Effects of Different Drying Methods on Nutritional Compositions, Bioactive Substances, and Volatile Compounds of Radish (Raphanus sativus L.) Leaves

TL;DR

This study compares how different drying methods affect the nutrients and flavor compounds in radish leaves, finding that freeze-vacuum drying best preserves their quality.

Contribution

The study identifies freeze-vacuum drying as the most effective method for preserving nutrients and bioactive compounds in radish leaves.

Findings

Freeze-vacuum drying preserved proteins, lipids, minerals, and bioactive compounds better than hot-air and microwave drying.

FD-treated samples showed distinct glucosinolate and volatile profiles compared to HD- and MD-treated samples.

Multivariate analysis identified 12 key differential glucosinolates and 27 differential odor-active compounds among the drying methods.

Abstract

Radish leaves are a nutrient-rich yet underutilized byproduct containing abundant fiber, minerals, and phytochemicals; however, their quality is highly affected by drying methods. This study systematically investigated the effects of three drying methods—hot-air drying (HD), microwave drying (MD), and freeze-vacuum drying (FD)—on the nutritional components, bioactive substances, and volatile compounds of radish leaves. A comparative analysis was conducted on their proximate composition, amino acid profiles, mineral contents, antioxidant capacities, glucosinolate profiles, and volatile profiles. Among the three methods, FD exhibited superior preservation of proteins, lipids, minerals (K, Mg, P, Fe, Zn, and Mn), and bioactive components, including polyphenols, flavonoids, glucosinolates, and vitamin C. In contrast, HD and MD led to significant reductions in these nutrients and bioactive compounds. A total of 33 glucosinolates and 779 volatile compounds, including 164 odor-active compounds, were identified collectively across the three treatments. The FD-treated samples exhibited distinct glucosinolate and volatile profiles, whereas HD- and MD-treated samples showed greater similarity. Multivariate analysis further revealed 12 key differential glucosinolates and 27 differential odor-active compounds among the three groups. This study provides a scientific basis for optimizing drying strategies to improve the nutritional quality and flavor characteristics of processed radish leaves.

Figures

Full text

1. Introduction

The radish (Raphanus sativus L.) is a globally significant vegetable crop extensively cultivated across Asia, Europe, and North America [1]. It is widely consumed in many countries, particularly in China, Japan, Korea, and Southeast Asia [2]. Owing to its high content of essential minerals and phytochemicals, as well as low levels of antinutritional factors, radish offers substantial nutritional value and potential health benefits [3]. The taproot is traditionally the primary edible portion of radish, as it is rich in dietary fiber and amylase, which collectively enhance digestive function. Moreover, its low caloric density is advantageous for dietary management and weight control.

Despite the widespread consumption of radish taproots, the leaves are often considered low-value byproducts and are typically discarded during post-harvest handling. However, accumulating evidence suggests that the nutritional value of radish leaves has been substantially underestimated [4,5]. Radish leaves are rich in essential minerals such as potassium (K), calcium (Ca), magnesium (Mg), iron (Fe), zinc (Zn), as well as vitamin C (ascorbic acid), flavonoids, polyphenols, glucosinolates and other bioactive compounds. These components contribute to a broad spectrum of biological activities, including antioxidant [6,7], anticancer [8], antihypertensive [9], and acetylcholinesterase inhibitory effects [10]. Furthermore, they enhance gastrointestinal motility and support the treatment of gastric ulcers [11,12]. In Traditional Chinese Medicine (TCM), radish leaves are characterized by sweet, bitter, and neutral properties; they are primarily used to aid digestion and regulate “Qi”. The American Center for Public Science has evaluated the nutritional value of vegetables based on parameters such as calorie content, vitamin K, lutein, vitamin C, and other nutrients, revealing that radish leaves rank third among 85 vegetable types [5]. Moreover, recent nutritional studies indicate that radish leaves exhibit nutritional profiles superior to those of taproots [13,14]. For instance, the vitamin C content in radish leaves is over twice that of the taproots, whereas the contents of Ca, Mg, Fe, Zn, vitamin B_2_, and folic acid are 3 to 10 times higher [4].

Despite their high nutritional value, the elevated moisture content of fresh radish leaves presents a major post-harvest challenge, resulting in rapid spoilage and considerable resource wastage. This not only represents an economic loss but also limits the valorization of a valuable agricultural byproduct. With growing emphasis on waste reduction and value-added utilization, effective processing methods are crucial to mitigate these losses and transform this underutilized resource into stable, value-added ingredients suitable for industrial supply chains.

Drying is an effective preservation method for addressing these challenges that removes moisture to suppress microbial growth and enzymatic activity, thereby significantly extending shelf life. Common drying methods include hot-air drying (HD), microwave drying (MD), and freeze-vacuum drying (FD), each with distinct advantages and limitations. HD is a simple and cost-effective method; however, its high processing temperatures often lead to color alteration, texture hardening, and nutrient loss [15]. MD enables rapid processing through volumetric heating and offers high energy efficiency; however, it is prone to non-uniform heating, which can cause localized overheating and uneven dehydration, thereby compromising product quality [16]. FD operates via low-temperature sublimation, effectively preserving bioactive compounds and minimizing oxidative damage to yield, thereby achieving superior product quality. However, it involves prolonged processing times and high operational costs [17,18]. Different drying methods significantly influence the nutritional quality, bioactive substances, and volatile compounds of the final product [19,20,21]. The selection of an appropriate drying method is critical, as it directly determines the retention of beneficial compounds and the functional properties of the dried leaves, which in turn dictates their suitability for specific industrial applications, such as in functional foods, nutraceuticals, or as a traditional medicine ingredient. Hence, the optimal selection of a drying strategy that balances nutrient retention with comprehensive aroma characterization remains challenging. Although the impacts of drying methods have been investigated in various leafy vegetables, systematic studies specifically focusing on radish leaves remain scarce.

Therefore, the objective of this study is to systematically evaluate and compare the effects of three different drying methods (HD, MD and FD) on the nutrients, bioactive substances, and volatile compounds of radish leaves. The findings will provide a scientific basis for optimizing post-harvest processing techniques, maintaining product quality, and ultimately advancing the valorization and commercial utilization of radish leaves as a sustainable and functional agricultural resource.

2. Materials and Methods

2.1. Chemicals and Reagents

Hydrochloric acid (HCl), nitric acid (HNO_3_), perchloric acid (HClO_4_), metaphosphoric acid (HPO_3_), sodium hydroxide (NaOH), sodium nitrite (NaNO_2_), aluminum nitrate (Al(NO_3_)3), sodium chloride (NaCl), sodium carbonate (Na_2_CO_3_), ferric chloride hexahydrate (FeCl_3_·6H_2_O), ferrous sulfate heptahydrate (FeSO_4_·7H_2_O), sodium acetate (CH_3_COONa), ethanol, and phenol were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Chromatographic-grade methanol and formic acid were purchased from Shanghai Aladdin Bio-Chemical Technology Co., Ltd. (Shanghai, China). Sodium tetrachloropalladate(II) (Na_2_PdCl_4_), potassium persulfate (K_2_S_2_O_8_), dithiothreitol (DTT), L-ascorbic acid, Folin–Ciocalteu reagent, gallic acid, rutin, sinigrin, 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS), 6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid (Trolox), and 2,4,6-tris(2-pyridyl)-s-triazine (TPTZ) were purchased from Merck (Darmstadt, Germany).

2.2. Raw Material and Drying Methods

2.2.1. Raw Material

Fresh radish leaves of the ‘Weixianqing’ cultivar (a common green radish variety in China) were harvested 70 days post-sowing from the experimental field of the Institute of Vegetables, Shandong Academy of Agricultural Sciences. Uniform leaves with intact surfaces and no visible defects or insect damage were selected. Leaves from three individual plants were pooled to form one biological sample. The samples were then washed thoroughly with distilled water to remove surface impurities without detergents. Following rinsing, surface moisture was removed by blotting with sterile cotton fabric prior to drying treatments. Each drying treatment was performed in three independent experimental replicates.

2.2.2. Drying Procedure

For HD, radish leaves were evenly spread and dried in an electric constant-temperature drying oven (6CHT-7.0, Zhejing Chunjiang Tea Machinery Co., Ltd., Hangzhou, China) at 60 °C until constant weight was achieved (weight change of less than 0.1 g over a 1 h interval), which typically required approximately 12 h.

For MD, radish leaves were uniformly arranged on trays and dried in a microwave (KH-6HPTN, Shandong Kehong Microwave Energy Co., Ltd., Jinan, China) at a power of 500 W under continuous operation until constant weight was achieved (as per the same criterion), with a processing time of approximately 2 h.

For FD, radish leaves were freeze-dried in a vacuum freeze dryer (LGJ-25C, Sihuan Furuikeyi Technology Development Co., Ltd., Beijing, China) at −50 °C under a vacuum pressure of 0.1 kPa for 48 h. The sample was removed once the instrument temperature reached room temperature and the sample weight stabilized (constant weight criterion).

Subsequently, all dried radish leaves were ground using a blade mixer (FW100, Taisite Instrument Co., Ltd., Tianjin, China) and sieved through 60-mesh screens.

2.3. Determination of Proximate Composition

The concentrations of nutrients (moisture, ash, protein, fat, and dietary fiber) in HD-, MD- and FD-treated samples were determined according to the methods specified in the Chinese national food safety standards: GB 5009.3-2016 (for moisture), GB 5009.4-2016 (for ash), GB 5009.5-2025 (for protein), GB 5009.6-2016 (for fat), and GB 5009.88-2023 (for dietary fiber). Carbohydrate content was calculated based on the following formula that complies with AOAC (Association of Official Analytical Chemists) guidelines: Carbohydrates (g) = Total Mass (g) − (Protein (g) + Fat (g) + Moisture (g) + Ash (g) + Dietary Fiber (g)) [22].

2.4. Determination of Amino Acid Profiles

The amino acid composition in HD-, MD- and FD-treated samples was determined using an LA8080 amino acid analyzer (Hitachi High-Tech Corporation, Tokyo, Japan). Briefly, 0.2 g of the sample was hydrolyzed with 10 mL of 6 mol/L HCl and three drops of 1% (v/v) phenol in an amino acid hydrolysis tube. The mixture was frozen at −20 °C for 5 min, followed by rapid sealing under a nitrogen atmosphere. The sealed tube was incubated at 110 °C for 22 h. After hydrolysis, the hydrolysate was filtered and diluted to 250 mL with purified water, and a 1 mL aliquot was transferred to a centrifuge tube. The residue was concentrated to near-dryness, re-dissolved in 1 mL of pure water, and evaporated to dryness again. The final residue was then dissolved in 0.02 mol/L HCl and filtered through a 0.22-μm membrane filter to obtain the solution for analysis. Amino acids were separated on a sulfonic acid-type cation exchange resin column and detected at 570 nm and 440 nm. The results were expressed as mg/g dry weight (DW).

2.5. Determination of Mineral Elements

The HD-, MD- and FD-treated samples were subjected to wet digestion [23]. Briefly, 0.2 g of each sample was accurately weighed and combined with 10 mL of a mixed acid solution (HNO_3_:HClO_4_ = 4:1, v/v) in conical flasks. The mixture was soaked overnight in a fume hood, followed by digestion on an adjustable hot plate at 120 °C for 2–3 h until the solution volume reduced to 1–2 mL. The residue was then diluted to a final volume of 20 mL with pure water. The concentrations of K, sodium (Na), Ca, Mg, Fe, Zn, manganese (Mn), and copper (Cu) were determined by flame atomic absorption spectrometry, conducted in accordance with the following Chinese national food safety standards: GB 5009.91-2017 (for K and Na), GB 5009.92-2016 (for Ca), GB 5009.241-2017 (for Mg), GB 5009.90-2016 (for Fe), GB 5009.14-2017 (for Zn), GB 5009.242-2017 (for Mn), and GB 5009.13-2017 (for Cu). Phosphorus (P) content was quantified by molybdenum blue spectrophotometry according to national standard GB 5009.87-2016.

2.6. Determination of the Main Bioactive Substances

2.6.1. Total Phenolic Content (TPC) and Total Flavonoid Content (TFC)

Total phenolic and total flavonoid extracts were prepared according to the method of Li et al. [24], with minor modifications. Briefly, 0.1 g of the sample was mixed with 0.5 mL of 60% (v/v) ethanol, followed by ultrasonic-assisted extraction at 60 °C for 30 min. After extraction, the mixture was centrifuged at 12,000 rpm for 10 min at 25 °C, and the resulting supernatant was collected for subsequent analysis.

TPC was determined using the Folin–Ciocalteu method [25]. In brief, 10 μL of extract was mixed with 50 μL of Folin–Ciocalteu reagent and incubated at room temperature for 2 min. Then, 50 μL of 12% (w/v) Na_2_CO_3_ and 90 μL of pure water were added. After reacting at room temperature for 10 min, the absorbance of the mixture was measured at 765 nm using a SpectraMax ABS Single-Mode Reader (Molecular Devices, San Jose, CA, USA). A standard curve was constructed using gallic acid concentrations ranging from 0.02 to 0.80 mg/mL (y = 3.546x + 0.0407; R^2^ = 0.9998). The results were expressed as mg gallic acid equivalent/g dry weight (mg GAE/g DW).

TFC was determined using the NaNO_2_-Al(NO_3_)3 colorimetric method [26]. Briefly, 20 μL of extract was mixed with 6 μL of 5% NaNO_2_ and incubated at room temperature for 6 min. Subsequently, 6 μL of 10% Al(NO_3_)3 solution was added, mixed, and incubated at room temperature for an additional 6 min. Finally, 80 μL of 4% NaOH solution and 88 μL of pure water were added, followed by thorough mixing and a final incubation of 15 min under the same conditions. The absorbance was measured at 510 nm. A standard curve was prepared using rutin concentrations ranging from 0.1 to 1.0 mg/mL (y = 0.6453x + 0.0235; R^2^ = 0.9993). The results were expressed as mg rutin equivalent/g dry weight (mg RE/g DW).

2.6.2. Total Glucosinolate Content (TGC)

Total glucosinolates were extracted and quantified following the method of Aghajanzadeh et al. [27]. Briefly, a 0.1 g sample was extracted with 3 mL of boiling 90% (v/v) methanol for 2 min. The extract was centrifuged at 2500 rpm for 2 min, and the residue was re-extracted twice with 3 mL of boiling 70% (v/v) methanol. Glucosinolate content was determined based on the reaction with Na_2_PdCl_4_. Specifically, a reaction mixture containing 60 µL of the extract and 1800 µL of 2 mmol/L Na_2_PdCl_4_ (58.8 mg Na_2_PdCl_4_ + 170 µL concentrated HCl + 100 mL pure water) was incubated at room temperature for 30 min, and the absorbance of the mixture was measured at 450 nm. A standard calibration curve was constructed using sinigrin concentrations ranging from 0.02 to 0.80 mg/mL (y = 0.8372x − 0.0285; R^2^ = 0.9978). The results were expressed as µmol sinigrin equivalent/g dry weight (µmol SE/g DW).

2.6.3. Vitamin C Content

Vitamin C was identified and quantified according to the method described by Varo et al. [28]. Briefly, 0.1 g of the sample was mixed with 1 mL of 4.5% (w/v) HPO_3_, followed by ultrasonication for 5 min and centrifugation at 4000 rpm for 5 min. Subsequently, 0.5 mL of the supernatant was mixed with 0.2 mL of DTT solution. The mixture was incubated in darkness at room temperature for 2 h to ensure complete reduction of dehydroascorbic acid to L-ascorbic acid. After conversion, the sample was filtered through a 0.45-μm nylon membrane. Ascorbic acid content was determined by HPLC with detection set at 243 nm. The ascorbic acid content was expressed as mg/g DW.

2.7. Determination of Antioxidant Activities

The antioxidant activities of HD-, MD- and FD-treated samples were evaluated by measuring the ABTS radical scavenging capacity and ferric reducing antioxidant power (FRAP). The extracts prepared for both assays followed the method of Siriamornpun et al., with minor modifications [29]. Briefly, 0.1 g of sample was homogenized with 1 mL of 80% (v/v) ethanol, vigorously vortexed, and centrifuged at 10,000 rpm for 10 min at 4 °C. The resulting supernatant was collected and maintained on ice until further analysis.

The ABTS assay was performed according to the method of Van der Werf et al. [30] with minor modifications. Briefly, 5 mL of 7 mmol/L ABTS aqueous solution was mixed with 88 μL of 140 mmol/L K_2_S_2_O_4_ solution and incubated in the dark at room temperature for 12–16 h to generate the ABTS^+^ radical cation. The resulting reaction solution was diluted with anhydrous ethanol to an absorbance of 0.70 ± 0.02 at 734 nm prior to use. Subsequently, 10 μL of the extract was added to 200 μL of the diluted ABTS^+^ solution, vortexed gently, and allowed to react at room temperature for 10 min in the dark. The absorbance was measured at 734 nm. A standard curve was constructed using Trolox at concentrations ranging from 0.05 to 0.80 μmol/mL (y = 0.7610x + 0.0004; R^2^ = 0.9991). The results were expressed as µmol Trolox equivalent/g of dry weight (µmol TE/g DW).

The FRAP assay was determined based on a modified method of Sompong et al. [31]. The FRAP working solution was freshly prepared by mixing 300 mM acetate buffer (pH = 3.6), 10 mM TPTZ solution, and 20 mM FeCl_3_ in a 10:1:1 ratio (v/v/v), followed by incubation at 37 °C in a water bath for 5 min. Then, 6 μL of the extract was mixed with 180 μL of the FRAP working solution and 14 μL of pure water, vortexed thoroughly, and allowed to react in the dark at room temperature for 10 min. The absorbance was measured at 593 nm. A standard curve was computed using FeSO_4_ at concentrations ranging from 0.005 to 0.100 μmol/mL (y = 13.591x − 0.0043, R^2^ = 0.9994). The results were expressed as µmol of FeSO_4_ equivalent/g dry weight (µmol FeSO_4_ eq./g DW).

2.8. Determination of Glucosinolate Components

The determination of glucosinolate components was performed using ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS). A total of 50 mg of sample was extracted with 1.2 mL of 70% methanol (v/v) pre-cooled at −20 °C. The extract was vortexed for 30 s at 30 min intervals over 6 cycles, followed by centrifugation at 12,000 rpm for 6 min at 4 °C. The resulting supernatant was filtered through a 0.22-μm membrane prior to LC-MS/MS analysis.

Sample extracts were analyzed using a UPLC-ESI-MS/MS system (UPLC: ExionLC™ AD, SCIex, Framingham, MA, USA; MS: Applied Biosystems 6500 Triple Quadrupole, Applied Biosystems, Foster City, CA, USA) equipped with an Agilent SB-C18 column (1.8 μm, 2.1 mm × 100 mm). The mobile phases consisted of water containing 0.1% formic acid (Solvent A) and methanol containing 0.1% formic acid (Solvent B). The elution gradient was as follows: 5% B at 0 min, increasing linearly to 95% B within 9 min and held at 95% B for 1 min. From 10 to 11.1 min, the proportion of B decreased to 5%, followed by re-equilibration at 5% B until 14 min. The flow rate was set to 0.35 mL/min, the column temperature to 40 °C, and the injection volume to 2 μL. The effluent was alternatively connected to an ESI-triple quadrupole-linear ion trap (QTRAP)-MS.

The ESI source parameters were as follows: source temperature, 500 °C; ion-spray voltage (IS), 5500 V in positive ion mode and −4500 V in negative ion mode; ion source gas I (GSI), gas II (GSII), and curtain gas (CUR) set to 50, 60, and 25 psi, respectively. Collision-activated dissociation energy was set to high, and data were acquired in multiple reaction monitoring (MRM) mode. Nitrogen was used as the collision gas, with declustering potential (DP) and collision energy (CE) optimized for each MRM transition. A specific set of MRM transitions was monitored for each period based on the metabolites eluted in that period. Data acquisition and quantification were carried out using Analyst software v1.6.3 and Multiquant software v3.0.3.

2.9. Determination of Volatile Compounds

The detection of volatile compounds was performed using headspace-solid phase microextraction coupled with gas chromatography-mass spectrometry (HS-SPME-GC-MS), following the method described by Luo et al. [32]. For HS-SPME, 0.5 g of sample was placed into a headspace vial, and 5 mL of saturated NaCl solution along with 20 µL of internal standard solution (10 µg/mL 3-Hexanone-2,2,4,4-d4) were added. After shaking at 60 °C for 5 min, a 120 μm DVB/CWR/PDMS extraction fiber was inserted into the headspace for 15 min of extraction, followed by desorption at 250 °C for 5 min. Prior to sampling, the fiber was conditioned at 250 °C for 5 min.

GC-MS analysis was conducted using a GC-MS system equipped with a DB-5MS capillary column. High-purity helium (purity ≥ 99.999%) served as the carrier gas at a constant flow rate of 1.2 mL/min. The inlet temperature was maintained at 250 °C with a solvent delay of 3.5 min. The initial oven temperature was set at 40 °C for 3.5 min, gradually increased to 100 °C at a rate of 10 °C/min, then to 180 °C at a rate of 7 °C/min, and finally to 280 °C at a rate of 25 °C/min, and then the temperature was held for 5 min. The mass spectrometry conditions were as follows: electron impact ion source (EI) with an ion source temperature of 230 °C, quadrupole temperature of 150 °C, and mass spectrometry interface temperature of 280 °C; electron energy was set to 70 eV. Data acquisition was performed in selected ion monitoring (SIM) mode to enable precise detection of target qualitative and quantitative ions.

Raw mass spectral data were processed using MassHunter software vB.07.00 and compared against a self-constructed MWGC database for compound identification and quantification. Quantitative analysis was performed using a semi-quantitative method [33,34]. The relative content of each metabolite was calculated by comparing the peak areas of a given component to those of the internal reference compound. The odor activity value (OAV) of each volatile compound was calculated as OAV = Ci/Ti, where Ci was the content of compound i (µg/kg) and Ti was the threshold of compound i (µg/kg).

2.10. Statistical Analysis

All assays were performed in triplicate, and the results were expressed as mean ± standard deviation (SD). Statistical analysis was conducted using one-way ANOVA followed by Tukey’s honestly significant difference (HSD) post hoc test for multiple comparisons with SPSS Statistics v30.0.0 (IBM, New York, NY, USA). Principal component analysis (PCA) was performed using the MetaboAnalyst online platform (https://www.metaboanalyst.ca/, accessed on 21 September 2025), and partial least squares discriminant analysis (PLS-DA) was carried out using SIMCA software v14.1.0 (Umetrics, Umeå, Sweden).

3. Results and Discussion

3.1. Proximate Composition

The ash, protein, lipid, dietary fiber, and total carbohydrate contents of radish leaves subjected to different drying methods are presented in Table 1. Among the three samples, dietary fiber showed the highest content (32.57–34.89 g/100 g), followed by total carbohydrates (27.84–28.22 g/100 g), protein (13.46–15.41 g/100 g), and ash (13.49–15.05 g/100 g). Lipid showed the lowest content (0.20–0.24 g/100 g). Hence, dried radish leaves exhibited a fiber-rich, low-lipid profile with abundant protein and minerals, suggesting their potential as a valuable nutritional source for high-fiber foods, plant-based products, and functional ingredients [4].

The drying method significantly affected the ash, protein, lipid, and dietary fiber contents of radish leaves (p < 0.05), whereas total carbohydrate content did not differ significantly (p > 0.05) among the three drying methods. The FD-treated samples (FD group) exhibited significantly higher ash, protein, and lipid contents than both the HD- and MD-treated samples (HD group and MD group, respectively) (p < 0.05). For dietary fiber, the FD group showed a significantly higher content than the HD group, while it did not differ significantly (p > 0.05) from the MD group. The HD group had the lowest content of ash, lipid, and dietary fiber, whereas the MD group had the lowest protein content (p < 0.05).

The relatively higher protein and lipid contents in the FD group were likely due to the combined effects of biochemical preservation and physical structural integrity [35]. Firstly, the low-temperature conditions of FD effectively inhibited protease and lipase activities, thereby reducing enzymatic degradation. Secondly, FD better preserved cellular structure through sublimation, which minimized tissue collapse and reduced the leakage of soluble solids. In contrast, hot-air drying and microwave drying are performed at higher temperatures, which can accelerate lipid oxidation and protein denaturation and may also induce microstructural damage (e.g., cell rupture) [36]. Such structural disruption likely increased drip loss and facilitated the leaching of water-soluble proteins and other solutes, thereby reshaping the composition of the retained dry matter (even on a dry-weight basis) [37]. Moreover, during microwave heating, rapid accumulation of thermal energy within the samples generated local hot spots, which accelerated protein denaturation and Maillard reactions, ultimately resulting in reduced protein content [37,38]. This mechanism might contribute to the significantly lower measured protein content in the MD group compared with the HD group.

3.2. Analysis of Amino Acid Profiles

Amino acids are important intermediates in plant metabolism and play a key role in the nutritional value and flavor quality of foods [39]. The total amino acid content of the FD, MD, and HD groups ranged from 154.09 to 173.23 mg/g, with the highest content observed in the FD group and the lowest in the HD group (p < 0.05) (Table 2). The essential amino acid content was 60.92–64.73 mg/g, accounting for 37–40% of the total amino acids. This proportion is comparable to values reported for Coccinia abyssinica leaves (TEAA/TAA ≈ 36.8%) [40], Moringa oleifera leaves (≈40%) [41], purslane (≈41%) [42], and spinach (≈41%) [43], but slightly lower than those reported for dried kale (42–44%) [43,44] and quinoa green leaves (≈47.4%) [45], suggesting that radish leaves have a relatively balanced amino acid composition and potential for nutritional applications.

For essential amino acids, the FD group had the highest content of lysine (11.14 mg/g), isoleucine (8.92 mg/g), and total essential amino acids (64.73 mg/g), whereas the HD group had the lowest content (p < 0.05). The contents of leucine, phenylalanine, threonine, and valine did not differ significantly (p > 0.05) among the different drying groups. Moreover, methionine showed the highest content in the MD group (1.36 mg/g) (p < 0.05), exhibiting a distinct change pattern compared with the other essential amino acids.

For non-essential amino acids, the FD group had significantly higher contents (p < 0.05) of glutamic acid, aspartic acid, histidine, alanine, arginine, proline, and serine than the HD group. Among these amino acids, aspartic acid, alanine, arginine, proline, and serine showed the lowest content in the HD group, whereas histidine did not differ significantly (p > 0.05) between the HD and MD groups. In the FD group, the contents of glutamic acid, histidine, arginine, and proline were significantly higher than those in the MD group (p < 0.05), whereas the contents of aspartic acid, alanine, and serine did not differ significantly (p > 0.05) between them. Given that glutamic acid and aspartic acid are typical umami-related amino acids, the FD group might have better flavor potential than both HD and MD groups. Glycine and tyrosine did not differ significantly (p > 0.05) among the three drying groups, indicating that they were relatively insensitive to the drying conditions.

From a nutritional perspective, the amino acid profile of dried radish leaves suggests their potential as a valuable plant-protein ingredient. Moreover, the specific amino acid pattern highlighted its functional relevance for food applications. In particular, the higher contents of flavor-enhancing amino acids (glutamic acid and aspartic acid) in the FD group implied that radish leaf could serve as a natural umami enhancer in soups, seasonings, and ready-to-eat products. Additionally, the appreciable levels of branched-chain amino acids (leucine, isoleucine, and valine) supported their potential incorporation into plant-based meat analogs to enhance nutritional value and savory attributes [46].

3.3. Analysis of Mineral Elements

In this study, as shown in Table 3, K had the highest content among the macrominerals (K, Na, Ca, Mg, and P) in radish leaves subjected to different drying methods, with a content of approximately (2.18–3.22) × 10^4^ mg/kg, followed by Na and Ca. Among the trace elements (Fe, Zn, Mn, Cu, and Se), Fe had the highest content, at about (1.87–2.24) × 10^2^ mg/kg, whereas Zn, Mn, and Cu were present at relatively lower contents. Overall, dried radish leaves represented a plant-based mineral source rich in K, Na, Ca, and Fe while also providing several essential trace elements.

In terms of macrominerals, the FD group had the highest K content and the lowest Na content (p < 0.05), resulting in the highest K/Na ratio. Ca showed only minor variation, with no significant differences among the three drying groups (p > 0.05). The contents of Mg and P in the FD group were significantly higher than those in both HD and MD groups (p < 0.05). For trace elements, the FD group generally showed higher contents of Fe, Zn, and Mn than the HD group (p < 0.05), with the MD group mostly exhibiting intermediate values. The content of Cu was comparable among the three drying methods (p > 0.05). Overall, freeze-vacuum drying showed a comparative advantage in preserving most trace elements.

Since mineral elements remain chemically stable under thermal, photic, and oxidative stress, the observed variations in mineral content likely resulted from the migration and physical loss of water-soluble salts rather than elemental degradation [47]. K and Na are mainly present as ions in vacuoles and apoplastic spaces [48,49]. During hot-air and microwave drying, the coexistence of elevated temperatures and liquid water likely caused disruption of cell structures and leakage of cell sap, leading to the loss of some soluble ions together with exuded surface liquids or residues [50,51]. Freeze-vacuum drying achieved dehydration via sublimation of ice under low-temperature and vacuum conditions, which better preserved cell integrity and reduced liquid exudation [52], thereby improving the retention of K, Na, and other water-soluble minerals. Therefore, radish leaves processed by freeze-vacuum drying not only had higher levels of K, Mg, P, and several trace elements but also exhibited a significantly higher K/Na ratio (2.65) than those processed by hot-air and microwave drying, suggesting greater potential for optimizing the mineral composition of final products.

3.4. Impact of Drying Methods on Bioactive Substances and Antioxidant Capacity

3.4.1. Analysis of Phenolic, Flavonoid, and Glucosinolate Contents

As shown in Figure 1A–C, the drying methods had a significant effect on the contents of bioactive substances in radish leaves. Freeze-vacuum drying achieved the highest retention of polyphenols, flavonoids, and glucosinolates, whereas hot-air drying resulted in the greatest losses.

The TPC of the FD group reached 8.29 ± 0.23 mg GAE/g, which was significantly higher than that of the MD group (7.13 ± 0.15 mg GAE/g) and the HD group (6.22 ± 0.25 mg GAE/g) (p < 0.05) (Figure 1A). These differences were attributed to distinct patterns of enzymatic reactions and thermal degradation under different drying kinetics [53,54]. Despite higher operating temperatures, microwave drying resulted in a significantly higher TPC than hot-air drying. This was likely due to rapid dehydration shortening the thermal exposure period, combined with microwave-induced cell disruption that facilitated the release of bound phenolics.

TFC of the FD group was significantly higher than that of the HD group (p < 0.05), whereas the MD group did not differ significantly (p > 0.05) from both the FD and HD groups (Figure 1B). TGC was much more sensitive to the drying methods (Figure 1C). TGC in the FD group (24.75 µmol SE/g) was approximately 2.3 times higher than that of the HD group (10.84 µmol SE/g). The pronounced loss of TGC in the HD group was likely due to enzymatic hydrolysis of glucosinolates in the early heating stage, when damage to cellular membranes led to the formation of volatile isothiocyanates that were subsequently removed by the hot-air stream [55].

3.4.2. Vitamin C Stability and Antioxidant Mechanisms

The vitamin C content in the FD group was the highest (approximately 5.20 mg/g) and was significantly greater than that in both MD and HD groups (p < 0.05) (Figure 1D). Antioxidant capacity was evaluated using ABTS scavenging activity and FRAP assays (Figure 1E,F). The results showed that the antioxidant activity closely mirrored the retention patterns of bioactive compounds, following the order FD > MD > HD. The FD group exhibited the strongest ABTS scavenging capacity (544.63 µmol TE/g) and FRAP value (62.91 µmol FeSO_4_ eq./g) (p < 0.05), which could be attributed not only to its higher concentrations of phenolic compounds but also to the synergistic antioxidant contribution of vitamin C [56].

In summary, freeze-vacuum drying utilizes sublimation to bypass liquid water-mediated enzymatic reactions and high-temperature thermal degradation, making it the optimal technique for preserving bioactive compounds in radish leaves. While microwave drying offers superior processing efficiency, it is insufficient for retaining heat-labile constituents like glucosinolates and vitamin C compared to freeze-vacuum drying. Nevertheless, the application of freeze-vacuum drying is accompanied by significant challenges, including high energy consumption, prolonged processing time, and limitations in industrial scalability, which must be carefully considered for large-scale practical implementation. In addition, these findings provide a foundational comparison of drying methods under controlled conditions using a single radish cultivar (‘Weixianqing’). To fully translate these insights into broad industrial application, future work will aim to validate the observed effects across a wider range of radish cultivars, varying agricultural practices, and multiple harvest periods.

3.5. Analysis of Glucosinolate Profiles

Glucosinolates are sulfur-containing compounds predominantly found in cruciferous plants. These compounds are considered beneficial to human health primarily due to the bioactive derivatives generated upon hydrolysis by the endogenous enzyme myrosinase. The findings in this study demonstrate that different drying methods significantly affect TGC in radish leaves. However, further investigation is required to elucidate compositional changes in individual glucosinolates. Thirty-three glucosinolates were identified by UPLC-ESI-MS/MS, including 23 aliphatic, 5 aromatic, and 5 indole glucosinolates (Table S1). A clustering heatmap based on the glucosinolate profiles revealed clear separations among the three drying groups (Figure S1). Specifically, the HD and MD groups clustered closely together, indicating a high similarity in their compositional patterns, whereas both were distinctly separated from the FD group. To further explore variations in glucosinolate composition across the groups, PCA was performed; the first two components (PC1 and PC2) accounted for 64.20% and 26.12% of the total variance, respectively (Figure 2A). All samples were clearly separated into three distinct clusters according to their respective treatment groups. Notably, the FD group was well distinguished from the HD and MD groups along PC1, consistent with the clustering heatmap. These findings collectively indicate that the drying method markedly influences the glucosinolate composition in radish leaves, with freeze-vacuum drying leading to a distinct profile compared to hot-air drying and microwave drying.

To identify differential glucosinolate compositions resulting from different drying methods, both multivariate analysis using PLS-DA and univariate analysis via one-way ANOVA were performed. The PLS-DA model showed clear separation among the three treatment groups, with R^2^X, R^2^Y, and Q^2^ values of 0.933, 0.997, and 0.983, respectively (Figure 2B). Since both R^2^Y and Q^2^ exceed 0.5, the model exhibits strong explanatory and predictive capacity. Furthermore, a permutation test with 200 iterations was conducted to evaluate potential overfitting. The results showed that the intercepts of R^2^ and Q^2^ were within acceptable ranges (R^2^ = [0, 0.460], Q^2^ = [0, −0.221]), confirming the validity and goodness of fit of the model (Figure 2C). Based on the model outcomes, variables with a variable importance in projection (VIP) > 1.0 and one-way ANOVA p < 0.05 were considered statistically significant. Consequently, 12 glucosinolates were identified as significantly different across the three drying treatments (Figure 2D). The five glucosinolates with the highest VIP scores were 2-ethylbutyl glucosinolate, glucohirsutin, 1-methylbutyl glucosinolate, tert-butyl glucosinolate, and 6-hydroxyhexyl glucosinolate.

3.6. Analysis of Volatile Compounds

Odor plays an important role in food quality evaluation and consumer preference [57]. To better understand how odor changes under different drying treatments, volatile compounds were analyzed by HS-SPME-GC-MS. A total of 779 volatile compounds were identified, including 146 esters, 101 terpenoids, 96 ketones, 89 heterocyclic compounds, 70 alcohols, 61 hydrocarbons, 56 aldehydes, 34 amines, 34 acids, 27 aromatics, 27 phenols, and 38 other compounds (Figure 3A, Table S2). The MD group showed the highest total volatile content (424.24 μg/g), followed by the HD group (355.00 μg/g), and the FD group had the lowest content (293.93 μg/g) (p < 0.05) (Figure 3B). These differences may result from the volatilization of low-volatility compounds and interconversion among various chemical species. As shown in Figure 3C, the MD group consistently yielded the highest contents of most volatile compounds relative to the other groups, with ketones, esters, and terpenoids all reaching their highest levels under this treatment. In contrast, heterocyclics, amines, and phenols were present at lower contents in the FD group. The influence of treatment type on volatile composition varied markedly across categories, with MD treatment generally promoting the accumulation of most compound classes. To further investigate the differences in volatile compounds among the three drying groups, a clustering heatmap analysis and PCA were conducted. The results revealed that the FD group was distinctly separated from the HD and MD groups (Figure 3D and Figure S2), which were closely clustered—a pattern consistent with the glucosinolate profiles.

It is worth noting that not all volatile compounds contribute to food aroma [58]. The sensory impact of a volatile compound depends not only on its concentration and odor characteristics but also critically on its odor threshold—the minimum concentration at which the compound can be perceived [59]. The odor activity value (OAV), defined as the ratio of the concentration of a compound to its odor threshold, is widely used to assess its potential contribution to overall aroma [60,61]. Compounds with an OAV > 1 are generally regarded as having aroma activity and play a significant role in shaping the overall aroma profile [62]. A total of 164 odor-active compounds with OAV > 1 were identified across the three drying groups, including 27 heterocyclic compounds, 26 esters, 23 ketones, 20 aldehydes, 16 alcohols, 15 terpenoids, 14 aromatics, 11 phenols, 3 amines, 2 hydrocarbons, and 7 other compounds (Table S3), indicating the important contribution of these compounds to the odors of radish leaves.

To further identify key differential odor-active compounds among the three drying groups, PLS-DA and one-way ANOVA were performed. The PLS-DA model based on the 164 compounds with OAV > 1 demonstrated robust performance (R^2^X = 0.884, R^2^Y = 0.991, Q^2^ = 0.974), showing clear separation among the three treatment groups (Figure 4A). A permutation test with 200 iterations indicated that the model was reliable with no overfitting (R^2^ = [0, 0.321], Q^2^ = [0, −0.210]) (Figure 4B). Through screening with VIP > 1 and p < 0.05, 27 key differential compounds were identified (Figure 4C), which are likely responsible for the aroma differences observed across the three drying methods. The top five compounds ranked by VIP were naphthalene, decahydro-; 2,3-diethylpyrazine; cyclohexaneacetic acid, ethyl ester; 2-ethoxy-3-methylpyrazine; and 1,2-cyclopentanedione, 3-methyl-.

4. Conclusions

This study systematically evaluated the effects of hot-air drying, microwave drying, and freeze-vacuum drying on the nutritional quality, bioactive constituents, and volatile compounds of radish leaves. Freeze-vacuum drying resulted in higher protein and lipid contents and better retention of key minerals (e.g., K, Mg, P, and Fe), thereby maintaining a favorable K/Na balance. Furthermore, the FD group maximized the retention of vitamin C, polyphenols, flavonoids, and glucosinolates, yielding the strongest antioxidant capacity among the three treatments. Moreover, the three drying methods markedly reshaped the glucosinolate patterns and volatile profiles of radish leaves. Collectively, these results highlight drying as a critical lever that simultaneously regulates nutritional preservation and aroma characteristics in leafy byproducts, providing a mechanistic basis for quality-oriented process selection. Building on these findings, drying methods could be selected strategically according to product goals. Freeze-vacuum drying is recommended for premium applications where nutrient and bioactive retention are prioritized, such as functional powders, clean-label fortification of bakery and noodle products, and plant-based foods requiring enhanced nutritional quality. By contrast, hot-air drying and microwave drying remain practical options for large-scale, cost-sensitive production of blended vegetable powders and seasoning bases, provided that process conditions are optimized to mitigate thermal/oxidative deterioration. Overall, this work provides process-selection guidance to support the industrial upcycling of radish leaves and their incorporation into sustainable food systems.