Maillard reaction of heme protein hydrolysate and xylose: A strategy for improving protein solubility, umami taste, and texture in plant-based patties

Abstract

The Maillard reaction generates key flavor compounds that enhance the palatability and overall quality of plant-based meats (PBMs). We utilized the Maillard reaction on heme protein hydrolysate (HH) to improve functionality and diversify flavor while maintaining its meat-like bloody taste. The optimized condition was an HH–xylose ratio of 2:1, heated for 6 h at 80 °C. Three samples were compared: HH, HH–xylose mixture (HHX), and HH–xylose Maillard reaction product (MHHX). MHHX exhibited higher grafting degree and browning intensity, along with stable structural modifications. MHHX improved protein solubility, emulsifying capacity, and thermal stability. Changes in MHHX amino acid composition contributed to enhanced umami and reduced bitterness. In PBMs, 1% MHHX increased hardness from 12.54 to 15.95 N while enhancing umami and maintaining bloody taste. Taken together, Maillard reaction modified HH structure and improved its physicochemical and sensory properties while maintaining bloody taste, making MHHX a suitable PBM additive.

Graphical abstract

Highlights

• •Heme protein hydrolysate–xylose effectively yielded Maillard reaction product (MRP).
• •Heme protein hydrolysate–xylose MRP (MHHX) improved solubility and emulsification.
• •Maillard reaction altered free amino acids to enhance umami and reduce bitterness.
• •MHHX improved the texture and overall taste profile of plant-based meat.
• •Maillard reaction retained heme-derived bloody taste while improving palatability.

1. Introduction

Plant-based meats (PBMs) are typically prepared from plant proteins, water, oils, and polysaccharides to mimic traditional meat. However, plant proteins often have an undesirable beany or bitter taste due to compounds such as hexanal and phenols, requiring the masking of these off-flavors to create a more meaty taste (Karolkowski et al., 2023; Su et al., 2024). Consequently, recent research has focused on replicating the taste, flavor, appearance, and texture of real meat. For instance, Kothuri et al. (2024) improved the texture, oxidative stability, and juiciness with an emulsion gel of chia flour, oat bran, and soy protein in PBMs. Similarly, Rasul et al. (2024) showed that adding date fruit fibers enhanced structural and thermal stability while creating a beef-like reddish-brown color. While synthetic additives can also improve flavor, this method may produce harmful compounds such as benzopyrene and polycyclic hydrocarbons, posing potential health risks (Variyar & Mishra, 2024). Despite ongoing efforts to find effective food additives that improve flavor and texture, a single substance with diverse functional properties that significantly enhances PBM quality has yet to be optimized.

The Maillard reaction is a nonenzymatic chemical reaction in which reducing sugars react with proteins, peptides, and amino acids to form various Maillard reaction products (MRPs). The Maillard reaction generally occurs through early, intermediate, and final phases. During the early phase, protein degradation generates free amino acids and other reactive nitrogenous compounds that can participate in subsequent glycation reactions. The carbonyl group of reducing sugar reacts with the amino group of amino acid to form a Schiff base, which subsequently undergoes Amadori rearrangement to produce Amadori or Heyns products (Liu et al., 2022). During the intermediate phase, Amadori or Heyns products are degraded into α-dicarbonyl compounds, which serve as precursors for flavor-active substances such as pyrazines and furans (Shakoor et al., 2022). In the final phase, the intermediate products undergo polymerization to form high-molecular-weight brown pigments, including melanoidins, along with polyphenolic compounds that contribute to antioxidant activity (Shakoor et al., 2022). However, excessive progression of the Maillard reaction can lead to the formation of advanced glycation end products (AGEs), which are associated with an increased risk of diabetes, cardiovascular disease, and cancer (Liu et al., 2020). One strategy to suppress AGE formation is to use protein hydrolysates as reactants in the Maillard reaction (Arasteh et al., 2023). This strategy enhances the reactivity of the Maillard reaction and promotes the formation of MRPs (Fu et al., 2020).

The use of monosaccharides rather than oligo- or polysaccharides has been shown to be more effective in promoting MRP formation. A previous study demonstrated that glucose induces a more pronounced Maillard reaction with soy protein than with chitosan oligosaccharide, resulting in a greater color change and elevated lysine/arginine content (Xu et al., 2019). Monosaccharide type also influences reaction extent and flavor formation. Pentoses, including xylose, are generally more reactive than hexoses such as glucose and fructose and have been shown to promote Maillard reaction progression more effectively in protein hydrolysate-based reaction (Jeon et al., 2025). In addition, xylose is also widely used in protein hydrolysate-based Maillard reaction to generate meat-like flavor compounds, including furans, furanthiols, and pyrazines (Sun et al., 2023). Collectively, xylose is a suitable reducing sugar for enhancing Maillard reactivity and meat-like flavor development in protein hydrolysate-based reactions.

Plant protein-based MRPs have been investigated for improving the flavor and structure of PBMs. In 3D-printable PBMs formulated with mung bean protein, xylose promoted the Maillard reaction, leading to changes in rheological and textural properties (Ma et al., 2026). Soy protein-derived MRPs have been used as emulsion-based fat substitutes in PBMs, improving textural properties without significantly affecting color (Yuan et al., 2026). In addition, incorporating xylose and sulfur-containing amino acids during high-moisture extrusion promoted the Maillard reaction, thereby enhancing fibrous structure and textural properties while generating roasted aromas (Ma et al., 2026). These findings highlight the potential of MRPs to improve overall PBM quality. However, reliance on plant-derived components limits the reproduction of meat-specific sensory attributes, including bloody and metallic tastes.

Heme proteins are globular proteins found in muscle tissue that are capable of reversible oxygen binding through an iron atom coordinated within the heme group. Recent advances in recombinant technology have enabled the production of heme proteins in yeast, which are used as food additives to enhance the meaty flavor profile of plant-based products (Carlsson et al., 2020). Given their distinct bloody and metallic taste, which is primarily attributed to their iron content, heme proteins play a key role in imparting a meat-like flavor to PBMs (Gerhard, 2020). These flavors are perceived through interactions between iron ions in heme and T2R-type taste receptors in the human oral cavity. Iron in heme proteins is the main source of flavor, whereas amino acids influence taste by binding to receptors (Gerhard, 2020). Given the diverse amino acid composition of the heme protein, we hypothesized that the Maillard reaction could enhance its flavor functionality. Compared with plant protein-based MRPs, heme protein-derived MRPs may offer a distinct advantage by combining bloody and metallic taste attributes with Maillard reaction-induced umami enhancement and improved functionality. Accordingly, the resulting MRPs may serve as multifunctional additives in PBMs by improving flavor, color, and texture while retaining characteristic bloody taste.

Although MRPs are increasingly recognized for their potential to enhance the quality of various foods, research in this area, especially concerning PBMs, remains limited. Notably, almost no studies explore the Maillard reaction in the context of heme proteins. Therefore, our study aimed to improve the taste attributes and functionality of heme protein to enhance the overall quality of PBMs. In our study, we used heme protein hydrolysates (HH) to produce MRPs while simultaneously limiting AGE formation. The optimal Maillard reaction conditions for HH and xylose were established. The generated MRPs were analyzed for physicochemical and sensory characteristics to evaluate their potential as functional additives for PBM formulations.

2. Materials and methods

2.1. Materials

Heme protein was provided by Professor Pil Kim at the Catholic University of Korea (Seoul, Korea). Heme proteins were isolated from Corynebacterium glutamicum. C. glutamicum was cultured in YS medium containing glucose (40 g/L), yeast extract (10 g/L), soytone (10 g/L), MgSO₄ (1 g/L), (NH₄)₂SO₄ (5 g/L), K₂HPO₄ (1.5 g/L), NaH₂PO₄ (0.5 g/L), CaCl₂ (0.4 g/L), and FeSO₄ (0.02 g/L). The medium was supplemented with kanamycin at 25 mg/L, and cultivation was performed at 30 °C and 220 rpm for 96 h. After cultivation, heme protein was purified according to the method described by Yang and Kim (2019). Briefly, 1 mL of fermentation broth was centrifuged at 12,000×g for 10 min at 4 °C to collect the cells. The cell pellet was resuspended in 1 mL of 1 N NaOH and transferred to a 1.5 mL tube containing 0.2 g of acid-washed glass beads (212–300 μm; Sigma-Aldrich, USA). Cell disruption was performed using a bead-beater (Minibeadbeater-16; BioSpec Products, USA) for 1 min and repeated five times. The lysate was centrifuged again under the same conditions, and the resulting supernatant was lyophilized to purify the heme protein. Alcalase and Flavourzyme, used for enzymatic hydrolysis, were purchased from Daejongzymes Co. (Seoul, Korea). Xylose served as the reducing sugar in the Maillard reaction and was supplied by Junsei Chemical Co., Ltd. (Tokyo, Japan). For the plant-based patty formulation, methyl cellulose (LOTTE Fine Chemical Co., Ulsan, Korea) and textured pea protein (TPP; Sotexpro Co., Paris, France) were used as ingredients. κ-Carrageenan and isolated pea protein were obtained from ESfood Co. (Gyeonggi, Korea) and Hyangrim Co. (Seoul, Korea), respectively. Starch and canola oil were purchased from Daehan Flour Mills Co. (Seoul, Korea) and CJ Cheiljedang (Seoul, Korea), respectively. Beet powder and salt were obtained from ESfood Co. (Gyeonggi, Korea), whereas coconut oil was purchased from Palmtop Vegeoil Products Sdn. Bhd. (Johor, Malaysia). As analytical reagents, O-phthalaldehyde (OPA; Sigma–Aldrich, St. Louis, MO, USA), urea (Duksan General Science, Seoul, Korea), sodium tetraborate (Sigma–Aldrich), β-mercaptoethanol (VWR Amresco, Radnor, PA, USA), sodium dodecyl sulfate (SDS; VWR Amresco), NaCl (Daejung, Busan, Korea), 1,10-phenanthroline (Tokyo Chemical Industry Co., Tokyo, Japan), sodium nitrite (Duksan General Science), FeSO₄ (Sigma-Aldrich), and hydroxylamine hydrochloride (Sigma-Aldrich) were used.

2.2. Preparation of heme protein hydrolysate

HH was prepared using sequential enzymatic hydrolysis with Alcalase and Flavourzyme (Liu et al., 2025). Heme protein was dispersed in distilled water at 5% (w/w) and held at 55 °C for 2 h to induce protein denaturation. The pH was adjusted to 8.5, and Alcalase was added at an enzyme-to-substrate (E/S) ratio of 2%. After hydrolysis at 55 °C for 4 h, the reaction mixture was adjusted to pH 4.0 with HCl and maintained at 50 °C for 20 min to inactivate Alcalase. The pH was then readjusted to 7.5, followed by the addition of Flavourzyme (E/S 2%). The second hydrolysis step was conducted at 55 °C for 4 h and terminated by heating at 90 °C for 20 min. The hydrolysate was centrifuged at 4000×g for 20 min at 4 °C, and the supernatant was collected as HH for the subsequent Maillard reaction.

To determine the degree of hydrolysis (DH) of HH, the OPA method was employed (Parandi et al., 2024). The samples were prepared by dissolving heme protein and HH in distilled water at a concentration of 10 mg/mL. The OPA solution was prepared immediately before analysis. Briefly, OPA (80 mg) was first dissolved in methanol (2 mL), and the solution was combined with 0.1 M sodium tetraborate buffer (50 mL, pH 9.7), β-mercaptoethanol (200 μL), and 20% (w/w) SDS solution (5 mL). Distilled water was then added to bring the final volume to 100 mL. For analysis, 400 μL of each sample was mixed with 3 mL of the OPA reagent, incubated at 35 °C for 2 min, and the absorbance was measured at 340 nm using a spectrophotometer (BioTek Instruments, VT, USA). The total amide bond content of the heme protein was determined using the same procedure following complete hydrolysis with 6 N HCl at 110 °C for 24 h. OPA reagent with 400 μL deionized water was used as the blank. The degree of hydrolysis of HH was calculated using Eq. (1) and was found to be 21.70 ± 0.38%.

$$\text{Degree of hydrolysis}\left(\%\right)=\left(\left({\mathrm{Ab}}_{\mathrm{HH}}-{\mathrm{Ab}}_{\mathrm{HP}}\right)/\left({\mathrm{Ab}}_{\text{total}}-{\mathrm{Ab}}_{\text{blank}}\right)\right)\times 100$$ (1)

Ab_HH: Absorbance of HH; Ab_HP: Absorbance of heme protein; Ab_total: Absorbance of heme protein after being treated with 6 N HCl; Ab_blank: Absorbance of distilled water.

2.3. Preparation of HH-xylose MRPs and determination of optimal Maillard reaction conditions

HH and xylose were reacted under wet-heating conditions to generate MRPs. The preliminary reaction conditions were selected based on a previously reported method (Zhang et al., 2024), with an HH concentration of 2% (w/w), a heating time of 6 h, and a treatment temperature of 80 °C. HH-xylose mixtures were prepared in a 50 mL conical tube and subjected to the Maillard reaction in a water bath (BS-11; Jeio Tech, Daejeon, Korea) at 80 °C for 6 h with gentle mixing. To determine the optimal condition for HH–xylose MRP formation, the HH-to-xylose mass ratio was initially varied at 4:1, 2:1, 1:1, 1:2, and 1:4. The ratio showing favorable GD and browning intensity was then applied to the subsequent screening of heating time (2, 4, 6, 8 h, and 10 h) and temperature (70, 80, and 90 °C). The samples were classified as HH, HHX, and MHHX, representing native heme protein hydrolysate, unheated HH–xylose mixture, and MRP synthesized from HH–xylose, respectively. All samples were freeze-dried prior to analysis.

2.4. Characterization of MRPs

2.4.1. Determination of GD and browning intensity

The GD value was calculated from the reduction in OPA-reactive free amino groups according to a previously reported method (Han et al., 2025). OPA (80 mg) was first dissolved in methanol (2 mL), and the solution was combined with 0.1 M sodium tetraborate buffer (50 mL, pH 9.7), β-mercaptoethanol (200 μL), and 20% (w/w) SDS solution (5 mL). Distilled water was then added to bring the final volume to 100 mL. The OPA reagent was freshly prepared prior to use. For analysis, sample solution (100 μL, 10 mg/mL) was mixed with 4 mL of freshly prepared OPA reagent and allowed to react at 35 °C for 2 min. The reaction absorbance was then recorded at 340 nm using a UV–Vis spectrophotometer (BioTek Instruments, VT, USA). Native HH served as the reference sample. The GD value was determined using Eq. (2), and all measurements were performed in triplicate.

$$\mathrm{GD}\left(\%\right)=\left(\left({\mathrm{C}}_0-{\mathrm{C}}_{\mathrm{c}}\right)/{\mathrm{C}}_0\right)\times 100$$ (2)

C₀: free amino group concentration of native HH; C_c: free amino group concentration retained in HH–xylose MRPs.

Browning intensity was determined from the absorbance of each sample at 294 (A₂₉₄) and 420 nm (A₄₂₀). Sample solutions were prepared at 1 mg/mL and measured using a UV–Vis spectrophotometer (BioTek Instruments, VT, USA). All measurements were performed in triplicate.

2.4.2. Fluorescence spectra

Fluorescence spectra were measured to evaluate the formation of fluorescent Maillard intermediates. Samples were prepared in distilled water at 1 mg/mL prior to fluorescence analysis. Spectral data were acquired with a Varioskan LUX spectrophotometer (Thermo Fisher Scientific, MA, USA) with the excitation wavelength set to 347 nm, and the emission range was set to 400–480 nm (Han et al., 2025). All measurements were performed in triplicate.

2.4.3. Fourier transform-infrared (FT-IR) spectroscopy

The functional groups of HH, HHX, and MHHX were analyzed using an FT-IR spectrophotometer (FT/IR-4700; JASCO, Tokyo, Japan). The spectra were recorded over the range of 500–4000 cm⁻¹ at a resolution of 1 cm⁻¹. Spectral fitting was performed over the 1600–1700 cm⁻¹ range with PeakFit v4.12 (SeaSolve Software Inc., Framingham, USA). The proportion of each structure was calculated from the Gaussian peak area corresponding to α-helix (1650–1665 cm⁻¹), β-sheet (1615–1640 and 1690–1700 cm⁻¹), β-turn (1665–1690 cm⁻¹), and random coil (1640–1650 cm⁻¹) (Han, Keum, et al., 2024). Each sample was analyzed three times.

2.4.4. Intermolecular forces

A stepwise solvent extraction method was used to compare the intermolecular forces contributing to the structures of HH, HHX, and MHHX (Kim, Lee, et al., 2024). Samples were prepared at 2 mg/mL in five dissociating solutions: 0.05 M NaCl (S1), 0.6 M NaCl (S2), 0.6 M NaCl with 1.5 M urea (S3), 0.6 M NaCl with 8 M urea (S4), and 0.6 M NaCl with 8 M urea plus 0.5 M β-mercaptoethanol (S5). The dispersions were centrifuged at 8000×g for 15 min, and protein concentration in the supernatant was determined by the BCA assay using a bovine serum albumin calibration curve. The relative contribution of each intermolecular force was estimated from changes in soluble protein content during sequential solvent extraction. The increase observed at each solvent transition was used to represent ionic bond, hydrogen bond, hydrophobic interaction, and disulfide bond in the order of the S1–S5 extraction sequence. Measurements were performed in triplicate.

2.4.5. Particle size and zeta-potential

Mean particle size and zeta potential were evaluated to characterize the dispersion behavior of MRPs. Freeze-dried samples were prepared as an aqueous dispersion at 1 mg/mL and subjected to dynamic light scattering analysis using a Zetasizer Nano ZS instrument (Malvern Instruments Ltd., Malvern, UK). The instrument settings were fixed at a 90° scattering angle, 633 nm laser wavelength, and 25 °C during analysis. All samples were analyzed in triplicate.

2.4.6. Scanning electron microscopy (SEM)

For SEM analysis, the lyophilized samples were mounted on conductive carbon tape, followed by gold coating with an MCM-200 sputter coater (SEC Co., Ltd., Gyeonggi, Korea). The morphology of each sample was visualized at 300× magnification using a TM 3000 SEM (Hitachi Ltd., Tokyo, Japan).

2.5. Evaluation of MRPs

2.5.1. Surface hydrophobicity

Surface hydrophobicity was determined using a bromophenol blue (BPB)-binding assay. Briefly, 1 mL of sample solution (1 mg/mL) was combined with 200 μL of BPB solution (1 mg/mL) and allowed to react for 10 min. Following the reaction, the suspension was centrifuged (8000×g, 15 min). The supernatant was collected, diluted 10-fold, and analyzed at 595 nm using a UV–Vis spectrophotometer (BioTek Instruments, VT, USA). Distilled water was used as the blank. All samples were analyzed in triplicate. Surface hydrophobicity of MRPs was calculated using Eq. (3):

$$\text{Bound}\mathrm{BPB}=\left(200\mathrm{\mu g}\times \left({\mathrm{Ab}}_{\text{blank}}-{\mathrm{Ab}}_{\text{sample}}\right)\right)/{\mathrm{Ab}}_{\text{blank}}$$ (3)

Ab_blank: absorbance of the blank; Ab_sample: absorbance of sample.

2.5.2. Protein solubility

To evaluate the water-dispersibility of HH, HHX, and MHHX, the soluble protein fraction was quantified according to a modified procedure based on Han et al. (2025). Each sample was dispersed in distilled water at 2 mg/mL. The amount of soluble protein in the supernatant was measured using the BCA assay, and the values were determined from a bovine serum albumin calibration curve. All samples were analyzed in triplicate.

2.5.3. Emulsifying properties

The emulsifying activity index (EAI) and emulsifying stability index (ESI) of HH and MRPs were determined using a previously reported method with slight modifications (Zhao et al., 2022). Freeze-dried samples were dissolved in 7.2 mL of distilled water to obtain a final protein concentration of 1%. Sunflower oil (0.8 mL) was added, and the mixture was homogenized at 10,000 rpm for 2 min using a homogenizer (IKA-T18, IKA Staufen, Germany) to prepare the emulsion. Equal volumes of the emulsion were collected at 0 and 10 min after preparation and subsequently diluted 100-fold with a 1% SDS solution (w/v). The absorbance of the diluted emulsions was measured at 500 nm using a UV–Vis spectrophotometer (BioTek Instruments, VT, USA), with SDS solution serving as a blank. The EAI and ESI values were calculated using Eqs. (4), (5).

$$\mathrm{EAI}\left({\mathrm{m}}^2/\mathrm{g}\right)=\left(2\times 2.303\times {\mathrm{Ab}}_{\text{0min}}\times 100\right)/\left(\mathrm{C}\times 1\times \mathrm{\theta }\times \mathrm{10,000}\right)$$ (4)

$$\mathrm{ESI}\left(\min \right)=10\times {\mathrm{Ab}}_{\text{0min}}/\left({\mathrm{Ab}}_{\text{0min}}-{\mathrm{Ab}}_{10\min }\right)$$ (5)

Ab_0min: absorbance of the emulsion measured 0 min after preparation; C: protein concentration; θ: volume fraction of the oil phase (0.1); Ab_10min: absorbance of the emulsion measured 10 min after preparation.

2.5.4. Differential scanning calorimetry (DSC)

The thermal properties of the HH, HHX, and MHHX samples were analyzed using DSC (DSC 4000; PerkinElmer, MA, USA). HH, HHX, and MHHX were prepared in distilled water at 10 mg/mL and loaded into aluminum pans. A sealed empty pan was used for reference. The samples were heated from 20 °C to 80 °C at 10 °C/min, and onset temperature, peak temperature, end temperature, and enthalpy change (ΔH) were calculated using Pyris data analysis software (PerkinElmer, MA, USA). All samples were analyzed in triplicate.

2.5.5. Free amino acid composition

The free amino acid compositions of HH and MRP were analyzed using trichloroacetic acid (TCA) precipitation. To extract free amino acids, 0.1 g of the sample was mixed with 1.9 mL of 5% TCA and incubated for 1 h. The mixture was then centrifuged (LZ-1312, Labogene, Lillerød, Denmark) at 10,000 rpm for 10 min. The resulting supernatant was filtered through a 0.2 μm microfilter, and the absorbance was measured at 440 and 570 nm wavelengths using an amino acid autoanalyzer (LA8080, Hitachi, Tokyo, Japan). The free amino acid content was analyzed by measuring the following amino acids: aspartic acid (Asp), glutamic acid (Glu), serine (Ser), histidine (His), glycine (Gly), threonine (Thr), arginine (Arg), alanine (Ala), tyrosine (Tyr), cysteine (Cys), valine (Val), methionine (Met), phenylalanine (Phe), isoleucine (Ile), leucine (Leu), lysine (Lys), and proline (Pro).

2.5.6. Electronic tongue (E-tongue)

The taste attributes of HH, HHX, and MHHX were evaluated using the Astree II E-tongue (Alpha MOS, Occitanie, France). The E-tongue sensor array consisted of five taste sensors, AHS, PKS, CTS, NMS, and ANS, along with two reference sensors, SCS and CPS. HH, HHX, and MHHX were prepared in distilled water (10 mg/mL), and 25 mL of each solution was measured at 25 °C for 120 s in five replicates. The sensors were conditioned and calibrated with 0.01 M HCl solution and rinsed with distilled water between measurements. Data were expressed on a relative scale of 0–10 and processed using AlphaSoft 17 software (Alpha MOS, Occitanie, France).

2.6. Preparation of plant-based patties (PPs)

Four formulations were prepared to evaluate the effect of HH and MRPs on PP (Table S1): PP-Con was prepared without HH and MRPs, whereas PP-HH, PP-HHX, and PP-MHHX contained 1% (w/w) HH, HHX, and MHHX, respectively. TPP was hydrated in distilled water at 4 °C for 1 h, then mixed with the ingredients for 6 min. Subsequently, 100 g of the mixture was shaped into patties using a patty presser (Manual Burger Press 4; Spikomat Ltd., Nottingham, UK). The patties were heated on an electric pan (DW-1530, Daewon Home Electric Co., Ltd., Gyeonggi, Korea) at 150 °C for 3 min on each side until the core temperature reached 80 °C. Cooked samples were then equilibrated at room temperature for 30 min before physicochemical and sensory analyses.

2.7. Evaluation of physicochemical and sensory properties of PPs

2.7.1. Physicochemical properties of PPs

Raw and cooked patties were evaluated for surface color using a CR-210 colorimeter (Konica Minolta, Ltd., Osaka, Japan). Prior to measurement, the instrument was standardized with a white calibration plate (L^⁎ = +97.27, a^⁎ = +5.21, b^⁎ = −3.40). Color was expressed as CIE L^⁎, a^⁎, and b^⁎ values, corresponding to lightness, redness, and yellowness, respectively. The delta E (ΔE) was measured to evaluate the color differences, with PP-con as a reference. Measurements were repeated five times for each sample.

Heme-associated absorbance was determined using a previously described acidified acetone extraction method with slight modifications (Yin et al., 2026). Cooked patties were mixed with acidified acetone solution (40 mL of acetone, 9 mL of water, and 1 mL of concentrated hydrochloric acid) at a sample-to-solvent ratio of 1:10 (w/v), followed by homogenization at 10,000 rpm for 1 min. The mixture was kept in the dark for 1 h and centrifuged at 6000×g for 10 min. Absorbance was measured at 640 nm using acidic acetone as the blank. The Absorbance of PP-Con was used for background correction. PP-HH, containing 1% HH without xylose addition and Maillard reaction, was used as the heme-positive reference. Relative heme-associated absorbance was calculated using Eq. (6).

$$\text{Relative heme}-\text{associated absorbance}\left(\text{fold}\right)=\left({\mathrm{Ab}}_{\text{sample}}-{\mathrm{Ab}}_{\mathrm{PP}-\mathrm{Con}}\right)/\left({\mathrm{Ab}}_{\mathrm{PP}-\mathrm{HH}}-{\mathrm{Ab}}_{\mathrm{PP}-\mathrm{Con}}\right)$$ (6)

Ab_sample: absorbance of the sample including PP-HHX and PP-MHHX; Ab_PP-con: absorbance of PP-Con; Ab_PP-HH: absorbance of PP-HH.

Extractable non-heme iron profiles were determined using the 1,10-phenanthroline colorimetric method with slight modifications (Mawouma et al., 2024). Cooked patties were mixed with distilled water at a sample-to-solvent ratio of 1:10 (w/v), followed by homogenization at 10,000 rpm for 1 min. The samples were centrifuged at 6000×g for 10 min, and the supernatant (2 mL) was mixed with 0.16 mL of 0.39% sodium nitrite, 1 mL of 10% TCA in 1 N HCl, and 6.84 mL of distilled water. The mixture was heated at 100 °C for 10 min and centrifuged again. Fe²⁺ was determined by reacting the resulting supernatant with 1,10-phenanthroline, whereas total extractable non-heme iron was determined after reduction of Fe³⁺ to Fe²⁺ using hydroxylamine hydrochloride. After incubation for 15 min, Fe²⁺ and total extractable non-heme iron contents were calculated using a FeSO₄ standard curve. Fe³⁺ content was calculated by subtracting Fe²⁺ from total extractable non-heme iron.

The rheological properties of the samples were analyzed using a rheometer (MCR 92, Anton Paar, Graz, Austria) equipped with a parallel plate measuring 25 mm in diameter. For rheological measurement, raw patty samples were positioned on the lower plate with the plate gap set to 1 mm. A temperature sweep test was performed at a constant frequency of 1 Hz and a strain amplitude of 1% to simulate thermal gelation conditions. The samples were heated from 25 °C to 80 °C at a constant rate of 2 °C/min, maintained at 80 °C for 5 min, and then cooled to 25 °C at a rate of 5 °C /min. Storage modulus (G′) and loss modulus (G″) were continuously recorded during the heating–cooling cycle. The rheological test was conducted in triplicate.

Cooked patty samples were prepared as 1.5 cm × 1.5 cm × 1 cm cubes and subjected to texture profile analysis using a TA-XT Plus texture analyzer (Stable Micro Systems Ltd., Godalming, UK). The evaluated texture parameters included hardness, springiness, cohesiveness, chewiness, and gumminess. The test conditions were set as follows: pre-test speed, 2.0 mm/s; test speed, 1.0 mm/s; post-test speed, 1.0 mm/s; and trigger force, 5 g. Five measurements were performed.

2.7.2. Sensory evaluation of PPs

Sensory attributes were evaluated as previously described (Han et al., 2023). After three training sessions over 2 weeks, seven trained panelists (three males and four females, aged 25–33 years) evaluated the intensity of umami, bloody taste, bitterness, flavor, texture, and color. Each attribute was scored on a 9-point intensity scale, where 1 represented “not perceived” and 9 represented “extremely intense.” Training was performed using PP samples to familiarize the panelists with the sensory attributes and sample characteristics. Patties were served as 1.5 × 1.5 × 1.0 cm pieces with random three-digit codes. Between samples, the panelists rinsed their mouths with water. This procedure was approved by the Institutional Review Board (KKUIRB-202505-HR-070), and written informed consent was obtained from all participants.

2.8. Statistical analysis

Results are presented as mean ± standard deviation. Data processing and statistical analysis were carried out using SPSS-PASW Statistics software version 22.0 (SPSS Inc., Chicago, IL, USA). Significant differences among groups were assessed using one-way analysis of variance (ANOVA), followed by Duncan's multiple range test for post-hoc comparison. Differences were considered statistically significant at P < 0.05.

3. Results and discussion

3.1. Determination of optimal Maillard reaction conditions

To determine the optimal Maillard reaction conditions for the HH–xylose system, GD values and browning intensity were used as reaction indicators. During the Maillard reaction, protein grafting contributes to pigment generation and browning, which serve as key indicators of reaction progress. An OPA assay was used to quantify the extent of grafting. The browning intensity was evaluated by measuring A₂₉₄ for the early phase MRPs and A₄₂₀ for the final phase MRPs (Wang et al., 2024). As the HH proportion increased within the HH–xylose mixture, GD values gradually increased and reached a maximum at the 2:1 ratio (HH: 2%, xylose: 1%) (Fig. 1A). However, the GD decreased at HH-xylose mass ratio of 4:1 (HH: 2%, xylose: 0.5%). This reduction was likely due to insufficient xylose content, which limited its binding to free amino groups and subsequently hindered the formation of early Maillard reaction products such as Schiff bases. Meanwhile, a lower proportion of HH, accompanied by a higher xylose content, resulted in a decreasing trend in GD. This decline might be attributed to increased xylose levels, elevating the solution viscosity, which could impede the Maillard reaction by reducing the diffusion and interaction of reactive molecules (Ma et al., 2021). The browning intensity (A₂₉₄ and A₄₂₀ values) of the MRPs showed a trend similar to that of the GD, with the highest values observed at an HH-xylose mass ratio of 2:1 (Fig. 1B). Therefore, the optimal HH-xylose mass ratio of 2:1 was selected for further study.

Fig. 1.

Influence of different treatment parameters on the grafting degree and browning intensity of heme protein hydrolysate (HH) in the Maillard reaction. (A–B) HH–xylose mass ratio (Fixed HH concentration: 2 wt%; reaction time: 6 h; treatment temperature: 80 °C), (C–D) reaction time (HH concentration: 2 wt%; xylose concentration: 1 wt%; treatment temperature: 80 °C), and (E–F) treatment temperature (HH concentration: 2 wt%; xylose concentration: 1 wt%; reaction time: 6 h). Error bars indicate standard deviations (n = 3). Sample groups are distinguished by line color.

Under varying reaction times and temperatures, the MRPs exhibited the highest GD at 6 h and 80 °C (Fig. 1C and E). These results suggest that the optimal reaction time and temperature conditions might enhance grafting by increasing the molecular energy and promoting interaction between xylose and free amino groups. However, excessive heating time (8 h and 10 h) and temperature (90 °C) likely induced protein aggregation, thereby reducing the availability of amino groups in the Maillard reaction. Browning intensity, as measured by the absorbance at A₂₉₄ and A₄₂₀, increased with longer reaction times and higher temperatures (Fig. 1D and F). However, elevated A₄₂₀ levels are correlated with the accumulation of AGEs, which are associated with adverse health effects such as renal impairment and carcinogenesis (Lan et al., 2020). Therefore, the reaction conditions of 6 h and 80 °C were determined to be optimal for our samples. Based on these results, the optimal reaction conditions were set as 2% HH, 1% xylose, and heating at 80 °C for 6 h.

3.2. Characterization of MRP structure

3.2.1. GD, browning intensity, and fluorescence spectra of MRPs

Fluorescent Maillard intermediates can serve as indicators of reaction progress in the HH–xylose system under wet-heating conditions (Sun et al., 2023). To determine whether HH participates in the Maillard reaction with xylose under the selected wet heating conditions, the GD, browning intensity, and fluorescence spectra of the MRPs were evaluated (Fig. 2A–C). In addition, to test whether enzymatic hydrolysis enhances the Maillard reaction, the GD and browning intensity of the non-hydrolyzed heme protein-xylose mixture (HX), hydrolyzed heme protein-xylose mixture (HHX), MRPs of HX (MHX), and MRPs of HHX (MHHX) were evaluated. Results showed that MHX and MHHX exhibited significantly higher GD values than HX and HHX, with MHHX showing the highest value of 49.25% (P < 0.05) (Fig. 2A). These results suggest that protein denaturation during the Maillard reaction exposed free amino groups, facilitating their interaction with the carbonyl groups of xylose (Liu et al., 2022). Notably, enzymatic hydrolysis prior to the Maillard reaction likely promoted the exposure of free amino acids, enabling MHHX to bind more actively to carbonyl groups than MHX, thereby resulting in a higher GD (Hu et al., 2023). In addition, HHX demonstrated a significantly higher GD than HX (P < 0.05), which may be attributed to the enzymatic degradation of the heme protein structure, increasing the availability of free amino acids. Overall, the optimized Maillard reaction conditions effectively promoted grafting between xylose and HH, with enzymatic hydrolysis enhancing the reactivity between the free amino and carbonyl groups.

Fig. 2.

Degree of Maillard reaction and structural characteristics of heme protein hydrolysate and Maillard reaction products (MRPs). (A) Grafting degree, (B) browning intensity, (C) fluorescence spectra, (D) FT-IR spectra, (E) secondary structure, and (F) scanning electron microscopy images. Error bars indicate standard deviations (n = 3). Different letters indicate significant differences among groups at P < 0.05. Sample groups are distinguished by line color. White arrows indicate the porous matrix observed in the microstructure. HX: heme protein and xylose mixture, MHX: MRPs synthesized from heme protein and xylose, HH: heme protein hydrolysate, HHX: heme protein hydrolysate and xylose mixture, MHHX: MRPs synthesized from HH and xylose.

The browning intensity under optimized Maillard reaction conditions was also evaluated. A₂₉₄ reflects the formation of early phase MRPs, including Amadori and Heyns products, and dicarbonyl compounds. These compounds are associated with the generation of flavor-active molecules that contribute to food taste and aroma. A₄₂₀ indicated melanoidin accumulation during the final phase of the Maillard reaction and played a key role in food browning and visual quality (Kathuria et al., 2023). In our study, both the A₂₉₄ and A₄₂₀ values were significantly higher in MHX and MHHX than in HX and HHX (P < 0.05) (Fig. 2B). Among these groups, MHHX showed the highest values. The higher GD and browning intensity observed in MHHX indicated effective MRP formation under the optimal reaction condition. The stronger response of MHHX than MHX further suggests that enzymatic hydrolysis increased the availability of reactive groups in heme protein.

In the initial phase of the Maillard reaction, fluorescent compounds formed as intermediates. These compounds, which contribute to the formation of browning products, display an excitation peak at 347 nm and emit light in the 400–500 nm range (Zhao et al., 2024). Our data demonstrates that the fluorescent compounds formed during the Maillard reaction exhibited strong emission at 440–450 nm upon excitation at 347 nm (Fig. 2C). MHHX showed greater fluorescence intensity than HH and HHX, likely because of the enhanced early phase Maillard reaction, where exposed free amino acids in HH reacted with xylose, facilitating the formation of fluorescent compounds (Zhao et al., 2024). Following the reaction, the maximum fluorescence emission wavelength (λ_max) shifted from 450 nm to 440 nm. The reduction in λ_max was associated with the progression of the Maillard reaction, which promoted the formation of fluorescent compounds including pyrrole, pyridine, argpyrimidine, and pentoside (Sun et al., 2023; Wang et al., 2020). Furthermore, alterations in the intrinsic fluorescence of MHHX suggested modifications to its tertiary protein structure, likely due to changes in tryptophan (Li & Li, 2023). Overall, MHHX exhibited a fluorescence response with a lower λ_max, suggesting the formation of fluorescent Maillard intermediates and accompanying changes in the protein structure.

3.2.2. FT-IR, secondary structure, and intermolecular forces of MRPs

The FT-IR spectra revealed characteristic peaks that were consistent across all groups: C—C, C—O, and C—H bending at 1025 cm⁻¹, C—N stretching and N—H deformation at 1227 cm⁻¹ (amide III), COO⁻ symmetric stretching and/or C—H bending at 1392 cm⁻¹ (Fig. 2D). In addition, a broad band in the 1564–1641 cm⁻¹ region was attributed to overlapping amide II and amide I vibrations, mainly arising from N—H bending and peptide C O stretching, respectively. In MHHX, an increase in C—O stretching (1025 cm⁻¹), C—H bending (1392 cm⁻¹), and C—N stretching (1227 cm⁻¹) was observed compared to HH and HHX, indicating the formation of covalent bonds between HH and xylose. This increase led to the production of MRPs such as pyrazines, Amadori compounds, and Schiff base (Zhang, Wang, & Adhikari, 2022). Compared with HH and HHX, MHHX showed distinct changes in the broad amide I/II region at 1564–1641 cm⁻¹. These spectral changes indicate that the Maillard reaction modified the peptide bond environment and may have induced changes in the secondary structure of HH (Han, Keum, et al., 2024). Compared with HH and HHX, MHHX showed increased intensity at 2883 cm⁻¹ and 2968 cm⁻¹, which is associated with C—H stretching vibrations of —CH₂ and —CH₃ groups (Zhang, Wang, & Adhikari, 2022). The increased intensity of the bands at 2883 and 2968 cm⁻¹ indicate strengthened hydrophobic interactions from the Maillard reaction (Sun et al., 2024). MHHX also showed changes in the broad 3000–3500 cm⁻¹ region corresponding to O—H and N—H stretching vibrations, suggesting that the Maillard reaction altered hydrogen bonding in HH (Yang et al., 2020; Zhang, Wang, & Adhikari, 2022). Therefore, under the optimal Maillard reaction conditions, MHHX exhibited structural modifications associated with changes in hydrogen bonding and hydrophobic interactions.

The structural changes observed in HH through fluorescence spectroscopy led to the testing of its protein secondary structure and intermolecular forces. All samples showed more than 60% β-sheet, α-helix, and β-turn structures, suggesting the protein conformation is relatively stable and ordered (He et al., 2024). The structure of HH was found to be 14.81% α-helix, 34.39% β-sheet, 38.68% β-turn, and 12.12% random coil. HHX showed no significant structural differences compared to HH (Fig. 2E). In MHHX, β-turn and random coil represented 26.60% and 6.86% of the secondary structure, respectively. In contrast, β-sheet accounted for 46.63%, indicating that β-sheet was the predominant structural component after the reaction. These modifications were likely caused by thermal processing and xylose attachment, which facilitated conformational changes such as unfolding, solubilization, and deformation, ultimately increasing β-sheet formation. These conformational changes likely made nonpolar residues and thiol-related reactive sites more available, supporting the formation of HH–xylose conjugates (Jang et al., 2025). The higher proportion of β-sheet in MHHX suggests that part of the β-turn structure was rearranged into a more ordered β-sheet conformation (Li et al., 2021). Meanwhile, the α-helix content of MHHX was 19.91%, which was higher than that of HH (14.81%) (Fig. 2E). This elevated α-helix content may influence the dipole moment of the α-helices, which in turn could promote the partial self-assembled aggregation of unfolded protein fragments (Kim, Shin, et al., 2024). Overall, these findings indicate that the Maillard reaction can lead to MHHX denaturation, specifically by transforming β-turn into β-sheet structures and increasing the self-assembly of unfolded protein fragments.

Intermolecular forces were analyzed to clarify the chemical interactions involved in HH–xylose conjugate formation. In HH, disulfide bond showed the largest contribution to the intermolecular forces, whereas hydrophobic interaction, ionic bond, and hydrogen bond contributed to a lesser extent (Table 1). Compared with HH, HHX showed significantly increased hydrophobic interaction and ionic bond (P < 0.05). These data suggest that mixing HH and xylose led to the exposure of hydrophobic groups, thereby enhancing hydrophobic interaction. A previous study also showed that casein–maltodextrin could increase hydrophobic interaction (Yu et al., 2024). MHHX showed a significant increase in hydrophobic interaction, hydrogen bond, and ionic bond, accompanied by a reduction in disulfide bond (P < 0.05) (Table 1). Notably, MHHX exhibited the most pronounced increase in hydrophobic interaction from 0.21 mg/mL to 0.46 mg/mL, constituting the highest proportion among all intermolecular interaction types (P < 0.05). The Maillard reaction promoted conformational rearrangement of HH, which increased the availability of nonpolar residues and contributed to enhancement of hydrophobic association in MHHX (Kim, Yong, et al., 2024). Furthermore, steric hindrance provided by xylose can inhibit disulfide bond formation in polypeptide chains, leading to their disruption and potential conversion into free sulfhydryl groups on the exterior (Zhao et al., 2024). Overall, these findings demonstrate that the Maillard reaction can modulate the chemical interactions between HH and xylose, particularly by increasing hydrophobic interaction.

Table 1.

Intermolecular forces of heme protein hydrolysate and Maillard reaction product (MRP).

Chemical interactions		HH	HHX	MHHX
Intermolecular forces (mg/mL)	Ionic bond	0.15 ± 0.01c	0.18 ± 0.01b	0.24 ± 0.00a
	Hydrogen bond	0.15 ± 0.00b	0.14 ± 0.02b	0.25 ± 0.03a
	Hydrophobic interaction	0.21 ± 0.01c	0.25 ± 0.02b	0.46 ± 0.03a
	Disulfide bond	1.10 ± 0.01a	1.01 ± 0.01b	0.69 ± 0.01c

HH: heme protein hydrolysate, HHX: HH and xylose mixture, MHHX: MRPs synthesized from HH and xylose.

Different lowercase letters (^a-c) within the same row indicate significant differences among formulations (P < 0.05).

Values represent mean ± standard deviation (n = 3).

3.2.3. Particle size, zeta-potential, and scanning electron microscopy of MRPs

Protein stability and solubility are commonly evaluated using key factors, such as particle size, surface charge, and microscopic structure (Kim, Yong, et al., 2024). Additionally, changes in particle size may serve as an indicator of structural modifications, including conjugation between HH and xylose (Li, Zhong, et al., 2024). HH showed a mean particle size of 692.13 nm (Table S2). In HHX, the mean particle size increased slightly to 770.00 (P < 0.05), indicating that partial conjugation could have occurred as a result of simple mixing of HH with xylose. The mean particle size of MHHX was significantly higher at 1212.67 nm compared with HH and HHX (P < 0.05) (Table S2), suggesting that the heating process facilitated the formation of covalent bonds between HH and xylose, leading to the development of larger molecular structures (Hu et al., 2020). Therefore, these findings indicate that HH reacted with xylose through a Maillard reaction, generating larger HH–xylose conjugates via covalent bonding.

Zeta-potential analysis was used to compare the surface charge states of the sample solutions, which are closely related to colloidal stability. HH exhibited a negative zeta-potential of −21.70 mV (Table S2). An absolute zeta-potential value greater than 20 mV indicates that repulsive forces surpass attractive forces, resulting in a relatively stable system (Tirgarian et al., 2023). No significant difference in zeta potential was found between HH and HHX (P > 0.05), suggesting that adding xylose without thermal treatment had minimal effect on the electrical properties of HH (Table S2). MHHX exhibited significantly higher absolute zeta potential than HH and HHX (P < 0.05) (Table S2). These results indicate that the covalent attachment of negatively charged xylose to HH leads to a reduction in the positively charged amino acid residues (Zhang et al., 2023). In addition, the observed increase in the absolute zeta potential value might result from protein unfolding during the Maillard reaction, which could expose internally located negatively charged groups (Zhao et al., 2022). This exposure leads to a more negative zeta potential, enhancing electrostatic repulsion and suppressing protein aggregation, which could be advantageous for various applications (Carpentier et al., 2021). Collectively, the Maillard reaction facilitated the binding of xylose to HH, which increased the negative surface charge and contributed to the formation of a more stable dispersed phase.

SEM has been employed to investigate the morphological characteristics of HH and MRPs, which provide functional properties through alterations in the protein microstructure (He et al., 2021). The surface morphologies of HH, HHX, and MHHX were visualized by SEM. Both HH and HHX displayed coarse, fragmented surfaces typical of lyophilized powders (Fig. 2F). However, MHHX exhibited a large, irregular, and highly porous matrix formed through cross-linking between HH and xylose (Fig. 2F), facilitating water interaction and enhancing solubility (He et al., 2021). The SEM image revealed structural changes in MHHX, which were further supported by an increase in its absolute zeta potential, suggesting a more stable morphology. Collectively, SEM data revealed that HH–xylose conjugate formation was accompanied by morphological changes, which were likely associated with secondary-structure rearrangement and chemical interactions.

3.3. Evaluation of MRPs functionality

3.3.1. Surface hydrophobicity and protein solubility of MRPs

Surface hydrophobicity reflects the availability of nonpolar regions exposed at the protein interface. Because these regions affect protein behavior at air–water and oil–water interfaces, surface hydrophobicity is often related to emulsifying and foaming properties (Han et al., 2025). Surface hydrophobicity was evaluated based on the absorbance generated by the interaction between BPB and the surface-exposed hydrophobic residues (Zahir et al., 2021). No significant difference in surface hydrophobicity was observed between HH and HHX (P > 0.05; Fig. 3A). In contrast, the surface hydrophobicity of MHHX was significantly lower than that of HH and HHX (P < 0.05). This reduction may be attributed to conformational changes during the Maillard reaction, where protein unfolding exposes hydrophobic residues in HH along with structural modifications of xylose. Although the exposure of hydrophobic residues typically increases surface hydrophobicity, the covalent bonding of xylose to HH likely introduced hydrophilic hydroxyl groups that masked surface-exposed hydrophobic sites. This masking promoted self-assembled aggregation among buried hydrophobic residues, leading to a decrease in surface hydrophobicity (Kim, Shin, et al., 2024). Meanwhile, enzymatic hydrolysis of HH likely induced partial protein unfolding, leading to the exposure of hydrophobic residues (Knežević-Jugović et al., 2022). However, through the Maillard reaction, MHHX underwent more extensive protein unfolding than HH, resulting in increased exposure of hydrophobic residues on the protein surface. Therefore, in MHHX, the masking effect of hydroxyl groups and extensive exposure of hydrophobic residues may have enhanced hydrophobic interaction, leading to aggregation and decreased surface hydrophobicity (Li et al., 2025). These observations were supported by the secondary structure and intermolecular forces, which showed higher α-helix and β-sheet proportions, and hydrophobic interaction in MHHX (Fig. 2E, Table 1).

Fig. 3.

Protein functionality of heme protein hydrolysate and Maillard reaction products (MRPs). (A) Surface hydrophobicity, (B) protein solubility, (C) emulsifying ability index, (D) emulsifying stability index, and (E) electronic tongue. Error bars indicate standard deviations (n = 3). Different letters indicate significant differences among groups at P < 0.05. Sample groups are distinguished by line color. HH: heme protein hydrolysate, HHX: heme protein hydrolysate and xylose mixture, MHHX: MRPs synthesized from HH and xylose.

Protein solubility in water influences the visual appearance of various foods and plays a key role in improving the gelling, foaming, and emulsifying properties (Han et al., 2025). In this study, the protein solubilities of HH, HHX, and MHHX were measured. HHX did not show a significant difference in protein solubility compared with HH (P > 0.05) (Fig. 3B). The intermolecular forces and particle size results suggested that simple mixing of HH and xylose led to partial conjugation and structural changes (Tables 1 and S2). However, these changes were insufficient to affect the protein solubility. The protein solubility of MHHX was significantly higher than those of HH and HHX (P < 0.05). The enhanced solubility was attributed to protein unfolding, which facilitated covalent bonding between HH and xylose, thereby enhancing protein–water interactions and inhibiting aggregation (Tirgarian et al., 2022). This protein unfolding was evidenced by an increase in the β-sheet structure of MHHX (Fig. 2E). The high absolute zeta potential of MHHX enhances electrostatic repulsion, thereby inhibiting protein–protein aggregation and promoting the formation of water-soluble complexes that contribute to enhanced protein solubility (Kim, Shin, et al., 2024) (Table S2). In addition, SEM analysis revealed the porous structure of MHHX, which likely facilitated water interactions and further enhanced protein solubility (He et al., 2021) (Fig. 2F). Although MHHX showed lower surface hydrophobicity than HH due to the aggregation of hydrophobic residues, covalent bonding with xylose likely enhanced protein–water interactions, resulting in improved protein solubility. Collectively, the Maillard reaction enhanced MHHX protein solubility by promoting protein–water interactions and inhibiting aggregation, likely due to increased electrostatic repulsion and porous structure formation.

3.3.2. Emulsion properties of MRPs

EAI and ESI measure the protein adsorption capacity at the oil–water interface and the stability of the droplet interface over time, respectively (Kim et al., 2024). These measurements were used to assess the emulsifying properties of HH, HHX, and MHHX. The EAI and ESI values of HHX were significantly higher than those of HH (P < 0.05) (Fig. 3C and D). HHX is a mixture of HH and xylose, which is a hydrophilic saccharide due to its multiple hydroxyl (–OH) groups. In this study, an emulsion was prepared with a composition of 90% water and 10% oil. Therefore, the enhanced EAI and ESI values of HHX were likely influenced by the increased hydrophilicity imparted by xylose, which promoted protein adsorption at the water interface of the emulsion. A previous study reported that even a simple mixture of chickpea protein and citrus pectin led to increased EAI and ESI (Zhang et al., 2025). MHHX exhibited significantly higher EAI and ESI values than HH and HHX (P < 0.05) (Fig. 3C and D). The Maillard reaction forms covalent bonds between the hydrophilic groups of xylose and HH, thereby increasing the hydrophilicity and generating a porous structure that enhances protein solubility (Figs. 2F and 3B). Because water is the primary solvent in the emulsion, the enhanced solubility of MHHX facilitates a more effective distribution at the interface, contributing to a thicker interfacial layer and improved EAI and ESI (Yan et al., 2021). Additionally, structural modifications in MHHX increased β-sheet content and exposed hydrophobic groups, facilitating self-assembled aggregation and promoting α-helix formation (Fig. 2E). The increase in α-helix content may enhance interfacial adsorption capacity, ultimately improving emulsifying properties (Kim, Shin, et al., 2024), thereby increasing the EAI and ESI of MHHX. Collectively, structural rearrangement of MHHX through the Maillard reaction promoted adsorption and stabilization at the water–oil interface, leading to improved emulsifying performance.

3.3.3. DSC of MRPs

DSC analysis is used to characterize protein thermal transition behavior based on heat-flow changes during controlled heating, enabling comparison of thermal stability among HH, HHX, and MHHX (Chu et al., 2024). Peak temperatures ranging from 60 °C to 63 °C were observed across all samples (Table 2), suggesting the occurrence of thermal transitions (Zhao et al., 2022). No significant difference in the peak temperature of HHX was observed compared to that of HH (P > 0.05) (Table 2), implying that xylose incorporation had no measurable impact on the molecular environment of HH. MHHX exhibited significantly higher peak temperatures than HH and HHX (P < 0.05) (Table 2). A high peak temperature, which represents the thermal denaturation point of a protein, indicates high thermal stability (Han et al., 2025). Thermal stability is associated with GD, which reflects the conjugation between HH and xylose; consequently, MHHX exhibited high peak temperatures because of its higher GD resulting from the Maillard reaction (Zhao et al., 2022). MHHX showed the lowest ΔH value among the samples (P < 0.05). ΔH represents the energy required for protein conformational transition. The decreased ΔH of MHHX suggests that prior unfolding and conjugation during the Maillard reaction partially denatured the protein, reducing the enthalpy required for thermal denaturation (Chu et al., 2024). Higher peak temperatures coupled with lower ΔH reflected fewer intramolecular disruptions, improving thermal stability (Zhong et al., 2019). Improved thermal stability is associated with various functional properties of proteins including emulsion stability, foaming stability, and solubility (Chu et al., 2024). Collectively, these findings suggest that the Maillard reaction enhanced the thermal stability of MHHX through protein modification and potentially contributed to improved functionality.

Table 2.

Thermal stability of heme protein hydrolysate and Maillard reaction products (MRPs).

	HH	HHX	MHHX
Onset temperature (°C)	58.48 ± 0.13ns	58.69 ± 0.36ns	58.70 ± 0.12ns
Peak temperature (°C)	60.24 ± 0.02b	60.41 ± 0.43b	62.06 ± 0.55a
End temperature (°C)	62.49 ± 0.41b	63.34 ± 0.15a	63.92 ± 0.06a
ΔH (mJ/g)	22.87 ± 2.81a	24.27 ± 2.55a	18.87 ± 1.46c

HH: heme protein hydrolysate, HHX: HH and xylose mixture, MHHX: MRPs synthesized from HH and xylose.

Different lowercase letters (^a-c) within the same row indicate significant differences among formulations (P < 0.05).

Values represent mean ± standard deviation (n = 3).

ns: not significant (P > 0.05).

3.3.4. Free amino acids content of MRPs

During enzymatic hydrolysis, proteins are partially broken down into peptides and free amino acids, which contributes to flavor development. These hydrolysis products can act as precursors for the Maillard reaction, enhancing the flavor of MRPs (Qiu et al., 2024). The free amino acid contents of HH and MRPs are presented in Table 3, with MHHX showing the highest total free amino acid concentration (P < 0.05). This increase was likely due to the release of free amino acids from heme proteins during hydrolysis, followed by additional degradation during the Maillard reaction (Qiu et al., 2024). In HH, Leu was the most abundant, followed by Phe. As Leu and Phe residues at the C-terminus of peptides are known to impart bitterness, HH may exhibit a strong bitter taste (Li, Liu, et al., 2024). Asp and Glu are the most abundant amino acids in MHHX, both of which contribute to its umami taste (Qiu et al., 2024). The content of these amino acids was also significantly higher than in HH (P < 0.05). The elevated levels of umami amino acids resulting from the Maillard reaction may have enhanced the umami characteristics of MHHX. Meanwhile, the content of hydrophobic amino acids (e.g., Ala, Val, Leu, Ile, Met, Phe, Tyr, Cys, and Pro) decreased from 56.36 mg/mL in HH to 36.79 mg/mL following the Maillard reaction. This reduction suggested enhanced hydrophobic interaction during the Maillard reaction (Li, Liu, et al., 2024), consistent with the intermolecular force results (Table 1). Notably, Leu and Pro levels were significantly lower in the MHHX group than in the HH group (P < 0.05). As these hydrophobic amino acids contribute to bitterness, their lower levels of MHHX may have resulted in a reduced bitter taste (Li, Liu, et al., 2024). Taken together, the Maillard reaction altered the free amino acid content and composition of HH, which may have enhanced the umami taste and suppressed the bitterness of MHHX.

Table 3.

Free amino acid contents of heme protein hydrolysate (HH) and Maillard reaction products (MRPs).

Free amino acid contents (mg/mL)	HH	HHX	MHHX
Asp	6.81 ± 0.22b	6.80 ± 0.65b	9.26 ± 1.08a
Glu	4.93 ± 0.07b	4.97 ± 0.73b	7.11 ± 0.44a
Ser	1.42 ± 0.02c	2.65 ± 0.06b	4.85 ± 0.05a
His	0.65 ± 0.03b	0.44 ± 0.02c	1.92 ± 0.13a
Gly	2.64 ± 0.09a	2.70 ± 0.03a	1.82 ± 0.05b
Thr	6.25 ± 0.31a	5.19 ± 0.18b	4.78 ± 0.11c
Arg	4.22 ± 0.16a	3.80 ± 0.09b	1.66 ± 0.01c
Ala	4.03 ± 0.11b	4.14 ± 0.09b	7.55 ± 0.13a
Tyr	4.08 ± 0.06a	4.01 ± 0.10a	2.21 ± 0.03b
Cys	0.46 ± 0.04b	0.50 ± 0.02b	0.93 ± 0.03a
Val	7.73 ± 0.28a	7.49 ± 0.22a	5.36 ± 0.15b
Met	4.94 ± 0.15a	4.33 ± 0.04b	3.23 ± 0.06c
Phe	9.31 ± 0.34b	9.81 ± 0.14a	7.44 ± 0.20b
Ile	6.90 ± 0.13a	6.47 ± 0.24b	3.39 ± 0.08c
Leu	10.36 ± 0.64a	9.73 ± 0.32a	6.89 ± 0.12b
Lys	1.55 ± 0.04b	1.66 ± 0.08b	4.98 ± 0.06a
Pro	7.54 ± 0.30a	7.34 ± 0.18a	4.79 ± 0.09b
Total	83.84 ± 2.34a	82.03 ± 2.10a	78.17 ± 0.92b

HH: heme protein hydrolysate, HHX: HH and xylose mixture, MHHX: MRPs synthesized from HH and xylose.

Asp = aspartic acid, Glu = glutamic acid, Ser = serine, His = histidine, Gly = glycine, Thr = threonine, Arg = arginine, Ala = alanine, Tyr = tyrosine, Cys = cysteine, Val = valine, Met = methionine, Phe = phenylalanine, Ile = isoleucine, Leu = leucine, Lys = lysine, Pro = proline.

Different lowercase letters (^a-c) within the same row indicate significant differences among formulations (P < 0.05).

Values represent mean ± standard deviation (n = 3).

3.3.5. E-tongue analysis of MRPs

E-tongue analysis was used to compare the taste profiles of HH, HHX, and MHHX (Fig. 3E). AHS, PKS, CTS, NMS, and ANS are taste sensors that are responsive to sourness, sweetness, saltiness, umami, and bitterness, respectively (Han et al., 2025). HH showed the highest taste intensity for bitterness, followed in descending order by sourness, sweetness, saltiness, and umami. In the free amino acid analysis, HH showed the highest concentrations of Leu and Phe, which imparted a bitter taste (Table 3). Consequently, HH exhibited the highest response intensity in the ANS (bitterness) sensor, due to the elevated levels of Leu and Phe. Compared to HH, HHX exhibited lower sourness and bitterness intensities, along with an increased umami intensity. The saltiness and sweetness intensities in HHX were comparable to those observed in HH. These results suggest that xylose may independently influence taste perception, potentially contributing to the suppression of bitterness and sourness. However, the amount of xylose added to HHX was insufficient to enhance its sweetness and saltiness. In the Maillard reaction sample, the umami intensity of MHHX was 7.6, which was higher than those of HH and HHX. The elevated umami intensity in MHHX was likely associated with changes in taste-active compounds, particularly the increased levels of Glu and Asp and the formation of small peptides during the Maillard reaction (Li, Liu, et al., 2024). Indeed, the levels of Asp and Glu in MHHX increased significantly following the Maillard reaction (P < 0.05, Table 3). The intensities of sourness and bitterness of MHHX were lower than those of HH and HHX. Protein bitterness is influenced by the presence of hydrophobic amino acids (e.g., Ala, Val, Leu, Ile, Met, Phe, Tyr, Cys, and Pro) and specific bitter peptides. During the Maillard reaction, these compounds may have undergone thermal degradation or formed complexes through hydrophobic interaction, resulting in the loss of their inherent bitterness (Li, Liu, et al., 2024). This change was linked to the reduced bitterness intensity observed for MHHX. The results of free amino acid analysis and intermolecular forces for MHHX showed a reduction in hydrophobic amino acid content and an increase in hydrophobic interaction (Table 1, Table 3). In addition, the reduced surface hydrophobicity of MHHX suggests that conjugation with xylose may have masked surface-exposed hydrophobic sites through the introduction of hydrophilic hydroxyl groups (Fig. 3A). These changes collectively support the lower bitterness intensity observed in E-tongue analysis. The sweetness intensity of MHHX was similar to that of HH, whereas the saltiness intensity was markedly higher. A previous study on peanut protein–galactose Maillard products showed elevated umami and saltiness intensities in both sensory evaluation and E-tongue analysis (Xing et al., 2024). These findings suggest that umami amino acids may influence saltiness perception, as observed in MHHX. Overall, E-tongue analysis revealed that the Maillard reaction led to a decrease in the sourness and bitterness of HH, accompanied by an increase in umami and saltiness.

3.4. Physicochemical and sensory characteristics of PPs

To assess the potential of MHHX as a functional additive for PBMs, its surface hydrophobicity, solubility, emulsifying ability, thermal stability, and taste profile were analyzed. To examine the applicability of MHHX in PPs, we compared its color, heme-associated absorbance, extractable non-heme iron profiles, rheological properties, texture, and flavor with those of HH and HHX.

3.4.1. Physicochemical properties of PPs

To characterize the physicochemical properties of PPs containing HH, HHX, and MHHX, visual appearance, color, rheological behavior, and texture parameters were evaluated (Fig. 4A–C, Table 4). Meat color is a key quality attribute that plays a significant role in determining consumer preferences (Han et al., 2023). Heme proteins appear bright red when in the Fe²⁺ state and bound to oxygen; however, when oxidized to the Fe³⁺ state, forming metmyoglobin, the color changes to brown or grayish-brown (Han, Wang, et al., 2024). Although the Maillard reaction produces brown melanoidin pigments, excessive amounts can lead to an undesirable color in meat products. Although heme proteins and MRPs possess distinct inherent colors, our data showed that the addition of 1% HH, HHX and MHHX did not cause significant color differences among all raw patties (P > 0.05) (Fig. 4A, Table 4). Consistent with this finding, the ΔE values, reflecting overall color differences relative to PP-Con, did not differ significantly among PP-HH, PP-HHX, and PP-MHHX (P > 0.05). After cooking, PP-HHX and PP-MHHX showed significantly lower L^⁎ values than PP-Con and PP-HH, with PP-MHHX exhibiting the lowest L^⁎ values (P < 0.05) (Fig. 4B, Table 4). The a^⁎ and b^⁎ values of the cooked patties were not significantly different among the groups (P > 0.05). Because of Maillard reactions and surface charring during the cooking process, PP-HHX exhibited a dark brown color, leading to a reduction in its L^⁎ value. However, MHHX underwent Maillard reactions during the MRP preparation step and additional reactions occurred during cooking, resulting in a lower L^⁎ value (42.91) than that of PP-HHX. Based on our previous study, the L^⁎ value of cooked beef patties was 41.11 (Han et al., 2023). The ΔE value of PP-MHHX was 4.02, which was significantly higher than those of the other groups. Therefore, MHHX can contribute to the development of a meat-like color in PP by reducing the L^⁎ value after cooking.

Fig. 4.

Physicochemical and sensory properties of plant-based patties (PPs) supplemented with heme protein hydrolysate (HH) and Maillard reaction products (MRPs). Visual appearance of (A) raw and (B) cooked PPs, (C) viscoelasticity during temperature sweep at 25–80 °C, 1 Hz, and 1% strain, and (D) sensory evaluation conducted with seven panelists. Sample groups are distinguished by line color. PP-Con: Control formulation without HH-derived ingredients, PP-HH: Formulation containing 1% HH, PP-HHX: Formulation containing 1% HH–xylose mixture 1%, PP-MHHX: formulation containing 1% MRPs synthesized from HH and xylose.

Table 4.

Color and texture profile analysis of plant-based patties (PPs) supplemented with heme protein hydrolysate (HH) and Maillard reaction products (MRPs).

Parameters		PP-Con	PP-HH	PP-HHX	PP-MHHX
Raw	L⁎	53.45 ± 0.61ns	52.56 ± 1.09ns	52.60 ± 0.91ns	52.37 ± 0.83ns
	a⁎	23.69 ± 1.03ns	22.97 ± 0.36ns	23.05 ± 0.52ns	22.78 ± 0.92ns
	b⁎	10.31 ± 0.54ns	10.22 ± 0.26ns	10.75 ± 0.46ns	10.20 ± 0.55ns
	ΔE		1.65 ± 0.43ns	1.58 ± 0.75ns	1.78 ± 0.66ns
Cooked	L⁎	46.83 ± 1.43a	46.22 ± 0.59a	44.67 ± 1.29b	42.91 ± 0.59c
	a⁎	11.89 ± 0.59ns	11.45 ± 0.55ns	11.20 ± 0.41ns	11.39 ± 0.73ns
	b⁎	13.60 ± 1.05ns	14.43 ± 1.60ns	13.06 ± 1.95ns	13.85 ± 1.58ns
	ΔE		1.74 ± 0.66b	2.53 ± 0.57b	4.02 ± 0.87a
Hardness (N)		12.54 ± 1.18c	12.59 ± 1.13c	13.93 ± 0.84b	15.95 ± 0.72a
Springiness		0.74 ± 0.05ns	0.73 ± 0.01ns	0.76 ± 0.01ns	0.77 ± 0.04ns
Cohesiveness		0.38 ± 0.01c	0.42 ± 0.02b	0.44 ± 0.02b	0.47 ± 0.02a
Chewiness (N)		3.32 ± 0.52c	3.66 ± 0.93bc	4.27 ± 0.40ab	4.99 ± 0.56a
Gumminess (N)		4.53 ± 0.38b	4.30 ± 1.04b	5.80 ± 0.39a	6.58 ± 0.52a

PP-Con: Control formulation without HH-derived ingredients, PP-HH: Formulation containing 1% HH, PP-HHX: Formulation containing 1% HH–xylose mixture 1%, PP-MHHX: formulation containing 1% MRPs synthesized from HH and xylose.

Different lowercase letters (^a-c) within the same row indicate significant differences among formulations (P < 0.05).

ns: not significant (P > 0.05).

Values represent mean ± standard deviation (n = 5).

To evaluate changes in heme- and iron-related factors associated with bloody taste, heme-associated absorbance and extractable non-heme iron content were measured in PPs containing HH, HHX, or MHHX (Table S3). Cooked PPs were analyzed to reflect the heme and iron status at the point of consumption, as thermal processing may alter heme stability and iron release. The relative heme-associated absorbance of PP-HHX and PP-MHHX was significantly lower than that of PP-HH, with PP-MHHX showing the lowest value (P < 0.05; Table S3). In contrast, the total extractable non-heme iron content of PP-MHHX was significantly higher than that of PP-HH (P < 0.05). These results may reflect heat-induced structural changes in HH and subsequent heme degradation. Thermal denaturation can expose porphyrin-bound iron and facilitate degradation of the porphyrin ring (Yin et al., 2026). In PP-MHHX, prior heat exposure during the Maillard reaction, together with subsequent cooking, may further promote heme degradation and iron release, resulting in lower heme-associated absorbance and higher non-heme iron content. Despite the absence of prior Maillard reaction heating, PP-HHX showed lower heme-associated absorbance after cooking than PP-HH. This difference may result from interactions between xylose and HH that altered the microenvironment surrounding the heme chromophore (Raupbach et al., 2020). PP-MHHX exhibited the highest Fe³⁺ ratio, suggesting that repeated thermal exposure during the Maillard reaction and cooking promoted the oxidation of Fe²⁺. Overall, structural changes in HH induced by the Maillard reaction and cooking process reduced heme-associated absorbance and increased non-heme iron release in PP-MHHX.

The heating–cooling cycle induces protein denaturation and gelation, enabling evaluation of the viscoelastic and physicochemical properties of PBMs under varying temperature conditions. The temperature sweep was carried out across both the heating and cooling phases, with G′ and G″ tracked to characterize gel matrix formation of PPs (Fig. 4C). Changes in the rheological properties of PPs were primarily reflected in G', while variations in the G″ were negligible. An initial increase in G' was observed around 60 °C during the heating phase, followed by a continuous rise during the 80 °C holding stage. In the cooling phase, the G' value further increased (Fig. 4C). To maintain the cohesiveness and shape of the formulation, PPs were prepared with binders including methylcellulose, κ-carrageenan, pea protein, and starch. The presence of multiple binders increased the difficulty of determining which component predominantly contributed to the changes in G' during heating. In addition, identifying the binder that undergoes structural transitions at specific temperatures remains challenging due to potential interactions among the components. Nevertheless, the increase in G' of PPs during the cooling phase reflected the gelation behavior of pea protein, starch, and κ-carrageenan, which possess the ability to form gels upon cooling after thermal treatment (Ko et al., 2021). PP-MHHX exhibited the highest G' values throughout the heating-cooling cycle (Fig. 4C). Rheological properties are influenced by droplet morphology (size and shape) and particle–particle interactions (Han et al., 2025). In MRP, the increased content of β-sheet led to the exposure of hydrophobic groups, thereby promoting protein aggregation and inducing structural transition to α-helix formation (Kim et al., 2024). This stabilized structure, characterized by increased α-helix content and solubility, facilitates interfacial adsorption, leading to an enhanced emulsifying capacity (Yan et al., 2021). Furthermore, the high peak temperature and ΔH of MHHX indicate superior thermal stability and its potential to improve water–protein–oil interactions, contributing to gel network development during heating (Qi et al., 2024). Collectively, the enhanced functional properties of MHHX led to improved rheological properties of PP during the heating-cooling cycle.

Texture, along with taste and aroma, plays a key role in food quality. Replicating the texture of meat in PBMs is crucial to achieve comparable organoleptic properties. Table 4 shows the texture properties of the PPs supplemented with HH, HHX, and MHHX. PP-MHHX exhibited significantly higher hardness than PP-Con, PP-HH, and PP-HHX (P < 0.05). The high β-sheet content in MHHX could promote HH-xylose interactions and expose internal hydrophobic residues (Kim, Yong, Chun, Kim and Lee, 2024). In MHHX, exposure to hydrophobic residues induces self-assembled aggregation, leading to an increase in α-helix content and a decrease in surface hydrophobicity. Nevertheless, the increased α-helix content may enhance the emulsifying capacity due to its strong interfacial adsorption ability (Kim, Shin, et al., 2024). In addition, the high protein solubility of MHHX may facilitate protein migration toward the oil–water interface, thereby supporting interfacial film development (Yan et al., 2021). The thermal properties of MHHX facilitate the enhancement of water–protein–oil interactions during heating, thereby improving the gel strength (Qi et al., 2024). As a result, MHHX demonstrated higher EAI and ESI than the other groups, and the patty formulated with MHHX exhibited significantly stronger gel strength after the heating–cooling cycle (Figs. 3C and D, 4C). The formation of a strong gel network is presumed to be a key factor in the increased hardness observed in the PPs. Notably, PP-HHX also demonstrated significantly higher hardness than PP-Con and PP-HH (P < 0.05). The cooking process, performed at 150 °C until the patty core reached 80 °C, may have provided sufficient thermal conditions to promote Maillard reactions in HHX. However, MHHX, which was subjected to the Maillard reaction, likely exhibited stronger protein-water binding interactions than HHX, contributing to its higher hardness. No significant differences in springiness were observed among groups (P > 0.05). PP-HH, PP-HHX, and PP-MHHX exhibited significantly higher cohesiveness than that of PP-Con (P < 0.05), with PP-MHHX exhibiting the highest cohesiveness. MHHX may strengthen the PP gel matrix through interactions with proteins, water, and binders, resulting in a more compact network with improved structural integrity. PP-MHHX also exhibited significantly higher chewiness and gumminess than PP-Con and PP-HH (P < 0.05). Overall, MHHX improved the textural characteristics of PPs by promoting a more cohesive and compact gel network.

3.4.2. Sensory evaluation of PPs

MHHX showed higher umami and lower bitterness intensities in the E-tongue analysis than HH and HHX (Fig. 3E). Therefore, the sensory contribution of MHHX was further evaluated after its incorporation into PPs. Sensory evaluation was performed focusing on color, umami, bloody taste, bitterness, flavor, and texture. PP-MHHX exhibited a significantly higher color score than PP-Con (P < 0.05) (Fig. 4D). The color data showed that the L^⁎ value of PP-MHHX decreased significantly after heating (Table 4). These color changes impart a meat patty–like appearance to the PPs, leading to significantly higher color scores for PP-MHHX than for PP-Con. Umami and flavor scores were significantly higher for PP-MHHX than for PP-Con (P < 0.05). This outcome is consistent with other data obtained using MHHX. The E-tongue analysis revealed that MHHX exhibited higher umami and saltiness levels than those of HH and HHX (Fig. 3E). Additionally, free amino acid analysis revealed that MHHX had the highest Asp and Glu levels among the samples (P < 0.05; Table 3). Therefore, the higher umami and flavor scores of PP-MHHX may be related to the free amino acid profile of MHHX. Bitterness scores were lowest in PP-MHHX. This result is consistent with the lower bitterness intensity of MHHX in the E-tongue analysis (Fig. 3E), and its lower content of hydrophobic amino acids (Table 3). Accordingly, the addition of MHHX, characterized by high umami and low bitterness, likely masked the bitterness of the PP and improved its palatability, resulting in the lower bitterness score of PP-MHHX (Go et al., 2022). The HH-supplemented groups showed significantly higher bloody taste scores than the PP-Con group (P < 0.05), with no significant differences among the PP-HH, PP-HHX, and PP-MHHX groups (Fig. 4D). Although PP-MHHX showed a higher ΔE after cooking (Table 4), the difference was mainly associated with reduced L^⁎, indicating that bloody taste scores were unlikely to be driven by redness-related visual cues. Moreover, PP-HH showed higher bloody taste intensity without significant changes in color parameters or sensory color score, further supporting that the bloody taste score was not substantially biased by sample color. During heating, bounded-Fe²⁺ in heme proteins could be oxidized to Fe³⁺ or released as free iron; the released free iron could bind to T2R-type taste receptors, inducing metallic or bloody taste (Zhang, Tian, et al., 2022). Moreover, intact heme structures may contribute to bloody taste characteristics, with heme proposed as a taste-active molecule involved in the characteristic taste of meat (Gerhard, 2020). In PP-MHHX, thermal denaturation and the Maillard reaction increased extractable non-heme iron content, whereas structural alteration of heme reduced heme-associated absorbance (Table S3). Therefore, the balance between released iron and residual heme-associated structures maintained the overall bloody taste characteristics of PP-MHHX, resulting in no perceptible difference from PP-HH and PP-HHX. Excessive bloody taste can adversely affect the overall quality of food products. However, incorporating MHHX enhanced the umami taste while maintaining the characteristic bloody taste of heme proteins at a desirable level. PP-MHHX had the lowest bitterness score, which was associated with reduced total hydrophobic amino acid content during the Maillard reaction (Table 3). Additionally, the elevated umami content of PP-MHHX may enhance palatability and suppress bitterness through a masking effect (Go et al., 2022). PP-MHHX received a higher texture score than PP-Con (Fig. 4D). The favorable texture perception of PP-MHHX may be associated with its stronger gel network, which contributed to improved textural properties in the PP (Fig. 4C, Table 4). Taken together, PP-MHHX showed significantly improved overall sensory characteristics, particularly in umami and texture, while maintaining a desirable level of bloody taste. The desirable sensory characteristics of PP-MHHX identified by the trained panel should be further validated through consumer preference testing involving a diverse consumer panel.

4. Conclusions

The objective of this study was to use the Maillard reaction to alter and diversify the taste characteristics of HH, thereby making it suitable for application in PPs. Optimal Maillard reaction conditions were identified as an HH–xylose mass ratio of 2:1, reaction time of 6 h, and reaction temperature of 80 °C. Under the optimized condition, MHHX showed significantly higher GD and browning intensities, together with improved structural stability (P < 0.05). The improvements in protein solubility, emulsifying capacity, and thermal stability of MHHX were attributed to the development of hydrophobic interaction and the formation of β-sheet and α-helix structures. The free amino acid composition of MHHX showed a tendency toward decreased levels of hydrophobic amino acids associated with bitterness and increased levels of umami amino acids. The addition of MHHX to PPs led to significant improvements in texture properties and provided a more meat-like color. In addition, PP-MHHX retained the characteristic bloody taste of heme protein while enhancing the umami flavor and reducing bitterness. In conclusion, the Maillard reaction effectively improved the physicochemical and sensory characteristics of HH. This study expands the applicability of HH as a food additive in plant-based meats by enhancing their umami and texture while retaining their characteristic bloody taste.

CRediT authorship contribution statement

Jong Hyeon Han: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Conceptualization. Hyun Ju Lee: Writing – review & editing, Methodology. Ji Hwan Ryoo: Writing – review & editing, Methodology. Hyun Su Jung: Writing – review & editing, Methodology. Hyuk Cheol Kwon: Writing – review & editing, Methodology, Investigation. Dong-Min Shin: Writing – review & editing, Methodology, Investigation. Chang Hee Jeong: Writing – review & editing, Methodology, Investigation. Yun-Sang Choi: Writing – review & editing, Methodology, Investigation. Sung Gu Han: Writing – review & editing, Writing – original draft, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data will be made available on request.