Probing the unfolding region of ribonuclease A by site-directed mutagenesis Jens Ko¨ ditz, Renate Ulbrich-Hofmann and Ulrich Arnold Department of Biochemistry and Biotechnology, Martin-Luther University Halle-Wittenberg, Halle, Germany Ribonuclease A contains two exposed loop regions, around Ala20 and Asn34. Only the loop around Ala20 is sufficiently flexible even under native conditions to allow cleavage by nonspecific proteases. In contrast, the loop around Asn34 (together with the adjacent b-sheet around Thr45) is the first region of the ribonuclease A molecule that becomes sus- ceptible to thermolysin and trypsin under unfolding condi- tions. This s econd region therefore has been sugges ted to b e involved in early steps of unfolding and was designated as the unfolding region of the ribonuclease A molecule. Con- sequently, modifications in this region should have a great impact on the unfolding and, thus, on the thermodynamic stability. Also, i f the Ala20 loop contributes to the stability o f the ribonuclease A molecule, rigidification of this flexible region should stabilize the entire protein molecule. We sub- stituted several residues in both regions without any dra- matic effects on the native conformation and catalytic activity. As a result of their remarkably differing stability, the variants fell into two groups carrying the mutations: (a) A20P, S21P, A20P/S21P, S21L, or N34 D; (b) L35S, L35 A, F46Y, K31A/R33S, L35S/F46Y, L35A/F46Y, or K31A/ R33S/F46Y. The first group showed a thermodynamic and kinetic stability similar to wild-type ribonuclease A, whereas both stabilities of the variants in the second group were greatly decreased, suggesting that the decrease in DG can be mainly attributed to an increased unfolding rate. Although rigidification of the Ala20 loop by introduction of proline did not result in stabilization, disturbance of the network of hydrogen bonds and hydrophobic interactions that interlock the proposed unfolding region dramatically destabiliz ed the ribonuclease A molecule. Keywords: limited proteolysis; local unfo lding; protein engineering; ribonuclease A; stability. In contrast with the ensemble of conformational species in the unfolded state [1], the native state of proteins is generally characterized by a uniform overall global conformation [2]. Whereas larger proteins t hat consist of structural subunits or domains often behave very complexly during the processes of unfolding and refolding, most small proteins can be considered as a single unit [3]. Apart from local fluctuations of the p rotein str ucture in t he native state, the protein molecule unfolds highly co-operatively when exposed to denaturing conditions. The stability of the natively folded protein molecule is not determined by a single feature, but a number of internal and external factors contribute to the formation and stabilization of the native protein structure [4]. Studies on a large variety of proteins led t o t he assumption that confined regions of the 3 D protein s tructure are crucial for the conservation o f its folded, native state. A local disruption of the most labile region, which was referred to as unfolding region, was postulated to initiate the co-operative unfolding of the protein molecule [5,6]. This assumption was supported by the identification of a region in the neutral protease from Bacillus stearothermophilus that responds most sensitively to changes in the amino-acid composition by site-directed mutagenesis [7–9]. Consistent with this Ôcritical regionÕ,a Ôweak pointÕ in Arthrobacter D -xylose isomerase was postu- lated based on results from proteolysis experiments under thermal denaturation [10]. More recently, Machius et al. [11] deduced a Ôweak regionÕ in a-amylase and Gaseidnes et al. [12] identified a Ôweak spotÕ or a Ônucleation site for unfoldingÕ in chitinase by mutational analysis. Because of the decreased number of hydrogen bonds, loo p regions at the surface of the protein molecule are candidates for such Ôu nfolding regionsÕ. In fact, loops that are tethered by irregular hydrogen bonds or hydrophobic patches were found to be crucial for either the folding or unfolding of proteins (for a review see [13]). Ribonuclease A (RNase A) is a small, compact, and rather stable enzyme which is cross-linked by four disulfide bonds [14]. Nevertheless, even under native conditions the loop region around Ala20 is highly flexible [15], which leads to proteolytic cleavage by non-specific proteases. In con- trast, in spite of increased mobility detected for residues 37–42 by NMR [ 16], the fl exibility of the loop around Asn34, which contains potential cleavage sites for trypsin and thermolysin, is obviously not sufficient to allow Correspondence to U. Arnold, Department of Biochemistry and Bio- technology, Martin-Luther University Halle-Wittenberg, Kurt- Mothes Strasse 3, 06120 Halle, Germany. Fax: +49 3 455527303, Tel.: +49 3455524865, E-mail: arnold@biochemtech.uni-halle.de Abbreviations: 6-FAM-dArU(dA) 2 -6-TAMRA, 6-carboxyfluorescein- dArU(dA) 2 -6-carboxytetramethylrhodamine; GdnHCl, guanidine hydrochloride; RNase A, ribonuclease A. Enzymes: bovine pancreatic ribonuc lease A (EC 3.1.27.5); thermolysin (EC 3.4.24.27) Note: a website is available at h ttp://www.biochemtech.uni-halle.de/ biotech (Received 30 June 2004, revised 27 August 2004, accepted 3 September 2004) Eur. J. Biochem. 271, 4147–4156 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04355.x cleavage by these proteases. As soon as the protein molecule starts to unfold globally, however, the RNase A molecule becomes susceptible to these proteases, too. The primary cleavage sites were found in the loop region around Asn34 and the adjacent b-strand around Thr45, suggesting this section as the unfolding region [17]. To investigate the contribution of the two loop regions to the stability of the entire RNase A molecule, we replaced several amino-acid residues b y site-directed mutagenesis (A20P, S21P, A20P/S21P, S21L, N34D, L35S, L35A, F46Y, K31A/R33S, L35A/F46Y, L35S/F46Y, and K31A/ R33S/F46Y) and studied the effect of the mutations on the thermodynamic and kinetic stability. The similarity of the impact on both the thermodynamic and kinetic stabilities suggests a predominant effect on the native state by these mutations. Experimental procedures Proteins and chemicals RNase A from Sigma (St Louis, MO, USA) was purified on a Mono S FPLC column (Amersham Biosciences, Uppsala, Sweden). Thermolysin (from Calbiochem, Schwalbach, Germany) was used without further purification. Oligonu- cleotides were from MWG Biotech (Ebersberg, Germany) and restriction enzymes AvaI, BsmI, EcoRI, HindIII, NdeI, and SacI w ere f rom N ew England Biolabs (Frankfurt/ Main, Germany). Growth media were from Difco Laboratories (Detroit, MI, USA). Escherichia coli strains XL-1 Blue and BL21(DE3) were from Stratagene (Heidelberg, Germany). 6-Carboxyfluorescein-dArU(dA) 2 - 6-carboxytetramethylrhodamine (6-FAM-dArU(dA) 2 -6- TAMRA) was purchased from Integrated DNA Technologies (Coralville, IA, USA). All other chemicals were of purest grade commercially available. Site-directed mutagenesis The RNase A variants A20P, S21P, and A20P/S21P had been produced previously [18]. For other variants, the rnase A gene in the plasmid pBXR [19], a gift from R. T. Raines (University of Wisconsin, Madison, WI, USA), was modified by use of the Q uikChange TM site-directed mutagenesis kit (Stratagene) to obtain the mutations N34D, K31A/R33S, L35A, L35S, and F46Y. For the mutations K31A/R33S/F46Y, L35A/F46Y, a nd L35S/F46Y, site-directed mutagenesis was started from the rnase A genes that carry the mutations for K31A/R33S, L35A, or L35S using the oligonucleotides for the F46Y mutation. The oligonucleotides and the introduced restric- tion sites to facilitate the selection of positive clones are shown in Table 1. The mutations were verified by DNA sequencing as described by Sanger et al. [20] (SequiTherm- Excel TM LongRead TM DNA sequen cing kit, Biozym, Hess, Oldendorf, Germany, and Li-COR 4000 DNA-sequencer, MWG Biotech, Ebersberg, Germany). The p lasmids carrying the correct DNA sequence were each transformed into E. coli expression host strain BL21(DE3). Expression, renaturation, and purification of the enzyme variants The experimental procedure was performed as described previously [18]. Briefly, cultures of E. coli strain BL21(DE3) that had b een t ransformed with a plasmid directing the expression of the corresponding RNase A variant were g rown in terrific broth containing 50 lgÆmL )1 kanamycin [variants A 20P, S 21P, and A20P/S21P in vector pET 26b(+)] or 400 lgÆmL )1 ampicillin [the other variants i n vector pET 22b(+)] at 37 °CtoanA 600 of 2. Gene expression was induced by 1m M isopropyl thio-a- D -galactoside, and cells were grown a dditionally for 4 h before being harvested. Cell lysis was performed by treatment with lysozyme and homogenization with a Gaulin homogenizer. T he inclu- sion bodies were isolated by centrifugation followed by resolubilization ( 20 m M Tris/HCl, 7 M guanidine h ydro- chloride (GdnHCl), 100 m M dithiothreitol, 1 0 m M EDTA, pH 8.0) and dialysis of the protein solution against 20 m M acetic acid. Precipitates formed during dialysis were removed by centrifugation. After renatura- tion of the protein [100 m M Tris/HCl, pH 8.5, 100 m M NaCl, 1 m M glutathione (reduced), 0.2 m M glutathione ( oxidized), 10 m M EDTA at room tempera- ture for 24 h], it was purified on a Mono S column (50 m M Tris/HCl, pH 7.5, with a linear gradient of 0–500 m M NaCl). Table 1. Oligonucleotides for site-directed mutagenesis. The replaced nucleotides are bold-face and the introduced restriction sites are underlined. Mutation Oligonucleotides Restriction site S21L fw 5¢-C TCC AGC ACT TCC GCC GCC CT G AGC TCC AAC TAC TG3¢ SacI rev 5¢-CA GTA GTT G GA GCT CAG GGC GGC GGA AGT GCT GGA G-3¢ N34D fw 5¢-C CAG ATG ATG AAG AGC CGG GAC CTG ACC AAA GAT CGA TGC-3¢ No restriction site rev 5¢-GCA TCG ATC TTT GGT CAG GTC CCG GCT CTT CAT CAT CTG G-3¢ K31A/R33S fw 5¢-C TGT AAC CAG ATG ATG GC G AGC TCG AAC CTG ACC AAA GAT C3¢ SacI rev 5¢-G ATC TTT GGT CAG GTT C GA GCT CGC CAT CAT CTG GTT ACA G-3¢ L35A fw 5¢-G ATG ATG AAG AGC CG GAAT GCC ACC AAA GAT CGA TGC AAG C-3¢ BsmI rev 5¢-G CTT GCA TCG ATC TTT GGT G GC ATT CCG GCT CTT CAT CAT C-3¢ L35S fw 5¢-G ATG ATG AAG AGC CG GAAT TCC ACC AAA GAT CGA TGC AAG C-3¢ EcoRI rev 5¢-G CTT GCA TCG ATC TTT GGT G GA ATT CCG GCT CTT CAT CAT C-3¢ F46Y fw 5¢-GC AAG CCA GTG AAC A CA TAT GTG CAC GAG TCC CTG G-3¢ NdeI rev 5¢-C CAG GGA CTC GTG CA CATA TGT GTT CAC TGG CTT GC-3¢ 4148 J. Ko ¨ ditz et al.(Eur. J. Biochem. 271) Ó FEBS 2004 Determination of the protein concentration The prote in concentration of RNase A and the F46Y-free variants was determined using the molar absorption coef- ficient of 9800 M )1 Æcm )1 at 278 nm [21]. For the F46Y-containing RNase A variants, a molar absorption coefficient of 11 300 M )1 Æcm )1 at 278 nm, determined as described by Thannhauser et al.[22],wasused. For activity measurements, the concentration of the RNase stock solutions was determined by use of the BCA protein assay kit (Pierce, Bonn, Germany) with BSA as calibration standard according to the instructions of the manufacturer. The absorbance of the samples was measured at 560 nm after incubation at 37 °C for 30 min using a micro plate reader MR 7000 (Dynatech, Denkendorf, Germany). Activity assay Values of k cat /K m of wild-type RNase A and its variants were determined at 25 °Cin100m M Mes/NaOH, pH 6.0, containing 100 m M NaCl, 50 n M 6-FAM-dArU(dA) 2 - 6-TAMRA, and 0.2 5–0.5 ngÆmL )1 RNase A as described by Kelemen et al. [23]. The increase in fluorescence emission at 515 nm (band width 10 nm), on excitation at 490 nm (bandwidth1nm),wasfollowedina1· 0.4 cm fluores- cence cuvette using a Fluoro-Max-2 spectrometer (Jobin Yvon, Grasbrunn, Germany). The values of k cat /K m were determined using the follow- ing equation: k cat =K m ¼ m v ðF end À F start Þ½E where m v is the initial velocity calculated from the linear increase in the flu orescence signal, F start is the signal of the substrate before the addition of enzyme, F end is the signal after cleavage of all substrate, and [E] is the concentration of RNase A . CD spectroscopy CD spectra of RNase A and its variants were recorded in 10 m M Tris/HCl, pH 8.0, containing 1–2 mgÆmL )1 RNase on a CD s pectrometer 62 A DS (Aviv, Piscat away, NJ, USA) at 25 °C. Cuvettes of 1 cm and 0.01 cm path length were used for CD spectroscopy in the near-UV region (250– 340 nm) and in the far-UV region (200–260 nm), respect- ively. GdnHCl-induced transition curves GdnHCl-induced transition curves of RNase A and its variants were obtained by fluorescence spectroscopy on a Fluoro-Max-2 spectrometer (Jobin Yvon) at 25 °Cusing a c uvette of 1 · 0.4 cm. Protein concentration was 50 lgÆmL )1 in 50 m M Tris/HCl, pH 8.0, containing 0–6 M GdnHCl. After equilibration, the fluorescence signal was recorded at 303 nm and averaged over 40 s. The band width was 1 nm for excitation at 278 nm and 10 nm for emission. To calculate values of [D] 50% (the concentration of denaturant [D] at which 50% of the protein is denatured) and m DG (the measure of the dependence of t he free energy on denaturant concentration) the linear function, DG ½D ¼ DG water À m DG ½D was used where DG [D] is the free energy of unfolding at a given denaturant c oncentration, and DG water is the free energy of unfolding in the absence of denaturant [24]. The fluorescence signals y were fitted by nonlinear regression using the program SIGMA PLOT as described by Santoro & Bolen [25] with the modification by Clarke & Fersht [26], DG ½D ¼ m DG ð½D 50% À½DÞ leading to the equation; y ¼ ðy N 0 þ m N ½DÞ þ ðy D 0 þ m D ½DÞ exp ðÀm DG ð½D 50% À½DÞ RT 1 þ expð Àm DG ð½D 50% À½D RT Þ where y N 0 and y D 0 are the intercepts, and m N and m D the slopes in the pre-transition and post-transition region, respectively, in the y vs. [D] graph. The fraction of native protein (f N ) was calculated from the fi tted values using equation; f N ¼ y D À y y D À y N with y D ¼ y D 0 + m D [D]andy N ¼ y N 0 + m N [D], where y N and y D are the signals of the native and the denatured protein as a function of the denaturant concentration. The effect of mutations on the free energy was calculated as described by Clarke & Fersht [26], DDG ½D ¼ DG ½D À DG 0½D ¼ m DG ð½D 50% À½DÞ À m 0 DG ð½D 0  50% À½DÞ where DDG [D] is the change i n the free energy on mutation at a defined concentration of denaturant, DG [D] , m DG ,and[D] 50% are the values for wild-type RNase A, and DG¢ [D] , m¢ DG ,and[D¢] 50% are the values of the respective variant. Thermal transition curves Thermal transition c urves of wild-type, F46Y-RNase A, and L35A/F46Y-RNase A were obtained by measuring the absorbance at 287 nm and 25–80 °C a fter equilibration using a U-2000 spectrophotometer (Hitachi, Tokyo, Japan) and a water-jacketed cuvette (1 cm) connected to a W K14 thermostat (Colora, Lorch, Germany). The protein concen- tration was 0.5–1.0 mgÆmL )1 in 50 m M Tris/HCl, pH 8.0, containing 100 m M NaCl. The signal of absorbance w as fitted a s described by Santoro & Bolen [25] to obtain the transition midpoint T m . By use of the van’t Hoff equation, dðln K D Þ dð1=TÞ ¼À DH R DH m , the free enthalpy at T m , was calculated. Usingavalueof9.4kJÆK )1 Æmol )1 for DC p of wild-type RNase A [27], DG T can be calculated by the Gibbs– Helmholtz equation; Ó FEBS 2004 Unfolding region of ribonuclease A ( Eur. J. Biochem. 271) 4149 DG T ¼ DH m 1 À T T m  À DC p ðT m À TÞþT ln T T m  Values of DDG 25°C were estimated by the relation DDG T ¼ DG T À DG 0T where DG T and DG¢ T are the values of wild-type RNase A and its variant, respectively. Proteolysis Proteolysis was carried out at 35.0–57.5 °C with final concentrations of 0.1 mgÆmL )1 wild-type RNase A or its variants and 0.2 mgÆmL )1 thermolysin in 50 m M Tris/ HCl, pH 8.0, containing 1 m M CaCl 2 .Inatypical experiment, 20 lL thermolysin solution (2 mgÆmL )1 in 50 m M Tris/HCl, pH 8.0, containing 10 m M CaCl 2 )were mixedwith160lL50m M Tris/HCl, pH 8.0, and equilibrated at a defined t emperature. The reaction was started by addition of 20 lL RNase solution (1 mgÆmL )1 in 50 m M Tris/HCl, pH 8.0). After defined time intervals, samples of 20 lL were withdrawn, mixed immediately with 5 lL50m M EDTA, dried under nitrogen, and analyzed by SDS/PAGE. SDS/PAGE and determination of the rate constants of unfolding ( k U ) Electrophoresis was carried out on a Midget Electro- phoresis Unit (Hoefer, San Francisco, CA, USA) as described by L aemmli [28] using 10% and 15% acryl- amide for stacking and separating gels, respectively. The gels we re stained with Coomassie B rillant Blue R 250. After b eing s tained, the gels w ere evaluated at 5 60 nm using a densitometer CD 60 (Desaga, Heidelberg, Germany). The rate constants o f proteolysis (k p ) were calculated from the decrease in the peak areas of the intact RNase band as a function of time of proteolysis, which followed pseudo-first-order kinetics. The determ ination of k p was performed at least twice. If the protease can degrade the unfolded protein only and the unfolding reaction is the rate-limiting step for proteolysis, as was the case in our experiments, k p corresponds to the rate constant of unfold- ing, k U [27,29]. From the linear function ln(k U /T)vs.1/T in the Eyring plot, the k U values at 25 °C were calculated. On the basis of the Eyring equation, DG # U ¼ RT ln K h  À ln k U T  where DG # U , K, h,andR are free activation e nergy for the unfolding reaction, Boltzmann’s, Planck’s, and the gas constant, t he change in activation e nergy for unfolding DDG # U on mutation is given by, DDG # U ¼ DG # U À DG 0# U ¼ RT ln k 0 U k U where k¢ U is the rate constant for the variant and k U the rate constant of the wild-type enzyme [26]. Results Design of the RNase A variants Analysis by use o f t he program FIRST [30] (http:// firstweb.asu.edu/) indicates the highest flexibility of the peptide backbone of native RNase A at the N-terminus and in the loop region between helices I and II (around Ala20), followed by the region from the end of helix II spanning the loop to the adjacent b-strand (Lys31–Phe46; Fig. 1), which had a lso shown low stability i n both refolding [31] and unfolding [17] experiments. We replaced various amino-acid residues as both single and multiple mutations in the two loop regions to investigate their contribution to the overall stability of the RNase A molecule. To maintain RNase A folding and activity, we refrained from dramatic interference with the protein structure such as charge-reversal mutations or the intro- duction or deletion of disulfide bonds. In the A la20 loop region the exposed, proteolytically sensitive residues A la20 and/or Ser21 were replaced b y proline t o rigidify the flexible loop. As a control, Ser21 was replaced by leucine to introduce a cleavage site for thermolysin, which was also used to determine the unfolding rate constants of RNase A. Thus, both the local unfolding of this loop (vi a the cleavage a t Ala20–Leu21) and t he global unfolding of the RNase A molecule (via the cleavage at Asn34–Leu35/Thr45–Phe46) can be detected. In the proposed unfolding region, we selected residues with side chains found to be involved in intramolecular interactions (analysis using t he program WHAT IF [32]) ( Table 2). Lys31, Arg33, Leu35, and Phe46 (Fig. 1) were replaced as both single and multiple mutations (K31A/R33S, L35S, L35A, F46Y, L35S/F46Y , L35A/ F46Y, and K31A/R33S/F46Y). As the crystal structure reveals, the side chains of these residues a re involved in intramolecular interactions that form either a hydrophobic patch (Leu35 and Phe46 with Met29 and Met30; Fig. 2A) or a hydrogen bond network (Arg33, Fig. 2B). Furthermore, these residues had proven to be crucial in the proteolytic degradation of the RNase A molecule on unfolding [17]. A s a control for replacing Fig. 1. Tertiary structure of RNase A. The model (7rsa) was taken from the Brookhaven protein databank and drawn with Swiss PDB - VIEWER v3.7. The replaced residues are marked in red for the region around Ala20 and green for the proposed unfolding region (Lys31– Phe46). 4150 J. Ko ¨ ditz et al.(Eur. J. Biochem. 271) Ó FEBS 2004 solvent-exposed amino-acid residues, Asn34 was replaced by aspartate. Expression, renaturation, and purification All RNase A variants were expressed as inclusion bodies. Even though they differed in their tendency to form aggregates during renaturation, all variants could be obtained in sufficient amounts (up to 30 mgÆL )1 culture medium). The purified proteins proved to be homogeneous by SDS/PAGE and rechromatography on a Mono S column. CD spectra As detected by CD spectroscopy, a ll RNase A variants revealed a tertiary and secondary structure comparable to that of wild-type RNase A (not shown) with a marginal disturbance of the secondary structure in A20P/S21P- RNase A. An increased signal in the CD spectra in the near-UV region of F46Y containing RNase A variants is attributed to the introduction of the additional tyrosine. Activity Enzymatic activity provides a sensitive measure of the impact of modifications on the native structure of an enzyme [33]. The k cat /K m values for RNase A and its variants, determined w ith 6-FAM-dArU(dA) 2 -6-TAMRA as substrate [23], revealed that all RNase A variants are active (Table 3). However, while the RNase A variants with mutations in the Ala20 loop region a s well as N 34D- RNase A and L35A-RNase A showed an activity compar- able to that of wild-type RNase A, the variants with mutations in the unfolding region (except for N34D and L35A) showed a more s ignificant decrease in the k cat /K m values, with the lowest activity (% 20%) for L35S/F46Y- RNase A and L35A/F46Y-RNase A. Thermodynamic stability To study the effect of the mutations on the thermodynamic stability o f t he RNase A molecule, G dnHCl-induced Table 2. Relative solvent accessibility o f amino acid residues of wild-type RNase A. The relative accessibility was calculated using the program WHAT IF [32]andrelatestheaccessibilityofthesidechainoftheresidueintheprotein to the accessibility in a G ly-XXX-Gly peptide in vacuu m w hich is a good approximation for the accessibility in the unfolded state of the protein. Residue Relative accessibility (%) Known side chain interactions and effects by modification Met29 18 Hydrophobic core with Met30, Leu35, and Phe46 Met30 0 Hydrophobic core with Met29, Leu35, and Phe46 Lys31 76 No interactions; K31C slightly decreases T m [44] Ser32 66 No interactions; S32C slightly decreases T m [44] Arg33 23 H bonds with the backbone of Arg10 and Met13 Asn34 48 No interactions; attached carbohydrate moiety in the related RNase B increases T m by 1.5 °C [27] Leu35 9 Hydrophobic core with Met29, Met30, and Phe46 Thr36 12 No interactions but in proximity to Met30, Tyr97, and the disulfide bond Cys40–Cys95 Lys37 60 No interactions Asp38 68 No interactions; D38R decreases T m by 4 °C [52] Arg39 69 No interactions Cys40 11 Disulfid bond with Cys95; C40A/C95A decreases T m by 20 °C [53] Lys41 21 P1 subsite; H bond to the side chain of Asn44; K41R strongly decreases the activity but does not affect T m [54]; a chemical crosslink K7–K41 increases both DG (H2O) and DG # U by about 12 kJ mol )1 [51] Pro42 43 No interactions; P42A does not affect the thermodynamic stability [55] Val43 44 No interactions Asn44 3 H bond with Gln11 and Lys41 Thr45 8 B1 subsite; T45G decreases T m by 10 °C [56] Phe46 0 Hydrophobic core with Met29, Met30, and Leu35; exchange by Leu, Val, Glu, Lys, or Ala greatly decreases DG (H2O) [42,43] Fig. 2. Tertiary structure of the unfolding region of wild-type RNase A. The model (7rsa) was taken from the Br ookhaven protein databank and drawn with Swiss PDB - VIEWER v3.7. (A) Hydrophobic cluster formed by residues Phe46, Leu35, Met29 and Met30. The ribbon at positions Phe46 and Leu35 is marked in green. (B) Hydroge n b onds between the side ch ains o f Arg33 and the backbone of Met13 and Arg10 and the hyd rogen bo nd be tween t he side ch ain of Arg10 and the backbone of Arg33. The hydrophobic residues are marked as green balls. Ó FEBS 2004 Unfolding region of ribonuclease A ( Eur. J. Biochem. 271) 4151 transition curves were recorded (Fig. 3). The values for [D] 50% and m DG as well as the change in free energy by the mutation at the transition midpoint of wild-type RNase A DDG [D]50% were determined (Table 3). All mutations in the A la20 loop as well as the control N34D did not significantly affect the GdnHCl-induced unfolding so that the transition c urves of t hose RNase A variants resemble t hat of w ild-type RNase A (group I; Fig. 3) with a mean value for [D] 50% of 2.85 ± 0.10 M (Table 3). In contrast, a ll other mutations in the proposed unfolding region resulted in a considerable decrease in the transition midpoint (group II). Interestingly, the variants K31A/R33S, L35S, L35A, and F46Y show a remarkably coincident decrease in the thermodynamic stability ([D] 50% ¼ 1.8 5 ± 0.10 M ; Table 3, Fig. 3) whereas the variants obtained by the combination of destabilizing mutations in the unfolding region (K31A/R33S/F46Y, L35S/F46Y, and L35A/F46Y) are characterized by a further slight but uniform decrease in stability ([D] 50% ¼ 1.59 ± 0.10 M ; Table 3, Fig. 3). A similar destabilizing effect by the mutations was observed in thermal tran sition curves determined for wild- type RNase A and the variants F46Y and L35S/F46Y (not shown). Values of T m were 62.0 ± 0.1 °C, 53.0 ± 0.1 °C, and 48.0 ± 0.1 °C, respectively, corresponding to values of DDG 25°C of 10.4 ± 1.5 kJÆmol )1 and 23.6 ± 1.4 kJÆ mol )1 caused by the mutations F4 6Y and L35A/F36Y (calculat ed with DC p ¼ 9.4 kJÆK )1 Æmol )1 for wild-type RNase A [27] and the DH m values of 544 ± 14 kJÆmol )1 , 481±8kJÆmol )1 ,and341±5kJÆmol )1 for wild-type RNase A and the variants F46Y and L35S/F46Y, respect- ively, obtained from the van’t Hoff plot). Kinetic stability The decreased thermodynamic s tability of the RNase A variants with mutations in the proposed unfolding region could arise from faster unfolding or slower refolding (or both). To dissect the effect of the mutations, rate constants of unfolding of wild-type RNase A and its variants were determined. Owing to the isomerization of natively cis proline peptide bonds in the unfolded state [34,35] refolding of RNase A is known to be rather complex [36] and the introduction of further proline residues at positions 20 and 21 is expected to further increase this complexity, as indicated by the decreased m DG values f or the A20P and A20P/S21P variants (Table 3). Hence, we refrained from refolding experiments. Unfolding rate c onstants w ere d etermined b y limited proteolysis with thermo lysin at three different temperatures between 35.0 °C and 57.5 °C (Fig. 4). By linear extrapola- tion of the Eyring plots [29], v alues of k U at 25 °Cwere obtained which were used to calculate values of DDG # U at 25 °C (Table 3). Wild-type RNase A and all variants of group I with respect to their thermodynamic stability unfold with the same rate constants at 47.5–57.5 °C (Fig. 4), Table 3. Activity and thermodynamic and kinetic parameters of wild-type RNase A and its variants at 25 °C. Values of k cat /K m were determined as described in E xperimental Procedures with 6-FAM-dArU(dA) 2 -6-TAMRA as substrate i n 100 m M Mes/NaOH, pH 6.0, containing 100 m M NaCl. The thermodynamic parameters were determined from the GdnHCl-induced transition curves at 25 °C as described in Experimental Procedures. Values of DDG # U (25 °C) were calculated from parameters obtained from the Eyring plot (Fig. 4) as described in Experimental Procedures. RNase A variant 10 )7 · k cat /K m (s )1 Æ M )1 ) [D] 50% ( M ) m DG (kJÆmol )1 Æ M )1 ) DDG [D]50% (kJÆmol )1 ) DDG # U 25  C (kJÆmol )1 ) Wild-type 3.9 ± 0.7 2.79 ± 0.03 13.7 ± 1.6 – – A20P 2.4 ± 0.4 2.89 ± 0.03 10.2 ± 0.9 ) 1.0 ± 0.5 0.3 ± 1.0 S21P 3.5 ± 0.3 2.90 ± 0.02 13.0 ± 0.8 ) 1.4 ± 0.5 ) 0.6 ± 0.4 S21L 2.5 ± 0.2 2.78 ± 0.02 16.4 ± 1.8 0.2 ± 0.5 1.8 ± 0.4 A20P/S21P 2.8 ± 0.4 2.82 ± 0.02 11.1 ± 0.8 ) 0.3 ± 0.5 ) 1.0 ± 1.6 N34D 2.9 ± 0.1 2.86 ± 0.02 13.1 ± 1.2 ) 0.9 ± 0.5 ) 0.9 ± 0.6 L35A 3.1 ± 0.3 1.93 ± 0.02 10.8 ± 0.6 9.3 ± 0.7 5.9 ± 0.5 L35S 0.9 ± 0.2 1.84 ± 0.02 16.1 ± 1.9 15.3 ± 1.9 10.2 ± 0.4 F 46Y 1.8 ± 0.1 1.88 ± 0.02 11.9 ± 0.9 10.8 ± 0.9 7.1 ± 0.4 K31A/R33S 1.2 ± 0.2 1.92 ± 0.02 13.5 ± 1.3 11.7 ± 1.2 13.1 ± 0.2 L35A/F 46Y 0.8 ± 0.1 1.59 ± 0.01 14.1 ± 0.6 16.9 ± 0.8 11.7 ± 0.4 L35S/F 46Y 0.8 ± 0.2 1.51 ± 0.02 12.8 ± 1.1 16.4 ± 1.5 12.6 ± 0.2 K31A/R33S/F 46Y 1.2 ± 0.3 1.64 ± 0.02 13.4 ± 0.9 15.4 ± 1.1 16.3 ± 0.3 [GdnHCl] (M) 012345 f N 0.0 0.5 1.0 Fig. 3. GdnHCl-induced transition curves. The transition curves of wild-type RNase A (teal) and its variants A20P (black), S21P (grey), S21L (bright green), A20P/S21P (blue), N34D (red), L35A (cyan), L35S (green), F46Y (dark red), K31A/R33S (pink), L35A/F46Y (dark yellow), L35S/F46Y (dark blue), and K31A/R33S/F46Y (violet) were determined by fluorescence spectroscopy in 50 m M Tris/ H Cl , pH 8 .0 , at 25 °C. 4152 J. Ko ¨ ditz et al.(Eur. J. Biochem. 271) Ó FEBS 2004 indicating that the kinetic stability is also not affected by these mutations. All the thermodynamically less stable RNase A variants also show a large increase in k U .Even though the effects are not as uniform as for the thermo- dynamic stability, the comparison of DDG [D]50% and DDG # U 25  C shows that the decrease in the thermodynamic stability is mainly caused by an increase in the unfolding rate constant, i.e. a decrease in the kinetic stability. Whereas the introduction of a cleavage site for thermolysin in the control variant S21L-RNase A facilitated degradation of the RNase A molecule under native conditions (not shown), this variant was degraded like wild-type RNase A under denaturing conditions. Discussion As in the folding of proteins, confined regions of the protein structure have a crucial role in the unfolding process and are, thus, particularly important for kinetic stability [8,9,12]. These r egions are m ostly located on the surface of the protein molecule, and loops in particular often represent critical spots [8,11,17]. RNase A possesses two structural sections that might function as such a critical region (Fig. 1): (a) the loop region around Ala20, which is highly flexible under native condi- tions [14,15] as reflected in efficient proteolytic attack by nonspecific proteases such as proteinase K and subtilisin Carlsberg [37–39]; (b) the region from the end of helix II to the a djacent a-sheet (Lys31–Ph e46), which becomes access- ible to an H–D exchange [40] and to t he proteases thermolysin and trypsin when the molecule starts to unfold [17]. Furthermore, this region (residues 31–39) is the last one that becomes protected against tryptic attack during RNase A folding [31]. RNase A variants with amino-acid substitutions in the two regions fell into two classes w ith respect to thermo- dynamic stability (Fig. 3). The RNase A variants with similar unfolding transition curves to wild-type RNase A (group I) are obtained by mutations in the loop region around Ala20 or by the control mutation N34D. These amino-acid residues are not involved in interactions like hydrogen bonds, salt b ridges or hydrophobic clusters, as reflected in great flexibility of the loop region around Ala20 [15,16]. So even the replacement of two adjoined residues i n this r egion by proline (A20P/S21P) was tolerated. On the other hand, the introduction of the proline residues, i.e. the decrease in loop flexibility, did not increase the global stability of the RNase A molecule. By introducing a cleavage site for thermolysin (S21L- RNase A), the flexibility of the Ala20 loop became traceable for this protease. Nevertheless, the unfolding rate constants of this RNase A variant correspond to those of wild-type RNase A (Fig. 4), indicating that the local unfolding of the Ala20 loop is independent of the global unfolding of the RNase A molecule. In the control variant Asn34-RNase A, a s olvent-exposed residue that belongs to the unfolding region (Lys31–Phe46) and serves as anchor for the stabil- izing carbohydrate moiety in the related RNase B [41], was replaced. As expected, the mimicked deamidation does not affect interactions essential for stability. In contrast, the less stab le RNase A variants of group II (L35S, L35A, F46Y, K31A/R33S, L35S/F46Y, L35A/ F46Y, and K31A/R33S/F46Y) all of which were obtained by mutations in the region Lys31–Phe46 indicate a consid- erable contribution of this region to the thermodynamic stability of the entire RNase A molecule. The coincidence of the degree o f destabilization in these variants points to an effect on the stability of the entire region rather than on a particular interaction. A similar d estabilization was also found by Chatani et al. [42] and Kadonosono et al. [43] by replacement of Phe46 with Leu, Val, Ala, Lys, or Glu. The authors concluded that Phe46 has a n important role in the folding reaction through hydrophobic interactions and by the correct packing of the amino-acid side ch ains between two structural domains [42]. However, from the rate of oxidative protein folding, they concluded t hat there was a decreased k U , i.e. kinetic stabilization of the F46L, F46V, and F46A variants. In contrast, we found an acceleration of the unfolding reaction, i.e. kinetic destabilization, for the variants and the similarity of DDG [D]50% and DDG # U indicates that the decrease in the thermodynamic stability is mainly caused by an in crease in k U . Furthermore, our results suggest that Leu35, the side chain of which is buried in the interior of the molecule like that of Phe46 (Table 2), is involved in the formation of a hydrophobic cluster with Phe46, Met29 and Met30 (Fig. 2A) and consequently plays a similar role to Phe46. Molecular modeling revealed that any mutation in position 35 d estabilizes the e ntire molecule by disturbing these complex hydrophobic inter- actions (G. Vriend, University of Nijmegen, personal communication). In addition to these hydrophobic interactions, this region is stabilized by a network of hydrogen bonds between the side chain of Arg33 and the backbone of Met13 and Arg10 (three hydrogen bonds) and between the side chain of Arg10 and the backbone of Arg33 (one hydrogen bond; Fig. 2B). Because no hydrogen bonds were identified for the side chain of Lys31 of RNase A (analysis using the program WHAT IF [32]) and its exchange with Cys results in only a 1000 / T (K -1 ) 3.00 3.05 3.10 3.15 3.20 3.25 ln (k U / T) -14 -12 -10 Fig. 4. Eyring plot for the unfolding of wild-type RNase A and its vari- ants. Values for k U of wild-type RNase A (teal) and its variants A20P (black), S21P (grey), S21L (bright green), A20P/S21P (blue), N34D red), L35A (cyan), L35S (green), F46Y (dark red), K31A/R33S (pink), L35A/F46Y (dark yellow), L35S/F46Y (dark blue), and K31A/R33S/ F46Y (violet) were determined by limited proteolysis with thermolysin in 50 m M Tris/HCl, pH 8.0, at 35.0–57.5 °C as described in Experi- mental procedures. Ó FEBS 2004 Unfolding region of ribonuclease A ( Eur. J. Biochem. 271) 4153 slight decrease in the stability [44], the d estabilizing effect of the mutation K31A/R33S is probably caused by t he mutation of Arg33. Generally, changes in the thermodynamic s tability by mutations can be caused by effects on t he native and/or the unfolded state, whereas changes in the kinetic stability are due to a change in the native and/or transition state. The determination of the unfolding rate constants of wild-type RNase A and its variants (Fig. 4) allowed differentiation between the several possibilities. The RNase A variants with GdnHCl-induced transition curves similar to that of wild-type R Nase A, i.e. t he members of g roup I, also show thermal unfoldin g rate constants and consequently DG # U values comparable to that of wild-type RNase A (Table 3), indicating that the native state, relative to the transition state, is not affected by the mutations. The labile RNase A variants show a large increase in the unfolding rate constants. For the variants K31A/R33S and K31A/R33S/ F46Y, a value of DDG # U was obtained that corresponds very well to that of DDG [D]50% (Table 3), indicating that the decrease in the thermodynamic stability is caused by destabilization of the native state relative to th e unfolded state. The decrease in the thermodynamic s tability of the other less s table v ariants, all of which were exclusively obtained by exchanges of the hydrophobic residues Leu35 and/or Phe46, is not solely attributable to faster unfolding. The differences between DDG [D]50% and DDG # U also point to slower refolding, e.g. by disturbance of the formation of a hydrophobic cluster [42]. Nevertheless, the decrease in the thermodynamic stability is m ainly c ause d by t he faster unfolding resulting from destabilization of the native state relative to the transition state, underlining the predominant importance of this region for maintaining the natively folded structure of the RNase A molecule. Interestingly, the hydrophobic nature of residues 29, 30, 35, and 46 is conserved throughout the members of the ribonuclease A superfamily (Fig. 5). While Phe46 and Met30 (numbered by the RNase A sequence) are found in all members, Met29 and Leu35 can be occupied by Met, Ile, or Ala and Leu, Met, or Ile, respectively (Fig. 5, cf [45]). Furthermore, with the exception of mammalian ribonuc- leases 2 and frog ribonucleases, the charged residue Arg33 is conserved (Fig. 5). Altogether, whereas the loop region between helices I and II, i.e. around Ala20, does not contribute to the stability of the RNase A molecule and local flexibility does not lead to global unfolding, the interface between helix II and the adjacent a-sheet is stabilized by a multitude of interactio ns and is very sensitive to mutations. Connecting regions between different folding motifs have also been found to be crucial for the stability of other proteins [46–49]. Despite a vast number o f RNase A variants p roduced by protein engineering (for a review, see [50]), only two variants concerning this region are more stable than wild-type RNase A: the naturally occurring glycosylated RNase B (at Asn34 [41]) and the chemically cross-linked RNase A (Lys7–Lys41) [51]. F or both variants, the thermodynamic stabilization is comparable to the kinetic stabilization [27,51]. Also the effect of the mutations reported here on the thermodynamic stability can mainly be attributed to effects o n the kinetic stability o f the protein, providing further evidence for the validity of the concept of t he unfolding region. Acknowledgements We are grateful to Professor R. T. Raines (University of Wisconsin, Madison, WI, USA) for the gift of the plasmid pBXR, to Professor G. Vriend (University of Nijmegen, the Netherlands) for molecular modeling, and to Y. Markert for providing the plasmids for the variants A20P, S21P, and A20P/S21P. J. K. was supported by a grant from the Max-Buchner-Forschungsstiftung, Frankfurt, Germany. References 1. Shortle, D. (1996) The denatu red state (the oth er half of the folding equation) and its role in protein stability. FASEB J. 10, 27–34. 2. Daggett, V. & Fersht, A.R. (2003) Is there a unifying mechanism for protein folding? Trends Biochem. Sci. 28, 18–25. 3. Jackson, S.E. (1998) How do small single-d omain prot eins fold? Fold. Des. 3, R81–R91. 4. Jaenicke, R. (1999) Stability and folding of domain proteins. Prog. Biophys. Mol. Biol. 71, 155–241. 5. Schellenberger, A. & Ulbrich, R. (1989) Protein stabilization by blocking the native unfolding nucleus. Biomed. Biochim. Acta 48, 63–67. 6. Ulbrich-Hofmann, R., Arnold, U. & Mansfeld, J. (1999) The concept of the un folding region for approaching the mechanisms of enzyme stabilization. J. Mol. Catal. B Enzym. 7, 125–131. 7. Vriend, G. & Eijsink, V. (1993) Prediction and analysis of struc- ture, stability and unfolding of thermolysin-like proteases. J. Comput. Aided Mol. Des. 7, 367–396. 8. Eijsink, V.G., Veltman, O.R., Aukema, W., Vriend, G. & Venema, G. (1995) Structural determinants of the stability of thermolysin-like proteinases. Nat. Struct. Biol. 2, 374–379. Fig. 5. Alignment of the amino-acid sequence (residues 1–58) for members of the RNase A superfamily. RNase A (P00656) from bovine pancreas, RNase 1 (P07998) from human pancreas, and bs RNase (P00669) from bovine semen belong to the mammalian ribonucleases 1 (pancreatic type). Human EDN (P10153, eosinophil-derived neurotoxin) belongs to the mammalian (nonsecretory or neurotoxin-type) ribonucleases 2, human RNase 4 (P34096) belongs to the mammalian ribonucleases 4, human angiogenin (P03950) belongs to the mammalian angiogenins, and onconase (P22069) from Rana pipiens oocytes belongs to frog ribonucleases [45]. Residues are numbered according to RNase A. Ident ical residues and sim ilar residues are u nderlaid in black and grey, respectively. Deletions in the sequences are indicated by $. 4154 J. Ko ¨ ditz et al.(Eur. J. Biochem. 271) Ó FEBS 2004 9. Mansfeld, J., Vriend, G., Van den Burg, B., Eijsink, V.G. & Ulbrich-Hofmann, R. (1999) Probing the unfolding region in a thermolysin-like protease by site-specific immobilization. Biochemistry 38, 8240–8245. 10. Siddiqui, K.S., Rangarajan, M., Hartley, B.S., Kitmitto, A., Panico, M., Blench, I.P. & Morris, H.R. (1993) Arthrobacter D -xylose isomerase: partial proteolysis with thermolysin. Biochem. J. 289, 201–208. 11. Machius, M., Declerck, N., Huber, R. & Wiegand, G. (2003) Kinetic stabilization of Bacillus licheniformis alpha-amylase through introduction of hydrophobic residues at the surface. J. Biol. Chem. 278, 11546–11553. 12. Gaseidnes, S., Synstad, B., Jia, X., Kjellesvik, H., Vriend, G . & Eijsink, V.G. (2003) Stabilization of a chitinase from Serratia marcesce ns by Gly fi Ala and Xxx fi Pro mutations. Protein Eng. 16, 841–846. 13. Fetrow, J.S. (1995) Omega loops: nonregular secondary struc- tures significant in protein function and stability. FASEB J. 9, 708–717. 14. Wlodawer, A., Bott, R. & Sjolin, L. (1982) The refined crystal structure of ribonuclease A at 2.0 A ˚ resolution. J. Biol. Chem. 257, 1325–1332. 15. Hubbard, S.J., Beynon, R.J. & Thornton, J.M. (1998) Assessment of conformational parameters as pre dictors of limited proteolytic sites in native protein structures. Protein Eng. 11, 349–359. 16. Santoro, J., Gonzalez, C., Bruix, M., Neira, J.L., Nieto, J.L., Herranz, J. & Rico, M. (1993) High-resolution three-dimensional structure o f ribonuclease A in solution by nu clear magnetic resonance spectroscopy. J. Mol. Biol. 22 9, 722–734. 17. Arnold, U., Ru ¨ cknagel, K.P., Schierhorn, A. & U lbrich- Hofmann, R. (1996) Thermal unfolding and proteolytic suscept- ibility of ribonuclease A. Eur. J. Biochem. 237, 862–869. 18. Markert, Y., Ko ¨ ditz, J., Ulbrich-Hofmann, R. & Arnold, U. (2003) Proline versus charge concept for protein stabilization against proteolytic attack. Pr otein Eng. 16, 1041–1046. 19. delCardayre ´ , S.B., Ribo, M., Yokel, E.M., Quirk, D.J., Rutter, W.J. & Raines, R.T. (1995) Engineering ribonuclease A: production, purification and characterization of wild-type enzyme and mutants at Gln11. Protein Eng. 8, 261–273. 20. Sanger, F., Nicklen, S. & Coulson, A.R. (1977) DNA sequencing with chain-terminating inhibitors. Proc. Natl Acad. Sci. USA 74, 5463–5467. 21. Sela, M. & Anfinsen, C.B. (1957) Some spectrophotometric and polarimetric experiments with ribonuclease. Biochim. Biophys. Acta 24, 229–235. 22. Thannhauser, T.W., Konishi, Y. & Scheraga, H.A. (1984) Sensi- tive quantitative analysis of disulfide bonds in polypeptides and proteins. Anal. Biochem. 138, 181–188. 23. Kelemen, B.R., Klink, T.A., Behlke, M.A., Euban ks, S .R., Leland, P.A. & Raines, R.T. (1999) Hypersensitive substrate for ribonucleases. Nucleic Acids Res. 27, 3696–3701. 24. Pace, C.N. (1986) Determination and analysis of u rea and guanidine hydrochloride de naturation curves. I n Methods in Enzymology (Hirs, C.H.W. & Timasheff, S.N., eds), pp. 266–280. Academic Press, New York. 25. Santoro, M.M. & Bolen, D.W. (1988) Unfolding free energy changes determined by the linear extrapolation method. 1. Unfolding of phenylmethanesulfonyl alpha-chymotrypsin using different denaturants. Biochemistry 27, 8063–8068. 26. Clarke, J. & Fersht, A.R. (1993) Engineered disulfide bonds as probes of the folding pathway of barnase: increasing the stability of proteins against the rate of denaturation. Biochemistry 32, 4322–4329. 27. Arnold, U. & Ulbrich-Hofmann, R. (1997) Kinetic and thermo- dynamic thermal stabilities of ribonuclease A and ribonuclease B. Biochemistry 36, 2166–2172. 28. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227, 680–685. 29. Arnold, U. & Ulbrich-Hofmann, R. (2001) Proteolytic degrada- tion of ribonuclease A in the pretransition region of thermally and urea-induced unfolding. Eur. J. Biochem. 268, 93–97. 30. Jacobs,D.J.,Rader,A.J.,Kuhn,L.A.&Thorpe,M.F.(2001) Protein flexibility predictions using graph theory. Proteins 44, 150–165. 31. Lang, K. & Schmid, F.X. (1986) Use of a trypsin-pulse method to study the refolding pathway of ribonuclease. Eur. J. Biochem. 159, 275–281. 32. Vriend, G. (1990) WHAT IF: a molecular modeling and drug design program. J. Mol. Graph. 8, 52–56. 33. Knowles, J.R. (1987) Tinkering with e nzymes: what a re we learning? Science 236, 1252–1258. 34. Brandts, J.F., Halvorson, H.R. & Brennan, M. (1975) Con- sideration of the possibility that the slow step in protein dena- turation reactions is due to cis-trans isomerism of proline residues. Biochemistry 14, 4953–4963. 35. Juminaga, D., Wedemeyer, W.J. & Scheraga, H.A. (1998) Proline isomerization in bovine pancreatic ribonuclease A. 1. Unfolding conditions. Biochemistry 37, 11614–11620. 36. Wedemeyer, W.J., Welker, E. & Scheraga, H.A. (2002) Proline cis-trans isomerization and protein fo lding. Bioc hemistry 41, 14637–14644. 37. Rauber, N.R., Jany, K.D. & Pfleiderer, G. (1978) Ribonuclease A digestion by proteinase K. Z. Naturforsch. [C] 33, 660–663. 38. Richards, F.M. & Vithayathil, P.J. (1959) The preparation of subtilisin-modified ribonuclease and the separation of the peptide and protein components. J. Biol. Chem. 234, 1459–1465. 39. Markert, Y., Ko ¨ ditz, J., Mansfeld, J., Arnold, U. & Ulbrich-Hofmann, R. (2001) Increased proteolytic resistance of ribonuclease A by protein engineering. Protein Eng. 14, 791– 796. 40. Loh,S.N.,Rohl,C.A.,Kiefhaber,T.&Baldwin,R.L.(1996)A general tw o-process model d escribes the hyd rogen e x change behavior of RNase A in unfolding conditions. Proc.NatlAcad. Sci. USA 93, 1982–1987. 41. Plummer, T.H. Jr & Hirs, C.H. (1964) On the structure of bovine pancreatic ribonuclease B. Isolation of a glycopeptide. J. Biol. Chem. 239, 2530–2538. 42. Chatani, E., N onomura, K., Hayashi, R., Balny, C. & Lange, R. (2002) 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. Biochemistry 41, 4567–4574. 43. Kadonosono, T., Chatani, E., Hayashi, R., Moriyama, H. & Ueki, T. (2003) Minimization of cavity size ensures protein sta- bility and folding: structures of Phe46-replaced bovine pancreatic RNase A. Biochemistry 42, 10651–10658. 44. Catanzano, F., Graziano, G., Cafaro, V., D’Alessio, G., Di Donato, A. & Barone, G. (1997) From ribonuclease A toward bovine seminal ribonuclease: a step by step thermodynamic ana- lysis. Biochemistry 36, 14403–14408. 45. Beintema, J.J., Breukelman, H.J., Carsana, A. & Furia, A. (1997) Evolution of vertebrate r ibonucle ases: ribonuclease A superfamily. In Ribonucleases: Structures and Functions (D’Alessio, G. & Riordan, J.F., eds), pp. 245–269. Academic Press, New York. 46. Ahrweiler, P.M. & Frieden, C . (1991) Effects of point mutations in a hinge region on the stability, folding, and enzymatic activity of Escherichia coli dihydrofolate reductase. Bioche mistr y 30, 7801– 7809. 47. Collinet, B., Garcia, P., Minard, P. & Desmadril, M. (2001) Role of loops in the folding and stability of yeast phosphoglycerate kinase. Eur. J. Biochem. 268, 5107–5118. Ó FEBS 2004 Unfolding region of ribonuclease A ( Eur. J. Biochem. 271) 4155 48. Gu, H., Kim, D. & Baker, D. (1997) Contrasting roles for sym- metrically dispo sed beta-turns in the f olding of a small protein. J. Mol. Biol. 274, 588–596. 49. Urfer, R. & Kirschner, K. (1992) The importance of surface loops for stabilizing a n eightfold beta alpha barrel protein. Protein Sci. 1, 31–45. 50. Chatani, E. & Hayashi, R. (2001) Functional and structural roles of constituent a mino acid resid ues of bovine pancreatic ribo- nuclease A. J. Biosci. Bioeng. 92, 98–107. 51. Lin, S.H., Konishi, Y., Nall, B.T. & Scheraga, H.A. (1985) Influence of an extrinsic cross-link on the folding pathway of ribonuclease A. Kinetics of folding-unfolding. Biochemistry 24, 2680–2686. 52. 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. 53. Klink, T.A., Woycechowsky, K.J., Taylor, K.M. & Raines, R.T. (2000) Contribution of disulfide bonds to the conformational stability and catalytic activity of ribonuclease A. Eur. J. Biochem. 267, 566–572. 54. Haigis, M.C., Kurten, E.L., Abel, R.L. & Raines, R.T. (2002) KFERQ sequence in ribonuclease A-mediated cytotoxicity. J. Biol. Chem. 277, 11576–11581. 55. Dodge, R.W. & Scheraga, H.A. (1996) Folding and unfolding kinetics of the proline-to-alanine mutants of bovine pancreatic ribonuclease A. Biochemistry 35, 1548–1559. 56. delCardayre ´ ,S.B.&Raines,R.T.(1995)Aresiduetoresidue hydrogen bond mediates the nucleotide specificity of ribonuclease A. J. Mol. Biol. 252, 328–336. 4156 J. Ko ¨ ditz et al.(Eur. J. Biochem. 271) Ó FEBS 2004 . ATG ATG AAG AGC CG GAAT TCC ACC AAA GAT CGA TGC AAG C-3¢ EcoRI rev 5¢-G CTT GCA TCG ATC TTT GGT G GA ATT CCG GCT CTT CAT CAT C-3¢ F46Y fw 5¢-GC AAG CCA. GTT ACA G-3¢ L3 5A fw 5¢-G ATG ATG AAG AGC CG GAAT GCC ACC AAA GAT CGA TGC AAG C-3¢ BsmI rev 5¢-G CTT GCA TCG ATC TTT GGT G GC ATT CCG GCT CTT CAT CAT C-3¢ L35S