Proteomics reveals that quinoa bioester promotes replenishing effects in epidermal tissue

The continuous search for natural products that attenuate age-related losses has increasingly gained notice; among them, those applicable for skin care have drawn significant attention. The bioester generated from the Chenopodium quinoa’s oil is a natural-origin ingredient described to produce replenishing skin effects. With this as motivation, we used shotgun proteomics to study the effects of quinoa bioester on human reconstructed epidermis tridimensional cell cultures after 0, 3, 6, 12, 24, and 48 h of exposure. Our experimental setup employed reversed-phase nano-chromatography coupled online with an Orbitrap-XL and PatternLab for proteomics as the data analysis tool. Extracted ion chromatograms were obtained as surrogates for relative peptide quantitation. Our findings spotlight proteins with increased abundance, as compared to the untreated cell culture counterparts at the same timepoints, that were related to preventing premature aging, homeostasis, tissue regeneration, protection against ultraviolet radiation and oxidative damage.

www.nature.com/scientificreports/ Essential fatty acids have been shown to regulate several cell signaling pathways involved in skin inflammation, dehydration, and tissue degradation 21 . To date, the effects of quinoa derivatives have been investigated only through DNA and mRNA sequencing 12,22 .
Proteomics has served at the forefront as a tool for the development and evaluation of innovative cosmetic products 23,24 . Here, we investigated the proteomic alterations in human epidermis tridimensional (3D) cell culture (RHE), co-cultivated with dermal fibroblasts, after 0 h, 3 h, 6 h, 12 h, 24 h, and 48 h of exposure to quinoa bioester (QB); comparisons were performed with unexposed counterparts at the same time points. The results allowed us to draw important conclusions about how this potent ingredient provides beneficial effects, at the protein level, and pinpointed key proteins related to homeostasis and other key beneficial epithelial tissue functions.

Results and discussion
Wrinkles, sagging, spots, and tissue dehydration are associated with skin aging and are aggravated without proper care habits such as frequent use of cosmetics and a healthy diet. Skin health impacts the aesthetic appearance and quality of life; unhealthy skin may lead to the development of dermatoses, itching, depigmentation, fungal or bacterial infections 25 . Natural ingredients serve as a treasure trove to form the basis of new cosmetics 24 . The Chenopodium Quinoa oil, evaluated in this study, is rich in essential fatty acids, minerals, and amino acids; these are known to be highly emollient and replenishing to the skin. In addition, quinoa oil is a source of antioxidant tocopherols and a potent anti-inflammatory complex that helps replenish the barrier function of the epidermis and prematurely combat tissue damage 26 . With this as motivation, we performed a time-course experiment to verify the effects QB on the epidermis in vitro 27 . We opted for the Reconstructed Human Epidermis (RHE) cell culture as this system mimics the in vivo 3D structure of epidermal tissue as well as the conditions and processes that occur in exposure to exogenous factors 23 . Microscopy assessment of RHE differentiation. Keratinocyte migration time from basal layer to stratum corneum was evaluated through phase-contrast microscopy. In accordance with previous results, our results show that it is possible to observe the complete differentiation of epidermal layers after 15 days of in vitro culture ( Fig. 1) 28,29 . Proteomic identifications. Our proteomic analysis identified up to 3981 proteins with 32,897 peptides.
The complete list of identified proteins is found in Supplementary Usage of T0 as experimental quality control for our bioinformatics pipeline. The 0 h time-point (T0) of our cell culture was used to verify the effectiveness of our experimental and computational pipeline as more biological replicates were acquired. The assessment for chromatographic reproducibility of technical replicates was done using RawVegetable ( Supplementary Fig. S1) 30 . It is expected that no (or almost no) statistically differentially abundant proteins should be identified when comparing groups of biological replicates from the same biological condition. As expected, no differential proteins were shortlisted in PatternLab's T-Fold analysis ( Fig. 2A). In contrast, all other timepoints listed differential proteins when comparing the QB-treated cell culture with its QB-free counterpart for the same timepoint. Figure 2B  QB induces Acyl-Coa that is linked with homeostasis. Acyl-Coa is part of the Fatty acid metabolism pathway. The stratum corneum (SC) serves as a barrier between the deeper layers of the skin and the external environment and controls homeostasis 31,32 ; improper hydration impairs certain enzymatic functions resulting in www.nature.com/scientificreports/ the adhesion and accumulation of corneocytes 33 . The SC is commonly described as the skin's brick-and-mortar, where the anucleate corneocytes, mainly composed of keratin, are within a matrix rich in lipids containing cholesterol, ceramides, esters, and fatty acids 34 . The Acyl-CoA protein (ACBP) binds to esters with high specificity and affinity and acts as an intracellular carrier in various enzymatic systems. ACBP is abundant in the epidermis, the suprabasal layers, which are highly active in lipid synthesis. According to Bloksgaard et al., the silencing of the ACBP gene in mice caused oiliness, development of alopecia, and skin scaling. Moreover, it compromised the function of the epidermal barrier, causing an increase in the loss of transepidermal water, indicating this consequence of reduced levels of non-esterified fatty acids in the stratum corneum 29 . Our results demonstrate QB stimulates Acyl-Coa after 3 h (59%), 6 h (62%), 12 h (85%), and 24 h (99%) (p < 0,01), according to the TFold analyses.
QB stimulates glutamine synthetase that has been linked with skin regeneration. Glutamine synthetase is part of the metabolism pathway and is the only enzyme able to catalyze the synthesis of Glutamine (Gln) from ammonia and glutamate; this amino acid is essential for various tissue functions. GS plays an essential role in the acid-base balance and is used as an energy source. In cellular division, it acts as a precursor for the synthesizing of several biologically active compounds, such as purines and pyrimidines 35 . A deficiency of glutamine leads to maleficent responses such as the appearance of erythema and the formation of blisters on the tegument 36,37 . The importance of glutamine in the role of recovery from burn injuries has also been described 38 . GS plays key roles throughout the various layers of the epidermis: in the basal layer, for housekeeping the keratinocyte accumulation cells and in the stratum corneum as a physical and chemical barrier against UV and pollution 31 . GS also controls the homeostasis in the epidermis 39 , provides tissue resistance, and reduces the paracellular permeability 40 . Our results showed that QB favors an increase in the abundance of GS in our 3D cell culture after 3 h (26%), 6 h (38%), and 12 h (34%) (p < 0.01), according to TFold analyses for all time points.
QB stimulates Actin-related protein 2/3 subunits 2 and 4 that are linked with epidermal morphogenesis and homeostasis. The Actin-related protein 2/3 (Arp 2/3) is part of the EPH-Ephrin signaling pathway. Skin barrier alterations, hyperproliferation, and epithelial hypertrophy are characteristic of epidermal homeostasis changes, leading to different diseases, such as psoriasis 41,42 . The actin-related protein (Arp2/3) complex consists of two actin-related proteins and five additional actin-associated protein complex subunits (Arpc1-5); Arpc2 and Arpc4 as a core subunit. The Arp2/3 complex regulates actin-associated processes, such as endocytosis, cell migration, vesicle trafficking, organelle remodeling, and cell-matrix and cell-cell adhesion 43 .
Studies have shown that the downregulation of the Arp2/3 complex in mouse epidermis causes interference in morphogenesis and homeostasis 44 . Our data analysis showed that QB increased the abundance Arpc4 after 3 h (24%), 12 h (47%), and Arpc2 in 12 h (68%) and 24 (63%) with p < 0.01, thus suggesting that it could be beneficial for those influences epidermal morphogenesis and homeostasis, according to TFold analyses for all time points.
Cellular retinoic acid-binding protein-II (CRABP-II) is more abundant in QB-exposed cells and has been linked to preventing premature aging. The CRABP-II is part of the retinoic acid signaling pathway. Skin aging is classified into extrinsic aging, by environmental exposure, such as UV radiation and intrinsic determined by genetic factors 45 . CRABP-II is expressed by suprabasal fibroblasts and keratinocytes and defines a family of proteins that bind to all-trans-retinoic acid (atRA) 46 . atRA's have a profound effect on the growth and differentiation of human epidermal cells in vivo and in vitro 47 playing a crucial role in skin homeostasis 48 , controlling the epithelial width, thickening the epidermal by boosting the proliferation of keratinocytes and thus serving as UV protection 49 and ultimately for preventing carcinogenesis 50 .
The biologically active form of retinoic acid is vitamin A, also known as retinol (ROL), a precursor of retinoic acid. Human skin can convert ROL into its biologically active retinoic metabolite. When used topically on human skin, ROL permeates it, becomes converted to retinaldehyde and then to retinoic acid 51,52 . The signaling of retinoic acid (RA) is essential for epidermal differentiation 53 . The regulation of intracellular retinoid bioavailability is made by the presence of specific retinol and retinoic acid-binding proteins, such as CRAPBS 50 . Our results showed that QB increases the levels of the cellular retinoic acid-binding protein-II; effects were especially notable after 48 h presenting an increase in abundancy of ~ 50% (p < 0.01), according to the TFold analyses for all time points. QB induces downregulation in S100-A2 that is linked with oxidant defense. S100 proteins belong to a family of cytosolic calcium-binding proteins, composed of 25 members 54 with different intracellular and extracellular functions. The S100A2 protein is located in the basal layer of the human epidermis 55 , having its overexpression in epidermal dysfunctions of morphogenesis and homeostasis 56 . According to Zhang et al., downregulation of S100-A2 is associated with defense against oxidants in epithelial tissue 57 . Our results demonstrate the S100-A2 downregulation at 3 h (15%) and 12 h (29%) (p < 0.01), according to the TFold analyses.
Time-course analysis suggests that QB favors cornification. PatternLab's TrendQuest module was applied to group proteins that shared a similar abundancy profile over our time-course experiment. The software converged to five clusters for QB-Free and another five for QB-Treated cells. Only proteins found in three or more time points were considered. Table 1 provides a bird's-eye view for all clusters; each one is presented side-by-side with its most enriched pathway. Supplementary Table S7 includes detailed information and plots for all clusters. In general, the enriched pathways suggest that QB favors cornification. Cornification refers to the formation of a dead cell (corneocyte) layer that serves as a protective physical barrier for the skin 58 . Several pathways are activated during homeostatic keratinocyte differentiation to control the keratinocytes from pre- www.nature.com/scientificreports/ mature apoptosis and necrosis to enable the keratinization process 59,60 . Among the enriched pathways displayed in Table 1, we highlight the "The citric acid (TCA) cycle and respiratory electron transport" (Table 1-Q4) and the "Cholesterol biosynthesis" (Table 1- Our results demonstrate that both pathways were triggered after 12 h, suggesting that QB favors the renewal of essential elements to enable homeostasis and the integrity of the skin barrier. In contrast, these pathways were far from topping the list in the QB-Free cell line. In fact, in QB-free cells, the "Formation of the cornified envelope" pathway (Table 1-F4) decreases over time as the "Methylation" (Table 1-F3) pathway increased. Such an inverse correlation is well described in the literature; increased methylation suppresses the differentiation and maintains cell proliferation at baseline levels 62 . Figure 3 contrasts profiles from clusters Q3 and F4; a joint Reactome analysis (Fig. 4) shows that both are related to the cornification process and yet, in our results, their general abundancy trend is indirectly correlated. Finally, we note a common enriched pathway to both QB-free and QB cell-lines: "Formation of the ternary complex, and subsequently, the 43S complex" (Table 1, Q1 and F1). The aforementioned pathway is found in clusters that share a similar protein profile distribution for both the QB and QB-free cells; it is related to essential tasks and thus remains with its relative quantitation mostly unaltered throughout.  63 . Although this work focused on evaluating protein expressions by keratinocytes, fibroblasts were co-cultivated in a different experimental compartment to simulate epidermal and dermal communication.

Material and methods
Fibroblasts were cultured in DMEM supplemented with 10% FBS and 1% P/S, while keratinocytes were cultured with KBM Gold medium. For subcultures, the confluent monolayers were gently washed with PBS and after brief 3-min trypsinization, the cells were suspended in the complete culture medium. For the formation of monolayers, fibroblasts were cultivated into three 12-well plates. After 3 days, the medium was exchanged, and ThinCert was placed in each well with keratinocytes added to its superior face. Finally, 3 days after, the entire liquid content of the insert and the wells were removed, and then performed at air-liquid interphase filled with 1 mL Grupo Boticário's propriety differentiation medium into plates. The medium was changed every 3 days, during 15 days for epidermal differentiation.  The software parameters were Benjamin Hochberg q-value (FDR) of 0.05, F-Stringency 0.10, and L-Stringency 0.60. Each dot represents a protein that is mapped according to its − log2(P-value) (x-axis) and log 2 (fold change) (y-axis). Red dots are proteins that do not satisfy the fold change cutoff and the q-value cutoff. Green dots are proteins that satisfy the fold change but not the q-value cutoff. Orange dots are proteins that satisfy both the fold change cutoff and q-value cutoff but received very low quantitative values and therefore were disregarded from the analysis. Finally, the blue dots are proteins that satisfy all statistical filters and the ones we consider as statistically differentially abundant. (A) Comparison of groups of 3D cell cultures not exposed to QB. Sample preparation. RHE proteins were extracted with RapiGest detergent at a concentration of 0.1% according to the manufacturer's recommendations. According to the manufacturer's instructions, protein concentrations were determined using the fluorimetric assay from the Qubit platform (Invitrogen). One hundred micrograms of proteins from each sample were reduced with dithiothreitol (DTT) (final concentration of 10 mM) for 30 min, at 60 °C. After being cooled to room temperature, the samples were alkylated with iodoacetamide (final concentration of 30 mM) for 25 min at room temperature, in the dark, and finally digested with sequence grade modified trypsin in the proportion of 1/50 (E/S) for 20 h, at 37 °C.

Quinoa bioester (QB) treatment. QB was diluted in
Desalting and sample quantification. In due course, the enzymatic reaction was stopped by adding trifluoroacetic (0.4% v/v final) and the peptides were incubated for additional 40 min to degraded the RapiGest. Afterward, the samples were centrifuged at 18.000×g for 10 min to remove any insoluble materials. Subsequently, the peptides were quantified using the fluorometric assay-Qubit 2.0 (Invitrogen) according to the manufacturer's recommendations. Each sample was desalted and concentrated using Stage-Tips (STop and Go-Extraction TIPs) according to Rappsillber and collaborator 65 . Table 1. Top enriched pathways per cluster. The "Node" column presents Q1-Q5 and F1-F5 for clusters of protein profiles obtained from the quinoa bioester treated and from the Quinoa Free cell lines, respectively. The "Pathway Identifier" and "Pathway Name" columns refer to the corresponding entries per Reactome. The "Entities Found and "Entities Total" is the number of proteins identified in this study belonging to the respective pathway and the total number of proteins cataloged for that pathway, as per Reactome, respectively. Finally, the "Consensus" column represents the consensus protein profile of all identified proteins belonging to the respective cluster; the y-axis representing relative abundancy and the x-axis related to the different time points.

Node
Pathway identifier Pathway name #Entities found #Entities total Consensus Peptide spectrum matching (PSM). The data analysis was performed with the PatternLab for proteomics 4 software that is freely available at https ://www.patte rnlab forpr oteom ics.org 67 . Homo sapiens' sequences were downloaded on June 6th, 2020 from the Swiss-Prot and then a target-decoy database was generated to include a reversed version of each sequence plus those from 104 common mass spectrometry contaminants. The Comet 2019.01 rev. 5 search engine was used for identifying the mass spectra 68 . The search parameters considered: fully and semi-tryptic peptide candidates with masses between 550 and 5500 Da, up to two missed cleavages, 40 ppm for precursor mass, and bins of 1.0005 m/z for MS/MS with an offset of 0.4. The modifications were carbamidomethylation of cysteine and oxidation of methionine as fixed and variable, respectively.
Validation of PSMs. The validity of the PSMs was assessed using Search Engine Processor (SEPro) 69 . The identifications were grouped by charge state (2 + and ≥ 3 +), and then by tryptic status, resulting in four distinct subgroups. For each group, the XCorr, DeltaCN, DeltaPPM, and Peaks Matches values were used to generate a Bayesian discriminator. The identifications were sorted in nondecreasing order according to the discriminator score. A cutoff score accepted a false-discovery rate (FDR) of 2% at the peptide level based on the number of decoys 70 . This procedure was independently performed on each data subset, resulting in an FDR independent of charge state or tryptic status. Additionally, a minimum sequence length of five amino-acid residues and a protein score greater than 3 were imposed. Finally, identifications deviating by more than 10 ppm from the theoretical mass were discarded. This last filter led to FDRs, now at the protein level, to be lower than 1% for all search results 71 .
Proteomic data analysis. We quantitated, independently, three biological replicates for each of our sixtime points (i.e., T0h, T3h, T6h, T12h, T24h, T48h), with two technical replicates. Quantitation was performed according to PatternLab's Normalized Ion Abundance Factors (NIAF) as a relative quantitation strategy and as described in our bioinformatics protocol 67 . We recall that NIAF is the equivalent to NSAF 72 , but applied to extracted ion chromatogram (XIC). Differentially abundant proteins were listed by using PatternLab's TFold module to compare time point zero with the other time points 73 . We also performed a TFold analysis comparing the two batches of biological replicates acquired for timepoint 0 h to serve as a quality control step; we expected www.nature.com/scientificreports/ to find no differentially abundant proteins as all 3D cell cultures originated from the same biological condition. PatternLab's TrendQuest module was also employed to group proteins that share the same temporal abundancy patterns over the time-course experiment 74 . Finally, we used the Reactome 75 tools to help interpret the data.

Conclusion.
Here, we pinpointed proteomic alterations that 3D keratinocyte cell cultures undergo when exposed (or not) to QB at several timepoints. Our results shortlisted up-regulated proteins that are known to be beneficial for skin replenishing. We opted for performing our work on 3D cell cultures as they have been described to better mimic in vivo as when compared to 2D cell cultures 76 , thus our results suggest that the application of QB could be beneficial to human skin; nevertheless, in vivo studies should be performed to validate such hypothesis.

Data availability
The mass spectrometry data have been deposited to the ProteomeXchange Consortium via the PRIDE 77 partner repository with the dataset identifier PXD020893. www.nature.com/scientificreports/