Influence of Solvent Properties on Phenolic Recovery and Bioactivity of Eugenia protenta Leaf Extracts

Abstract

The influence of extraction solvents on the recovery of phenolic constituents from Eugenia protenta McVaugh leaves was investigated using solvent selection criteria based on the Snyder triangle and the linear solvation energy relationships (LSER) solvatochromic parameters (hydrogen bond donor acidity (α), hydrogen-bond acceptor basicity (β) and dipolarity/polarizability (π*)). Extracts obtained with methanol, ethanol, isopropanol, acetone, and ethyl acetate were evaluated in terms of extraction yield, gallic acid content, and antioxidant and antiglycation activities. Gallic acid was identified by nuclear magnetic resonance (¹H NMR) and quantified by quantitative nuclear magnetic resonance (qNMR) using the pulse length-based concentration determination (PULCON) method as a marker of the phenolic fraction. Methanol provided the highest extraction yield (30.6 ± 4.0 mg g^-1) and gallic acid content (36.8 ± 0.2 mg g^-1 dry extract), while alcoholic extracts showed the best overall antioxidant and antiglycation performance. The results indicate that extraction efficiency depends not only on solvent polarity, but also on specific solute-solvent interactions related to hydrogen bonding and dipolarity/polarizability, supporting the use of protic solvents to recover phenolic rich bioactive extracts.

Keywords:
Eugenia protenta; gallic acid; qNMR; biological activities; solvation parameters

Introduction

The genus Eugenia (Myrtaceae) comprises more than 1,000 species and is widely recognized as an important source of phenolic compounds with pharmacological relevance.¹,² Phenolic compounds with antioxidant and antiglycation properties have been consistently associated with the bioactive potential of different species within the genus.³,⁴ These metabolites play a central role in the modulation of oxidative stress and in the inhibition of advanced glycation end-product (AGE) formation, processes directly related to the development of metabolic complications, including type 2 diabetes mellitus and cellular aging.⁵,⁶

This association has been observed in several species of the genus, such as Eugenia pyriformis, whose extracts exhibited marked antioxidant and antiglycation activities attributed to the presence of major phenolic compounds.⁷ Similarly, Eugenia uniflora has been described as a species rich in flavonoids and tannins, compounds related to its antioxidant potential.⁸ In Eugenia dysenterica, the contribution of phenolic compounds to antioxidant and antiglycation activities has also been highlighted.⁹ In addition, studies on Eugenia punicifolia demonstrated that the biological response of the extracts is directly associated with the phenolic compounds present and may vary according to the extraction system employed.³,⁴ Taken together, these data indicate that both chemical composition and methodological choice are determinant factors in the expression of the bioactive potential of species of this genus.

Eugenia protenta McVaugh remains poorly investigated, with scarce reports on its chemical composition and biological activity, which contrasts with the phytochemical diversity already described for other species of the genus and reinforces the need to investigate its metabolites potentially associated with relevant biological effects.¹⁰ Among the phenolic compounds frequently related to the bioactivity of plant extracts, gallic acid (3,4,5-trihydroxybenzoic acid) stands out as a metabolite widely distributed in plant species. Its polyphenolic structure confers high antioxidant capacity, especially through the neutralization of reactive oxygen species (ROS), and recent evidence indicates that this compound may also contribute to the inhibition of AGE formation, either by scavenging reactive intermediates or by modulating oxidative stress.³,¹¹,¹² Considering the relevance of these processes in metabolic disorders, the quantification of gallic acid represents an effective strategy for correlating chemical composition and biological activity in plant extracts.

However, the obtainment of chemically representative extracts depends decisively on the choice of the extraction system, since different solvents modulate the selectivity for the extraction of specific classes of metabolites.¹³ To understand and systematize this influence, the Snyder triangle constitutes a useful tool, as it classifies solvents according to their polarity characteristics, thereby favoring comparability among different extraction systems.¹⁴ This approach is complemented by the linear solvation energy relationships (LSER) model, which, through the solvatochromic parameters hydrogen-bond donor acidity (α), hydrogen-bond acceptor basicity (β) and dipolarity/polarizability (π*), allows the independent quantification of the contributions of dipolarity/polarizability and hydrogen bonding to solute-solvent interactions, providing a more detailed analysis of solubilization.¹⁵,¹⁶ Given the compositional complexity of the obtained extracts, nuclear magnetic resonance (NMR) stands out as a selective and robust spectroscopic technique for the dereplication and quantification of organic compounds in complex matrices, making it possible to assess more accurately the contribution of specific metabolites, such as gallic acid, to antioxidant and antiglycation activities.¹⁷

Accordingly, the present study aimed to investigate the influence of different extraction systems on the gallic acid content of E. protenta leaf extracts, as well as to evaluate their antioxidant and antiglycation activities, integrating solvent characterization tools and quantitative NMR analysis to achieve a broader understanding of the relationship between chemical composition and bioactivity.

Experimental

Materials

The solvents used for plant material extraction were high performance liquid chromatography (HPLC)-grade methanol, absolute ethanol, isopropanol, ethyl acetate, and acetone, purchased from Sigma-Aldrich (St. Louis, MO, USA), QHemis (São Paulo, SP, Brazil), and Tedia High Purity Solvents (Fairfield, OH, USA), respectively. Deuterated methanol (MeOD-d₄, 99.9%) and deuterated chloroform (CDCl₃), used for NMR analyses, were obtained from Cambridge Isotope Laboratories Inc. (Andover, MA, USA). Dimethyl terephthalate (DMT), a certified reference material, was supplied by the Division of Chemical and Thermal Metrology of Inmetro (Rio de Janeiro, Brazil) under certification No. DIMCI 1507/2019 (certified purity: 99.988 ± 0.060%). The reagents used in the antioxidant assays, including 6-hydroxy-2,5,7,8-tetramethylchromane-2-carboxylic acid (Trolox), 2,2-diphenyl-1-picrylhydrazyl (DPPH^•), and the ammonium salt of 2,2’-azino-bis(3 ethylbenzothiazoline-6-sulfonic acid) (ABTS^•+), were obtained from Sigma-Aldrich (St. Louis, MO, USA). Likewise, albumin, glyoxal, bergenin, fructose, and sodium azide used in the AGE inhibition assay were also purchased from Sigma-Aldrich (St. Louis, MO, USA).

Plant material

E. protenta leaves were collected in the Adolpho Ducke Forest Reserve (2°55’42.41”S, 59°58’37.34”W), Manaus-AM, Brazil (SisGen No. AFDFDDC). This plant material was identified by Prof Dr Maria Anália D. Souza, and a voucher specimen (No. 264190) was deposited in the Herbarium of the National Institute of Amazonian Research (INPA).

Extraction procedure

The extraction systems used were selected based on the methodology described by Neves et al.,³ with adaptations to the experimental conditions of the present study. For the extractions, 1.0 g of plant material was submitted, in triplicate, to 10 mL of five different extraction systems: methanol, ethanol, isopropanol, ethyl acetate, and acetone. The extraction procedure was carried out exhaustively in four successive cycles and consisted of sonication in an ultrasonic bath for 15 min, followed by centrifugation at 4,000 rpm (4,226 g) for 10 min. After centrifugation, the supernatant was carefully separated and dried under a nitrogen stream until complete solvent removal.

Identification and quantification of gallic acid by NMR

For the ¹H NMR analysis, 10.0 mg of each extract was dissolved, in triplicate, in 550 μL of MeOD-d₄, sonicated in an ultrasonic bath for 10 min, and transferred to a 5 mm NMR tube. The analysis was performed on an NMR spectrometer (Bruker Avance III HD, 500.13 MHz, BBI probe) at the Nuclear Magnetic Resonance Laboratory (NMRLab) of the Federal University of Amazonas (UFAM). The pulse sequence used was zgpr with the following acquisition parameters: 32k time-domain data points (TD), spectral width (SW) of 8 kHz, acquisition time (AQ) of 2.04 s, receiver gain (RG) of 128, number of scans (NS) equal to 16, central frequency (O1) set to 2438.81 Hz, and presaturation power (PLW9) of 4.304 e^-005 W. The P1 value was automatically and individually calculated for each sample using the pulsecal sn command. The relaxation delay (D1 = 31.76 s) was calculated for the gallic acid signal at d_H 7.04 (s, 2H) using equation 1, in which the longitudinal relaxation time (T1) was determined by means of the t1ir1d pulse sequence.

(1) $$\text{D1}=7\times \text{T1}-\text{AQ}$$

DMT used as the external reference standard, was prepared in triplicate at a concentration of 18.65 mM in CDCl₃ (D, 99.9%), containing TMS (0.05% v/v) as the internal reference standard in 0.0 ppm (Figure S1, Supplementary Information (SI) section). For acquisition of the quantitative ¹H NMR spectrum, the 90° pulse of DMT (7.79 ms) was calculated for the signal at d_H 8.10 (s, 4H) using the 90° pulse experiment (zg). The longitudinal relaxation time (T1) was determined by the inversion-recovery experiment (t1ir1d) for the signal at d_H 8.10. After the T1 value had been obtained, D1 (23.20 s) was estimated using equation 1, in which the acquisition time (AQ) used was 2.04 s. Except for the P1 and D1 parameters, the remaining acquisition parameters used for the samples were also applied to the acquisition of the DMT spectrum.

Phase and baseline corrections, as well as signal integration, were performed manually using TopSpin 3.6.3 (Bruker Optics GmbH & Co. KG, Ettlingen, Germany, 2021). Heteronuclear single quantum coherence (HSQC) edit and heteronuclear multiple bond correlation (HMBC) NMR experiments were acquired to verify the absence of signal overlap with the signals of interest in the samples, and the ¹H-¹³C correlations in the 2D NMR experiments were acquired using ¹J_CH and ³J_CH coupling constants of 145 and 8 Hz, respectively. Quantitative analysis by the pulse length-based concentration determination (PULCON) method was performed using the electronic reference to access in vivo concentrations (ERETIC2) tool, employing the gallic acid singlet at d_H 7.04 (2H) for quantification and the aromatic signal of DMT at d_H 8.10 (4H) as the external reference.¹⁸,¹⁹

Determination of antioxidant activities

DPPH radical scavenging capacity

The experiments were carried out according to the methodology described by Samaniego-Sánchez et al.,²⁰ with adaptations by Mar et al.²¹ The free-radical scavenging capacity of E. protenta leaf extracts was evaluated using the DPPH^• assay. A 100 μM methanolic DPPH^• solution was prepared. The samples were prepared at a concentration of 1.0 mg mL^-1, and 10 μL of each sample solution was mixed with 1,900 μL of the methanolic DPPH^• solution. Trolox, at concentrations ranging from 100 to 2000 μM, was used as positive control. The mixture was incubated in the dark at 25 °C for 30 min. Absorbance was measured at 515 nm using a microplate reader (BioTek Instruments, Winooski, VT, USA). Antioxidant capacity was quantified in Trolox equivalents, and the assay was performed in triplicate. The relationship between absorbance and Trolox concentration was determined as y = -0.0004x + 0.7264, with an R² (coefficient of determination) value of 0.9944. All measurements were performed in triplicate, and the results were expressed as micromolar Trolox equivalents (μM Trolox g^-1).

ABTS radical cation scavenging capacity

The ABTS^•+ scavenging assay involves monitoring the discoloration of the ABTS^•+ solution in the presence of antioxidant extracts at a concentration of 1.0 mg mL^-1. The methodology described by Samaniego-Sánchez et al.²⁰ was used, with adaptations by Mar et al.²¹ After a 6 min reaction between the sample and the radical at a 1:10 ratio, absorbance was measured at 750 nm using a microplate reader (BioTek Instruments, Winooski, VT, USA). Trolox was used to construct the standard curve (y = 0.0003x + 0.7349, R² = 0.9989), and the results were expressed as mean ± standard deviation (n = 3) in micromolar Trolox equivalents (μM Trolox g^-1).

Determination of AGE inhibition potential

Oxidative pathway

Antiglycation activity was determined according to Kiho et al.,²² with modifications. The reaction was performed in triplicate using bovine serum albumin (BSA, 8.0 mg mL^-1), glyoxal (30 mM), and sample (1.0 mg mL ¹, dissolved in dimethyl sulfoxide (DMSO)). Glyoxal and BSA solutions were prepared in phosphate buffer (20 mM, pH 7.4) containing sodium azide (3 mM) as an antimicrobial agent. The total reaction mixture (300 μL) consisted of BSA (135 μL), glyoxal (135 μL), and DMSO or sample solution (30 μL), and was incubated at 37 °C for 24 h under sterile conditions and protected from light. After incubation, each sample was analyzed in a microplate reader by measuring fluorescence intensity (excitation at 230 and emission at 420 nm). Quercetin (100 μM) was used as the standard, and DMSO was used as the negative control. The results were calculated using equation 2 and expressed as mean ± standard deviation (n = 3) of the percentage inhibition.

(2) $$\text{Inhibition (\%)}=100-[(\text{FluorA}/\text{FluorC})\times 100]$$

where FluorA = sample fluorescence and FluorC = control fluorescence.

Non-oxidative pathway

Antiglycation activity was determined according to Kiho et al.,²² with modifications. The same procedure used for the oxidative pathway assay was followed, except that glyoxal was replaced with fructose (100 mM) and the incubation time was extended to 120 h. Aminoguanidine (100 μM) was used as the inhibition standard. The results were expressed as mean ± standard deviation (n = 3) of the percentage inhibition.

Statistical analysis

The distribution of the data related to extraction yield, gallic acid quantification, and biological activity was initially evaluated using the Kolmogorov-Smirnov normality test. Comparisons among multiple groups with normal distribution were performed by parametric ANOVA, followed by Tukey’s test, adopting a significance level of 5%. Pearson correlation coefficients were calculated to evaluate the relationships between quantitative data and biological activity data (Table S1, SI section), considering p < 0.05. All statistical analyses were performed using Minita^TM 13.0 (Minitab Inc., State College, PA, USA, 2000).

Results and Discussion

Effect of solvent on extraction yields

The extraction yields obtained for the different solvent systems are presented in Table 1. The methanol based extraction system showed the highest mean yield (30.6 mg g^-1 of dry extract), differing statistically from the other solvents evaluated (p < 0.05), as indicated by Tukey’s test. In contrast, the systems based on acetone and ethyl acetate showed the lowest yields, with no significant difference between them.

Table 1
Extraction yields of E. protenta leaf extracts obtained using different extraction systems

Extraction system	Yield / (mg g-1)	Extraction yield / %	P’
Methanol	30.6 ± 4.0a	3.1 ± 0.4a	5.1
Ethanol	12.0 ± 0.9b	1.2 ± 0.1b	4.3
Isopropanol	7.6 ± 0.2b,c	0.8 ± 0.0b,c	3.9
Acetone	6.4 ± 1.1c	0.6 ± 0.1c	5.1
Ethyl acetate	4.9 ± 0.7c	0.5 ± 0.1c	4.4

• ^a,b,c
  Grouping of extraction yields using Tukey’s test at a 95% confidence level. P’: polarity index; Yield expressed as mg g^-1 of plant material.

When the solvents were compared according to the polarity index (P’) derived from the Snyder triangle (Table 1), a general trend of increasing extraction yield with increasing solvent polarity was observed, suggesting that more polar extraction systems favored the recovery of polar constituents from E. protenta leaves.³,²³ In this context, methanol, which presents a relatively high polarity index (P’ = 5.1), afforded the highest extraction yield among the evaluated solvents, whereas isopropanol and ethyl acetate, with lower P’ values, resulted in lower yields. This behavior supports the usefulness of the Snyder polarity index as an initial parameter for rationalizing the influence of solvent properties on extraction performance.¹⁵,²⁴-²⁶ However, the relationship between P’ and extraction yield was not strictly linear. Methanol and acetone have the same polarity index (P’ = 5.1) but exhibited markedly different extraction yields, indicating that overall solvent polarity alone cannot fully account for the extraction behavior observed. This result suggests that, although P’ provides a useful first approximation of solvent effects, other physicochemical properties and more specific solute-solvent interactions also contribute to the selective recovery of constituents from the plant matrix. These additional interactions are examined in greater detail in the following section using the solvatochromic parameters of the LSER model.

Gallic acid content and its interpretation based on the LSER model

The analysis of the different E. protenta leaf extracts were carried out by means of ¹H NMR (Figure S2, SI section), ¹H-¹³C HSQC-edit, and ¹H-¹³C HMBC experiments. The ¹H NMR spectrum revealed the presence of an intense singlet at d_H 7.04 (Figure 1), which, in the HMBC experiment (Figure S3, SI section), showed correlations with five carbon signals: d_C 144.8 (C-3/C-5), 138.2 (C-4), 120.8 (C-1), 109.0 (C-2/C-6), and 168.8 (C(O)O-7). Combined, these spectroscopic data allowed the structure of gallic acid (3,4,5-trihydroxybenzoic acid) to be confirmed. In addition, the HSQC-edit experiment (Figure S4, SI section) showed that the signal at d_H 7.04 did not overlap with signals from other compounds present in the extracts, which made it possible to quantify this compound in the different extracts by the PULCON method (Table 2).

Table 2
Quantification of gallic acid in E. protenta leaf extracts by ¹H NMR using the PULCON method

Extraction system	Gallic acid / (mg g-1)A	Gallic acid / (mg kg-1)B
Methanol	36.8 ± 0.2ª	1130.0 ± 150.0a
Ethanol	28.1 ± 0.1b	340.0 ± 30.0b
Isopropanol	6.1 ± 0.1c	46.0 ± 1.0c
Acetone	9.5 ± 0.1d	61.0 ± 10.0d
Ethyl acetate	5.8 ± 0.0e	29.0 ± 4.0e

• ^a,b,c,d,e
  Grouping of the quantitative gallic acid values using Tukey’s test at a 95% confidence level;
• ^A
  gallic acid content expressed as mg g^-1 of dry extract;
• ^B
  gallic acid content expressed as mg kg^-1 of plant material.

Figure 1
¹H NMR spectrum (500 MHz, MeOD-d₄) of the methanolic extract of E. protenta leaves. Expansion of the region between d_H 7.45 and 6.66 highlights the singlet at d_H 7.04 assigned to the equivalent aromatic protons H-2 and H-6 of gallic acid. The absence of signal overlap in this region allowed the resonance at d_H 7.04 to be selected for quantification by the PULCON method.

Although gallic acid was selectively identified and quantified by NMR, a more comprehensive characterization of the extracts could be achieved by combining NMR with complementary techniques such as UHPLC-HRMS. Owing to its high sensitivity and mass accuracy, UHPLC HRMS could enable the detection and tentative annotation of metabolites present at lower concentrations that may not be readily observed by NMR, thereby providing a broader characterization of the chemical composition of the extracts.⁴,¹⁰ Within the scope of the present study, NMR was employed as a targeted approach for the identification and quantification of gallic acid. Therefore, gallic acid was considered a quantitative chemical marker rather than the sole determinant of biological activity, and the antioxidant and antiglycation responses should be interpreted as resulting from the combined contribution of multiple constituents present in the complex plant matrix.

Among the evaluated extracts, the methanolic extract showed the highest gallic acid concentration (36.8 ± 0.2 mg g^-1 dry extract), differing statistically from the other extraction systems (p < 0.05). The ethanolic extract showed an intermediate concentration (28.1 ± 0.1 mg g^-1 dry extract), whereas the extracts obtained with isopropanol, acetone, and ethyl acetate exhibited the lowest levels of this compound. As reported by Arya et al.,¹³ solvent selection strongly influences the recovery of phenolic compounds, and methanol was particularly effective for gallic acid extraction, providing the highest gallic acid content among the evaluated solvents. This finding is consistent with the results obtained for E. protenta, in which the methanolic extract also showed the highest gallic acid concentration.

The solvent-dependent differences observed in gallic acid recovery can be further interpreted based on the solvation parameters of the employed solvents.¹⁶,²⁷ Gallic acid is a highly polar phenolic compound containing three hydroxyl groups and one carboxyl group, whose solubilization is favored by solvents capable of establishing intense and specific intermolecular interactions.²⁸ In this context, the solvents were characterized by the solvatochromic parameters α, β, and π*, proposed by Taft et al.¹⁵ and consolidated in the LSER approach.¹⁶ This model describes solute-solvent interactions in terms of non-specific contributions, represented by π*, and specific contributions associated with hydrogen-bond formation, quantified by α and β.

Among the evaluated solvents, methanol presented the highest α value (0.98), together with relatively high β (0.66) and π* (0.60) values (Table 3).¹⁶ This combination indicates a strong hydrogen-bond donor ability, together with hydrogen-bond acceptor capacity and appreciable dipolarity/polarizability, which may favor interactions with the hydroxyl and carboxyl groups of gallic acid.¹⁵,¹⁶ Ethanol and isopropanol are also protic solvents and exhibit high β values, however, their lower α and π* values compared with methanol indicate a different balance of solvation properties, which may contribute to the lower gallic acid recovery observed in these extracts. In contrast, acetone and ethyl acetate exhibit very low α values (0.08 and 0.00, respectively), despite presenting moderate to high β and π* values.¹⁶ Notably, acetone showed the highest π* value (0.71) among the evaluated solvents but did not provide the highest gallic acid recovery, demonstrating that dipolarity/polarizability alone cannot account for the extraction behavior. Overall, these results indicate that gallic acid extraction cannot be attributed to a single solvent parameter, but rather to the combined contribution of hydrogen-bonding interactions and dipolarity/polarizability described by the LSER parameters.¹⁴-¹⁶

Table 3
Solvatochromic parameters α, β, and π^* of the solvents employed in the extraction systems of E. protenta leaves

Solvent	α	β	π*
Methanol	0.98	0.66	0.60
Ethanol	0.86	0.75	0.54
Isopropanol	0.76	0.84	0.48
Acetone	0.08	0.43	0.71
Ethyl acetate	0.00	0.45	0.55

Evaluation of the biological activities of E. protenta

DPPH^• and ABTS^•+ assays are widely used to evaluate the antioxidant capacity of pure substances and plant extracts because they are rapid, sensitive, and low-cost methods.²⁹,³⁰ Among the evaluated extraction systems, the extracts obtained with methanol and ethanol showed the best results (Figures 2a and 2b). In addition, the Pearson correlation between the two assays was 0.961 (p < 0.05), indicating a strong association between the results obtained by both methods and supporting the antioxidant potential of the analyzed extracts. Although DPPH^• and ABTS^•+ differ in the nature of the radical species and the reaction environment, these differences make the assays complementary for evaluating antioxidant capacity. DPPH^• is based on the reduction of a stable radical, whereas ABTS^•+ employs a radical cation that is more accessible to structurally diverse antioxidant compounds.²⁰ Consequently, the combined use of both assays provides complementary information on the radical-scavenging capacity of the extracts. The high correlation between DPPH^• and ABTS^•+ suggests that the compounds recovered by the different extraction systems exhibited similar antioxidant responses in both assays, despite the mechanistic differences between the two methods.

Figure 2
Antioxidant and antiglycation activities of E. protenta leaf extracts obtained using different extraction systems. (a) DPPH^• radical scavenging capacity; (b) ABTS^•+ radical cation scavenging capacity; (c) inhibitory capacity against AGE formation via the oxidative pathway; and (d) via the non oxidative pathway. Results are expressed as mean ± standard deviation (n = 3). *, **, ***, ****, ***** indicate grouping for antioxidant and antiglycation activities according to Tukey’s test at a 95% confidence level (p < 0.05).

This consistent antioxidant response across both assays suggests that compounds extracted by more polar solvents exhibit a greater free-radical scavenging capacity, which is consistent with the higher efficiency of these systems in recovering phenolic metabolites.³,¹⁴ In particular, the results obtained for the methanolic and ethanolic extracts indicate that solvent polarity directly influenced the recovery of constituents with antioxidant activity, reinforcing the relationship between chemical composition and biological response.

Considering that oxidative stress is also directly related to the formation of AGEs, the evaluation of the antiglycation potential of the extracts represents a complementary step toward understanding their biological activity.⁶,³¹ In this context, the ability of E. protenta leaf extracts to inhibit AGE formation through the non-oxidative and oxidative pathways was investigated (Figures 2c and 2d, respectively). The results showed that, although all extracts exhibited some inhibitory activity, none of them reached 50% inhibition in either pathway, indicating a low to moderate effect under the experimental conditions employed. In the oxidative pathway, the highest inhibition percentages were observed for the extracts obtained with ethanol (44.74 ± 2.12%), isopropanol (43.71 ± 7.33%), and methanol (43.32 ± 0.65%), which differed from the extracts obtained with acetone and ethyl acetate, both of which showed lower performance. In the non-oxidative pathway, the methanolic extract showed the highest inhibitory capacity (41.74 ± 0.99%), whereas the other extracts exhibited lower values. Although these assays provide complementary information on the antioxidant and antiglycation potential of the extracts, they are limited to in vitro assessments and do not encompass the broader biological effects that could be investigated using more complex experimental models.

Despite this limitation, the results show that alcoholic solvents favored the recovery of extracts with greater bioactive potential, expressed by their better performance in the antioxidant and antiglycation assays. This behavior may be explained not only by the higher polarity of these solvents, but also by the more favorable combination of the solvatochromic parameters α, β, and π*, which contribute to more efficient interactions with bioactive phenolic compounds.

Correlation between gallic acid content and biological activities

When Pearson correlations (p < 0.05) between gallic acid content and the results of the biological assays were evaluated, a moderate to strong positive correlation was observed between the concentration of this phenolic compound and the antioxidant and antiglycation activities of E. protenta extracts. Exploratory correlation analysis indicated that gallic acid content showed a strong association with the antioxidant activity determined by DPPH^• (r = 0.845) and a moderate association with the ABTS^•+ assay (r = 0.681). Regarding antiglycation activity, a moderate correlation was observed for the oxidative pathway (r = 0.603) and a moderately strong correlation for the non-oxidative pathway (r = 0.711). These results support the interpretation that gallic acid may contribute to the antioxidant and antiglycation properties of the extracts. In particular, the stronger correlations observed for DPPH^• scavenging activity and inhibition of the non-oxidative glycation pathway are consistent with the known bioactive properties of this phenolic compound. This pattern further supports the hypothesis that more polar solvents favored the recovery of phenolic-rich fractions, including gallic acid, which may be associated with the enhanced biological responses observed. However, the biological activities cannot be attributed exclusively to gallic acid, since plant extracts are complex mixtures containing multiple phenolic constituents that may act through additive or synergistic effects. Therefore, gallic acid should be regarded as a relevant quantitative chemical marker of the phenolic fraction rather than as the sole determinant of the observed biological activities. These findings reinforce the usefulness of the marker-directed qNMR approach for monitoring solvent-dependent enrichment of bioactive phenolic constituents in E. protenta extracts.

Conclusions

The present study demonstrates that solvent choice directly affects extraction yield, targeted phenolic recovery, and the biological responses of E. protenta leaf extracts. The combined use of the Snyder triangle and LSER parameters demonstrated that extraction performance depends not only on overall polarity, but also on specific solvation properties, especially those related to hydrogen-bonding interactions. Methanol proved to be the most efficient solvent among the systems evaluated, affording the highest extraction yield and the highest gallic acid content. However, due to the known toxicity of methanol, methanol-based extraction is more appropriate for the preparation of dried extracts or for analytical purposes. In this context, qNMR proved suitable for the targeted evaluation of these extracts, allowing gallic acid to be established as a useful chemical marker for monitoring solvent-dependent differences in the recovery of the phenolic fraction from a complex plant matrix, without implying a complete molecular characterization of the extracts. In addition, the better antioxidant and antiglycation performance of the alcoholic extracts was consistent with their greater ability to recover polar bioactive phenolic compounds. These findings provide analytical and chemical support for the rational selection of extraction systems for E. protenta and other phenolic-rich plant materials.

Supplementary Information

Supplementary Information (dataset used for Pearson correlation analyses, ¹H NMR spectrum of dimethyl terephthalate, comparative ¹H NMR profiles of the extracts, and HMBC and HSQC-edit spectra) is available free of charge at http://jbcs.sbq.org.br as a PDF file.

Supplementary PDF

Supplementary PDF

Acknowledgments

The authors gratefully acknowledge the CAPES (88887.506792/2020-00, finance code 001), CNPq (421935/2023 5, 318684/2025-0), FINEP and FAPEAM (No. 010/2021 - CT&I Priority Areas). The authors further acknowledge the Analytical Center of the Federal University of Amazonas (UFAM) for providing the infrastructure for the NMR analyses.

Data Availability Statement

The datasets generated and/or analyzed during the current study are fully available within the article and its Supplementary Information files.

References

• ¹ de Souza, A. M.; de Oliveira, C. F.; de Oliveira, V. B.; Betim, F. C. M.; Miguel, O. G.; Miguel, M. D.; Planta Med 2018, 84, 1232. [Crossref]
  » Crossref
• ² Araújo, F.; Numa, I. N.; Farias, D.; Cunha, G.; Pastore, G.; Food Res. Int 2019, 121, 57. [Crossref]
  » Crossref
• ³ Neves, K. O. G.; Silva, S. O.; Cruz, M. S.; Mar, J. M.; Bezerra, J. A.; Sanches, E. A.; Cassani, N. M.; Antoniucci, G A.; Jardim, A. C.; Chaves, F. C. M.; Acho, L. D. R.; Lima, E. S.; Machado, M. B.; Santos, A. D. C.; Molecules 2025, 30, 713. [Crossref]
  » Crossref
• ⁴ Oliveira, E. S. C.; Pontes, F. L. D.; Acho, L. D. R.; da Silva, B. J. P.; do Rosário, A. S.; Chaves, F. C. M.; Campos, F. R.; Bezerra, J. A.; Lima, E. S.; Machado, M. B.; Phytochem. Anal. 2024, 35, 552. [Crossref]
  » Crossref
• ⁵ Harris, C. S.; Beaulieu, L. P.; Fraser, M. H.; McIntyre, K. L.; Owen, P. L.; Martineau, L. C.; Cuerrier, A.; Johns, T.; Haddad, P. S.; Bennett, S. A.; Arnason, J. T.; Planta Med. 2011, 77, 196. [Crossref]
  » Crossref
• ⁶ Sadeghi, M.; Miroliaei, M.; Kamyabiamineh, A.; Taslimi, P.; Ghanadian, M.; Arabian J. Chem. 2023, 16, 105143. [Crossref]
  » Crossref
• ⁷ Haminiuk, C. W. I.; Plata-Oviedo, M. S. V.; Mattos, G. D.; Carpes, S. T.; Branco, I. G.; J. Food Sci. Technol. 2014, 51, 2862. [Crossref]
  » Crossref
• ⁸ Samy, M. N.; Sugimoto, S.; Matsunami, K.; Otsuka, H.; Kamel, M. S.; J. Nat. Prod 2014, 7, 37. [Link] accessed in August 2026
  » Link
• ⁹ Justino, A. B.; Silva, H. C. G.; Franco, R. R.; Pimentel, I. O. C.; Silva, N. F.; Saraiva, A. L.; Espindola, F. S.; J. Ethnopharmacol 2022, 285, 114902. [Crossref]
  » Crossref
• ¹⁰ Sarges, F. N.; Campos, M. M.; Moura, S. L.; Gonçalves, G. M. S. P.; Sarquis, E. O.; Zaranza, M. G. B.; Alencar, E. H. A.; Rodrigues, A. P. D.; Costa, B. F.; Figueiredo, R. N. M.; Nat. Prod. Res 2019, 33, 879. [Crossref]
  » Crossref
• ¹¹ Hadidi, M.; Liñán-Atero, R.; Tarahi, M.; Christodoulou, M. C.; Aghababaei, F.; Antioxidants 2024, 13, 1001. [Crossref]
  » Crossref
• ¹² Wianowska, D.; Olszowy-Tomczyk, M.; Molecules 2023, 28, 1186. [Crossref]
  » Crossref
• ¹³ Arya, O. P.; Bhatt, I. D.; Mohanty, K.; J. Appl. Res. Med. Aromat. Plants 2024, 40, 100547. [Crossref]
  » Crossref
• ¹⁴ Johnson, A. R.; Vitha, M. F.; J. Chromatogr. A 2011, 1218, 556. [Crossref]
  » Crossref
• ¹⁵ Taft, R. W.; Abboud, J.-L. M.; Kamlet, M. J.; Abraham, M. H.; J. Solution Chem 1985, 14, 153. [Crossref]
  » Crossref
• ¹⁶ Deng, Z.; Li, F.; Zhao, G.; Yang, W.; Hu, Y.; J. Chem. Eng. Data 2020, 65, 3579. [Crossref]
  » Crossref
• ¹⁷ Gowda, G. A. N.; Raftery, D.; Anal. Chem 2017, 89, 490. [Crossref]
  » Crossref
• ¹⁸ Mak, J. Y. W.; Aust. J. Chem. 2022, 75, 160. [Crossref]
  » Crossref
• ¹⁹ TopSpin, version 3.6.3; Bruker Optics GmbH & Co. KG; Ettlingen, Germany, 2021.
• ²⁰ Samaniego-Sánchez, C.; Inurreta-Salinas, Y.; Quesada Granados, J. J.; Blanca-Herrera, R.; Villalón Mir, M.; Serrana, H. L. G.; Martínez, M. C. L.; J. Food Compos. Anal 2011, 24, 79. [Crossref]
  » Crossref
• ²¹ Mar, J. M.; da Silva, L. S.; Moreira, W. P.; Biondo, M. M.; Pontes, F. L. D.; Campos, F. R.; Kinupp, V. F.; Campelo, P. H.; Sanches, E. A.; Bezerra, J. A.; Food Chem 2021, 356, 129723. [Crossref]
  » Crossref
• ²² Kiho, T.; Usui, S.; Hirano, K.; Aizawa, K.; Inakuma, T.; Biosci., Biotechnol., Biochem. 2004, 68, 200. [Crossref]
  » Crossref
• ²³ de Araújo, F. H. S.; Nogueira, C. R.; Trichez, V. D. K.; Guterres, Z. R.; Pinto, L. S.; Velter, S. Q.; Ferreira, G. A. M.; Machado, M. B.; Neves, K. O. G.; Vieira, M. C.; Cardoso, C. A. L.; Heredia-Vieira, S. C.; Oliveira, K. M. P.; Piva; R. C.; Oesterreich, S. A.; J. Ethnopharmacol 2023, 303, 115991. [Crossref]
  » Crossref
• ²⁴ Dai, J.; Mumper, R. J.; Molecules 2010, 15, 7313. [Crossref]
  » Crossref
• ²⁵ Ignat, I.; Volf, I.; Popa, V. I.; Food Chem 2011, 126, 1821. [Crossref]
  » Crossref
• ²⁶ Snyder, L. R.; J. Chromatogr. A 1974, 92, 223. [Crossref]
  » Crossref
• ²⁷ Kamlet, M. J.; Abboud, J. L.; Taft, R. W.; J. Am. Chem. Soc. 1977, 99, 6027. [Crossref]
  » Crossref
• ²⁸ Park, Y.-R.; Lee, B.-S.; Separations 2024, 11, 36. [Crossref]
  » Crossref
• ²⁹ Huang, D. J.; Ou, B. X.; Prior, R. L.; J. Agric. Food Chem. 2005, 53, 1841. [Crossref]
  » Crossref
• ³⁰ Dudonne, S.; Vitrac, X.; Coutiere, P.; Woillez, M.; Merillon, J. M.; J. Agric. Food Chem. 2009, 57, 1768. [Crossref]
  » Crossref
• ³¹ Dil, F. A.; Ranjkesh, Z.; Goodarzi, M. T.; Diabetes Metab. Syndr 2019, 13, 1225. [Crossref]
  » Crossref

*

e-mail: [e-mail]

Editor handled this article:

João Henrique Ghilardi Lago (Associate)

Publication Dates

• Publication in this collection
  28 Sept 2026
• Date of issue
  2026

History

• Received
  08 July 2026
• Accepted
  28 Aug 2026