Spatial distribution of glucose and amino acids within all-aqueous emulsions directs the Maillard reaction and oxidation pathways

Spatial distribution of glucose and amino acids within all-aqueous emulsions directs the Maillard reaction and oxidation pathways | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Spatial distribution of glucose and amino acids within all-aqueous emulsions directs the Maillard reaction and oxidation pathways Antonio Dario Troise, Kangni Chen, Ashkan Madadlou, Sabrina De Pascale, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7307960/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 21 Mar, 2026 Read the published version in Communications Chemistry → Version 1 posted You are reading this latest preprint version Abstract All-aqueous emulsions represent a versatile platform for studying and controlling chemical reactions in foods and biological systems. Through the compartmentalization and segregation of reactants, distinct molecularly crowded microenvironments enable the generation of unique reaction products. We explored how spatial organization within all-aqueous emulsions composed of polyethylene glycol (PEG) and sodium sulfate (Na 2 SO 4 ) modulates the Maillard reaction and oxidation reactions between glucose and amino acids. Using untargeted metabolomics and molecular networking, we characterized the chemical diversity of reaction products formed when the reactants were either co-encapsulated within the droplet phase or distributed (segregated) between the two phases of the emulsions. Over 500 compounds were annotated across both systems, revealing distinct molecular profiles driven by reactant localization and phase partitioning. When the precursors were segregated (tryptophan and glucose), oxidation products such as β-carbolines and N -formylkynurenine accumulated preferentially in the PEG phase. Conversely, when the reactants were co-encapsulated (asparagine and glucose) within Na 2 SO 4 droplets, enhanced formation of the Amadori products and dipeptides was observed, guided by phase-specific microenvironment. Our results demonstrate that the reactant location, in addition to time and temperature, plays a critical role in modulating food-relevant reactions, with a new framework for controlling the formation of glycation compounds via emulsion-based microreactors. Physical sciences/Chemistry/Biochemistry/Metabolomics Physical sciences/Chemistry/Organic chemistry/Reaction mechanisms Maillard reaction Oxidation Water-in-water emulsions Metabolomics Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 1. Introduction Droplet reactors, particularly those based on all-aqueous (water-in-water, W/W) emulsions, have emerged as versatile platforms for carrying out chemical reactions within molecularly crowded biphasic systems 1 . The physicochemical properties of all-aqueous emulsions influence the location of reactants (precursors), directing specific chemical reactions to occur either within the dispersed droplets (i.e., compartmentalized), at the interface between the two phases or in the continuous bulk phase. Chemical reactions that can be studied using such droplet-based systems include hydrolysis, as well as acid-base, redox and enzymatic reactions. Each of these is essentially influenced by the chemical nature of the reactants and by the tendency of precursors to partition between the two aqueous phases or accumulate at the interface 2 . Among the many chemical transformations that can occur in such biphasic systems, the Maillard reaction (MR) is of particular interest due to its complexity and importance in food chemistry. The MR produces a wide variety of volatile and non-volatile compounds. Volatile products include nitrogen- and oxygen-containing heterocycles (e.g., pyrazines, thiazoles, furans), as well as Strecker aldehydes (e.g., isovaleraldehyde) 3–5 . Non-volatile products range from early-stage compounds like Amadori products to melanoidins and advanced glycation end-products 6,7 . It is known that their formation depends on factors such as precursor type, temperature, pH, and water activity 8 . However, little is known about how the localization, i.e., the spatial distribution and dynamics of reactants and products within multiphase systems affect their retention, concentration, and migration of these molecules across phase boundaries. Other reactions, including Strecker degradation, retro-aldol cleavage, and oxidation may occur concurrently with the MR or interact with its intermediates contributing to this complex chemical profile 9 . Within this scenario, effective analysis of diverse compounds generated during the Maillard reaction in all-aqueous emulsions requires specialized analytical methods tailored to their dynamic microenvironment. Untargeted metabolomics provides a holistic profiling of the molecular species, enabling detailed characterization of both temporal (i.e., over time) and spatial (i.e., across compartments) variations in reaction intermediates and products. This systems-level approach is particularly valuable in deciphering not only the pathways of the MR but also concurrent processes, such as oxidation, that may occur even in simple sugar–amino acid model systems. Similar strategies have been applied in studying the peroxidation of lipids and volatile compounds formation during food processing 10,11 . In the context of untargeted metabolomics, the application of hydrophilic interaction liquid chromatography high-resolution mass spectrometry (HILIC-HRMS) combined with droplet-based reactors, especially all-aqueous emulsions, for decoding the MR pathways remains the technique of choice to match the hydrophilic nature of the most relevant MR precursors. Our prior work demonstrated that all-aqueous emulsions can control the spatial distribution of amino acids, reducing sugars, and the corresponding reaction intermediates, thereby directing MR pathways 12 . Although that study established the role of the emulsions in modulating early-stage reactions through compartmentalization and segregation, two critical unanswered questions persist: (I) how the spatial organization tackles the molecular diversity of end-products, and (II) how their formation and migration through different phases depends on amino acid partitioning behavior. Here, untargeted metabolomics was used to annotate and identify the end-products formed and elucidate the MR and oxidation pathways in all-aqueous emulsions through multivariate data analysis. All-aqueous emulsions were prepared, and the reactants were partitioned either inside the droplets (co-encapsulation, for asparagine and glucose) or segregated between the interior and exterior phases of the emulsions (yet interfacially touching, for tryptophan and glucose) according to the amino acid side chain 12 . HILIC-HRMS analysis coupled to molecular networking enabled comprehensive profiling and pathway elucidation. Our findings offer new insights into the role of compartmentalization in steering chemical reactions and present a novel platform for their control under food-relevant conditions. 2. Material and methods 2.1 Chemicals and reagents Acetonitrile and water were purchased from Merck-Sigma-Aldrich (Darmstadt, Germany); D-glucose, L-tryptophan, L-asparagine, polyethylene glycol (PEG) 8 kDa, sodium sulfate (Na 2 SO 4 ), phosphate buffer powder, ammonium formate and formic acid were purchased from Merck-Sigma-Aldrich. N -(1-deoxy-D-fructos-1-yl)-L-tryptophan (Fru-Trp) and N -(1-deoxy-D-fructos-1-yl)-L-asparagine (Fru-Asn) were purchased from Toronto Research Chemicals (Toronto, Canada). All chemicals used in this study were of analytical grade, except for liquid chromatographic solvents, water and acetonitrile which were of mass spectrometry grade. 2.2 Preparation of all-aqueous droplet reactors All-aqueous droplet reactors were prepared according to the method described in our prior report 12 . In short, aqueous stock solutions of Na 2 SO 4 (15 wt%) and PEG (50 wt%) were prepared by the addition of either Na 2 SO 4 or PEG to phosphate buffer stock solution (0.1 M, pH 7.4). All-aqueous droplet reactors (Na 2 SO 4 /PEG) were prepared by mixing the Na 2 SO 4 and PEG stock solutions at a volume ratio of 1:4, with stirring at 1000 rpm for 1 h. Subsequently, the droplet reactors were supplemented with the Maillard reactants (glucose and either asparagine or tryptophan), achieving a final concentration of 40 mM. The mixtures were stirred overnight at 25°C for complete dissolution of the reactants. After phase separation, PEG phase (EP) and Na 2 SO 4 phase (ES) were collected from samples. For comparison purposes, in addition to the droplet reactor samples (two-phase systems), single-phase solutions were also prepared. These included mixtures of phosphate buffer solution and Na 2 SO 4 or PEG stock solutions (referred to as "Na 2 SO 4 " and "PEG", respectively). All these solutions were supplemented with the same reactants. 2.3. Running the MR in droplet reactors and single-phase solutions The reactants-supplemented emulsions (and single-phase solutions, 5 mL) were transferred into gas chromatography vials (20 mL), and securely sealed. Subsequently, these vials were subjected to heating at 95°C for 0 h, 3 h, and 5 h while being stirred at 1000 RPM. Next, the samples were rapidly cooled within an ice bath, then kept at room temperature (25°C) for 4 h. The top and bottom phases were then subjected to diafiltration (molecular weight cut-off 3000 Da) at 14,000 × g, repeated five times, with each cycle lasting 30 minutes. Both the retentate and permeate fractions were collected and subsequently freeze-dried (Alpha 2–4 LD plus, Martin Christ, Osterode am Harz, Germany) for further analysis. For comparison purposes, the single-phase samples supplemented with the reactants were also stored at 25°C for 4 h and subjected to diafiltration. 2.4 Liquid chromatography high resolution mass spectrometry The formation of reaction products of asparagine and tryptophan was screened in untargeted mode in emulsion systems after phase separation by using PEG solution and Na₂SO₄ solution as controls. Freeze-dried permeates were redissolved in water and both the upper and lower phases were diluted in 50% aqueous acetonitrile then centrifuged (18,000 x g, 4°C, 10 min). LC-MS/MS data were acquired using an Exploris 120 quadrupole Orbitrap high-resolution mass spectrometer interfaced to a Vanquish Core liquid chromatographic system (Thermo Fisher Scientific, Bremen, Germany). Compounds were separated at 40°C through a zwitterionic sulfobetaine column (Atlantis Premier BEH, Z-HILIC, 100 x 2.1, 1.7 µm, Waters) with the following gradient of solvent B (minutes/%B): (0/5), (1.5/5), (10/50), (13/50). Mobile phases consisted of 0.1% formic acid in acetonitrile (v/v, solvent A) and 0.1% formic acid in water (v/v, solvent B) and the flow rate was 0.3 mL/min. Samples were acquired in profile data dependent scanning (ddMS 2 ) mode by using polarity switching. For positive ion (and negative ion) mode, H-ESI interface parameters were as follows: static spray voltage 3.3 kV (-2.9 kV), while ion transfer tube and vaporizer temperature were both at 280°C; sheath gas flow and auxiliary gas flow were 38 and 8 arbitrary units, respectively, for both positive and negative ion mode. Along with full scan acquisition in the m/z range 60–900 (resolution 60,000 at m/z 200 FWHM), ddMS 2 were performed on the top 4 scans by using an intensity threshold of 60,000 (area counts), a customized dynamic exclusion window of 3.5 s and a mass tolerance of 5 ppm. A targeted mass exclusion list was incorporated within the method upon the analysis of two separated full scan acquisitions of blank samples and the removal of the top 600 most intense background signals, for both positive and negative ion mode. Finally, ddMS 2 scan properties included an isolation window of m/z 1.5, a normalized higher energy collisional dissociation (HCD) of 20%, 40% and 80% and the Orbitrap resolution fixed at 30,000. Data were acquired in profile mode by using a standardized AGC (all gain control) target and EASY-IC® (fluoranthene as internal calibrant) at the beginning of each run. Identification of precursors, putative intermediates and end-products was achieved by injecting quality control samples at regular intervals of 6 runs and by analyzing reference mixes through differential scanning ranges ( m/z 60–200; m/z 190–400, m/z 390–600 and m/z 590–900) in both positive and negative mode by using the same chromatographic layout as described above. 2.5 Untargeted metabolomics An untargeted metabolomic workflow based on end-products identification was used to find and characterize differences between the buffer control samples and fully aqueous emulsions in the presence of asparagine and glucose or tryptophan and glucose by importing raw files in Compound Discoverer software (v. 3.3 Thermo Fisher Scientific, San José, CA). The procedure involved the retention time alignment and detection of expected and unknown compounds for samples grouping. Upon definition of the elemental composition, exact masses, chemical formulas and fragmentation spectra, compounds were matched with analytical standards as glucose, tryptophan, asparagine and their respective Amadori compounds. The procedure encompassed the generation of five separated libraries including end-products with a similar fragmentation pattern. This strategy generated molecular networks that were combined with analytical information reported in publicly available databases, as mzCloud ( www.mzcloud.org ) and ChemSpider ( www.chemspider.com ). A supplementary search was performed in Reaxys® (Elsevier, Amsterdam, the Netherlands) by using as reactants the two amino acids and glucose, with temperature higher than 50°C, to generate an internal mass list of putative chemical structures arising from the model systems. After correction for quality control samples, post-processing nodes performed descriptive statistics and differential analysis working on two technical replicates arising from two replicates representative of four observations for each time point/condition. To further investigate significative differences between samples, hypothesis test was performed by one-way ANOVA model with Tukey as post-hoc test. For volcano plots and ratio among sample groups, p -values were adjusted by Benjamini-Hochberg algorithm. Tandem MS spectra, principal component analysis (PCA), histograms, scatter plot, loadings plots, molecular network 13 , statistical test, and log fold changes were obtained in Compound Discoverer. Reaction pathways and chemical structures were built in ChemDraw (Revvity Signals Software, Waltham, MA). 3. Results and Discussion 3.1 The analytical background for the definition of the chemical space behind reaction mechanisms in all-aqueous emulsions The spatial distribution of reaction precursors including amino groups and reducing sugars within all-aqueous emulsions leads to the formation of thousands of compounds with distinct properties. We opted for zwitterionic HILIC as a flexible technique to appropriately separate polar charged compounds present in an aqueous environment. This method also provided a good response for the separation of chromophores and brown pigments or polymers (usually separated in reversed phase chromatography) formed in both tryptophan- and asparagine-supplemented samples as a result of condensation, oxidation, and polymerization reactions 14 . Then, high resolution tandem mass spectrometry-based metabolomics provided the analytical context for the annotation and identification of unknown end-products formed 10 . Control samples and procedural blanks were used to define the chemical space in the untargeted analysis: in full scan MS1 mode, a total of 524 signals were detected in the system with segregated reactants compared to 917 signals in the system where reactants were co-encapsulated. To improve the reliability of chemical feature responses, several filters were used starting with the removal of signals that were not associated with any chemical formulas due to poor matching with theoretical isotopic pattern distribution. Next, we considered only compounds in tandem mode that generated a fragmentation mass spectrum and added structural information to the isotopic pattern. The third filter included a chromatographic peak score higher than 5 as the result of the profiles associated with the scan points below each peak in all the replicates. Finally, the metabolomic workflow included two separate procedures: each run was screened first in polarity switching mode by considering both ionization modalities. Annotated compounds in polarity switching experiments were further checked in positive or negative ion mode. Wherever possible, spectra were manually curated in both positive and negative mode by excluding co-eluting compounds with overlapping peak shape that could result from in-source ionization or fragmentation typical of analytical procedures that consider broad dynamic ranges 15 . Upon all the filtering procedures, we annotated 265 compounds in all the tryptophan-supplemented samples and 302 compounds in all asparagine-supplemented samples. In Supplementary Table S1 and S2 , we reported for each compound an identification level ranging from 1 (compound identification upon matching to analytical standards as the amino acids and glucose as precursors and the two Amadori compounds as intermediate, reported in bold) to 2 (compounds annotation based on mass spectra, isotopic pattern, chemical formula, mass accuracy and matching with databases when compound names are present) according to Metabolomics standard initiative (MSI) level 16 . In this view, we decided to keep tryptophan-supplemented and asparagine-supplemented samples separated, and study the type of compounds formed, then explore the overall reactivity 12 . 3.2 Tryptophan and glucose reactivity in reactants segregation. Figure 1 provides a sample distribution overview of tryptophan-supplemented samples at time 0 and at 5 h in the form of a PCA, accompanied by a loadings plot and a molecular network of the annotated compounds. Figure 1 A shows that the overall variance explained was 47.8% with the first component accounting for 27.9% and the second component for 19.9%; samples Na₂SO₄_0, ES_0, and ES_5 were in the first quadrant. It is worth mentioning that ES remained in the first quadrant even after 5 h of heating, suggesting that the key features of its major reaction products did not change significantly. This preliminary screening was in line with the results already reported by our group: after the initial reaction, part of the products migrated into the EP phase instead of undergoing continuous reaction within the ES. EP_0 and PEG_0, both containing PEG were in the second quadrant indicating marked differences toward the other two Na 2 SO 4 -containing samples. In the third quadrant, EP and PEG at 5 h were predominant, and both clustered within the same region, indicating that they generated similar reaction product profiles after heating. One possible explanation for this observation is that EP functions as the main site for subsequent transformations, resulting in similar reaction pathways and product distributions as observed in the PEG samples. This two-dimensional distribution corresponded to the molecular loadings shown in Fig. 1 B, where each grey point represents one of the 320 compounds annotated when reactants were segregated between the emulsion phases. As expected, dark gray datapoints predominantly clustered within the second and third quadrants (Q2-Q3), corresponding to compounds exhibiting significantly elevated response intensities in PEG_5, EP_5, and Na₂SO₄_5 samples. This pattern is likely attributed to the effects of prolonged thermal processing (95°C for 5 h), which promoted advanced glycation and oxidation reactions, as well as cleavage and polymerization processes. Figure 1 C presents information on the molecular interrelationships between glucose and tryptophan reaction products. The molecular network illustrates overall chemical patterns independently of model systems, with green points representing annotated compounds and blue points indicating unknown compounds. The red box highlights degradation products associated with glucose, including putative intermediates involved in dicarbonyl formation and small organic acids. In contrast, the blue box contains compounds with fragmentation spectra similar to those of tryptophan, such as indole derivatives and β-carbolines. These two compartments were connected by N -(1-deoxy-D-fructos-1-yl)-tryptophan (Fru-tryptophan), which exhibits fragmentation patterns consistent with those of glucose, showing consecutive losses of water from secondary alcohol groups, as well as with tryptophan, characterized by indole ring formation and diagnostic ions at m/z 146.0601 and 188.0706. The molecular network and the observed spectral similarities reflected the theoretical grouping of different MR pathways, as previously hypothesized by Yaylayan 17 . In particular, the Amadori compounds appear to serve as a key bridging point between amino acids and glucose, and they provide a basis for studying the spatial organization of the resulting degradation products. To gain insight into the qualitative and quantitative relationships among the four model systems (two buffer systems and the emulsion systems after phase separation), we employed a supervised discriminant analysis based on fold change in logarithmic scale and post-hoc test. The results were summarized using a composite volcano plot of the emulsion system, comparing the two phases (PEG and Na₂SO₄) at 5 h (Fig. 2 ). In the green region, light blue points highlighted 33 discriminating compounds that were over-represented in PEG_5, as identified by a separate differential analysis between PEG_5 and Na₂SO₄_5 ( Supplementary Figure S1 ). Hence, among the 60 compounds over-represented in the green region, 27 were found at significantly higher levels. We subsequently performed manual curation of fragmentation spectra for target analytes present in EP_5 (Fig. 2 with a magnified area). These analytes were highlighted using boxes of different shapes, corresponding to the reaction mechanisms illustrated in Figs. 3 and 4 . Specifically, analytes were marked with yellow, green, blue, and red circles, as well as green, red, and blue rectangles, to represent distinct mechanistic categories. Considering the evidence that Amadori compound rearrangement is likely to occur at the interface, we propose that indole oxidation may predominate when reactants were segregated between the emulsion phases. Figure 3 illustrates the oxidation pathways initiated by tryptophan, along with trend lines showing the normalized area counts of the target analytes over 5 h. We observed that oxidation products such as N -formylkynurenine and N -formylanthranilic acid were formed at significantly higher levels in the PEG phase, while their concentrations were lower in the PEG solution, indicating that the all-aqueous emulsion system plays a crucial role in the synthesis of these compounds. Conversely, N -formylkynurenine and N -formylanthranilic were detected only in trace amounts in Na 2 SO 4 solution; however, due to their specific chemical properties, their concentrations increased over time in the Na 2 SO 4 phase. Furthermore, we spotted the formation of four structural isomers peaking at m/z 361.1426 with molecular formula C21H19N3O3, different retention times, similar kinetic profiles and only one eluting at 4.8 min characterized by the predominance of typical fragmentation spectra of tryptophan (C11H10NO2, m/z 188.0705; C9H8NO, m/z 146.0600; C8H8N, m/z 118.0650, as suggested by the molecular network in Fig. 1 C and by the spectra in Supplementary Figure S2 A ). In Supplementary Figure S2 B , fragment ion search (FiSH) scoring confirmed a structural match with N -(1H-indol-3-ylacetyl)tryptophan, suggesting the formation of an amide bond between 3-indoleacetic acid and tryptophan. This condensation is likely promoted under alkaline conditions and local dehydration. Figure 4 summarizes the reaction pathways involved in the formation of condensation and polymerization products. The key intermediates include tryptamine, formed via decarboxylation of tryptophan, and indole-3-pyruvic acid, produced through oxidative deamination 18 . Both intermediates can undergo condensation, leading to the formation of polymerization products. In the presence of reactive α-dicarbonyls, a specific class of β-carbolines can also be synthesized through intramolecular cyclization 19 ; β-carbolines are characterized by different logP depending on the type of α-dicarbonyls. In Supplementary Figure S3 , we reported the distribution of two compounds formed from the interaction between reactive carbonyls and tryptophan. Although 1-carboxyethyl-β-carboline shares strong structural similarities with its propanoic acid derivative, their partitioning behavior differs significantly. Furthermore, tryptophan fragmentation products might contribute to the generation of other compounds including ethylamine and indole derivatives and can serve as indole building blocks for the formation of brown pigments and volatiles. The approach for PEG phase was extended to the study of Na 2 SO 4 phase, and Fig. 5 presents the composite volcano plot comparing ES and EP at 5 h. The 42 light blue compounds represent those over-represented in the Na 2 SO 4 phase compared to the PEG buffer phase (Supplementary Figure S1 ). As expected, none of the compounds that were overrepresented in the Na 2 SO 4 solution at 5 h maintained the same pattern in the emulsion system. All were distributed below the green and red regions, indicating very low –log₁₀ p-values and, thus, weak discriminative power. This analytical strategy enabled the identification of compounds significantly enriched in the Na 2 SO 4 phase after phase separation and facilitated the investigation of how compounds migrate from the PEG phase to the Na 2 SO 4 phase. Such compounds were highlighted using red, purple, blue, and yellow triangles. The quantitative trends shown in Fig. 5 were consistent across datasets: a pronounced and statistically significant increase in the Na 2 SO 4 phase after phase separation, with negligible signals detected in the other model systems. One exception was 3-amino-2,3-dideoxy-scyllo-inosose, a putative dehydration product of glucosamine, which was formed at comparable levels in both Na 2 SO 4 solution and Na 2 SO 4 phase (Na 2 SO 4 and ES). Here, most of the generated compounds originate from the glucose moiety, with some amino-sugar derivatives tentatively formed at the interface, while none of the indole derivatives were detected at significative levels among the over-represented molecules. 3.3 Asparagine and glucose reactivity in reactants co-encapsulation. Figure 6 provides an overview of the 2D distribution of samples and the 423 compounds annotated in when reactants were co-encapsulated. This analysis compares data from time 0 and after 5 h of thermal treatment at 95°C. The PCA shown in Fig. 6 A explains a total variance of 53.0%, with a more complex sample distribution than observed in reactants segregation. At time 0, all samples were in the third and fourth quadrants. After 5 h, the samples shifted along the second principal component (accounting for 20.9% of the variance), all exhibiting positive PC2 values. This distribution highlights key differences already present at time 0, suggesting that the preparation of the model system (including pre-incubation, mixing, and dissolution) may have induced the formation of early-stage reaction products. The first principal component (PC1) accounted for 32.1% of the total variance. This interpretation is supported by the loadings plot in Fig. 6 B, where grey data points representing annotated compounds clustered around black points along both axes, indicating a greater contribution from Na 2 SO 4 -containing systems (Na 2 SO 4 solution and ES at time 0 and 5 h). Figure 6 C depicts the interconnection between asparagine, glucose, and their respective reaction products. Again, the Amadori compound of asparagine is positioned between two molecular clusters representing amino acid degradation products and glucose degradation products, in red and blue boxes, respectively. This distribution confirmed what observed for tryptophan: Maillard reaction and the Amadori rearrangement remains the key driver in both reactants segregation and co-encapsulation in the presence of free amino group and reducing sugars. Besides the linkages between the two clusters, a closer relationship was observed between the Amadori compound of asparagine and several degradation products, most of which are small organic acids. Here, we hypothesize that the carbohydrate moiety of the Amadori compound contributes more efficiently to the generation of fragmentation products than tryptophan, likely due to reduced intramolecular reactivity of indole rings with dicarbonyl compounds. As in reactant segregation, the co-encapsulation molecular network was constructed across the four model systems without using time as a discriminating variable. To provide information on the spatial organization of reactants co-encapsulation and segregation, a composite volcano plot was developed in Fig. 7 (log₂ fold change > 1 and log₁₀ p -value > 1.5), identifying 92 compounds over-represented in EP after 5 h at 95°C. As a preliminary step, 38 compounds that were over-represented in PEG compared to Na 2 SO 4 solution at 5 h ( Supplementary Figure S4 ) were highlighted in light blue. This strategy enabled an initial screening of compounds significantly enriched or accumulated in the PEG phase after phase separation. Among the 16 compounds over-represented in EP (green region in Fig. 7 ), four annotated compounds were further highlighted using red, blue, green, and yellow circles. Three compounds exhibited similar fragmentation spectra characterized by an intense signal at m/z 96.96. This conserved and prominent peak is indicative of a sulfate ion fragment. We hypothesize that reactions involving sulfate moieties occur in the Na 2 SO 4 phase, with subsequent cyclization products migrating into the PEG phase. In addition, 1H-pyrrole-2-carboxamide serves as a key example of a pyrrole derivative that moves between the two regions, exhibiting accumulation in the PEG phase. In parallel, Fig. 8 presents the same composite volcano plot shown in Fig. 7 , with light blue-labeled points derived from the discriminant analysis between Na 2 SO 4 and PEG solutions at 5 h and pinpointed in the red region. According to this layout, we identified the compounds that accumulated in the Na 2 SO 4 phase after phase separation, excluding the 29 compounds over-represented in Na 2 SO 4 based on the PEG vs. Na 2 SO 4 discriminant analysis (Supplementary Figure S4). Figures 8 A and 8 B illustrate two examples of pyridine derivatives that accumulated significantly in the Na 2 SO 4 phase, while remaining at very low concentrations in both PEG and EP. Figure 8 C summarizes the time-dependent trend of 3-deoxyglucosone. Although 3-deoxyglucosone is hydrophilic, its normalized area counts were comparable between the ES and EP. Notably, an increasing trend over time was observed only in the Na 2 SO 4 phase after phase separation, suggesting its distribution among both phases. These results suggest that compounds formed in the Na 2 SO 4 phase may continue to migrate between phases, influenced by reaction conditions such as time and temperature. In addition to asparagine degradation products, we identified other markers of asparagine oxidation: aspartyl-asparagine and asparaginyl-aspartic acid, two dipeptide isomers that eluted at different retention times but displayed identical fragmentation spectra ( Supplementary Figure S5A ). In Supplementary Figure S5B , we reported the MS/MS spectrum with FiSH scoring for the putative chemical compound including the formation of a peptide bond between aspartic acid and asparagine. Both isomers may originate from the initial oxidation of asparagine to aspartic acid, followed by further peptide bond formation and oxidation. Despite amine acylation in water is a common reaction in the presence of dedicated enzymes, this process occurs rarely in the absence of catalysts; this is because the high concentration of water shifts the equilibrium toward hydrolysis. In the present work, we observed the formation of the peptide bond probably as result of the alteration of the local water concentration due the different reagent partitioning or the occurrence of a base-catalyzed process at the hydrophobic interfacial compartment that promoted the formation of the necessary tetrahedral reaction intermediate 20 . 3.4 Reactants co-encapsulation and segregation introduce the space as a control parameter for chemical reactions. All-aqueous emulsions influence the spatial distribution of reactants based on the intrinsic chemical properties of the precursors, thereby affecting overall chemical reactivity and contributing to system stabilization 21 . The dichotomy between co-encapsulation and segregation, as specific modes of compartmentalized reactions, enabled the spatial separation of glucose, asparagine, and tryptophan within the two immiscible aqueous phases of the all-aqueous emulsions studied here. Comparing their separation across the phases with their coexistence within a single phase allowed for the controlled execution of sequential transformations characteristic of the Maillard cascade, including the formation of the Amadori compounds and related oxidation processes 12 . For the first time, we demonstrated that melanoidin building blocks can migrate from one phase to another, representing an extreme case of on-droplet formation. The selective partitioning of the two amino acids into specific phases or at the interface increased their local concentrations at the reaction sites, thereby enhancing reaction rates in accordance with the principles of mass action. Through our HILIC-HRMS and multivariate data analysis protocol, we highlighted how precursors with distinct physicochemical properties behave differently in all-aqueous emulsions and underscores the potential of metabolomics to elucidate the complex interplay among time, temperature, reactant localization, and the final distribution of end-products within the emulsion system. As previously reported by our group, the partition coefficients of amino acids regulate the localization of side chains in all-aqueous emulsions and influence the formation of reaction end-products 12 . We observed that the localization of precursors, whether coexisting within droplets as in the case of asparagine and glucose or spatially separated at the interface as with tryptophan and glucose, can lead to the formation of distinct and unique reaction products because of the intrinsic chemical nature of the precursors (reactants location) and the subsequent distribution of the resulting products within all-aqueous emulsions. Specifically, tryptophan and glucose first reacted at the interface to form the corresponding glycosylamine. This intermediate subsequently migrated into the PEG phase. In addition to participating in the Maillard reaction, tryptophan was also subject to oxidation. In this context, it has been reported that tryptophan can form 3-hydroperoxytryptophan and a dioxoethane intermediate through oxidation pathways, both of which are proposed precursors to N -formylanthranilic acid 18 . Furthermore, tryptophan may undergo oxidative deamination to yield indole-3-pyruvic acid and decarboxylation to produce tryptamine, as here reported for reactant segregation mode. The condensation of these two compounds can lead to the formation of 1-(methylenindole)-tetrahydro-β-carboline-3-carboxylic acid, a β-carboline derivative. Notably, these transformation pathways are typically activated under high-temperature or in anaerobic conditions (up to 140°C), as described by Bellmaine, Schnellbaecher and Zimmer 18 . Maillard reaction intermediates and advanced products, such as carbonyl compounds, pyridines, quinones, and pyrroles, often carry reactive functional groups including α-dicarbonyls, aldehydes, ketones, and quinones 22 . These groups possess oxidative properties and can generate free radicals, contributing to the chemical complexity of the system with propagation reactions, consistent with previous observations in oil-in-water emulsions 23 . Indeed, the co-localization of these MR products and tryptophan within the PEG phase likely promoted the oxidative transformation of tryptophan, either by reaction with α-dicarbonyls or through the generation of reactive oxygen species, which can also be formed from the degradation of PEG. These processes may have led to modifications in the overall supramolecular arrangement of the all-aqueous system. The results suggest that reactants segregation not only directs initial Maillard reaction steps but also modulates the fate of amino acid-derived intermediates through selective partitioning and the resulting localized reactivity. The reaction between glucose and asparagine in all-aqueous emulsion is influenced when reactants were segregated between the emulsion phases. The co-encapsulation of the reaction precursors within the Na₂SO₄ phase enabled the Maillard reaction to proceed in a compartmentalized manner. Specifically, glycosylamines were significantly partitioned into the Na₂SO₄ phase and subsequently underwent the Amadori rearrangement. The resulting Amadori products then degraded to form reactive α-dicarbonyl compounds, such as 3-deoxyglucosone 24 . We hypothesize that under these conditions, glucose may have undergone rearrangement, partial deoxygenation, and subsequent nucleophilic addition by amines, potentially leading to the formation of a stable six-membered ring compound, putatively identified as 3-amino-2,3-dideoxy-scyllo-inosose. Asparagine is known to undergo non-enzymatic deamidation under physiological or elevated temperature conditions. This process involves the formation of a cyclic imide intermediate, which is unstable and subsequently hydrolyzed to yield a mixture of L-aspartic acid (Asp) and L-isoaspartic acid. Here, we propose that Asp and Asn may undergo a condensation reaction to form dipeptides such as Asn–Asp or Asp–Asn. However, peptide bond formation from free amino acids in aqueous solution is thermodynamically unfavorable due to the high-water activity and the lack of catalyzing agents 25 . Nevertheless, under our experimental conditions, factors such as reduced water activity, selective partitioning, alkaline pH, and local microenvironments with limited hydration may have helped in shifting the equilibrium toward peptide bond formation. It is therefore possible that the high salt concentration in the Na₂SO₄ phase created a dehydrating environment that facilitated this reaction 26 . Taken together, the co-encapsulation of glucose and asparagine within the Na₂SO₄ phase created a microenvironment with reduced water activity and elevated local reactant concentrations, thereby facilitating condensation reactions such as the Amadori rearrangement and possibly peptide bonds formation in the absence of conventional catalyzing agents. Of note, a similar pathway was observed also in the case of segregation mode with the formation of N -(1H-indol-3-ylacetyl)tryptophan, as a reaction product between 3-indoleacetic acid and tryptophan. In both segregation and co-encapsulation cases, the use of labelled precursors, advanced techniques as NMR and testing the system in absence of glucose can provide further structural confirmation to those observed here by tandem mass spectrometry. 4. Conclusions Overall, our findings highlight how all-aqueous emulsions can drive distinct chemical trajectories for amino acids and sugars, with broad implications across various reaction types. The reaction intermediates and products generated within these systems can, in turn, influence other processes, as shown here for oxidation and peptide bonds formation. Indeed, the selective partitioning of reactants and intermediates into PEG or Na₂SO₄ phases not only directs canonical glycation reactions but also fosters oxidation and condensation transformations that are otherwise suppressed in homogeneous aqueous systems. These findings provide new insight into how microenvironmental factors, such as water activity and reactant concentration, can be harnessed to modulate reaction pathways and product profiles in complex food matrices. Finally, we finally demonstrated that spatial compartmentalization, beyond traditional parameters like time, pH, viscosity, and temperature, is a critical factor in governing all the above-mentioned reactions. Declarations Competing interests The authors declare no competing financial interest. Author contributions K.C. prepared the samples, carried out the experiments, curated the data, and wrote the original draft. S.D.P. and A.D.T. performed the HILIC-HRMS analysis and data interpretation. All co-authors jointly discussed the results. A.M., A.S., and V.F. contributed to manuscript revision and supervision. Acknowledgments This study was financially supported by the Food Quality and Design group in Wageningen University and Research and China Scholarship Council. A part of this study was funded by the National Recovery and Resilience Plan, mission 4, component 2, investment 1.3, call n. 341/2022 of Italian Ministry of University and Research funded by the European Union - NextGenerationEU for the project “ON Foods-Research and innovation network on food and nutrition Sustainability, Safety and Security-Working ON Foods”, project PE00000003, concession decree n. 1550/2022, CUP B83C22004790001. References Ruiz-Lopez, M. F., Francisco, J. S., Martins-Costa, M. T. C. & Anglada, J. M. Molecular reactions at aqueous interfaces. Nature Reviews Chemistry 4 , 459-475 (2020). Madadlou, A., Saggiomo, V., Schroen, K. & Fogliano, V. All-aqueous emulsions as miniaturized chemical reactors in the food and bioprocess technology. Current Opinion in Food Science 33 , 165-172 (2020). Shakoor, A., Zhang, C., Xie, J. & Yang, X. Maillard reaction chemistry in formation of critical intermediates and flavour compounds and their antioxidant properties. Food Chemistry 393 , 133416 (2022). Gao, Y., Miao, J. & Lai, K. Study on Maillard reaction mechanism by quantum chemistry calculation. Journal of molecular modeling 29 , 81 (2023). Mottram, D. S. in Flavours and Fragrances: Chemistry, Bioprocessing and Sustainability (ed Ralf Günter Berger) 269-283 (Springer Berlin Heidelberg, 2007). Murata, M. Browning and pigmentation in food through the Maillard reaction. Glycoconj J 38 , 283-292 (2021). Bork, L. V., Haase, P. T., Rohn, S. & Kanzler, C. Structural characterization of polar melanoidins deriving from Maillard reaction intermediates - A model approach. Food Chem 395 , 133592 (2022). van Boekel, M. A. Kinetic aspects of the Maillard reaction: a critical review. Nahrung 45 , 150-159 (2001). Hellwig, M. & Henle, T. Baking, ageing, diabetes: a short history of the Maillard reaction. Angewandte Chemie International Edition 53 , 10316-10329 (2014). Weidner, L., Cannas, J. V., Rychlik, M. & Schmitt-Kopplin, P. Molecular characterization of cooking processes: A metabolomics decoding of vaporous emissions for food markers and thermal reaction indicators. Journal of Agricultural and Food Chemistry 71 , 17442-17454 (2023). Zhou, Z. et al. Unraveling the Thermal Oxidation Products and Peroxidation Mechanisms of Different Chemical Structures of Lipids: An Example of Molecules Containing Oleic Acid. Journal of Agricultural and Food Chemistry 70 , 16410-16423 (2022). Chen, K. et al. Compartmentalization vs. segregation of reactants: Accomplishment of the Maillard reaction at the water-water interface. Food Chemistry 465 , 142089 (2025). Schmid, R. et al. Ion identity molecular networking for mass spectrometry-based metabolomics in the GNPS environment. Nature communications 12 , 3832 (2021). Yan, Y., Hemmler, D. & Schmitt-Kopplin, P. Discovery of glycation products: Unraveling the unknown glycation space using a mass spectral library from in vitro model systems. Anal Chem 96 , 3569-3577 (2024). Kaufmann, A. High-resolution mass spectrometry for bioanalytical applications: Is this the new gold standard? Journal of Mass Spectrometry 55 , e4533 (2020). Sumner, L. W. et al. Proposed minimum reporting standards for chemical analysis: chemical analysis working group (CAWG) metabolomics standards initiative (MSI). Metabolomics 3 , 211-221 (2007). Yaylayan, V. A. Classification of the Maillard reaction: A conceptual approach. Trends in Food Science & Technology 8 , 13-18 (1997). Bellmaine, S., Schnellbaecher, A. & Zimmer, A. Reactivity and degradation products of tryptophan in solution and proteins. Free Radical Biology and Medicine 160 , 696-718 (2020). Herraiz, T., Peña, A., Mateo, H., Herraiz, M. & Salgado, A. Formation, characterization, and occurrence of β-carboline alkaloids derived from α-dicarbonyl compounds and L-tryptophan. Journal of Agricultural and Food Chemistry 70 , 9143-9153 (2022). David, R., Tuñón, I. & Laage, D. Competing reaction mechanisms of peptide bond formation in water revealed by deep potential molecular dynamics and path sampling. Journal of the American Chemical Society 146 , 14213-14224 (2024). Pavlovic, M., Plucinski, A., Zeininger, L. & Schmidt, B. V. K. J. Temperature sensitive water-in-water emulsions. Chemical Communications 56 , 6814-6817 (2020). Rizzi, G. P. Formation of strecker aldehydes from polyphenol-derived quinones and alpha-amino acids in a nonenzymic model system. J Agric Food Chem 54 , 1893-1897 (2006). Shi, Y. et al. The antioxidant mechanism of Maillard reaction products in oil-in-water emulsion system. Food Hydrocolloids 87 , 582-592 (2019). Gobert, J. & Glomb, M. A. Degradation of glucose: reinvestigation of reactive α-dicarbonyl compounds. Journal of Agricultural and Food Chemistry 57 , 8591-8597 (2009). Sauer, F. et al. From amino acid mixtures to peptides in liquid sulphur dioxide on early Earth. Nat Commun 12 , 7182 (2021). Schwendinger, M. G. & Rode, B. M. Possible Role of Copper and Sodium Chloride in Prebiotic Evolution of Peptides. Analytical Sciences 5 , 411-414 (1989). Additional Declarations There is NO Competing Interest. Supplementary Files Supplementarytable1.xlsx Supplementary table 1 Supplementarytable2.xlsx Supplementary table 2 supportinginformationCommChem29072025.docx Supporting information file Cite Share Download PDF Status: Published Journal Publication published 21 Mar, 2026 Read the published version in Communications Chemistry → Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7307960","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":498996557,"identity":"657103e9-24fc-4180-a07f-02385eb3104b","order_by":0,"name":"Antonio Dario Troise","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA/ElEQVRIiWNgGAWjYNACAwbGBhACATYGBsYHUAYOwIysJQGsktkArAWnHmYwCbUiAWKRBMw+bIB/dv/BxwUFNrL9sw83PmD8cU+Oj//wsYqfO+4x8Mk3YNUicecws/EMgzTjGecSmw0YEoqN2STS0m72ninG7bAbyWzSPAaHExvOMLZJMCQkJLZJ8JjdZmxLwKlFHqZlPlwL//lvxfi0GMC0bIBrYchhY8anxfBGsrExD9AvG88wNhskpCWA/GIs2duWwMPGloBVi9yNxIePef7YyM47w/7wwQebBDn5/sMPP/xsAzKaD+DwPzJANpaHCPWjYBSMglEwCnAAAN4bUcGfYsaeAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0001-7635-5244","institution":"National Research Council - CNR","correspondingAuthor":true,"prefix":"","firstName":"Antonio","middleName":"Dario","lastName":"Troise","suffix":""},{"id":498996558,"identity":"43607bdc-c947-4807-a2d6-1dfb2c70bf63","order_by":1,"name":"Kangni Chen","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Kangni","middleName":"","lastName":"Chen","suffix":""},{"id":498996559,"identity":"e188de53-495e-4d61-b8e1-3a8852e6f189","order_by":2,"name":"Ashkan Madadlou","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Ashkan","middleName":"","lastName":"Madadlou","suffix":""},{"id":498996560,"identity":"b1eadcd7-890d-420d-8f62-6ac7624dc6f8","order_by":3,"name":"Sabrina De Pascale","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Sabrina","middleName":"","lastName":"De Pascale","suffix":""},{"id":498996561,"identity":"aa75d8fc-c5c1-49f4-beb2-403f5e45aa71","order_by":4,"name":"Andrea Scaloni","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Andrea","middleName":"","lastName":"Scaloni","suffix":""},{"id":498996562,"identity":"d5fbf23c-5ff3-48a1-bdb7-c2af1af99d0e","order_by":5,"name":"Vincenzo Fogliano","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Vincenzo","middleName":"","lastName":"Fogliano","suffix":""}],"badges":[],"createdAt":"2025-08-06 09:20:43","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7307960/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7307960/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s42004-026-01951-6","type":"published","date":"2026-03-21T04:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":89411838,"identity":"364fa130-dd45-4c28-ad40-044cd0669f4f","added_by":"auto","created_at":"2025-08-19 16:23:32","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":231577,"visible":true,"origin":"","legend":"\u003cp\u003e2D sample distribution, compound loadings plot, and molecular networking upon filtering procedure of ddMS\u003csup\u003e2\u003c/sup\u003e experiments for reactants segregation reaction mode. Panel A shows a PCA of the four model systems at time 0 and after 5 h at 95 °C. Colors indicate different time points and systems: EP (Emulsion PEG phase), ES (Emulsion Na₂SO₄ phase), PEG (PEG solution), and Na₂SO₄ (Na₂SO₄ solution). Panel B displays the compound loadings plot, with dark grey indicating compound accumulation after 5 h of thermal treatment. Panel C represents the chemical space, including glucose and tryptophan degradation products, linked via the Amadori compound of tryptophan (indicated as Fru-tryptophan). Annotated molecules are shown in green according to their level of identifications based on publicly available databases or internal mass list, unknowns in blue. Yellow lines connect compounds with similar fragmentation patterns in ddMS² according to specific chemical transformation. The network was built across all four systems without using time as a grouping variable, to reveal relationships among precursors, intermediates, and end-products.\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-7307960/v1/868a2ef4e507b9b9a8541f72.png"},{"id":89411230,"identity":"cfd06a69-df30-414e-8201-942c1f607e7e","added_by":"auto","created_at":"2025-08-19 16:15:32","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":187140,"visible":true,"origin":"","legend":"\u003cp\u003eComposite volcano plot including the log\u003csub\u003e2\u003c/sub\u003e fold change and log\u003csub\u003e10\u003c/sub\u003e p value between EP_5 vs ES_5 for reactants segregation. Green areas report the compounds significantly over-represented in EP_5. Light blue-labeled points indicate compounds significantly over-represented in PEG_5 based on the discriminant analysis between PEG_5 and Na₂SO₄_5 (Supplementary figure S1). Dark green points represent compounds significantly over-represented in EP_5, but not significantly over-represented in PEG_5 according to the same discriminant analysis. Red, blue, green, and yellow circles denote selected target compounds involved in reaction mechanisms, including direct tryptophan oxidation products, such as hydroperoxytryptophan, anthranilic acid, and formylkynurenine. Blue, green, and red rectangles mark degradation and condensation products that involve tryptamine as a key reaction intermediate.\u003c/p\u003e","description":"","filename":"CommChemF2.png","url":"https://assets-eu.researchsquare.com/files/rs-7307960/v1/d932287f66e82e7ce509f118.png"},{"id":89411232,"identity":"ba81878a-c4cc-478c-86c8-b6316938ace5","added_by":"auto","created_at":"2025-08-19 16:15:32","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":262621,"visible":true,"origin":"","legend":"\u003cp\u003eTryptophan degradation products formed upon oxidation, including the generation of 3-hydroperoxytryptophan and a dioxoethane intermediate according to Bellmaine et al.\u003csup\u003e18\u003c/sup\u003e These intermediates lead to the formation of \u003cem\u003eN\u003c/em\u003e-formylkynurenine and \u003cem\u003eN\u003c/em\u003e-formylanthranilic acid. Additionally, dioxyindolylalanine (diOia) was annotated as a precursor for polymerization product formation. Panels A–I display scatter plots of the normalized signal intensities of selected target analytes, obtained from full-scan experiments following annotation and identification via tandem MS spectra. Molecular formulas were used in the case of annotation level. For structural isomers, the retention time was combined with the name or with the molecural formula.\u003c/p\u003e","description":"","filename":"Figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-7307960/v1/ccd5206a9ddef069a616119f.png"},{"id":89411236,"identity":"06dcd489-9761-48a3-84ba-a325698f9843","added_by":"auto","created_at":"2025-08-19 16:15:32","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":234390,"visible":true,"origin":"","legend":"\u003cp\u003eTryptophan degradation products formed exclusively after heat exposure, with tryptamine, a decarboxylation product, as the key intermediate. Indole-3-pyruvic acid is included as a building block for further condensation reactions leading to the formation of carboline intermediates. The pathway also outlines the generation of fragmentation products such as ethylamine, which may react with reducing sugars to form amino-sugar compounds. Scatter plots in panels A–D present the normalized areas of target analytes identified in full-scan experiments following annotation and tandem MS spectrum identification.\u003c/p\u003e","description":"","filename":"Figure4.png","url":"https://assets-eu.researchsquare.com/files/rs-7307960/v1/a4f5be87988addb6cfabfd4f.png"},{"id":89411840,"identity":"b7f1ff64-77d3-44cf-bfad-9e0690657905","added_by":"auto","created_at":"2025-08-19 16:23:32","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":300499,"visible":true,"origin":"","legend":"\u003cp\u003eComposite volcano plot including the log\u003csub\u003e2\u003c/sub\u003e fold change and log\u003csub\u003e10\u003c/sub\u003e p value between EP_5 vs ES_5 for reactants segregation, following the arragment reported in Figure 2 and in Supplementary Figure S1. Red areas report the compounds significantly over-represented in ES_5. Light blue-labeled points indicate compounds significantly over-represented in Na₂SO₄_5 based on the discriminant analysis between PEG_5 and Na₂SO₄_5 (Supplementary Figure S1). Dark red points represent compounds significantly over-represented in ES_5, but not significantly enriched in Na₂SO₄_5 according to the same discriminant analysis. Red, blue, purple, green, and yellow triangles mark target compounds involved in proposed reaction mechanisms, including direct glucose oxidation products and other intermediate degradation products. Scatter plots in panels A–E show the normalized areas of target analytes identified in full-scan experiments following annotation and tandem MS spectral identification.\u003c/p\u003e","description":"","filename":"Figure5.png","url":"https://assets-eu.researchsquare.com/files/rs-7307960/v1/9746484e2af021bdfb1dd5d8.png"},{"id":89411240,"identity":"ed610295-5ea1-49b0-9cd5-ea885259e80d","added_by":"auto","created_at":"2025-08-19 16:15:32","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":217952,"visible":true,"origin":"","legend":"\u003cp\u003e2D sample distribution, compound loadings plot, and molecular networking upon filtering procedure of ddMS\u003csup\u003e2\u003c/sup\u003e experiments for reactants co-encapsulation. Panel A displays a PCA of the four model systems at time 0 and after 5 hours at 95 °C. Colors indicate different time points (0 vs 5 h) and systems: EP (Emulsion PEG phase), ES (Emulsion Na₂SO₄ phase), PEG (PEG solution), and Na₂SO₄ (Na₂SO₄ solution). Panel B shows the compound loadings plot, with dark grey indicating compound accumulation after 5 h at 95 °C. Panel C represents the chemical space, including glucose, asparagine, and their putative degradation products, linked via the Amadori compound of asparagine (Fru-Asn). Identified molecules are shown in green scale according to their level of identification, unknowns in blue. Yellow lines connect compounds with similar fragmentation patterns in ddMS². The network was built across all four systems without using time as a grouping variable, to reveal relationships among precursors, intermediates, and end-products.\u003c/p\u003e","description":"","filename":"Figure6.png","url":"https://assets-eu.researchsquare.com/files/rs-7307960/v1/bbb16134245d682189bf1ddf.png"},{"id":89411841,"identity":"a37ae8f0-8ff1-4a6c-82b6-85b8723b6f83","added_by":"auto","created_at":"2025-08-19 16:23:32","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":280662,"visible":true,"origin":"","legend":"\u003cp\u003eComposite volcano plot including the log\u003csub\u003e2\u003c/sub\u003e fold change and log\u003csub\u003e10\u003c/sub\u003e p value between EP_5 vs ES_5 for reactants co-encapsulation. Green area reports the compounds significantly over-represented in EP_5. Light blue-labeled points indicate compounds significantly over-represented in PEG_5 based on the discriminant analysis between PEG_5 and Na₂SO₄_5 (Supplementary Figure S4). Dark green points represent compounds significantly over-represented in EP_5, but not significantly over-represented in PEG_5 according to the same discriminant analysis. Red, blue, green and yellow circles are representative of the target compounds part of the reaction mechanisms including direct asparagine cyclyzation and oxidation products as pyrrole derivatives. Scatter plots in panels A-D include the normalized areas of target analytes screened in full scan experiments upon annotation and identification of tandem MS spectra.\u003c/p\u003e","description":"","filename":"Figure7.png","url":"https://assets-eu.researchsquare.com/files/rs-7307960/v1/51ef18d1ec58d2dd26898ce3.png"},{"id":89413202,"identity":"2b1a2846-7446-48b4-8481-ce81bc759665","added_by":"auto","created_at":"2025-08-19 16:39:32","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":348389,"visible":true,"origin":"","legend":"\u003cp\u003eComposite volcano plot including the log\u003csub\u003e2\u003c/sub\u003e fold change and log\u003csub\u003e10\u003c/sub\u003e p value between EP_5 vs ES_5 for reactants segregation, following the layout used in Figure 6. Red areas report the compounds significantly over-represented in ES_5. Light blue-labeled points indicate compounds significantly over-represented in Na₂SO₄_5 based on the discriminant analysis between PEG_5 and Na₂SO₄_5 (Supplementary Figure S4). Dark red points represent compounds significantly over-represented in ES_5, but not significantly enriched in Na₂SO₄_5 according to the same discriminant analysis. Red, blue, purple, green, yellow, and grey triangles mark target compounds involved in proposed reaction mechanisms, including direct glucose oxidation products and other intermediate degradation products. Scatter plots in panels A–H show the normalized areas of target analytes identified in full-scan experiments following annotation and tandem MS spectral identification.\u003c/p\u003e","description":"","filename":"Figure8.png","url":"https://assets-eu.researchsquare.com/files/rs-7307960/v1/b5b2d95a137a28264007bbb3.png"},{"id":109158094,"identity":"dbf0e307-9587-4894-91b5-22594052046f","added_by":"auto","created_at":"2026-05-13 07:08:04","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2382962,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7307960/v1/e1d3aecb-264f-497d-b5a8-18a8b97c456d.pdf"},{"id":89411227,"identity":"e91ad537-9b1f-4a4e-b2ad-fb1355d9ada1","added_by":"auto","created_at":"2025-08-19 16:15:32","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":28199,"visible":true,"origin":"","legend":"Supplementary table 1","description":"","filename":"Supplementarytable1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7307960/v1/a774a593dcbb254b145b1ab2.xlsx"},{"id":89411228,"identity":"c170ae1a-76a1-41ad-8c72-8042137a8f9a","added_by":"auto","created_at":"2025-08-19 16:15:32","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":30825,"visible":true,"origin":"","legend":"Supplementary table 2","description":"","filename":"Supplementarytable2.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7307960/v1/545bb72b0ef51dc7b94a11da.xlsx"},{"id":89411233,"identity":"7e99a8a2-0b43-409e-8ed7-3a589e755128","added_by":"auto","created_at":"2025-08-19 16:15:32","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":343404,"visible":true,"origin":"","legend":"Supporting information file","description":"","filename":"supportinginformationCommChem29072025.docx","url":"https://assets-eu.researchsquare.com/files/rs-7307960/v1/9cd507ff058144cd96d8eac7.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Spatial distribution of glucose and amino acids within all-aqueous emulsions directs the Maillard reaction and oxidation pathways","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eDroplet reactors, particularly those based on all-aqueous (water-in-water, W/W) emulsions, have emerged as versatile platforms for carrying out chemical reactions within molecularly crowded biphasic systems\u003csup\u003e1\u003c/sup\u003e. The physicochemical properties of all-aqueous emulsions influence the location of reactants (precursors), directing specific chemical reactions to occur either within the dispersed droplets (i.e., compartmentalized), at the interface between the two phases or in the continuous bulk phase. Chemical reactions that can be studied using such droplet-based systems include hydrolysis, as well as acid-base, redox and enzymatic reactions. Each of these is essentially influenced by the chemical nature of the reactants and by the tendency of precursors to partition between the two aqueous phases or accumulate at the interface\u003csup\u003e2\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eAmong the many chemical transformations that can occur in such biphasic systems, the Maillard reaction (MR) is of particular interest due to its complexity and importance in food chemistry. The MR produces a wide variety of volatile and non-volatile compounds. Volatile products include nitrogen- and oxygen-containing heterocycles (e.g., pyrazines, thiazoles, furans), as well as Strecker aldehydes (e.g., isovaleraldehyde)\u003csup\u003e3\u0026ndash;5\u003c/sup\u003e. Non-volatile products range from early-stage compounds like Amadori products to melanoidins and advanced glycation end-products\u003csup\u003e6,7\u003c/sup\u003e. It is known that their formation depends on factors such as precursor type, temperature, pH, and water activity\u003csup\u003e8\u003c/sup\u003e. However, little is known about how the localization, i.e., the spatial distribution and dynamics of reactants and products within multiphase systems affect their retention, concentration, and migration of these molecules across phase boundaries. Other reactions, including Strecker degradation, retro-aldol cleavage, and oxidation may occur concurrently with the MR or interact with its intermediates contributing to this complex chemical profile\u003csup\u003e9\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eWithin this scenario, effective analysis of diverse compounds generated during the Maillard reaction in all-aqueous emulsions requires specialized analytical methods tailored to their dynamic microenvironment. Untargeted metabolomics provides a holistic profiling of the molecular species, enabling detailed characterization of both temporal (i.e., over time) and spatial (i.e., across compartments) variations in reaction intermediates and products. This systems-level approach is particularly valuable in deciphering not only the pathways of the MR but also concurrent processes, such as oxidation, that may occur even in simple sugar\u0026ndash;amino acid model systems. Similar strategies have been applied in studying the peroxidation of lipids and volatile compounds formation during food processing\u003csup\u003e10,11\u003c/sup\u003e. In the context of untargeted metabolomics, the application of hydrophilic interaction liquid chromatography high-resolution mass spectrometry (HILIC-HRMS) combined with droplet-based reactors, especially all-aqueous emulsions, for decoding the MR pathways remains the technique of choice to match the hydrophilic nature of the most relevant MR precursors.\u003c/p\u003e\u003cp\u003eOur prior work demonstrated that all-aqueous emulsions can control the spatial distribution of amino acids, reducing sugars, and the corresponding reaction intermediates, thereby directing MR pathways\u003csup\u003e12\u003c/sup\u003e. Although that study established the role of the emulsions in modulating early-stage reactions through compartmentalization and segregation, two critical unanswered questions persist: (I) how the spatial organization tackles the molecular diversity of end-products, and (II) how their formation and migration through different phases depends on amino acid partitioning behavior. Here, untargeted metabolomics was used to annotate and identify the end-products formed and elucidate the MR and oxidation pathways in all-aqueous emulsions through multivariate data analysis. All-aqueous emulsions were prepared, and the reactants were partitioned either inside the droplets (co-encapsulation, for asparagine and glucose) or segregated between the interior and exterior phases of the emulsions (yet interfacially touching, for tryptophan and glucose) according to the amino acid side chain\u003csup\u003e12\u003c/sup\u003e. HILIC-HRMS analysis coupled to molecular networking enabled comprehensive profiling and pathway elucidation. Our findings offer new insights into the role of compartmentalization in steering chemical reactions and present a novel platform for their control under food-relevant conditions.\u003c/p\u003e"},{"header":"2. Material and methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003e2.1 Chemicals and reagents\u003c/h2\u003e\u003cp\u003eAcetonitrile and water were purchased from Merck-Sigma-Aldrich (Darmstadt, Germany); D-glucose, L-tryptophan, L-asparagine, polyethylene glycol (PEG) 8 kDa, sodium sulfate (Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e), phosphate buffer powder, ammonium formate and formic acid were purchased from Merck-Sigma-Aldrich. \u003cem\u003eN\u003c/em\u003e-(1-deoxy-D-fructos-1-yl)-L-tryptophan (Fru-Trp) and \u003cem\u003eN\u003c/em\u003e-(1-deoxy-D-fructos-1-yl)-L-asparagine (Fru-Asn) were purchased from Toronto Research Chemicals (Toronto, Canada). All chemicals used in this study were of analytical grade, except for liquid chromatographic solvents, water and acetonitrile which were of mass spectrometry grade.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\u003ch2\u003e2.2 Preparation of all-aqueous droplet reactors\u003c/h2\u003e\u003cp\u003eAll-aqueous droplet reactors were prepared according to the method described in our prior report\u003csup\u003e12\u003c/sup\u003e. In short, aqueous stock solutions of Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e (15 wt%) and PEG (50 wt%) were prepared by the addition of either Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e or PEG to phosphate buffer stock solution (0.1 M, pH 7.4). All-aqueous droplet reactors (Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e/PEG) were prepared by mixing the Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e and PEG stock solutions at a volume ratio of 1:4, with stirring at 1000 rpm for 1 h. Subsequently, the droplet reactors were supplemented with the Maillard reactants (glucose and either asparagine or tryptophan), achieving a final concentration of 40 mM. The mixtures were stirred overnight at 25\u0026deg;C for complete dissolution of the reactants. After phase separation, PEG phase (EP) and Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e phase (ES) were collected from samples. For comparison purposes, in addition to the droplet reactor samples (two-phase systems), single-phase solutions were also prepared. These included mixtures of phosphate buffer solution and Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e or PEG stock solutions (referred to as \"Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e\" and \"PEG\", respectively). All these solutions were supplemented with the same reactants.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\u003ch2\u003e2.3. Running the MR in droplet reactors and single-phase solutions\u003c/h2\u003e\u003cp\u003eThe reactants-supplemented emulsions (and single-phase solutions, 5 mL) were transferred into gas chromatography vials (20 mL), and securely sealed. Subsequently, these vials were subjected to heating at 95\u0026deg;C for 0 h, 3 h, and 5 h while being stirred at 1000 RPM. Next, the samples were rapidly cooled within an ice bath, then kept at room temperature (25\u0026deg;C) for 4 h. The top and bottom phases were then subjected to diafiltration (molecular weight cut-off 3000 Da) at 14,000 \u0026times; g, repeated five times, with each cycle lasting 30 minutes. Both the retentate and permeate fractions were collected and subsequently freeze-dried (Alpha 2\u0026ndash;4 LD plus, Martin Christ, Osterode am Harz, Germany) for further analysis. For comparison purposes, the single-phase samples supplemented with the reactants were also stored at 25\u0026deg;C for 4 h and subjected to diafiltration.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\u003ch2\u003e2.4 Liquid chromatography high resolution mass spectrometry\u003c/h2\u003e\u003cp\u003eThe formation of reaction products of asparagine and tryptophan was screened in untargeted mode in emulsion systems after phase separation by using PEG solution and Na₂SO₄ solution as controls. Freeze-dried permeates were redissolved in water and both the upper and lower phases were diluted in 50% aqueous acetonitrile then centrifuged (18,000 x g, 4\u0026deg;C, 10 min). LC-MS/MS data were acquired using an Exploris 120 quadrupole Orbitrap high-resolution mass spectrometer interfaced to a Vanquish Core liquid chromatographic system (Thermo Fisher Scientific, Bremen, Germany). Compounds were separated at 40\u0026deg;C through a zwitterionic sulfobetaine column (Atlantis Premier BEH, Z-HILIC, 100 x 2.1, 1.7 \u0026micro;m, Waters) with the following gradient of solvent B (minutes/%B): (0/5), (1.5/5), (10/50), (13/50). Mobile phases consisted of 0.1% formic acid in acetonitrile (v/v, solvent A) and 0.1% formic acid in water (v/v, solvent B) and the flow rate was 0.3 mL/min. Samples were acquired in profile data dependent scanning (ddMS\u003csup\u003e2\u003c/sup\u003e) mode by using polarity switching. For positive ion (and negative ion) mode, H-ESI interface parameters were as follows: static spray voltage 3.3 kV (-2.9 kV), while ion transfer tube and vaporizer temperature were both at 280\u0026deg;C; sheath gas flow and auxiliary gas flow were 38 and 8 arbitrary units, respectively, for both positive and negative ion mode. Along with full scan acquisition in the \u003cem\u003em/z\u003c/em\u003e range 60\u0026ndash;900 (resolution 60,000 at \u003cem\u003em/z\u003c/em\u003e 200 FWHM), ddMS\u003csup\u003e2\u003c/sup\u003e were performed on the top 4 scans by using an intensity threshold of 60,000 (area counts), a customized dynamic exclusion window of 3.5 s and a mass tolerance of 5 ppm. A targeted mass exclusion list was incorporated within the method upon the analysis of two separated full scan acquisitions of blank samples and the removal of the top 600 most intense background signals, for both positive and negative ion mode. Finally, ddMS\u003csup\u003e2\u003c/sup\u003e scan properties included an isolation window of \u003cem\u003em/z\u003c/em\u003e 1.5, a normalized higher energy collisional dissociation (HCD) of 20%, 40% and 80% and the Orbitrap resolution fixed at 30,000. Data were acquired in profile mode by using a standardized AGC (all gain control) target and EASY-IC\u0026reg; (fluoranthene as internal calibrant) at the beginning of each run. Identification of precursors, putative intermediates and end-products was achieved by injecting quality control samples at regular intervals of 6 runs and by analyzing reference mixes through differential scanning ranges (\u003cem\u003em/z\u003c/em\u003e 60\u0026ndash;200; \u003cem\u003em/z\u003c/em\u003e 190\u0026ndash;400, \u003cem\u003em/z\u003c/em\u003e 390\u0026ndash;600 and \u003cem\u003em/z\u003c/em\u003e 590\u0026ndash;900) in both positive and negative mode by using the same chromatographic layout as described above.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\u003ch2\u003e2.5 Untargeted metabolomics\u003c/h2\u003e\u003cp\u003eAn untargeted metabolomic workflow based on end-products identification was used to find and characterize differences between the buffer control samples and fully aqueous emulsions in the presence of asparagine and glucose or tryptophan and glucose by importing raw files in Compound Discoverer software (v. 3.3 Thermo Fisher Scientific, San Jos\u0026eacute;, CA). The procedure involved the retention time alignment and detection of expected and unknown compounds for samples grouping. Upon definition of the elemental composition, exact masses, chemical formulas and fragmentation spectra, compounds were matched with analytical standards as glucose, tryptophan, asparagine and their respective Amadori compounds. The procedure encompassed the generation of five separated libraries including end-products with a similar fragmentation pattern. This strategy generated molecular networks that were combined with analytical information reported in publicly available databases, as mzCloud (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e\u003ca href=\"http://www.mzcloud.org\" target=\"_blank\"\u003ewww.mzcloud.org\u003c/a\u003e\u003c/span\u003e\u003cspan address=\"http://www.mzcloud.org\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) and ChemSpider (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e\u003ca href=\"http://www.mzcloud.org\" target=\"_blank\"\u003ewww.chemspider.com\u003c/a\u003e\u003c/span\u003e\u003cspan address=\"http://www.chemspider.com\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). A supplementary search was performed in Reaxys\u0026reg; (Elsevier, Amsterdam, the Netherlands) by using as reactants the two amino acids and glucose, with temperature higher than 50\u0026deg;C, to generate an internal mass list of putative chemical structures arising from the model systems. After correction for quality control samples, post-processing nodes performed descriptive statistics and differential analysis working on two technical replicates arising from two replicates representative of four observations for each time point/condition. To further investigate significative differences between samples, hypothesis test was performed by one-way ANOVA model with Tukey as \u003cem\u003epost-hoc\u003c/em\u003e test. For volcano plots and ratio among sample groups, \u003cem\u003ep\u003c/em\u003e-values were adjusted by Benjamini-Hochberg algorithm. Tandem MS spectra, principal component analysis (PCA), histograms, scatter plot, loadings plots, molecular network\u003csup\u003e13\u003c/sup\u003e, statistical test, and log fold changes were obtained in Compound Discoverer. Reaction pathways and chemical structures were built in ChemDraw (Revvity Signals Software, Waltham, MA).\u003c/p\u003e\u003c/div\u003e"},{"header":"3. Results and Discussion","content":"\u003cp\u003e\u003cb\u003e3.1 The analytical background for the definition of the chemical space behind reaction mechanisms in all-aqueous emulsions\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThe spatial distribution of reaction precursors including amino groups and reducing sugars within all-aqueous emulsions leads to the formation of thousands of compounds with distinct properties. We opted for zwitterionic HILIC as a flexible technique to appropriately separate polar charged compounds present in an aqueous environment. This method also provided a good response for the separation of chromophores and brown pigments or polymers (usually separated in reversed phase chromatography) formed in both tryptophan- and asparagine-supplemented samples as a result of condensation, oxidation, and polymerization reactions\u003csup\u003e14\u003c/sup\u003e. Then, high resolution tandem mass spectrometry-based metabolomics provided the analytical context for the annotation and identification of unknown end-products formed \u003csup\u003e10\u003c/sup\u003e. Control samples and procedural blanks were used to define the chemical space in the untargeted analysis: in full scan MS1 mode, a total of 524 signals were detected in the system with segregated reactants compared to 917 signals in the system where reactants were co-encapsulated. To improve the reliability of chemical feature responses, several filters were used starting with the removal of signals that were not associated with any chemical formulas due to poor matching with theoretical isotopic pattern distribution. Next, we considered only compounds in tandem mode that generated a fragmentation mass spectrum and added structural information to the isotopic pattern. The third filter included a chromatographic peak score higher than 5 as the result of the profiles associated with the scan points below each peak in all the replicates. Finally, the metabolomic workflow included two separate procedures: each run was screened first in polarity switching mode by considering both ionization modalities. Annotated compounds in polarity switching experiments were further checked in positive or negative ion mode. Wherever possible, spectra were manually curated in both positive and negative mode by excluding co-eluting compounds with overlapping peak shape that could result from in-source ionization or fragmentation typical of analytical procedures that consider broad dynamic ranges\u003csup\u003e15\u003c/sup\u003e. Upon all the filtering procedures, we annotated 265 compounds in all the tryptophan-supplemented samples and 302 compounds in all asparagine-supplemented samples. In \u003cb\u003eSupplementary Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e and S2\u003c/b\u003e, we reported for each compound an identification level ranging from 1 (compound identification upon matching to analytical standards as the amino acids and glucose as precursors and the two Amadori compounds as intermediate, reported in bold) to 2 (compounds annotation based on mass spectra, isotopic pattern, chemical formula, mass accuracy and matching with databases when compound names are present) according to Metabolomics standard initiative (MSI) level\u003csup\u003e16\u003c/sup\u003e. In this view, we decided to keep tryptophan-supplemented and asparagine-supplemented samples separated, and study the type of compounds formed, then explore the overall reactivity\u003csup\u003e12\u003c/sup\u003e.\u003c/p\u003e\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\u003ch2\u003e3.2 Tryptophan and glucose reactivity in reactants segregation.\u003c/h2\u003e\u003cp\u003eFigure \u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e provides a sample distribution overview of tryptophan-supplemented samples at time 0 and at 5 h in the form of a PCA, accompanied by a loadings plot and a molecular network of the annotated compounds. Figure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA shows that the overall variance explained was 47.8% with the first component accounting for 27.9% and the second component for 19.9%; samples Na₂SO₄_0, ES_0, and ES_5 were in the first quadrant. It is worth mentioning that ES remained in the first quadrant even after 5 h of heating, suggesting that the key features of its major reaction products did not change significantly. This preliminary screening was in line with the results already reported by our group: after the initial reaction, part of the products migrated into the EP phase instead of undergoing continuous reaction within the ES. EP_0 and PEG_0, both containing PEG were in the second quadrant indicating marked differences toward the other two Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e-containing samples. In the third quadrant, EP and PEG at 5 h were predominant, and both clustered within the same region, indicating that they generated similar reaction product profiles after heating. One possible explanation for this observation is that EP functions as the main site for subsequent transformations, resulting in similar reaction pathways and product distributions as observed in the PEG samples.\u003c/p\u003e\u003cp\u003eThis two-dimensional distribution corresponded to the molecular loadings shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB, where each grey point represents one of the 320 compounds annotated when reactants were segregated between the emulsion phases. As expected, dark gray datapoints predominantly clustered within the second and third quadrants (Q2-Q3), corresponding to compounds exhibiting significantly elevated response intensities in PEG_5, EP_5, and Na₂SO₄_5 samples. This pattern is likely attributed to the effects of prolonged thermal processing (95\u0026deg;C for 5 h), which promoted advanced glycation and oxidation reactions, as well as cleavage and polymerization processes.\u003c/p\u003e\u003cp\u003eFigure \u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC presents information on the molecular interrelationships between glucose and tryptophan reaction products. The molecular network illustrates overall chemical patterns independently of model systems, with green points representing annotated compounds and blue points indicating unknown compounds. The red box highlights degradation products associated with glucose, including putative intermediates involved in dicarbonyl formation and small organic acids. In contrast, the blue box contains compounds with fragmentation spectra similar to those of tryptophan, such as indole derivatives and β-carbolines. These two compartments were connected by \u003cem\u003eN\u003c/em\u003e-(1-deoxy-D-fructos-1-yl)-tryptophan (Fru-tryptophan), which exhibits fragmentation patterns consistent with those of glucose, showing consecutive losses of water from secondary alcohol groups, as well as with tryptophan, characterized by indole ring formation and diagnostic ions at \u003cem\u003em/z\u003c/em\u003e 146.0601 and 188.0706. The molecular network and the observed spectral similarities reflected the theoretical grouping of different MR pathways, as previously hypothesized by Yaylayan \u003csup\u003e17\u003c/sup\u003e. In particular, the Amadori compounds appear to serve as a key bridging point between amino acids and glucose, and they provide a basis for studying the spatial organization of the resulting degradation products.\u003c/p\u003e\u003cp\u003eTo gain insight into the qualitative and quantitative relationships among the four model systems (two buffer systems and the emulsion systems after phase separation), we employed a supervised discriminant analysis based on fold change in logarithmic scale and post-hoc test. The results were summarized using a composite volcano plot of the emulsion system, comparing the two phases (PEG and Na₂SO₄) at 5 h (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). In the green region, light blue points highlighted 33 discriminating compounds that were over-represented in PEG_5, as identified by a separate differential analysis between PEG_5 and Na₂SO₄_5 (\u003cb\u003eSupplementary Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e\u003c/b\u003e). Hence, among the 60 compounds over-represented in the green region, 27 were found at significantly higher levels. We subsequently performed manual curation of fragmentation spectra for target analytes present in EP_5 (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e with a magnified area). These analytes were highlighted using boxes of different shapes, corresponding to the reaction mechanisms illustrated in Figs.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e and \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. Specifically, analytes were marked with yellow, green, blue, and red circles, as well as green, red, and blue rectangles, to represent distinct mechanistic categories.\u003c/p\u003e\u003cp\u003eConsidering the evidence that Amadori compound rearrangement is likely to occur at the interface, we propose that indole oxidation may predominate when reactants were segregated between the emulsion phases. Figure\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e illustrates the oxidation pathways initiated by tryptophan, along with trend lines showing the normalized area counts of the target analytes over 5 h. We observed that oxidation products such as \u003cem\u003eN\u003c/em\u003e-formylkynurenine and \u003cem\u003eN\u003c/em\u003e-formylanthranilic acid were formed at significantly higher levels in the PEG phase, while their concentrations were lower in the PEG solution, indicating that the all-aqueous emulsion system plays a crucial role in the synthesis of these compounds. Conversely, \u003cem\u003eN\u003c/em\u003e-formylkynurenine and \u003cem\u003eN\u003c/em\u003e-formylanthranilic were detected only in trace amounts in Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e solution; however, due to their specific chemical properties, their concentrations increased over time in the Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e phase. Furthermore, we spotted the formation of four structural isomers peaking at \u003cem\u003em/z\u003c/em\u003e 361.1426 with molecular formula C21H19N3O3, different retention times, similar kinetic profiles and only one eluting at 4.8 min characterized by the predominance of typical fragmentation spectra of tryptophan (C11H10NO2, \u003cem\u003em/z\u003c/em\u003e 188.0705; C9H8NO, \u003cem\u003em/z\u003c/em\u003e 146.0600; C8H8N, \u003cem\u003em/z\u003c/em\u003e 118.0650, as suggested by the molecular network in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC and by the spectra in \u003cb\u003eSupplementary Figure \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003eA\u003c/b\u003e). In \u003cb\u003eSupplementary Figure \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003eB\u003c/b\u003e, fragment ion search (FiSH) scoring confirmed a structural match with \u003cem\u003eN\u003c/em\u003e-(1H-indol-3-ylacetyl)tryptophan, suggesting the formation of an amide bond between 3-indoleacetic acid and tryptophan. This condensation is likely promoted under alkaline conditions and local dehydration.\u003c/p\u003e\u003cp\u003eFigure \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e summarizes the reaction pathways involved in the formation of condensation and polymerization products. The key intermediates include tryptamine, formed via decarboxylation of tryptophan, and indole-3-pyruvic acid, produced through oxidative deamination\u003csup\u003e18\u003c/sup\u003e. Both intermediates can undergo condensation, leading to the formation of polymerization products. In the presence of reactive α-dicarbonyls, a specific class of β-carbolines can also be synthesized through intramolecular cyclization\u003csup\u003e19\u003c/sup\u003e; β-carbolines are characterized by different logP depending on the type of α-dicarbonyls. In \u003cb\u003eSupplementary Figure \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e\u003c/b\u003e, we reported the distribution of two compounds formed from the interaction between reactive carbonyls and tryptophan. Although 1-carboxyethyl-β-carboline shares strong structural similarities with its propanoic acid derivative, their partitioning behavior differs significantly. Furthermore, tryptophan fragmentation products might contribute to the generation of other compounds including ethylamine and indole derivatives and can serve as indole building blocks for the formation of brown pigments and volatiles.\u003c/p\u003e\u003cp\u003eThe approach for PEG phase was extended to the study of Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e phase, and Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e presents the composite volcano plot comparing ES and EP at 5 h. The 42 light blue compounds represent those over-represented in the Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e phase compared to the PEG buffer phase (Supplementary Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). As expected, none of the compounds that were overrepresented in the Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e solution at 5 h maintained the same pattern in the emulsion system. All were distributed below the green and red regions, indicating very low \u0026ndash;log₁₀ p-values and, thus, weak discriminative power. This analytical strategy enabled the identification of compounds significantly enriched in the Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e phase after phase separation and facilitated the investigation of how compounds migrate from the PEG phase to the Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e phase. Such compounds were highlighted using red, purple, blue, and yellow triangles. The quantitative trends shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e were consistent across datasets: a pronounced and statistically significant increase in the Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e phase after phase separation, with negligible signals detected in the other model systems. One exception was 3-amino-2,3-dideoxy-scyllo-inosose, a putative dehydration product of glucosamine, which was formed at comparable levels in both Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e solution and Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e phase (Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e and ES). Here, most of the generated compounds originate from the glucose moiety, with some amino-sugar derivatives tentatively formed at the interface, while none of the indole derivatives were detected at significative levels among the over-represented molecules.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e\u003ch2\u003e3.3 Asparagine and glucose reactivity in reactants co-encapsulation.\u003c/h2\u003e\u003cp\u003eFigure \u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e provides an overview of the 2D distribution of samples and the 423 compounds annotated in when reactants were co-encapsulated. This analysis compares data from time 0 and after 5 h of thermal treatment at 95\u0026deg;C. The PCA shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA explains a total variance of 53.0%, with a more complex sample distribution than observed in reactants segregation. At time 0, all samples were in the third and fourth quadrants. After 5 h, the samples shifted along the second principal component (accounting for 20.9% of the variance), all exhibiting positive PC2 values.\u003c/p\u003e\u003cp\u003eThis distribution highlights key differences already present at time 0, suggesting that the preparation of the model system (including pre-incubation, mixing, and dissolution) may have induced the formation of early-stage reaction products. The first principal component (PC1) accounted for 32.1% of the total variance. This interpretation is supported by the loadings plot in Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eB, where grey data points representing annotated compounds clustered around black points along both axes, indicating a greater contribution from Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e-containing systems (Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e solution and ES at time 0 and 5 h).\u003c/p\u003e\u003cp\u003eFigure \u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eC depicts the interconnection between asparagine, glucose, and their respective reaction products. Again, the Amadori compound of asparagine is positioned between two molecular clusters representing amino acid degradation products and glucose degradation products, in red and blue boxes, respectively. This distribution confirmed what observed for tryptophan: Maillard reaction and the Amadori rearrangement remains the key driver in both reactants segregation and co-encapsulation in the presence of free amino group and reducing sugars. Besides the linkages between the two clusters, a closer relationship was observed between the Amadori compound of asparagine and several degradation products, most of which are small organic acids. Here, we hypothesize that the carbohydrate moiety of the Amadori compound contributes more efficiently to the generation of fragmentation products than tryptophan, likely due to reduced intramolecular reactivity of indole rings with dicarbonyl compounds. As in reactant segregation, the co-encapsulation molecular network was constructed across the four model systems without using time as a discriminating variable.\u003c/p\u003e\u003cp\u003eTo provide information on the spatial organization of reactants co-encapsulation and segregation, a composite volcano plot was developed in Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e (log₂ fold change\u0026thinsp;\u0026gt;\u0026thinsp;1 and log₁₀ \u003cem\u003ep\u003c/em\u003e-value\u0026thinsp;\u0026gt;\u0026thinsp;1.5), identifying 92 compounds over-represented in EP after 5 h at 95\u0026deg;C. As a preliminary step, 38 compounds that were over-represented in PEG compared to Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e solution at 5 h (\u003cb\u003eSupplementary Figure S4\u003c/b\u003e) were highlighted in light blue. This strategy enabled an initial screening of compounds significantly enriched or accumulated in the PEG phase after phase separation. Among the 16 compounds over-represented in EP (green region in Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e), four annotated compounds were further highlighted using red, blue, green, and yellow circles. Three compounds exhibited similar fragmentation spectra characterized by an intense signal at \u003cem\u003em/z\u003c/em\u003e 96.96. This conserved and prominent peak is indicative of a sulfate ion fragment. We hypothesize that reactions involving sulfate moieties occur in the Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e phase, with subsequent cyclization products migrating into the PEG phase. In addition, 1H-pyrrole-2-carboxamide serves as a key example of a pyrrole derivative that moves between the two regions, exhibiting accumulation in the PEG phase.\u003c/p\u003e\u003cp\u003eIn parallel, Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e presents the same composite volcano plot shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e, with light blue-labeled points derived from the discriminant analysis between Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e and PEG solutions at 5 h and pinpointed in the red region. According to this layout, we identified the compounds that accumulated in the Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e phase after phase separation, excluding the 29 compounds over-represented in Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e based on the PEG vs. Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e discriminant analysis (Supplementary Figure S4). Figures\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eA and \u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eB illustrate two examples of pyridine derivatives that accumulated significantly in the Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e phase, while remaining at very low concentrations in both PEG and EP. Figure\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eC summarizes the time-dependent trend of 3-deoxyglucosone. Although 3-deoxyglucosone is hydrophilic, its normalized area counts were comparable between the ES and EP. Notably, an increasing trend over time was observed only in the Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e phase after phase separation, suggesting its distribution among both phases. These results suggest that compounds formed in the Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e phase may continue to migrate between phases, influenced by reaction conditions such as time and temperature.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eIn addition to asparagine degradation products, we identified other markers of asparagine oxidation: aspartyl-asparagine and asparaginyl-aspartic acid, two dipeptide isomers that eluted at different retention times but displayed identical fragmentation spectra (\u003cb\u003eSupplementary Figure S5A\u003c/b\u003e). In \u003cb\u003eSupplementary Figure S5B\u003c/b\u003e, we reported the MS/MS spectrum with FiSH scoring for the putative chemical compound including the formation of a peptide bond between aspartic acid and asparagine. Both isomers may originate from the initial oxidation of asparagine to aspartic acid, followed by further peptide bond formation and oxidation. Despite amine acylation in water is a common reaction in the presence of dedicated enzymes, this process occurs rarely in the absence of catalysts; this is because the high concentration of water shifts the equilibrium toward hydrolysis. In the present work, we observed the formation of the peptide bond probably as result of the alteration of the local water concentration due the different reagent partitioning or the occurrence of a base-catalyzed process at the hydrophobic interfacial compartment that promoted the formation of the necessary tetrahedral reaction intermediate\u003csup\u003e20\u003c/sup\u003e.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003e3.4 Reactants co-encapsulation and segregation introduce the space as a control parameter for chemical reactions.\u003c/h2\u003e\u003cp\u003eAll-aqueous emulsions influence the spatial distribution of reactants based on the intrinsic chemical properties of the precursors, thereby affecting overall chemical reactivity and contributing to system stabilization\u003csup\u003e21\u003c/sup\u003e. The dichotomy between co-encapsulation and segregation, as specific modes of compartmentalized reactions, enabled the spatial separation of glucose, asparagine, and tryptophan within the two immiscible aqueous phases of the all-aqueous emulsions studied here. Comparing their separation across the phases with their coexistence within a single phase allowed for the controlled execution of sequential transformations characteristic of the Maillard cascade, including the formation of the Amadori compounds and related oxidation processes\u003csup\u003e12\u003c/sup\u003e. For the first time, we demonstrated that melanoidin building blocks can migrate from one phase to another, representing an extreme case of on-droplet formation. The selective partitioning of the two amino acids into specific phases or at the interface increased their local concentrations at the reaction sites, thereby enhancing reaction rates in accordance with the principles of mass action. Through our HILIC-HRMS and multivariate data analysis protocol, we highlighted how precursors with distinct physicochemical properties behave differently in all-aqueous emulsions and underscores the potential of metabolomics to elucidate the complex interplay among time, temperature, reactant localization, and the final distribution of end-products within the emulsion system.\u003c/p\u003e\u003cp\u003eAs previously reported by our group, the partition coefficients of amino acids regulate the localization of side chains in all-aqueous emulsions and influence the formation of reaction end-products \u003csup\u003e12\u003c/sup\u003e. We observed that the localization of precursors, whether coexisting within droplets as in the case of asparagine and glucose or spatially separated at the interface as with tryptophan and glucose, can lead to the formation of distinct and unique reaction products because of the intrinsic chemical nature of the precursors (reactants location) and the subsequent distribution of the resulting products within all-aqueous emulsions. Specifically, tryptophan and glucose first reacted at the interface to form the corresponding glycosylamine. This intermediate subsequently migrated into the PEG phase. In addition to participating in the Maillard reaction, tryptophan was also subject to oxidation. In this context, it has been reported that tryptophan can form 3-hydroperoxytryptophan and a dioxoethane intermediate through oxidation pathways, both of which are proposed precursors to \u003cem\u003eN\u003c/em\u003e-formylanthranilic acid\u003csup\u003e18\u003c/sup\u003e. Furthermore, tryptophan may undergo oxidative deamination to yield indole-3-pyruvic acid and decarboxylation to produce tryptamine, as here reported for reactant segregation mode. The condensation of these two compounds can lead to the formation of 1-(methylenindole)-tetrahydro-β-carboline-3-carboxylic acid, a β-carboline derivative. Notably, these transformation pathways are typically activated under high-temperature or in anaerobic conditions (up to 140\u0026deg;C), as described by Bellmaine, Schnellbaecher and Zimmer \u003csup\u003e18\u003c/sup\u003e. Maillard reaction intermediates and advanced products, such as carbonyl compounds, pyridines, quinones, and pyrroles, often carry reactive functional groups including α-dicarbonyls, aldehydes, ketones, and quinones\u003csup\u003e22\u003c/sup\u003e. These groups possess oxidative properties and can generate free radicals, contributing to the chemical complexity of the system with propagation reactions, consistent with previous observations in oil-in-water emulsions\u003csup\u003e23\u003c/sup\u003e. Indeed, the co-localization of these MR products and tryptophan within the PEG phase likely promoted the oxidative transformation of tryptophan, either by reaction with α-dicarbonyls or through the generation of reactive oxygen species, which can also be formed from the degradation of PEG. These processes may have led to modifications in the overall supramolecular arrangement of the all-aqueous system. The results suggest that reactants segregation not only directs initial Maillard reaction steps but also modulates the fate of amino acid-derived intermediates through selective partitioning and the resulting localized reactivity.\u003c/p\u003e\u003cp\u003eThe reaction between glucose and asparagine in all-aqueous emulsion is influenced when reactants were segregated between the emulsion phases. The co-encapsulation of the reaction precursors within the Na₂SO₄ phase enabled the Maillard reaction to proceed in a compartmentalized manner. Specifically, glycosylamines were significantly partitioned into the Na₂SO₄ phase and subsequently underwent the Amadori rearrangement. The resulting Amadori products then degraded to form reactive α-dicarbonyl compounds, such as 3-deoxyglucosone\u003csup\u003e24\u003c/sup\u003e. We hypothesize that under these conditions, glucose may have undergone rearrangement, partial deoxygenation, and subsequent nucleophilic addition by amines, potentially leading to the formation of a stable six-membered ring compound, putatively identified as 3-amino-2,3-dideoxy-scyllo-inosose.\u003c/p\u003e\u003cp\u003eAsparagine is known to undergo non-enzymatic deamidation under physiological or elevated temperature conditions. This process involves the formation of a cyclic imide intermediate, which is unstable and subsequently hydrolyzed to yield a mixture of L-aspartic acid (Asp) and L-isoaspartic acid. Here, we propose that Asp and Asn may undergo a condensation reaction to form dipeptides such as Asn\u0026ndash;Asp or Asp\u0026ndash;Asn. However, peptide bond formation from free amino acids in aqueous solution is thermodynamically unfavorable due to the high-water activity and the lack of catalyzing agents\u003csup\u003e25\u003c/sup\u003e. Nevertheless, under our experimental conditions, factors such as reduced water activity, selective partitioning, alkaline pH, and local microenvironments with limited hydration may have helped in shifting the equilibrium toward peptide bond formation. It is therefore possible that the high salt concentration in the Na₂SO₄ phase created a dehydrating environment that facilitated this reaction\u003csup\u003e26\u003c/sup\u003e. Taken together, the co-encapsulation of glucose and asparagine within the Na₂SO₄ phase created a microenvironment with reduced water activity and elevated local reactant concentrations, thereby facilitating condensation reactions such as the Amadori rearrangement and possibly peptide bonds formation in the absence of conventional catalyzing agents. Of note, a similar pathway was observed also in the case of segregation mode with the formation of \u003cem\u003eN\u003c/em\u003e-(1H-indol-3-ylacetyl)tryptophan, as a reaction product between 3-indoleacetic acid and tryptophan. In both segregation and co-encapsulation cases, the use of labelled precursors, advanced techniques as NMR and testing the system in absence of glucose can provide further structural confirmation to those observed here by tandem mass spectrometry.\u003c/p\u003e\u003c/div\u003e"},{"header":"4. Conclusions","content":"\u003cp\u003eOverall, our findings highlight how all-aqueous emulsions can drive distinct chemical trajectories for amino acids and sugars, with broad implications across various reaction types. The reaction intermediates and products generated within these systems can, in turn, influence other processes, as shown here for oxidation and peptide bonds formation. Indeed, the selective partitioning of reactants and intermediates into PEG or Na₂SO₄ phases not only directs canonical glycation reactions but also fosters oxidation and condensation transformations that are otherwise suppressed in homogeneous aqueous systems. These findings provide new insight into how microenvironmental factors, such as water activity and reactant concentration, can be harnessed to modulate reaction pathways and product profiles in complex food matrices. Finally, we finally demonstrated that spatial compartmentalization, beyond traditional parameters like time, pH, viscosity, and temperature, is a critical factor in governing all the above-mentioned reactions.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eCompeting interests\u003c/h2\u003e\u003cp\u003eThe authors declare no competing financial interest.\u003c/p\u003e\u003ch2\u003eAuthor contributions\u003c/h2\u003e\u003cp\u003eK.C. prepared the samples, carried out the experiments, curated the data, and wrote the original draft. S.D.P. and A.D.T. performed the HILIC-HRMS analysis and data interpretation. All co-authors jointly discussed the results. A.M., A.S., and V.F. contributed to manuscript revision and supervision.\u003c/p\u003e\u003ch2\u003eAcknowledgments\u003c/h2\u003e\u003cp\u003eThis study was financially supported by the Food Quality and Design group in Wageningen University and Research and China Scholarship Council. A part of this study was funded by the National Recovery and Resilience Plan, mission 4, component 2, investment 1.3, call n. 341/2022 of Italian Ministry of University and Research funded by the European Union - NextGenerationEU for the project \u0026ldquo;ON Foods-Research and innovation network on food and nutrition Sustainability, Safety and Security-Working ON Foods\u0026rdquo;, project PE00000003, concession decree n. 1550/2022, CUP B83C22004790001.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eRuiz-Lopez, M. F., Francisco, J. S., Martins-Costa, M. T. C. \u0026amp; Anglada, J. M. Molecular reactions at aqueous interfaces. \u003cem\u003eNature Reviews Chemistry\u003c/em\u003e \u003cstrong\u003e4\u003c/strong\u003e, 459-475 (2020).\u003c/li\u003e\n\u003cli\u003eMadadlou, A., Saggiomo, V., Schroen, K. \u0026amp; Fogliano, V. All-aqueous emulsions as miniaturized chemical reactors in the food and bioprocess technology. \u003cem\u003eCurrent Opinion in Food Science\u003c/em\u003e \u003cstrong\u003e33\u003c/strong\u003e, 165-172 (2020).\u003c/li\u003e\n\u003cli\u003eShakoor, A., Zhang, C., Xie, J. \u0026amp; Yang, X. Maillard reaction chemistry in formation of critical intermediates and flavour compounds and their antioxidant properties. \u003cem\u003eFood Chemistry\u003c/em\u003e \u003cstrong\u003e393\u003c/strong\u003e, 133416 (2022).\u003c/li\u003e\n\u003cli\u003eGao, Y., Miao, J. \u0026amp; Lai, K. Study on Maillard reaction mechanism by quantum chemistry calculation. \u003cem\u003eJournal of molecular modeling\u003c/em\u003e \u003cstrong\u003e29\u003c/strong\u003e, 81 (2023).\u003c/li\u003e\n\u003cli\u003eMottram, D. S. in \u003cem\u003eFlavours and Fragrances: Chemistry, Bioprocessing and Sustainability\u003c/em\u003e (ed Ralf G\u0026uuml;nter Berger) 269-283 (Springer Berlin Heidelberg, 2007).\u003c/li\u003e\n\u003cli\u003eMurata, M. Browning and pigmentation in food through the Maillard reaction. \u003cem\u003eGlycoconj J\u003c/em\u003e \u003cstrong\u003e38\u003c/strong\u003e, 283-292 (2021).\u003c/li\u003e\n\u003cli\u003eBork, L. V., Haase, P. T., Rohn, S. \u0026amp; Kanzler, C. Structural characterization of polar melanoidins deriving from Maillard reaction intermediates - A model approach. \u003cem\u003eFood Chem\u003c/em\u003e \u003cstrong\u003e395\u003c/strong\u003e, 133592 (2022).\u003c/li\u003e\n\u003cli\u003evan Boekel, M. A. Kinetic aspects of the Maillard reaction: a critical review. \u003cem\u003eNahrung\u003c/em\u003e \u003cstrong\u003e45\u003c/strong\u003e, 150-159 (2001).\u003c/li\u003e\n\u003cli\u003eHellwig, M. \u0026amp; Henle, T. Baking, ageing, diabetes: a short history of the Maillard reaction. \u003cem\u003eAngewandte Chemie International Edition\u003c/em\u003e \u003cstrong\u003e53\u003c/strong\u003e, 10316-10329 (2014).\u003c/li\u003e\n\u003cli\u003eWeidner, L., Cannas, J. V., Rychlik, M. \u0026amp; Schmitt-Kopplin, P. Molecular characterization of cooking processes: A metabolomics decoding of vaporous emissions for food markers and thermal reaction indicators. \u003cem\u003eJournal of Agricultural and Food Chemistry\u003c/em\u003e \u003cstrong\u003e71\u003c/strong\u003e, 17442-17454 (2023).\u003c/li\u003e\n\u003cli\u003eZhou, Z.\u003cem\u003e et al.\u003c/em\u003e Unraveling the Thermal Oxidation Products and Peroxidation Mechanisms of Different Chemical Structures of Lipids: An Example of Molecules Containing Oleic Acid. \u003cem\u003eJournal of Agricultural and Food Chemistry\u003c/em\u003e \u003cstrong\u003e70\u003c/strong\u003e, 16410-16423 (2022).\u003c/li\u003e\n\u003cli\u003eChen, K.\u003cem\u003e et al.\u003c/em\u003e Compartmentalization vs. segregation of reactants: Accomplishment of the Maillard reaction at the water-water interface. \u003cem\u003eFood Chemistry\u003c/em\u003e \u003cstrong\u003e465\u003c/strong\u003e, 142089 (2025).\u003c/li\u003e\n\u003cli\u003eSchmid, R.\u003cem\u003e et al.\u003c/em\u003e Ion identity molecular networking for mass spectrometry-based metabolomics in the GNPS environment. \u003cem\u003eNature communications\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 3832 (2021).\u003c/li\u003e\n\u003cli\u003eYan, Y., Hemmler, D. \u0026amp; Schmitt-Kopplin, P. Discovery of glycation products: Unraveling the unknown glycation space using a mass spectral library from in vitro model systems. \u003cem\u003eAnal Chem\u003c/em\u003e \u003cstrong\u003e96\u003c/strong\u003e, 3569-3577 (2024).\u003c/li\u003e\n\u003cli\u003eKaufmann, A. High-resolution mass spectrometry for bioanalytical applications: Is this the new gold standard? \u003cem\u003eJournal of Mass Spectrometry\u003c/em\u003e \u003cstrong\u003e55\u003c/strong\u003e, e4533 (2020).\u003c/li\u003e\n\u003cli\u003eSumner, L. W.\u003cem\u003e et al.\u003c/em\u003e Proposed minimum reporting standards for chemical analysis: chemical analysis working group (CAWG) metabolomics standards initiative (MSI). \u003cem\u003eMetabolomics\u003c/em\u003e \u003cstrong\u003e3\u003c/strong\u003e, 211-221 (2007).\u003c/li\u003e\n\u003cli\u003eYaylayan, V. A. Classification of the Maillard reaction: A conceptual approach. \u003cem\u003eTrends in Food Science \u0026amp; Technology\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, 13-18 (1997).\u003c/li\u003e\n\u003cli\u003eBellmaine, S., Schnellbaecher, A. \u0026amp; Zimmer, A. Reactivity and degradation products of tryptophan in solution and proteins. \u003cem\u003eFree Radical Biology and Medicine\u003c/em\u003e \u003cstrong\u003e160\u003c/strong\u003e, 696-718 (2020).\u003c/li\u003e\n\u003cli\u003eHerraiz, T., Pe\u0026ntilde;a, A., Mateo, H., Herraiz, M. \u0026amp; Salgado, A. Formation, characterization, and occurrence of \u0026beta;-carboline alkaloids derived from \u0026alpha;-dicarbonyl compounds and L-tryptophan. \u003cem\u003eJournal of Agricultural and Food Chemistry\u003c/em\u003e \u003cstrong\u003e70\u003c/strong\u003e, 9143-9153 (2022).\u003c/li\u003e\n\u003cli\u003eDavid, R., Tu\u0026ntilde;\u0026oacute;n, I. \u0026amp; Laage, D. Competing reaction mechanisms of peptide bond formation in water revealed by deep potential molecular dynamics and path sampling. \u003cem\u003eJournal of the American Chemical Society\u003c/em\u003e \u003cstrong\u003e146\u003c/strong\u003e, 14213-14224 (2024).\u003c/li\u003e\n\u003cli\u003ePavlovic, M., Plucinski, A., Zeininger, L. \u0026amp; Schmidt, B. V. K. J. Temperature sensitive water-in-water emulsions. \u003cem\u003eChemical Communications\u003c/em\u003e \u003cstrong\u003e56\u003c/strong\u003e, 6814-6817 (2020).\u003c/li\u003e\n\u003cli\u003eRizzi, G. P. Formation of strecker aldehydes from polyphenol-derived quinones and alpha-amino acids in a nonenzymic model system. \u003cem\u003eJ Agric Food Chem\u003c/em\u003e \u003cstrong\u003e54\u003c/strong\u003e, 1893-1897 (2006).\u003c/li\u003e\n\u003cli\u003eShi, Y.\u003cem\u003e et al.\u003c/em\u003e The antioxidant mechanism of Maillard reaction products in oil-in-water emulsion system. \u003cem\u003eFood Hydrocolloids\u003c/em\u003e \u003cstrong\u003e87\u003c/strong\u003e, 582-592 (2019).\u003c/li\u003e\n\u003cli\u003eGobert, J. \u0026amp; Glomb, M. A. Degradation of glucose: reinvestigation of reactive \u0026alpha;-dicarbonyl compounds. \u003cem\u003eJournal of Agricultural and Food Chemistry\u003c/em\u003e \u003cstrong\u003e57\u003c/strong\u003e, 8591-8597 (2009).\u003c/li\u003e\n\u003cli\u003eSauer, F.\u003cem\u003e et al.\u003c/em\u003e From amino acid mixtures to peptides in liquid sulphur dioxide on early Earth. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 7182 (2021).\u003c/li\u003e\n\u003cli\u003eSchwendinger, M. G. \u0026amp; Rode, B. M. Possible Role of Copper and Sodium Chloride in Prebiotic Evolution of Peptides. \u003cem\u003eAnalytical Sciences\u003c/em\u003e \u003cstrong\u003e5\u003c/strong\u003e, 411-414 (1989).\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Maillard reaction, Oxidation, Water-in-water emulsions, Metabolomics","lastPublishedDoi":"10.21203/rs.3.rs-7307960/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7307960/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eAll-aqueous emulsions represent a versatile platform for studying and controlling chemical reactions in foods and biological systems. Through the compartmentalization and segregation of reactants, distinct molecularly crowded microenvironments enable the generation of unique reaction products. We explored how spatial organization within all-aqueous emulsions composed of polyethylene glycol (PEG) and sodium sulfate (Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e) modulates the Maillard reaction and oxidation reactions between glucose and amino acids. Using untargeted metabolomics and molecular networking, we characterized the chemical diversity of reaction products formed when the reactants were either co-encapsulated within the droplet phase or distributed (segregated) between the two phases of the emulsions. Over 500 compounds were annotated across both systems, revealing distinct molecular profiles driven by reactant localization and phase partitioning. When the precursors were segregated (tryptophan and glucose), oxidation products such as β-carbolines and \u003cem\u003eN\u003c/em\u003e-formylkynurenine accumulated preferentially in the PEG phase. Conversely, when the reactants were co-encapsulated (asparagine and glucose) within Na\u003csub\u003e2\u003c/sub\u003eSO\u003csub\u003e4\u003c/sub\u003e droplets, enhanced formation of the Amadori products and dipeptides was observed, guided by phase-specific microenvironment. Our results demonstrate that the reactant location, in addition to time and temperature, plays a critical role in modulating food-relevant reactions, with a new framework for controlling the formation of glycation compounds via emulsion-based microreactors.\u003c/p\u003e","manuscriptTitle":"Spatial distribution of glucose and amino acids within all-aqueous emulsions directs the Maillard reaction and oxidation pathways","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-19 16:15:27","doi":"10.21203/rs.3.rs-7307960/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"communications-chemistry","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"commschem","sideBox":"Learn more about [Communications Chemistry](http://www.nature.com/commschem/)","snPcode":"","submissionUrl":"","title":"Communications Chemistry","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Communications Series","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"73d1c37c-1713-45a7-8f31-82f5ef5b4994","owner":[],"postedDate":"August 19th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":53001230,"name":"Physical sciences/Chemistry/Biochemistry/Metabolomics"},{"id":53001231,"name":"Physical sciences/Chemistry/Organic chemistry/Reaction mechanisms"}],"tags":[],"updatedAt":"2026-05-13T07:07:58+00:00","versionOfRecord":{"articleIdentity":"rs-7307960","link":"https://doi.org/10.1038/s42004-026-01951-6","journal":{"identity":"communications-chemistry","isVorOnly":false,"title":"Communications Chemistry"},"publishedOn":"2026-03-21 04:00:00","publishedOnDateReadable":"March 21st, 2026"},"versionCreatedAt":"2025-08-19 16:15:27","video":"","vorDoi":"10.1038/s42004-026-01951-6","vorDoiUrl":"https://doi.org/10.1038/s42004-026-01951-6","workflowStages":[]},"version":"v1","identity":"rs-7307960","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7307960","identity":"rs-7307960","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}