Thermal processing under mild acidic conditions modulates metabolite profile and emulsifying functionality of Moringa oleifera leaf proteins

Abstract

This study investigated the effects of thermal treatment (30, 55, and 85 °C) under mild acidic conditions (pH 4.5) during protein precipitation on structural characteristics, emulsifying properties, metabolite profile and protein composition of Moringa leaf protein (MoLP) from two cultivars (local Thai and PKM-1). Moderate heating at 55 °C resulted in the highest protein recovery, reaching 4.59% and 2.47% for Thai and PKM-1 cultivars, respectively. In contrast, treatment at 85 °C reduced protein content to 54.8% and 50.1%. Structural analyses indicated heat-induced unfolding, thiol–disulfide rearrangement, decreased thermal stability, and increased β-sheet formation, indicating protein aggregation. SDS-PAGE further confirmed protein alterations, particularly affecting RuBisCO-associated proteins. LC–MS/MS analysis demonstrated selective proteomic remodeling, with retention of metabolic enzymes and enrichment of stress-related proteins, whereas thermolabile photosynthetic proteins declined after heating. Metabolomic analysis showed cultivar-dependent responses, with PKM-1 exhibiting greater thermal resilience. Heat treatment at 55 °C substantially improved emulsion stability, increasing from 20 to 86 min in local Thai cultivar and from 119 to 208 min in PKM-1 cultivar. Overall, moderate thermal treatment at 55 °C provided the balance between protein recovery, structural integrity, and emulsifying functionality of MoLP, demonstrating the potential of MoLP as plant-based emulsifier.

Graphical abstract

Highlights

• •Thermal treatment under mild acid condition reshaped Moringa leaf protein structure.
• •Heat during protein precipitation enhanced metabolite release and transformation.
• •PKM-1 showed stronger induction of heat shock proteins and antioxidant enzymes.
• •Metabolite profiling revealed cultivar-dependent differences.
• •Moderate thermal treatment at 55 °C optimized the structure and function of MoLP.

1. Introduction

Leaf protein concentrates, also known as leaf protein extracts, are a promising source of sustainable, high-nutrition protein for human foods and animal feeds due to their favorable nutritional composition and low environmental footprint. Among potential leaf protein sources, Moringa oleifera leaves have received considerable attention because of its relatively high protein content (22–34% dry basis), balanced amino acid profile, and abundance of bioactive compounds, including phenolics, flavonoids, and phytosterols (Benhammouche et al., 2021; Cattan et al., 2022; Silva et al., 2020). In addition to its nutritional value, M. oleifera exhibits considerable intraspecific genetic and biochemical diversity among cultivars originating from different geographical regions. These cultivar-specific variations can influence protein yield, amino acid composition, structural characteristics, and the abundance of associated bioactive compounds, ultimately affecting the functional performance of the resulting protein concentrates (Buyel et al., 2025). Consequently, evaluating distinct cultivars is imperative to understanding and optimizing the downstream techno-functional and structural properties of M. oleifera leaf proteins. Functional studies have demonstrated that Moringa leaf proteins possess favorable solubility and emulsifying properties comparable to those of conventional plant proteins such as soy protein, highlighting their potential as sustainable ingredients for food applications (Cattan et al., 2022; Sultana, 2020). In addition to proteins, Moringa leaves contain substantial amounts of phenolic compounds and flavonoids that may remain associated with protein concentrates after extraction. These endogenous metabolites can interact with proteins through hydrogen bonding, hydrophobic interactions, and covalent modifications, thereby influencing protein structure, thermal stability, and functional properties. Thermal processing may further modify these protein–phenolic interactions, affecting both protein functionality and metabolite retention. Despite their potential importance, the simultaneous responses of proteins and associated metabolites during Moringa protein processing remain poorly understood (Buyel et al., 2025; Pop et al., 2022; Wang et al., 2025).

Alkaline extraction followed by isoelectric precipitation is mostly employed technique for recovering leaf proteins because of its simplicity, scalability, and effectiveness. Protein precipitation is commonly performed near pH 4 – pH 5, corresponding to the isoelectric region of many plant proteins, where protein solubility is minimized and recovery is maximized (Benhammouche et al., 2021; Cao et al., 2023; Nyobe et al., 2025). Consequently, proteins are exposed to mildly acidic conditions during processing, and in industrial operations this step may coincide with thermal treatment. The selection of pH 4.5 in the present study was therefore based on its practical relevance to leaf protein production. In addition to promoting protein recovery, mild acidic conditions can influence protein–phenolic interactions and modulate heat-induced conformational changes, making pH 4.5 an industrially relevant processing condition for evaluating structure–function relationships in Moringa leaf protein concentrates (Gonzalez-Ortega et al., 2023; Wang et al., 2025).

Heat treatment is one of the most common modifications applied during plant protein processing and can substantially affect protein solubility, structural integrity, and functionality. Thermal exposure may induce protein unfolding, aggregation, or conformational rearrangements, leading to exposure of hydrophobic residues, alterations in secondary structure, and formation of new intermolecular interactions (Wang et al., 2025). Previous studies on soy and other plant proteins have demonstrated that moderate thermal treatment can improve molecular flexibility and interfacial properties, thereby enhancing emulsifying performance, whereas excessive heating often promotes aggregation and functional deterioration (Liu et al., 2026). Similar mechanisms may occur in Moringa leaf proteins; however, current knowledge remains limited. Although several studies have investigated the effects of thermal processing on Moringa proteins, these investigations have primarily focused on structural or functional characteristics of isolated proteins or seed proteins (Bakwo Bassogog et al., 2022; Xie et al., 2025). Little information is available regarding how thermal treatment during mildly acidic precipitation simultaneously affects protein structure, associated metabolites, proteomic composition, and functional properties in leaf protein concentrates. Furthermore, whether these processing-induced responses differ among genetically distinct Moringa cultivars remains largely unknown despite the recognized diversity within Moringa germplasm (Buyel et al., 2025). Thus, a comprehensive understanding of the interactions among thermal processing, protein structure, metabolite composition, and functionality is still lacking. Therefore, the objective of this study was to investigate how thermal treatment under mild acidic conditions (pH 4.5) influences the structural, metabolomic, proteomic, and emulsifying properties of Moringa oleifera leaf protein concentrates. Unlike previous studies that primarily focused on isolated structural or functional aspects of Moringa proteins, the present work integrates physicochemical characterization with LC–MS/MS-based metabolomic and proteomic analyses to simultaneously evaluate processing-induced changes in both protein and metabolite fractions. In addition, two genetically distinct cultivars, a local Thai cultivar and PKM-1 (Indian cultivar), were compared to elucidate cultivar-dependent thermal responses. This multi-omics approach provides a more comprehensive understanding of how thermal processing influences the structure–composition–function relationships of Moringa leaf protein concentrates and contributes new insights for the development of cultivar-specific plant protein ingredients.

2. Materials and methods

2.1. Raw materials and chemicals

The mature green leaves from two Moringa oleifera cultivars—local Thai cultivar (called as Thai moringa) and PKM-1 (called as Indian moringa)—were harvested at the end of rainy season (October 2020) from Nan Province, Thailand. All chemicals and reagents used were of analytical grade except those for SDS-PAGE were of electrophoresis grade.

2.2. Sample preparation

The mature green leaves of both Thai moringa and PKM-1 were harvested and dried in a tray dryer at 40–45 °C until the moisture content reached approximately 6.0%. The dried leaves were ground and sieved using an 80-μm mesh sieve, then sealed in polyethylene plastic bags and stored in a dark place at room temperature (30 °C) for protein extraction.

2.2.1. Extraction of Moringa oleifera leaf (Moringa leaf) protein

The alkali-acid extraction process was adapted from Benhammouche et al. (2021) with modifications. Moringa leaf powder was suspended in distilled water at a 1:10 (w/v) ratio, and 1 M NaOH was used to adjust the pH to 9.0 while stirring for 1 h, with pH monitored and readjusted every 15 min as required. The solution was then centrifuged at 9000 xg for 10 min, yielding a brownish supernatant containing soluble proteins. The supernatant was adjusted to pH 4.5 using 1 M HCl to initiate isoelectric precipitation. Subsequent precipitation and thermal treatments were described in Section 2.2.2.

2.2.2. Heat treatment under mild acidic conditions

To evaluate the effects of thermal processing during isoelectric precipitation, protein suspensions adjusted to pH 4.5 were subjected to different precipitation conditions. For the control samples, the acidified suspensions of both cultivars were maintained at room temperature (30 °C) for 30 min for protein precipitation. For the thermal treatments, the acidified suspensions were first allowed to precipitate at room temperature for 30 min and were subsequently heated at either 55 °C or 85 °C for an additional 20 min. Following treatment, all suspensions were immediately cooled in an ice bath and centrifuged at 9000 ×g for 10 min. The resulting protein pellets were collected and resuspended in a minimal volume of distilled water to facilitate pH adjustment to 7.0, using 1 M NaOH. The resulting leaf protein concentrates were frozen, freeze-dried, vacuum-packed, and stored at −30 °C until further analysis. In total, six Moringa leaf protein (MoLP) samples were prepared and coded as follows: Thai–RT, Thai–55 °C, Thai–85 °C, Indian–RT, Indian–55 °C, and Indian–85 °C. Indian represents PKM-1 while Thai represents local Thai cultivar.

2.3. Chemical composition and some physicochemical properties of MoLP

2.3.1. Proximate analysis

Moisture content, crude fat, crude protein, ash, and crude fiber contents of the samples were determined using the methods described in AOAC (2019). The crude protein was calculated by multiplying % N with a factor of 6.25. Carbohydrate was calculated by differences.

2.3.2. Surface hydrophobicity (H₀)

Surface hydrophobicity was assessed using the fluorescence probe method with 1-anilino-8-naphthalene sulfate (ANS). MoLP stock solutions (1000 μg/mL) in 0.1 M sodium phosphate buffer at pH 7 were serially diluted to concentrations of 25–250 μg/mL. The fluorescent intensity of ANS-protein conjugates was monitored at 390 nm (excitation) and 470 nm (emission) using a 96-microplate reader (SpectraMax® iD3, Molecular devices, CA) as described in Sangsukiam and Duangmal (2022). The surface hydrophobicity of each sample was calculated from the slope of the relative fluorescence intensity to protein concentration (μg/mL).

2.3.3. Free amino group content

Free amino group of each moringa protein sample was determined using the reaction of free amino group with 2,4,6-trinitrobenzenesulfonic acid (TNBS) at 420 nm (Sangsukiam & Duangmal, 2017). The value was expressed as mg leucine per g sample.

2.3.4. Total sulfhydryl and free sulfhydryl group content

Moringa protein sulfhydryl group (total and free) was assayed using the Ellman's reagent (DTNB) method according to the method of Hao et al. (2024). The sulfhydryl group in terms of total and free in the sample and disulfide bond (-S-S-) were calculated as shown in eq. (1) and eq. (2), respectively and expressed as μmol per g protein.

$$-\mathrm{SH}=73.53\times {\mathrm{OD}}_{412\mathrm{nm}}\times \frac{\text{dilution factor}}{\text{sample concentration}}$$ (1)

$$\text{Disulfide bond}=\frac{\text{Total sulfhydryl group content}-\text{free sulfhydryl group content}}{2}$$ (2)

2.3.5. Total phenolic content (TPC)

TPC in free form of each sample was determined using Folin-Ciocalteu method according to Sritongtae et al. (2017) and expressed as mg gallic acid equivalent per g sample.

2.4. SDS-PAGE analysis of protein profiles

MoLP powder was dissolved in distilled water and mixed with sample buffer (60 mM Tris-HCl, pH 6.8, containing 40 g/L SDS, 200 g/L glycerol, 14.4 mM β-mercaptoethanol and 1 g/L bromophenol blue) at a ratio of 1:1 and boiled for 2 min to denature proteins. WIDE-VIEW™ prestained protein size marker III, MW 245 to 11 kDa (FUJIFILM Wako Pure Chemical Corporation, Japan), was used as markers. The aliquots containing 10 μg proteins from each sample were subjected to the reduced SDS-PAGE using a 12.5% (w/v) acrylamide separating gel. The electrophoresis was conducted at a constant current of 20 mA per gel using a Criterion™ Vertical Electrophoresis Cell (BioRad, CA.). After this process, the gel was stained with Coomassie Brilliant Blue R-250 solution and destained.

2.5. Structural characterization

2.5.1. Fourier transform infrared (FTIR)

FTIR spectra of samples were analyzed using a Fourier transform infrared spectrophotometer (JASCO FT-IR 6300, Japan) and recorded at a resolution of 4 cm⁻¹ in the range of 400 to 4000 cm⁻¹ (Karami, Butkinaree, et al., 2022).

2.5.2. Differential scanning calorimetry (DSC)

The thermal transition behavior of Moringa protein concentrates was evaluated using a differential scanning calorimeter (DSC 25, TA Instruments, DE), based on the method of Liu et al. (2022) with slight modifications. Approximately 2 mg of freeze-dried sample powder was combined with 6 μL of distilled water in a hermetically sealed aluminum crucible (TA Q20, Φ 5.4 × 2.6 mm). The samples were heated from 25 °C to 150 °C at a constant rate of 10 °C/min under a nitrogen atmosphere. The onset (To) and peak (Tp) temperatures of thermal transitions were recorded from the resulting thermograms to determine the denaturation behavior of proteins.

2.5.3. Thermogravimetric analysis (TGA)

The thermal stability and decomposition characteristics of MoLP were assessed by a thermogravimetric analyzer (TG 209F1 Libra, NETZSCH, Germany). Approximately 9–11 mg of freeze-dried sample was accurately weighed and placed in an open alumina crucible (85 μL). Samples were heated from 28 °C to 600 °C at a rate of 10 °C/min. The thermogravimetric (TG) and derivative thermogravimetric (DTG) curves were recorded. The main decomposition stages, temperatures and percentages of weight loss are also determined to compare thermal treatments.

2.6. Metabolites profiling

Metabolite profiling of MoLP was performed using a comparative metabolomics approach. Samples were prepared by dissolving each MoLP sample in 50% acetonitrile to a final volume of 1 mL and vortexed. The suspension was centrifuged at 13,000 ×g for 10 min at 4 °C, and the supernatant was collected and adjusted to 500 ppm with 0.1% formic acid in 50% acetonitrile. The metabolite extracts were analyzed using an ultra-high-performance liquid chromatography system (Vanquish UHPLC, Thermo Fisher Scientific) coupled with a high-resolution Quadrupole-Orbitrap mass spectrometer (Q Exactive HF-X, Thermo Fisher Scientific). Separation was achieved on an Ascentis® Express C18 column (4.6 × 150 mm, 2.7 μm; Merck) maintained at 40 °C. The mobile phases consisted of 0.1% formic acid +10 mmol/L ammonium formate in water (phase A) and 0.1% formic acid in acetonitrile (phase B), at a flow rate of 0.4 mL/min. Under a step gradient elution (starting from 1% phase B for 5 min. Then 95% phase B for the next 20 min. and 1% phase B for the last 5 min. Mass spectrometry data were acquired in both positive and negative ionizatiohn modes at a resolution of 120,000 (Full MS) and 30,000 (dd-MS²). The scan range was m/z 100–1500, with automatic gain control (AGC) = 1.0 × 10⁵ and stepped normalized collision energies of 20, 50, and 80 eV. The spray voltage was set to 3.5 kV (positive mode) and 2.5 kV (negative mode).

Compound identification was performed using Compound Discoverer 3.3 (Thermo Fisher Scientific) with comparison against the mzCloud, mzVault, and ChemSpider databases (Human Metabolome Database, LIPID MAPS, Natural Product Updates, and Phenol-Explorer). Only compounds with a mass tolerance of ±5 ppm were retained for analysis. Statistical evaluation included descriptive statistics, principal component analysis (PCA), and volcano plot analysis to identify significantly altered metabolites between treatments.

2.7. Protein identification

A total of 30 μg of protein, quantified by the Bradford assay kit (Thermo Fisher Scientific), was reduced with 10 mM dithiothreitol prepared in 10 mM ammonium bicarbonate at 65 °C for 20 min and alkylated with iodoacetamide at room temperature for 30 min. The alkylated proteins were cleaned using the Zeba Spin Desalting Columns, 7 K MWCO, 0.5 mL (Thermo Fisher Scientific). The recovered proteins were digested with MS-grade trypsin (Thermo Fisher Scientific) at an enzyme-to-protein ratio of 1:50 (w/w) and incubated at 37 °C for 16 h. Peptide mixtures were dried using a centrifugal concentrator (TOMY, Japan), reconstituted in 0.1% formic acid, and transferred to total recovery LC–MS vials (Waters™). Bovine serum albumin (0.2 mg; n = 2) was processed in parallel as a quality control to evaluate digestion efficiency and sample preparation consistency. Tryptic peptides (1.2 μg) were analyzed using an Orbitrap Exploris 240 mass spectrometer with an UltiMate 3000 LC system. Peptide were trapped on a C18 PepMapTM 100 column and separated on a C18 analytical column using a gradient 0.1% formic acid in water (A) and 95% acetonitrile containing 0.1% formic acid (B). Data dependent acquisition (DDA) was performed over an m/z range of 400 to 1600. The 15 most abundant precursor ions with charge states ranging from 2+ to 5+ were selected for fragmentation. Dynamic exclusion was applied with an exclusion duration of 15–18 s. Full MS spectra were collected a resolution of 120,000 with an automatic gain control (AGC) target of 3 × 10⁶ ions. MS/MS spectra were recorded once an AGC target of 1 × 10⁵ ions was reached with dynamic exclusion set to15 s. Raw data were processed using Proteome Discoverer 2.4 (Thermo Fisher Scientific), which SEQUEST, Percolator and Minora algorithms against the UniProtKB of Mesangiospermae database (17,463,292 sequences, accessed 01 October 2025). Search parameters included two missed tryptic cleavages, precursor mass tolerance of 10 ppm and fragment mass tolerance of 0.1 Da. Carbamidomethylation of cysteine residues, whereas methionine oxidation was included as a variable modification. Peptide and protein identifications were filtered using a false discovery rate (FDR) threshold of 1% Relative protein abundance ratios were normalized usingtotal peptide intensity.

2.8. Emulsifying properties of MoLP

Emulsions were prepared following Karami et al. (2022) with minor modifications. MoLPC (0.4 g) was dispersed in 20 mL of 0.1 M sodium phosphate buffer (pH 6, 7 and 8) and stirred for 15 min at 25 °C. Corn oil (10 mL) was then added and the mixture was homogenized at 20,000 rpm for 1 min.

2.8.1. Emulsifying capacity (EC)

EC was determined based on conductivity changes during oil addition. Oil was added (0.1 mL/s) to the emulsion under continuous homogenization and the phase inversion point was identified by a sudden drop in conductivity. EC was expressed as g of oil per g of crude protein.

2.8.2. Emulsifying activity index (EAI)

EAI was determined using a turbidimetric method (Karami et al., 2022). The emulsion was diluted with 0.1% SDS to a final protein concentration of 0.002 g/mL and absorbance at 500 nm was measured immediately using a 1 cm pathlength cuvette. EAI was calculated as shown in eq. 4.

2.8.3. Emulsifying stability index (ESI)

Determination of ESI was performed according to the turbidimetric method by Karami et al. (2022). Absorption of the emulsion diluted (0.002 g/ml) for ESI was measured at 500 nm on the 0th minute and 10th minute interval immediately after homogenization. ESI was calculated as shown in equation below.

$$\mathit{Turbidity}\left(T\right)=\frac{2.303\times A_{500}\times F}{l}$$ (3)

$$\mathit{EAI}\left(m^2/g\right)=\frac{2\times T}{\Phi \times C}$$ (4)

$$\mathit{ESI}\left(\mathit{mins}\right)=\frac{T\times ∆T}{∆T}$$ (5)

Where: A₅₀₀ = sample absorbance, F = dilution factor, l = cuvette pathlength (cm), Φ = oil volume fraction, C = weight of protein per unit volume of aqueous phase, t = time (mins).

2.9. Statistical analysis

All experiments were run in triplicates. The statistic difference analysis was conducted by ANOVA followed by Tukey's test and p ≤ 0.05 was considered statistically significant.

3. Results and discussion

3.1. Yield and proximate composition analysis of MoLP

Both cultivar type and extraction conditions significantly influenced protein recovery and proximate composition (Table 1). For both cultivar types, the highest yield was obtained at 55 °C, followed by slight reductions at 85 °C and room temperature (RT), indicating a temperature-dependent response during protein precipitation (Cheng et al., 2021). The results observed in this study were consistent with previous reports showing that moderate heating increase protein solubilization, while excessive heating causes denaturation, aggregation and reduced extractability (Bhattarai et al., 2024; Cao et al., 2023).

Table 1.

Extraction yield and proximate composition of Moringa leaf protein concentrates from PKM-1 and local Thai cultivars under different thermal treatments.

Condition	Yield (%db.)	Crude Protein (%)	Crude Fat (%)	Crude Fiber (%)	Ash (%)	Carbohydrates (%)⁎
Indian leaf powder	–	23.93 ± 0.45e	7.22 ± 0.48ab	40.0 ± 3.10a	0.73 ± 0.08a	28.8 ± 3.92a
Indian-55${}^{°}\mathit{C}$	2.47 ± 0.20bc	57.19 ± 0.88b	2.88 ± 0.02c	9.28 ± 0.07c	0.05 ± 0.01c	29.07 ± 1.08a
Indian-85 °C	2.30 ± 0.18c	50.10 ± 1.50c	2.50 ± 0.10d	8.00 ± 0.30c	0.04 ± 0.02c	31.16 ± 1.20a
Indian-RT	2.00 ± 0.20c	51.82 ± 1.73c	5.06 ± 0.90b	14.20 ± 0.16b	0.07 ± 0.01c	29.34 ± 1.70a
Thai leaf powder	–	29.01 ± 0.03d	8.54 ± 1.21a	46.96 ± 5.50a	0.21 ± 0.08b	19.74 ± 5.45b
Thai-55${}^{°}\mathit{C}$	4.59 ± 0.13a	61.48 ± 1.04a	3.52 ± 0.56c	11.40 ± 1.30bc	0.01 ± 0.04c	22.62 ± 2.53b
Thai-85 °C	4.20 ± 0.15a	54.80 ± 1.20b	3.00 ± 0.15c	9.80 ± 0.50bc	0.02 ± 0.01c	28.38 ± 1.60a
Thai-RT	3.16 ± 0.47b	60.57 ± 0.88ab	6.03 ± 0.13b	12.30 ± 1.28bc	0.05 ± 0.03c	20.94 ± 0.60b

Values are expressed as mean ± standard deviation (n = 3). Different superscript letters within the same column indicate significant differences (p < 0.05) according to Tukey's multiple comparison test.

“Indian” represents PKM-1 while “Thai” represents local Thai cultivar.

^⁎Carbohydrates were calculated by difference. RT: room temperature; db.: dry basis.

Across all heat treatments, the local Thai cultivar consistently showed higher protein yields than the PKM-1. This result correlated with the crude protein content of the local Thai leaf powder, which was relatively higher than that of PKM-1. The highest protein content among the leaf powders was found in Thai–55 °C, followed by Thai–RT, but the high-temperature treatment at 85 °C resulted in lower protein content in both cultivars. Thermal and mild acid treatments also lowered crude fat and fiber amounts in MoLP samples, particularly at 55 and 85 °C, most likely due to disruption of lipid–protein and fiber–protein interactions during the precipitation step. Moringa protein concentrates exhibited comparable reductions in non-protein fractions. The ash content remained low and virtually unchanged by heat treatment but carbohydrate content changed only minimally, showing relative compositional shifts rather than absolute variations (Tranquilino-Rodríguez et al., 2025).

3.2. Effects of thermal treatment on total phenolic content and protein functional groups of MoLP

Compared to RT condition, thermal treatment caused significant differences (p < 0.05) in total phenolic content in both cultivars (Table 2). Metabolite profiling, as explained in Section 3.5, showed increases and decreases in individual phenolic compounds (Fig. 3, Fig. 4; Tables S1 and S2), suggesting that thermal treatment influenced phenolic composition but not total extractable phenolics. Similar behavior has also been observed for plant proteins, where thermal treatment is known to influence phenolic-protein interactions without impacting TPC in free form values; these phenomena have been attributed to partial thermal degradation of phenolic compounds, phenolic-protein complexes, and reduced phenolic extractability from protein matrices (Gonzalez-Ortega et al., 2023; Karami, Akbari-adergani, & Duangmal, 2022; Sritongtae et al., 2022). The amount of free amino groups significantly increased following thermal treatment, reflecting heat-induced protein unfolding and partial cleavage of peptide bonds that enhance exposure of amino termini. Moderate heating at 55 °C resulted in intermediate values, suggesting limited conformational rearrangement compared with high-temperature processing. Surface hydrophobicity exhibited significant cultivar- and temperature-dependent changes. The highest hydrophobicity was observed in PKM-1 at room temperature, indicating a greater proportion of exposed hydrophobic residues in the native protein structure. Thermal treatments at 55 °C and 85 °C generally reduced surface hydrophobicity, particularly in the local Thai cultivar. This behavior suggests that, although thermal unfolding initially exposes buried hydrophobic residues, prolonged or more severe heating promotes intermolecular association through hydrophobic interactions, leading to aggregate formation and subsequent burial of hydrophobic domains within the protein matrix. Consequently, fewer hydrophobic sites remain accessible to the ANS probe used for hydrophobicity determination. This behavior differs from the findings of Xi et al. (2025) who reported increased surface hydrophobicity with increasing temperature in a purified Moringa allergen preparation. The discrepancy likely arises from fundamental differences in protein system complexity. Xie et al. (2025) investigated a relatively purified protein fraction, where unfolding was the dominant structural event and exposed hydrophobic residues remained accessible. In contrast, the present study examined crude Moringa leaf protein concentrates containing proteins, carbohydrates, phenolic compounds, and other endogenous constituents. Under these conditions, exposed hydrophobic regions can readily participate in protein–protein, protein–phenolic, and protein–carbohydrate interactions, accelerating aggregate formation and reducing the measurable surface hydrophobicity. This interpretation was supported by the concurrent decrease in protein band intensity using SDS-PAGE, reduced thermal stability observed by DSC, and changes in sulfhydryl/disulfide balance, all of which indicate aggregation-driven structural reorganization following thermal treatment. These results are consistent with the concept presented by Nyobe et al. (2025), who showed that the hydrophobicity of MoLP depends on intrinsic properties, such as amino acid composition and molecular conformation, as well as extrinsic factors, such as heat treatment. Nyobe et al. (2025) reported that moderate heating promotes exposure of hydrophobic residues through protein unfolding, whereas excessive heating favors aggregation and hydrophobic group burial. Accordingly, the high surface hydrophobicity observed in Indian samples at room temperature likely reflects naturally exposed hydrophobic regions, while the pronounced reduction at 85 °C in Thai samples indicates aggregation-dominated behavior.

Table 2.

Effects of thermal treatment on total phenolic content and protein functional groups of Moringa leaf protein concentrates from local Thai and PKM-1 cultivars.^⁎

Condition	Total phenolic compounds	Free amino group	Total sulfhydryl groupns (-SH)	Free sulfhydryl group (-SHfree)	Disulfide bond (-S-S-)	Surface Hydrophobicity
	(mg gallic acid /g sample)	(mmole/ g protein)	($\mu$mol/g protein)	(μmol/ g protein)	(μmol/ g protein)	(H0)
Indian- RT	277.93 ± 19.08a	245.98 ± 13.08c	1454.71 ± 28.20	108.93 ± 3.02ab	672.89 ± 15.30b	2445.7
Indian-55${}^{°}C$	205.70 ± 26.16b	271.44 ± 19.39b	1457.80 ± 16.40	91.05 ± 2.17c	683.37 ± 9.28b	1575.2
Indian-85${}^{°}C$	169.82 ± 25.86b	303.94 ± 22.21a	1532.05 ± 51.76	110.40 ± 3.17ab	710.83 ± 27.46a	1717.8
Thai-RT	255.07 ± 28.82a	236.64 ± 8.67c	1481.28 ± 11.11	116.81 ± 1.42a	682.24 ± 5.19b	1848.4
Thai-55 °C	181.21 ± 13.46b	264.82 ± 22.74b	1517.60 ± 32.62	109.56 ± 4.27ab	704.02 ± 16.54ab	1707.1
Thai-85 °C	198.53 ± 25.69b	299.57 ± 24.81a	1528.13 ± 28.95	86.89 ± 3.31c	720.62 ± 13.88a	264.35

Values are expressed as mean ± standard deviation (n = 3). Different superscript letters within the same column indicate significant differences (p < 0.05) according to Tukey's multiple comparison test while ns means non significant difference.

^⁎“Indian” represents PKM-1 while “Thai” represents local Thai cultivar.

Fig. 3.

Volcano plots of differential metabolites in local Thai Moringa leaf protein concentrate following thermal treatment at 55 °C and 85 °C compared with RT, analyzed in positive (a) and negative (b) electrospray ionization modes. Red and green dots indicate significantly up- and down-regulated metabolites, respectively (p < 0.05). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 4.

Volcano plots of differential metabolites in PKM-1 Moringa leaf protein concentrate following thermal treatment at 55 °C and 85 °C compared with RT, analyzed in positive (a) and negative (b) electrospray ionization modes. Red and green dots indicate significantly up- and down-regulated metabolites, respectively (p < 0.05). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Thermal treatment significantly affected both disulfide bond content and free sulfhydryl group in both local Thai and PKM-1 Moringa leaf protein concentrates in a cultivar- and temperature-dependent manner (Table 2), indicating that heat-induced unfolding exposed thiol residues previously buried within the protein interior and cause the formation of disulfide bond. Sulfhydryl group/disulfide bond interchange reactions occurred during heating under mild acid condition. These structural modifications are consistent with the differences observed in protein band intensities on SDS-PAGE (Fig. 1). In contrast, no significant differences were observed in total sulfhydryl content, indicating that the overall thiol pool, comprising cysteine residues in proteins and low-molecular-weight thiols such as glutathione, remained relatively stable across treatments in both cultivars. In contrast to the present findings, Xi et al. (2025) reported a decline in total sulfhydryl content with increasing temperature in a purified M. oleifera allergen, attributed to enhanced oxidation and disulfide formation. This discrepancy likely reflects differences in protein system complexity, as thermal unfolding in multi-protein concentrates may expose previously inaccessible thiol groups while simultaneously promoting thiol–disulfide interchange reactions.

Fig. 1.

SDS–PAGE profiles of Moringa leaf protein concentrates from local Thai and PKM-1 (Indian) cultivars subjected to different thermal treatments (RT, 55 °C, and 85 °C).

3.3. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis

Protein patterns of MoLP (local Thai and PKM-1) subjected to different treatments are presented in Fig. 1. In both cultivars, two major protein bands (52–53 kDa and 44–45 kDa) were consistently observed with different intensities depending on the treatment. The variation in band intensity of these two major bands may be associated with heat-induced protein denaturation, aggregation, and/or fragmentation, as reflected by changes in free amino groups and free sufhydryl content (Table 2). The observed banding patterns suggest that thermal treatment affected protein organization differently in two cultivars, with evidence of greater band disappearance in local Thai cultivar and increased high-molecular-weight aggregation tendencies in PKM-1. Moreover, a protein band with a MW of approximately 60 kDa was detected only in PKM-1 indicating differences between the protein profiles of the local Thai cultivar and PKM-1. Previous reports on Moringa leaf proteins identified dominant bands around 60, 56, 45, and 36 kDa (Cattan et al., 2022; Lu et al., 2025; Xi et al., 2025). The persistence of proteins in this range in all treatments suggested that these proteins are intrinsic to the leaf. Increasing the temperature across all treatments reduced band resolution and intensity, particularly at 85 °C. This suggested that there was unfolding and aggregation of proteins due to heat treatment rather than proteolysis. This interpretation was consistent with the FTIR results which showed a heat-dependent red shift of the amide I band, indicative of alterations in peptide bonding and secondary structure (Limpisophon et al., 2024), as well as with DSC data indicating decreased denaturation temperatures under higher thermal exposure. Collectively, these findings confirm the progressive loss of protein conformational stability upon heating (Kuang et al., 2023; Xi et al., 2025).

In the local Thai cultivar, the sample at room temperature showed the most complex protein profile, with two major protein bands (52–53 kDa and 44–45 kDa) and another major high MW band around 180–245 kDa accompanied by faint intensity bands with MW around 25–35 kDa. The presence of this high MW fraction likely reflects native oligomeric or aggregated protein assemblies that remain stable under non-thermal conditions but dissociate or further aggregate upon heating as previously reported for Moringa and other plant leaf proteins (Cattan et al., 2022; Xi et al., 2025).

Proteins in the 45–55 kDa range, typically associated with the large subunit of RuBisCO, although definitive protein identification requires proteomic analysis, were consistently detected in both cultivars. It is known that RuBisCO as well as other soluble proteins from leaves is highly temperature-sensitive at temperatures above 60–70 °C, undergoing unfolding and aggregation that reduce electrophoretic visibility (Kong et al., 2026; Law et al., 2001). Interestingly, PKM-1 leaf proteins retained clearer band profiles at 55 °C compared to the local Thai sample, suggesting cultivar dependent differences in thermal stability, potentially related to intrinsic protein composition and interactions with endogenous compounds. In brief, the SDS-PAGE analysis findings consistent with those from the FTIR and DSC analyses showed that processing at room temperature and moderate heating at 55 °C maintained better the original protein composition in Moringa leaf protein concentrates whereas treatment at 85 °C led to substantial structural destabilization and reduced protein extractability in both local Thai and PKM-1cultivars.

3.4. Structural characterization of MoLP under thermal and mild acid treatments

3.4.1. Fourier transform infrared (FTIR) spectra

The effect of heat and acid treatment on the structure of MoLP from local Thai and PKM-1 cultivars was investigated using FTIR spectrum shown in Fig. 2a. It can be seen as a similar pattern with typical characteristic peak of amide I spectrum (around 1630 cm⁻¹) and aliphatic secondary amine (around 3300 cm⁻¹). The amide A band (N—H stretching) and amide I regions were evaluated to monitor protein backbone vibrations. No change in character of N—H stretching band was observed while a clear heat-dependent shift of the amide I band toward lower wavenumbers was observed for both cultivars. The red shift of the amide I band suggested alterations in the protein backbone environment and peptide bond interactions induced by thermal treatment under mild acidic conditions (pH 4.5) indicating increased structural perturbation and possible partial hydrolysis at elevated temperatures (Limpisophon et al., 2024).

Fig. 2.

FTIR spectra (a), DSC chromatogram (b), and TGA chromatogram (c) of Moringa leaf protein concentrates from PKM-1 (Indian) and local Thai cultivars subjected to thermal precipitation at RT, 55 °C, and 85 °C.

To further evaluate structural modifications, the amide I region was deconvoluted and the relative proportions of α-helix, β-sheet, β-turn, and random coil structures were estimated (Table 3). In both cultivars, increasing thermal treatment resulted in a gradual decrease in α-helix content and a corresponding increase in β-sheet structures. The increase in β-sheet structures together with the reduction in random coil content suggested that thermal unfolding was followed by protein aggregation and reorganization into more ordered intermolecular β-sheet assemblies. These findings are consistent with the observed reductions in surface hydrophobicity and alterations in sulfhydryl/disulfide balance (Table 2), which indicate that protein unfolding and aggregation occurred simultaneously during heating. Similar transitions from α-helical structures to β-sheet-rich aggregates have been reported in thermally treated plant proteins and are commonly associated with enhanced intermolecular interactions and reduced conformational flexibility (Limpisophon et al., 2024; Wang et al., 2025).

Table 3.

Secondary structure composition and thermal transition parameters of Moringa leaf protein concentrates from local Thai and PKM-1 cultivars subjected to different thermal treatments, determined by FTIR Amide I band deconvolution and differential scanning calorimetry.^⁎

Samples	FTIR Secondary Structure (%)				DSC Thermal Parameters
	α-Helix (%)	β-sheet (%)	β-Turn (%)	Random coil (%)	Onset Temperature (${}^{°}C$)	Td (${}^{°}C$)	Endset Temperature (${}^{°}C$)	ΔH (J/g)
Thai-RT	16.3 ± 0.4	46.6 ± 0.5	18.9 ± 0.3	18.2 ± 0.3	61.5 ± 0.8	83.2 ± 0.4	105.0 ± 1.2	1.42 ± 0.08
Thai-55 °C	15.7 ± 0.3	48.3 ± 0.6	18.4 ± 0.4	17.7 ± 0.4	58.0 ± 0.7	81.7 ± 0.5	108.5 ± 1.5	2.15 ± 0.11
Thai-85 °C	15.1 ± 0.4	49.8 ± 0.5	18.2 ± 0.3	16.9 ± 0.3	54.5 ± 0.6	78.4 ± 0.3	102.0 ± 1.0	0.84 ± 0.04
Indian-RT	16.2 ± 0.3	46.9 ± 0.5	18.8 ± 0.4	18.0 ± 0.4	57.0 ± 0.8	80.3 ± 0.4	104.5 ± 1.1	1.88 ± 0.09
Indian-55 °C	15.6 ± 0.4	48.5 ± 0.6	18.2 ± 0.3	17.7 ± 0.3	62.0 ± 0.9	83.3 ± 0.5	112.0 ± 1.6	1.06 ± 0.05
Indian-85 °C	15.3 ± 0.3	49.3 ± 0.5	18.4 ± 0.3	17.0 ± 0.4	55.0 ± 0.7	78.4 ± 0.4	101.5 ± 1.2	0.62 ± 0.03

Values are expressed as mean ± standard deviation.

^⁎“Indian” represents PKM-1 while “Thai” represents local Thai cultivar.

3.4.2. Differential scanning calorimetry (DSC) thermograms

The DSC thermograms presented in Fig. 2b indicate that both local Thai and PKM-1 cultivars showed broad endothermic transitions within a comparable temperature range (≈78–83 °C). This broad nature of these endothermic peak may be attributed to the relatively low purity of the protein concentrate samples, which contained only 50–60% protein and 20–30% carbohydrates, resulting in overlapping thermal events from multiple protein fractions and associated macromolecules. Although the 85 °C-treated samples had already been exposed to thermal processing, residual endothermic transitions were still observed during DSC analysis. These transitions are unlikely to represent thermal decomposition, which generally occurs at substantially higher temperatures and is more commonly detected by thermogravimetric analysis. Instead, they likely reflect further unfolding of thermally resistant protein fractions, dissociation or rearrangement of heat-induced protein aggregates, and disruption of protein–carbohydrate or protein–phenolic interactions formed during processing. Similar residual thermal transitions have been reported in complex plant protein systems where heat treatment induces partial rather than complete denaturation (Kuang et al., 2023; Ma et al., 2017). In both cultivars, the denaturation temperature (T_d) decreased at the 85 °C treatment compared with the room temperature one (The control), indicating reduced thermal stability and a greater degree of structural modification induced by thermal processing.

The local Thai cultivar exhibited a decrease in T_d from 83.2 °C to 78.4 °C, while PKM-1 decreased from 80.3 °C to 78.4 °C. These reductions suggest that a substantial proportion of the native protein structure had already undergone conformational changes during thermal processing, resulting in proteins that required less thermal energy for subsequent denaturation during DSC analysis. This interpretation is consistent with the FTIR results showing heat-induced alterations in the amide I region and with SDS-PAGE results indicating reduced band intensity at elevated temperatures. The denaturation enthalpy (ΔH) further supported these structural changes (Table 3). In the local Thai cultivar, moderate heating at 55 °C increased ΔH from 1.42 to 2.15 J/g, suggesting the formation of partially reorganized protein structures that retained or enhanced molecular order. Such behavior has been reported in plant proteins undergoing controlled thermal treatment, where limited unfolding promotes structural rearrangement without extensive denaturation. In contrast, treatment at 85 °C reduced ΔH to 0.84 J/g, indicating a substantial loss of native structure and a reduced requirement for thermal energy during denaturation. A similar trend was observed for PKM-1, where ΔH progressively decreased from 1.88 J/g in the untreated sample to 1.06 and 0.62 J/g after treatment at 55 and 85 °C, respectively. The lower enthalpy values observed after severe heating suggest that protein unfolding and aggregation had already occurred during processing, leaving fewer intact ordered structures available for thermal transition during DSC measurement. Cultivar-specific thermal behavior was also evident. While moderate heating appeared to promote structural reorganization in the local Thai cultivar, PKM-1 exhibited a progressive decline in thermal stability with increasing treatment temperature. These differences likely reflect cultivar-dependent variations in protein composition, intermolecular bonding, and interactions with endogenous metabolites such as phenolic compounds. Similar cultivar-specific thermal behavior was reported in legume and plant protein systems, where differences in protein architecture and non-protein constituents influence denaturation behavior and thermal resistance (Wanniarachchi et al., 2026; Zink et al., 2024).

3.4.3. Thermogravimetric analysis (TGA) thermograms

All MoLP samples exhibited similar TGA thermograms (Fig. 2c), characterized by a major mass loss occurring at approximately 300 °C, which is associated with the thermal decomposition of organic components in the protein matrix. The comparable TGA profiles among samples indicated similar thermal stability, composition, and degradation mechanisms, suggesting that neither cultivar type nor processing conditions significantly affected the thermal degradation behavior of the MoLP. Comparable trends in DSC and TGA analyses were widely reported for plant protein systems, in which overlapping denaturation and degradation profiles reflect similar macromolecular organization and thermal stability of the protein matrix (Ricci et al., 2018).

3.5. Mass spectrometry–based profiling of MoLP under thermal and mild acid treatments

3.5.1. Thermal modulation of metabolite profiles in local Thai MoLP

Thermal processing markedly altered the metabolite composition of local Thai MoLP in a temperature- and mild acid–dependent manner, as revealed by volcano plots (Fig. 3) and supported by pairwise abundance ratios (55 °C/RT, 85 °C/RT, and 85 °C/55 °C) presented in Table S1. Severe heating at 85 °C under mild acid condition (pH 4.5) for 20 min. Resulted in a predominance of downregulated metabolites and a clear narrowing of the metabolite profile, highlighting degradation-dominated behavior. In particular, cordycepin, cholesteryl acetate, and 1-β-D-arabinofuranosylcytosine were significantly decreased at 85 °C (Fig. 3a, b). These results demonstrate that nucleoside-type and lipid-related metabolites are very sensitive to temperature. High temperature under mild acid condition (pH 4.5) may accelerate the degradation of these metabolites. Similar preferential losses of nucleosides and lipid-associated compounds under intense thermal treatments were reported in sterilized plant juices, where oxidation-prone functional groups and labile ester or glycosidic bonds undergo rapid degradation (Kumar et al., 2016; Li et al., 2025). Table S1 shows that the amounts of different phenolic and flavonoid-type compounds are substantially lower at 85 °C than at RT, which is consistent with the trends shown by the volcano plot. These losses are most likely caused by a combination of thermal degradation, oxidative processes, and disruption of protein-metabolite stabilization, all of which have been extensively described in heat-treated plant systems (Pop et al., 2022; Razzak et al., 2022). Comparable studies have shown that high-temperature treatments disproportionately affected phenylpropanoid-derived metabolites because they are susceptible to heat-induced oxidation and stress-related metabolic disturbances (Hýsková et al., 2021).

Interestingly, despite the overall decline in metabolite content at 85 °C, certain metabolites such as guanosine monophosphate and chamuvaritin were significantly enriched compared to RT (Fig. 3 and Table S1). This selective upregulation reflects studies on thermally sterilized plant products where extreme heat destroys strong matrix or protein linkages, allowing the release of deeply buried or conjugated metabolites while overall destruction predominates (Pop et al., 2022; Wang et al., 2024; Wang & Zhang, 2026). In the present study, such strong matrix or protein associations might be disrupted under these severe conditions. These types of responses demonstrate a shift from global preservation to compound specific liberation under severe thermal conditions. In contrast, moderate heating at 55 °C under mild acid conditions promoted a broad and consistent enrichment of bioactive metabolites, particularly phenolics, isoflavonoids and sphingolipid-related compounds, including gartryprostatin B, 3-ketosphingosine, bracteatin, pseudosindorin and genistein (Fig. 3 and Table S1). This enrichment pattern is the characteristic of controlled thermal treatments that enhance metabolite extractability without inducing extensive oxidative damage, as widely reported for moderately heated plant tissues and protein-rich matrices (Pop et al., 2022; Wang et al., 2024). In these conditions, moderate heating at 55 °C did not significantly speed up the disruption of matrix structure. Mechanistically, this effect is consistent with partial protein unfolding below the denaturation temperature of Moringa leaf proteins (∼71–75 °C), which reduces non-covalent protein metabolite interactions and enhances protein bound secondary metabolite accessibility. Similar thermal optima were described in Moringa leaves and Curcuma longa rhizomes, where moderate heating (∼55–60 °C) maximized secondary metabolite retention, while higher temperatures accelerated degradation (Lai et al., 2025; Wang et al., 2024; Zagorska & Jaworowska, 2024). Beyond their thermal behavior, these metabolites possess important biological activities. Genistein is a well-known isoflavone associated with antioxidant and anti-inflammatory effects, while flavonoids such as bracteatin and pinocembrin have been linked to radical-scavenging activity and cellular protection against oxidative stress. Therefore, their enrichment at 55 °C suggests that moderate thermal processing may improve not only metabolite extractability but also the potential bioactive value of the resulting Moringa protein concentrate (Hou et al., 2025).

Table S1 displays further compound class specific thermal responses. Aglycone flavonoids (e.g., genistein, bracteatin, pinocembrin and pseudosindorin) were especially vulnerable to degradation at 85 °C in mild acidic conditions, whereas glycosylated flavonoids and primary metabolites, such as vitexin and luteolin derivatives, remained relatively stable across treatments. This result was consistent with recent findings that glycosylation enhances flavonoid thermal stability by protecting the aglycone backbone from oxidative and structural degradation. Overall, these findings indicated that the indigenous Thai cultivar possessed limited resistance to high-temperature processing under mild acid conditions, with 55 °C indicating the best balance between metabolite release and preservation. Overall, Table S1 shows that gray-shaded metabolites represent compounds exhibiting significant temperature-dependent enrichment, whereas non-shaded metabolites display progressive declines with increasing thermal severity, as reflected by Thai 85 °C/55 °C, 85 °C/RT and 55 °C/RT ratios below unity. This pattern is closely aligned with volcano plot based metabolomic interpretations reported for thermally processed plant foods (Li et al., 2025; Wang et al., 2026).

3.5.2. Thermal effects on metabolite profiles of PKM-1 leaf protein

Thermal treatment has also significantly modulated the metabolite profile of PKM-1 leaf protein depending on temperature and compound (Fig. 4a, b and Table S2). In contrast with the local Thai cultivar, the PKM-1 has demonstrated greater thermal resistance, with both moderate and high temperatures under mild acid processing condition predominantly promoting selective metabolite enrichment rather than widespread degradation.

At 85 °C, only a small subset of metabolites, including A32390A, D-gluconic acid, 2-keto-3-deoxy-D-gluconic acid and N-desisopropyl delavirdine, were downregulated, indicating limited thermal sensitivity. By comparison, numerous phenolic-related and sugar derived metabolites most notably 4-hydroxycoumarin, pentahydroxycyclohexyl α-D-galactopyranoside, quinic acid, melleolide M and staphyloferrin B were significantly enriched relative to RT, as confirmed by volcano plots and abundance ratios (Table S2). A similar enrichment under extreme heat conditions has been found in thermally sterilized plant systems, where robust cellular matrices maintained metabolite integrity while allowing the release of tightly bound phenolics and conjugated sugars (Pop et al., 2022; Wang et al., 2024).

The most noticeable metabolic changes in PKM-1 were found at 55 °C under mild acid processing condition, which resulted in a wide enrichment of secondary metabolites, including coumarin- and limonoid-related compounds, hydroxyluteolin 7-O-HMG-glucoside, 11-deoxy-16,16-dimethyl-PGE₂, demethylwedelolactone, tetrahydroxycyclohexanecarboxylic acid, repensol 3,7,9-trihydroxycoumestan, and fumiquinazoline F. Only a small percentage of organic acids and amino acids decreased. This type of reaction is similar to that found in moderately processed plant foods. Regulated heat input increased metabolites extractability and reduced degradation (Wang et al., 2024; Zhang et al., 2025). Several of the enriched metabolites have recognized functional significance. Quinic acid derivatives contribute to antioxidant capacity and serve as intermediates in phenylpropanoid metabolism, whereas hydroxycoumarin and coumestan derivatives such as repensol have been associated with antioxidant and anti-inflammatory activities. Consequently, the enhanced abundance of these compounds following thermal treatment may contribute to preservation of the functional and nutraceutical potential of PKM-1 leaf protein concentrates (Choo et al., 2025; Ninkuu et al., 2025).

These findings indicated cultivar-specific variation in thermal response. While the local Thai cultivar indicated degradation-dominated behavior at high thermal treatment, the PKM-1 cultivar maintained metabolic integrity and benefited from both moderate and severe heating under mild acid processing conditions by the release of certain metabolites. Table S2 further confirms that gray-shaded metabolites exhibited significant temperature-dependent increases, whereas non-shaded metabolites declined with increasing thermal severity, as indicated by Indian 85 °C/55 °C, 85 °C/RT, and 55 °C/RT ratios below unity. These results underscore the importance of cultivar tailored thermal under mild acid processing conditions strategies for optimizing metabolite retention and functional potential in MoLP.

3.6. Protein identification of MoLP under thermal and mild acid treatments

LC-MS/MS results showed that temperature- and mild-acid-dependent changes occurred significantly in the protein profiles of MoLP from both local Thai and PKM-1 cultivars after thermal treatment at 85 °C (Table S3 and S4). Since 85 °C is the most severe condition, it is worthwhile comparing the proteomic profiles at this temperature with that at RT. In the local Thai cultivar, several highly abundant chloroplastic and metabolic proteins—including ribulose-1,5-bisphosphate carboxylase/oxygenase large chain (RuBisCO), RuBisCO activase, transketolase, and ATP synthase subunit β were retained after heating, indicating substantial thermal stability. The persistence of these proteins is consistent with their known structural robustness and high cellular abundance, which frequently provides resistance to heat denaturation (Zhu et al., 2013). ATP synthase subunit α had a genotype- and isoform-specific responses. The subunit α was lost at 85 °C for local Thai samples while a chloroplastic isoform accumulated under high temperature. This selective behavior suggests subunit specific heat sensitivity within the ATP synthase complex. The differential stability of ATP synthase subunits under heat stress has been related to changes in folding kinetics, membrane association and aggregation susceptibility (Guo et al., 2017).

Several enzymes involved in central carbon metabolism, including glyceraldehyde-3-phosphate dehydrogenase (GAPDH), malate dehydrogenase, fructokinase, and phosphopyruvate hydratase, exhibited increased sequence coverage and/or peptide detection following the 85 °C treatment. Importantly, this increase was not observed uniformly across all identified proteins. While several metabolic enzymes showed enhanced coverage, other proteins, including ATP synthase α subunit, heat shock protein 70, and photosystem II D2 protein, exhibited reduced detection or were no longer identified after heating. This selective response suggests that the observed changes are unlikely to result from analytical artifacts or contributions from digestion-enzyme-derived peptides. Instead, the enhanced sequence coverage likely reflects heat-induced structural modifications that improved proteolytic accessibility and peptide recovery during LC-MS/MS analysis. Thermal unfolding can expose previously buried cleavage sites and disrupt protein complexes, increasing the number of detectable peptides generated from specific proteins without necessarily increasing their abundance. Similar improvements in peptide detectability following thermal denaturation have been reported in plant proteomic studies and are generally attributed to enhanced extractability and accessibility of protein regions rather than de novo protein synthesis (Zhu et al., 2013). The concurrent loss of several heat-sensitive proteins further supports the conclusion that thermal treatment induced selective structural remodeling of the proteome rather than a generalized increase in peptide generation (Liu et al., 2013).

Antioxidant-related proteins, such as catalase and thioredoxin-dependent peroxiredoxin, were found to be more abundant in heated local Thai cultivar samples, suggesting activation or stabilization of oxidative stress defense mechanisms under thermal treatment (Zhu et al., 2013). Heat shock proteins (HSPs) displayed compartment-specific responses. In the local Thai cultivar, cytosolic HSP70 and HSP83-like proteins were reduced, whereas mitochondrial and chloroplastic HSP90 isoforms were maintained or enriched. Chaperonin CPN60–2 was detected only after treatment at 85 °C, indicating organelle-targeted proteostasis under severe heat stress (Guo et al., 2017). Photosynthesis-related proteins exhibited marked thermal sensitivity. Several photosystem II (PSII) core components, including the D2 protein, were reduced or destroyed at 85°$C$, whereas the 33 kDa oxygen-evolving complex subunit remained detectable. This pattern suggests preferential destabilization of membrane-associated reaction center proteins compared with peripheral or luminal components (Liu et al., 2013).

In the PKM-1 cultivar, stress-associated proteins-including HSP90–5, Clp R domain-containing proteins, catalase, and small heat shock proteins-were prominently detected after heating at 85°$C$ (Table S4). Core metabolic enzymes such as transketolase, GAPDH, malate dehydrogenase, and RuBisCO-associated proteins were largely preserved, indicating comparatively greater proteomic stability than the local Thai cultivar (Guo et al., 2017). Despite cultivar-dependent differences, ATP synthase subunit α was consistently absent at 85 °C in both cultivars, underscoring its intrinsic sensitivity to combined thermal and mild acid conditions and suggesting its potential as a marker of processing-induced proteomic remodeling.

Therefore, thermal treatment at 85 °C under mild acid conditions selectively remodeled the Moringa leaf proteome by favoring the retention or enhanced detectability of proteins associated with stress tolerance and oxidative defense while reducing thermolabile photosynthetic and energy-conversion proteins. The appearance or enrichment of heat shock proteins (HSP90, stromal HSPs, HSP15-like proteins and CPN60–2), together with the retention of catalase and thioredoxin-dependent peroxiredoxin, suggests preferential preservation of proteostasis- and antioxidant-related proteins during thermal processing (Dündar et al., 2025; Raza et al., 2025). In contrast, the disappearance or reduced detection of ATP synthase α subunits and photosystem II D2 proteins demonstrates the greater thermal sensitivity of photosynthetic reaction-center and energy-transduction components. Notably, PKM-1 showed a greater retention and induction of stress-associated proteins than the local Thai cultivar, indicating genotype-dependent differences in proteomic stability under processing conditions. Collectively, these findings demonstrate that thermal treatment does not affect all proteins uniformly but selectively reshapes the detectable proteome according to protein structure, cellular localization, and intrinsic thermal stability. The stronger retention and appearance of stress-associated proteins in PKM-1 further suggest genotype-dependent differences in proteomic stability during thermal processing. It should be noted that the changes observed by LC–MS/MS do not necessarily reflect changes in protein abundance alone. Thermal treatment can induce partial unfolding of protein structures, exposing previously buried proteolytic cleavage sites and increasing protein accessibility during enzymatic digestion. As a result, proteins such as GAPDH, transketolase, and malate dehydrogenase exhibited higher sequence coverage and peptide recovery after heating, likely due to improved extractability and enhanced proteolytic accessibility rather than increased protein concentration (Kong et al., 2026). Conversely, thermal denaturation may also promote protein aggregation and insolubilization, particularly for membrane-associated and photosynthetic proteins, thereby reducing their extraction efficiency and detectability. This phenomenon likely contributed to the reduced detection or disappearance of proteins such as ATP synthase α subunits and photosystem II D2 proteins following treatment at 85 °C (Chen et al., 2025). These proteomic observations are consistent with the structural analyses obtained in the present study. FTIR results demonstrated alterations in protein secondary structure, DSC revealed reduced thermal stability, and SDS-PAGE showed decreased protein band intensity and loss of specific protein fractions after heating. Collectively, these findings indicate that thermal processing primarily altered protein conformation, accessibility, and extractability rather than causing extensive changes in the overall protein composition. Such structural modifications are expected to influence functional properties by altering protein–protein and protein–water interactions, surface hydrophobicity, and interfacial behavior, which ultimately affect the emulsifying and stabilization capacity of Moringa leaf protein concentrates.

3.7. Emulsifying properties of MoLP

The emulsifying properties of MoLP concentrates were significantly affected by both cultivar type and heat treatment (Table 4). Among all samples, the PKM-1 control (Indian-RT) exhibited the highest emulsifying capacity (EC) and emulsifying activity (EAI), which may be attributed to its relatively higher surface hydrophobicity (Table 2). Moderate exposure of hydrophobic groups enhances protein affinity for the oil–water interface, facilitating rapid adsorption and reducing interfacial tension during emulsion formation (Liu et al., 2026). However, hydrophobicity alone does not determine emulsifying performance. Effective emulsification requires a balance between surface hydrophobicity, solubility, and molecular flexibility, which collectively govern the diffusion and rearrangement of proteins at the interface (Karami et al., 2022). Thermal treatment at 55 °C reduced EC and EAI in the PKM-1 samples, likely due to partial protein aggregation and reduced molecular mobility, which limited interfacial adsorption. Similar decreases in emulsifying activity following moderate heat treatment have been reported for plant proteins where protein–protein interactions restrict conformational flexibility and diffusion to the oil–water interface (Guo et al., 2017). In contrast, the 55 °C treatment markedly improved emulsion stability, particularly in the PKM-1 cultivar at pH 8 and in the Thai cultivar at pH 6. This improvement suggests that moderate thermal unfolding promoted stronger intermolecular interactions and the formation of more cohesive and viscoelastic interfacial films surrounding oil droplets. Such films provide enhanced steric stabilization and resistance to droplet coalescence, thereby improving long-term emulsion stability (Guo et al., 2024; Lam & Nickerson, 2013). Interestingly, the Thai cultivar exhibited improved EAI following treatment at 55 °C compared with the untreated control, indicating that moderate structural modification enhanced its interfacial functionality. This cultivar-dependent response is consistent with the observed differences in protein composition, secondary structure, and surface hydrophobicity between Thai and PKM-1 samples. Previous studies have similarly demonstrated that moderate heat-induced unfolding can expose functional groups that enhance emulsifying performance, whereas excessive heating promotes aggregation and loss of functionality (Guo et al., 2024). These findings also agree with previous reports on Moringa proteins, where favorable emulsifying properties have been attributed to their higher solubility and balanced hydrophobic-hydrophilic nature (Cattan et al., 2022). Notably, the relationship between hydrophobicity and emulsifying performance is not linear. While moderate surface hydrophobicity facilitates adsorption and improves EAI, excessive exposure of hydrophobic regions can increase aggregation and reduce protein solubility, ultimately impairing emulsification efficiency (Lam & Nickerson, 2013). Therefore, the superior emulsion stability observed after moderate thermal treatment likely reflects an optimal balance between interfacial adsorption and the formation of mechanically robust interfacial films.

Table 4.

Effect of thermal treatment on emulsifying properties of Moringa leaf protein concentrates from PKM-1 and local Thai cultivars.

Samples⁎	Emulsifying Capacity (g oil/g of crude protein)			Emulsifying Stability (min)			Emulsifying Activity (m2 of oil interface/g of crude protein)
	pH 6	pH 7	pH 8	pH 6	pH 7	pH 8	pH 6	pH 7	pH 8
Indian-RT	243.3 ± 8.1a	257.9 ± 5.7a	252.0 ± 3.6a	58.0 ± 14.0b	23.0 ± 0.7b	119.0 ± 17.0b	82.1 ± 3.3a	96.7 ± 1.0a	108.9 ± 1.9a
Indian-55${}^{°}\mathit{C}$	222.6 ± 0.6b	233.6 ± 4.6b	225.8 ± 3.1b	16.0 ± 1.0d	27.0 ± 3.0a	208.0 ± 12.0a	63.2 ± 1.6b	65.0 ± 1.9c	64.9 ± 0.2c
Indian-85${}^{°}\mathit{C}$	205.4 ± 3.2c	211.0 ± 1.9c	207.6 ± 2.8c	14.0 ± 1.2d	22.0 ± 0.5b	95.0 ± 10.3c	49.1 ± 1.3d	48.2 ± 1.5d	50.0 ± 1.7d
Thai-RT	188.7 ± 4.0d	195.5 ± 5.0d	173.1 ± 7.3e	20.0 ± 0.4c	21.0 ± 0.3c	17.0 ± 0.4e	56.8 ± 1.0c	63.0 ± 3.4c	67.1 ± 0.5c
Thai-55 °C	211.6 ± 3.7b	213.6 ± 1.3c	214.3 ± 12.6b	86.0 ± 4.0a	26.0 ± 1.0a	23.0 ± 1.0d	58.4 ± 0.2c	79.7 ± 1.0b	82.2 ± 2.4b
Thai-85 °C	182.3 ± 2.9d	190.0 ± 2.1d	185.7 ± 3.5d	18.0 ± 0.8c	20.2 ± 0.6c	18.5 ± 0.9e	46.5 ± 1.6d	47.2 ± 1.8d	48.9 ± 1.2d

^⁎“Indian” represents PKM-1 while “Thai” represents local Thai cultivar. Values presented as mean ± standard deviation (n = 3) Different letters down the column for each emulsifying property shows mean values that are significantly different at p < 0.05 using Tukey mean comparison.

Across the pH range of 6–8, only minor changes in EC were observed, suggesting that the ability of MoLP to adsorb at the oil–water interface was relatively unaffected within this pH range. In contrast, emulsion stability exhibited greater pH dependence, with improved stability observed at pH 6 for the Thai cultivar and at pH 8 for PKM-1. These differences may be related to pH-induced changes in protein charge distribution, electrostatic repulsion, and solubility, all of which influence protein–protein interactions and interfacial film formation. The relatively lower performance observed near pH 7 for several samples may indicate reduced electrostatic repulsion, leading to increased aggregation and destabilization of the emulsion system (Lam & Nickerson, 2013). Overall, the results demonstrate that thermal treatment was the primary factor governing emulsifying behavior, while pH exerted a secondary but important influence through modulation of protein interactions and interfacial stability.

4. Conclusions

Thermal treatment under mild acidic conditions significantly influenced the structural characteristics, metabolite composition, proteomic profile, and emulsifying properties of Moringa oleifera leaf protein concentrates in a cultivar-dependent manner. Moderate heating at 55 °C provided the most favorable balance between protein recovery, structural integrity, metabolite retention, and emulsion stability, whereas treatment at 85 °C promoted protein unfolding, aggregation, and cultivar-specific metabolite changes. Proteomic analysis indicated that thermal processing primarily affected protein extractability and detectability through structural remodeling rather than changes in protein composition. Overall, the results demonstrate that both cultivar selection and processing conditions are critical determinants of the structure–function relationships of Moringa leaf proteins and should be optimized together to maximize functionality. A limitation of this study is that only two Moringa cultivars and a limited range of thermal processing conditions were evaluated. Furthermore, the biological functionality and digestive behavior of the modified proteins were not investigated. Future studies should examine a broader range of genotypes and processing parameters, while also evaluating protein digestibility, bioactive peptide generation, and performance in complex food systems. Such investigations would provide a more comprehensive understanding of the potential of Moringa leaf proteins as sustainable plant-based food ingredients.

CRediT authorship contribution statement

Zohreh Karami: Writing – original draft, Visualization, Methodology, Investigation, Formal analysis. Thasanporn Sangsukiam: Visualization, Methodology, Formal analysis. Annelise Halafihi: Methodology. Yodying Yingchutrakul: Writing – review & editing, Resources. Yi Zhang: Writing – review & editing, Conceptualization. Chutikarn Butkinaree: Writing – review & editing, Visualization, Resources, Methodology, Conceptualization. Kiattisak Duangmal: Writing – review & editing, Writing – original draft, Supervision, Resources, Project administration, Funding acquisition, Conceptualization.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Kiattisak Duangmal reports financial support was provided by The National Research Council of Thailand. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

Electronic Supplementary material.

Data availability

Data will be made available on request.