Báo cáo khoa học: Structure–cytotoxicity relationships in bovine seminal ribonuclease: new insights from heat and chemical denaturation studies on variants ppt

Structure–cytotoxicity relationships in bovine seminal ribonuclease: new insights from heat and chemical denaturation studies on variants Concetta Giancola 1 , Carmine Ercole 1 , Iolanda Fotticchia 1 , Roberta Spadaccini 2 , Elio Pizzo 3 , Giuseppe D’Alessio 3 and Delia Picone 1 1 Department of Chemistry ‘Paolo Corradini’, University of Naples ‘Federico II’, Italy 2 Department of Biological and Environmental Sciences, Universita’ degli Studi del Sannio, Benevento, Italy 3 Department of Structural and Functional Biology, University of Naples ‘Federico II’, Italy Keywords calorimetric analysis; chemical denaturation; cytotoxic ribonucleases; domain-swapping; structure–activity relationships Correspondence D. Picone or C. Giancola, Dipartimento di Chimica, Universita ` di Napoli ‘Federico II’, Complesso Universitario di Monte Sant’Angelo, Via Cintia, 80126 Napoli, Italy Fax: +39 081674409; +39 081674499 Tel: +39 081674406; +39 081674266 E-mail: delia.picone@unina.it; concetta.giancola@unina.it (Received 29 June 2010, revised 17 September 2010, accepted 25 October 2010) doi:10.1111/j.1742-4658.2010.07937.x Bovine seminal ribonuclease (BS-RNase), a homodimeric protein displaying selective cytotoxicity towards tumor cells, is isolated as a mixture of two isoforms, a dimeric form in which the chains swap their N-termini, and an unswapped dimer. In the cytosolic reducing environment, the dimeric form in which the chains swap their N-termini is converted into a noncovalent dimer (termed NCD), in which the monomers remain intertwined through their N-terminal ends. The quaternary structure renders the reduced pro- tein resistant to the ribonuclease inhibitor, a protein that binds most ribo- nucleases with very high affinity. On the other hand, upon selective reduction, the unswapped dimer is converted in two monomers, which are readily bound and inactivated by the ribonuclease inhibitor. On the basis of these considerations, it has been proposed that the cytotoxic activity of BS-RNase relies on the 3D structure and stability of its NCD derivative. Here, we report a comparison of the thermodynamic and chemical stability of the NCD form of BS-RNase with that of the monomeric derivative, together with an investigation of the thermal dissociation mechanism revealing the presence of a dimeric intermediate. In addition, we report that the replacement of of Arg80 by Ser significantly decreases the cytotoxic activity of BS-RNase and the stability of the NCD form with respect to the parent protein, but does not affect the ribonucleolytic activity or the dissociation mechanism. The data show the importance of Arg80 for the cytotoxicity of BS-RNase, and also support the hypothesis that the reduced derivative of BS-RNase is responsible for its cytotoxic activity. Abbreviations BS-RNase, bovine seminal ribonuclease; DSC, differential scanning calorimetry; GSH, glutathione; hA-BS-RNase, G16S ⁄ N17T ⁄ P19A ⁄ S20A variant of bovine seminal ribonuclease; hA-mBS, G16S ⁄ N17T ⁄ P19A ⁄ S20A variant of the monomeric N67D variant of bovine seminal ribonuclease with Cys31 and Cys32 linked to glutathione moieties; mBS, monomeric N67D variant of bovine seminal ribonuclease with Cys31 and Cys32 linked to glutathione moieties; MxM, dimeric form of bovine seminal ribonuclease in which the chains swap their N-termini; M=M, unswapped dimer of bovine seminal ribonuclease; NCD, noncovalent dimer; PDB, Protein Data Bank; RI, ribonuclease inhibitor; RNase A, bovine pancreatic ribonuclease; S 80 -BS-RNase, R80S variant of bovine seminal ribonuclease; S 80 -hA-BS-RNase, R80S ⁄ G16S ⁄ N17T ⁄ P19A ⁄ S20A variant of bovine seminal ribonuclease; S 80 -hA-mBS, R80S ⁄ G16S ⁄ N17T ⁄ P19A ⁄ S20A variant of the monomeric N67D variant of bovine seminal ribonuclease with Cys31 and Cys32 linked to glutathione moieties; S 80 -mBS, R80S variant of the monomeric N67D variant of bovine seminal ribonuclease with Cys31 and Cys32 linked to glutathione moieties. FEBS Journal 278 (2011) 111–122 ª 2010 The Authors Journal compilation ª 2010 FEBS 111 Introduction The outstanding feature of bovine seminal ribonucle- ase (BS-RNase), proposed initially by Piccoli et al., (1992) [1] on the basis of biochemical data and struc- turally proven a few years later [2,3], is the formation of a dimeric form in which the chains swap their N-termini (MxM). This phenomenon was later found in many proteins, and is now well known under the name of ‘3D domain swapping’. In most cases, the swapping is associated with new biological functions, and it has also been proposed as a possible mecha- nism for protein aggregation in misfolding-associated pathologies [4]. To date, more than 150 structures of swapped proteins are present in the Protein Data Bank (PDB). Among these, BS-RNase still represents a special case, because the native protein is isolated as a mixture of two dimeric isoforms, MxM and an M=M [1], with or without the exchange of N-termini respectively, in a molar ratio of about 2 : 1. Therefore, only for this protein, the swapping is a physiological, equilibrium process consisting of a dimer-to-dimer interconversion; that is, it is not associated with varia- tion in the quaternary structure. The swapping is considered to be a prerequisite for most additional biological properties accompanying the basal enzy- matic activity, including a selective cytotoxic activity towards malignant tumor cells [5]. However, the X-ray structures of the two isoforms have revealed only minor differences [2,3], located essentially at the loop connecting the dislocated arms to the main body of the protein. On the other hand, a so-called ‘buried diversity’ [6] becomes evident when the protein is con- sidered under different environments, such as cytosolic reducing conditions. In vitro, under mild reducing conditions, the two interchain disulfides bridging the subunits of BS-RNase undergo selective cleavage, so that M=M is converted into two monomers, whereas MxM maintains a dimeric structure, stabilized by noncovalent interac- tions of the N-termini [1]. The monomeric form of BS-RNase is readily neutralized by the ribonuclease inhibitor (RI) [7], a protein that is abundant in mam- malian cells, whereas the quaternary structure of the reduced dimer, henceforth called the noncovalent dimer (NCD), allows this protein to evade RI binding [8]. It has been proposed that RI prevents endogenous RNA degradation by binding monomeric ribonucleases with very high affinity [9]. A schematic representation of the different forms that BS-RNase can adopt in different environments is given in Fig. 1, to emphasize that, in the reducing conditions of the cytosol, BS-RNase exists both as a monomer and as swapped NCD stabilized by noncovalent interactions. NCD, which is a transient species because, in solution, it dis- sociates into two monomers, is considered to be the form responsible for the cytotoxic activity of the enzyme, given its resistance to RI. Furthermore, it has been reported that the structural determinants that favor the proper quaternary structure [6,10] and the stability in solution of the swapped form play a significant role in the additional biological properties of the enzyme. In this study, we have investigated the relationship between the cytotoxic activity and the stability of the NCD form of BS-RNase in comparison with variants obtained by replacing the 16–20 hinge loop region and ⁄ or Arg80 with the corresponding residues of Structured digital abstract l MINT-8050499: BS-RNase (uniprotkb:P00669) and BS-RNase (uniprotkb:P00669) bind ( MI:0407)bybiophysical (MI:0013) l MINT-8050482: BS-RNase (uniprotkb:P00669) and BS-RNase (uniprotkb:P00669) bind ( MI:0407)byclassical fluorescence spectroscopy (MI:0017) l MINT-8050435: BS-RNase (uniprotkb:P00669) and BS-RNase (uniprotkb:P00669) bind ( MI:0407)bycircular dichroism (MI:0016) Fig. 1. Crystal structures of the multiple forms of BS-RNase: mBS, PDB code 1N1X; M=M, PDB code 1R3M; MxM, PDB code 1BSR; NCD, PDB code 1TQ9. Stability–activity relationships in a cytotoxic dimeric ribonuclease C. Giancola et al. 112 FEBS Journal 278 (2011) 111–122 ª 2010 The Authors Journal compilation ª 2010 FEBS bovine pancreatic ribonuclease (RNase A). It is worth noting that none of these residues is actually implicated in the catalytic activity. We have already reported that neither substitution, i.e. the changes in the 16–20 region or the change at Arg80, significantly affects the swapping propensity of BS-RNase [11,12]. However, when the R80S mutation is inserted into the construct containing the 16–20 hinge loop region of RNase A [R80S⁄ G16S⁄ N17T⁄ P19A⁄ S20A variant of BS-RNase (S 80 -hA-BS-RNase)], the MxM ⁄ M=M molar ratio in the equilibrium mixture is changed from 2 : 1 to 1 : 2 [12]. On the basis of the hypothesis that ascribes the special functions of BS-RNase to the swapped form, in this study we investigated the biolog- ical activity of this mutant, and found that it loses almost all of the cytotoxic activity. Interestingly, the single R80S substitution is sufficient to reduce the anti- tumor activity almost to the same extent, leading us to assign to this residue a pre-eminent role in the cyto- toxic activity of BS-RNase, independently of the hinge sequence. Furthermore, evaluation of the chemical and thermal stability of the NCD forms of the variant pro- teins in comparison with those of the parent one sup- ports the hypothesis that the reduced swapped dimer represents the bioactive form of BS-RNase. Results BS-RNase and its R80S variant (S 80 -BS-RNase), its G16S ⁄ N17T ⁄ P19A ⁄ S20A variant (hA-BS-RNase) and R80S/G16S/N17T/P19A/S20A variant (S 80 -hA-BS-RNase) were expressed in monomeric form, with Cys31 and Cys32 linked covalently and reversibly to two glutathi- one (GSH) moieties, as already reported [12,13]. The correctness of the fold of each monomeric protein was confirmed by comparing their CD spectra and Kunitz enzymatic activity on yeast RNA [14] with those of parent monomeric BS-RNase (data not shown). Fur- thermore, we compared the 2D NMR spectra of all the monomeric variant proteins with that of the parent monomeric N67D variant with Cys31 and Cys32 linked to GSH moieties (mBS), and found that the res- onances of most amide signals are almost coincident, the differences being essentially restricted to the back- bone and side chain of the mutated residue(s) and of those closest (in space). This is illustrated in detail for the R80S variant of mBS (S 80 -mBS) in Fig. 2: panel (A) shows the overlay of the 1 H- 15 N-HSQC spectrum with that of mBS, and panel (B) gives a detail of the 3D structure of monomeric BS-RNase (PDB code 1N1X), showing the local environment of Arg80. It is very evident that the shifted residues belong to the region encompassing Arg80 (78–84), to the hinge (15 and 16) and to regions 45–49 and 101–103, which are less than 4 A ˚ from the Arg80 side chain. The overlay of the 1 H- 15 N-HSQC spectrum of mBS with those of its G16S ⁄ N17T ⁄ P19A ⁄ S20A mBS and R80S ⁄ G16S ⁄ N17T ⁄ P19A ⁄ S20A variants, indicated as hA-mBS and S 80 -hA-mBS respectively, is shown in Fig. S1. Biological activity The monomeric proteins were converted into dimers by removal of the protecting GSH moieties and 15 N (p.p.m.) 1 H (p.p.m.) A B Fig. 2. (A) Overlay of the 1 H– 15 N-HSQC spectra of the monomeric derivatives of BS-RNase (black) and S 80 -BS-RNase (red) at 300 K. Residues whose resonances are shifted are labeled. (B) Details of the 3D structure of monomeric BS-RNase (PDB code 1N1X), show- ing the local environment of Arg80. Residues that are less than 4 A ˚ from the Arg80 side chain are indicated. C. Giancola et al. Stability–activity relationships in a cytotoxic dimeric ribonuclease FEBS Journal 278 (2011) 111–122 ª 2010 The Authors Journal compilation ª 2010 FEBS 113 reoxidation of the intersubunit disulfides, followed by incubation at 37 °C to allow the interconversion of M=M and MxM to reach equilibrium. The cytotoxic activity of the proteins towards tumor cells was mea- sured by adding increasing concentrations (ranging from 12.5 to 100 lgÆmL )1 ) of each variant to malig- nant SVT2 cells, using BS-RNase as a positive control. For a negative control, the proteins were also assayed on nontumor 3T3 cell cultures at a final concentration of 100 lgÆmL )1 , and found to be nontoxic (Fig. S2). The percentage of SVT2 cells surviving, illustrated in Fig. 3, show that the replacement of Arg80 by Ser induced a significant drop in cytotoxic activity, inde- pendently of the hinge sequence. In contrast, we found that changes in the 16–20 hinge region had only small effects, as the cytotoxic activity of BS-RNase was only slightly higher than that of hA-BS-RNase, and that of S 80 -BS-RNase was very close to that of S 80 -hA-BS- RNase. Stability of NCD versus monomeric forms In the search for the molecular basis for the induction of the loss of cytotoxic activity of BS-RNase variants reported in Fig. 3, we followed by CD the thermal denaturation process of NCD derivatives, measuring the molar ellipticity at 222 nm as a function of temper- ature (Fig. 4A). As a comparison, the melting curves of the corresponding monomers are reported in Fig. 4B. The melting temperatures (T m values) of NCD derivatives, which represent the midpoint of the denaturation curve (Fig. 4A), were 59.4 °C for BS-RNase and 59.0 °C for hA-BS-RNase, i.e. very close to each other. In turn, significantly lower, and comparable, T m values of 54.3 °C and 53.6 °C were found for S 80 -hA-BS-RNase and S 80 -BS-RNase, respec- tively. A similar trend was observed for the CD melting temperatures of monomeric derivatives (Fig. 4B) and [12], which can be separated into two groups, corre- sponding to T m values around 58.0 °C and 54.0 °C for the proteins with Arg80 or Ser80, respectively. The CD melting curves of the monomers were used to calculate the denaturation enthalpy changes by using the van’t Hoff equation (Eqn 3 in Experimental procedures), which describes two-state NMD transi- tions. The data obtained, reported in Table 1, indicate that DH 0 v:H: values of the monomers follow the same trend of T m values, with those of mBS and hA-mBS close to each other and higher than the DH 0 v:H: values of S 80 -mBS and S 80 -hA-mBS, which, in turn, are close to each other. As a further step, we performed a calorimetric analysis of all the proteins by standard differential 0 25 50 75 100 12.5 25 50 100 (µg·mL –1 ) Cell survival (%) Fig. 3. SVT2 cell survival after 48 h of incubation with different amount of BS-RNase ( ), S 80 -BS-RNase (h), hA-BS-RNase ( ) and S 80 -hA-BS-RNase ( ). 10 20 30 40 50 60 70 80 90 0 1 A B Temperature (°C) 10 20 30 40 50 60 70 80 90 Temperature (°C) Fraction unfolded 0 1 Fraction unfolded Fig. 4. Thermal unfolding curves obtained following the change of CD signal at 222 nm of NCDs (A) and monomeric derivatives (B) of BS-RNase ( ), hA-BS-RNase (x), S 80 -BS-RNase (•), and S 80 -hA-BS- RNase (h). The unfolded fraction of protein was calculated as (Q – Q min ) ⁄ (Q max – Q min ); Q is the ellipticity at 222 nm at a given temper- ature, and Q max and Q min are the maximum and minimum values of ellipticity corresponding to the denaturated state and native state of proteins, respectively. Stability–activity relationships in a cytotoxic dimeric ribonuclease C. Giancola et al. 114 FEBS Journal 278 (2011) 111–122 ª 2010 The Authors Journal compilation ª 2010 FEBS scanning calorimetry (DSC) measurements. The calori- metric profiles of all proteins are reported in Figs S3 and S4 for monomers and NCDs respectively. The denaturation enthalpies of the monomeric derivatives obtained from the DSC curves, reported as DH 0 cal in Table 1, were in good agreement with the van’t Hoff enthalpies derived from DSC and CD curves, thus confirming that the thermal denaturation process for these proteins is a two-state transition process. An inspection of the whole set of thermodynamic parame- ters of monomeric forms, collected in Table 1, shows that hA-mBS and mBS have comparable stabilities, indicating that the substitution of four residues in the hinge region does not significantly perturb the global stability of the monomeric form of BS-RNase. On the other hand, the single mutation R80S leads to decreases of about 4 °C in the melting temperature and of about 100 kJÆmol )1 in the value of DH 0 cal . This destabilization is well reflected in DG 0 values, showing that Arg80 is crucial for the stability of monomeric form of BS-RNase. For the NCD forms, thermal denaturation was found to be an irreversible process, because there was no refolding upon cooling of the protein solutions. The irreversibility of the denaturation process does not allow Gibbs energy calculations, but only a compari- son of the melting temperatures and unfolding enthalpy changes. DH 0 cal and DH 0 v:H: . values, both calcu- lated from calorimetric profiles, are very similar and the DH 0 cal =DH 0 v:H: ratio is in the range 0.98–1.07, sug- gesting that the unfolding of the dimers is close to being a one-step process (Table 1). The DH 0 v:H: from CD and DSC curves, relative to the unfolding of the secondary and tertiary structure respectively, are for each dimer very close, indicating simultaneous collapse of both structures. As also indicated in Table 1, the enthalpy changes for the NCD forms are very similar to each other. For a comparison of the DH 0 cal values of the dimers with those of the corresponding monomers, the enthalpy changes of NCD forms were calculated at the melting temperatures of the corresponding monomers, using the Kirchhoff equation. Values of 607 kJÆmol )1 , 642 kJÆ mol )1 , 581 kJÆ mol )1 and 605 kJÆmol )1 were obtained for NCD forms of BS-RNase, hA-BS-RNase, S 80 -BS-RNase and S 80 -hA-BS-RNase, respectively. All values are less than twice those of the respective monomers. This indicates a loss of interac- tions of the monomers in the dimeric structures. In conclusion, all of the reported data indicate that the R80S mutation is crucial for the loss in the enthalpic content of NCDs of BS-RNase, engendering a lower melting temperature of the R80S variants. Urea denaturation of NCD forms The conformational stability of the NCD forms against the denaturing action of urea in comparison with the corresponding monomers was investigated by means of steady-state fluorescence and CD measure- ments at pH 7.0. Monomeric proteins showed sigmoidal transition curves when the change in fluorescence intensity was Table 1. Thermodynamic melting parameters of the unfolding process of monomers and NCDs of BS-RNase mutants. T m , denaturation temperature; DH 0 (T m ), calorimetric enthalpy change; DH 0 v:H: , van’t Hoff enthalpy change; DC 0 p ðT m Þ, excess heat capacity change; DS 0 (T m ), entropy change; DG 0 298 , denaturation Gibbs energy change at 298 K. T m (°C) DH 0 (T m ) (kJÆmol )1 ) DH 0 v:H: (kJÆmol )1 ) DC 0 p ðT m Þ (kJÆmol )1 ÆK )1 ) DS 0 (T m ) (kJÆmol )1 ÆK )1 ) DG 0 298 (kJÆmol )1 ) mBS 58.0 ± 0.5 428 ± 12 456 ± 20 408 ± 16 a 5.3 ± 0.5 0.72 ± 0.03 36.9 ± 5.5 hA-mBS 58.5 ± 0.5 405 ± 13 397 ± 21 420 ± 17 a 4.7 ± 0.6 0.69 ± 0.04 41.9 ± 6.3 S 80 -hA-mBS 53.9 ± 0.5 334 ± 10 330 ± 16 328 ± 13 a 4.8 ± 0.4 0.57 ± 0.03 24.9 ± 3.7 S 80 -mBS 54.5 ± 0.5 331 ± 9 325 ± 18 316 ± 13 a 5.4 ± 0.6 0.50 ± 0.03 25.4 ± 3.8 NCD BS-RNase 59.4 ± 0.5 619 ± 18 570 ± 23 528 ± 21 a ––– NCD hA-BS-RNase 59.0 ± 0.5 646 ± 19 609 ± 20 588 ± 17 a ––– NCD S 80 -hA-BS-RNase 54.3 ± 0.5 585 ± 17 597 ± 23 553 ± 22 a ––– NCD S 80 -BS-RNase 53.6 ± 0.5 597 ± 18 590 ± 22 547 ± 21 a ––– a DH 0 v:H: from CD measurements. C. Giancola et al. Stability–activity relationships in a cytotoxic dimeric ribonuclease FEBS Journal 278 (2011) 111–122 ª 2010 The Authors Journal compilation ª 2010 FEBS 115 recorded at the wavelength maximum, I max , as a func- tion of urea concentration (insets in Fig. 5), whereas the curves of NCDs displayed two transitions (Fig. 5). The values of urea concentration at half-completion of transition, C ½ , are shown in Table 2, which also reports the values found by monitoring the molar ellip- ticity at 222 nm with CD measurements; these reflect conformational changes of the secondary structures for monomers (insets in Fig. 6) and NCDs (Fig. 6). Also in this case, two distinct C ½ values, the first value in the 2–3 m range and the second close to the C ½ value of the corresponding monomeric form (Table 2), were observed for the dimers. To investigate in more detail the mechanism of urea denaturation, we followed this process at different protein concentrations, but focus- ing on the parent BS-RNase and on the single-point R80S variant. The results are reported in Fig. 7, where the folded fraction is reported as a function of the urea concentration at four different protein concentrations, in the range 0.1–25 lm. According to Rumfeldt et al. [15], the variation in the curve shape from sigmoidal to biphasic observed when the protein concentration increases confirms the presence of a dimeric intermedi- ate in the dissociation process of both NCD variants. The biphasic curves for the NCD forms of BS-RNase and S 80 -BS-RNase at the highest concentration, where the intermediate is present in significant amounts, were analyzed according to the three-state equilibrium model (N 2 MI 2 M2U) [16]. The following values for the Gibbs energy changes and m-values were obtained: DG 1 =14kJÆmol )1 , m 1 =5kJÆmol )1 Æm )1 , DG 2 =80 kJÆmol )1 , m 2 =19kJÆmol )1 Æm )1 for NCD BS-RNase; and DG 1 =12kJÆmol )1 , m 1 =6kJÆmol )1 Æm )1 , DG 2 = 43 kJÆmol )1 , m 2 =15kJÆmol )1 Æm )1 for S 80 -BS-RNase. The Gibbs energy values indicate that the perturbative action of the urea is greater for the second transition, I 2 M2U, than for the first transition, N 2 MI 2 , for both dimers. The m-values also indicate that the surface area exposed to solvent in the first transition is smaller than that in the second transition. A comparison between the two NCD forms shows that the R80S mutation decreases the stability mainly in the step I 2 M2U, and, if we assume that the final state is the same for both NCD forms, this suggests that the R80S mutation decreases the stability of the intermediate. Structural models of the NCD forms In the search for the possible origin of the reduced activity of S 80 variants in both aggregation states, i.e. in the monomeric and noncovalent swapped dimeric forms, we examined the corresponding 3D structures. All attempts to obtain crystals suitable for X-ray anal- ysis of any form of the two S 80 -BS-RNase variants had hitherto been unsuccessful. Supported by the similarity of NMR spectra (Figs 2 and S1) of the monomers, suggesting that the global architecture of all the variant proteins is very similar to that of the parent BS-RNase, and by the close similarity among the X-ray structures of swapped isoforms of hA-BS-RNase and BS-RNase [11], models of the 3D structures of all proteins were obtained starting from the X-ray structure of the corresponding form of the parent BS-RNase, i.e. the monomeric derivative (PDB 02468 0.0 0.2 0.4 0.6 0.8 1.0 Fraction unfolded [Urea] M 02468 0.0 0.2 0.4 0.6 0.8 1.0 Fraction unfolded [Urea] M 02468 0.0 0.2 0.4 0.6 0.8 1.0 Fraction unfolded [Urea] M 02468 0.0 0.2 0.4 0.6 0.8 1.0 Fraction unfolded [Urea] M AB CD Fig. 5. Urea-induced transition curves for NCDs of BS-RNase variants and for the corresponding monomers (inset), followed by fluorescence spectroscopy. (A) BS-RNase. (B) hA-BS-RNase. (C) S 80 -BS-RNase. (D) S 80 -hA-BS-RNase. The unfolded fraction represents the fraction of denaturated protein, calculated as (I ) I min ) ⁄ (I max ) I min ); I is the fluorescence intensity at a given temperature, and I max and I min are the maximum and minimum values of fluorescence intensity corresponding to the denaturated state and native state of proteins, respectively. Stability–activity relationships in a cytotoxic dimeric ribonuclease C. Giancola et al. 116 FEBS Journal 278 (2011) 111–122 ª 2010 The Authors Journal compilation ª 2010 FEBS code 1N1X) and the noncovalent swapped dimer (PDB code 1TQ9). A representation of the structural models built for S 80 -hA-BS-RNase, which, among the variants examined in this study, is the one hosting the highest number of substitutions, is reported in Fig.8. A careful inspection of these structures reveals that all of the variant proteins examined, and notably both of the S 80 variants, are characterized by the presence of a decreased number of hydrogen bonds with respect to the parent protein. This trend is even more evident in the NCD derivatives: in these forms, all of the mutants have fewer intersubunit hydrogen bonds than the native protein. In particular, focusing on residue 80, a contact between the side chains of Arg80 and Ser18 is detectable only in parent BS-RNase. Fur- thermore, the same protein and hA-BS-RNase are sta- bilized by a contact between a core residue of one subunit (Gln101) and the hinge residues of the other subunit (Ser20 in the case of BS-RNase and Ser18 in the case of hA-BS-RNase). Discussion The antitumor activity of dimeric ribonucleases relies on their quaternary structure, which enables the pro- teins to avoid inhibition by RI and provides good sta- bility in solution. We have already shown that Pro19, Leu28 and, possibly, Gly16 play a relevant role in the cytotoxicity, because they ensure the correct quater- nary assembly of the NCD derivative of BS-RNase [17]. However, the hinge residues and Leu28 have a synergistic effect, because to observe a drastic reduc- tion of the cytotoxic activity they have to be replaced simultaneously [6,18]. In contrast, the data reported in this article show that the substitution of the whole hinge region induces only a small reduction in the basal cytotoxic activity of BS-RNase, as in the case of the single mutants P19A and L28Q [6]. Surprisingly, the replacement of Arg80 by Ser significantly reduces the cytotoxic activ- ity, as both S 80 variants are less active than the parent BS-RNase. On the one hand, this shows the impor- tance of Arg80, although it is irrelevant for the cata- lytic activity or the swapping extent of BS-RNase [12]; on the other hand, it indirectly confirms that the Table 2. Urea-induced denaturation parameters of monomers and dimers of BS-RNase mutants, monitored by fluorescence and CD spectroscopy. [Urea] 1 ⁄ 2 values were the denaturant concentrations at half-completion of the transition. [Urea] 1 ⁄ 2 (M) CD [Urea] 1 ⁄ 2 (M) Fluorescence mBS 5.70 ± 0.20 5.70 ± 0.08 NCD BS-RNase 3.00 ± 0.06 5.60 ± 0.04 3.00 ± 0.06 5.70 ± 0.03 S 80 -hA-mBS 4.10 ± 0.15 4.20 ± 0.20 NCD S 80 -hA-BS-RNase 1.60 ± 0.05 4.50 ± 0.15 1.60 ± 0.20 4.60 ± 0.20 hA-mBS 5.31 ± 0.03 5.36 ± 0.15 NCD hA-BS-RNase 2.20 ± 0.04 5.60 ± 0.01 2.62 ± 0.03 5.64 ± 0.01 S 80 -mBS 4.76 ± 0.01 5.04 ± 0.03 NCD S 80 -BS-RNase 2.07 ± 0.02 5.12 ± 0.02 2.17 ± 0.03 5.09 ± 0.02 02468 0.0 0.2 0.4 0.6 0.8 1.0 Fraction unfolded [Urea] M 02468 0.0 0.2 0.4 0.6 0.8 1.0 Fraction unfolded [Urea] M 02468 0.0 0.2 0.4 0.6 0.8 1.0 Fraction unfolded [Urea] M 02468 0.0 0.2 0.4 0.6 0.8 1.0 Fraction unfolded [Urea] M AB CD Fig. 6. Urea-induced transition curves for NCDs of BS-RNase variants and for the corresponding monomers (inset), followed by CD spectroscopy. (A) BS-RNase. (B) hA-BS-RNase. (C) S 80 -BS-RNase. (D) S 80 -hA-BS-RNase. The unfolded fraction of protein was calculated as (Q ) Q min ) ⁄ (Q max ) Q min ); Q is the ellipticity at 222 nm at a given temperature, and Q max and Q min are the maximum and minimum values of ellipticity corresponding to the denaturated state and native state of proteins, respectively. C. Giancola et al. Stability–activity relationships in a cytotoxic dimeric ribonuclease FEBS Journal 278 (2011) 111–122 ª 2010 The Authors Journal compilation ª 2010 FEBS 117 exchange of the N-terminal arms in BS-RNase-like proteins is not sufficient to elicit the antitumor activity [10]. We are aware that, in principle, the substitution of a basic residue on the protein surface might reduce the cytotoxic activity, by affecting the electrostatic interaction with the cell membrane and perhaps the internalization process [19], but, as shown by Notomi- sta [20], the side of BS-RNase with the strongest posi- tive potential is the one hosting the N-termini, which is located opposite to Arg80 (Fig. 8). The observation of models built with the X-ray structure of the reduced dimer of BS-RNase (PDB code 1TQ9) as template indicate that all of the variants maintain a quaternary structure very close to that of the parent protein, but are characterized by weaker interactions between subunits. This result is also in agreement with thermodynamic data and thermal and chemical dissociation, which indicate a lower stability of the variant proteins with respect to BS-RNase, prin- cipally for both Ser80 variants. We also investigated the dissociation mechanism, and found that it is not affected by the mutations investigated here. The experimental data related to thermal denaturation processes of both Ser80 variant proteins can be interpreted by surmising a simple two-step process, from native dimer to denatured monomers, whereas the presence of biphasic chemical denaturation profiles, exhibited by both fluorescence and CD curves, suggests the existence of a thermody- namically stable intermediate induced by urea. This discrepancy is not unexpected, because the two dena- turation processes are induced by different perturbing agents, proceed through different mechanisms (in the case of the thermal denaturation process, the interme- diate state is not present in significant amounts) and end with completely distinct denatured states [21]. It is possible that the chaotropic effect of urea initially causes a small perturbation of the secondary and ter- tiary structures of the proteins (DG 1 for the N 2 MI 2 transition is smaller than DG 2 of the I 2 M2U transi- tion), and urea then stabilizes the intermediate through its hydrogen-bonding ability. Furthermore, the com- parison of the values obtained for the first step indi- cates that, in all variants examined, the C ½ value of the first step of the denaturation process is decreased with respect to that of the parent protein, indicating 02468 0.0 0.2 0.4 0.6 0.8 1.0 Fraction unfolded [Urea] M A 0 24 68 0.0 0.2 0.4 0.6 0.8 1.0 Fraction unfolded [Urea] M B Fig. 7. Urea-induced transition curves of NCDs of BS-RNase (A) and S 80 -BS-RNase (B), followed by fluorescence spectroscopy at different protein concentrations: (•), 0.1 l M;( ), 1 lM;(h), 7 lM; ( ), 25 lM. A B Fig. 8. (A) Crystal structure of the NCD form of BS-RNase (PDB code 1TQ9). (B) Homology model of S 80 -hA-BS-RNase. The Arg80– Ser18 and Ser20–Gln101 hydrogen bonds, which are detectable only in BS-RNase, are indicated. Stability–activity relationships in a cytotoxic dimeric ribonuclease C. Giancola et al. 118 FEBS Journal 278 (2011) 111–122 ª 2010 The Authors Journal compilation ª 2010 FEBS that the interactions between the hinge region and resi- due 80 are involved in the early stages of the chemical denaturation process. As a consequence, S 80 -hA-BS- RNase is the most prone to unfolding. Our data suggest a correlation between the cytotoxic activity of BS-RNase and its derivatives and the stability of the corresponding reduced swapped forms. This is also in agreement with the finding that cyto- toxic RNases are, in general, very stable enzymes, and with the relationship between resistance to unfolding and cytotoxic activity observed for different variants of RNase A [22,23]. In conclusion, the enhancement of the conformational stability of the NCD derivative represents a good approach to increase the toxicity of BS-RNase towards cancer cells. In addition, the find- ing that the structure and stability of the dimeric inter- mediate, shown by urea denaturation studies, play a key role in the dissociation process suggests further investigations of the dissociation mechanism that may help in the design of new cancer chemotherapeutic agents based on BS-RNase. Experimental procedures Protein samples All of the experimental procedures for obtaining significant amounts of BS-RNase and its variants starting from the cor- responding pET-22b(+) plasmid cDNA have already been described in detail elsewhere [12,13]. As in the previous stud- ies, all of the constructs were coding for an Asp residue at position 67, instead of an Asn as in the wild-type protein, to avoid side effects caused by the spontaneous deamidation of Asn67 [24,25]. All of the proteins were expressed in Escheri- chia coli cells and purified in monomeric form, with Cys31 and Cys32 linked to two GSH molecules. Monomers with Cys31 and Cys32 in the reduced form, prepared as described previously [12], were either carboxyamidomethylated with iodoacetamide [26], to obtain the monomeric proteins used directly for analysis, or dialyzed against 0.1 m Tris ⁄ acetate (pH 8.4) for 20 h at 4 °C, to obtain dimers. In both cases, the last step of the purification procedure was gel filtration chromatography on a G-75 column (75 · 3 cm). Freshly prepared dimeric proteins were essentially made by M=M isomers: they were incubated at 37 °C for at least 72 h, to allow the mixture to reach equilibrium. The protein concentration was measured by UV spectro- photometry, assuming e = 0.5 (0.1%, 278 nm, 1 cm) for monomers and e = 0.465 for dimers. NCDs The NCDs were prepared according to the protocol previ- ously described by Piccoli et al. [1], from the exchanged form of the corresponding dimer. The reduction of the interchain disulfide bridges was confirmed by SDS ⁄ PAGE, under nonreducing conditions. The NCD forms were kept at 4 °C until used for kinetic or thermodynamic analyses. Cytotoxicity studies Cytotoxicity was evaluated by performing the 3-(4,5-dim- ethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide reduc- tion inhibition assay as described previously [27]. Simian virus-40-transformed mouse fibroblasts (SVT2 cells) and the parental nontransformed Balb ⁄ C 3T3 line (3T3 cells) were obtained from the ATCC (Manassas, VA, USA). The cells were plated on 96-well plates at a density of 2.5 · 10 3 cells per well in 100 lL of medium containing BS-RNase or one of its variants (12.5, 25, 50 or 100 lgÆmL )1 ), and incubated for 24 and 48 h at 37 °C. Cell survival is expressed as the absorbance of blue formazan measured at 570 nm [27] with an automatic plate reader (Victor3 Multilabel Counter; Per- kin Elmer, Shelton, CT, USA). Each curve reports the aver- age of three independent assays. Standard deviations in all assays were in the range of 5–10%. DSC DSC measurements were performed on a third-generation Setaram micro-DSC instrument at 1 °CÆmin )1 . DSC data were analyzed with a previously described program [28]. The excess heat capacity of the protein in solution in the sample cell was measured against a reference cell filled with the buf- fer solution in the temperature range of 4–80 °C. The excess heat function hDC 0 P i. was obtained after baseline subtraction, assuming that the baseline is given by the linear temperature dependence of the native state heat capacity. The denatur- ation enthalpies, DH 0 (T m ), were obtained by integrating the area under the heat capacity versus temperature curves. T m is the melting temperature, and corresponds to the maximum of each DSC peak. The entropy changes corresponding to the thermal denaturation of the monomers, DS 0 (T m ), were determined by integrating the curve obtained by dividing the heat capacity curve by the absolute temperature. Thermodynamic analysis The van’t Hoff enthalpies were obtained by DSC profiles, utilizing the equation [29]: DH 0 v:H: ¼ aRT 2 m DC 0 P ðT m Þ=DH 0 ðT m Þð1Þ where a originates from the stoichiometry of the reaction (a =4 or a = 6 for monomer or dimer denaturation, respectively), T m is the maximum of the DSC peak, DC 0 P ðT m Þ is the value of the excess heat capacity function at T m , and DH 0 (T m ) is the calorimetric enthalpy calculated by direct integration of the area under the DSC peak. For the C. Giancola et al. Stability–activity relationships in a cytotoxic dimeric ribonuclease FEBS Journal 278 (2011) 111–122 ª 2010 The Authors Journal compilation ª 2010 FEBS 119 monomers, the entropy changes, DS 0 (T m ), were determined by integrating the curve obtained by dividing the heat capacity curve by the absolute temperature, and the dena- turation Gibbs energies at 298 K were calculated by com- bining the classical Kirchoff equations: DG 0 ð298Þ¼DH 0 ðT m Þ T m À 298 T m  À DC 0 p ðT m Þ ðT m À 298Þþ298 DC 0 p ðT m Þ ln T m 298  ð2Þ NMR Two-dimensional NMR spectra were acquired at 300 K on on Bruker DRX600 spectrometer by using standard pulse sequence libraries. For the natural abundance 1 H– 15 N- HSQC spectra, the protein concentration was 2.5 mm in 95% H 2 O ⁄ 5% D 2 O (pH 5.65). 1 H chemical shifts are rela- tive to the water signal at 4.70 p.p.m. at 300 K, and 15 N chemical shifts were indirectly referenced to the 1 H chemi- cal shifts according to gyromagnetic ratios [30]. CD spectroscopy CD measurements were performed on a JASCO 715 CD spectrophotometer equipped with a thermoelectrically con- trolled cell holder (JASCO PTC-348) that allows measure- ments at a controlled temperature. Quartz cuvettes with 0.1 cm optical pathlength were used. Unless otherwise reported, the protein concentration was 0.2 mgÆmL )1 . The CD spectra were recorded from 250 to 200 nm at 4 °C, and normalized by subtraction of the buffer spectrum. Molar ellipticity per mean residue [h] in deg cm 2 Ædmol )1 was calcu- lated from the equation [h] = 100[h] obs ⁄ lC, where [h] obs is the ellipticity measured in degrees, l is the pathlength of the cell (cm) and C is the protein molar concentration. CD spectra were recorded with a response of 4 s, a 1.0 nm bandwidth and 20 nmÆmin )1 scan rates. Thermal denatur- ation curves were recorded in temperature mode at 222 nm, with heating of the protein solution from 4 °Cto80°C and a scan rate of 1.0 °CÆmin )1 . The enthalpy changes were calculated using origin 7.5 to fit CD melting curves by the van’t Hoff equation: @ ln K @ 1 = TðÞ  P ¼À DH 0 R ð3Þ where K is the equilibrium constant and R is the gas con- stant. The urea-induced transition curves at 4 °C were obtained by recording the CD signal at 222 nm for each independent sample. All of the measurements were performed after over- night incubation of the samples at 4 °C. The values of urea concentration at half-completion of transition, [urea] 1 ⁄ 2 , were calculated with the Boltzmann equation of origin 7.0. Fluorescence analyses Intrinsic protein fluorescence was recorded with a Perkin Elmer LS50B spectrofluorimeter equipped with a circulating water bath. Unless otherwise stated, the protein concentra- tion was 0.2 mgÆmL )1 , corresponding to 14 lm for the monomers and 7 lm for the dimers. The excitation wave- length was set at 274 nm, and the emission was measured between 250 and 400 nm. The spectra were recorded at 4 °C with a 1 cm cell and a 10 nm emission slit width, and cor- rected for background signal. The urea-induced transition curves were obtained by recording the change in fluorescence intensity at maximum wavelength as a function of denatur- ant concentration. The maximum emission wavelength of proteins was recorded at 303 nm. Measurements were per- formed after overnight incubation of samples at 4 °C. The values of urea concentration at half-completion of transi- tion, [urea] 1 ⁄ 2 , were calculated as for the CD measurements. The biphasic curves at 25 lm were analyzed according to the three-state equilibrium model with a dimeric intermedi- ate, N2MI2M2U, where N2 is the native dimeric state, I2 the dimeric intermediate and U the unfolded monomer, respectively [16]. Fitting of the data was performed with the matlab 5.3 package. The model gives the Gibbs energy changes, DG 1 and DG 2 , and the values of m 1 and m 2 for the N2MI2 and I2M2U transitions, respectively. The m-value is a measure of the dependence of DG on denatur- ant concentration [31], and is proportional to the amount of protein surface area exposed upon unfolding [32]. Molecular modeling The structures of monomeric and dimeric variants of BS-RNase were calculated from the NMR and X-ray struc- tures deposited at the PDB. In particular, the NMR struc- ture of mBS (PDB code 1QWQ) was used for S 80 -mBS, whereas for the dimeric NCD form three crystallographic structures were used, corresponding respectively to the NCD BS-RNase derivative (PDB code 1TQ9), to the N-dimer of RNase A (PDB code 1A2W) and to the PM8 human pancreatic RNase variant (PDB code 1H8X). The atomic coordinates of the above-mentioned protein struc- tures were used as a template to predict the 3D structure of the variants, using modeller 8v5 [33]. The quality of the structural models was evaluated with modeller, using the score of variable target function method [34]. Model analy- ses were performed with molmol [35] and pymol [36]. Acknowledgements C. Ercole is a fellow of the Department of Chemistry ‘Paolo Corradini’ of University of Naples ‘Federico II’, supported by a training grant from the ‘Compag- nia di San Paolo di Torino’. We thank T. Tancredi for help with NMR spectra acquisition, L. Petraccone for Stability–activity relationships in a cytotoxic dimeric ribonuclease C. Giancola et al. 120 FEBS Journal 278 (2011) 111–122 ª 2010 The Authors Journal compilation ª 2010 FEBS [...]... T & Picone D (2003) The swapping of terminal arms in ribonucleases: comparison of the solution structure of monomeric bovine seminal and pancreatic ribonucleases Biochemistry 42, 8704–8711 Kunitz M (1946) A spectrophotometric method for the measurement of ribonuclease activity J Biol Chem 164, 563–568 Rumfeldt JA, Galvagnion C, Vassall KA & Meiering EM (2008) Conformational stability and folding mechanisms... Mazzarella L (2005) The role of the hinge loop in domain swapping The special case of bovine seminal ribonuclease J Biol Chem 280, 13771– 13778 12 Ercole C, Spadaccini R, Alfano C, Tancredi T & Picone D (2007) A new mutant of bovine seminal 14 15 16 17 18 19 20 21 22 23 24 25 ribonuclease with a reversed swapping propensity Biochemistry 46, 2227–2232 Avitabile F, Alfano C, Spadaccini R, Crescenzi O, D’Ursi AM,... structure, and its resistance to the cytosolic ribonuclease inhibitor Biochemistry 40, 3492– 3496 8 Sica F, Di Fiore A, Merlino A & Mazzarella L (2004) Structure and stability of the non-covalent swapped dimer of bovine seminal ribonuclease: an enzyme tailored to evade ribonuclease protein inhibitor J Biol Chem 279, 36753–36760 9 Kobe B & Deisenhofer J (1996) Mechanism of ribonuclease inhibition by ribonuclease... G (1997) Ribonucleases: Structures and Functions Academic Press, NewYork 6 Merlino A, Ercole C, Picone D, Pizzo E, Mazzarella L & Sica F (2008) The buried diversity of bovine seminal ribonuclease: shape and cytotoxicity of the swapped non-covalent form of the enzyme J Mol Biol 376, 427– 437 7 Antignani A, Naddeo M, Cubellis MV, Russo A & D’Alessio G (2001) Antitumor action of seminal ribonuclease, its... of barnase by heteronuclear NMR: implications for the initiation of protein folding J Mol Biol 254, 305–321 Klink TA & Raines RT (2000) Conformational stability is a determinant of ribonuclease A cytotoxicity J Biol Chem 275, 17463–17467 Leland PA, Schultz LW, Kim BM & Raines RT (1998) Ribonuclease A variants with potent cytotoxic activity Proc Natl Acad Sci USA 95, 10407–10412 Di Donato A & D’Alessio... D’Alessio G (1981) Heterogeneity of bovine seminal ribonuclease Biochemistry 20, 7232– 7237 Di Donato A, Galletti P & D’Alessio G (1986) Selective deamidation and enzymatic methylation of seminal ribonuclease Biochemistry 25, 8361–8368 FEBS Journal 278 (2011) 111–122 ª 2010 The Authors Journal compilation ª 2010 FEBS 121 Stability–activity relationships in a cytotoxic dimeric ribonuclease 26 D’Alessio G, Malorni... A (1993) Bovine seminal ribonuclease: structure at 1.9 A resolution Acta Crystallogr D Biol Crystallogr 49, 389–402 3 Berisio R, Sica F, De Lorenzo C, Di Fiore A, Piccoli R, Zagari A & Mazzarella L (2003) Crystal structure of the dimeric unswapped form of bovine seminal ribonuclease FEBS Lett 554, 105–110 4 Bennett MJ, Sawaya MR & Eisenberg D (2006) Deposition diseases and 3D domain swapping Structure... Sica F (2009) Toward an antitumor form of bovine pancreatic ribonuclease: the crystal structure of three noncovalent dimeric mutants Biopolymers 91, 1029–1037 Ercole C, Avitabile F, Del Vecchio P, Crescenzi O, Tancredi T & Picone D (2003) Role of the hinge peptide and the intersubunit interface in the swapping of N-termini in dimeric bovine seminal RNase Eur J Biochem 270, 4729–4735 Leich F, Stohr N,... Malorni MC & Parente A (1975) Dissociation of bovine seminal ribonuclease into catalytically active monomers by selective reduction and alkylation of the intersubunit disulfide bridges Biochemistry 14, 1116–1122 27 Wang X, Ge J, Wang K, Qian J & Zou Y (2006) Evaluation of MTT assay for measurement of emodin-induced cytotoxicity Assay Drug Dev Technol 4, 203–207 28 Barone G, Del Vecchio P, Fessas D, Giancola... CA, USA Supporting information The following supplementary material is available: Fig S1 Overlay of the 1H–15N-HSQC of the monomeric derivatives of: mBs (black), hA-mBS (magenta) and S80-hA-mBS (green) Fig S2 Cytotoxicity of BS-RNase and its variants, all assayed at a final concentration of 100 lgÆmL)1, on 3T3 cells after 48 h of incubation Fig S3 DSC profiles of monomers of BS-RNase variants: (A) mBS; . Structure–cytotoxicity relationships in bovine seminal ribonuclease: new insights from heat and chemical denaturation studies on variants Concetta. the monomeric N67D variant of bovine seminal ribonuclease with Cys31 and Cys32 linked to glutathione moieties; mBS, monomeric N67D variant of bovine seminal