In the present work, for the first time the expression protein Response Surplus or RS is introduced for the purpose of designating the increase of antioxidant capacity of a protein when it undergoes the effect of a structural stressor. With the objective of obtaining a measurement of the RS, the antioxidant capacity (AC) of a protein sample was determined before (ACb) and after (ACa) an oxidative or structural challenge; AC and ACb are then equivalents. The results obtained in Trolox concentration (Trolox Equivalent Units, nM TEU) with the use of standard curves are transformed to %, considering ACb to be 100%. In this way the values reported as RS represent the percentage of antioxidant capacity surplus above the average antioxidant capacity of proteins before treatment (native protein). In the same way the antioxidant capacity accumulate % (ACA %) represents the sum of antioxidant capacity AC plus the RS value in percentage. In the present work, the ACA % appears above the bars as % or as a series of % values, considering the average antioxidant capacity of the native sample tested to be 100%. In this manner, a quick way to obtain RS from the results is to subtract 100 from ACA %.

AC, Antioxidant Capacity. Trolox nM concentration (Trolox Equivalents Units, TEU)

ACA%, Antioxidant Capacity Accumulate %  =  (AC after) (100) ÷ AC before

RS%, Response Surplus %  =  (ACA %)−100

Antioxidant capacities (AC) were measured and Response Surplus (RS) calculated in human serum albumin (HSA) with the use of a chemiluminiscence system described in the Material and Methods section, and the results are summarized in Figure 1 . The native HSA showed an average AC of close to 90 nM TEU (Trolox Equivalent Units), but it more than doubled when the protein was previously incubated for 10 minutes with two strong oxidants: sodium Hypochlorite (NaOCl) 2 mM and hydrogen peroxide (H₂O₂) 500 µM (250 and 225 nM TEU respectively). The same pattern was obtained when protein was exposed to UV light (254 nm) for 15 seconds (p<0.01) (250 nM TEU). The Antioxidant Capacity Accumulated in % (ACA %) corresponds to 270%, 226%, and 273% respectively, and appear in parentheses above the bars. RS % corresponds to ACA % -100, as was described previously. There are significant differences between the treated groups and the native protein antioxidant capacity (p<0.001), but there are no significant differences between the treated groups in terms of TEU, ACA % or RS % values.

Such as oxidative stress, thermal treatment produces changes in the AC and RS of albumin. The thermal denaturation process of HSA is comprised of two stages [20]. The first is a reversible thermal structural alteration of the protein, and the second a state that results in an irreversible structural change at different temperatures dependent of albumin concentration [21]. We find albumin aggregation (light scattering at 400 nm) close to 50°C with an albumin concentration of 100 mg/ml. In the present work, with the purpose of testing the antioxidant properties derived from thermal structural changes, we studied the antioxidant behavior of albumin in a temperature range of 10 to 50°C. The results demonstrate clearly that albumin changes its antioxidant properties within a reversible thermal denaturation process in a temperature-dependent manner ( Figure 2 ). When the native HSA was set at 10°C, it yielded an AC of close to 50 TEU (which corresponds to 100% of ACA in the graph). From this point the antioxidant capacity increased gradually with the temperature, until it reached almost 200% ACA as it approached 50°C, just before it was possible to see evidence of aggregation. From this point the antioxidant capacity apparently diminished, probably as a consequence of the formation of intermolecular disulphide bridges by free SH groups, which plays an essential role in aggregation process [21]. However, some extent of chemiluminiscence light interference with turbidity cannot be discarded.

In independent experiments, when the denaturation temperature was maintained at 65°C, the antioxidant capacity showed a particular and repetitive pattern of fluctuation several seconds before protein aggregation and light scattering at 400 nm ( Figure 3 ). We speculate that the fluctuations observed could correlate with a transient fluctuation pattern of structural unfolding and refolding. However, in contrast to the ACA of close to 200% at about 50°C (Figure 2), the changes observed here never exceed 10% of ACA.

Experiments to explore the effect of proteolysis and hence the structural integrity of the AC and RS of albumin showed that AC is increased as a result of hydrolytic activity of proteinase K ( Figure 4 ). However, it was clear that the magnitude of the response was partial compared to the result observed for normal albumin (see Figure 1). Incubation of 15 minutes with proteinase K increases the AC from 187 (native) to 228 (treated) TEU (21% of RS) and 60 minutes of incubation with the proteinase K changes the AC from 187 (native) to 270 (treated) TEU (44% of RS). This means that responses of 1.21 and 1.44 times those of native albumin (including the contribution of the same proteinase K) were obtained. Although 60 minutes of incubation with proteinase K virtually totally hydrolyzed the albumin in the experimental conditions used, the maximum antioxidant capacity achieved in this way does not explain the magnitude of the change in AC reached with stressors like Reactive Oxygen Species (ROS) and UV light ( Figure 1 ). It is possible to suppose that with the hydrolysis reactive residues from the structure of a protein such as the SHs groups could be liberated, promoting reactions of these with free radicals as described in the next section. In any case, these results implied that structural molecular integrity is important to proteins if they work as antioxidants (AC) and when they face a stressor (RS).

It has been proposed that in living cells the protein thiols could be directly involved in the cellular defense mechanism against oxidants [22]. However, that would imply that sufficient SHs were available to be oxidized. New reports confirm that most cellular protein SHs equivalents are found in the reduced state and that protein thiols are directly involved in the cellular defense mechanism against oxidants [23]. In the present study, the contribution of the reducing equivalents on the antioxidant capacity of albumin was achieved by using the reducer agent DTT ( Figure 5 ). As was expected, results indicated that reduced albumin increased its AC to a high magnitude (ACA% = 600%); an equivalent amount was reproduced with an exposure of 35 seconds to UV light. It is noteworthy that the partial contribution of both stressors and the cumulative character of these effects (DTT+UV) reached an antioxidant capacity close to 13 times that of normal albumin. This data indicate that although SHs groups are important reactive residues and key to the RS phenomenon, the stressor (and the structural response induced on proteins) is necessary to explain the origin of the antioxidant capacity remaining after DTT exposure. An hour of UV light (plus DTT) did not further increase the antioxidant capacity.

With respect to the previous results, in the present work RNase A (ribonuclease A, 124 residues, ∼13.7 kDa) was used as a means of understanding better the nature of RS and its relation with the SHs groups in the presence of UV light as stressor ( Figure 6 ). The antioxidant capacity of ribonuclease challenged with UV light yielded relatively low RS %, reaching around 16% with 5 minutes of exposure time (ACA% = 116%), but surprisingly, ribonuclease in native state presented a high level of AC (300 nM TEU) ( Figure 6 , internal graph). It is noticeable because in this state ribonuclease does not normally present free SHs groups. Indeed, when an attempt was made to measure the SHs, detectable amounts were not obtained ( Figure 6 , t = 0). These results indicated that the stressor liberates SHs groups from the protein structure.

The behaviors of ribonuclease in relation to a high AC and relatively low RS can be explained if one aspect of the system is taken in account: all systems for measuring the antioxidant activity are sources of free radicals and stressors. It is possible to hypothesize that some proteins should be more sensitive to the free radicals derived from the same system. If the former is true, a rapid induction of RS could be produced when the ribonuclease is placed with the reaction mix. This would be an unavoidable consequence of the use of the system for some proteins. However, it seems clear that passive intervention of SHs groups does not by itself explain the high antioxidant capacity observed (around 300 TEU) because there were no detectable free SHs groups in native ribonuclease.

With the objective of looking for quantitative differences from other proteins in relation to AC and RS, three non-related proteins were tested and the results are presented in Figure 7 . The AC resulted in 27 nM TEU for the carbonic anhydrase (Isoform CA-I, 29 KD), 85 nM TEU for the Bovine Serum Albumin (Fraction V, 66.4 KD) and 106 nM TEU for Human Insulin (recombinant 5.8 KD). The change of AC by exposure to UV light (254 nm) for 1.5 minutes resulted in 66, 220 and 283 TEU for Carbonic anhydrase, BSA and Insulin respectively. The corresponding values of ACA % for the three proteins in the same order were 240%, 259% and 264% (shown above the bars in Figure 7 ), and correspond to % of RS of 40, 25.9 and 26.4. There are significant differences between the antioxidant capacities of the proteins before and after exposure to UV light (p<0.01). No correlation between the antioxidant capacity and the molecular size or amino acid sequence of the proteins tested could be identified. However, it is possible to see first, an inverse relationship between antioxidant capacity and the protein's half life (Figure S1) and second, a correlation between antioxidant capacity and the total amount of cysteine residues per gram of protein calculated (Table S1). However, the former result does not explain completely the differences presented in antioxidant capacity, because no single or minimal amount of cysteins is normally presented in reduced state. As in the previously explained case of ribonuclease, we do not exclude the possible intervention of the ROS derived from the system in the induction of the antioxidant response. We hope data presented here will stimulate interest in investigating this aspect more deeply.

In the present work two chemical modifications of albumin were tested to learn how the changes introduced affect the antioxidant capacity and RS in response to a stressor such as UV light. Incubation of albumin with acrolein results in the adduct formation with lysine [24], whereas incubation of albumin with glucose produces several modifications to protein residues grouped under the term glycation or non-enzymatic glycosylation. Previous reports of in vitro modifications of bovine serum albumin (BSA) by methylglyoxal [15] and glucose [16] indicate that these modifiers decrease content of the sulphydryl groups, impairing the antioxidant properties of albumin and inducing conformational changes attributed to diminishing of the fluorescence of the tryptophan residues.

The evidence indicates that the protein modifiers used here increase the antioxidant capacity of the albumin ( Figure 8 ). The structural changes produced by protein-acrolein adduct formation and the non-enzymatic attachment of glucose to free primary amine residues seem to be stressors capable of inducing an incipient antioxidant response, increasing the AC by up to 4 times the native albumin. In contrast, these same changes produced a drastic reduction in the RS of the albumin ( Figure 9 ) when it was previously challenged with the exposure of UV light (254 nm). One minute of exposure to UV light increased the ACA % 8 times, to 800%, whereas albumin treated previously with acrolein and glycosilated albumin in the same conditions increased the ACA % by only 1.9 and 2 times respectively (ACA% = 190 and 200%). When the previous exposure time to UV light was increased to 3 minutes native albumin reached more than 11 times the antioxidant capacity of native albumin (ACA% = 1100), whereas acrolein-modified and glycosilated albumin achieved only 2.2 and 3.2 times the antioxidant capacity presented before UV exposure (ACA% = 220 and 320% respectively). The results previously described emphasized the importance and complexities of the relationship between antioxidant behavior and the protein structure, which produce a partial increase but a remarkable total decrease of protein antioxidant capacity.

The reciprocal relationship between molecular stability and flexibility, two elements of the dynamic of molecular changes, used to explain loss of biological activity for proteins [25], [26], could help us to understand the induction of an incipient antioxidant response and the important reduction in the RS component observed here with albumin modified with acrolein and glucose. If the changes introduced in albumin increase the protein stability (rigidity of the molecule), a reduction of flexibility and then a corresponding reduction in the RS component can be expected, as was observed here ( Figure 9 ). The loss of flexibility apparently does not exclude the initial increased AC produced by the structural modifiers, possibly because it was derived from some intact Cys 34 residues.

With the objective of revealing changes produced in the native structure of albumin by the modifications described previously, binding studies were performed with the hydrophobic compound 8-Anilino-1-naphthalene sulfonate (ANS), a sensitive probe for partially folded intermediates in protein-folding pathways. Results reveal how the fluorescence of ANS increases substantially when it is bound to normal albumin, (providing evidence of a partially folded state in the commercial albumin). In contrast, the same albumin modified with acrolein and glycosilated albumin both reduced substantially the fluorescence of ANS excited by UV light. This was probably due to structural changes and the reduction of hydrophobic core regions that are inaccessible to the dye ( Figure 10 ). Independent experiments confirm quantitative reductions in the ANS fluorescence (Ex: 355 Em: 460) of modified albumin ( Figure 11 ). A reduction in hydrophobicity of albumin showed a positive correlation with the AC; however, a drastic reduction in the RS accompanied the same change in hydrophobicity ( Figures 8 and Figure 9 , respectively). Changes in ANS fluorescence bonding to albumin as a result of previous albumin transient UV light exposure were negligible. Taken together, these data indicate how albumin responds to different stressors with more antioxidant capacity and how pronounced and permanent changes in the structure can negatively modify the antioxidant response (RS).

It has been demonstrated that the free thiol groups in HSA (Cys-34) can interact with reactive oxygen species (ROS) like hydrogen peroxide (H₂O₂) and peroxinitrite (ONOO-) to form a sulfenic acid derivative (HSA-SOH) [10]. Sulfenic acid has been described as very probably the most potent of all peroxyradical trapping antioxidants [27]. The sulfenic acid can be oxidized to sulfinic acid (RSO₂H) and sulfonic acid (RSO₃H), and can also react with thiols, providing a mechanism for disulfide formation [28]. In addition, sulfenic acid can undergo self-condensation to yield the correspondent diallyl thiosulfinate; these can undergo Cope elimination at room temperature to give 2-propensulfenic acid and thioacrolein [27]. From the diversity of possible reactions, some of these make a hypothetical cycle possible.

Using some previously described information and evidence from the present work it is possible to construct a theoretical model to explain the possible mechanism of the antioxidant capacity and RS of proteins. We present a specific alternative, and try to simplify and summarize the available information.

In accordance with this model (Figure 12), a stressor introduces instability into the core of a protein that promotes the transition (structural perturbation) between disulfide and SH groups. In the present work, thermal treatment and UV light as stressors and DTT all increased the antioxidant capacity of albumin by a high magnitude ( Figure 1 , Figure 2 and Figure 5 ), whereas in addition ribonuclease exposure to UV light clearly liberates SHs groups that then were detected ( Figure 6 ). Thiols will then interact with reactive oxygen species (ROS), re-oxidize or form mixed disulfides. Although it is a critical point in the mechanism and represents a primer, it is strictly part of the passive antioxidant capacity of a protein. In contrast, in a second dynamic phase, a sulfenic acid derivative formed from the reaction between ROS and thiols can reduce free radicals such as peroxyl radicals (ROO) with high efficiency, as has been demonstrated [27], generating sulfinyl radical derivatives (RSO) as the product of the reaction. Then, a second stock of sulfenic acid can be recycled from RSO^. if enough SHs are available, which will be converted to a new S-S group ( Figure 12 ). It is possible to assume that flexibility is necessary, allowing the proposed instability induced by the stressor. In addition, data from experiments with ANS it is possible to suggest that hydrophobic molecular regions should be implicated in the antioxidant response of proteins, although quantitative differences should exist between native and non-native albumins.

If the native configuration represents the more favorable energy state for a protein in accordance with Anfinsen's thermodynamic hypothesis [18],[19], then it is possible to hypothesize that the thermodynamically most stable native state must represent the more favorable antioxidant state and the driving force that moves the protein thiol groups towards the most favorable location. It is very important to explain the tight interdependence between structure and antioxidant function, but in addition it could help us to understand why a protein assumes its particular configuration.

No matter what the precise mechanism of protein antioxidant capacity is, including the RS component, the existence of this extraordinary behavior has several implications by itself. The entire antioxidant capacity of a system such as serum, plasma or any tissue may be conveniently evaluated only if the entire antioxidant capacity of proteins is taken into account. The antioxidant capacity of any biological component is apparently much more complex than was previously supposed. A biological sample should present some of the several active elements with antioxidant capacity at the same time: antioxidant enzymes (SOD, glutathione peroxidase, etc.); primary antioxidants such as ceruloplasmin and ferritin; the soluble non-protein-dependent antioxidants (vitamin C, glutathione, etc.); a protein antioxidant capacity dependent on SH and other residues (passive component); and the Response Surplus (RS) described here (active component).

The importance of recognizing all the components of the antioxidant capacity becomes critical when the antioxidant status of a biological sample from patients suffering an illness is evaluated. A large number of studies support the hypothesis that oxidative damage to DNA, lipids and proteins may contribute to the development of cardiovascular disease, cancer, and other degenerative disorders such as arteriosclerosis, diabetes, Alzheimer's and also in ischemia/reperfusion injury, ethanol intoxication, liver steatosis and ageing. If a new component of the biological antioxidant capacity not previously considered is finally recognized, it will likely radically change the knowledge obtained from this theme. For example: the destabilization of a native protein in a critical site in the body should result in the drastic reduction of its antioxidant potentialities and should become the preliminary step before aggregation.

Additionally, RS should help us understand inconsistencies in previous reports when human plasma total antioxidant activity has been evaluated. Although some reports have emphasized the loss of albumin antioxidant activity as the result of structural changes [17], [29], it seems clear that the evaluated component in these works correspond to the relationship between mercaptalbumin (reduced form) and non mercaptalbumin (oxidized) alternative forms of HSA. From the present work, we can deduce that the evaluation referred to corresponds to the passive antioxidant component, attributable only to the oxidation of SH groups and in particular Cys 34. When in the work of Kawakami et. al. the hydroxyl radical scavenging activity of reduced and oxidized HSA was studied using ESR spectroscopy (ESR), reduced albumin [HSA (red) % = 73%] quenched up to 68.7% of the ESR hydroxyl radical signal, while the oxidized form [HSA (red) % = 8%] reduced the signal by 54.4% [17], and the authors concluded that oxidation of HSA diminished its radical scavenging activity. We suggest that the scavenging properties of HSA described are only partially interpreted. If the reduced albumin quenched 68.7%, then oxidized albumin must quench only 8% of the ESR signal and not 54.4% if the total antioxidant potential of the protein had been dependent only on the oxidation of SHs groups. We can see in these results indirect evidence of the existence of the antioxidant component RS. Oxidized albumin reduced the ESR signal more than predicted by the SHs proportion because an unpredictable component was present. An additional component such as RS is necessary to explain how a reduction of 65% in SHs groups of HSA can reduce less than 15% of its radical scavenging activity.

Finally, we found that some texts include reports of “unidentified antioxidant(s)” that represent a percentage of the total antioxidant capacity of the plasma [30], [31], [32], [33]. From the data presented in this work it is possible to suggest that the value of unidentified and not yet clearly defined antioxidant reported previously corresponds to the indirect influence of components of the protein antioxidant capacity and in particular RS. This is extremely important because even though a soluble non-protein component of a biological sample could be evaluated alone, the influence of a protein antioxidant reserve as a part of a whole should be decisive and must be taken into account.

In the present work the antioxidant capacity (AC) was measured using an enhanced chemiluminescence based assay. The reaction mechanism (Hydrogen Atom Transfer or HAT) allowed measurement of the potential of an antioxidant to quench free radicals by hydrogen donation. In real time it is possible visualize the individual contribution of each component to the total AC. Horseradish peroxidase (HRP) is an enzyme that catalyses the decomposition of peroxides and forms free radicals. Peroxidase/H₂O₂ mixtures have been used to generate free radicals [34] from several substrates. Hydrogen peroxide removes two electrons of the enzyme, and each of these is replaced in two one-electron steps, in each of which a substrate molecule forms a radical [35]. Oxygen consumption for the sodium horseradish peroxidase/hydrogen peroxide system and free radicals generation both have been documented [36]. Radicals formed, peroxides and oxygen oxidize luminol, producing light that can be detected and measured. Given that a constant rate of radical generation can be achieved, the light emission depends on constant production of free radicals. Enhanced chemiluminiscence has been used to measure the antioxidant capacity in biological fluids [37]. The system used in the present work contains Horseradish peroxidase (5 µU/l), hydrogen peroxide 30% w/w (3.0 µM), substrate p-iodophenol (25 µM) and luminol (300 µM from a stock solution of 10 mM in DMSO). A reaction mixture containing perborate base buffer prepared with sodium tetraborate (100 mM) and sodium carbonate (100 mM) was used. Intra-assay variation for fifteen samples analyzed in triplicate was 0.29–5.2%, with a mean of 4.05%. The luminometer display and complementary information are depicted in supporting information Figure S2.

Using the procedure described, a standard curve was achieved using increasing amounts of the water soluble tocopherol analogue Trolox (6- hydroxy-2, 5, 7, 8-tetramethylchroman 2-carboxylic acid). Standards of Trolox were prepared in milli-Q water (Millipore Corporation, Bedford, MA 01730, USA) before initiating each experiment and starting with a stock solution of 80 µM. The antioxidant capacity of test solutions through the experiments in the present work is expressed as nM of Trolox (Trolox Equivalent Units or TEU nM). An approximate amount of 320 nM of Trolox is enough to suppress completely the chemiluminiscent signal and correspond to the maximal antioxidant capacity (C) achieved. One standard curve was prepared before each experiment.

After the treatment, all protein samples were processed to remove the low molecular components of the reaction mixture. Before the incubation period, the protein samples were placed in microcentrifuge filters (10 KDa cutoff; 5 KDa cutoff for insulin) (Microcon, Millipore, Bedford, MA, USA), and centrifuged at 5000 g for 1 hour. The filter residues were washed and dialyzed two more times and dissolved in phosphate buffer pH 7.4. With the objective of removing possible components adsorbed to the proteins, samples were precipitated with 1.5 ml of chloroform and methanol (2∶1). After agitating thoroughly, samples were centrifuged at 3000 g for 10 minutes and the organic phase discarded. The protein residue was dried (nitrogen) and dissolved in PBS pH 7.4.

Finally, the samples were centrifuged at 5000 g for an additional 10 minutes and soluble residues separated; the protein concentration was determined using the method of Bradford [38]. Aliquots containing 60–500 µg of protein were added to the reaction mixture in the luminometer cuvette after the signal was stable.

The residues of solvent remaining after dialyze samples were collected and placed to react with Folin-Ciocalteu reagent to monitor protein fragmentation. In addition to amino acid residue observation, possible light scattering due to aggregation was measured at 400 nm in a spectrophotometer with the objective of taking control of possible aggregation of irradiated proteins.

For evaluation of the SHs punctual contribution to the antioxidant capacity of a protein, the free thiol groups of ribonuclease resulting from the UV treatment were measured using Ellman's reagent. Briefly, 5,5′-Dithiobis(2-nitrobenzoic acid) 2.5 mM in 0.2 M phosphate buffer pH 8.0 was mixed with 50 µl of sample (ribonuclease 1 mM) and 750 µl of 50 mM phosphate buffer pH 8.0. Freshly prepared Ellman's reagent (250 µl) was added and the reaction proceeded for 15 minutes at room temperature in the dark. The release of 5-mercapto-2-nitrobenzoic acid (MNB) was monitored at 412 nm in a Beckman DU-600 spectrophotometer. MNB was quantified using an extinction coefficient of 14150 mol-1 cm-1 [41]. The results were expressed in µmol/mg protein.

The binding of the hydrophobic probe ANS to protein was verified to assess small changes produced in the native structure of albumin [42]. Albumin samples prepared with acrolein and glucose as described previously were suspended in buffer at a concentration of 0.5 mg/ml in 50 mM HEPES, 100 mM KCL, pH 7.8. ANS was prepared just before assays and was used at a final concentration of 0.1 nM. The image of aqueous solutions of ANS excited by ultraviolet light was achieved using a BioDoc-It System with UV transilluminator (UVP, Inc., Upland, California, U.S.A.). Fluorescence measurements were performed on a BioTek plate reader (BioTek Instruments, Inc.). The excitation wavelength was set at 355 nm and the emission taken in 460 nm. The temperature was fixed at 30°C.

In the present work, groups were compared using the one-way ANOVA test followed by Dunet's test. Results were considered to be statistically different when p<0.01.