in P93A appears to be sufficient to bias the local unfolding free energy requirements in favor of exposing the (40 – 95) disulfide bond for reduction to such an extent that only one dominant intermediate (des [40 – 95] P93A) is seen. (From the rate constant40 for the reduction of the (40 – 95) disulfide bond in P93A, the calculated free energy for exposing this bond is 4.0 kcal/mol, as opposed to 6.6 kcal/mol for the same bond in WT RNase A: 1 cal ¼ 4.184 J.) Furthermore, a closer inspection of the structure of the WT protein (Figure 5A) revealed that the hydroxyl group of the aromatic moiety hydrogen bonds with the backbone carbonyl group of Lys37.32 From Figure 7A and an examination of the relevant distances (the Oh[Y92] – O[K37] ˚ in RNase A to 5.4 A ˚ distance changes from 2.7 A in the P93A mutant), it is clear that the hydrogen bond between the Tyr92 side-chain and the backbone carbonyl group of Lys37 in WT RNase A (Figure 5A) is missing from P93A (Figure 7A) due to the slightly different orientation of the Tyr92 side-chain (perhaps due to the loss of ring stacking interactions with Pro93 and the increase in the flexibility of the loop containing these residues). The effect of this hydrogen bond was tested on the reductive unfolding kinetics of RNase A by the use of a Y92F mutant.35 The kinetics and unfolding pathways were identical with those of the WT protein, which indicates that the lack of the hydrogen bond (described above) does not seem to affect the orientation of the aromatic

805

Reductive Unfolding Pathways of Two Homologues

side-chain or the rigidity around the (40 – 95) disulfide bond.

Table 4. Reduction rate constant, and solventaccessibility of cysteine sulfur atoms, of the first disulfide bond to be reduced in various proteins (using DTTred as the reducing agent)

Conclusion In conclusion, we demonstrate here that reductive unfolding studies of homologous proteins are fruitful in revealing the structural determinants of their reductive unfolding pathways. The reductive unfolding of ONC begins with the reduction of the (30 – 75) disulfide bond to form des [30 –75]. The formation of this single intermediate by the sole reduction of the (30 –75) disulfide bond appears feasible, given the relatively large exposed surface area of the sulfur atoms of Cys30 and Cys75 compared to the exposed surface areas of its other cysteine sulfur atoms (Cys19, Cys48, Cys68, Cys87, Cys90 and Cys104), which would translate into a lower free energy cost for exposing the (30 – 75) disulfide bond compared to the other three disulfide bonds, (19 – 68), (48 –90) and (87 – 104). All three disulfides of des [30 – 75] appear to be relatively buried, which would explain the slow rate of reduction of des [30 –75] compared to its rate of formation. No intermediate is seen in the reduction of des [30 –75]. Furthermore, a comparison of the residues in the vicinity of the (30 – 75) disulfide bond of ONC with those of the homologous (40 – 95) disulfide bond of WT RNase A suggested that the presence of Tyr92 in RNase A (which is replaced by Arg73 in ONC) may be responsible for the different reductive unfolding pathways of the two mutants. Removal of Tyr (in a Y92A mutant) resulted in an “ONClike” reductive unfolding behavior of the mutant bovine variant indicating that the bulky aromatic residue plays an important role in hindering the reduction of the (40–95) disulfide bond in RNase A. In ONC, Arg73 is situated away from the (30 – 75) disulfide bond, resulting in a large exposed surface area for its cysteine sulfur atoms in contrast to Tyr92 in RNase A, which buries the analogous (40 – 95) disulfide bond. The large exposed surface area of the cysteine sulfur atoms of the (30 – 75) disulfide bond of ONC compared to the exposed surface area of the cysteine sulfur atoms of its other disulfide bonds is consistent with the rapid reduction of the (30 – 75) disulfide. The very fast reduction rate of one disulfide compared to other disulfides in a protein containing multi-disulfides leads to the formation of a single intermediate, and demonstrates the role of environment on the reactivity of a covalent bond.18 Our analysis of the P93A mutant of RNase A indicates that Tyr92 is held rigidly in WT RNase A because of Pro93 (ring stacking interactions and flexibility constraints). Its absence translates into a preferential increase in the flexibility of the (40 – 95) disulfide (over the (65 – 72) disulfide bond), facilitating the exposure of this bond for reduction. Thus, only one intermediate, des

Proteins a-Lactalbumin BPTI ONC RNase A a b c d

Disulfide bond that is the first to be reduced

Exposed surface ˚ 2) area (A

Reduction rate constant (min-1 M-1)

(6–120) (14–38) (30–75) (40–95) (65–72)

29.5 27.7 23.8 5.8 5.2

4.56(^0.05) £ 104a 2.3 £ 102b 2.3 ^ 0.2c 6.8(^0.5) £ 1023d 1.8(^0.1) £ 1023d

At pH 8.5 and 25 8C, from Kuwajima et al.8 At pH 8.7 and 25 8C, from Creighton.52 At pH 8.0 and 15 8C, from Xu et al.53 At pH 8.0 and 15 8C, from Li et al.13

[40 – 95] P93A is observed in its reductive unfolding. Importantly, our study provides a structural understanding for the difference in the reductive unfolding pathways and kinetics of two homologous proteins. The exposed surface areas of cysteine sulfur atoms (above a certain value) may be critical in determining the kinetics and reductive unfolding pathways of proteins. The reductive unfolding pathways of a-lactalbumin and BPTI are similar to those of ONC and P93A, in that a single fast-forming species is the first intermediate to be formed in these proteins.8,9,14 Both BPTI and a-lactalbumin possess one disulfide bond with far larger exposed surface area compared to their other disulfide bonds (the (14 –38) disulfide bond in BPTI has an exposed surface˚ 2 and the (6 – 120) disulfide bond in area of 27.7 A a-lactalbumin has an exposed surface-area of ˚ 2), which is very easily reduced by an 29.5 A external reducing agent (Table 4)).8,9,14 Conservation (or lack of) key residues around the homologous disulfide bonds may be critical factors in the reductive unfolding pathways of the proteins.18

Materials and Methods Materials WT Onconase cDNA in a pet11 expression vector was kindly provided by R. J. Youle.19 The cDNA was amplified by PCR and cloned to a pet22b(þ) vector in-frame with the pelB signal sequence without the starting methionine residue. WT ONC was expressed in BL21 cells and purified as described;45 the level of expression was compared to that from another wild-type plasmid of ONC (pONC) which was kindly provided by R. T. Raines.20 Conversion of the N-terminal glutamine residue of the uncyclized residue to pyro-glutamic acid was carried out by dialyzing the folded protein against either 0.1 M HCl or 200 mM phosphate buffer (pH 7) at room temperature for two days and checked by mass spectrometry. The protein was purified by

806 cation-exchange HPLC as described,45 and lyophilized. Y92F RNase A was obtained from a previous study.35 AEMTS (. 99% pure) was purchased from Anatrace and used without further purification. DTTred was obtained from Sigma and used without further purification. All other chemicals were of the highest grade commercially available. Native RNase A was obtained from Sigma and purified as described.30 Reductive unfolding of ONC Lyophilized native ONC was dissolved in 50 mM acetic acid at room temperature to obtain a stock solution (5 mg/ml), and aliquots of this solution were stored frozen at 2 20 8C. Reduction experiments were initiated by introducing one such aliquot (0.2 ml) of ONC into a solution (20 mM Tris – HCl (pH 8), 0.2 mM EDTA, 15 8C) containing DTTred (final concentration 10 mM) such that the final protein concentration was 0.5 mg/ml. Aliquots of the reaction mixture were withdrawn periodically, and free thiol groups in each aliquot were blocked immediately with excess AEMTS (final concentration 40 mM).5 After five minutes, the pH of the mixture was reduced to 3 by addition of 20 ml of glacial acetic acid (a necessary step, consistent with the procedure followed when working with RNase A, in order to prevent it from being deamidated at higher pH46). The AEMTS-blocked, acid-quenched samples were then diluted by addition of a fivefold excess of 1 mM acetic acid, and analyzed by HPLC. The HPLC runs were carried out on a Rainin-Hydropore (SCX) cation-exchange column using a salt gradient (50 mM – 150 mM NaCl in 150 minutes). Reduction of ONC was carried out also at a higher concentration of the reducing agent (100 mM DTTred, 20 mM Tris – HCl (pH 8), 0.2 mM EDTA, 15 8C). Additionally, control reduction experiments on RNase A were carried out with 10 mM DTTred (20 mM Tris – HCl (pH 8), 0.2 mM EDTA, 15 8C). Identity of the intermediate appearing in the partial reductive unfolding pathway of ONC Only two peaks were found in the HPLC chromatogram from the reduction of ONC using 10 mM DTTred. One of these two peaks co-elutes with a control sample that contained only WT ONC. Nevertheless, both peaks were collected separately and desalted on a reversedphase C-18 column. Water and organic solvents were removed by a Speed-vac. The lyophilized sample was dissolved in water/methanol/acetic acid (50:48:2 by vol.) and electrosprayed at 1 – 50 nl/minute with a nanospray emitter. Mass spectra were acquired on a 6 T modified Finnigan FTMS as described.47 Fragmentation was achieved by infrared multiphoton dissociation (IRMPD) and nozzle skimmer dissociation. The MS/MS spectra are averages of 20 – 100 scans. Assignment of the fragment masses and compositions were made with the computer program THRASH.48 After each mass value, the mass difference (in units of 1.00235 Da) between the most abundant isotopic peak and the monoisotopic peak is denoted in italics (Figure 4). Isolation of the (unblocked) intermediate, Ir, appearing in the reduction pathway of ONC ONC (1 mg/ml) was reduced with 100 mM DTTred (100 minutes reduction time, 20 mM Tris – HCl (pH 8),

Reductive Unfolding Pathways of Two Homologues

0.2 mM EDTA, 15 8C). The reaction mixture was quenched by addition of glacial acetic acid and the (unblocked) sample was introduced onto a reversedphase for HPLC separation. Two peaks were observed (one minor in area as compared to the other peak). Since the elution time of ONC on this reversed-phase system was determined previously, it was possible to identify ONC in the reduction mixture as the minor peak, appearing 100 minutes after initiation of the reduction process. The major peak appearing in the reversed-phase HPLC column was collected, and a small volume of the collected sample was blocked with AEMTS. Mass spectrometry and SCX-HPLC confirmed that this was the same intermediate that appears in the reductive unfolding of ONC using 10 mM DTTred. Any remaining (unblocked) intermediate was lyophilized (to remove water and organic solvent) and then dissolved in acetic acid (1 mM, 0.5 mg/ml final concentration) prior to freezing at 2 20 8C, to be used in a future investigation.

Assignment of disulfide bonds in Ir Two independent methods were used to determine the disulfide-bond distribution in Ir. In the first method, the sample was subjected to proteolytic digestion, followed by mass spectrometric analysis of the HPLC-separated peptide fragments, as described.26 Trial experiments demonstrated that Ir is very resistant to proteolytic digestion. Therefore, an alternative method was used to facilitate enzymatic digestion. First, unblocked Ir was carboxymethylated by iodoacetamide (100 mM, pH 8, room temperature, 20 minutes, in the dark). This procedure blocks primarily the free cysteine residues of this intermediate.49 The carboxymethylated Ir was desalted using reversed-phase HPLC and all solvents were removed by lyophilization. Then, the carboxymethylated sample was fully-reduced using a high concentration of DTTred (100 mM reductant, 100 mM Tris – HCl (pH 8), 1 mM EDTA, 25 8C) under denaturing conditions (6 M guanidinium hydrochloride (GdnHCl)). The two previously alkylated cysteine residues are unaffected by this procedure. Finally, the remaining six thiol groups were blocked using excess AEMTS. The AEMTS-blocked carboxymethylated Ir was desalted using reversed-phase HPLC and lyophilized prior to enzymatic digestion. The digestion conditions were similar to those used previously in our laboratory for determining the disulfide-bond distributions of unstructured ensembles that are populated during the oxidative folding of RNase A.26,27 AEMTS-blocked carboxymethylated Ir was subjected to chymotrypsin digestion at pH 8 and 37 8C. The digestion buffer contained 2 M GdnHCl, 100 mM Tris – HCl, 1 mM EDTA, and the ratio of enzyme to substrate was , 1/50. The sample was quenched to pH 2 with 5% (v/v) trifluoroacetic acid (TFA) after two hours. The digested mixture was fractionated by reversedphase HPLC with the detector set at 222 nm. The fractionated fragments were collected separately, lyophilized, and analyzed further using a MALDI-TOF mass spectrometer. The masses of the digested fragments were then matched to the theoretically obtained digest masses in order to deduce their identities. In the second method, AEMTS-blocked carboxymethylated Ir was subjected to peptide mapping using ESI/FTMS as described above.

Reductive Unfolding Pathways of Two Homologues

807

Exposed surface area calculations for ONC and RNase A, and a comparison of residues around RNase A, ONC, and P93A RNase A

3. Welker, E., Narayan, M., Wedemeyer, W. J. & Scheraga, H. A. (2001). Structural determinants of oxidative folding in proteins. Proc. Natl Acad. Sci. USA, 98, 2312– 2316. 4. Welker, E., Wedemeyer, W. J., Narayan, M. & Scheraga, H. A. (2001). Coupling of conformational folding and disulfide-bond reactions in oxidative folding of proteins. Biochemistry, 40, 9059– 9064. 5. Rothwarf, D. M. & Scheraga, H. A. (1993). Regeneration of bovine pancreatic ribonuclease A. 1. Steadystate distribution. Biochemistry, 32, 2671– 2679. 6. Creighton, T. E. (1985). The problem of how and why proteins adopt folded conformations. J. Phys. Chem. 89, 2452– 2459. 7. Goldenberg, D. P. (1988). Kinetic analysis of the folding and unfolding of a mutant form of bovine pancreatic trypsin inhibitor lacking the cysteine-14 and -38 thiols. Biochemistry, 27, 2481– 2489. 8. Kuwajima, K., Ikeguchi, M., Sugawara, T., Hiraoka, Y. & Sugai, S. (1990). Kinetics of disulfide bond reduction in a-lactalbumin by dithiothreitol and molecular basis of superreactivity of the Cys6Cys120 disulfide bond. Biochemistry, 29, 8240– 8249. 9. Ewbank, J. J. & Creighton, T. E. (1993). Pathway of disulfide-coupled unfolding and refolding of bovine a-lactalbumin. Biochemistry, 32, 3677– 3693. 10. Mendoza, J. A., Jarstfer, M. B. & Goldenberg, D. P. (1994). Effects of amino acid replacements on the reductive unfolding kinetics of pancreatic trypsin inhibitor. Biochemistry, 33, 1143– 1148. 11. Creighton, T. E. (1994). The protein folding problem. In Mechanisms of Protein Folding (Pain, R. H., ed.), pp. 1 – 25, Oxford University Press, New York. 12. Yamashita, H., Nakatsuka, T. & Hirose, M. (1995). Structural and functional characteristics of partially disulfide-reduced intermediates ovotransferrin N lobe. Cystine localization by indirect end-labeling approach and implications for the reduction pathway. J. Biol. Chem. 270, 29806– 29812. 13. Li, Y.-J., Rothwarf, D. M. & Scheraga, H. A. (1995). Mechanism of reductive protein unfolding. Nature Struct. Biol. 2, 489– 494. 14. Ma, L.-C. & Anderson, S. (1997). Correlation between disulfide reduction and conformational unfolding in bovine pancreatic trypsin inhibitor. Biochemistry, 36, 3728 –3736. 15. Chang, J.-Y. (1997). A two-stage mechanism for the reductive unfolding of disulfide-containing proteins. J. Biol. Chem. 272, 69 – 75. 16. Singh, R. R. & Rao, A. G. A. (2002). Reductive unfolding and oxidative refolding of a Bowman– Birk inhibitor from horsegram seeds (Dolichos biflorus): evidence for “hyperreactive” disulfide bonds and rate-limiting nature of disulfide isomerization in folding. Biochim. Biophys. Acta, 1597, 280–291. 17. Yan, Y.-B., Zhang, R.-Q. & Zhou, H.-M. (2002). Biphasic reductive unfolding of ribonuclease A is temperature dependent. Eur. J. Biochem. 269, 5314 –5322. 18. Laskowski, M. & Scheraga, H. A. (1956). Thermodynamic considerations of protein reactions. II. Modified reactivity of primary valence bonds. J. Am. Chem. Soc. 78, 5793– 5798. 19. Boix, E., Wu, Y.-N., Vasandani, V. M., Saxena, S. K., Ardelt, W., Ladner, J. & Youle, R. J. (1996). Role of the N terminus in RNase A homologues: differences

The contributions to the molecular surface areas of cysteine sulfur atoms of ONC, RNase A, and the P93A mutant of RNase A were computed using the Connolly algorithm,37 as implemented in the InsightII program (Accelrys Inc.)†. The conformations of RNase A, ONC, and the P93A mutant of RNase A correspond to their three-dimensional structures determined by X-ray crystallography and available in the RCSB PDB38,39 (PDB codes 7RSA, 1ONC, and 1A5Q). In the case of ONC, hydrogen atoms were absent from the PDB files and were added to the structures using the program InsightII, before the surface areas were computed. To compare the residues around cysteine residues 40 and 95 of RNase A (7RSA), and P93A (1A5Q) with the residues around cysteine residues 30 and 75 of ONC (1ONC), their three-dimensional structures were examined with the visualization program Rasmol50,51 and InsightII. Reductive unfolding of Y92F35 and Y92A RNase A These mutants were prepared as described.45 The free energy of unfolding was determined as described,45 and was found to be comparable to the WT protein. The lyophilized mutants were dissolved separately in 50 mM acetic acid at room temperature to obtain stocks solution (5 mg/ml), and aliquots of these solutions were stored frozen at 2 20 8C. Reduction experiments were carried out as described above. A control experiment to determine the kinetics of reduction of WT RNase A under identical conditions was carried out using 0.5 mg/ml of native RNase A.

Acknowledgements We thank Dr N. Manoj for help with Figure 7(B) and Drs Stanislaw Oldziej and Jorge Vila for helpful discussions. This work was supported by the National Institute of General Medical Sciences of the National Institutes of Health (grant nos GM-24893, GM-16609). Support was received also from the National Foundation for Cancer Research. Part of this research was carried out with the resources of the Cornell Theory Center, which receives funding from Cornell University, the State of New York, Federal agencies, Foundations, and Corporate partners. K.B. acknowledges funding from the Austrian Science Foundation (FWF grant P15767).

References 1. Narayan, M., Welker, E., Wedemeyer, W. J. & Scheraga, H. A. (2000). Oxidative folding of proteins. Accts Chem. Res. 33, 805– 812. 2. Wedemeyer, W. J., Welker, E., Narayan, M. & Scheraga, H. A. (2000). Disulfide bonds and protein folding. Biochemistry, 39, 4207– 4216.

† http://www.accelrys.com/insight/

808

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32. 33.

34.

in catalytic activity, ribonuclease inhibitor interaction and cytotoxicity. J. Mol. Biol. 257, 992 –1007. Leland, P. A., Schultz, L. W., Kim, B.-M. & Raines, R. T. (1998). Ribonuclease A variants with potent cytotoxic activity. Proc. Natl Acad. Sci. USA, 95, 10407– 10412. Ardelt, W., Mikulski, S. M. & Shogen, K. (1991). Amino acid sequence of an anti-tumor protein from Rana pipiens oocytes and early embryos—homology to pancreatic ribonucleases. J. Biol. Chem. 266, 245– 251. Saxena, S. K., Sirdeshmukh, R., Ardelt, W., Mikulski, S. M., Shogen, K. & Youle, R. J. (2002). Entry into cells and selective degradation of tRNAs by a cytotoxic member of the RNase A family. J. Biol. Chem. 277, 15142– 15146. Mosimann, S. C., Ardelt, W. & James, M. N. G. ˚ X-ray crystallographic structure (1994). Refined 1.7 A of P-30 protein, an amphibian ribonuclease with antitumor activity. J. Mol. Biol. 236, 1141– 1153. Leland, P. A., Staniszewski, K. E., Kim, B.-M. & Raines, R. T. (2000). A synapomorphic disulfide bond is critical for the conformational stability and cytotoxicity of an amphibian ribonuclease. FEBS Letters, 477, 203– 207. Notomista, E., Catanzano, F., Graziano, G., Dal Piaz, F., Barone, G., D’Alessio, G. & Di Donato, A. (2000). Onconase: an unusually stable protein. Biochemistry, 39, 8711 – 8718. Xu, X., Rothwarf, D. M. & Scheraga, H. A. (1996). Nonrandom distribution of the one-disulfide intermediates in the regeneration of ribonuclease A. Biochemistry, 35, 6406– 6417. Volles, M. J., Xu, X. & Scheraga, H. A. (1999). Distribution of disulfide bonds in the two-disulfide intermediates in the regeneration of bovine pancreatic ribonuclease A: further insights into the folding process. Biochemistry, 38, 7284– 7293. Wedemeyer, W. J., Xu, X., Welker, E. & Scheraga, H. A. (2002). Conformational propensities of protein folding intermediates: distribution of species in the 1S, 2S, and 3S ensembles of the [C40A, C95A] mutant of bovine pancreatic ribonuclease A. Biochemistry, 41, 1483– 1491. Shin, H.-C., Narayan, M., Song, M.-C. & Scheraga, H. A. (2003). Role of the [65 – 72] disulfide bond in oxidative folding of bovine pancreatic ribonuclease A. Biochemistry, 42, 11514– 11519. Rothwarf, D. M., Li, Y.-J. & Scheraga, H. A. (1998). Regeneration of bovine pancreatic ribonuclease A: identification of two native like three-disulfide intermediates involved in separate pathways. Biochemistry, 37, 3760– 3766. Rothwarf, D. M., Li, Y.-J. & Scheraga, H. A. (1998). Regeneration of bovine pancreatic ribonuclease A: detailed kinetic analysis of two independent folding pathways. Biochemistry, 37, 3767– 3776. Wlodawer, A., Svensson, L. A., Sjo¨lin, L. & Gilliland, G. L. (1988). Structure of phosphate-free ribonuclease ˚ . Biochemistry, 27, 2705– 2717. A refined at 1.26 A Pearson, M. A., Karplus, P. A., Dodge, R. W., Laity, J. H. & Scheraga, H. A. (1998). Crystal structures of two mutants that have implications for the folding of bovine pancreatic ribonuclease A. Protein Sci. 7, 1255– 1258. Xiong, Y., Juminaga, D., Swapna, G. V. T., Wedemeyer, W. J., Scheraga, H. A. & Montelione, G. T. (2000). Solution NMR evidence for a cis TyrAla peptide group in the structure of [Pro93Ala]

Reductive Unfolding Pathways of Two Homologues

35.

36.

37. 38.

39. 40.

41.

42. 43. 44. 45.

46.

47.

48.

49. 50. 51.

bovine pancreatic ribonuclease A. Protein Sci. 9, 421– 426. Juminaga, D., Wedemeyer, W. J., Gardun˜o-Juarez, R., McDonald, M. A. & Scheraga, H. A. (1997). Tyrosyl interactions in the folding and unfolding of bovine pancreatic ribonuclease A: a study of tyrosine-tophenylalanine mutants. Biochemistry, 36, 10131– 10145. Ge, Y., Lawhorn, B. G., ElNaggar, M., Strauss, E., Park, J.-H., Begley, T. P. & McLafferty, F. W. (2002). Top down characterization of large proteins (45 kDa) by electron capture dissociation mass spectrometry. J. Am. Chem. Soc. 124, 672–678. Connolly, M. L. (1983). Analytical molecular surface calculation. J. Appl. Crystallog. 16, 548– 558. Bernstein, F. C., Koetzle, T. F., Williams, G. J. B., Meyer, E. F., Jr, Brice, M. D., Rodgers, J. R. et al. (1977). The Protein Data Bank: a computer-based archival file for macromolecular structures. J. Mol. Biol. 112, 535– 542. Berman, H. M., Westbrook, J., Feng, Z., Gilliland, G., Bhat, T. N., Weissig, H. et al. (2000). The Protein Data Bank. Nucl. Acids Res. 28, 235– 242. Cao, A., Welker, E. & Scheraga, H. A. (2001). Effect of mutation of proline 93 on redox unfolding/folding of bovine pancreatic ribonuclease A. Biochemistry, 40, 8536– 8541. Trueblood, K. N., Bu¨rghi, H.-B., Burzlaff, H., Dunitz, J. D., Gramaccioli, C. M., Schulz, H. H. et al. (1996). Atomic displacement parameter nomenclature. Report of a subcommittee on atomic displacement parameter nomenclature. Acta Crystallog. sect. A, 52, 770– 781. Beintema, J. J., Schu¨ller, C., Irie, M. & Carsana, A. (1988). Molecular evolution of the ribonuclease superfamily. Prog. Biophys. Mol. Biol. 51, 165– 192. Leland, P. A. & Raines, R. T. (2001). Cancer chemotherapy—ribonucleases to the rescue. Chem. Biol. 8, 405– 413. MacArthur, M. W. & Thornton, J. M. (1991). Influence of proline residues on protein conformation. J. Mol. Biol. 218, 397– 412. Laity, J. H., Lester, C. C., Shimotakahara, S., Zimmerman, D. E., Montelione, G. T. & Scheraga, H. A. (1997). Structural characterization of an analog of the major rate-determining disulfide folding intermediate of bovine pancreatic ribonuclease A. Biochemistry, 36, 12683– 12699. Thannhauser, T. W. & Scheraga, H. A. (1985). Reversible blocking of half-cystine residues of proteins and an irreversible specific deamidation of asparagine-67 of S-sulforibonuclease under mild conditions. Biochemistry, 24, 7681– 7688. Beu, S. C., Senko, M. W., Quinn, J. P., Wampler, F. M. & McLafferty, F. W. (1993). Fourier-transform electrospray instrumentation for tandem high-resolution mass-spectrometry of large molecules. J. Am. Soc. Mass Spectrom. 4, 557– 565. Horn, D. M., Zubarev, R. A. & McLafferty, F. W. (2000). Automated reduction and interpretation of high resolution electrospray mass spectra of large molecules. J. Am. Soc. Mass Spectrom. 11, 320– 332. Creighton, T. E. (1986). Disulfide bonds as probes of protein folding pathways. Methods Enzymol. 131, 83 – 106. Saqi, M. A. S. & Sayle, R. (1994). Pdbmotif—a tool for the automatic identification and display of motifs in protein structures. Comput. Appl. Biosci. 10, 545– 546. Sayle, R. A. & Milnerwhite, E. J. (1995).

Reductive Unfolding Pathways of Two Homologues

Rasmol—biomolecular graphics for all. Trends Biochem. Sci. 20, 374– 376. 52. Creighton, T. E. (1975). Interactions between cysteine residues as probes of protein conformation: the disulphide bond between Cys-14 and Cys-38 of the pancreatic trypsin inhibitor. J. Mol. Biol. 96, 767– 776.

809

53. Xu, G., Narayan, M., Welker, E. & Scheraga, H. A. (2004). Characterization of the fast-forming intermediate, des [30 –75], in the reductive unfolding of onconase. Biochemistry, 43, 3246 –3254.

Edited by C. R. Matthews (Received 11 December 2003; received in revised form 27 February 2004; accepted 2 March 2004)