New Factors Enhancing the Reactivity of Cysteines in Molten Globule-Like Structures

Protein cysteines often play crucial functional and structural roles, so they are emerging targets to design covalent thiol ligands that are able to modulate enzyme or protein functions. Some of these residues, especially those involved in enzyme mechanisms—including nucleophilic and reductive catalysis and thiol-disulfide exchange—display unusual hyper-reactivity; such a property is expected to result from a low pKa and from a great accessibility to a given reagent. New findings and previous evidence clearly indicate that pKa perturbations can only produce two–four-times increased reactivity at physiological pH values, far from the hundred and even thousand-times kinetic enhancements observed for some protein cysteines. The data from the molten globule-like structures of ribonuclease, lysozyme, bovine serum albumin and chymotrypsinogen identified new speeding agents, i.e., hydrophobic/electrostatic interactions and productive complex formations involving the protein and thiol reagent, which were able to confer exceptional reactivity to structural cysteines which were only intended to form disulfides. This study, for the first time, evaluates quantitatively the different contributions of pKa and other factors to the overall reactivity. These findings may help to clarify the mechanisms that allow a rapid disulfide formation during the oxidative folding of many proteins.


Introduction
Protein cysteines are involved in many critical structural and functional roles that can be performed thanks to the peculiar properties of their sulfhydryl group, which, in its deprotonated form, becomes an active nucleophile. Due to the physiological importance of this residue, it represents an attractive and emerging target for the development and design of covalent ligands, which are able to modulate the function of specific proteins and enzymes. Therefore, it is of paramount importance that a precise knowledge of the factors that influence the reactivity of these residues is obtained [1,2]. It is a common opinion (universally accepted) that this is mainly controlled by their pK a and by the accessibility of a given reagent [1,[3][4][5][6][7]. In fact, the sulfhydryl group of cysteines is almost inert in its protonated form (except in free-radical reactions), while the thiolate form is the true reactive form. A relevant number of cysteines with functional roles in catalysis have been found to react a hundred-and even thousand-times faster than a free cysteine, but it is only rarely that a quantitative and reasoned analysis of the contribution of a low pK a to these unusual reactivities has been made. This study, based on novel experiments and considering old findings, shows that pK a is not the main determinant, given that, at physiological pH values, the highest increment of the reactivity due to pK a variations cannot exceed two-four-times. Thus, other important factors will be considered in order to discover that, factors will be considered in order to discover that, in a few proteins, they assume an almost exclusive prevalence in modulating the reactivity of these residues. Moreover, we demonstrate that hyperreactivity is not an exclusive feature of functional cysteines and even structural cysteines may have extraordinary reactivity toward many thiol reagents or natural disulfides, which are possibly finalized to a correct and rapid formation of native disulfide bridges during the nascent phase.

Poor Enhancement of the Cysteine Reactivity is Due to pKa Perturbations
An extended study, which examined the pKa values of cysteines in many polypeptides, determined 9.1 to be a likely pKa value for an unperturbed residue [8]. In fact, this value is close to that (8.9-9) which was found for reduced glutathione [8], which contains a cysteine residue in which its carboxylic and amino groups are not free, but are rather engaged in covalent links with glycine and glutamic acid. At the physiological pH value of 7.4, an unperturbed protein cysteine should be present, like active thiolate, in a small but not negligible fraction (α) evaluated as 2% (see Equations (1)-(4)). Thus, the forced deprotonation of a protein cysteine due to a low pKa, possibly caused by near-positively charged residues, cannot increase its reactivity more than 50-times (see Section 4 for details).
However, the nucleophilicity of the thiolate decreases by lowering its pKa. At pH 7.4, for a reaction between alkylating compounds and various thiols with different pKa, a linear decrease should be observed by plotting the logarithm of the second order kinetic constant of the thiolate (kRS − ) onto the corresponding pKa of the thiols, according to Equation (5).
In our case, this linear relation was confirmed by observing the dependence of log kRS − on pKa in the reaction of different thiols (with different pKa) with 4-chloro-7-nitrobenzofurazan (NBD-Cl) and with 1-chloro-2,4-dinitrobenzene (CDNB): two well-known alkylating reagents (insets in Figure 1A,B at pH = 7.4 and Figure 1C,D at pH 5.0). (NBD-Cl) with thiols with different pK a (see Section 4 for details). Continuous lines of bell-shaped curves coming from Equation (6) are obtained from the best-fit of the linear regressions, as shown in the corresponding insets. (E,F) Dependence of the second order kinetic constants (α k RS − ) on pK a for the reaction of several thiols with different pK a with taurine chloramine and with different disulfides. The data were obtained from previous studies [9,10]. The experimental points of the linear regressions (insets) are the Mean ± S.D. of the three independent experiments. The red arrows mark the maximum values of the bell-shaped graphs. The pK a of the unperturbed protein cysteine is labelled with the green arrows.
The values of the Brønsted coefficient (β nuc ) and C (a constant applicable to a specific reaction involving various thiols and alkylating reagents) were 0.46 ± 0.07 (β nuc ) and −3.99 ± 0.56 (C) for CDNB and 0.53 ± 0.08 (β nuc ) and −2.13 ± 0.67 (C) for NBD-Cl at pH = 7.4. These values were inserted into the Equation (6), obtaining a bell-shaped behavior with a maximum at pK a ≈ 7.4 ( Figure 1A,B). This plot also indicates that an unperturbed protein cysteine with pK a = 9.1 cannot increase its reactivity more than three-four-times by lowering its pK a from 9.1 to 7.4. A further decrease of pK a below 7.4 results in a relevant decrease of reactivity. If these reactions are performed at pH = 5.0 (β nuc = 0.69 ± 0.07 and C = −5.48 ± 0.57 for CDNB, β nuc = 0.72 ± 0.06 and C = −3.10 ± 0.46 for NBD-Cl), a similar bell-shaped dependence is also observed, but the maximum of the reactivity was shifted to pK a ≈ 5.4 ( Figure 1C,D). These results parallel those obtained by Ferrer-Sueta et al. for the reaction of the different thiols involved in taurine chloramine reduction at pH = 7.4 ( Figure 1E) [9].
A similar variation of the nucleophilicity of the sulfhydryl group was also observed during the reaction of a few disulfides with free thiols or cysteine containing peptides showing different pK a . Even in that case, a linear decrease was found by plotting the logarithm of the second order kinetic constant of the thiolate onto the corresponding pK a of the thiols [9][10][11][12]. From these data, a bell-shaped graph was derived ( Figure 1F) as described in Section 4. That study reveals that an unperturbed protein cysteine with a pK a = 9.1 cannot increase its reactivity toward disulfides more than three-times at pH = 7.4.
Overall, our experimental data, together with preceding results, indicate that protein cysteines showing pK a ≤ 5.0 are almost unreactive toward various thiol reagents at physiological pH values.

Hydrophobic Interactions
In an interesting experimental set, performed on the reduced forms of a few disulfide-containing proteins and enzymes taken in their corresponding molten globule-like conformations, an extraordinary enhancement of the reaction rates of a few cysteines in their reaction with two thiol reagents with hydrophobic characters, i.e., CDNB and NBD-Cl (Table 1), was observed. As reported in Figure 2A, four cysteines in the reduced lysozyme (rLyz) which also display a low pK a (7.1) react with a thousand and a hundred-times higher reactivity with CDNB and NBD-Cl, respectively [13]. After removing the positive contribution of a lowered pK a ( Figure 2B) (about three-times), 572 and 121 are the enhanced reactivities normalized to the one of an unperturbed protein cysteine ( Figure 2C and Table 1). This extra-reactivity is probably due to hydrophobic interactions, as will be confirmed below. a E.R., enhanced reactivity is the second order kinetic constant for a given reaction, normalized to that of an unperturbed protein cysteine; b from Ref. [14]; c from Ref. [13]; d from Ref. [15]; e from Ref. [16]; f E.R. due to pK a from Ref. [17]; g unperturbed protein cysteine from Ref. [8]. Abbreviations: rBSA, reduced bovine serum albumin; CDNB, 1-chloro-2,4-dinitrobenzene; rChTg, reduced chymotrypsinogen; DTNB, 5,5 -dithiobis-(2-nitrobenzoic acid); GSH, reduced glutathione; GSSG, oxidized glutathione; rLyz, reduced lysozyme; NBD-Cl, 4-chloro-7-nitrobenzofurazan; rRNase, reduced ribonuclease.
The circle diagrams of Figure 2D compare the percent values of reactivity due to pK a variation and the hypothesized hydrophobic effects for both of these reagents.
The reduced form of chymotrypsinogen (rChTg) also displays a very similar enhanced reactivity that involves all the ten cysteine residues of this enzyme showing a lowered pK a of 8.1 ( Figure 3A-C) [16]. For this protein, the experiments were performed at pH = 5.0 due to the irreversible aggregation occurring at higher pH values [16]. The net contribution of the hypothesized hydrophobic interaction may be evaluated as 339-and 197-times enhanced reactivity for CDNB and NBD-Cl, respectively, while that which was due to the lowered pK a of these cysteines is only two-times ( Table 1).
The reduced form of chymotrypsinogen (rChTg) also displays a very similar enhanced reactivity that involves all the ten cysteine residues of this enzyme showing a lowered pKa of 8.1 ( Figure 3A-C) [16]. For this protein, the experiments were performed at pH = 5.0 due to the irreversible aggregation occurring at higher pH values [16]. The net contribution of the hypothesized hydrophobic interaction may be evaluated as 339-and 197-times enhanced reactivity for CDNB and NBD-Cl, respectively, while that which was due to the lowered pKa of these cysteines is only two-times ( Table 1).
The circle diagrams allow a percentage comparison of the two effects on the total reactivity of these cysteines with these two reagents ( Figure 3D). A support for the presence of hydrophobic interactions as the cause of the hyper-reactivities found in rChTg and rLyz toward CDNB and NBD-Cl was given by the relevant decrease of the reactivity observed by lowering the ionic strength ( Figure 4A,B), as was theoretically predicted [18].
A different and opposite behavior was observed for the reduced ribonuclease (rRNase) in its reactions with CDNB and NBD-Cl at different ionic strengths ( Figure 4C). This evidence and the modest hyper-reactivity of this reduced protein toward these reagents, indicates the absence of a pK a and hydrophobic effects on the observed hyper-reactivity in the reduced chymotrypsinogen. (A) Experimental kinetic data for the reaction of rChTg with CDNB and NBD-Cl at pH = 5.0, normalized to those of an unperturbed protein cysteine (enhanced reactivity) [16]. (B) Enhanced reactivity due to the experimental lower pK a for 10 cysteines (pK a = 8.1). (C) Enhanced reactivity due to hydrophobic interaction. (D) The circle diagrams show the percent contribution of the two factors: the hydrophobic effect (blue) and the pK a effect (red). The 100% represents the sum of the values reported in panels (B,C). The error bars in panels (A-C) are derived from the propagation of uncertainties (see Section 4).
A support for the presence of hydrophobic interactions as the cause of the hyper-reactivities found in rChTg and rLyz toward CDNB and NBD-Cl was given by the relevant decrease of the reactivity observed by lowering the ionic strength ( Figure 4A,B), as was theoretically predicted [18].
A different and opposite behavior was observed for the reduced ribonuclease (rRNase) in its reactions with CDNB and NBD-Cl at different ionic strengths ( Figure 4C). This evidence and the modest hyper-reactivity of this reduced protein toward these reagents, indicates the absence of a hydrophobic interaction which is able to speed up these reactions.  Finally, seven cysteines in the reduced form of bovine serum albumin (rBSA) also display 140times enhanced reactivity toward CDNB, when compared to that of an unperturbed protein cysteine ( Figure 5A) [14]. The expected maximum implementation, calculated on the basis of the pKa variations (pKa = 7.8 for 7 Cys), is about four-times ( Figure 5B). Thus, 35-times is the enhanced reactivity due to hydrophobic interactions ( Figure 5C and Table 1). Finally, seven cysteines in the reduced form of bovine serum albumin (rBSA) also display 140-times enhanced reactivity toward CDNB, when compared to that of an unperturbed protein cysteine ( Figure 5A) [14]. The expected maximum implementation, calculated on the basis of the pK a variations (pK a = 7.8 for 7 Cys), is about four-times ( Figure 5B). Thus, 35-times is the enhanced reactivity due to hydrophobic interactions ( Figure 5C and Table 1 The circle diagram visualizes the percent contributions of the hydrophobic and pKa effects ( Figure 5D).

The Peculiar Interaction with 5,5′-Dithiobis-(2-Nitrobenzoic Acid)
The 5,5′-dithiobis-(2-nitrobenzoic acid) (DTNB), a well known thiol reagent, displays both polar and non-polar characteristics in the same molecule (two aromatic rings and two ionized carboxylate functions). This particular nature is the origin of its interesting interaction with all of the tested reduced proteins in their molten globule-like structures. As reported in Table 1, a few cysteines of rLyz (7 Cys), rChTg (10 Cys) and rBSA (7 Cys) are hyper-reactive toward this reagent, with a hundred-times enhanced reactivity. The different numbers of hyper-reactive cysteines and the different extents of their increased reactivity compared to those seen with CDNB and NBD-Cl is probably due to the presence of the charged carboxylate. The influence of ionic interactions in the protein cysteine reactivity has been described in the past [5,19] and it appears to be crucial in the case of the rRNase. This protein does not display any hyper-reactivity toward CDNB and NBD-Cl, but almost all of its eight cysteines react with DTNB with 125-times increased reactivity. This phenomenon is likely due only to the electrostatic positive interactions of this reagent with charged residues surrounding all the protein cysteines. The effect of the ionic strength on this hyperreactivity-exactly the opposite of that shown with CDNB and NBD-Cl ( Figure 4C)-confirms a prevalent electrostatic effect. A similar conclusion can be drawn for rChTg and rLyz (Figure 4 A,B).

CD Analyses and ANS Fluorescence Fulfill Further Insights
The circular dichroism (CD) spectra for albumin, lysozyme, ribonuclease and chymotrypsinogen, both in their native and in their reduced molten globule-like status, have been reported in previous studies [13][14][15][16]. Now, a more detailed analysis, performed using the BeStSel program [20], can give us further insights about the molten globule-like structures. As shown in Table  2, the conversion of the native oxidized form of albumin, chymotrypsinogen and lysozyme into a The circle diagram visualizes the percent contributions of the hydrophobic and pK a effects ( Figure 5D).

The Peculiar Interaction with 5,5 -Dithiobis-(2-Nitrobenzoic Acid)
The 5,5 -dithiobis-(2-nitrobenzoic acid) (DTNB), a well known thiol reagent, displays both polar and non-polar characteristics in the same molecule (two aromatic rings and two ionized carboxylate functions). This particular nature is the origin of its interesting interaction with all of the tested reduced proteins in their molten globule-like structures. As reported in Table 1, a few cysteines of rLyz (7 Cys), rChTg (10 Cys) and rBSA (7 Cys) are hyper-reactive toward this reagent, with a hundred-times enhanced reactivity. The different numbers of hyper-reactive cysteines and the different extents of their increased reactivity compared to those seen with CDNB and NBD-Cl is probably due to the presence of the charged carboxylate. The influence of ionic interactions in the protein cysteine reactivity has been described in the past [5,19] and it appears to be crucial in the case of the rRNase. This protein does not display any hyper-reactivity toward CDNB and NBD-Cl, but almost all of its eight cysteines react with DTNB with 125-times increased reactivity. This phenomenon is likely due only to the electrostatic positive interactions of this reagent with charged residues surrounding all the protein cysteines. The effect of the ionic strength on this hyper-reactivity-exactly the opposite of that shown with CDNB and NBD-Cl ( Figure 4C)-confirms a prevalent electrostatic effect. A similar conclusion can be drawn for rChTg and rLyz ( Figure 4A,B).

CD Analyses and ANS Fluorescence Fulfill Further Insights
The circular dichroism (CD) spectra for albumin, lysozyme, ribonuclease and chymotrypsinogen, both in their native and in their reduced molten globule-like status, have been reported in previous studies [13][14][15][16]. Now, a more detailed analysis, performed using the BeStSel program [20], can give us further insights about the molten globule-like structures. As shown in Table 2, the conversion of the native oxidized form of albumin, chymotrypsinogen and lysozyme into a reduced molten globule-like structure causes a loss of the alpha helix and an accompanying increase in the beta sheet content. This evidence only suggests some changes of the secondary structures, but nothing about an increased exposition of hydrophobic residues which are able to promote a productive binding of CDNB and NBD-Cl. The considerable fluorescence emission of 8-anilinonaphthalene-1-sulfonic acid (ANS) in the presence of the reduced protein, compared to that observed with the oxidized protein (Table 3), represents a consistent proof that the semi-flexible nature of the molten globule status permits some internal non-polar residues to become exposed to the solvent, making the surface more hydrophobic [21]. Thus, this conformation can bind non-polar molecules much more strongly than the native one, thereby favoring the observed enhanced reactivity toward CDNB and NBD-Cl. Conversely, the rRNase displays a secondary structure similar to the oxidized form and only a negligible binding to ANS. These finding could explain the absence of hyper-reactivity toward CDNB and NBD-Cl (Table 1).

Productive Transient Complex
Beside hydrophobic and electrostatic interactions, other factors may be present in the reduced forms of the proteins, which are able to greatly increase the cysteine reactivity. By examining the reaction of the reduced molten globule-like structures of the four tested proteins with oxidized glutathione (GSSG), we found that a single specific cysteine in each protein displays thousands-times higher reactivity than an unperturbed protein cysteine (Table 1). This phenomenon has been observed in rBSA, rLyz, rRNase and rChTg, for Cys75, Cys94, Cys95 and Cys1, respectively [13][14][15][16]. This hyper-reactivity is specific for GSSG, because a normal reactivity has been found toward other natural disulfides like cystine and cystamine (see Table 1).
This phenomenon is mainly due to a productive transient protein-GSSG complex, as proven by fluorescence and kinetic experiments which also defined the dissociation constants of this interaction (K D = 0.3 mM for rLyz, K D = 0.12 mM for rRNase and K D = 1.5 mM for rChTg) [13][14][15][16]. In other words, a specific region of the molten globule-like conformations of these proteins resembled an enzyme active site which was able to bind productively with GSSG in order to react with a selected cysteine.
In Figure 6A-C, we report the reactivity toward GSSG of a single cysteine of these four enzymes, compared to that of an unperturbed protein cysteine (pK a = 9.1) together with the enhanced reactivity subtracted by the contribution of a lowered pK a . In particular, these were pK a = 6.6 for Cys75 (rBSA) and Cys94 (rLyz), pK a 7.9 = for Cys95 (rRNase) and pK a = 8.1 for Cys1 (rChTg) ( Table 1) The circle diagrams ( Figure 6D) allow a percent comparison of the effects due to a transient complex formation and pKa on the total reactivity of these cysteines with GSSG.

Discussion
All of the above data lead to the conclusion that pKa perturbations represent only a very small contribution for the enhancement of the reactivity in many types of reaction involving a protein cysteine [22]. Previous data [9][10][11][12] gave this indication for thiol-disulfide reactions and for the interaction of thiols with peroxide, but now we have extended the analysis to the reactions involving alkylating compounds. For all of these types of reactions, only three-four-times increased reactivity can be achieved; these are very modest implementations when compared to those found for many functional cysteines exhibiting thousand-times increased reactivity.
Paradoxically, forced deprotonations leading pKa < 7.0 provide significant decreases in reactivity. Many studies have often referred to this parameter to justify unusual kinetic properties and this conclusion can be revised. In this paper, we have shown an evident extraordinary effect due to a hydrophobic interaction between two thiol reagents (CDNB, NBD-Cl) and a few structural cysteines devoted to the formation of disulfides in the native proteins (i.e., bovine serum albumin, lysozyme and chymotrypsinogen). It appears that the unknown structures of the molten globule-like conformations of these proteins expose most of their cysteines to these reagents with a concomitant nearness of hydrophobic amino acids, which in the native conformations are normally masked and confined inside the polypeptide radius. The relevant increase of beta sheets in the molten globulelike proteins (Table 2) and the increase of ANS fluorescence in the presence of rChTg and rLyz ( Table  3) are signals of the widespread exposed hydrophobicity of these structures, favoring the interaction Figure 6. pK a and hydrophobic effects on the observed hyper-reactivity toward GSSG in different proteins. (A) Experimental kinetic data for the reaction of Cys75 for rBSA, Cys94 for rLyz, Cys95 for rRNase and Cys1 for rChTg with GSSG normalized to those of an unperturbed protein cysteine (enhanced reactivity) [13][14][15][16]. (B) Enhanced reactivity due to the experimental lower pK a of the protein cysteines of panel (A) (pK a values: 6.6 for Cys75, 6.6 for Cys94, 7.9 for Cys95 and 8.1 for Cys1) [13][14][15][16]. The circle diagrams ( Figure 6D) allow a percent comparison of the effects due to a transient complex formation and pK a on the total reactivity of these cysteines with GSSG.

Discussion
All of the above data lead to the conclusion that pK a perturbations represent only a very small contribution for the enhancement of the reactivity in many types of reaction involving a protein cysteine [22]. Previous data [9][10][11][12] gave this indication for thiol-disulfide reactions and for the interaction of thiols with peroxide, but now we have extended the analysis to the reactions involving alkylating compounds. For all of these types of reactions, only three-four-times increased reactivity can be achieved; these are very modest implementations when compared to those found for many functional cysteines exhibiting thousand-times increased reactivity.
Paradoxically, forced deprotonations leading pK a < 7.0 provide significant decreases in reactivity. Many studies have often referred to this parameter to justify unusual kinetic properties and this conclusion can be revised. In this paper, we have shown an evident extraordinary effect due to a hydrophobic interaction between two thiol reagents (CDNB, NBD-Cl) and a few structural cysteines devoted to the formation of disulfides in the native proteins (i.e., bovine serum albumin, lysozyme and chymotrypsinogen). It appears that the unknown structures of the molten globule-like conformations of these proteins expose most of their cysteines to these reagents with a concomitant nearness of hydrophobic amino acids, which in the native conformations are normally masked and confined inside the polypeptide radius. The relevant increase of beta sheets in the molten globule-like proteins (Table 2) and the increase of ANS fluorescence in the presence of rChTg and rLyz ( Table 3) are signals of the widespread exposed hydrophobicity of these structures, favoring the interaction with CDNB and NBD-Cl. The particular effect of the ionic strength on the reactivity of these cysteines toward these reagents is a further proof of the presence of hydrophobic interactions as the cause of the observed hyper-reactivity. A second accelerating factor must be mentioned, i.e., the electrostatic interaction, which seems to be the origin of the hyper-rectivity toward DTNB observed in rRNase, rLyz and rChTg.
As a third and more evident kinetic-enhancing factor, we discovered the specific interaction of GSSG with only one cysteine residue of each of these proteins. Here, the increased reactivity is evaluated in terms of thousand-times with a stringent specificity: no other small natural disulfides, like cystine and cystamine, show similar reactivities [13][14][15][16]. This hyper-reactivity toward GSSG could reveal a primordial event inside the oxidative folding of many disulfide containing proteins, given that GSSG is present at millimolar concentrations in the endoplasmic reticulum [23]. This should indicate that the glutathionylation of these cysteines could represent a useful "incipit" of their oxidative folding avoiding protein aggregation, as demonstrated for lysozyme [13], or to promote a hierarchical disulfide bond formation, as found in albumin [14].
In conclusion: i. pK a is not the main determinant in the enhancement of the reactivity of protein cysteines toward various reagents. Conversely, a very low pK a , as well as a very high pK a , may render unreactive these residues (see Figures 1 and 7). What is the utility of some functional cysteine showing very low pK a , such as selected residues in DsbA (pK a = 3.5), DsbC (pK a = 4.1) and Grx1 (pK a = 3.5) [24]? One reasonable explanation is that this property accelerates the reaction of the oxidized form of these enzymes with the thiol substrates stabilizing the products [24]. Another possibility is that a very low pK a that makes the thiolate less reactive and may preserve it against some unproper modifications. This may be the case for GSTP1-1, where the thiolate of Cys47 (pK a = 3.5) is bound to Lys54 in an ion-pair, which is important for the enzyme mechanism and a correct binding of the substrate [25]. ii.
Cysteine hyper-reactivity is not an exclusive property of functional cysteines involved in catalysis and even structural cysteines devoted to the formation of disulfides may display hundred-or thousand-times increased reactivity toward GSSG and various thiol reagents. iii.
Hydrophobic interactions are the main determinant factors triggering hyper-reactivity toward CDNB and NBD-Cl for rBSA, rChTg and rLyz, while electrostatic interactions are the prominent factors for the reactivity of DTNB toward rRNase, rLyz and rChTg. iv.
A specific binding site for GSSG is surprisingly present in the reduced molten globule-like conformations of albumin, lysozyme, chymotrypsinogen and ribonuclease. It is the main determinant for the observed hundred-and even thousand-times increased reactivity of one specific cysteine. This phenomenon raises the question of whether a rapid glutathionylation may be the early step of their oxidative pathway. v.
Methods for the proteomic identification of cysteines, like the isoTOP-ABPP procedure [1,26], should be used with caution, because they only identify hyper-reactive cysteines toward a specific reagent (i.e., a modified iodacetamide) and this property cannot be referred to an 'intrinsic reactivity' because it may be not present in reactions with different thiol reagents. Conversely, some protein cysteines, which are normo-reactive toward the modified iodoacetamide probe, can be hyper-reactive toward some natural intracellular compounds. Our data, in fact, likely indicate that one cysteine may have extraordinary hyper-reactivity toward a specific disulfide (GSSG) and normo-reactivity toward other small disulfides, like cystine and cystamine. Conversely, many cysteines which are present in rBSA and rLyz are hyper-reactive toward hydrophobic reagents like CDNB and NBD-Cl, but (except for one residue) are normo-reactive toward GSSG and other small disulfides. In other words, the "intrinsic" reactivity for a protein cysteine is only determined by its pK a and by the nucleophilicity of its deprotonated form, but it cannot be increased more than three-four-times, as demonstrated in this paper. An evident hyper-reactivity can only be generated by "extrinsic" factors like the protein environment surrounding the cysteine, which may productively and often selectively bind a specific reagent through hydrophobic or electrostatic interactions. This selective hyper-reactivity should be of particular interest in the elucidation of the early step of the oxidative folding of these proteins. The hyper-reactivity of protein cysteines appears to be an open puzzle whose pieces, even now, have not been completely identified. This paper may be a useful contribution to this scenario.

Reactivity of rChTg Cysteines Toward Alkylating Reagents and DTNB Varying the Ionic Strength
The chymotrypsinogen (ChTg) reduction was performed as previously reported [16].  This selective hyper-reactivity should be of particular interest in the elucidation of the early step of the oxidative folding of these proteins. The hyper-reactivity of protein cysteines appears to be an open puzzle whose pieces, even now, have not been completely identified. This paper may be a useful contribution to this scenario.

Reactivity of rChTg Cysteines toward Alkylating Reagents and DTNB Varying the Ionic Strength
The chymotrypsinogen (ChTg) reduction was performed as previously reported [16]. In brief, 8 mg protein was solubilized in 8 M urea, 1 mM EDTA, 10 mM sodium borate buffer (pH = 8.5) and the ChTg cystine was reduced by adding dithiotreitol (DTT) (ChTg:DTT = 1:21) at 60 • C for 50 min. The DTT excess was removed through a Sephadex G-25 column (1 × 20 cm) equilibrated with 8 M urea, 1 mM EDTA and 10 mM acetate buffer (pH = 5.0). An aliquot of this reduced protein was assayed with DTNB. All of the 10 -SH/mole were titrated at pH = 5.0 in the presence of 0.2 M urea.
The second order kinetic constants for the reactions of the rChTg with CDNB were determined spectrophotometrically, continuously at 340 nm, where the Cys-DNB adduct absorbs (ε = 9.6 mM −1 cm −1 ). The concentrations were from 1 to 5 µM rChTg and 0.

Reactivity of rLyz Cysteines toward Alkylating Reagents and DTNB Varying the Ionic Strength
The lysozyme (Lyz) reduction was previously reported in [13]. In brief, Lyz (0.

Reactivity of rRNase Cysteines toward Alkylating Reagents and DTNB Varying the Ionic Strength
The ribonuclease (RNase) reduction was performed as previously reported in [15]. In brief, RNase

Data Analysis and Graphical Representation
The only active form of a thiol group is its deprotonated form; thus, for an unperturbed protein cysteine with a pK a = 9.1 [8] [ At pH = 7.4, α is 0.020; thus, a lowered pK a of the sulfhydryl group, which makes an almost fully dissociated cysteine (α = 1), cannot cause a kinetic increase higher than 50-times at physiological pH values.
A typical reaction of a protein or free cysteine with alkylating reagents can be schematized as: CysSH + RX → no reaction The pH-independent rate constant k RS − of the sulfhydryl group is obtained by dividing the observed/experimental rate constant (k obs ) at a given pH by the fraction of the dissociated sulfhydryl (α) at the same pH: k RS − = k obs/α = k obs 1 + 10 pK a −pH (4) The logarithm of the rate constant is linearly correlated with the thiol pK a according to the Brønsted relationship: log k RS − = β nuc pK a + C where β nuc is the Brønsted coefficient and C is a constant applicable to a specific reaction involving various thiols and an alkylating reagent. The data of the linear regressions (insets of Figure 1) are expressed as Mean ± Standard Deviation (S.D.). The data were obtained from three independent experiments. To obtain the bell-shaped graphs reported in Figure 1, Equations (4) and (5) were combined: αk RS − = k obs = 10 β nuc pK a +C−log (1+10 pKa−pH ) Parameters (β nuc and C) derived from the linear correlations between log k RS − and peptide thiols pK a in the thiol-disulfide exchange reactions [10] were inserted into Equation (6), thus obtaining a bell-shaped graph ( Figure 1F). A similar procedure was used for the analysis of the taurine chloramine reduction by several small thiols [9], obtaining the bell-shaped graph shown in Figure 1E. The experimental data from the literature [9,10] were digitalized using GetData Graph Digitizer software (v2.24) (ShareIt, Germany). The propagation of uncertainties was analyzed according to the classical statistical methods [28]. The graphics and the results visualization were obtained using GraphPad Prism software v5.0 (GraphPad Company, La Jolla, CA, USA).

Acknowledgments:
The authors would like to thank Jens Z. Pedersen for helpful discussion and Silvia Ticconi and Sara Notari for lab assistance.

Conflicts of Interest:
The authors declare no conflict of interest.