Crystalline cysteine (Sigma Aldrich) was solubilized in double distilled water (ddH₂O) to a final concentration of 2 mM on a daily basis. For short-term storage, blue ice was used to avoid pH value distortions during freeze-thawing cycles [57]. After respective plasma or irradiation treatments, solutions were immediately subjected to high-resolution mass spectrometry, multiple reaction monitoring mass spectrometry, or ion chromatography. Fresh porcine ears were received from Landmetzgerei Urich (Bad Koenig, Germany) on blue ice, serving as a well-accepted replacement model for human skin [58]. The ears were washed carefully, shaved, and the superficial corneocyte layer was removed with a single CorneoFix strip (Courage and Khazaka electronic GmbH, Cologne, Germany) to increase homogenicity. Plasma treatment was performed in selected clean and homogenous areas of a 2 cm × 2 cm dimension.

The kINPen, an argon-driven (99.999%, Air Liquide) dielectric barrier plasma jet with a flow rate of 3 standard liters per minute (slm) served as plasma source. If desired, the working gas was modified by 1% admixture of molecular oxygen (99.998, Air Liquide). Its central electrode is powered by an alternating current with a sinusoidal waveform, 2–6 kV peak–peak voltage, and a frequency of around 1 MHz. The outer electrode is insulated by a ceramic tube. The dissipated electrical power is around 1.1 W. For most experiments, a gas curtain created by a nozzle and 5 slm nitrogen (99.999%, Air Liquide) shielded the effluent from the ambient air. For details about the design and working principle of the jet, refer to Reuter and colleagues [9] and citations therein. To investigate the plasma-derived products and emitted radiation by optical emission spectroscopy, the plasma jet was positioned on axis at a distance of 9 mm to the front of a spectrometer (AvaSpec-2048; Avantes, Germany) allowing the observation of both UV and VIS/NIR range (195–980 nm) with a spectral resolution of 0.7 nm. For the VUV spectral measurements, a single grating monochromator (Acton VM-502, grating 1200 g/ mm) was used. This system was set to a spectral resolution of 0.2 nm and the spectral range of 100–200 nm was observed. Furthermore, the system was under low pressure (2.2*10⁻⁶ mbar) and connected via an MgF₂ window for VUV transmissions down to 100 nm. The kINPen was placed at 9 mm distance in front of the MgF₂ window.

The distance between the nozzle and the target was kept at 9 mm. Targets were either 750 µL aqueous solution in a 24 well-plate, fresh porcine skin prepared as described in 2.1, or a 25 mm diameter radiation chamber. The chamber could be equipped with windows that transmit different parts of the VUV/UV radiation and was filled with 80 µL solution forming a 500 µm thick layer or sections of porcine skin (Figure 1A,B). After treatments (20 s–180 s) samples were submitted to reactive species analysis (Section 2.3). Liquids containing the tracer molecule cysteine were also analyzed via liquid chromatography coupled to mass spectrometry for the detection of oxidative modifications (Section 2.4). The first 3 layers of treated porcine skin tissues were sampled using three consecutive CorneoFix strips that were immediately deposited in a protein solubilization buffer and subjected to shotgun proteomics (as described in Section 2.5). Each experiment was performed in triplicate.

For the quantification of plasma-generated, short-lived ROS (hydroxyl radicals, atomic oxygen) solution of 5 mM terephthalic acid in 25 mM phosphate buffer, pH 7.4 was used despite the limited selectivity [59]. The reaction yield to the fluorescent compound 2-hydroxyterephthalic acid (HTPA) that could be quantified at 318 nm excitation and 426 nm emission using an external calibration curve. Hydrogen peroxide deposited in treated liquids was determined using the ferrous oxidation–xylenol orange (FOX) assay according to the manufacturer’s protocol (Thermo Scientific, Dreieich, Germany). The reaction yield to a purple product, which absorbance at 595 nm was measured in a spectrophotometer (Tecan M200 multi-plate reader, Männedorf, Switzerland).

The kINPen spectra of emission goes from the vacuum UV (100 nm) to the near infrared (1,000 nm) [61]. The intensity of the UV irradiation was determined to be around 100 μJ cm⁻². Optical emission spectroscopy was performed in order to characterize the emission spectra of the applied plasma treatment conditions (Figure 1). In this case, the emission spectra were measured applying the three different windows filtering various radiation ranges (Figure 2). Additionally, the vacuum UV range was recorded using a vacuum setup (Figure 3) [9,37]. In pure argon, emission lines from impurities of •OH and N₂ were measured between 300 and 350 nm, and small atomic oxygen lines at 777 and 844 nm. The infrared region was dominated by argon emission lines, while between the range 400–700 nm no emission was observed. In the vacuum UV region (<195 nm), a dominant continuum centered at 126 nm can be measured for argon excimer (Ar₂ ^*), which includes also small absorption lines from ozone and O₂. The modulation of working or shielding gases, induced changes in the emission spectra, e.g., the addition of molecular oxygen in the working gas leads to a reduction of Ar₂ ^* emission lines and increase those oxygen-based [9,37]; (Figure 2). The presence of molecular O₂ in the working gas interfered with the gas phase chemistry and UV emission of various species, such as ^•OH, N, ^•NO, and Ar₂ ^*. Suprasil and Borofloat-33 scavenged the emission of argon excimers, while in the range between 195 and 465 more subtle changes were observed, reflecting the characteristics of normal glass (Figure 2). Particularly in conditions with pure argon, which are certified for medical applications with kINPen MED, the role of vacuum UV radiation was dominant (Figure 3). As discussed previously, this event could be due to the lower formation of gaseous reactive species and therefore less reactions of argon metastable and excimers with surrounding gases to form further species, e.g., ozone, atomic oxygen, singlet oxygen, etc. [9,37]; (Figures 2, 4).

The formation of ^•OH and H₂O₂ in water was compared using full plasma treatments (3 slm Ar and Ar + 1% O₂) and treatments using windows filtering different radiation ranges (Figure 5). The dissociation of water yielding OH radicals was favored in full argon plasma treatments and when VUV radiation was admitted (MgF₂-window). For short treatments only small differences in ^•OH production was observed, but enlarging with treatment time. Full argon plasma produced around 30 µmol of ^•OH more than Ar plasma/MgF₂-window. It may be argued that either OH radicals from the gas phase contributed here or atomic oxygen generated from impurities. A recent report showed the sensitivity of terephatalic acid to this reactive short-lived species. Also liquid dynamic effects increasing the contact area between emitted radiation and target in contrast to the static treatments performed in the micro-chamber can contribute. Upon addition of molecular oxygen to the working gas, a decrease in ^•OH and H₂O₂ production was observed, especially when the MgF₂-window was used, blocking interphase chemistry. In the near-complete absence of VUV radiation from the argon excimers, water dissociation did not occur. Corroborating this observation, almost no OH/H₂O₂ formation was sparked by the longer UV ranges (UV-C and B and UV-A, respectively Suprasil-1 and Borofloat 33). Therefore, oxygen-base emission lines (atomic oxygen in NIR, ozone and O₂ in VUV), which increases in presence of molecular oxygen in the working gas, did not led to water dissociation. However, their potential direct impact on cysteine structures was investigated [55]. Overall, being highly energetic, VUV radiation emitted by Ar₂ ^* were able to induce water ionization, in contrast to the other UV ranges. Therefore, the direct and indirect (e.g., water-derived species production) impact of VUV radiation resulted as predominant responsible of the effects induced on the liquid target. In the presence of oxygen in the gas phase, species derived from interphase chemistry and/or chemistry in liquid bulk are dominant, while VUV radiation is eliminated (Figure 3A).

The impact of full plasma treatments and plasma-generated UV/VIS/NIR light on cysteine molecules is shown in Figure 6. Table 2 shows the structures, the quantification methods, and the major plasma components involved in the generation of each cysteine derivative. As discussed, argon plasma conditions stimulated predominantly pathways in liquid via impact of VUV radiation emitted by argon excimers, leading to ^•OH and H₂O₂ production. The same experiments were performed with cysteine containing solutions. The formation of products that derive from the loss of the thiol group in cysteine (alanine, cysteine-S-sulfonate, sulfite ions, sulfate ions, and hydrogen sulfide) was almost identical in treatments with full argon plasmas and filtered VUV radiation of argon plasmas (MgF₂). In contrast, these products almost disappeared when VUV-impermeable oxygen was introduced to the working gas and gas–liquid phase chemistry was blocked by an MgF₂-window. This confirmed the key role of VUV radiation, leading to a cleavage of the carbon-sulfur bond [63], that was not observed by using other windows/radiation ranges. The energy of this bond is 272 kJ mol⁻¹, weaker than other bonds in cysteine or water (H-O 465 kJ mol⁻¹, C-C 347 kJ mol⁻¹, S-H 347 kJ mol⁻¹). The energy of the impinging UV photons of the argon excimer lines is much higher, 949 kJ mol⁻¹ (126 nm), allowing the cleavage of all bonds present in the target, including the oxygen-hydrogen bond in water molecules. Accordingly, it might be argued that OH radical formation is the first step ultimately yielding in thiol moiety abstraction. However, the presence of hydrogen sulfide (H₂S) clearly indicates that a direct cleavage of the C-S bond contributes significantly or is even dominant considering the weaker bond energy compared to the H-O bond. Interestingly, no indications of a C-C bond breakage in cysteine, leading e.g. to the formation formic acid, was observed. Summarizing, conditions with pure argon showed a significant contribution of the VUV radiation, stimulating i) the production of OH radicals, and ii) C-S bond breakages. Under VUV radiation, derivatives such as cystine and cysteine acids were also produced in consistent amounts in relation to full argon plasmas. The origin of the oxygen incorporated in structures such as sulfite, sulfate, cysteine acids and S-sulfonate, in this case, are water-derived species [55].

As previously discussed, due to the controlled pH (7.2), only low amounts of thiolate were available to react with H₂O₂ since the pKa of the cysteine thiol group is 8.18—allowing less than 5% deprotonation. Therefore, a minimal role in thiol oxidation could be attributed to hydrogen peroxide. Furthermore, a two-step reaction would be needed to form cystine: a first reaction of H₂O₂ with thiolate, with formation of cysteine sulfenic acid, and a second reaction of RSOH with another thiolate [63]; [64]. More likely, the reaction of ^•OH with cysteine generates cystine by first formation of thiyl radicals (RS^•), which rapidly recombine to form cystine. The formation of cysteinyl radicals was detected using BMPO/EPR spectroscopy earlier [54].

The reaction of RS^• with ^•OH would lead to cysteine sulfenic acid, which immediately is oxidized into sulfinic and further to sulfonic acid by ^•OH or H₂O₂ [65–67]. The production of S-sulfonate (RSSO₃H), in absence of atomic oxygen, could be promoted by first cystine oxidation by two ^•OH, with following C-S breakage promoted by VUV and final incorporation of another oxygen. Sulfite can be generated by cut of the C-S bond in cysteine sulfonic acids, or most likely, cysteine S-sulfonate (S-S bond dissociation energy 414.6 kJ mol⁻¹) [68]. Supplementary experiments have been performed by treating cysteine sulfonic acid in the radiation chamber, as shown in the supplementary material (Supplementary Figure S1). Despite the high concentration of the reference compound cysteine sulfonic acid, only small amounts of sulfite and sulfate ions are formed, indicating that the majority of C-S bond cleavages takes place at the cysteine or cystine level. With that, most sulfite and sulfate ions were generated by the oxidation of H₂S. This pathway is favored in the presence of reactive oxygen species produced by the cleavage water molecules (e.g. H₂O₂, ^•OH). Clearly, a number of chemical pathways were active in liquids under the influence of vacuum UV radiation, in contrast to radiation >195 nm (UV-C).

The cysteine products generated by UV-C (Suprasil) or UVB/UVA (Borofloat) were clearly different. In these cases, almost no production of the water-derived species OH/H₂O₂ (Figure 5) and cysteine derivatives generated by C-S bond cleavage (SO₄ ^2-, SO₃ ^2-, alanine/RH; Figure 6) were observed.

The formation of cystine via hydrogen abstraction in cysteine (type I photo-oxidation mechanisms) and RS^• recombination [50,51]; was observed independent from the window, but with lower extent when vacuum UV is blocked (Borofloat/Suprasil). The production of sulfonic acid (RSO₃H, Figure 6) was measured in conditions with oxygen in the working gas, suggesting the potential role of radiation emitted in the near infrared and depending by the presence of O₂. The origin of oxygen incorporated by sulfonic acid, in this case, could be due to the reaction of thiyl radical with water, generating ^•OH by hydrogen abstraction. Even though the radiation emitted in these ranges are not ionizing, due to the lower energy, it was shown that also the UV radiation >200 nm can induce oxidative stress in cellular compartments in relation to the exposure time, with increase in the cellular production of reactive species [49,50].

The analysis of the oxidative modifications of cysteine belonging to the porcine epidermal proteome was performed via tape stripping assay in combination with high-resolution mass spectrometry. The oxidation was monitored in the first three stratum corneum layers of the porcine epidermis, reflecting a penetration depth of 5–7 µm. Table 3 and Figure 7 give an overview of the detected modifications, while Figure 8 shows a quantitative comparison of the most relevant oxidative modifications. Unstable oxidative modifications, such as S-nitrosylation, may have been underestimated. According to proteomics standard procedures, thiols were reduced and alkylated during sample workup. Some losses to the plasma chemistry on protein thiols (e.g., sulfenic acids, S-nitrosylation) cannot be excluded although stable modifications (e.g., cysteine sulfonic acid, cysteine-S-sulfonate) were retained. Overall, five types of modifications were found to be introduced into the epidermal proteins with a statistical significance. These are the apparent replacement of cysteine by alanine (sulphur loss), the formation of cysteine sulfonic acid (trioxidation), the formation of cysteine-S-sulfonate (sulfonylation), the formation of a disulfide with other cysteine moieties (+S2R), and the loss of two hydrogen atoms forming an unsaturated cysteine molecule (didehydrogenation) (Table 3). Alongside depth in the epidermis, the intensity and number of detected modifications decrease (Figure 7). The impact of plasma-derived UV or VUV light was lower than expected from the in vitro experiments (Table 3). Only nitrosylation was found to be significantly elevated by MgF₂ filtered kINPen irradiation, pointing at a limited role of (V)UV photons in vivo. The conversion of cysteine in cysteine S-sulfonate (+SO₃) and dehydrogenated cysteine (-2H) (Figure 8), was promoted especially in the third skin epidermis layer, regardless of the applied treatment condition. The results indicate that protein modifications result from complex dynamics and are not corresponding to the action of only one plasma components produced in specific conditions.

Some modifications appear in special conditions only. For example, the conversion of cysteine into cystine was observed in treatment with Ar/O₂ full plasmas, with a maximum in the first layer and a progressive decrease in the second and third. In this condition, the effluent contains significant amounts of singlet oxygen and atomic oxygen, especially at short distances from the nozzle. Atomic oxygen is able to form thyil radicals by hydrogen abstraction to the protein cysteines, which rapidly may recombine forming a disulfide bond. This event causes conformational changes of the protein, alongside a potential gain or loss of function. The oxidation of cysteine into sulfonic acid by the incorporation of three oxygen atoms, and the breakage of the C-S bond with conversion of cysteine to alanine were events observed predominantly in treatments with full argon plasma in the first layer of the skin. When using windows, even in case of the VUV-transmitting MgF₂ and more prominent in suprasil and Borofloat-33 windows, only few such modifications were detected. This indicates that the VUV radiation, although most prominent in Ar plasma, does in soft targets not contribute in the same manner to biomolecule modification than in liquid targets. In this case, the direct impact of argon metastables and other gas phase species and the subsequent formation of secondary reactive species is more prominent. This is in line with a report investigating the oxidation of human skin lipids that showed a limited impact of the argon plasma on lipid side chain oxidation in the absence of water [33]. Obviously, the highly energetic VUV radiation is unable to penetrate deeper into layers of biomolecules such as the model described here or the sebum lipids, limiting its ability to contribute significantly to the plasma chemistry in situ.

The proteomics data connected to this paper have been uploaded to the ProteomeXchange servers under the project name “Gas plasma and (V)UV impact on porcine epidermis using a tape strip assay approach” (Project accession: PXD028915, username: reviewer_pxd028915@ebi.ac.uk/Password: OPKS8hii).