biomolecules Article Structural Lesions of Proteins Connected to Lipid Membrane Damages Caused by Radical Stress: Assessment by Biomimetic Systems and Raman Spectroscopy Armida Torreggiani 1, * , Anna Tinti 2 , Zuzana Jurasekova 3,4 , Mercè Capdevila 5 , Michela Saracino 1 and Michele Di Foggia 2 1 Istituto I.S.O.F. (C.N.R.), via P. Gobetti 101, 40129 Bologna, Italy; michela.saracino@isof.cnr.it 2 Dipartimento di Scienze Biomediche e Neuromotorie, Università di Bologna, via Belmeloro 8/2, 40126 Bologna, Italy; anna.tinti@unibo.it (A.T.); michele.difoggia2@unibo.it (M.D.F.) 3 Department of Biophysics, Faculty of Science, P.J. Safarik University, Jesenna 5, 04001 Kosice, Slovakia; zuzana.jurasekova@upjs.sk 4 Center for Interdisciplinary Biosciences, Technology and Innovation Park, P. J. Safarik University, Jesenna 5, 04001 Kosice, Slovakia 5 Departament de Quimica, Facultat de ciencies, Universitat Autònoma de Barcelona, 08193 Bellaterra, Catalonia, Spain; merce.capdevila@uab.cat * Correspondence: armida.torreggiani@isof.cnr.it; Tel.: +39-0516399821  Received: 31 October 2019; Accepted: 27 November 2019; Published: 27 November 2019  Abstract: Model systems constituted by proteins and unsaturated lipid vesicles were used to gain more insight into the effects of the propagation of an initial radical damage on protein to the lipid compartment. The latter is based on liposome technology and allows measuring the trans unsaturated fatty acid content as a result of free radical stress on proteins. Two kinds of sulfur-containing proteins were chosen to connect their chemical reactivity with membrane lipid transformation, serum albumins and metallothioneins. Biomimetic systems based on radiation chemistry were used to mimic the protein exposure to different kinds of free radical stress and Raman spectroscopy to shed light on protein structural changes caused by the free radical attack. Among the amino acid residues, Cys is one of the most sensitive residues towards the attack of free radicals, thus suggesting that metal-Cys clusters are good interceptors of reactive species in metallothioneins, together with disulfides moieties in serum albumins. Met is another important site of the attack, in particular under reductive conditions. Tyr and Phe are sensitive to radical stress too, leading to electron transfer reactions or radical-induced modifications of their structures. Finally, modifications in protein folding take place depending on reactive species attacking the protein. Keywords: free radical stress; lipid domains; protein damage; Raman spectroscopy; desulfurization; protein:lipid interaction; cis-trans isomerization; tandem lesions; thiyl radicals 1. Introduction The cis geometry of double bonds in lipids is a crucial feature for the functionality of cell membrane because it regulates the membrane properties such as permeability and fluidity, and membrane homeostasis is obtained by a precise balance between saturated and cis-unsaturated structures. On the contrary, trans geometry is unusual in nature and only some bacteria make use of an enzymatic cis-trans isomerization during the adaptation response [1]. The cis-trans isomerization can affect lipid assembly. Since the trans lipid geometry resembles that of saturated lipids [2], the geometrical isomerization can be considered equivalent to the decrease Biomolecules 2019, 9, 794; doi:10.3390/biom9120794 www.mdpi.com/journal/biomolecules Biomolecules 2019, 9, 794 2 of 18 of the degree of unsaturation in the cell membrane [3]. In addition, the trans lipid incorporation in membranes is also linked to a recently growing concern in nutrition [4]. The real culprits for the cis-trans isomerization in vivo have not been demonstrated yet, although many results support the correlation between the origin of endogenous formation of trans lipids and radical stress. Thiyl radicals RS• are the most likely candidates because of the efficiency of the thiyl radical-catalyzed cis-trans isomerization in vitro, and the presence of many sulfur-containing compounds within the cell. The role of thiyl radicals as radical stress inducers has emerged in the last decade [5,6], counteracting the long-standing and most well-known beneficial effects of thiols. The damage effect starts from the consideration that, in the so-called ‘repair reaction’, Cys residues can efficiently stop the radical cascade, by donating a hydrogen atom and trapping the radical intermediate of the chain. However, after this hydrogen donation, the thiol (RSH) is transformed into its corresponding thiyl radical (RS• ), which can, in turn, attack a substrate. Possible scenarios for the production of diffusible thiyl radicals were expanded by the discovery of free-radical damage to sulfur-containing proteins resulting in consequent isomerization of membrane lipids. In fact, a radical reaction with a primary target amino acid could lead to secondary radicals. These may engage in radical- and/or electron-transfer processes with additional protein amino acids, so that the ultimate location of a radical damage may evolve potentially far from the site of initial radical attack [7]. Thus, tandem lesions can occur from a single radical event, and at the same time produce reactive species able to damage another molecule. This tandem process can involve two different cell compartments, such as in case of the lipids with DNA molecules [8] or with proteins [9]. Under radical stress, thiyl radicals can be readily generated from a free thiol group by hydrogen abstraction, by cleavage of disulfide linkages and by the attack to methionine (Met) residues and metal-sulfur clusters. Thus, a desulfurization process takes place, coupled with the formation of S-centered radicals (CH3 S• or S•– ), which in turn can migrate from the aqueous to the lipid bilayer, causing cis-trans isomerization of unsaturated fatty acid residues through their reversible addition to double bonds (Scheme 1a) [10–15]. The desulfurization processes correspond to a mutation of the natural sequence of proteins. For example, Cys residues can be converted into dehydroalanine residue, or if a disulfide bods is attacked, one cysteine (Cys) involved in the S-S bridge is mutated into alanine (Ala), whereas Met is mutated into α-aminobutyric acid (Aba) (Scheme 1b) [16]. This amino acid is not genetically coded, therefore the conversion of Met into Aba is analogous to a post-translational modification into another natural structure, whose biological consequences are not known. Biomolecules 2019, 9, 794 3 of 17 Biomolecules 2019, 9, 794 3 of 18 Scheme 1. (a) The reaction mechanism for the cis-trans isomerization catalyzed by S-centered radicals. Schemeof It consists 1. a(a) The reaction reversible mechanism addition for theradicals of S-centered cis-transto isomerization double bondscatalyzed by to giving rise S-centered radicals. trans geometry, It consists of a reversible addition of S-centered radicals to double bonds giving rise the most thermodynamically favorable disposition of double bond. (b) The diffusible sulfur-centered to trans geometry, the most radicals, thermodynamically able favorable to migrate to the lipid disposition bilayer and induceof double bond. cis-trans (b) The diffusible isomerization sulfur-centered of unsaturated fatty radicals, able to migrate to the lipid bilayer and induce cis-trans isomerization of acid residues, can be produced by the desulfurization of cysteine (Cys) residues into dehydroalanine, unsaturated fatty acid residues, cystine into Cys and can be produced alanine (Ala),by themethionine and desulfurization (Met)ofinto cysteine (Cys) residues α-aminobutyric acidinto dehydroalanine, (Aba). cystine into Cys and alanine (Ala), and methionine (Met) into α-aminobutyric acid (Aba). The most studied intermediates known to cause protein damage are reactive oxygen species (ROS) The most OH and in particular studied intermediates • radicals [17,18], whereasknownthe toreductive cause protein stress damage has been are lessreactive considered.oxygen species However, • not(ROS) only and in particular oxidative speciesOH radicals but also [17,18], reductive whereas ones, such the reductive atoms as hydrogen stress hasandbeen less considered. solvated electrons, However, have been foundnot only oxidative to cause specieseffects deleterious but also reductive in cells [19–21]. ones, Onsuch as hydrogen this basis, atomsthat it is evident andspecific solvated electrons, have been found to cause deleterious effects in cells [19–21]. research has to be conducted in order to connect the chemical reactivity to the biological environment, On this basis, it is evident that specific the including research scenariohasoftoreductive be conducted radicalinstress, orderless to connect considered thethanchemical reactivity the oxidative to the biological one. environment, including the scenario of reductive radical stress, less In order to elucidate some possible scenarios for the production of diffusible thiyl radicals, considered than the oxidative our one. research strategy combined different approaches. Biomimetic systems based on radiation chemistry were used In order to elucidate to mimic some possible protein exposure scenarios to different kindsfor the radical of free production stress. ofIn diffusible addition,thiyl model radicals, systems our research strategy combined different approaches. Biomimetic systems constituted by protein and unsaturated lipid vesicles were used to get information on how the initial based on radiation chemistry were damage radical used toon mimic proteinprotein can beexposure eventually to propagated different kinds to theoflipid freecompartment. radical stress.Finally, In addition, model the protein systems changes structural constituted by protein caused by theand freeunsaturated radical attack lipid werevesicles analyzedwereby used to get using information Raman on how spectroscopy, thecan that initial shedradical light on damage on protein the preferential can site of be theeventually radical attack propagated and on the toradical-induced the lipid compartment. Finally, modifications in the protein protein structural changes caused by the free radical attack were analyzed by using Raman folding. spectroscopy, Two kinds of that can shed light proteins, sulfur-containing on the preferential serum albumins site ofand themetallothioneins, radical attack and wereon chosen the radical- to induced modifications in protein folding. better connect the chemical reactivity with membrane lipid transformation. Human serum albumin (HSA) Two is onekinds of theofmost sulfur-containing important plasma proteins, serum albumins antioxidants and metallothioneins, in protecting were chosen key regulatory proteins, has to better connect the chemical reactivity with membrane lipid transformation. many functions (i.e., ligand-binding and antioxidant) and it is rich in sulfur-containing residues Human serum albumin (HSA) (Table 1). is It one of the contains Metmost important residues, plasma disulfide antioxidants bonds (17 S-S) and in one protecting free thiol key regulatory group (CySH).proteins, Its bovinehas many functions (i.e., ligand-binding and antioxidant) and it is rich in counterpart (BSA) slightly differs from HSA in the sequence (76% is identical), as reported in Table 1 [22]. sulfur-containing residues As(Table regards 1). metallothioneins It contains Met residues, (MTs), theydisulfide bonds are low (17 S-S) and molecular one free weight, thiol group sulfur-rich (CySH). proteins (thioIts bovine means counterpart sulfur), (BSA) slightly characterized by a highdiffers fromof content HSACysin(almost the sequence a third(76%of the is identical), total amino asacids) reported andina Table high 1 [22]. As regards metallothioneins (MTs), they are low molecular weight, sulfur-rich proteins (thio means sulfur), characterized by a high content of Cys (almost a third of the total amino acids) and a Biomolecules 2019, 9, 794 4 of 18 capacity for binding metals by formation of metal-thiolate (metal-Cys) clusters. They are present in all living organisms and have multifunctional roles [23,24], among them scavengers to various radicals and ROS [25,26]. Table 1. Total amino acid content, number of Met, Cys, and aromatic residues in the four samples studied, ZnII-QsMT2, ZnII-GmMT2, BSA, and HSA. Samples Total AA Cys Met Arom. AA ZnII -QsMT2 81 14 3 2 Phe, 1 Hys ZnII -GmMT2 77 14 4 2 Tyr, 1 Phe BSA 583 35 4 2 Trp, 20 Tyr, 27 Phe, 17 His HSA 585 35 6 1 Trp, 18 Tyr, 31 Phe, 14 His In this work we present a study of two homologous plant MTs (from cork oak and soybean) and two serum albumins under oxidative and reductive radical stress in order to compare their capability to scavenge reactive radicals, to transfer eventually a damage to lipid compartment, as well as to evaluate the consequence on the protein structure. In fact, the connection of protein reactivity with membrane transformation can give a contribution to the puzzling context of radical stress occurring to biomolecules. In native M-MTs complexes (M standing for metal) S-S bonds are almost completely absent because all the thiol groups of Cys residues are deprotonated and bound to metal ions. In addition, M-MTs contain Met residues, and sometimes also labile sulfide anions (S2− ) as non-proteic ligands [27] that could generate diffusible sulfhydryl radicals and promote isomerization of double bonds in membrane lipids. The presence of sulfur-containing amino acids, methionine (Met) and cysteine (Cys), as well as sulfide anions (S2− ) in the same molecular aggregate makes this system a very interesting case to study in the contest of the tandem protein-lipid damages [28]. Two plant MTs from cork oak (Quercus suber —QsMT) and soybean (Glycine max—GmMT) (the subfamily MT2) were chosen because of their structural homology: both metal-protein complexes present two Cys-rich domains separated by one Cys-devoid domain; the length of the two MTs and the Cys number is about the same (Table 1). In addition, Cys residues are in a similar sequence position and there are similarities also between their metal-clusters [29,30]. 2. Materials and Methods 2.1. Protein Models and Their Analysis Fatty acid-free human and bovine serum albumin (HSA and BSA) were purchased from Sigma-Aldrich (trought Merck Life Science S.r.l. 20149, Milan, Italy) and used without further purification. HSA or BSA (1 mg) were dissolved in 1mL of tri-distilled water or aqueous 0.2 M t-BuOH solution. The two homologous plant MTs from cork oak (Quercus suber—QsMT) and soybean (Glycine max —GmMT) (the subfamily MT2) were recombinantly synthesised in E.coli, following a rationale based on the heterologous synthesis of the protein in E.coli cells harbouring the cDNA encoding for the desired MT, in its wild-type form, in form of split domains or site-directed mutants. The metal bound to the aggregates depends on which metal is supplemented to the media where the bacteria are growing, so the synthesis of the metallic aggregates takes place inside the cell. The use of this genetic engineering approach is an optimal way to obtain significant amounts of highly pure MTs that does not require any further manipulation, being representative of in vivo folding processes since they are formed into a cell environment. Biomolecules 2019, 9, 794 5 of 18 By sonication, centrifugation, and chromatographic procedures, the M-MT aggregates are obtained in purity and amounts enough for spectroscopic structural studies. Full details of the synthesis and purification procedures have been extensively described in [31,32] The S and Zn content of all Zn-MTs preparations was analyzed by inductively-coupled plasma atomic emission spectroscopy (ICP-AES) in a Polyscan 61E (Thermo Jarrell Ash, Franklin, MA, USA) spectrometer, measuring S at 182.040 nm and Zn at 213.856 nm. Molecular mass determinations were performed by electrospray ionization time-of-flight mass spectrometry on a Micro Tof-Q instrument (Bruker, Bremen, Germany) interfaced with a Series 1100 HPLC Agilent pump (Agilent Technologies, Waldbronn, Germany). The content of acid-labile sulfide ligands (S2− ), referred to the MT concentration measured by acid ICP-AES, was quantified by gas chromatography—flame photometric detection (HP 5890 Series II gas chromatograph coupled to a FPD80 CE Instruments –Thermo Finnigan- detector, Thermo Electron Corporation) after strong acidification (generation of H2 S). Finally, the molecular mass of the M-MTs species is obtained by electrospray ionization mass spectrometry (MicroToF-Q Instrument – Bruker, Bremen, Germany) [30,31]. The recombinant Zn-MTs used in this work contain the same amount of metal (metal-to-protein ratio is 3.5), whereas the sulfide ions are present only in the isoform from Quercus suber (sulfide-to-protein ratio is 1). To study the structural properties of polypeptides and proteins many different chemico-physical techniques can be utilized, i.e., nuclear magnetic resonance and X-ray diffraction. The disadvantage of these two techniques is that they require large amounts of sample and are slower than the optical methods. To obtain information on the radical-induced products gas chromatography–mass spectrometry (GC/MS) and HPLC techniques have been generally used, but both methods require isolation and hydrolysis of the samples prior to product quantification. In this context Raman spectroscopy can give a useful contribution since it is capable of providing a rapid and non-destructive analysis of the samples, and it can detect several structural features at the same time. Raman spectroscopy involves the analysis of scattered photons from a laser beam focused into the sample. A small percentage of the scattered photons exchanges energy with the vibrational energy levels (or, crudely, the ‘vibrations’) of the molecules. These vibrations are a function of molecular conformation, of the distribution of electrons in the chemical bonds, and of the molecular environment. Thus, a Raman spectrum provide information on local sites in a much larger macromolecule, and can be of help in identifing the favourite sites of the radical attack and radical-induced modification in protein folding. One of the greatest advantages of Raman spectroscopy is the lack of sample preparation needed prior to analysis. Because Raman scatter originates from the surface of a sample, there is no concern with sample thickness or size, it requires a low amount of sample. To obtain samples suitable for the Raman analysis avoiding spectral interference by the Tris buffer used during the M-MTs synthesis, a dialysis-lyophilization protocol was implemented prior to the spectroscopic measurements [30]. Raman spectra were recorded before and after irradiation of the aqueous solutions by a Bruker IFS 66 spectrometer equipped with a FRA-106 Raman module and a cooled Ge-diode detector, in order to follow the changes in the protein structure and the microenvironment of amino acidic residues, resulting from the exposure to radical species. The excitation source was a Nd3+ -YAG laser (1064 nm), the spectral resolution was 4 cm−1 and the total number of scans for each spectrum was 6000. The laser power on the sample was about 100 mW. No carbonization effect was observed on focused part of the samples, as already tested in the previous studies [13–15,29,30]. Raman spectra were obtained on lyophilized samples, before and after irradiation of the aqueous solutions in order to improve the signal/noise ratio that was low in aqueous solutions due to the interference from protein fluorescence and to avoid any eventual post-irradiation effect in the protein samples, since in radiolysis of water, also nonradical oxidants (H2 O2 ), together with HO•, eaq- and H• short-lived species, are produced. Lyophilization was performed on a Modulo 4K Freeze Biomolecules 2019, 9, 794 6 of 18 Dryer equipped with a RV8 Rotary Vane Pump (Edwards, Fisher Scientific, Pittsburgh, PA, USA). The lyophilized products were kept at −80 ◦ C until use. The curve fitting analysis for the secondary structure determination was accomplished by using the OPUS/IR v5.0 program (Bruker Italia, Milan, Italy), using the Levenberg–Marquardt algorithm. Peak component number and position were identified by fourth-derivative spectra, smoothed with a 13-point Savitsky–Golay function. These band positions were used as the initial guess for the curve fitting of the original spectra. Best curve fitting was obtained at the lowest possible χ2 values. The Raman component profiles used in the curve fitting was a linear combination of Lorentzian and Gaussian functions. The content of secondary structure elements was calculated from the integrated intensities of the individually assigned bands and their fraction of the total intensity. All the components were assigned to a particular secondary structure on the basis of the literature [33–37]. 2.2. Radiolytic Production of Transients Reactive species generation, mimicking the conditions of an endogenous radical stress, was obtained by using γ-radiolysis of aqueous solutions (60 Co-Gammacell), which allows selecting the reacting species by changing the appropriate experimental conditions. Exposure to irradiation conditions, such as radiolysis, has been largely used for the study of oxidative stress, thus providing a very useful tool for investigating mechanistic and structural features of the damages, i.e., to amino acids [17,18]. When diluted aqueous solutions are irradiated, practically all the energy absorbed is deposited in water molecules giving rise to water radiolysis. Thus, the chemical changes in solutes are brought about indirectly by the three short-lived species produced, hydrated electron (eaq − ), hydroxyl radicals (• OH), and hydrogen atoms (H• ), together with H+ and H2 O2 , (Equation (1)) [38]. The first three reactive species are expected to react very fast with peptides and proteins. For example, the rate constants for eaq − and • OH species with serum albumin are 8.2 × 109 and 7.8 × 1010 M−1 s−1 , respectively [11,38]. Direct action due to energy deposited directly in the solute is not important when the solute concentration is below 100 mol/m3 [39]. γ • • + H2 O ∼∼∼ → e− aq (0.27), HO (0.28), H (0.06), H (0.27), H2 O2 (0.07) (1) The values in parentheses represent the radiation chemical yields expressed in terms of G-value (µmol·J−1 ), corresponding to the yield of radicals per 10 eV of absorbed energy. By saturating with N2 O, eaq − are efficiently converted into the oxidizing HO• radical (Equation (2)) and mainly oxidative stress conditions are obtained. • − e− aq + N2 O + H2 O → HO + N2 + HO (2) Under these experimental conditions, used in model studies of oxidative damages occurring in vivo [40,41], HO• and H• radicals account for 90% and 10%, respectively, of the reactive species. It is possible to investigate the reactivity of only reductive species (eaq − and H• ) and their consequences on protein structure by adding an alcohol, such as t-BuOH (0.2 M), to the protein solution, and irradiating in an oxygen-free atmosphere. Under these experimental conditions, HO• radicals are efficiently scavenged, and eaq − and H• account for about 80% and 20% of the reactive species, respectively. Finally, by using N2 O-saturated solutions in combination with t-BuOH as additive, the reactivity of only H• atoms can be investigatedsince the presence of t-BuOH limits the eventual observations of undesired polypeptide modifications resulting from HO• radicals. 2.3. Biomimetic Models Biomimetic models can be very useful to address lipid-protein damage, in order to test the capability of proteins under free radical stress to damage the lipid cell compartment and to extrapolate the mechanistic information to the more complex situations present in vivo. Monolamellar vesicles, Biomolecules 2019, 9, 794 7 of 18 formed by the extrusion methodology and made of phospholipid containing an unsaturated fatty acid moiety in the cis configuration (such as oleic acid), are simple but appropriate models of natural membranes. They are a good biomimetic model for the double bond isomerization process if used as aqueous suspensions in the presence of peptide/proteins and the reactive radical species since they are mimicking two biological compartments, the membrane and the extracellular protein domain. In detail, we used proteins at micromolar concentration levels added to a millimolar vesicle suspension; the vesicles being formed by 1-palmitoyl 2-oleoyl phosphatidylcholine (POPC), where one of the two fatty acid chains presents the double bond in the cis configuration. Aliquots of the suspensions are saturated with an appropriate gas (N2 O or Ar) prior to γ-irradiation at different doses, in order to selectively form the reactive radical species. After trans-esterification of the phospholipids to the corresponding fatty acid methyl esters [42], the trans isomer content, calculated as oleate/elaidate (cis/trans) ratio, is obtained by gas chromatographic (GC) analysis [43]. Based on the palmitic moiety of POPC, which served as the internal standard, trans isomers in our experiments were formed quantitatively and no side-reactions occurred. GC is the easiest and fastest technique for fatty acid analysis. GC analysis requires the fatty acid-containing lipids to be converted to more volatile compounds, such as methyl ester or other ester derivatives (FAMEs). For separation of all of FAMEs components in a mixture, the choice of the chromatographic column and the carrier gas play a crucial role. In this study, the GC analysis were performed by the Agilent 6850 (Milan, Italy) GC system equipped with a 60 m × 0.25 mm × 0.25 µm (50%-cyanopropyl)-methylpolysiloxane column (AgilentJ&W DB-23, Cernusco sul Naviglio MI, Italy) and a flame ionization detector. A constant pressure mode (29 psi) was chosen using helium as the carrier gas [44,45]. As regards the experiments with biomimetic models, 30 µM ZnII -MTs or 9.2 µM HSA/BSA aqueous solutions were added to 2 mM liposome suspensions of POPC containing large unilamellar vesicles (diameter = 100 nm) and γ-irradiated. 100 µL aliquots of the protein–lipid suspension were withdrawn at different irradiation times for lipid isolation, following the procedure described elsewhere [46]. When necessary, t-BuOH was also added to reach a final concentration of 0.2 M, a value known to be compatible with vesicle stability. Control experiments in the absence of proteins confirmed that trans-isomer formation is less than 0.2% after exposure to 500 Gy. All set of experiments were replicated twice (M-MTs) or three times (HSA/BSA), depending on the availability of sample amount for the anaysis. The error in the determination of the trans isomer percentage ranged from 2% to 6% (the lowest errors in the case of serum albumins). It is worth underlining that even very low levels of protein modifications produced by reactive species can be detected by the liposome vesicle reactivity. This is due to an amplification effect given by the catalytic cycle of the thiyl radical-based cis-trans isomerization [47]. 3. Results and Discussion 3.1. Damage Transfer from Protein to Lipid In order to obtain information on protein modifications caused by the HO• and H• attack and verify the eventual release of sulfur-centered radicals from the MTs isoforms and serum albumins, lipid-protein systems were used. All the four examined proteins under the attack of mainly HO• radicals species were able to cause the trans-isomerization of the naturally occurring fatty acids in a dose-dependent manner (Figure 1a,b) [48–51]. These results confirm once again that under radical stress thiyl radicals can be readily generated from the cleavage of disulfide linkages (the latter are abundant in serum albumins and almost absent in MTs), the attack to Met residues or metal-thiolate clusters, where the sulfur atom of one Cys can bind to one or two metals. Biomolecules 2019, 9, 794 8 of 17 Biomolecules 2019, 9, 794 8 of 18 (a) (b) (c) (d) Figure 1. Percentage of trans isomer formation in POPC vesicles in the presence of proteins and reactive Figure 1. Percentage of trans isomer formation in POPC vesicles in the presence of proteins and radical species at different dose radiation (Gy) under mainly oxidative (a,b) and reductive conditions reactive radical species atIIdifferent dose radiation (Gy) under mainly oxidative (a, b) and reductive (eaq − and H• ) (c,d); (a,c)•Zn -QsMT and ZnII -GmMT2, (b,d) BSA and HSA. conditions (eaq and H ) (c, d); (a) and (c) Zn -QsMT and ZnII-GmMT2, (b) and (d) BSA and HSA. - II The comparison between proteins with a high percentage of homology shows the importance of The comparison between proteins with a high percentage of homology shows the importance of the sequence for the reactivity and the potential to generate small radicals involved in tandem damage. the sequence IIfor the reactivity and the potential to generate small radicals involved in tandem In fact, both Zn -GmMT2 and HSA, containing a higher number of sulfur-containing residues than damage. In fact, both ZnII-GmMT2 and HSA, containing a higher number of sulfur-containing ZnII -QsMT2 and BSA, respectively, have a higher capacity to induce cis-trans isomerization than the residues than ZnII-QsMT2 and BSA, respectively, have a higher capacity to induce cis-trans other two proteins. This is particularly evident at the highest dose when serum albumins and the isomerization than the other two proteins. This is particularly evident at the highest dose when metal complexes have almost lost their native folding, making the sulfur-containing moieties more serum albumins and the metal complexes have almost lost their native folding, making the accessible to the free radical attack and the lipid damage strictly connected to the amino acid content in sulfur-containing moieties more accessible to the free radical attack and the lipid damage the sequence. In ZnII -MTs systems the sulfur-centered radicals can be generated by the attack of HO• strictly connected to the amino acid content in the sequence. In ZnII-MTs systems the sulfur- to Cys bound to the metal ions and H• to Met residues, respectively (Scheme 2). centered radicals can be generated by the attack of HO• to Cys bound to the metal ions and H• to Met residues, respectively (Scheme 2). Biomolecules 2019, 9, 794 9 of 17 Biomolecules 2019, 9, 794 9 of 18 Biomolecules 2019, 9, 794 9 of 17 Cys-Metal clusters Cys-Metal clusters Scheme 2. Diffusible sulfur-centered radicals, able to migrate to the lipid bilayer and induce cis-trans Scheme2.2.Diffusible Scheme isomerization sulfur-centered of unsaturated Diffusible fatty acidradicals, sulfur-centered radicals,able residues, canto able tomigrate be migrateto producedtothe by lipid lipidbilayer thesulfur-metal and andinduce bilayerclusters cis-trans affected induce by cis-trans • attack and isomerization theisomerization of unsaturated fatty acid residues, can be produced by sulfur-metal clusters affectedby HO of unsaturated by the fatty acid desulfurizationresidues, of can methioninebe produced residues. by sulfur-metal clusters affected by the • theHO HO• attack attack and andbybythe thedesulfurization desulfurizationofofmethionine methionineresidues. residues. As regards BSA and HSA, thiyl radicals under anaerobic conditions are mainly produced by the attackAs of regards As HO • andBSA regards BSA and H• to HSA, thiyl disulfide and radicals bonds, HSA, thiyl underinanaerobic as shown radicals Scheme 3,conditions under anaerobic are be but they can conditions mainly produced by also generated are mainly the producedfrom by the attack of HO • and H• to disulfide bonds, as shown in Scheme 3, but they can be also generated from • the attack of HO• and fragmentation • ofHMetto residues disulfidecaused bonds,byas the H attack shown (Scheme in Scheme 2).they can be also generated from 3, but the fragmentation of Met residues caused by the H••attack (Scheme 2). the fragmentation of Met residues caused by the H attack (Scheme 2). Scheme 3. Mechanism for the reactions of HO• , eaq − , and H• with cystines. Scheme 3. Mechanism for the reactions of HO•, eaq-, and H• with cystines. All the four protein models showed to reactions have a high •, eaq-, and capacity • with cystines. to induce cis-trans isomerization All the four protein Schememodels 3. Mechanismshowed for to the have a high of HOcapacity to Hinduce cis-trans isomerization under reductive radical stress conditions (Figure 1c,d). For example, the formation of trans isomer under All reductive the fourradical proteinstress conditions models showed (Figure to have1c,d). For capacity a high example,tothe formation induce of trans cis-trans isomer isomerization caused by the attack of mainly HO• • radicals towards ZnII II -GmMT2 is about 20%, whereas it increases caused underby the attack reductive of mainly radical stressHO radicals (Figure conditions towards1c,d). Zn -GmMT2 For example, is about the20%, whereas formation of ittrans increases isomer up to about 50% when eaq- − and H • • are the reactive species. This result suggests that reducing species • the reactive species.IIThis result suggests that reducing species upcaused to about 50% when e and H are by the attack of mainly HO radicals towards Zn -GmMT2 is about 20%, whereas it increases aq are very specific damaging -agents for • sulfur-containing residues, therefore contributing significantly to are upvery specific to about 50% damaging when eaqagents and Hfor are sulfur-containing the reactive species. residues, This therefore contributing result suggests significantly that reducing species the transfer of the protein damage towards the lipid domain. toare thevery transfer of the specific protein damage damaging agents for towards the lipid domain. sulfur-containing residues,•therefore contributing significantly This can be explained by considering that the attack of HO• radicals can occur at the level of to theThistransfer can beofexplained the protein by damage considering towards that the the lipid attack of HO radicals can occur at the level of domain. sulfur-containing residues, aromatic amino acids, aliphatic side chains, and, consequently, multiple sulfur-containing This can beresidues, explained aromatic amino acids, by considering aliphatic that the attackside of HOchains, and, consequently, • radicals can occur at the multiple level of products can be generated. For example, the attack towards sulfur-containing residues can produce products can be generated. sulfur-containing residues,For example, aromatic aminothe attack acids, towards aliphaticsulfur-containing side chains, and, residues consequently,can produce multiple oxidation derivatives such as disulfide bridges and sulphoxides, whereas the withdraw of an H atom oxidation productsderivatives such as disulfide can be generated. For example, bridges the and sulphoxides, attack whereas the withdraw towards sulfur-containing residues of can an Hproduce atom from the α-carbon can give rise to the release of thiyl radicals capable to diffuse into lipid bilayers and from the α-carbon oxidation can give derivatives such rise to the release as disulfide bridges ofand thiyl radicals capable sulphoxides, whereas to diffuse into lipid the withdraw of an bilayers H atom induce secondary damage. and from induce secondarycan the α-carbon damage. give rise to the release of thiyl radicals capable to diffuse into lipid bilayers Under reductive stress conditions the higher isomerization yields obtained from ZnIIII -GmMT2 andUnder inducereductive secondary stress conditions damage. the higher isomerization yields obtained from Zn -GmMT2 II -QsMT2 and BSA can be attributed to several factors, such as and HSA in comparison with Zn and HSA Underin comparison reductive stresswith conditions ZnII-QsMT2the and BSA can higher be attributed isomerization to several yields obtained factors, from such as the ZnII-GmMT2 the higher number of Met residues in IIZnII -GmMT2 and HSA and a slight different folding of the higher number of Met residues in II Zn and HSA in comparison with Zn -QsMT2 and BSA can be attributed to several factors, such asthe -GmMT2 and HSA and a slight different folding of the corresponding protein systems which can play a role in the accessibility to free radicals to reach corresponding higher number protein of Met systems residues which in Zn can play a role II-GmMT2 andinHSA the accessibility and a slighttodifferent free radicals foldingto reach of the sulfur-containing residues. In addition, the presence of two Tyr residues in ZnIIII -GmMT2, absent in sulfur-containing corresponding residues. protein In systemsaddition, which thecan presence play a of role two in Tyr the residues accessibility in Zn to -GmMT2, free radicals absent to in reach ZnII II -QsMT2, can contribute to the higher capability of ZnII II -GmMT2 to transfer the damage to lipid Znsulfur-containing -QsMT2, can contribute residues.toInthe higher the addition, capability presence of of Zntwo -GmMT2 to transfer Tyr residues in Zn the damage to II-GmMT2, lipidin absent domains. In fact, although the benzene ring of both Tyr and Phe is known to react with H•• atoms, domains. Zn -QsMT2, II In fact, canalthough contribute thetobenzene the higher ringcapability of both Tyr of Znand-GmMT2 II Phe is known to react to transfer thewith damage H atoms, to lipid leading to the formation of H-ring adducts, only Tyr has been reported to be involved in side-chain leading domains. to the In formation fact, althoughof H-ring adducts, the benzene ringonly of Tyr bothhasTyrbeenandreported Phe is known to be involved to react with H• atoms, in side-chain reactions where Tyr radicals can be repaired by Cys, yielding thiyl radicals [17]. reactions leading where Tyr radicals to the formation of can be repaired H-ring adducts,by Cys, only Tyryielding has been thiyl radicalsto[17]. reported be involved in side-chain reactions where Tyr radicals can be repaired by Cys, yielding thiyl radicals [17]. Biomolecules 2019, 9, 794 10 of 18 Biomolecules 2019, 9, 794 10 of 17 In order In order to to demonstrate demonstrate whichwhich reducing reducing species species is more active is more active regarding regarding isomerization isomerization yield, yield, • experiments only experiments only inin the the presence presence of of H H radicals • radicals have have been been performed performed with with BSA BSA andand HSA. HSA. It It clearly clearly appeared that appeared thatthis thisspecies speciesisisable abletotopromote promote the the isomerization isomerization in in an an extremely extremely efficient efficient wayway (i.e.,(i.e., at at the the highest dose the percentage of trans isomer is over 75% for both proteins) (Figure highest dose the percentage of trans isomer is over 75% for both proteins) (Figure 2). Thus, as regards 2). Thus, as regardsalbumins, serum serum albumins, there there is not is not acontribution a synergic synergic contribution of hydrogen- of hydrogen- and electron-adduct and electron-adduct on the protein on the protein system to the formation of diffusible species responsible for the cis-trans isomerization. system to the formation of diffusible species responsible for the cis-trans isomerization. On the contrary, On the this contrary, synergy was this found synergy for thewas found M-MTs for the[48]. systems M-MTs systems [48]. Figure Figure 2. Percentage of 2. Percentage trans-isomer (elaidate) of trans-isomer (elaidate) formation formation inin POPC POPC vesicles vesicles inin the the presence presence of of BSA BSA (black) t-BuOH • (black) and HSA (red) in aqueous Ar-flushed solutions in the presence of 0.2 M t-BuOH (only the and HSA (red) in aqueous Ar-flushed solutions in the presence of 0.2 M (only the H H• species species is is present). Moreover, the present). Moreover, the results results obtained obtained with with BSA BSA and and HSA HSA show show aa strong strong dependence dependence on on the the content contentofofMet Metresidues residues in the protein in the sequence, protein sincesince sequence, the lowest value value the lowest of isomerization was always of isomerization was obtained for BSA, for always obtained which BSA,contains whichless Met than contains lessHSA. Met This thansuggests HSA. Thisthatsuggests in the tandem albumin-lipid that in the tandem damage the attack to Met residues by reducing species plays a more relevant role albumin-lipid damage the attack to Met residues by reducing species plays a more relevant than that to role disulfide than bridges (the same amount in both proteins). that to disulfide bridges (the same amount in both proteins). 3.2. Protein Analysis by Raman Spectroscopy 3.2. Protein Analysis by Raman Spectroscopy The changes in the protein structure and the microenvironment of amino acidic residues, The changes in the protein structure and the microenvironment of amino acidic residues, resulting from the exposure to radicals, can be followed by Raman spectroscopy. Indeed, this resulting from the exposure to radicals, can be followed by Raman spectroscopy. Indeed, this vibrational technique has become a versatile tool in protein science and its application in characterizing vibrational technique has become a versatile tool in protein science and its application in protein in biotechnology, pharmaceutical production, and food industry has been developed [52]. characterizing protein in biotechnology, pharmaceutical production, and food industry has been In particular, it can be used to quantify the contribution of distinct secondary structure motif to the developed [52]. In particular, it can be used to quantify the contribution of distinct secondary overall protein structure. The best-studied Raman band of proteins is the amide I band appearing structure motif to the overall protein structure. The best-studied Raman band of proteins is the amide between 1630 cm−1 and 1700 cm−1 , which arises from the stretching vibration of the peptide C=O group. I band appearing between 1630 cm–1 and 1700 cm–1, which arises from the stretching vibration of the peptide 3.2.1. C=O group. Protein Secondary Structure Change 3.2.1.In order Secondary Protein to evaluateStructure the damage performed by HO• and H• radicals on the protein structure, Change Raman spectra were recorded after exposure at different doses under N2 O-saturated conditions • • (i.e., In Znorder to evaluate II -GmMT2 the damage at 100 and 300 Gy).performed The shapebyofHOAmideandI H bandradicals in theonZnthe protein spectrum II -GmMT2 structure, Raman spectra significantly were recorded changed with theafter exposureofatshoulders appearance different doses under at higher 2O-saturated conditions (i.e., andNlower wavenumbers respect Zn II-GmMT2 at 100 and 300 Gy). The shape of Amide I band in the ZnII-GmMT2 spectrum to the band maximum (Figure 3), indicating differences in the secondary structure contents after the significantly radical stresschanged exposure. with In the appearance particular, of shoulders the new Amide I at higher andgiving component, lower rise wavenumbers respect at to weak shoulder to the band about 1686maximum (Figure 3), cm−1 , is indicative of indicating an increasedifferences in the content in the beta-turns secondary [34].structure contents after the radical stress exposure. In particular, the new Amide I component, giving rise to weak shoulder at about 1686 cm−1, is indicative of an increase in the beta-turns content [34]. Biomolecules 2019, 9, 794 11 of 17 Biomolecules 2019, 9, 794 11 of 18 Biomolecules 2019, 9, 794 11 of 17 Figure 3. Raman spectra of ZnII-GmMT2 system before (black) and after γ-irradiation in N2O- saturated aqueous solution (mainly oxidative conditions) at 100 (red) and 300 Gy (blue) dose Figure 3. Raman spectra of ZnII -GmMT2 system before (black) and after γ-irradiation in N2 O-saturated exposure. Figure 3. aqueous Raman(mainly solution spectraoxidative of ZnII-GmMT2 system conditions) at 100before (black) (red) and and(blue) 300 Gy afterdose γ-irradiation exposure. in N2O- saturated aqueous solution (mainly oxidative conditions) at 100 (red) and 300 Gy (blue) dose Analogously, weak modifications in the Amide I spectral profile were observed in HSA Analogously, exposure. weak modifications in the Amide I spectral profile were observed in HSA undergone undergone to 300 Gy irradiation dose (Figure 4); under reductive radical stress conditions, the to 300 Gy irradiation dose (Figure 4); under−1reductive radical stress conditions, the appearance of a appearance of a shoulder Analogously, weak at about 1680 cm as well as the increaseprofile of the Amide III component at shoulder at about 1680 cm−1modifications as well as the in the increase Amide I spectral of the Amide III componentwere observed at about in−1HSA 1280 cm are about 1280 cm undergone −1 are Gy indicative of an dose increase in the4);beta-turns content (Figure 4). indicative of to an 300 increase irradiation in the beta-turns (Figure content (Figureunder 4). reductive radical stress conditions, the appearance of a shoulder at about 1680 cm as well as the increase of the Amide III component at −1 about 1280 cm−1 are indicative of an increase in the beta-turns content (Figure 4). Figure 4. Raman spectra of HSA system before (black) and after γ-irradiation in N2 O-saturated aqueous Figure 4.(mainly solution Ramanoxidative spectra of HSA system before conditions—blue) (black) and in and after Ar-flushed γ-irradiation solutions in N containing 2O-saturated 0.2M t-BuOH aqueousreductive (mainly solution (mainly oxidative conditions—blue) conditions—purple) and in Ar-flushed solutions containing 0.2M t- at 300 Gy dose exposure. BuOH (mainly reductive conditions—purple) at 300 Gy dose exposure. Figure 4. Raman spectra of HSA system before (black) and after γ-irradiation in N2O-saturated To obtain a semi-quantitative evaluation of the conformational changes upon irradiation, a curve aqueous solution (mainly oxidative conditions—blue) and in Ar-flushed solutions containing 0.2M t- fittingTo analysis obtain can be performed. The a semi-quantitative neat ZnIIof evaluation -GmMT2 system was found the conformational changes to contain mainly β-strands upon irradiation, a curve BuOH (mainly reductive conditions—purple) at 300 Gy dose exposure. II and fittingβ-turns [30],can analysis as be well as a significant performed. contribution The neat ZnII-GmMT2fromsystem was was β-strands foundfound also in mainly to contain Zn - andβ- Cd II -QsMT2 complexes [53]. Both mainly oxidative and reductive radical stress conditions cause IIa- strands and β-turns [30], as well as a significant contribution from β-strands was found also To obtain a semi-quantitative evaluation of the conformational changes upon irradiation, a curve in Zn loss in β-strands fitting analysis canandbea relevant gainThe performed. of β-turns neat Znin these protein II-GmMT2 systems system (Tableto was found 2).contain A moremainly extreme β- strands and β-turns [30], as well as a significant contribution from β-strands was found also in ZnII- Biomolecules 2019, 9, 794 12 of 18 unfolding, with a notable increase also of random coil structure, takes place in ZnII -GmMT2 under strictly reductive radical stress conditions (eaq − and H• —Table 2), similarly to the behavior of the ZnII -QsMT system, where the exposure to reductive species induces an increase in the percentages of β-turn conformation, according to the theoretical prediction reported for the normal modes of β-turns [53,54]. In the case of QsMT, the increase of random coil conformation in reducing conditions was observed for CdII -QsMT [53]. Table 2. Secondary structure percentages of ZnII -GmMT2 before and after γ-irradiation at 100 and 300 Gy under mainly oxidative conditions (HO• and H• ) and after γ-irradiation under reductive stress conditions (eaq − and H• ). The values were obtained by the curve fitting analysis of the Amide I band. Reactive Species Dose (Gy) Beta-Strands (%) Beta-Turns (%) Random (%) - 0 45 43 12 HO• and H• 100 25 55 19 HO• and H• 300 20 53 28 eaq − and H• 100 10 51 39 As regards serum albumins, the curve fitting analysis of Raman spectra showed that both mainly oxidative and reductive radical stress conditions cause a loss in α-helix and an increase in the β-sheets content (Table 3). Table 3. Secondary structure percentages of HSA before and after the attack of different reactive species (γ-irradiation at 300 Gy), obtained by the curve fitting analysis of Amide the I band. Reactive Species Alpha-Helix (%) Beta-Sheets (%) Beta-Turns (%) Random (%) - 54 22 14 10 HO• and H• 43 32 12 13 eaq − and H• 46 23 22 9 H• 38 31 18 13 Also in this case a more significant unfolding of the helical structure takes place under the attack of hydrogen atom than of hydroxyl radicals. This is in agreement with the literature where H• atoms have been found to have a high ability in denaturing protein structure [55,56]. The conformation changes induced by the reductive stress conditions can be due both to the addition of the solvated electron to the backbone carbonyl groups, that can eventually result in the cleavage of peptidic bonds [17], and to the selective attack of H• atoms at different sites. 3.2.2. Specific Amino Acid Residues Damage As regards the side chains, aromatic residues, in particular Tyr, have resulted to be among the most sensitive residues towards radical attack. This conclusion was drawn out from the analysis of the doublet at 850–830 cm−1 that depends on the state of the hydrogen bonding involving the OH group of Tyr and from other Tyr bands, such as that at 645 cm−1 [33]. For example, GmMT2 contains two Tyr and one Phe residues that give rise to many Raman bands due to aromatic ring vibrations (Figure 3). Exposure of ZnII -GmMT2 to radical stress causes significant changes in the I852 /I830 ratio, indicating that the environment around the Tyr residues has become more hydrophilic and/or Tyr are less bound to COO- of aspartic and glutamic acid residues. In addition, the intensity ratio between the ~645 and ~620 cm−1 bands, due respectively to Tyr and Phe residues, significantly decreased by increasing the free radicals concentration, showing a quite good linear correlation (R2 = 0.98) with the radiation dose used (Figure 5). This change could be due to a decrease in the Tyr number that, forming phenoxyl radicals by reacting with HO• radicals, give rise to bi-tyrosine cross-linked derivatives [57]. Biomolecules Biomolecules 2019, 9, x 13 of 18 2019,9,9,794 Biomolecules2019, 794 13 13 of of1817 5,0 4,5 4,0 3,5 I645 / I620 3,0 2,5 2,0 1,5 1,0 0,5 0,0 0 50 100 150 200 250 300 Irradiation Dose (Gy) Figure 5. I645/I620 ratio of ZnII-GmMT2 as a function of the γ-irradiation dose in N2O-saturated Figure 5. I645 /I620 ratio of ZnII -GmMT2 as a function of the γ-irradiation dose in N2 O-saturated aqueous aqueous solution (mainly oxidative conditions). Figure solution (mainly 5. Intensity oxidative ratio between the 645 and 620 cm -1 Raman band ( I645/I620) of ZnII-GmMT2 as conditions). Commented [m1]: Is it a “.” for the a function of the γ-irradiation dose in N2O-saturated aqueous solution (mainly oxidative conditions). A similar behavior was shown also by HSA when exposed to the eaq−- and H•• attack (the intensity A similar behavior similarwas shown also by alsoHSA whenwhen exposed exposed to to the eaq  and H(theattack (the intensity of the 645 cm −1A band behavior strongly was shown decreased, by HSA see Figure 4) [36]. the The eaq- and I645/IH620attack ratio hasintensity been connected to 620 /I of the 645 cmof −1 band strongly decreased, see Figure 4)4)[36]. TheI645I/I645 ratio has been connected to the 645 cm −1 band strongly decreased, see Figure [36]. The ratio 620 has been connected to the Tyr/Phe the content Tyr/Phe ratio in ratio content some in proteins some proteins[58], [58], soso this this change change couldcould be be indicative indicative of a of a radical-induced radical-induced the Tyr/Phe content ratio in some proteins [58], so this change could be indicative of a radical-induced modificationmodification of some Try of someresidues Try residuesintointo Phe Pheresidues. residues. HigherHigher Phe/Tyr Phe/Tyr ratio has ratio beenhasfoundbeen found in blood in blood modification plasmaof some Try residues into Phe residues. Higher Phe/Tyr ratio has been found in blood of patients with HIV infection, with cancer and trauma, and it has been associated with radical plasma of patients with HIV infection, with cancer and trauma, and it has been associated with radical plasma of patients with stress due HIV infection, to inflammation with cancer and immune activationand[59],trauma, although and the itmechanism has beenofassociated the disturbedwith radical stress due toPheinflammation metabolism has and not immune still completely activation understood. [59], although the mechanism of the disturbed stress due to inflammation and immune activation [59], although the mechanism of the disturbed Phe Phe metabolismBoth hasoxidative not stilland completely reductive radicalunderstood. stress causes significant changes in the spectral features of metabolism has not still completely the S-S stretching bands of HSA understood. (17 disulfide bridges), Both oxidative and reductive radical stress causeseven at the lowest significant irradiation changes indose the(Figure 4), features of spectral Both oxidative and bonds since disulfide reductive rapidlyradical react withstress causes HO radicals, Hsignificant atoms and solvated changes in the electrons, spectral leading under features of the S-S stretching anaerobicbands of HSA conditions (17 disulfide to thiyl bridges), RS (Scheme even at the lowest irradiation dose (Figure 4), the S-S stretching bands of HSA (17radicals disulfide • bridges), 2). The latterat even can regenerate the lowestdisulfide irradiationbridges,dose that (Figure 4), since disulfide can bonds rapidly have different react with conformation withHO radicals, respect • H• • atoms to the initial one (that and means solvated differentelectrons, Raman bands) leading under since disulfide bonds rapidly react with HO radicals, H atoms and solvated electrons, leading under or give rise to anaerobic conditions tolipid thiyl cis-trans radicals RS••(Scheme isomerization by producing 2). The •/S•– radicals [36]. These modifications HSlatter can regenerate disulfide bridges, that anaerobic conditions to thiylinradicals were also confirmed RS by the case of HSA, (Scheme 2). Theanalysis mass spectrometric latter whichcan regenerate disulfide bridges, evidenced the chemical can have different conformation with respect to theAlainitial one (that means different Raman bands) that can havetransformation different conformation of one Cys involved withinrespect S-S bond into to the initial as well as one a rapid(that means disulfide different scrambling Raman bands) events or give rise to lipid cis-trans isomerization by producing HS •/S•– radicals [36]. These modifications or give rise to lipid cis-trans isomerization by producing HS• /S•– radicals [36]. These modifications towards more stable disulfide/thiol population [50] werealso alsoconfirmed confirmedMoreover, inthe as regards the caseofofserumHSA, albumins, bymass mass thespectrometric spectral changes in the Phe bands analysis which canevidenced be connectedthe chemical were with the in relevant case HSA, modifications by occurring spectrometric in disulfide bridges analysis which under reductive evidenced radical stressthe chemical transformation transformation ofofone oneCys conditions. Cysinvolved involved It is known ininS-S that solvated S-Sbondbond into electroninto reacts Ala Alawith asas well well Phe asaarapid as generating rapid disulfidescrambling disulfide a cyclohexadienyl scrambling radical events events towards more towards morethat stable disulfide/thiol coulddisulfide/thiol stable be repaired by one-electronpopulation population oxidation[50] [50] to an acceptor molecule (Scheme 4). Disulfide Moreover, linkages as can be these regards serumacceptor moieties since albumins, the they are known spectral changesto act in as athemajor Phesink for electrons bands canbe beconnected connected Moreover, as regards arising from electron serumtransferalbumins, by reducingthe spectral species: changes in the Phe bands can with the relevant modifications occurring in disulfide bridges with the relevant modifications occurring in disulfide bridges under reductive radical stress conditions. under reductive radical stress Itconditions. is known that It issolvated known that electronsolvated reactselectron with Phe reacts with Phe generating generating a cyclohexadienyl a cyclohexadienyl radical that could radical be that could be repaired by one-electron oxidation to an acceptor repaired by one-electron oxidation to an acceptor molecule (Scheme 4). Disulfide linkages can be these molecule (Scheme 4). Disulfide linkagesmoieties acceptor can be since thesethey acceptor are known moieties to act since as a theymajorare sink known to act arising for electrons as a major fromsink for electrons electron transfer arising from by reducing species:electron transfer by reducing species: Scheme 4. Mechanism of the reaction between solvated electron and Phe residue, generating a Scheme 4. Mechanism of the reaction between solvated electron and Phe residue, generating a cyclohexadienyl radical. The latter can be successively involved in an electron transfer reaction with cyclohexadienyl radical. The latter can be successively involved in an electron transfer reaction with disulfide bonds, which can give rise to the reactions indicated in Scheme 3. disulfide bonds, which can give rise to the reactions indicated in Scheme 3. Thus, an initial addition of solvated electrons to Phe residues can result in the ultimate reduction Thus, of cystine an initial groups, addition which of solvated can then electrons lead to the to Phe residues fragmentation can resultbonds of the disulfide in the and ultimate reduction generation of of• cystine groups, which RS radicals (Scheme 3). can then lead to the fragmentation of the disulfide bonds and generation of RS• radicals (Scheme 3). Biomolecules 2019, 9, 794 14 of 18 The occurrence of an intra-protein electron transfer can contribute to obtaining a higher yield of isomerization for HSA, which contains more Phe residues than BSA (Figure 1d and Table 1). Also in M-MTs systems Cys is among the most sensitive residue towards radical attack, since the radical stress exposure causes a partial rearrangement and deconstruction of metal-thiolate clusters, as shown by the spectral modifications of some Zn-S stretching bands visible at low wavenumbers, typical of M-S stretching (280–350 cm−1 ) [53]. Demetallation reactions of the ZnII -MTs systems can also take place, as confirmed by various spectrometric techniques such as CD and ESI-MS [48,51]. As regards the Met residues, they are sensitive to the free radical attack in all conditions, as indicated for example by the decrease in intensity of the 726 cm−1 band due to the νC-S bond of Met residues in the ZnII -GmMT2 system (Figure 3). Thus, this investigation underlines the damages that oxidative and reductive radical attacks can cause to the secondary and tertiary structure of proteins and to specific amino acid residues. 4. Conclusions In the context of free radical processes with biological interest, tandem protein–lipid radical damage has been the subject of recent investigations that pointed out its harmfulness in the general scenario of establishing the consequences of radical stress. Lipid vesicle suspensions containing a protein have been found very useful in the initial identification of the coupled lipid–protein damage since the amplified membrane damage is the final effect of these reactive pathways. The results of γ-irradiation experiments reported in this paper can give a contribution to the knowledge of free radical-induced modifications in the structure of sulfur-containing proteins, resulting in consequent isomerization of membrane lipids. Radical damages are not randomly distributed along the polypeptide chain, but specific damages occur at sensitive amino acid sites. Met residues are one of the main sites of the attack and play an important role in the tandem albumin-lipid damage under reductive conditions by producing sulfur-centered radicals by means of a desulfurization reaction. In the case of serum albumins, this role becomes more relevant than that of disulfide bridges, which act as efficient trapping centers of reducing species, by direct trapping of electron transfer within the protein. Thus, an initial addition of solvated electrons to aromatic side chains of Tyr and Phe or to the backbone carbonyl groups of peptide bonds can result in the fragmentation of cystine groups, which can lead to HS• /S•– radicals able to promote lipid cis-trans isomerization in POPC suspensions or to reform disulfide bonds. In M-MTs systems, metal-clusters are good interceptors of free radicals. They react with free radical species, without a complete loss of metals, thus showing their capability to play a protective role against these stressors. As regards aromatic residues, aromatic side chains of Tyr and Phe have highlighted their sensitivity to radical stress, leading to electron transfer reactions or radical-induced modifications of their structures. Moreover, modifications in protein folding, depending on which reactive species attack the protein, are observed, although both oxidative and reductive radical stress are able to cause a significant loss in α-helix content in serum albumins and an increase in β-turns in M-MTs. Since some human pathologies have been associated with damage to unsaturated lipids [60,61], also in this case protein modifications should be used as marker events to eventually monitor pathological conditions. Thus, the chemical evidence acquired so far suggest to combine studies on oxidative and reductive processes to obtain a complete scenario of the reactivity of multifunctional substrates, like proteins, and the free radical stress consequences associated with cellular degeneration, related with aging and pathologies. Author Contributions: Conceptualization, A.T. (Armida Torreggiani); Methodology, A.T. (Armida Torreggiani), A.T. (Anna Tinti) and M.C.; Software, M.D.F. and Z.J.; Investigation, A.T. (Anna Tinti), M.S., Z.J.; Data curation, A.T. (Armida Torreggiani), A.T. (Anna Tinti), M.S., Z.J.; Writing—original draft preparation, A.T. (Armida Torreggiani), A.T. (Anna Tinti), and M.D.F.; Writing—review and editing, A.T. (Armida Torreggiani) and A.T. (Armida Torreggiani); Supervision, A.T. (Armida Torreggiani) and A.T. (Anna Tinti); Funding acquisition, A.T. (Armida Torreggiani). Biomolecules 2019, 9, 794 15 of 18 Funding: This research was funded by the project RM@Schools 3.0 of the EIT RawMaterials, Grant Number 17146. Acknowledgments: The authors are extremely grateful to Carla Ferreri (ISOF-CNR) who contributed to the set up of experiments with biomimetic systems; Chryssostomos Chatgilialoglu for the useful discussion on the reaction mechanisms; Mireia Tomas who performed experiments with MTs samples; Jordie Domènech who contributed to the conceptualization in the initial steps of this research; and Sílvia Atrian, friend and colleague, who worked hard to contribute to this research and passed away before the publication. Conflicts of Interest: The authors declare no conflict of interest. References 1. Heipieper, H.J.; Meinhardt, F.; Segyra, A. The cis-trans isomerase of unsaturated fatty acids in Pseudomonas and Vibrio: Biochemistry, molecular biology and physiological function of a unique stress adaptive mechanism. FEMS Microbiol. Lett. 2003, 229, 1–7. [CrossRef] 2. Wolff, R.L.; Entressangles, B. Steady-state fluorescence polarization study of structurally defined phospholipids from liver mitochondria of rats fed elaidic acid. Biochem. Biophys. Acta 1994, 1211, 198–206. [CrossRef] 3. Murakami, Y.; Tsuyama, M.; Kobayashi, Y.; Kodama, H.; Iba, K. Trienoic fatty acids and plant tolerance of high temperature. Science 2000, 287, 476–479. [CrossRef] [PubMed] 4. Ratnayake, W.M.N.; Chen, Z.-Y. Trans, ω-3, and ω-6 fatty acids in Canadian human milk. Lipids 1996, 31, S279–S282. [CrossRef] 5. Wardman, P. Thiyl radicals in biology: Their role as a ‘molecular switch’ central to cellular oxidative stress. In S-Centered Radicals; Alfassi, Z.B., Ed.; Wiley: Chichester, Great Britain, UK, 1999; pp. 289–309. 6. Sies, H. Glutathione and its role in cellular functions. Free-Radic. Biol. Med. 1999, 27, 916–921. [CrossRef] 7. Mozziconacci, O.; Williams, T.D.; Kerwin, B.A.; Schoneich, C. Intramolecular hydrogen transfer between protein cysteine thiyl radicals and α C−H bonds in insulin: Control of selectivity by secondary structure. J. Phys. Chem. B 2008, 112, 15921–15932. [CrossRef] 8. Douki, T.; Odin, F.; Caillat, S.; Favier, A.; Cadet, J. Predominance of the 1,N2-propano0 -deoxyguanosine adduct among 4-hydroxy -2-nonenal-induced DNA lesions. Free Radic. Biol. Med. 2004, 37, 62–70. [CrossRef] 9. Esterbauer, H.; Schaur, R.J.; Zollner, H. Chemistry and biochemistry of 4-hydroxynonenal malonaldehyde and related aldehydes. Free Radic. Biol. Med. 1991, 11, 81–128. [CrossRef] 10. Lykakis, I.N.; Ferreri, C.; Chatgilialoglu, C. The sulfhydryl radical (HS(.)/S(.- )): A contender for the isomerization of double bonds in membrane lipids. Angew. Chem. Int. Ed. 2007, 46, 1914–1916. [CrossRef] 11. Chatgilialoglu, C.; Ferreri, C.; Torreggiani, A.; Renzone, G.; Salzano, A.M.; Scaloni, A. Radiation-induced reductive modifications of sulfur-containing amino acids within peptides and proteins. J. Proteom. 2011, 74, 2264–2273. [CrossRef] 12. Chatgilialoglu, C.; Ferreri, C.; Melchiorre, M.; Sansone, A.; Torreggiani, A. Lipid Geometrical isomerism: From Chemistry to Biology and Diagnostics. Chem. Rev. 2014, 114, 255–284. [CrossRef] [PubMed] 13. Torreggiani, A.; Tamba, M.; Manco, I.; Faraone-Mennella, M.R.; Ferreri, C.; Chatgilialoglu, C. Radiation damage of lysozyme in a biomimetic model: Some insights by Raman spectroscopy. J. Mol. Struct. 2005, 744–747, 767–773. [CrossRef] 14. Torreggiani, A.; Tamba, M.; Manco, I.; Faraone-Mennella, M.R.; Ferreri, C.; Chatgilialoglu, C. Investigation of radical-based damage of RNase A in aqueous solution and lipid vesicles. Biopolymers 2006, 81, 39–50. [CrossRef] [PubMed] 15. Ferreri, C.; Manco, I.; Faraone-Mennella, M.R.; Torreggiani, A.; Tamba, M.; Manara, S.; Chatgilialoglu, C. The reaction of hydrogen atoms with methionine residues: A model of reductive radical stress-causing tandem protein-lipid damage. ChemBioChem 2006, 7, 1738–1744. [CrossRef] [PubMed] 16. Mee, L.K.; Adelstein, S.J.; Steinhart, C.M.; Lichtin, N.N. Reactions of aqueous L-methionine, L-phenylalanine, L-methionyl-L-phenylalanine, L-phenylalanyl-L-methionine and their mixtures with H atoms during steady radiolysis at pH 6.5. Radiat. Res. 1977, 71, 493–504. [CrossRef] [PubMed] 17. Hawkins, C.L.; Davies, M.J. Generation and propagation of radical reactions on proteins. Biochim. Biophys. Acta 2001, 1504, 196–219. [CrossRef] Biomolecules 2019, 9, 794 16 of 18 18. Garrison, W.M. Reaction mechanisms in the radiolysis of peptides, polypeptides, and proteins. Chem. Rev. 1987, 87, 381–398. [CrossRef] 19. Zhang, X.; Min, X.; Li, C.; Benjamin, I.J.; Qian, B.; Zhang, X.; Ding, Z.; Gao, X.; Yao, Y.; Ma, Y.; et al. Involvement of reductive stress in the cardiomyopathy in transgenic mice with cardiac-specific overexpression of heat shock Protein 27. Hypertension 2010, 55, 1412–1417. [CrossRef] 20. Bellezza, I.; Giambanco, I.; Minelli, A.; Donato, R. Nrf2-Keap1 signaling in oxidative and reductive stress. Biochim. Biophys. Acta–Mol. Cell Res. 2018, 1865, 721–733. [CrossRef] 21. Wadley, A.J.; Holliday, A.; Morgan, R.G.; Heesom, K.J.; Aldred, S.; Peters, D.M.; Bueno, A.A.; Coles, S.J. Preliminary evidence of reductive stress in human cytotoxic T-cells following exercise. J. Appl. Physiol. 2018, 125, 586–595. [CrossRef] 22. Majorek, K.A.; Porebski, P.J.; Dayal, A.; Zimmerman, M.D.; Jablonska, K.; Stewart, A.J.; Chruszcz, M.; Minor, W. Structural and immunologic characterization of bovine, horse, and rabbit serum albumins. Mol. Immunol. 2012, 52, 174–182. [CrossRef] 23. Vašák, M.; Meloni, G. Mammalian Metallothionein-3: New Functional and Structural Insights. Int. J. Mol. Sci. 2017, 18, 1117. 24. Capdevila, M.; Bofill, R.; Palacios, Ò.; Atrian, S. State-of-the-art of metallothioneins at the beginning of the 21st century. Coord. Chem. Rev. 2012, 256, 46–62. [CrossRef] 25. Aravindakumar, C.T.; Ceulemans, J.; De Ley, M. Nitric oxide induces Zn2+ release from metallothionein by destroying zinc-sulphur clusters without concomitant formation of S-nitrosothiol. Biochem. J. 1999, 344, 253–258. [PubMed] 26. Hartmann, H.J.; Weser, U. Copper-release from yeast Cu(I)-metallothionein by nitric oxide (NO). Biometals 2000, 13, 153–156. [CrossRef] 27. Capdevila, M.; Domènech, J.; Pagani, A.; Tío, L.; Villarreal, L.; Atrian, S. Zn- and Cd-metallothionein recombinant species from the most diverse phyla may contain sulfide (S2− ) ligands. Angew. Chem. Int. Ed. Engl. 2005, 44, 4618–4622. [CrossRef] 28. Maret, W.; Vallee, B.L. Thiolate ligands in metallothionein confer redox activity on zinc clusters. Proc. Natl. Acad. Sci. USA 1998, 95, 3478–3482. [CrossRef] 29. Domenech, J.; Tinti, A.; Capdevila, M.; Atrian, S.; Torreggiani, A. Structural study of the zinc and cadmium complexes of a type 2 plant (Quercus suber) metallothionein: Insights by vibrational spectroscopy. Biopolymers 2007, 86, 240–248. [CrossRef] 30. Tomas, M.; Tinti, A.; Bofill, R.; Capdevila, M.; Atrian, S.; Torreggiani, A. Comparative Raman study of four plant metallothionein isoforms: Insights into their Zn(II) clusters and protein conformations. J. Inorg. Biochem. 2016, 156, 55–63. [CrossRef] 31. Capdevila, M.; Cols, N.; Romero-Isart, N.; Gonzàlez-Duarte, R.; Atrian, S.; Gonzàlez-Duarte, P. Recombinant synthesis of mouse Zn3 -ß and Zn4 -α metallothionein 1 domains and characterization of their cadmium(II) binding capacity. Cell. Mol. Life Sci. 1997, 53, 681–688. [CrossRef] 32. Cols, N.; Romero-Isart, N.; Capdevila, M.; Oliva, B.; Gonzàlez-Duarte, P.; Gonzàlez-Duarte, R.; Atrian, S. Binding of excess cadmium(II) to Cd7 -metallothionein from recombinant mouse Zn7 -metallothionein 1. UV-VIS absorption and circular dichroism studies and theoretical location approach by surface accessibility analysis. J. Inorg. Biochem. 1997, 68, 157–166. [CrossRef] 33. Tu, A.T. Spectroscopy of Biological Systems; Clark, R.J.H., Hester, R.E., Eds.; Wiley: Chichester, UK, 1986; Volume 13, pp. 47–112. 34. Vass, E.; Hollosi, M.; Besson, F.; Buchet, R. Vibrational spectroscopic detection of beta and gamma-turns in synthetic and natural peptides and proteins. Chem. Rev. 2003, 103, 1917–1954. [CrossRef] 35. Torreggiani, A.; Tinti, A. Raman spectroscopy a promising technique for investigations of metallothioneins. Metallomics 2010, 2, 246–260. [CrossRef] [PubMed] 36. Jurasekova, Z.; Tinti, A.; Torreggiani, A. Use of Raman spectroscopy for the identification of radical-mediated damages in human serum albumin. Anal. Bioanal. Chem. 2011, 400, 2921–2931. [CrossRef] 37. Kramer, A.C.; Torreggiani, A.; Davies, M.J. Effect of Oxidation and Protein Unfolding on Cross-Linking of beta-Lactoglobulin and alpha-Lactalbumin. J. Agric. Food Chem. 2017, 65, 10258–10269. [CrossRef] Biomolecules 2019, 9, 794 17 of 18 38. Buxton, G.V.; Greenstock, C.L.; Helman, W.P.; Ross, A.B. Critical-review of rate constants for reactions of hydrated electrons, hydrogen atoms and hydroxyl radicals (• OH/• O− ) in aqueous-solution. J. Phys. Chem. Ref. Data 1988, 17, 513–886. [CrossRef] 39. Spinks, J.W.T.; Woods, R.J. An Introduction to Radiation Chemistry, 3rd ed.; Wiley & Sons: New York, NY, USA, 1990; pp. 243–313. 40. Winterbourn, C.C. Toxicity of iron and hydrogen peroxide: The Fenton reaction. Toxicol. Lett. 1995, 82/83, 969–974. [CrossRef] 41. Steinberg, D.; Witztum, J. Is the oxidative modification hypothesis relevant to human atherosclerosis? Do the antioxidant trials conducted to date refute the hypothesis? Circulation 2002, 105, 2107–2111. [CrossRef] 42. Kramer, J.F.K.; Fellner, V.; Dugan, M.E.R.; Sauer, F.D.; Mossoba, M.M.; Yurawecz, M.P. Evaluating acid and base catalysts in the methylation of milk and rumen fatty acids with special emphasis on conjugated dienes and total trans fatty acids. Lipids 1997, 32, 1219–1228. [CrossRef] 43. Ferreri, C.; Samadi, A.; Sassatelli, F.; Landi, L.; Chatgilialoglu, C. Regioselective cis-trans isomerization of arachidonic double bonds by thiyl radicals: The influence of phospholipid supramolecular organization. J. Am. Chem. Soc. 2004, 126, 1063–1072. [CrossRef] 44. Talamo, F.; D’Ambrosio, C.; Arena, S.; Del Vecchio, P.; Ledda, L.; Zehender, G.; Ferrara, L.; Scaloni, A. Proteins from bovine tissues and biological fluids: Defining a reference electrophoresis map for liver, kidney, muscle, plasma and red blood cells. Proteomics 2003, 3, 440–460. [CrossRef] [PubMed] 45. Melchiorre, M.; Ferreri, C.; Tinti, A.; Chatgilialoglu, C.; Torreggiania, A. A Promising Raman Spectroscopy Technique for the Investigation of trans and cis Cholesteryl Ester Isomers in Biological Samples. Appl. Spectrosc. 2015, 69, 613–622. [CrossRef] [PubMed] 46. Mihaljević, B.; Tartaro, I.; Ferreri, C.; Chatgilialoglu, C. Linoleic acid peroxidation vs. isomerization: A biomimetic model of free radical reactivity in the presence of thiols. Org. Biomol. Chem. 2011, 9, 3541–3548. 47. Chatgilialoglu, C.; Ferreri, C.; Ballestri, M.; Mulazzani, Q.G.; Landi, L. Cis-trans isomerization of monounsaturated fatty acid residues in phospholipids by thiyl radicals. J. Am. Chem. Soc. 2000, 122, 4593–4601. [CrossRef] 48. Torreggiani, A.; Domènech, J.; Orihuela, R.; Ferreri, C.; Atrian, S.; Capdevila, M.; Chatgilialoglu, C. Zinc and cadmium complexes of a plant metallothionein under radical stress: Desulfurisation reactions associated with the formation of trans-lipids in model membranes. Chem. Eur. J. 2009, 15, 6015–6024. [CrossRef] 49. Torreggiani, A.; Chatgilialoglu, C.; Ferreri, C.; Melchiorre, M.; Atrian, S.; Capdevila, M. Non-enzymatic modifications in metallothioneins connected to lipid membrane damages: Structural and biomimetic studies under reductive radical stress. J. Proteom. 2013, 92, 204–2015. [CrossRef] 50. Salzano, M.; Renzone, G.; Scaloni, A.; Torreggiani, A.; Ferreri, C.; Chatgilialoglu, C. Human serum albumin modifications associated with reductive radical stress. Mol. BioSyst. 2011, 7, 889–898. [CrossRef] 51. Tomas, M.; Jiménez-Martí, E.; Bofill, R.; Tinti, A.; Di Foggia, M.; Chatgilialoglu, C.; Torreggiani, A. Comparative analysis of the soybean metallothionein system under radical stress: Protein modification connected to lipid membrane damages. Metallomics 2018, 10, 1792–1804. [CrossRef] 52. Tuma, R. Raman spectroscopy of proteins: From peptides to large assemblies. J. Raman Spectrosc. 2005, 36, 307–319. [CrossRef] 53. Torreggiani, A.; Domenech, J.; Tinti, A. Structural modifications in metal complexes of a plant metallothionein caused by reductive radical stress: A Raman study. J. Raman Spectrosc. 2009, 40, 1687–1693. [CrossRef] 54. Krimm, S.; Bandekar, J. Vibrational analysis of peptides, polypeptides, and proteins. V. Normal vibrations of beta-turns. Biopolymers 1980, 19, 1–29. [CrossRef] 55. Torreggiani, A. Structural lesion to ribonuclease A caused by reductive stress: Assessment by Raman spectroscopy. Spectroscopy 2008, 22, 279–286. [CrossRef] 56. Mee, L.K.; Adelstein, S.J.; Stein, G. Inactivation of Ribonuclease by the Primary Aqueous Radicals. Rad. Res. 1972, 52, 588–602. [CrossRef] 57. Davies, M.J.; Fu, S.; Wang, H.; Dean, R.T. Stable markers of oxidant damage to proteins and their application in the study of human disease. Free Radic. Biol. Med. 1999, 27, 1151–1163. [CrossRef] 58. Liddle, W.K.; Tu, A.T. Evaluation of phenylalanine and tyrosine Raman lines for the amino acid ratio. Appl. Spectrosc. 1981, 35, 444–446. [CrossRef] 59. Ploder, M.; Neurauter, G.; Spittler, A.; Schroecksnadel, K.; Roth, E.; Fuchs, D. Serum phenylalanine in patients post trauma and with sepsis correlate to neopterin concentrations. Amino Acids 2008, 35, 303–307. [CrossRef] Biomolecules 2019, 9, 794 18 of 18 60. Ferreri, C.; Angelini, F.; Chatgilialoglu, C.; Dellonte, S.; Moschese, V.; Rossi, P.; Chini, L. Trans fatty acids and atopic eczema/dermatitis syndrome: The relationship with a free radical cis-trans isomerization of membrane lipids. Lipids 2005, 40, 661–667. [CrossRef] 61. Balazy, M.; Chemtob, S. Trans-arachidonic acids: New mediators of nitro-oxidative stress. Pharmacol. Ther. 2008, 119, 275–290. [CrossRef] © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).