From Urban Forest Pruning to Cosmetics: Bioactive Potential of Twig Extracts from Selected Woody Species

Abstract

Urban forest management practices generate substantial amounts of twig biomass that is commonly treated as waste, despite its potential as a source of bioactive compounds. Biological and chemical properties of methanolic extracts of 19 urban forest tree and shrub species were assessed using a multidisciplinary approach combining high-performance thin-layer chromatography (HPTLC) and HPTLC-DPPH bioautography with spectrophotometric determination of radical scavenging activity (RSA), total phenolic content (TPC), inhibition assays of skin aging-related enzymes (tyrosinase and elastase), and testing against skin pathogens Staphylococcus aureus and Pseudomonas aeruginosa. The results revealed marked differences in biological activity among extracts, driven primarily by specific phytochemical profiles. Torminalis glaberrima (Gand.) Sennikov & Kurtto (108.8 ± 6.6 μmol TE/mL) and Paliurus spina-christi Mill. (106.6 ± 1.6 μmol TE/mL) exhibited the highest RSA, correlating with elevated TPC. Acer campestre L. (51.6 ± 9.1%) showed the strongest elastase inhibition. The most pronounced tyrosinase inhibition was observed for Torminalis glaberrima (39.0 ± 3.5%), indicating a significant contribution of TPC. In contrast, the strongest antibacterial activity was recorded for Acer platanoides L. and Carpinus betulus L., despite their lower TPC values, suggesting the contribution of non-phenolics. Phenolic zones (R_F 0.10, 0.28, 0.57, 0.58) were identified as putative markers of the observed bioactivities. Overall, twigs emerge as an underexplored source with considerable potential for natural cosmetics development, warranting further investigations.

1. Introduction

Urban forests are increasingly recognized as multifunctional ecosystems delivering diverse ecosystem services, including climate regulation, biodiversity conservation, improved air quality, and enhanced human well-being [1]. Alongside these ecological and social benefits, urban forests are subject to continuous management practices such as pruning, thinning, and sanitation cutting, which generate substantial amounts of woody biomass. This lignocellulosic biomass, particularly twigs and small branches, is often treated as waste, despite its considerable potential for valorization within the frameworks of sustainable forest management and the circular bioeconomy [1]. Urban pruning has been estimated to generate up to 47 kg of biomass per capita annually, posing a challenge for waste management systems [2]. These residues are typically managed through chipping, composting, or energy recovery, but such methods do not fully exploit their chemical potential. According to the Forest Management Plan for the “Košutnjak” unit (2017–2026) [3], the planned annual harvesting volume is approximately 297 m³. Forest residues may account for around 20% of this volume (≈59.4 m³), with branches representing the dominant fraction (≈70%, ≈41.6 m³), providing a substantial and continuously available source of twig biomass suitable for valorization. This classification aligns with the Forest Management Regulation of the Republic of Serbia [4], defining logging residues as woody biomass not processed into standard timber assortments, typically including small-diameter branches. From a forestry perspective, these residues represent a renewable feedstock for the recovery of value-added compounds, supporting resource-efficient management and circular bioeconomy strategies.

Twig biomass represents a metabolically active plant tissue involved in defense responses and environmental adaptation. Consequently, twigs are frequently enriched in secondary metabolites, particularly phenolic compounds, which play a crucial role in plant protection against oxidative stress, pathogens, and herbivores [5,6]. While phenolic compounds derived from leaves, bark, and wood have been extensively investigated [7], twigs remain a largely underexplored forest resource. This is especially relevant in urban forests, where pruning residues are continuously produced and represent a readily available and sustainable source of potentially valuable bioactive compounds.

Phenolic secondary metabolites are of particular interest in cosmetic science due to their well-documented antioxidant, antimicrobial, and enzyme-inhibitory properties, which contribute to skin protection and anti-aging effects [5,6,8,9]. Antioxidants play a central role in mitigating oxidative damage induced by UV radiation and urban pollution, while natural inhibitors of skin aging-related enzymes, such as elastase and tyrosinase, are highly sought after for controlling pigmentation, preserving skin elasticity, and preventing wrinkle formation [10,11]. Consequently, plant-derived phenolics are increasingly favored as safer and more sustainable alternatives to synthetic cosmetic ingredients.

Xerothermophilous oak habitats, particularly in Southern Europe, are exposed to intense solar radiation and prolonged periods of drought, creating environmental conditions that favor light-demanding and drought-tolerant plant species. Such abiotic stressors are known to activate plant defense mechanisms, often resulting in enhanced biosynthesis of secondary metabolites [12,13]. A representative example of this ecological setting is the forest community of Hungarian oak and Turkey oak with butcher’s broom (Quercetum frainetto-cerridis Rudski (1940) 1949 var. geograph. Ruscus aculeatus B. Jovanović 1951) [14]. This forest type is characteristic of Belgrade and central Serbia, where most settlements were established on former habitats of this community, making it one of the forest types most closely and historically associated with human presence [15]. In the absence of extensive historical clearing, these forests would represent the dominant potential vegetation of the region. Consequently, they are among the forest types most closely linked to human presence throughout history.

The urban “Košutnjak Forest” was protected for its important spatial and bio-ecological values, including its forest habitats that support diverse mammals, birds, insects, reptiles, and amphibians, as well as geological features considered natural rarities significant for understanding Belgrade’s geological history. The flora of the Košutnjak area was extensively analyzed, with 188 species documented in Gajić (1952) [14]. The taxonomic analysis showed dominance of the following families: Fabaceae (12.2%), Asteraceae (10.1%), Poaceae (9.6%), Rosaceae (8.5%), Lamiaceae (6.9%), Caryophyllaceae (6 species, 3.2%), Apiaceae (3.2%), and Plantaginaceae (3.2%) [14].

Valorizing bioactives from urban forest pruning residues supports non-timber forest products and high-value biomass use without additional harvesting. Converting these low-value residues into cosmetic ingredients combines urban forest management, resource efficiency, and sustainable product development [1,16]. Despite growing interest in plant-based cosmetics, woody species in urban forests remain largely unexplored, and exploring their potential could enhance sustainable urban forestry and diversify forest-based value chains.

To our knowledge, the biological and chemical profiles of plant extracts from species in the urban Košutnjak forest have not yet been investigated using HPTLC and HPTLC-DPPH bioautography, combined with radical scavenging activity (RSA) and skin-aging enzyme inhibition assays. Previous studies focused mainly on taxonomy, floral composition, and life forms [14,15], or on phytochemical characterization of bark, flower, and leaf extracts from these species, while twig extracts remain largely understudied [17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32].

Therefore, this study aimed to evaluate the cosmetic-relevant bioactive potential of 19 twig extracts from selected urban forest tree and shrub species. A multidisciplinary approach combining HPTLC profiling with spectrophotometric assessment of radical scavenging capacity, total phenolic content, elastase and tyrosinase inhibition, and antibacterial activity against common skin pathogens (Staphylococcus aureus and Pseudomonas aeruginosa) was applied. Correlating phytochemical profiles with biological activities highlights urban forest twig biomass as a sustainable source of cosmetic bioactives, with the most potent species suitable as natural raw materials under sustainable sourcing principles. The results of this study include in vitro biochemical assays used to evaluate the effectiveness of applying twig biomass in cosmetic formulations, while future research will focus on assessing the safety potential of the investigated extracts for cosmetic use, including cytotoxicity assessments, cell-based assays, and dermatological evaluations.

2. Materials and Methods

2.1. Chemicals and Reagents

Kojic acid (≥98.5%, KA), epigallocatechin gallate (EGCG), 3,4-dihydroxy-L-phenylalanine (L-DOPA), mushroom tyrosinase, Tris-HCl (≥99%), porcine pancreatic elastase (PPE), and N-succinyl-Ala-Ala-Ala-p-nitroanilide (≥98%), 2-aminoethyl diphenylborinate (Natural Product reagent) and polyethylene glycol (PEG) were obtained from Sigma-Aldrich (St. Louis, MO, USA). Methanol and toluene were obtained from ZORKA Pharma (Šabac, Serbia). The free radical 2,2-diphenyl-1-picrylhydrazyl (DPPH) was purchased from Fluka (Buchs, Switzerland). Folin–Ciocalteu (FC) reagent, sodium carbonate, formic acid, sodium dihydrogen phosphate, sodium hydrogen phosphate and glass HPTLC plates silica gel 60 F₂₅₄ (Art. 1.05642.0001) were obtained from Merck (Darmstadt, Germany). Quercetin-3-O-glucoside (Q-3-O-G), chlorogenic acid (CA), and kaempferol-3-O-glucoside (K-3-O-G) were purchased from Sigma-Aldrich (Darmstadt, Germany). Ethyl acetate was obtained from BETAHEM (Belgrade, Serbia). Tripton LP0042 and yeast extract LP0021 were supplied by Oxoid LTD (Basingstoke, UK), and nutrient agar was provided by Lab M (Bury, UK). Streptomycin and sodium chloride were purchased from Sigma-Aldrich Chemie GmbH (Steinheim, Germany).

2.2. Plant Material

2.2.1. Study Area

Plant material was collected from the Natural Monument “Košutnjak Forest” located within the administrative territory of the City of Belgrade, Serbia (44.7668° N, 20.4353° E; Figure 1). “Košutnjak Forest” is an urban forest of state-owned land that covers 265.26 ha of protected land. This forest ranges in elevation from 79.0 to 216.9 m above sea level (City Assembly of Belgrade, 2014). The forest belongs to a complex of xerothermophilous oak forest types classified according to the European Nature Information System (EUNIS) habitat classification (types G1.7612; G1.7617; G1.7A1; G1.A1C), hosting species of diverse floristic elements [33].

2.2.2. Plant Material Collection

Twig samples from 19 plant species belonging to 12 different families (Table S1): Sapindaceae (Acer campestre L. (1), Acer platanoides L. (2), Acer tataricum L. (3)); Betulaceae (Carpinus betulus L. (4)); Oleaceae (Fraxinus ornus L. (5), Ligustrum vulgare L. (14)); Fagaceae (Quercus cerris L. (6), Quercus pubescens Willd. (7)); Rosaceae (Torminalis glaberrima (Gand.) Sennikov & Kurtto (8), Crataegus monogyna Jacq. (13), Rosa canina L. (16), Pyrus pyraster (L.) Burgsd. (19)); Malvaceae (Tilia tomentosa Moench (9), Tilia cordata Mill. (10)); Berberidaceae (Berberis vulgaris L. (11)); Cornaceae (Cornus mas L. (12)); Rhamnaceae (Paliurus spina-cristi Mill. (15)); Asparagaceae (Ruscus aculeatus L. (17)); and Viburnaceae (Viburnum lantana L. (18)) were collected during November 2025.

Twigs of approximately two years of age were selected based on morphological indicators of annual growth. For each species, samples were collected from three different individuals at different locations and then combined into a single composite sample. Since this study focuses on the urban forest ecosystem, composite samples were intentionally used to represent the forest-level chemical profile of twig biomass rather than variability among individuals of selected species. Samples were collected from the lower canopy layer, covering all sides of each sampled tree or shrub. After sampling, the twigs were packed in paper bags and air-dried at room temperature in a cool, well-ventilated environment for approximately 7 days before laboratory processing.

2.2.3. Twig Extraction

Dried twig materials were pulverized for 5 min to obtain a homogeneous powder using a household grinder (Gorenje, Belgrade, Serbia). Compounds present in pulverized twigs were extracted with methanol using an ultrasound-assisted extraction procedure (ELTA 90 Medical Science, Belgrade, Serbia). The extraction was performed for 30 min at room temperature, employing a solid-to-solvent ratio of 1:10 (m/v). After extraction, the solvent was removed under reduced pressure using a rotary evaporator (IKA RV 05, IKA Werke, Staufen, Germany). The dried extracts were then redissolved in methanol to obtain a final concentration of 100 mg/mL and kept refrigerated −18 °C until further use.

2.3. HPTLC Analysis

HPTLC analysis was performed on glass silica gel plates (Merck, Darmstadt, Germany) using a Linomat 5 applicator (CAMAG, Muttenz, Switzerland). Samples were applied as 6 mm wide bands, positioned 8 mm from the lower edge of the plate, with a minimum distance of 12 mm from each side. Chromatographic development was carried out up to a migration distance of 70 mm in a saturated chamber after 30 min of equilibration, using a mobile phase composed of toluene–ethyl acetate–formic acid (4:6:1, v/v/v). After development, the HPTLC chromatograms were dried and derivatized as required for the specific assay. Both chromatograms (Figure 2a,b) were documented using a smartphone camera (Samsung Galaxy A54, Samsung Electronics, Suwon, Republic of Korea) [34].

2.3.1. Natural Product (NP) Reagent Derivatization

Twig extracts (4 μL, 25 mg/mL, 1–19) and phenolic standard mixture (20) were applied onto the plates. After chromatographic development, the HPTLC chromatograms were derivatized by sequential immersion in 0.5% (w/v) NP reagent and 5% (w/v) methanolic PEG 400, with intermediate drying at 110 °C for 3 min after each derivatization step. After final drying, fluorescent bands were visualized under UV light at 366 nm and documented.

2.3.2. DPPH Derivatization

Twig extracts (2 μL, 25 mg/mL, 1–19) and phenolic standards (20) were applied onto the plates. After chromatographic development, the HPTLC chromatograms were dried with a stream of warm air for 5 min and subsequently derivatized by immersion in 0.1% (w/v) DPPH reagent in methanol. The HPTLC chromatogram was documented after 30 min, when no further formation of new yellow zones was observed, under white light.

Preparation of a Standard Solution Mixture

Phenolic reference compounds—namely CA, Q-3-O-G, and K-3-O-G were used for HPTLC analysis. A standard mixture was prepared by combining 30 µL of each phenolic reference solution (1 mg/mL concentration), yielding a final concentration of 0.33 mg/mL. The resulting standard mixture was applied to both HPTLC plates as 2 µL bands.

2.4. Spectrophotometric Assays

The radical scavenging (RSA) and total phenolic content (TPC) assays were conducted using a GBC Cintra 6 UV–visible spectrophotometer (GBC Scientific Equipment Ltd., Dandenong, VIC, Australia), with measurements recorded at the appropriate wavelengths. All analyses were performed in triplicate, and the results are expressed as mean values ± standard deviation.

2.4.1. Radical Scavenging Assay

The free RSA of the twig extracts was evaluated using the DPPH assay [9]. An aliquot of 0.1 mL of each extract was added to 3.9 mL of a methanolic DPPH-radical solution (71 μmol/L). The reaction mixtures were incubated for 60 min in the absence of light to prevent photodegradation of the radical. Following incubation, the decrease in absorbance was recorded at 517 nm. Antioxidant capacity was quantified by comparison with a Trolox standard calibration curve (100–600 μmol/L), and the results were expressed as micromoles of Trolox equivalents (TE) per milliliter of twig extract (μmol TE/mL).

2.4.2. Total Phenolic Content

TPC of the twig extracts was determined using the Folin–Ciocalteu colorimetric method as described in the literature [9]. Appropriately diluted extracts (0.5 mL) were mixed with distilled water (0.5 mL) and Folin–Ciocalteu reagent (2.5 mL, 10%, v/v). After 5 min at room temperature, sodium carbonate solution (2 mL, 7.5%, w/v) was added, and the mixtures were incubated for 2 h at room temperature. Quantification was performed using a gallic acid calibration curve (20–120 mg/L), and results were expressed as mg gallic acid equivalents per milliliter of twig extract (mg GAE/mL).

2.4.3. Skin Aging-Related Enzyme Assays

Both spectrophotometric assays for the determination of tyrosinase (TI) and elastase inhibition (EI) were performed using a BioTek 800 TS microplate spectrophotometer (Agilent Technologies, Inc., Headquarters, Santa Clara, CA, USA). The tested concentration of the investigated extracts (1–19) was 100 µg/mL (final concentration in the well). All assays were performed in triplicate for each sample, and results are presented as percentage inhibition mean (%) ± standard deviation. Kojic acid (KA) and epigallocatechin gallate (EGCG) were used as standard inhibitors for TI and EI assays, respectively.

Tyrosinase Inhibition Assay

Tyrosinase inhibition was determined using an established spectrophotometric method, with L-DOPA as the substrate [9]. The reaction mixture consisted of 20 mmol/L phosphate buffer (pH 6.8), 5 mM L-DOPA (in a buffer), and 480 U/mL mushroom tyrosinase (in a buffer), prepared with the tested sample (C) or without the sample (A). Control mixtures contained buffer and substrate with the sample but without enzyme (D) and without both enzyme and sample (B). All mixtures were pre-incubated for 15 min at 37 °C, after which the substrate was added to initiate the reaction. Following 20 min of incubation, the absorbance was measured at 475 nm using a microplate reader. Tyrosinase inhibitory activity was calculated according to the following equation:

$$I\%={\displaystyle \frac{\left(A-B\right)-(C-D)}{\left(A-B\right)}}\times 100$$

Elastase Inhibition Assay

Elastase inhibition was determined using our modified procedure based on a previously reported method in the literature, with slight modification [9]. The assay was performed in 96-well microplates. Reaction wells contained elastase enzyme (1 U/mL), elastase substrate (1 mmol/L), and buffer (50 mmol/L Tris–HCl, pH 8.0) and served as the control reaction (A). For sample evaluation, the same reaction mixture was prepared with the addition of the tested sample (C). Control wells contained either buffer and substrate only (B) or buffer, substrate, and sample without enzyme (D). The release of p-nitroanilide was monitored spectrophotometrically at 405 nm, 15 min after substrate addition, which was visually observed as the development of a yellow coloration of varying intensity in the wells. The percentage of EI was calculated analogously to the tyrosinase inhibition assay, using the same calculation formula.

2.5. Agar-Well Diffusion Test

The antibacterial activity of the extracts was evaluated using an agar well diffusion method against Staphylococcus aureus ATCC 6538 and Pseudomonas aeruginosa ATCC 15692. These strains were selected as representative Gram-positive and Gram-negative models, respectively, which are frequently implicated in secondary skin infections and are major targets in the development of topical dermo-cosmetic formulations [35,36]. The assay was performed according to the procedure described by Lazović et al. with minor modifications [34]. Briefly, bacterial cultures were grown in LB broth to an optical density (OD₆₀₀) of approximately 0.5, measured at 600 nm using a CINTRA 6 UV–Vis spectrophotometer (GBC Scientific Equipment Ltd., Dandenong, VIC, Australia). The LB medium was prepared by dissolving 10 g of tryptone, 5 g of yeast extract, and 5 g of NaCl in one liter of distilled water, followed by sterilization by autoclaving at 121 °C for 15 min. The resulting bacterial suspensions were mixed with nutrient agar at a ratio of 1:50 (v/v) and poured into sterile Petri dishes. Twig extracts, previously diluted in methanol (1:4, v/v), were applied to 10 mm diameter wells in the inoculated agar at a volume of 100 µL per well. The plates were incubated at 37 °C for 24 h, after which the inhibition zones were evaluated. Streptomycin (10 mg/mL in phosphate-buffered saline) served as a positive control, while methanol alone was used as a negative control. All experiments were performed in triplicate.

2.6. Statistical Analysis

Statistical analysis was performed for all spectrophotometric assays, including RSA, TPC, TI, and EI assays. For each assay, differences among extracts were evaluated using one-way analysis of variance (ANOVA), with the level of statistical significance set at p < 0.05. When a significant overall effect was observed, Tukey’s multiple comparisons test was applied as a post hoc procedure to identify statistically significant differences between individual samples. Assumptions of homogeneity of variances were assessed using Brown–Forsythe and Bartlett’s tests. All statistical analyses were conducted using GraphPad Prism software (version 6.04 (2014), GraphPad Software, San Diego, CA, USA).

2.7. Chromatogram Data Analysis

2.7.1. ImageJ Processing

HPTLC plates were documented as digital images in JPG format at 300 dpi (Figure 2b,c). The acquired images were resized and spatially standardized to 1210 × 442 pixels using the built-in image capture utility of the operating system (snipping tool). Subsequent image treatment and data extraction were carried out in ImageJ software (version 1.53t, National Institutes of Health, USA). To minimize background interference and improve the contrast of chromatographic zones, a background normalization procedure was applied employing the rolling ball algorithm (Process → Subtract Background→ radius: 50 pixels).

To enhance analytical selectivity and enable deeper signal interrogation, the color chromatogram (Figure 2b) was decomposed into individual red, green, and blue components, which were evaluated separately (Image → Color → Split Channels). In contrast, the chromatogram shown in Figure 2c was transformed into an 8-bit grayscale format before further analysis (Image → Type → 8-bit).

For each track, intensity profiles were extracted along the migration path, generating numerical datasets that describe variations in pixel intensity as a function of migration distance (i.e., R_F values). The migration path was discretized into 436 pixels, corresponding to the distance from the application point to the solvent front. Each pixel position was normalized by dividing it by the total number of pixels (436), yielding values in the range of 0–1, which correspond to R_F values. These datasets were subsequently subjected to signal pretreatment and multivariate chemometric analysis.

2.7.2. Principal Component Analysis (PCA)

The numerical data matrices obtained from HPTLC image processing were subjected to principal component analysis (PCA) to explore the underlying data structure, assess similarities among samples, and identify variables responsible for sample discrimination. All data preprocessing and chemometric analyses were carried out in MATLAB (version 7.4.0.287, R2007a; The MathWorks Inc., Natick, MA, USA) [34].

Prior to PCA, the data were preprocessed to enhance comparability between chromatographic profiles and to reduce non-informative variability [34]. A confidence level of 95% was defined for statistical evaluation. Variable alignment was initially applied using a peak-based approach (slack = 5) to correct minor shifts in migration distance across profiles. The aligned data were subsequently transformed using standard normal variate (SNV) to normalize signal intensities, followed by mean centering to place all variables on a common reference level before multivariate analysis.

PCA was computed using the singular value decomposition (SVD) algorithm on a dataset comprising 19 samples (extracts) × 437 variables (pixels). The number of models’ principal components (PCs) is selected according to commonly accepted criteria, taking into account the explained variance and model interpretability. Separate PCA models were constructed for each dataset, including individual color channels of the chromatogram presented in Figure 2b as well as the grayscale chromatogram shown in Figure 2c. Model reliability was evaluated using Venetian blinds cross-validation with 9 splits and a blind thickness of 1. Score plots were used to visualize grouping patterns and relationships among samples, while loading plots were examined to identify the variables contributing most strongly to the observed separation.

3. Results and Discussion

3.1. HPTLC Analysis

3.1.1. NP Reagent Derivatization

HPTLC enables detailed profiling of plant secondary metabolites, particularly phenolic compounds. Under UV-254 nm, aromatic ring-containing compounds (e.g., flavonoids, phenols, phenolic acids, alkaloids, and certain aromatic steroids) appear as dark quenching zones due to strong π–π transitions, while their R_F values provide preliminary information on polarity. In contrast, UV-366 nm reveals the intrinsic fluorescence of many conjugated structures, with phenolics emitting yellow, green, or pale blue signals [9]. These signals are further enhanced after NP derivatization, facilitating visualization and tentative structural characterization. The combined use of UV-254 and UV-366 provides insight into functional groups and the degree of natural product conjugation.

General Phenolic Profiles and Common Markers

The HPTLC analysis of 19 twig extracts revealed complex and varied phenolic profiles (Figure 2a,b). Based on comparison with reference standards, three phenolic compounds were identified: quercetin-3-O-glucoside (Q-3-O-G, orange band, R_F 0.03), chlorogenic acid (CA, light-blue band, R_F 0.07), and kaempferol-3-O-glucoside (K-3-O-G, light-green band, R_F 0.15) (Figure 2a,b). The standards Q-3-O-G, CA, and K-3-O-G were selected based on literature reports confirming their presence in the twigs and bark of the studied genera [5,25,27,32].

Most extracts exhibited multiple well-resolved bands, with samples 5 and 18 displaying the most intense zones. Extracts 4, 6, 7, 8, 10, 13, 16, and 18 appeared to contain the highest number of individual zones. In contrast, extracts 17 and 19 seemed relatively poor in phenolic content, with extract 11 showing the least prominent bands.

In the high-R_F region (R_F > 0.6), the extracts showed generally similar patterns, which may correspond to tentatively identified common twig constituents such as non-polar phenolic acids (blue zones at R_F 0.82 and 0.88), and possibly stilbenes or porphyrinic pigments (red zones at R_F ≈ 1.0), detectable under UV-366 nm [37].

A feature observed across numerous extracts (5, 6, 7, 8, 14, 17, 18 and 19) was the presence of chlorogenic acid (CA) at R_F 0.07. Its intensity varied among samples, being most pronounced in Fraxinus ornus (5) and Viburnum lantana (18). Fraxinus ornus is known for CA-rich leaves and bark, and our study provides the first evidence that its twigs are also a significant source [17].

Additionally, extracts 1, 2, 8–10, 12, 13, 15–16, and 18 shared a light-green band at R_F 0.57, while extract 5 showed a bright-green band at R_F 0.56, likely chemically related.

As anticipated, extracts belonging to the same botanical families exhibited pronounced chemical similarities, although subtle variations in the low- R_F region (<0.30) provided distinctive species-specific signatures. Within the Sapindaceae family (extracts 1–3), the profiles were largely comparable; however, extract 3 stood out due to intense bands at R_F 0.12 and 0.23, while extract 2 displayed the most prominent zones at R_F 0.44 and 0.71, along with a unique orange band at R_F 0.09 that was absent in the other two extracts. Similarly, Malvaceae extracts (9–10) showed generally consistent profiles, yet extract 9 was distinguished by a high-intensity orange phenolic band at R_F 0.18, tentatively assigned to a phenolic glycoside such as a myricetin derivative (e.g., myricitrin), which was not observed in extract 10 [38]; both extracts also contained a comparable polar phenolic compound at R_F 0.24, appearing orange after NP derivatization. In the case of Rosaceae (8, 13, 16), extracts 8 and 13 shared a higher degree of similarity, whereas extract 16 exhibited more intense zones in the lower R_F region (<0.40). Notably, in contrast to previous reports describing only two phenolic bands in Rosaceae species [39], the present results revealed a substantially richer chemical profile dominated by intense orange zones associated with flavonoid glycosides [38,40]. Conversely, extracts from the Oleaceae family (5, 14) displayed marked variability, with extract 14 being particularly distinctive due to an intense yellow band at R_F ≈ 0.75, likely corresponding to flavone aglycones such as apigenin or luteolin [40].

Specific Phenolic Markers

A considerable portion of the observed chemical diversity can be attributed to differences in the distribution of flavonoid glycosides. A prominent orange band at R_F 0.13, commonly associated with such compounds, was detected across several families (extracts 9 and 14–16). Among these, the zone corresponding to quercetin-3-O-glucoside (Q-3-O-G) was identified as a major constituent in extract 15, where it appeared with high intensity, while in extracts 12–16 it was present at lower intensity. In contrast, extract 17 exhibited a bright yellow zone at the application line, and extract 2 showed an orange band at R_F 0.09, both likely corresponding to related derivatives. Kaempferol-3-O-glucoside (K-3-O-G) was also widely distributed, being detected in extracts 6–9, 12, 14, 15, and 17, with the highest intensity observed in extract 12 (Cornus mas). Beyond these common flavonoid-related zones, several species exhibited highly specific markers. For instance, a distinctive purple phenolic band at R_F 0.71 was characteristic of extract 2 (Acer platanoides), with only trace amounts detected in extracts 1, 6, and 10. Extract 5 contained multiple blue phenolic bands (R_F 0.10, 0.18, 0.26, 0.63), likely corresponding to phenolic acids or methoxylated flavonoids, setting it apart from the other samples [40]. Additionally, extract 11 displayed a high-intensity blue zone at R_F 0.07, similar in polarity to caffeic acid but structurally distinct, while extracts 1–2 and 9–10 exhibited a blue zone at R_F 0.05, most likely corresponding to a caffeic acid isomer.

3.1.2. HPTLC-DPPH

The HPTLC–DPPH bioautographic analysis (Figure 2c) showed that radical scavenging activity (RSA) appeared to be associated with the overall complexity of the phenolic profiles (Figure 2a,b). This fact is consistent with the known radical-scavenging ability of phenolics via hydroxyl groups on aromatic rings [41]. A key observation was the distinction between “baseline” markers and those markers potentially responsible for observed RSA. While the ubiquitous phenolics (R_F > 0.60) are tentatively identified as common twig markers (Figure 2a,b), they did not seem to contribute significantly to the overall RSA. In contrast, the most pronounced RSA was concentrated in the lower-to-middle R_F region (R_F < 0.60) (Figure 2c). The variation in yellow intensity of these specific bands indicates that they may act as putative markers for the superior antioxidant performance of certain extracts.

Identification of Key Antioxidant Zones

The identification of key antioxidant zones revealed that the most significant radical scavenging (RS) activity was concentrated within specific R_F regions. In the R_F 0.55–0.60 range, light-green bands—previously observed in most extracts—demonstrated particularly strong activity. Notably, the zone at R_F 0.57 was a major contributor to antioxidant capacity in extracts 1, 2, 8–10, 13, 15, and 16. The most intense band in the HPTLC–DPPH chromatogram, appearing as a faint-green zone at R_F 0.58, was detected in extract 16 and was also prominent in extracts 6, 7, 15, 17, and 18. The pronounced activity of these bands is likely associated with favorable structural features, such as multiple hydroxyl groups or catechol moieties, which are known to enhance radical scavenging efficiency.

In the low-R_F region (<0.40), several highly active phenolic compounds were also identified. Extract 3 exhibited strong RS activity at R_F 0.12 and 0.23, while extracts 1 and 2 shared active zones at R_F 0.32 (faint-green) and 0.28 (blue). Additionally, extract 16 contained highly polar compounds at the application line that significantly contributed to its overall antioxidant profile. Common flavonoid-related zones, including those tentatively assigned to quercetin-3-O-glucoside (Q-3-O-G) and caffeic acid (CA), as well as its isomer at R_F 0.05, exhibited moderate RS capacity. However, visual assessment suggests that their overall contribution to total antioxidant activity was secondary compared to the highly active zones around R_F 0.57.

Interestingly, not all phenolic compounds detected under UV-366 nm displayed antioxidant activity in the DPPH assay. For example, bands at R_F 0.63 (extract 3) and R_F 0.75 (extract 14), despite their high intensity, showed no observable RS activity. This indicates that these compounds likely lack key structural features—such as appropriate hydroxyl group arrangements or extended conjugation—necessary for effective radical scavenging.

Overall, extracts 1–3, 15, and 16 displayed the most intense active zones, exhibited particularly high RSA and served as potential candidates for cosmetics, while extracts 11, 17 and 19 contained the least active HTPLC-DPPH profiles. A comparative analysis of the three chromatograms (Figure 2a–c) indicates that phenolic compounds appear to represent the predominant class of antioxidants in these methanolic twig extracts. Literature reports are consistent with our findings, identifying phenolics as the most abundant and bioactive constituents in woody twig samples [5,6].

To explore their cosmetic application, the 19 extracts were further screened for radical scavenging activity, skin-aging enzyme inhibition (elastase and tyrosinase) and antimicrobial efficacy. This comprehensive profiling links plausible identified phenolic markers with essential dermo-cosmetic functions.

3.2. Spectrophotometric Assays

3.2.1. RSA Assay

The RS capacity of the tested twig extracts ranged from 54.9 ± 1.1 μmol TE/mL (17) to 108.8 ± 6.6 μmol TE/mL (8), indicating statistically significant variability among samples (

Table 1. TPC and RSA assay results of investigated twig extracts (1–19), expressed as mean values of triplicate measurements ± standard deviation.

Extract No.	TPC (mg GAE/mL)	DPPH (µmol/mL TE)
1	8.2 ± 0.4 abc	94.2 ± 5.2 b
2	5.6 ± 0.1 def	91.8 ± 4.1 b
3	5.6 ± 0.8 def	92.9 ± 4.8 b
4	2.9 ± 0.2 g	58.1 ± 2.5 e
5	5.0 ± 0.1 de	62.4 ± 2.6 d
6	5.6 ± 0.2 def	69.5 ± 6.9 d
7	8.1 ± 0.1 abc	91.5 ± 2.1 b
8	10.1 ± 0.6 a	108.8 ± 6.6 a
9	3.5 ± 0.8 fg	94.8 ± 6.7 d
10	7.3 ± 0.5 bc	57.2 ± 0.2 b
11	1.0 ± 0.2 h	93.4 ± 8.1 e
12	5.7 ± 0.7 def	88.3 ± 3.3 b
13	7.4 ± 0.5 bc	85.1 ± 1.1 bc
14	3.2 ± 0.2 g	85.1 ± 1.1 c
15	12.4 ± 0.5 a	106.6 ± 1.6 a
16	11.7 ± 0.8 a	83.7 ± 0.8 c
17	1.2 ± 0.1 h	54.9 ± 1.1 e
18	2.6 ± 0.4 g	88.6 ± 4.4 bc
19	8.1 ± 0.1 abc	87.0 ± 3.8 bc

^a–h Different letters above the values indicate groups that are significantly different according to Tukey’s multiple comparison test (p < 0.05); TPC—Total Phenolic Content; RSA—Radical Scavenging Activity; GAE—Gallic Acid Equivalents; TE—Trolox Equivalents.

; p < 0.0001).

Most extracts exhibited comparable RS capacities and formed a broad intermediate-activity group, with values such as 94.2 ± 5.2, 91.8 ± 4.1, and 92.9 ± 4.8 μmol TE/mL for extracts 1, 2, and 3, respectively, and 83.7 ± 0.8—88.6 ± 4.4 μmol TE/mL for extracts 13, 14, 16, 18, and 19, which is consistent with the generally uniform intensity and spatial distribution of active zones observed on the HPTLC–DPPH profile. In contrast, several extracts showed markedly lower RS capacities, including extracts 4 (58.1 ± 2.5 μmol TE/mL), 5 (62.4 ± 2.6 μmol TE/mL), 6 (69.5 ± 6.9 μmol TE/mL), 10 (57.2 ± 0.2 μmol TE/mL), and 17 (54.9 ± 1.1 μmol TE/mL), and largely accounted for the observed statistical differentiation. Post hoc analysis allowed classification of the extracts into three activity groups, with the highest RS capacities observed for extracts 8 (108.8 ± 6.6 μmol TE/mL) and 15 (106.6 ± 1.6 μmol TE/mL), a moderately active group comprising the majority of samples (approximately 83–95 μmol TE/mL), and a low-activity group including extracts with RS capacities below 70 μmol TE/mL (4, 5, 6, 10, and 17).

Importantly, although CA, Q-3-O-G, and K-3-O-G, all previously reported in the literature as phenolics with confirmed RS capacity [8,42,43], were shown to contribute to the overall RS potential of the extracts, the exceptionally high RS activity of extract 8 cannot be attributed primarily to the identified phenolics. Instead, this pronounced activity is most likely driven by other constituents presented in Figure 2a–c. This observation strongly suggests the presence of highly active, taxon-specific phenolic compounds characteristic of twig matrices. The pronounced activity of extract 15 aligns with its HPTLC–DPPH profile, showing the most active zones (R_F 0.57/0.58), multiple active bands, as well as several unresolved active compounds at the application line, which together contribute to its overall activity (Figure 2c).

Consistent with this interpretation, the most active extracts were characterized by the presence of distinct DPPH-inhibiting zones at specific R_F values, namely R_F 0.57 (extracts 1, 2, 8, 13, 15, 16), R_F 0.58 (7, 16), R_F 0.44 (1, 2, 3, 7, 12, 15, 16), R_F 0.32 (1, 2, 8, 10, 13), and R_F 0.28 (1, 2, 8, 10, 13). These zones were either absent or markedly less intense in the low-activity extracts, indicating that they are associated with differences between and weakly active samples. Taken together, the combined spectrophotometric and HPTLC–DPPH results provide evidence that, while common phenolics contribute to baseline activity, the differentiation of RSA among the twig extracts is predominantly governed by a subset of specific phenolic compounds presented in Figure 2a–c. These constituents represent priority targets for further structural elucidation, as they appear to be associated with the superior antioxidant performance observed in selected extracts. Notably, extracts 8 and 15 showed the highest RSA in both methods, highlighting their potential as anti-aging candidates for cosmetic applications.

3.2.2. TPC Assay

TPC of the analyzed twig extracts varied from 1.0 ± 0.2 to 12.4 ± 0.5 mg GAE/mL, reflecting notable differences in the abundance of phenolic compounds among the samples (Table 1). Based on Tukey’s multiple comparison test, extracts 15 (12.4 ± 0.5 mg GAE/mL) and 16 (11.7 ± 0.8 mg GAE/mL) exhibited the highest TPC, which was statistically significantly greater than that of all other extracts. In contrast, samples 11 (1.0 ± 0.2 mg GAE/mL) and 17 (1.2 ± 0.1 mg GAE/mL) showed the lowest TPC. The remaining extracts displayed intermediate TPC values, with significant differences observed among several of these intermediate samples, illustrating the heterogeneity of TP abundance across the twig extracts. The overall TPC trend is in accordance with the phenolic profiles of the investigated twig extracts (Figure 2a,b).

A moderate to strong and statistically significant positive correlation was observed between TPC and RSA (Pearson’s r = 0.733, p < 0.001), with R² = 0.54, demonstrating that over half of the variability in RSA is explained by TPC, while also suggesting contributions from other non-phenolic metabolites. Extracts with the highest TPC (8, 15) exhibited strong RSA, those with moderate TPC (1, 2, 3, 6, 7, 10) showed intermediate activity, and the lowest TPC samples (4, 9, 11, 17) had the weakest RSA. Notably, extract 16 represents an outlier, reflecting a higher TPC relative to its RSA. This suggests that while the extract is rich in phenolics, a substantial portion of these compounds may possess only moderate-to-low RSA. This finding highlights that the overall antioxidant capacity is not determined solely by TPC but also by the specific composition, molecular structures, and potential synergistic interactions with other bioactive constituents, emphasizing the need for further systematic investigation.

The available literature reports that RSA measured in the twig extracts used in this study, such as those of Berberis vulgaris (11) and Rosa canina (16), as well as in Quercus spp., was closely associated with TP content [5,44,45], demonstrating that phenolic compounds are the main contributors to RS activity in these plant materials, consistent with our findings.

3.2.3. Skin Aging-Related Enzyme Assays

A thorough review of the available literature indicates that this study is the first to demonstrate the inhibitory potential of nearly all of the investigated plant species, highlighting their previously unexplored relevance as sources of skin-related enzyme inhibitors. These findings are significant for advancing the understanding of plant-derived compounds in the prevention of aesthetic skin disorders, such as melasma, hyperpigmentation, wrinkling, and solar lentigo, which are closely associated with tyrosinase activity [9]. Moreover, the observed inhibitory effects against enzymes involved in extracellular matrix degradation, such as elastase—responsible for collagen and elastin breakdown—suggest that these plant extracts may contribute to the preservation of skin structural integrity, maintenance of a youthful appearance, and prevention of skin aging [9].

Tyrosinase Inhibition (TI) Assay

Among the tested samples in the TI assay, extract 8 exhibited the strongest TI activity (39.0 ± 3.5%), indicating a moderate inhibitory potential (Figure 3a). This extract was statistically significantly different from the majority of the other samples (p < 0.05) and demonstrated approximately half of the inhibitory activity of the standard inhibitor (KA, 82.0 ± 2.0%). A review of the available literature revealed that only one study to date has reported TI activity for Torminalis glaberrima (8), but exclusively for the conventional extracts of its leaves, which were shown to be effective inhibitors [21]. This study provides the first evidence of moderate TI activity for Torminalis glaberrima (Sorbus torminalis) twigs, at low concentrations, indicating that the twigs of this species are a valuable source of tyrosinase inhibitors and warrant further examination.

Extracts 1 (28.6 ± 3.1%), 13 (29.2 ± 3.2%), and 15 (27.9 ± 6.2%) showed slightly lower TI than extract 8, with inhibition values ~2.8-fold lower than KA, forming a uniform statistical cluster distinct from the remaining samples. Extracts 1, 8, and 13 shared a phenolic zone at R_F 0.28 (Figure 2a,b), which may contribute to TI, although further investigation is needed. Extracts 16 (24.9 ± 2.0%) and 19 (25.4 ± 5.0%) were also statistically similar. Related phenolics at R_F 0.57/0.58 in extracts 1, 8, 13, 15, and 16 may contribute to TI and warrant further study. Chlorogenic acid (CA) was observed in extracts 8 and 19 and may contribute to TI, as in silico studies show binding to key active-site residues (ARG321, ARG374), and metabolic products like quinic acid also contribute [46,47]. In B16 melanoma cells, CA reduces TI activity and melanin production, supporting its anti-hyperpigmentation potential [48].

Previous studies have investigated different plant tissues within the Acer genus, indicating that bioactivity may vary depending on the plant part and extraction conditions [49,50]. However, this study provides the first evidence that Acer campestre (1) twigs are a promising source of tyrosinase inhibitors. A moderate positive correlation was observed between TPC and TI (r = 0.64, R² = 0.41), indicating that phenolics significantly contribute to anti-tyrosinase activity. Samples with high TPC (1, 8, 13, 15, 16, 19) showed the strongest TI effects, suggesting a direct relationship, although qualitative composition, structural features, and synergistic interactions also play roles.

Current literature reports TI activity of twig extracts for three Crataegus taxa, whereas no data is available for Crataegus monogyna (13). The referenced study indicates that twig extracts exhibited higher TI activity compared to leaves, emphasizing the relevance of twigs as a valuable material [51]. For Paliuris spina-cristi (15), literature reports investigations of the root, leaf, and fruit, whereas studies focusing on twig extracts are lacking [28]. In the case of Rosa canina (16), bioactivity has been predominantly examined in the fruit, while twigs appear to remain unexplored [52]. Furthermore, no literature data were found regarding the TI of Pyrus pyraster (19) for any plant part; in our study, the plant twigs showed moderate TI, warranting further investigations.

For Quercus cerris (6), the available literature includes one study reporting potent TI activity, attributed to its bark [53]. However, no data are currently available regarding the TI potential of its twigs, making the present study the first to provide evidence of TI associated with twig extract (17.4 ± 1.1%). The taxonomically related species Quercus pubescens (7) exhibited a comparable TI value (18.8 ± 3.5%) in our study, whereas the available literature reports no TI activity for the crude wood extract of this species [54].

In contrast, extracts 3 (11.0 ± 1.3%), 5 (9.6 ± 0.1%), 9 (7.7 ± 2.1%), 10 (9.5 ± 0.5%), 11 (9.2 ± 2.9%), 12 (13.5 ± 2.2%), 14 (10.4 ± 0.9%), and 17 (9.4 ± 1.4%) exhibited low TI activity, indicating a limited ability to interfere with the enzymatic activity of tyrosinase. For the tested species Fraxinus ornus (5), literature data are available regarding the TI activity of its bark exudate; however, the reported activity is negligible even at high concentrations. Our observation suggests that the twigs of Fraxinus ornus exhibit a higher TI inhibitory potential than the exudate [55].

Reported data indicate that TI activity has been documented only for extracts of the aerial parts of Tilia cordata (10; IC₅₀ = 0.67 ± 0.04 mg/mL), corresponding to a moderate TI potency [56]. Our results demonstrate that the twig extract of this plant exhibits low TI potential. Additionally, while no previous data are available for twigs of Tilia argentea (9), the literature reports the presence of tiliroside, a potent tyrosinase inhibitor, in the leaves of this species, supporting the relevance of Tilia argentea as a valuable source of bioactive compounds [57].

For Berberis vulgaris, moderate to low TI activity has been reported for bark and root extracts, and the present investigation of twigs corroborates these findings, suggesting a comparable TI potential [58].

For Cornus mas (12), available evidence indicates that iridoids isolated from the fruits exhibit stronger TI activity than the standard inhibitor [59]. However, no data have been reported to date regarding the TI potential of twig extracts of this species.

Finally, extracts 2 (0.6 ± 0.1%) and 4 (4.1 ± 0.5%) showed minimal TI effects and were statistically similar, suggesting that they can be considered non-inhibitory toward tyrosinase.

Elastase Inhibition (EI) Assay

Among the tested extracts, Acer campestre (1) exhibited the highest EI activity (51.6 ± 9.1%), although still statistically different from the standard inhibitor EGCG (79.1 ± 4.7%) (Figure 3b). Statistically, EGCG showed significantly stronger EI compared to all plant extracts (p < 0.05). While extracts 1–3 (Acer species) shared a broadly similar phenolic fingerprint, Acer campestre (1) exhibited elastase inhibitory activity (51.6 ± 9.1%) that far exceeded its congeners (0%). Although extract 1 showed slightly more intense zones at R_F 0.10, 0.28, and 0.57 (Figure 2a,b), this discrepancy might indicate that the superior potency of extract 1 may not solely arise from its phenolics visible under NP derivatization. Members of the Acer genus are known to accumulate diverse non-phenolic metabolites, including triterpenoids (such as friedelin and beta-sitosterol) and cyclitol derivatives, which have previously demonstrated enzyme inhibitory potential [60]. These compounds could contribute to the observed elastase inhibition through synergistic effects with the identified phenolic markers at R_F 0.10, 0.28, and 0.57. The fact that the zone at R_F 0.28 is also present in other elastase-inhibiting extracts (8 and 13) supports its role as a putative EI marker, but in the case of A. campestre, its activity is likely amplified by other compound classes that remain undetected by the current HPTLC-NP protocol.

Moderate EI was observed for T. glaberrima (8; 34.5 ± 1.6%) and Cornus mas (12; 29.5 ± 5.5%), which were statistically distinguishable from other extracts. No data are available on the elastase inhibitory activity of T. glaberrima; however, leaves of S. alnifolia from the same genus exhibited 87% inhibition at a concentration of 1 mg/mL [61]. Our findings further indicate that twigs also represent a valuable source of elastase inhibitors, highlighting the need for their detailed investigation.

No direct correlation was observed between EI and TPC values, as the highest activity was recorded for extract 1 despite its moderate phenolic levels, while other phenolic-rich samples remained inactive. However, in the case of extract 8, which emerged as the second-best inhibitor, the pronounced activity could indeed be attributed to its high TPC, suggesting that EI is governed by a complex interplay between specific bioactive markers (notably at R_F 0.10, 0.28, and 0.57) and overall phenolic density.

Previous studies have assessed the activity of Cornus mas fruit extracts suitable for cosmetic applications and reported that the fruit contains significant amounts of elastase-inhibiting compounds [62]. To the best of our knowledge, this study provides the first evaluation of the EI activity in the twig extract of Cornus mas and found them to be moderately potent elastase inhibitors, making them promising candidates for further investigation.

Lower inhibitory effects were detected for Pyrus pyraster (19; 16.4 ± 5.1%), R. canina (16; 14.1 ± 4.6%), Paliurus spina-cristi (19; 13.3 ± 6.7%), Carpinus betulus (4; 13.0 ± 3.4%), Crataegus monogyna (13; 12.6 ± 4.0%), and Tilia cordata (10; 10.6 ± 0.9%). These extracts generally did not differ significantly from each other, suggesting weak EI. Again, almost all detected elastase inhibitors (1, 8, 12, 13, 15, 16, 19) have potentially inhibitory activity at R_F 0.57/0.58 and warrant further investigation.

Several extracts, including those from Acer platanoides (2), Acer tataricum (3), Quercus cerris (6), Quercus pubescens (7), Tilia argentea (9), Berberis vulgaris (11), Lingustrum vulgare (14), and Ruscus aculeatus (17), showed no detectable EI activity under the applied experimental conditions. Data are available indicating that Quercus pubescens (its wood) contains a low concentration of elastase inhibitors, but data for the twig are missing [54].

Overall, only a few tested twig extracts showed notable EI effects, with A. campestre (1) being the most active, demonstrating strong anti-wrinkle potential at low concentrations. T. glaberrima (8) also showed promise, exhibiting both anti-pigmentation and anti-wrinkle activity, suggesting suitability for multifunctional cosmetics. Phenolic zones detected at R_F 0.28, 0.57, and 0.58 merit further investigation, as their correspondence with bioactivity suggests the presence of compounds with potential dual inhibitory effects.

3.3. Agar-Well Diffusion Assay

The antibacterial activity of the investigated twig extracts was assessed using the agar-well diffusion method, which allows clear visualization of both bactericidal effects and inhibition zones, allowing qualitative comparison of the in vitro antimicrobial potential of the extracts. Both Staphylococcus aureus and Pseudomonas aeruginosa are clinically significant pathogens associated with skin infections. S. aureus can rapidly develop antibiotic resistance, including MRSA, while P. aeruginosa is highly virulent, resistant to multiple antibiotics, and capable of forming biofilms [35,36], highlighting the importance of evaluating novel antibacterial agents against these strains. As shown in

Table 2. Antibacterial activity of twig extracts evaluated using the agar-well diffusion assay, expressed as mean inhibition zone diameters (mm) ± standard deviation from triplicate measurements.

Extract No.	S. aureus	P. aeruginosa
1	12.5 ± 0.5 *	/
2	16.5 ± 0.5	18.5 ± 0.5
3	/	/
4	18.5 ± 0.5	18.0 ± 1.0
5	/	/
6	14.5 ± 0.5 *	11.5 ± 0.5 *
7	14.0 ± 1.0 *	13.0 ± 1.0 *
8	14.5 ± 0.5 *	12.0 ± 1.0 *
9	/	11.0 ± 1.0 *
10	7.5 ± 0.5 *	12.0 ± 1.0 *
11	/	/
12	12.0 ± 1.0 *	11.0 ± 1.0 *
13	10.5 ± 0.5 *	10.0 ± 1.0 *
14	/	/
15	14.0 ± 1.0 *	13.0 ± 1.0 *
16	13.0 ± 1.0 *	12.0 ± 1.0 *
17	/	/
18	10.0 ± 1.0 *	/
19	13.5 ± 0.5 *	12.5 ± 0.5 *
MeOH	/	/
Streptomycin	17.5 ± 0.5 * 27.5 ± 0.5	15.5 ± 0.5 * 34.5 ± 0.5

* Asterisk indicates bactericidal activity, whereas inhibitory activity is shown without a symbol; MeOH—methanol.

, the majority of the extracts exhibited detectable antibacterial activity, generally stronger against S. aureus than against P. aeruginosa. This higher susceptibility of Gram-positive bacteria is consistent with previous reports and can be explained by differences in cell-wall architecture: Gram-positive bacteria possess a thick peptidoglycan layer but lack the outer membrane characteristic of Gram-negative bacteria, which limits penetration of antimicrobial compounds [63,64].

Among the tested samples, extracts 2, 4, 7, 8, 15, 16, and 19 exhibited the most pronounced antibacterial activity against both bacterial strains. Extract 2 was notably potent, surpassing streptomycin (17.5 ± 0.5 mm for S. aureus; 15.5 ± 0.5 mm for P. aeruginosa) with inhibition zones of 16.5 ± 0.5 mm and 18.5 ± 0.5 mm, respectively, indicating broad-spectrum potential. Extract 4 exhibited the highest activity against S. aureus (18.5 ± 0.5), while it ranked second in activity against P. aeruginosa (18.0 ± 0.5), immediately following extract 2, surpassing streptomycin. Several other extracts (6, 12, and 13) displayed moderate but consistent activity, while some extracts exhibited strain-dependent effects (e.g., 1 and 18 active (only against S. aureus). In contrast, extracts 3, 5, 11, 14, and 17 showed no detectable antibacterial activity.

Previous studies reported strong antibacterial activity for certain tissues of some species tested here: Quercus cerris cork (6) [65], Quercus pubescens bark (7) [66], and T. glaberrima inflorescences and leaves (8) [21,67]. The berries of Ligustrum vulgare (14) and extracts of Viburnum lantana (18) have also shown activity against one or both strains [68,69], while fruits and leaves of Paliurus spina-christi (15) have documented antibacterial effects [70]. Although literature lacks data specifically on twigs, our results indicate that twig extracts can be valuable, previously unexplored antimicrobial sources. Several moderately active extracts (1,8,15) also exhibited relatively high TPC values, suggesting a possible role of phenolic compounds in antibacterial activity. However, this trend was not observed for extracts 2 and 4, which showed pronounced effects. As previously noted, extract 2 contained the most intense zones at R_F 0.71 and 0.44, and a likely zone corresponding to flavonoid glycoside, at R_F 0.09 (Figure 2a,b). These compounds merit further investigation, as they are absent or present only in trace amounts in extract 3, which showed no antibacterial potential despite being from the same genus (Acer). This discrepancy indicates that antibacterial activity is likely driven by specific phenolic subclasses with high intrinsic potency or by non-phenolic constituents, as well as possible synergistic interactions between different compound classes. When antibacterial activity is considered together with the chemical profiles and previously demonstrated bioactivities, extracts 2 and 4 emerge as particularly promising candidates for cosmetic applications.

3.4. PCA

To further explore subtle differences among twig extracts and to objectively identify chromatographic zones associated with sample discrimination, PCA was applied to the NP- (red, green, blue channels) and DPPH (8-bit)- derivatized HPTLC chromatograms (Figure 2b,c). The red-channel PCA model exhibited poor discriminatory power and failed to achieve meaningful cluster resolution; consequently, due to its inadequate diagnostic performance, these results will not be discussed.

3.4.1. NP-Green Channel Model

The PCA model derived from the NP-green-channel (Figure 4a) effectively captured the major sources of phenolic variability. The first two principal components (PC1: 31.22% and PC2: 18.05%) accounted for 49.27% of the total variance, while the inclusion of PC3 (12.26%) and PC4 (8.58%) brought the cumulative explained variance to 70.11%. This high cumulative variance indicates the model’s reliability in reflecting the chemical diversity of the twig extracts.

When examining the separation along the PC1 axis (Figure 4a), a clear distinction emerges between strong tyrosinase inhibitors (6, 7, 8, 13, 15, 18, and 19; positive PC1 scores) and weaker inhibitors (2, 5, 9, 10, 11, and 14; negative PC1 scores). This differentiation is primarily associated with specific chromatographic zones: the zone at R_F 0.82 contributes positively to the separation of inhibitors, while zones at R_F 0.06, 0.12, 0.18, and 0.23 exert a negative influence, corresponding to zones that probably are not associated with TI in the model (Figure 5a).

3.4.2. NPR-Blue Channel Model

The PCA model derived from the NP-blue-channel intensities (PC1: 34.42%-; PC2: 17.36%-; PC3: 12.52%-; PC4: 8.49%-; PC5: 6.86% of variability), accounting for a total cumulative variance of 79.65%, did not yield significant diagnostic information for the majority of the dataset; however, it effectively grouped a specific cluster consisting of samples 6, 8, 13, and 18 which showed pronounced TI (Figure S1a), associated with characteristic twig-bands at R_F 0.82 and R_F 0.23, 0.28, and 0.57 (Figure S1b,c). Notably, an identical clustering trend for this particular subgroup was also observed in the green-channel model (Figure 4a), further supporting the association of these chemical zones with observed biological activity.

3.4.3. HPTLC-DPPH Model

The PCA model derived from HPTLC-DPPH chromatograms (PC1: 37.14%-; PC2: 13.18% of variability) successfully grouped the samples in accordance with the previously discussed visual profiling and the results of the RSA assay (Figure 4b). Extracts characterized with high and moderate RS abilities were clearly separated along the positive side of the PC1 axis, which appears to capture the variance associated with RS potency. The loading plots indicated that the compound at R_F 0.57 was strongly associated with this discrimination, which aligns with our bioautography findings (Figure 5c). The relatively low peak intensity observed in the graph of the scores plot explains the looser clustering within this group, indicating that, besides the main zone, other sample-specific zones also contribute to their individual RS capabilities. Furthermore, the most potent group of extracts (1, 2, 3, 15, and 16) was located within the quadrant defined by positive PC1 and negative PC2 values (Figure 4b). The separation of this potent group was primarily driven by variables at R_F 0.13, 0.18, 0.24, and 0.57, highlighting these zones as important markers for high RS capacity (Figure 5c,d), warranting further isolation and characterization. While extract 8 (Malvaceae) exhibited high RS capacity and aligned with this group along PC1, its slightly distinct spatial positioning on the PC2 axis reflects its unique phytochemical fingerprint compared to the Acer and Rosaceae/Rhamnaceae clusters (Figure 4b). This indicates that, besides the main shared marker (R_F 0.57), individual RS capabilities are also influenced by sample-specific secondary metabolites, which account for the looser clustering observed within the high-potency region.

However, the observed overlaps in the PCA models reflect the inherent complexity of biological activities, which often involve non-linear synergistic interactions rather than simple linear correlations with major phenolic peaks. While the unsupervised nature of PCA effectively captured the primary chemical variance, the slight divergence from bioactivity trends highlights the influence of minor bioactive constituents or matrix effects. Obtained associations provide initial insight into the compounds underlying the observed biological activity, with further confirmation achievable by complementary techniques such as LC–MS or isolation.

This integrated analytical platform—combining HPTLC fingerprinting, bioautography, and chemometric validation—provides a robust framework for the valorization of urban forest residues. By correlating complex chemical patterns with specific biological responses, this methodology enables the functional standardization of inherently heterogeneous waste materials. Such a multidisciplinary approach establishes a reproducible paradigm for transforming diverse urban green-waste streams into high-quality, standardized raw materials for the cosmetic and pharmaceutical industries.

4. Conclusions

This study demonstrates that twig biomass, commonly treated as low-value waste from urban forestry maintenance, represents an underexplored bioresource with potential relevance for the cosmetic industry. The valorization of pruning residues supports the concept of non-timber forest products, enabling the recovery of high-value bioactive compounds without additional harvesting pressure and linking urban forest management with sustainable product development. A multidisciplinary analytical approach integrating HPTLC profiling, spectrophotometric assays, and bioautography revealed that the biological activity of twig extracts is driven primarily by phytochemical composition rather than floristic origin. While ubiquitous phenolics such as chlorogenic acid contribute to baseline antioxidant effects, the superior performance of selected extracts was associated with distinct chromatographic features. Key zones at R_F 0.28, 0.57, and 0.58 consistently correlated with biological activity, identifying them as relevant contributors to the observed in vitro biological activities. Among the investigated samples, the extracts of Torminalis glaberrima (Gand.) Sennikov & Kurtto (8) and Paliurus spina-christi (L.) Burgsd. (15) exhibited the highest radical-scavenging activity, with extract 8 also showing pronounced tyrosinase inhibition (TI), indicating strong potential for anti-pigmentation applications. In contrast, Acer campestre L. (1) demonstrated the most potent elastase-inhibitory effect, suggesting notable relevance for anti-wrinkle activity. Antibacterial activity showed a different pattern, being driven by specific constituents rather than by total phenolic content. Acer platanoides L. (2) and Carpinus betulus L. (4) displayed the highest activity against P. aeruginosa and S. aureus, respectively. Overall, these findings identify selected urban forest residues as high-value substrates for further phytochemical isolation, mechanistic studies, and further biological validation relevant to formulation development. The chromatographic zones at R_F 0.10, 0.28, 0.57, and 0.58 should be prioritized in future structure–activity investigations. This work provides a scientific basis for the valorization of urban green waste as a bioactive source with potential relevance for sustainable cosmetics development, supporting circular bioeconomy principles and reinforcing the multifunctional role of urban forests. Importantly, these findings provide a basis for cosmetic relevance, with further validation through cytotoxicity, cell-based, and dermatological (skin irritation) studies supporting their potential application. Future research should focus on the isolation and structural characterization of the identified bioactive compounds, as well as on establishing their safe and effective concentration ranges in cosmetically relevant systems, which is essential for their translation into practical cosmetic applications.