Separation of Bioactive Compounds from Pfaffia glomerata: Drying, Green Extraction, and Physicochemical Properties

Abstract

Leaves (LV), stems (STs), and inflorescences (IFs) of Pfaffia glomerata are usually discarded despite containing various bioactive compounds, especially β-ecdysone saponin. The objective was to optimize by desirability (DI) the ultrasound-assisted extraction (UAE) of bioactive compounds (total phenolics (TPCs), antioxidant activity (AA), and total saponins) from the aerial parts (LV, ST, and IF) of P. glomerata. Ideal drying conditions were determined and the drying kinetics were evaluated. LV, STs, and IFs were dried and extracted (0.06 g/mL 80% EtOH) in a USS (6 cm × 12 mm, pulse 3/6 s) by Central Composite Design (CCD), varying sonication power (140–560 W) and time (11–139 min), with TPC, AA by DPPH, and total saponin content as responses. The DI indicated that the higher TPC, AA, and saponin levels were obtained at 136.5 min and 137.87 W (STs), and 138.6 min and 562.32 W (LV and IFs). IF extracts contained higher saponin, TPCs, and AA. Higher β-ecdysone levels (3.90 mg g⁻¹) were present in the leaves. Several phenolics were detected in area parts of P. glomerata, the most abundant being p-coumaric acid (LV) and nicotinic acid (STs and IFs). These compounds provide potential health benefits. Phytol was found in all extracts. Extracts by UAE from leaves have antibacterial potential, with demonstrated inhibitory effects against S. aureus, E. coli, L. monocytogenes, S. Typhi, and P. aeruginosa, and presented bactericidal effects against E. coli, L. monocytogenes, and S. Typhi. Aerial parts of P. glomerata can be used to obtain extracts by UAE rich in bioactive compounds, providing complete utilization of the plant and sustainability to cultivation. This work represents the first report on the application of ecofriendly UAE techniques to extract bioactive compounds from the aerial parts of Brazilian ginseng.

1. Introduction

Brazil has an enormous biodiversity of plants that can have great medicinal, food, and cosmetic potential, with more than 55,000 species being cataloged [1]. In this context, plant species need to be better researched to develop new natural products, technologies, and improve processes (cultivation, harvest, production, storage, drying, extraction, separation, etc.) in the aromatic and medicinal plants chain. An important aspect of this process is the extraction and separation of bioactive compounds from plants, which play a crucial role in determining the biological activity, efficacy, and safety of plant-derived products [2,3]. Various techniques can be employed for this purpose, including solvent extraction, supercritical fluid extraction, and chromatography-based methods [4,5,6]. These processes enable the isolation and purification of bioactive molecules, ensuring their effectiveness and stability in pharmaceutical, cosmetic, and food applications. Accordingly, Pfaffia glomerata (Spreng.) Pedersen is of great medicinal interest due to properties like tonic, aphrodisiac, anti-inflammatory, and analgesic [7]. Brazilian Indians have been using P. glomerata for centuries to cure and prevent diseases, but the first scientific studies to prove its medicinal properties took place when they were taken to Japan [8]. Pfaffia glomerata belongs to the Amaranthaceae family, which has a cosmopolitan distribution and includes about 170 genera and 2000 species, and in Brazil there are 20 native genera and approximately 100 species [9]. P. glomerata is an endemic plant found in the islands and floodplains of the Paraná River located between the states of Paraná, São Paulo, and Mato Grosso do Sul. The plant’s cultivation in this region began in the mid-1980s, by farmers remaining from the Landless Movement (MST). The cultivation of P. glomerata is considered a potential economic alternative for the Pantanal region [9,10]. It stands out for having chemical compounds with great potential for pharmaceutical applications and medicinal properties such as combating physical and mental exhaustion, lack of memory, and prevention of anemia, among others [11,12,13,14,15]. Studies associate these claims with the presence of saponins and ecdysteroids, particularly β-ecdysone [16,17]], to which the medicinal properties are attributed. P. glomerata has promising pharmacological properties and economic potential, and further research is important for unlocking its full potential since drying until separations of bioactive compounds for pharmaceutical, cosmetic, and food applications.

This work focused on optimizing the ultrasound-assisted extraction (UAE) process, utilizing desirability (DI), to obtain maximum yields of bioactive compounds, specifically total phenolics (TPCs), antioxidant activity (AA), and total saponins, from the leaves (LV), stems (STs), and inflorescences (IFs) of P. glomerata.

2. Materials and Methods

2.1. Sample Collection and Preparation

To conduct the tests, ten freshly collected individuals of P. glomerata, aged 4 years and exhibiting similar characteristics (development, appearance, and absence of pests), were selected from a local cultivation (23°45′51” S, 53°19′6” W; Umuarama, Paraná, Brazil). The average initial moisture content of Brazilian ginseng stems, leaves, and inflorescences were determined by vacuum drying at 105 °C for 8 h. The samples were separated into stems (STs), leaves (LV) and inflorescences (IFs), cleaned and cut into ~5 mm and stored at −4 ± 1 °C.

The samples (STs, LV, and IFs) were dried in a forced-air oven (Ethik Technology/400-4ND, Vargem Grande Paulista, Brazil) at different drying temperatures (40, 60, and 80 °C) and hot air speed (absence (A) or air circulation (C) and 1 m s⁻¹) to determine the ideal drying conditions. Mass reduction during drying was monitored using a semi-analytical balance with a resolution of 0.01 g (Shimadzu BL-3200H, Kyoto, Japan) until the samples reached a moisture equilibrium content of less than <0.2%. All the experiments were carried out in triplicate and the average moisture content was used to calculate the drying curves.

2.2. Study of Pfaffia Glomerata Stems, Leaves, and Inflorescences Drying

The moisture ratio (MR) of STs, LV, and IFs was determined by employing the following equation (Equation (1)) [18]:

$$\mathrm{M}\mathrm{R}={\displaystyle \frac{{\mathrm{M}}_{\mathrm{t}}-{\mathrm{M}}_{\mathrm{e}}}{{\mathrm{M}}_0-{\mathrm{M}}_{\mathrm{e}}}}$$

(1)

where, M₀, M_t, and M_e were the moisture content (kg kg⁻¹) at the initial stage, at time t, and the moisture content in equilibrium with the drying air, respectively.

To determine the better fit to explain the drying process, a modeling of drying curves process was performed using linear and nonlinear regression analyses using Statistica Software (version 7.1), between the models presented in Supplementary Table S1, according to Alkipinar [19] and Mujumdar [18]. The goodness of employed fitting models was tested by calculating statistical parameters, such as the coefficient of determination (R²) (Equation (2)), were

t—drying time (h);

k—drying constants (h s⁻¹);

a, b, c, and n—coefficients of the models.

$${\mathrm{R}}^2=1-{\displaystyle \frac{{\sum }_{\mathrm{i}=1}^{\mathrm{N}}{({\mathrm{M}\mathrm{R}}_{\mathrm{p}\mathrm{r}\mathrm{e},\mathrm{i}}-{\mathrm{M}\mathrm{R}}_{\mathrm{e}\mathrm{x}\mathrm{p},\mathrm{i}})}^2}{{\sum }_{\mathrm{i}=1}^{\mathrm{N}}{(\overline{{\mathrm{M}\mathrm{R}}_{\mathrm{p}\mathrm{r}\mathrm{e}}}-{\mathrm{M}\mathrm{R}}_{\mathrm{e}\mathrm{x}\mathrm{p},\mathrm{i}})}^2}}$$

(2)

The higher R² values were an indicator of the goodness of fit. The equations were fitted, and the data are presented in Supplementary Table S2. Also, the activation energy was determined by the Arrhenius model (Equation (3)) [19]. D₀ was the pre-exponential factor of the Arrhenius equation (m² s⁻¹), E_a was the activation energy (kJ mol⁻¹), R_g was the perfect gas constant (8.314 J mol⁻¹ K⁻¹), and T (K) was the temperature of the drying air.

$${\mathrm{D}}_{\mathrm{e}\mathrm{f}\mathrm{f}}={\mathrm{D}}_0 \exp \left(-{\displaystyle \frac{{\mathrm{E}}_{\mathrm{a}}}{{\mathrm{R}}_{\mathrm{g}}\mathrm{T}}}\right)$$

(3)

(a)

Color Measurement

In addition to the parameters related to drying kinetics, color parameters related to dried-product quality attributes [20] such as color variation were evaluated. The STs, LV, and IFs color difference (∆E*) was measured by comparing the samples before and after drying, using a colorimeter (D65 illuminant with CIELAB scale CR400, Konica Minolta, Japan). The colorimeter was calibrated by placing the tip of the measuring head flat against the surface of the white calibration plate. After standardization, L* (lightness), a* (redness/greenness), and b* (yellowness/blueness) values were measured on the surface of fresh and dried samples.

(b)

FTIR-ATR Analysis

The spectra analysis of the STs, LV, and IFs were carried out in the infrared spectrophotometer, model Cary 630 FTIR, (Agilent Technologies, Santa Clara, CA, United States) using an attenuated total reflectance accessory (ATR) Agilent Technologies, Santa Clara, CA, United States with a diamond ATR crystal. All the spectra were performed between 4000 to 400 cm⁻¹, scanning 64 scans and with 4 cm⁻¹ precision.

2.3. Ultrasound-Assisted Extraction

P. glomerata extracts of STs, LV, and IFs dried in the better condition were obtained through Central Composite Rotational Design (CCRD) (

Table 1. Planning and experimental response to obtain extracts with a higher content of total phenolic compounds (TPCs), antioxidant activity (AA), and saponins from stems, leaves, and inflorescences of Pfaffia glomerata.

	Experiment of Extraction		TPC (g GAE/100 g)			DPPH (µmol TE/g)			Saponins (g/100 g)
	Time (min)	Power (w)	ST	LV	IF	ST	LV	IF	ST	LV	IF
1	30	200	0.10 ± 0.03	0.31 ± 0.03	0.64 ± 0.03	2.81 ± 0.66	4.41 ± 0.03	2.29 ± 0.02	8.10 ± 0.02	3.21 ± 0.75	11.81 ± 1.64
2	30	500	0.09 ± 0.00	0.29 ± 0.03	1.26 ± 0.03	5.31 ± 0.09	1.19 ± 0.01	17.56 ± 0.13	11.76 ± 0.77	6.06 ± 0.10	22.07 ± 1.06
3	120	200	0.10 ± 0.03	0.28 ± 0.02	0.84 ± 0.01	6.10 ± 1.02	7.26 ± 0.04	20.94 ± 0.10	15.61 ± 1.35	6.95 ± 1.63	22.55 ± 1.41
4	120	500	0.13 ± 0.03	0.17 ± 0.01	0.82 ± 0.01	2.44 ± 0.34	10.55 ± 0.02	19.37 ± 0.09	13.83 ± 0.81	2.31 ± 0.26	19.04 ± 2.13
5	11	350	0.06 ± 0.02	0.35 ± 0.04	0.80 ± 0.01	1.94 ± 0.10	6.37 ± 0.01	21.05 ± 0.01	2.87 ± 0.22	2.12 ± 0.15	7.71 ± 0.72
6	139	350	0.08 ± 0.02	0.39 ± 0.03	0.72 ± 0.01	2.27 ± 0.32	14.18 ± 0.04	20.43 ± 0.02	11.58 ± 0.79	2.58 ± 0.33	20.24 ± 7.06
7	75	138	0.07 ± 0.03	0.42 ± 0.03	0.72 ± 0.01	1.57 ± 0.70	17.36 ± 0.02	22.37 ± 0.01	11.02 ± 0.71	1.65 ± 0.47	23.01 ± 1.28
8	75	562	0.09 ± 0.03	0.41 ± 0.02	0.72 ± 0.02	0.82 ± 0.51	27.62 ± 0.14	22.71 ± 0.01	14.10 ± 0.66	0.82 ± 0.74	16.65 ± 2.62
9 (CP)	75	350	0.11 ± 0.04	0.38 ± 0.03	0.77 ± 0.01	2.00 ± 0.23	18.01 ± 0.09	19.16 ± 0.02	11.76 ± 2.44	2.36 ± 0.31	16.67 ± 0.93

Results as mean ± standard deviation (or n = 6 (CP)).

), in which the variables of time (11 to 139 min) and power (138 to 562 W) were evaluated, with the addition of three central points to assess experimental error. The extractions were performed in duplicate. The equipment used for the extractions was the Ultrasonic Cell Disruptor (BIOBASE^®, Wolfenbuettel, Germany), with 1000 W of power and 60 Hz, operated in pulse mode (on/off, 3/6 s) and using a 6 cm × 12 mm probe. The samples (0, 6 g) were stored in 50 mL Falcon tubes, to which 10 mL of 80% ethanol was added as the solvent. During the sonication the maximum temperature was 25 °C. The extracts were separated after centrifugation (SOLAB^®, Paris, France) at 9056× g for 5 min and stored frozen at −15 °C until the time of analysis for the determination of saponins, antioxidant activity (AA) by DPPH, and total phenolic compounds (TPCs).

2.4. Extracts Characterization

2.4.1. Determination of Total Phenolic Compounds (TPCs) and Antioxidant Activity (AA)

The total phenolic compound (TPC) content was determined according to the methods of Singleton, Orthofer, and Lamuela-Raventos [20] and the results are expressed as µg of gallic acid equivalents (GAE) per mL. Antioxidant activity (AA) was evaluated by measuring the DPPH free radical scavenge method according to Brand-Wiliams et al. [21] and the results were expressed in μmol of Trolox per g of sample.

2.4.2. Total Saponin Content

In a tube, 1 mL of the extracts was mixed with 1 mL of 0.2% cobalt chloride and 1 mL of concentrated sulfuric acid. The mixture was vortexed (KASVI^®, Chennai, India) and, after 10 min, the reading was performed at 284 nm. The analytical curve (R² > 0.99) was prepared with saponin standard with concentrations of 0.08 to 0.28 mg/mL and the results were expressed as total saponins (Merck, Darmstadt, Germany) [22].

2.5. Characterization of the Extract Obtained in the Optimization of Extracting Parameters

2.5.1. Analysis of Phenolic, Organic Acids, and β-Ecdysone Composition

The sample was diluted in methanol (LC/MS grade, Merck, Darmstadt, Germany), filtered (PVDF filter, 25 mm and 0.45 μm), and analyzed by high-performance liquid chromatography (HPLC).

For the detection of phenolic compounds, HPLC (Shimadzu, model NEXERA X2, Kyoto, Japan) was used, coupled to the mass detector–MS/MS (Shimadzu, model 8050, Kyoto, Japan), using the Shimadzu C18 column (5 μm 150 × 4.6 mm). For elution, a linear gradient was used, composed of the mobile phases Milli Q water (A) and methanol (B), as follows: 1–9 min (20% B), 10–15 min (40% B) and 16–30 min (10% B). The temperature was set at 35 °C, injection in 1 μL sample. The MS/MS detector was operated in scan mode for 15 and it was monitored in positive and negative modes for each precursor ion and their respective transitions, which were identified by multiple reaction monitoring (MRM) mode. The quantification of compounds was performed based on the analytical curves (10–250 µg/L) of the following compounds: catechol, morin, isovanillin, gallic acid, quercetin, hydroxybenzaldehyde, naringenin, syringaldehyde, chlorogenic acid, syringic acid, protocatechuic acid, vanillic acid, salicylic acid, vanillin, ferulic acid, p-hydroxybenzoic acid, naringin, p-coumaric ac-id, caffeic acid, coniferyl aldehyde, sinapic acid, syringaldazine, catechin, sinaaldehyde, luteolin, rutin, theobromine, epicatechin, baicalin, chrysin, quinic acid, malic acid, kaempferol, coumarin, caffeine, resorcylic acid, nicotinic acid, and fumaric acid [23].

The analysis of β-ecdysone content was performed by the method described by Martins et al. [24], using HPLC (LC-10, Shimadzu, Kyoto, Japan) and ODS column (C18, 5 µm, 250.0 × 4.6 mm, Phenomenex, Torrance, CA, United States), at a temperature of 30 °C. The chromatographic run was performed in gradient mode using methanol and ultrapure water, with a flow rate of 1 mL/min, and the compound of interest was read at 245 nm (SPD-10A). The analytical curve was obtained using β-ecdysone (20-hydroxyecdysone, Sigma–Aldrich, Saint Louis, MO, USA) with concentrations of 30 to 250 mg/L.

2.5.2. GC-MS Analysis

The samples were analyzed by Gas Chromatography coupled to Mass Spectrometry. A Shimadzu QP2010 Ultra was used as SH-5MS capillary column (30 m × 0.25 mm and 0.25 µm), with helium gas for the drag. Two GC parameter types were used, for essential oil detection (1) and fatty acid detection (2), and the injection volume was 1 μL.

For essential oil detection, the temperature used was 230 °C in the inlet and detector. The initial temperature of the column was 50 °C (for one minute), being programmed to increase by 3 °C every minute, until reaching 185 °C (for one minute), to increase by 9 °C every minute, until reaching 275 °C (for two minutes), and to increase by 9 °C every minute, until reaching 300 °C (for one minute). For fatty acid detection, the temperature used was 230 °C in the inlet and detector. The initial temperature of the column was 40 °C (for two minutes), being programmed to increase by 3 °C every minute, until reaching 250 °C (for ten minutes).

In the mass detector, the temperature of the ionization chamber was 230 °C and the MS detection system was used in scan mode, operating in the mass/charge ratio (m/z) range of 35–550, with a solvent delay of 3 min. The compounds were identified by comparing the mass spectra with the NIST 14/14 s (Gaithersburg, MD, United States), Nistwebbook [25], and Adams [26] libraries.

2.5.3. Antibacterial Activity

(a)

Microorganisms and Inoculum Preparation

The antibacterial activity of the essential oil was tested against six bacterial strains: Listeria monocytogenes (ATCC 7644), Staphylococcus aureus (ATCC 25923), Staphylococcus epidermidis (ATCC 12228), Escherichia coli (ATCC 43893), Salmonella enterica subsp. enterica Typhi (ATCC 19214), and Pseudomonas aeruginosa (ATCC 27853). Bacterial cell mass dilution from 24 h culture was prepared for the assays.

The final bacterial cell concentration was adjusted according to the 0.5 McFarland Scale (1.5 × 10⁸ CFU mL⁻¹) in a spectrophotometer at 625 nm using 0.85% (m v⁻¹) sterile saline. Then, the suspension was diluted to 1:10 in Muller Hinton Broth culture medium (1.5 × 10⁷ CFU mL⁻¹), which was used as inoculum to determine the minimum inhibitory concentration (MIC).

(b)

Antibacterial Activity by Broth Microdilution Method

The antibacterial activity of the extracts was determined by serial microdilution in 96-well microplates according to the broth microdilution method [27], as modified for natural products. The tested concentrations ranged from 300 to 0.29 mg mL⁻¹. Sodium nitrite (100 to 1.25 mg mL⁻¹) was used as a positive control. All samples were evaluated in 100 μL of solution (culture medium and samples). After serial dilution, 50 µL of the inoculum prepared in saline solution, as described in the previous item, was added to each well (1.5 × 10⁵ CFU mL⁻¹) and subjected to incubation at 35 °C for 24 h. Reading was performed with the addition of 20 µL of 1.0% 2,3,5-triphenyltetrazolium chloride (Êxodo Científica^®, Sumaré, Brazil) developer in each well followed by the incubation of microplates at 35 °C for 20 min. The MIC was defined as the lowest concentration that resulted in visual growth inhibition, according to the developer. The minimum bactericidal concentration (MBC) was determined by subculturing 10 µL from each well on Muller Hinton agar plates and incubating at 35 °C for 24 h.

3. Results

3.1. Drying of Ginseng Stems, Leaves, and Inflorescence

The initial average moisture of the ginseng stems (STs), leaves (LV), and inflorescence (IF) were 69.3 ± 0.05 g 100 g⁻¹, 61.2 ± 0.04 g 100 g⁻¹, and 69.6 ± 0.02 g 100 g⁻¹, respectively.

At the end of drying at 40 °C with air circulation (40C), the STs, LV and IF showed an equilibrium humidity (UR) of 0.98 ± 0.08, 1.00 ± 0.12, and 1.05 ± 0.07 g 100 g⁻¹, respectively. On the other hand, the STs, LV and IF presented a UR of 1.06 ± 0.08, 1.04 ± 0.06, and 1.01 ± 0.07 g 100 g⁻¹, respectively, in the drying without air circulation at 40 °C (40A) (Supplementary Figure S1).

Increasing the drying temperature to 60 °C in drying without air circulation (60A) provided a UR of 0.88 ± 0.05, 0.79 ± 0.03, 1.04 ± 0.04 g 100 g⁻¹ (STs, LV, and FL, respectively), and at 60 °C with air circulation (60C) assumed values of 0.80 ± 0.02, 0.81 ± 0.02, and 0.95 ± 0.04 g 100 g⁻¹ for STs, LV, and IFs, respectively.

The STs, LV, and IFs presented at 80 °C without air circulation (80A) a UR of 0.85 ± 0.03, 0.83 ± 0.02, and 1.01 ± 0.03 g 100 g⁻¹, respectively. With air circulation at 80 °C, the URs were 0.83 ± 0.01, 0.85 ± 0.03, and 0.98 ± 0.03 g 100 g⁻¹, for STs, LV, and FL, respectively.

The time required to obtain UR was 360, 330, and 240 min for STs drying without air circulation at 40, 60, and 80 °C, respectively (Supplementary Figure S1). In ST drying with air circulation, it 330, 240, and 210 min were observed, respectively. The time reduction reaches about 8, 27%, and 12%, comparing the drying at 40, 60, and 80 °C, respectively. In the LV, it was observed that the use of circulating air reduced the drying time by 17, 27, and 30%, when drying was carried out at 40, 60, and 80 °C, respectively (Figure 1B). The reduction in RH time reached 8, 18, and 27% when the IFs was dried with circulating air at 40, 60, and 80 °C, respectively (Supplementary Figure S1).

These results point out that drying conducted at higher temperatures and with air circulating was faster and more effective than without air circulating. These results were expected, since at higher temperatures there is a higher rate of drying [28] and, consequently, leaving the water from STs, LV, and IFs of Brazilian ginseng. This mechanism also was observed in American ginseng drying [29] and in apples [30].

The initial hour presented, in general, a period of constant rate for Brazilian ginseng stems, leaves, and inflorescences (Figure 1A–C). According to Mujumdar [18], this behavior demonstrates that the main mechanism for removing water from the interior of the material is that of diffusion through the air–water interface. For the stems, as expected, the water output was more efficient in 60C and 80C experiments and was slower in 40A drying (Figure 1A). The leaves showed a higher water output in the processes in the first minutes of dehydration at 80C, 60C, and 80A conditions and, conversely, a lower drying rate in the 40A and 40C processes (Figure 1B). The other conditions tested showed a similar dehydration rate and water output behavior. According to Vigo et al. [31] and Louback et al. [32], the leaves and stems of P. glomerata are composed of tissues rich in cellulose, hemicellulose, and pectic substances. Higher lignification and high methoxylation and hydrophobic pectin were detected in P. glomerata stem tissues [32], which may explain the longer time required for stem dehydration in addition to the lower rate of water outflow in relation to the leaves and inflorescences.

In order to determine the moisture content as a function of drying, the MR data obtained at different drying conditions RH were fitted to five empirical models listed in Supplementary Table S1. The values of R² for different drying conditions determined by nonlinear regression analysis were presented in Table S2 and Table 1.

The coefficients of determination of most of the models were higher than 0.98 for all drying temperatures, which indicates a satisfactory fit of the mathematical models to the experimental data [33]. According to Draper & Smith [34], the capacity of a model to describe a certain physical process is inversely proportional to the standard deviation of the estimate. Based on this, the higher values of the R² revealed a better fit.

Among the models considered, the Page and modified Page models presented the highest R² values for most experiments, ranging from 0.9953–0.9996, although Page’s model did not fit the 40S experiment. Overall, the modified Page model generally obtained the highest R², notably under air circulation conditions, which are considered faster and more energy-efficient. As exceptions, the model that best fit was that of Henderson and Pabis for the ST-60A, LV-60C, IF-40A, and IF-60A experiments, while Lewis fitted the LV-60A.

Therefore, the modified Page model was selected to describe the drying behavior of the Brazilian ginseng STs, LV, and IFs during convective hot air drying within the experimental range. Similar findings were also reported by Ferreira et al. [35] in fermented grape pomace, by Akpinar [19] in basil and mint leaves, and by Oyefeso and Raji [36] in fresh tannia (Xanthosoma sagittifolium) corms. According to Onwude et al. [33], the models based on the New Law of cooling and Fick’s second law of diffusion, such as modified Page, Page, and others, have commonly been used to describe the drying behavior of various fruits and vegetables.

3.1.1. Activation Energy (Ae)

The drying Ae of STs, LV, and IFs of P. glomerata was 15.22, 16.67, and 15.33 kJ mol⁻¹, determined by the Arrhenius equation as shown in Equation (3). This result is in agreement with Santos et al. [37], which indicates that the higher the initial moisture content of the sample, the lower the Ae. The LV samples showed a higher initial moisture in relation to STs and IFs.

According to Akpinar [19] and Mujumdar [18], the Ae was an indication of the required energy to remove moisture from the product during the drying process. Zogzas, Maroulis and Marinos-Kouris [38] indicated a range between 12.7–110 kJ mol⁻¹ and with Onwude et al. [33], which indicated a range of 14.42 to 43.26 kJ mol⁻¹ for various biological materials. Ojediran and Raji [39] relate Ae to the energy required to initiate the diffusion of water from the inside of the tissue to the outside.

The aerial parts of P. glomerata showed a closer Ae when compared to crushed mass of jambu (16.61–16.97 kJ mol⁻¹) [40] and carrot (16.00 kJ mol⁻¹) [41] and lower than American ginseng roots (51.14 kJ mol⁻¹) and slices (46.64 kJ mol⁻¹) [29,42].

3.1.2. Color Variation During Drying Process

According to Xiao et al. [43], the color affects the overall impression and it is considered an important indicator of ginseng quality. Dried roots of American ginseng with higher L* and lower a* values are considered higher-quality products [29]. But there is no parameter for the aerial parts of P. glomerata.

The drying in higher temperatures provided a darkening of the LV (

Table 2. Variation in color parameters of Pfaffia glomerata stems (STs), leaves (LV), and inflorescences (IFs) ¹.

Experiment		L*	a*	b*
Fresh	Stems	39.1 ± 1.7 cB	−8.7 ± 0.9 cC	24.8 ± 2.1 aA
40C		39.9 ± 0.9 cB	−7.0 ± 0.5 bcB	19.8 ± 0.3 cB
60C		41.3 ± 2.3 bcB	−8.7 ± 0.8 cB	22.6 ± 1.9 abcB
80C		43.2 ± 3.8 bcB	−4.9 ± 1.2 aB	24.4 ± 1.0 abB
40A		43.8 ± 2.4 abcA	−5.0 ± 0.2 aA	21.4 ± 1.8 bcB
60A		48.7 ± 4.4 aA	−6.0 ± 0.3 abA	22.7 ± 1.0 abcB
80A		46.0 ± 1.4 abA	−5.0 ± 1.6 aA	22.2 ± 1.7 abcB
Fresh	Leaves	37.4 ± 0.3 aA	−4.0 ± 0.5 bcA	17.2 ± 0.1 bcdA
40C		34.4 ± 0.2 abB	−4.2 ± 0.4 bcA	17.1 ± 0.2 cdA
60C		31.9 ± 0.8 bB	−3.9 ± 0.2 bcA	16.3 ± 0.7 dA
80C		31.7 ± 3.5 bB	−2.9 ± 0.9 abA	20.4 ± 1.9 aA
40A		33.7 ± 2.2 abC	−4.3 ± 1.7 bcA	21.2 ± 1.2 aB
60A		31.6 ± 3.0 bC	−3.9 ± 1.5 abcA	19.6 ± 1.6 abB
80A		30.8 ± 2.3 cC	−2.9 ± 1.1 abA	18.9 ± 1.5 abcB
Fresh	Inflorescence	50.5 ± 0.7 aB	−4.1 ± 1.0 bcA	21.4 ± 0.3 bA
40C		52.5 ± 0.2 aB	−3.7 ± 0.0 bcB	24.5 ± 0.1 abB
60C		50.8 ± 2.1 aB	−2.8 ± 0.3 abB	24.3 ± 0.5 abB
80C		48.4 ± 1.3 aB	−1.4 ± 0.9 aB	24.5 ± 0.7 abB
40A		51.2 ± 3.4 aA	−3.4 ± 0.3 abcB	23.1 ± 1.1 abB
60A		21.7 ± 2.9 bA	−3.2 ± 2.1 bcB	23.0 ± 3.8 abB
80A		30.4 ± 1.1 bA	−2.5 ± 0.6 abB	23.9 ± 0.6 abB

¹ Results expressed as average (n = 6) ± standard deviation. Different lower-case letters in the same column represent significant differences between experiments (p < 0.05). Different upper-case letters in the same column represent significant differences with circulating air drying (p < 0.05). L*: luminosity: black (L* = 0) and white (L* = 100); a*: green color (−) and red color (+); b*: blue color (−) and yellow. Drying conditions: temperatures (40, 60, and 80 °C)/hot air speed (absence of circulation (A) or 1 m s⁻¹ (circulation: C)).

), possibly due to the degradation of pigments or non-enzymatic Maillard browning [43]. Compared with fresh STs and IFs (Table 2), there was no change in brightness when drying was conducted under air circulation. During static drying, the outermost layers of the STs dried out, which possibly led to an increase in luminosity. The use of step-down drying or using temperature or humidity gradients has been proposed by some authors [29,44,45,46,47].

For the STs and LV conditions, there was practically no variation in L*, a*, and b* when drying was carried out at 40 °C under air circulation. As an exception, only for STs there was a reduction in the blue-yellow component under this condition. For the flowers, temperatures of 40, 60, and 80 °C under air circulation had practically no influence on the color parameters, with changes only in the green-red component at 80 °C.

In general, compared to fresh samples, drying changed the color of the LV, STs, and IFs of P. glomerata. Drying under air circulation promotes a lower change in color components (L, a*, and b*) of STs, LV, and IFs. Thus, among the conditions tested, the best color parameters (L*, a*, and b*) were obtained by applying drying under air circulation, at temperatures of 40 °C or 60 °C for LV and STs, and at 60 or 80 °C for the IFs.

3.1.3. ATR–FTIR Analysis

Figure 2 shows the ATR-FTIR spectra of the β-ecdysone analytical standard (Sigma-Aldrich, ≥93%), fresh and dried stems (STs), leaves (LV), and inflorescences (IFs) of Brazilian ginseng.

In general, similar spectra are visualized for fresh and dried samples (STs, LV, and IFs) and for β-ecdysone standard, indicating the presence of this substance in the samples, considered the chemical marker for Brazilian ginseng. However, when comparing the ATR-FTIR spectra (Figure 2) of the fresh samples with the drying experiments, differences in peak intensities are observed. The spectra (Figure 2A–C) can be divided into two intervals: between 3700 to 2400 cm⁻¹ and another between 1800 to 800 cm⁻¹ (fingerprint region).

The broadband observed between 3500 and 3200 cm⁻¹ is mainly attributed to water, but some authors [48,49] also reported that this band is related to the presence of carbohydrates, carboxylic acids, alcohols, phenolic compounds, and saponins. In STs, LV, and IFs (Figure 2A, Figure 2B and Figure 2C, respectively), a reduction in the intensity of this band is observed when P. glomerata is dried, which is due to the loss of water present in the samples. Kareru et al. [49] and He et al. [50] related this band to β-ecdysone, since there are OH bonds in its molecule (Figure 3) [49,50].

The bands observed in the region from 2900 to 2840 are related to CH₂ and CH₃ symmetric and asymmetric stretch vibrations and can be attributed to lipids, proteins, carbohydrates, and nucleic acids [48]. These bands are present in β-ecdysone standard, fresh LV, and IFs, dried STs (40C, 40A, 60C, and 60A), and IFs (40C, 60C, 80C, 60A. and 80A). The bands in 2917 and 2849 cm⁻¹ are absent in LV and IFs, suggesting that the temperature increase included the degradation or volatilization of the chemical products. However, this does not happen in STs, probably because the stem cell structure is different compared to other aerial parts.

In the region between 2500 and 1800 cm⁻¹, there is no relevant data for chemical characterization, only bands referring to the interference of CO₂ present in the air [51]. The fingerprint region (1800 to 800 cm⁻¹) has a large amount of chemical information, with specific bands of molecular structures present in the samples, often superimposed on each other [48]. According to Xu and Wang [52], the peak observed at ~1725 cm⁻¹ can be associated with the C=O stretch of esters presents in phospholipids, cellulose, pectin, and hemicellulose. In the β-ecdysone pattern this peak was present and it was dislocated, possibly due to the C=C bond, which is close to the C=O in the β-ecdysone molecule (Figure 3). The same band shift also was observed in most of treatments (Figure 2A–C).

The band observed at ~1635 cm⁻¹ corresponds to C=C stretch vibrations related with phenolic compound and lignin [48,51]. Wathoni et al. [53] related this peak with pectin and the presence of a CH methyl group. According Abiramasundari, Gowda and Sreepriya [54], this peak can be related to β-ecdysone. This peak, with a small shift, is observed in β-ecdysone and in fresh samples of STs, LV, and IFs, and their intensities decrease during the drying process. It can be inferred, therefore, that drying can decrease β-ecdysone in P. glomerata stems, leaves, and inflorescences.

In the STs, there was a smaller decrease in this peak in the treatment 40A and 60A and in the LV in the 40C, 60C, 40A, and 60A. In the IF, it is observed that in the treatments 40C, 60C, 80C, and in the 60A, this peak is present. According to Lang et al. [54], the drying temperatures can cause the degradation of thermolabile compounds, such as phenolic compounds. Therefore, there was possibly a degradation of part of these compounds when the STs, LV, and FLs were dried. Furthermore, the disappearance of this peak may in some samples be related to the resonance phenomenon [55].

The region between 1200 and 900 cm⁻¹ is predominant of C-O-C and OH stretch vibrations, characteristic of cellulose and hemicellulose [56], and symmetrical acyclic stretch, CH₂, OH, and CO bonds of polysaccharides such as amylose and amylopectin [48,56]. It also covers the C-O-H, C-C and C-H bonds, which Kacuráková et al. [57] associate this region with pectin and C-O, C-C and C-C-O bonds of cellulose, hemicellulose, and lignin, which are present in the plant’s plant wall, in addition to monosaccharides such as ribose, glucose, and galactose [55]. In Figure 2A–C a peak is observed at 1012 cm⁻¹ in the β-ecdysone pattern and in fresh STs, LV, and IFs. Despite this peak being present in all treatments, there was a decrease in STs dried at 80 °C (80A, 80C), in LV dried at 80 °C (80C) and 40 °C (40A), and also in IFs dried at 40 °C (40C and 40A) and 60 °C (60C).

Among the treatments evaluated, the dryings conducted at 60C for the leaves and stems, and at 80C for the flowers, provide similar profiles of chemical compounds when compared to these fresh samples and to the standard of β-ecdysone.

3.2. Definition of the Conditions for Maximum Extraction of Total Phenolic Compounds, Saponins, and Antioxidant Activity by DPPH from Stems, Leaves, and Inflorescences of P. glomerata

The results obtained for the different UAE extraction conditions of P. glomerata compounds are shown in Table 2. The TPC levels ranged from 0.06 to 0.13 (STs), 0.28 to 0.42 (LV), and 0.64 to 1.26 (IFs) g GAE/100 g. Saponin levels varied from 2.87 to 15.61 (STs), 0.82 to 6.95 (LV), and 7.71 to 23.01 (IFs) g/100 g, and AA by DPPH ranged from 0.82 to 5.31 (STs), 1.19 to 27.62 (LV), and 2.29 to 22.71 (IFs) µmol TE/g. In general, there is significant variability in the composition of the extracts, depending on the extraction parameter used. Silva et al. [58] found TPC values of 70.30 mg/g to 118.60 mg/g in P. glomerata obtained from leaves that were manually reduced by cold maceration in ethanol with subsequent rotary evaporation at 40 °C. Vardanega et al. [59] found a variation of 100 to 170 mg/g of saponins in P. glomerata extracts obtained by subcritical water, which is consistent with the values found in this research.

The Central Composite Design (CCD) was used to optimize the UAE extraction process to obtain process parameters to achieve maximum values for TPC, AA, and saponin levels between the proposed interval of parameters. The parameters of the desirability methodology were validated through an analysis of variance (ANOVA) for the second-degree polynomial model (Table 2).

The R² of the models obtained was 0.74, 0.50, and 0.64 for the saponin content (STs, LV, and IFs, respectively) (

Table 3. ANOVA results of regression models from Central Composite Design for UAE extraction optimization of TPC, AA, and saponin content of Pfaffia glomerata stem, leaf, and inflorescence.

Source		X1-TPC			X2-AA by DPPH			X3-Saponin
		Mean	f-Value	p-Value	Mean	f-Value	p-Value	Mean	f-Value	p-Value
Stems	Time (Linear)	0.0012	0.3419	0.5645	0.1900	0.9993	0.3279	121.1230 *	70.3410 *	0.0000 *
	Time (Quadratic)	0.0023	0.6412	0.4315	7.5076 *	39.4779 *	0.0000 *	27.0030 *	15.6817 *	0.0006 *
	Power (L)	0.0004	0.1088	0.7444	1.2227 *	6.4294 *	0.0185 *	9.3602 *	5.4358 *	0.0288 *
	Power (Q)	0.0012	0.3389	0.5661	0.8361 *	4.3966 *	0.0472 *	27.4434 *	15.9375 *	0.0006 *
	Time x Power (L)	0.0004	0.1191	0.7331	18.9817 *	99.8130 *	0.0000 *	16.1969 *	9.4062 *	0.0054 *
	Lack of Fit	0.0019	0.2915	0.8311	8.4664 *	44.5197 *	0.0000 *	9.7530 *	5.6640 *	0.0047 *
	R2	0.0610			0.4911			0.7449
	Adjusted R2	0.0000			0.3933			0.6959
Leaves	Time (Linear)	0.0027	0.0352	0.8527	0.0885	0.3488	0.5605	0.0944	0.3828	0.5422
	Time (Quadratic)	0.0230	0.3010	0.5885	8.2133 *	32.3575 *	0.0000 *	7.8782 *	31.9396 *	0.0000 *
	Power (L)	0.0052	0.0684	0.7960	2.0965 *	8.2594 *	0.0086 *	2.1627 *	8.7679 *	0.0070 *
	Power (Q)	0.0037	0.0479	0.8287	0.2777	1.0941	0.3064	0.3279	1.3295	0.2607
	Time x Power (L)	0.0043	0.0561	0.8149	28.0198 *	110.3880 *	0.0000 *	28.0718 *	113.8080 *	0.0000 *
	Lack of Fit	0.0261	0.3415	0.7955	10.8409 *	42.7094 *	0.0000 *	10.7519 *	43.5903 *	0.0000 *
	R2	0.0207			0.5022			0.5040
	Adjusted R2	0.0000			0.4066			0.4086
Inflorescence	Time (Linear)	0.0870 *	9.2819 *	0.0057 *	95.7762 *	5.9177 *	0.0232 *	161.5823 *	43.7273 *	0.0000 *
	Time (Quadratic)	0.0121	1.2885	0.2680	10.9385	0.6759	0.4195	11.7099	3.1689	0.0883
	Power (L)	0.0158	1.6828	0.2074	50.2997	3.1079	0.0912	1.2374	0.3349	0.5684
	Power (Q)	0.0004	0.0451	0.8337	0.0862	0.0053	0.9424	68.9273 *	18.6530 *	0.0002 *
	Time x Power (L)	0.0914 *	9.7511 *	0.0048 *	141.6522 *	8.7523 *	0.0070 *	94.8969 *	25.6809 *	0.0000 *
	Lack of Fit	0.1625 *	17.3265 *	0.0000 *	110.4854 *	6.8266 *	0.0019 *	34.0340 *	9.2103 *	0.0003 *
	R2	0.2273			0.2980			0.6439
	Adjusted R2	0.0786			0.1630			0.5755

* Significant (p < 0.05).

). For the DPPH content (R² from 0.29 to 0.50) and TPC (0.02 to 0.22), the models showed a low R² and were not considered adequate. A high R² is related to the proximity between experimental and predicted results. It was found that the models obtained would not provide good prediction, mainly for the TPC and AA parameters; however, they could have a better fit for the saponin content. Several studies [5,6,7,58] indicate that saponins are the main bioactive compound present in P. glomerata, especially β-ecdysone, considered a chemical marker of this plant. Thus, the operating range used for extraction favored a better explanation of the total saponin content, which included β-ecdysone, and this can be considered sufficient for an optimization study.

Upon conducting the ANOVA, terms associated with the second-degree polynomial parameters model, exhibiting small p-values and large F-values, suggested a more substantial impact on the corresponding response variables. The software (Statistica 7.1) was employed using the desirability option to determine the critical values for extractive parameters to obtain extracts with maximum levels of saponins, AA, and TPCs, as presented in

Table 4. Predicted and experimental values, desirability index after optimization at maximum value for UAE extraction of Pfaffia glomerata stem, leaf, and inflorescence ¹.

Pfaffia glomerata Fraction	Bioactive Substance	Extractive Parameter	Parameters to Maximum Value	Predicted Value at Solution	Value of Global Desirability	Experimental Parameters	Experimental Value at Solution	Percentual Difference 2
Stems	TPC	Time (min)	136.50	0.1137	0.62735	137	0.14 ± 0.02	23%
		Power (W)	137.87			138
	Saponin	Time (min)	136.50	12.4980		137	11.99 ± 0.21	4%
		Power (W)	137.87			138
	DPPH	Time (min)	136.50	1.9993		137	2.13 ± 0.01	6.5%
		Power (W)	137.87			138
Leaves	TPC	Time (min)	138.63	0.3323	0.62896	139	0.36 ± 0.01	8.3%
		Power (W)	562.13			562
	Saponin	Time (min)	138.63	2.3746		139	2.53 ± 0.02	6.5%
		Power (W)	562.13			562
	DPPH	Time (min)	138.63	2.3738		139	2.49 ± 0.02	4.9%
		Power (W)	562.13			562
inflorescence	TPC	Time (min)	138.63	0.8364	0.80142	139	1.02 ± 0.01	21%
		Power (W)	562.13			562
	Saponin	Time (min)	138.63	17.6095		139	15.53 ± 0.23	11.7%
		Power (W)	562.13			562
	DPPH	Time (min)	138.63	19.9907		139	22.64 ± 0.20	13.3%
		Power (W)	562.13			562

¹ According to Program Statistica 7.1. ² Percentual difference from predicted to experimental value.

.

The desirability index (DI) was good, since the DI values were >0.60. The predicted values for TPCs, AA, and saponins after extraction using the indicated parameters are presented in Table 4. For validation of the models, new extractions were performed, and the results are presented in Table 4. Comparing the predicted values with the experimental values, it is noted that there was a greater difference, in decreasing order, generally for saponin < AA < TPC. Response surface and desirability graphs are presented in the Supplementary Material (Supplementary Figure S2). Based on the experimental responses, the response surface plot indicates that the chosen regions for the experiments are suitable. Thus, it can be seen that extracts rich in saponins do not necessarily have high AA or TPCs. To better understand which substances are present in the obtained extracts, chromatographic characterization analyses were conducted.

3.3. Optimized UAE Extract Characterization

The extracts were characterized regarding β-ecdysone content and phenolic compounds profile (

Table 5. Bioactive compounds in UAE extract from inflorescence, leaves, and stems of P. glomerata ¹.

Bioactive Compounds	Stems	Leaves	Inflorescence
β-ecdysone (mg/g db)	3.50 ± 0.01 b	3.90 ± 0.02 a	3.23 ± 0.01 c
Phenolic compounds (mg/100 g db)	0
Morin	n.d.	n.d.	1.03 ± 0.07
Gallic acid	n.d.	n.d.	0.48 ± 0.05
Quercetin	n.d.	n.d.	2.17 ± 0.07
Ferulic acid	n.d.	n.d.	35.55 ± 1.57
p-hydroxybenzoic acid	5.38 ± 0.01 c	n.d.	18.05 ± 1.69
Isovanillin	2.13 ± 0.06 a	1.54 ± 0.02 b	0.08 ± 0.01 c
Hydroxybenzaldehyde	0.52 ± 0.08 b	2.41 ± 0.03 a	2.29 ± 0.01 a
Syringaldehyde	3.59 ± 0.17 a	1.66 ± 0.0 b	3.83 ± 0.17 a
Coniferyl aldehyde	2.11 ± 0.00 b	10.22 ± 1.70 a	0.27 ± 0.07 b
Sinapaldehyde	6.33 ± 0.67 a	5.07 ± 0.47 b	1.65 ± 0.08 c
Cafein	0.54 ± 0.00 c	6.82 ± 0.08 a	2.00 ± 0.13 b
Nicotinic acid	11.35 ± 0.06 c	29.46 ± 1.77 a	17.47 ± 0.013 b
Chlorogenic acid	0.28 ± 0.01 b	0.96 ± 0.01 a	0.29 ± 0.03 b
Syringic acid	0.09 ± 0.0 c	0.48 ± 0.00 b	3.36 ± 0.04 a
Protocatechuic acid	0.85 ± 0.01 c	3.33 ± 0.51 b	11.42 ± 0.22 a
Vanillic acid	0.13 ± 0.0 b	1.05 ± 0.12 b	5.70 ± 0.61 a
Vanillin	9.79 ± 0.43 b	18.86 ± 1.15 a	4.47 ± 0.05 c
p-coumaric acid	3.41 ± 0.04 c	16.67 ± 0.77 b	45.28 ± 0.95 a
Rutin	8.32 ± 0.28 b	1.29 ± 0.01 c	28.81 ± 1.77 a
Quinic acid	1.26 ± 0.08 b	1.93 ± 0.07 a	1.91 ± 0.04 a
Theobromine	0.32 ± 0.0 b	0.58 ± 0.03 a	0.26 ± 0.04 b

¹ Results expressed as average (n = 6) ± standard deviation. Different lower-case letters in the same line represent significant differences between extracts (p < 0.05). n.d.—not detected.

). Also, the volatile and fatty acid compounds were investigated (

Table 6. Chemical composition of extracts from stems, leaves, and inflorescences of Pfaffia glomerata by GC-MS.

P. glomerata Fraction	Meto-Dology 1	Elution Order	Retention Time (min)	Identified Compound [25]	Area (%)
stem	1	1	3.51	Tetrachloroethylene	76.65
	1	2	44.83	1,2-Benzenedicarboxylic acid, bis(2-methylpropyl) ester	3.73
	1	3	48.905	Hexadecanoic acid, ethyl ester	2.17
	1	4	51.405	Phytol	7.61
	1	5	52.23	Linoleic acid ethyl ester	3.25
	1	6	52.34	9,12,15-Octadecatrienoic acid, ethyl ester, (Z,Z,Z)-	5.18
	2	1	3.52	Tetrachloroethylene	93.97
	2	2	51.425	Phytol	6.03
Leaves	1	1	3.34	Butanoic acid, ethyl ester	0.15
	1	2	3.51	Tetrachloroethylene	5.28
	1	3	4.015	Butanoic acid, 3-methyl-	0.19
	1	4	43.745	Neophytadiene	0.72
	1	5	44.83	1,2-Benzenedicarboxylic acid, bis(2-methylpropyl) ester	0.38
	1	6	48.08	n-Hexadecanoic acid	2.90
	1	7	48.905	Hexadecanoic acid, ethyl ester	2.35
	1	8	51.42	Phytol	44.11
	1	9	51.81	l-Norvaline, N-(2-methoxyethoxycarbonyl)-, hexyl ester	1.06
	1	10	51.84	9,12-Octadecadienoic acid (Z,Z)-, methyl ester	0.70
	1	11	51.955	9,12,15-Octadecatrienoic acid, (Z,Z,Z)-	11.49
	1	12	52.115	Hexadecanal, 2-methyl-	1.55
	1	13	52.24	Linoleic acid ethyl ester	3.19
	1	14	52.355	9,12,15-Octadecatrienoic acid, ethyl ester, (Z,Z,Z)-	6.02
	1	15	53.11	3,7,11,15-Tetramethylhexadec-2-en-1-yl acetate	0.20
	1	16	54.00	Octanoic acid, 2-dimethylaminoethyl ester	0.30
	1	17	54.185	Glycidyl palmitate	0.27
	1	18	54.405	Octanamide, N-(2-hydroxyethyl)-	0.48
	1	19	54.50	14,15,16-Trinor-8-.xi.-labd-5-ene, 8,13-epoxy-, (-)-	0.45
	1	20	54.945	4,8,12,16-Tetramethylheptadecan-4-olide	0.10
	1	21	55.39	2H-Pyran, 2-(7-heptadecynyloxy)tetrahydro-	0.14
	1	22	55.79	Phenol, 2,2′-methylenebis [6-(1,1-dimethylethyl)-4-methyl-	0.12
	1	23	56.095	3-Cyclopentylpropionic acid, 2-dimethylaminoethyl ester	0.11
	1	24	56.195	3-Cyclopentylpropionic acid, 2-dimethylaminoethyl ester	0.50
	1	25	56.305	Methyl 7,11,14-eicosatrienoate	0.07
	1	26	56.36	1,3,5-Trisilacyclohexane	0.23
	1	27	56.555	Fumaric acid, dodecyl tetradec-3-enyl ester	0.36
	1	28	56.735	Hexadecanoic acid, 2-hydroxy-1-(hydroxymethyl)ethyl ester	4.67
	1	29	56.94	Docosanal	0.40
	1	30	57.18	Bis(2-ethylhexyl) phthalate	0.27
	1	31	57.395	Tryptanthrine	0.53
	1	32	58.86	9,12-Octadecadienoic acid (Z,Z)-, 2,3-dihydroxypropyl ester	3.96
	1	33	58.98	9,12,15-Octadecatrienoic acid, ethyl ester, (Z,Z,Z)-	5.70
	1	34	60.71	Squalene	1.01
	2	1	3.52	Tetrachloroethylene	11.11
	2	2	51.435	Phytol	76.00
	2	3	52.25	Linoleic acid ethyl ester	3.29
	2	4	52.365	9,12,15-Octadecatrienoic acid, ethyl ester, (Z,Z,Z)-	9.60
inflorescence	1	1	3.5	Tetrachloroethylene	71.49
	1	2	50.84	9-Methyl-10-methylenetricyclo [4.2.1.1(2,5)]decan-9-ol	0.90
	1	3	51.41	Phytol	7.19
	1	4
	2	1	3.505	Tetrachloroethylene	93.29
	2	2	55.795	Phenol, 2,2′-methylenebis [6-(1,1-dimethylethyl)-4-methyl-	6.71

¹ Method of investigation: essential oil detection (1); fatty acid detection (2), according to item 2.5.2 GC-MS analysis.

). There is no similar characterization of aerial Pfaffia glomerata extracts in the literature.

The saponin β-ecdysone was present in the UAE extracts in a decreasing concentration: leaves > stems > inflorescence (Table 4), demonstrating that, among the aerial fractions, the leaves are the most promising for the extraction of saponins by UAE. Other, more toxic solvents may yield higher amounts of β-ecdysone, as presented by Franco et al. [60] (45 mg/g). These authors employed maceration extraction using methanol and fractionation with quantification of the dichloromethane fraction. These techniques promote good extraction but deviate from the principles of green chemistry, advocated in the present study.

Regarding the phenolic compounds investigated, there was, in general, a decreasing concentration in the flowers > leaves > stems. This ordering of the concentration is similar to that determined in the TPC methodology by Folin-Ciocalteu (Table 3).

In the extracts of aerial parts of P. glomerata, 21 phenolic compounds were detected. Morin, quercetin, gallic acid, and ferulic acid were detected only in inflorescence, while p-hydroxybenzoic acid was detected in the inflorescence and stems, but not in the leaves. The inflorescence extract had the highest content of syringic acid, protocatechuic acid, vanillic acid, and p-coumaric acid rutin. Cafein, coniferyl aldehyde, theobromine, vanillin, and nicotinic, chlorogenic, and quinic acids were predominant in leaves, while sinapaldehyde was found in a higher amount in stems. Chung et al. [61] also observed that the profile of phenolics varied in the different parts of Korean ginseng (Panax ginseng Meyer) evaluated, with the contents ranked as follows: fruit > leaves > roots. According to these authors, in the leaves, chlorogenic, genistic, and m- and p-coumaric acids were the main phenolic acids, and rutin the most abundant flavonoids, with the contents of p-coumaric acid (6.3–10.9 mg/100 g) and rutin (1.9–3.8 mg/100 g) being similar to those found in this study. Girotto et al. [62] found in leaves of P. glomerata gallic, trans-cinamic, ferulic, p-coumaric, and cafeic acids, whose contents are higher than those obtained for flavonoids catechin and quercetin, corroborating to the results found in the present work.

Rutin was found in a higher proportion among flavonoids, and the most abundant phenolic acids were ferulic, nicotinic acid, and p-coumaric. Among these compounds, ferulic and p-coumaric had a strong correlation with antioxidant activity [62], while rutin shows anti-inflammatory and anticancer properties [63].

In the GC-MS analysis, numerous compounds were identified in the extracts. A greater diversity of compounds was observed in the leaves. Phytol, a compound found abundantly in nature and derived from chlorophyll, particularly in plant extracts or essential oils, was present in all extracts. This compound is often associated with numerous biological activities, including anxiolytic, metabolism-modulating, cytotoxic, antioxidant, autophagy- and apoptosis-inducing, antinociceptive, anti-inflammatory, immune-modulating, and antimicrobial effects [64]. Phytol is a precursor of vitamin E and is used in the manufacture of synthetic forms of vitamin E and vitamin K1 [65]. Other compounds with reported biological activity were also detected, such as linoleic acid ethyl ester, an esterified form of linoleic acid, which prevents arteriosclerosis and is active against S. mutans, A. actinomycetemcomitans, P. gingivalis, S. gordonii, and S. sanguis and the fungus C. albicans (25 µg/mL) [66]. The n-Hexadecanoic acid and their ester had anti-inflammatory properties [67].

Due to the presence of phenolics, beta-ecdysone, phytol, and other bioactive compounds, the potential antimicrobial activity of UAE stem, leaf, and inflorescence extracts was investigated. The results are presented in

Table 7. Minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) of UAE extracts from stems, leaves, and inflorescence of Pfaffia glomerata.

Bacteria	Leaves (mg mL−1)	Stems (mg mL−1)	Inflorescence (mg mL−1)	Sodium Nitrite (mg mL−1)
	MIC	MIC	MIC	MIC
	MBC	MBC	MBC	MBC
Staphylococcus aureus	75 ± 0.00	150 ± 0.00	150 ± 0.00	12.5 ± 0.00
	>300 ± 0.00	150 ± 0.00	>300 ± 0.00	100 ± 0.00
Escherichia coli	75 ± 0.00	300 ± 0.00	150 ± 0.00	12.5 ± 0.00
	75 ± 0.00	>300 ± 0.00	>300 ± 0.00	25 ± 0.00
Listeria monocytogenes	150 ± 0.00	150 ± 0.00	150 ± 0.00	12.5 ± 0.00
	150 ± 0.00	>300 ± 0.00	>300 ± 0.00	>100 ± 0.00
Salmonella enterica Typhi	37 ± 0.00	150 ± 0.00	150 ± 0.00	12.5 ± 0.00
	37 ± 0.00	>300 ± 0.00	>300 ± 0.00	100 ± 0.00
Pseudomonas aeruginosa	75 ± 0.00	150 ± 0.00	150 ± 0.00	12.5 ± 0.00
	>300 ±0.00	>300 ± 0.00	>300 ± 0.00	50 ± 0.00
Staphylococcus epidermidis	>300 ± 0.00	300 ± 0.00	>300 ± 0.00	12.5 ± 0.00
	>300 ± 0.00	300 ± 0.00	>300 ± 0.00	>100 ± 0.00

Values are the mean ± standard deviation (n = 3).

.

The minimum inhibitory concentration (MIC) values for the UAE extracts ranged from 37 to >300 mg mL⁻¹ for the leaves, 150 to 300 mg mL⁻¹ for the stems, and 150 to >300 mg mL⁻¹ for the inflorescences. For the positive control sodium nitrite, the result was 12.50 mg mL⁻¹ for all samples evaluated.

The leaves exhibited the lowest MIC against Salmonella Typhi (37 mg mL⁻¹) and 75 mg mL⁻¹ against Staphylococcus aureus and Escherichia coli. The stems showed the lowest MIC against S. aureus, Listeria monocytogenes, S. Typhi, and Pseudomonas aeruginosa. Similarly, the inflorescences had the lowest MIC (150 mg mL⁻¹) against S. aureus, E. coli, L. monocytogenes, S. Typhi, and P. aeruginosa.

The average minimum bactericidal concentration (MBC) values ranged from 37 to >300 mg mL⁻¹ for the leaves and 150 to >300 mg mL⁻¹ for the stems. The inflorescences exhibited MBC values of >300 mg mL⁻¹. For the control, sodium nitrite’s MBC ranged to 25 (E. coli) to >100 mg mL⁻¹ (S. epidermidis and L. monocytogenes).

The lowest MBC values in the leaf extracts were against S. Typhi and E. coli (37 and 75 mg mL⁻¹, respectively), indicating significant bactericidal potential. The stem extracts presented a lower MBC against S. aureus (150 mg mL⁻¹) and S. Typhi (37 mg mL⁻¹). Stem MBC against S. Typhi had greater bactericidal potential than the sodium nitrite control.

Compared with the literature, these results represent an advancement in the study of the antimicrobial action of the aerial fractions of Pfaffia glomerata. Vigo et al. [23] concluded that hydroalcoholic extracts of the roots did not exhibit antimicrobial activity against E. coli ATCC 25922, P. aeruginosa ATCC 15442, Bacillus subtilis ATCC 6623, and S. aureus ATCC 25923 at concentrations up to >1000 μg mL⁻¹.

Generally, Gram-positive bacteria are more susceptible to essential oils and plant extracts than Gram-negative bacteria. This is due to the complexity of the Gram-negative cell wall, including the presence of lipopolysaccharides, which hinder substance penetration [68]. The UAE P. glomerata aerial extracts demonstrated activity against both Gram-negative and Gram-positive bacteria, even showing greater activity against the former group. This effect may be partially attributed to the presence of β-ecdysone, a saponin with detergent properties [69] that can enhance the solubilization of Gram-negative bacterial cell wall components. The different chemical components present in the UAE extracts explain the varying effects on bacterial growth inhibition. Phenolic acids compromise the stability of the bacterial cytoplasmic membrane, altering its permeability and inhibiting extracellular microbial enzyme activity. Similarly, undissociated organic acids diffuse into bacterial cells, leading to intracellular acidification [70,71]. The antibacterial activity can also be attributed to the potential synergistic effect between the extract components.

4. Conclusions

This research investigates the extraction of bioactive compounds from the leaves, inflorescences, and stems of Pfaffia glomerata.

The optimal drying conditions for these plant tissues were determined using a circulating air oven. The results demonstrated that higher drying temperatures combined with circulating air resulted in a faster drying rate. Drying with air circulation minimized alterations in color attributes and chemical profile, as evidenced by the FTIR-ATR analysis. Based on these findings, temperatures of 60 °C for leaves and stems, and 80 °C for the inflorescence are proposed as the most appropriate.

In addition, this study presents an optimal extraction potency and time for obtaining UAE extracts with high concentrations of saponins and phenolic compounds, exhibiting maximum antioxidant activity. This was determined through a Central Composite Design and subsequent desirability index optimization. Ultrasound-assisted extraction (UAE) operated at 137.8 W for 136.5 min (stems) and 562.32 W for 138.6 min (leaves and inflorescences), yielding extracts rich in β-ecdysone, phenolics, and other bioactive compounds, like phytol. Notably, the leaf extract demonstrated inhibitory effects against S. aureus, E. coli, L. monocytogenes, S. Typhi, and P. aeruginosa, and presented bactericidal effects against E. coli, L. monocytogenes, and S. Typhi. Extracts from other aerial parts (inflorescences and stems) showed lower bioactive potential compared to leaves. This work represents the first report on the application of ecofriendly UAE techniques to extract bioactive compounds from the aerial parts of Brazilian ginseng.

This research offers novel insights into the efficient and ecofriendly ultrasound-assisted extraction of bioactive compounds from the aerial parts of Pfaffia glomerata. The results highlight the potential for the full utilization of this underutilized plant.