Epilobium hirsutum L. (Onagraceae) is a perennial herb, usually 1.8–2 m tall, occasionally up to 2.5 m. The stem is strong and densely covered with soft, spreading hairs. Leaves sessile, clasping stem, lanceolate-elliptic to narrowly obovate or elliptic, hairy on both sides. Flowers large, 10–20 mm in diameter, with four pink or purple-pink petals. Stigmas white, with four lobes. Fruit a capsule, 3–9 cm long, usually hairy. Seeds dark brown, 0.8–1.2 mm long. This species is native to Europe, Africa and much of Asia, except most of Siberia. Naturalised in North America. Usually grows in a wide range of open, wet and damp, occasionally mesic habitats up to 2,500 m above sea level.

Three large stands of E. hirsutum, covering at least 50 m² and occurring in habitats of varying water availability, were selected for plant sampling in this study. All stands of the study species were selected in the vicinity of Wallington, London, United Kingdom, to eliminate the effect of meteorological conditions that may occur at distant sites (Table 1). The average temperature (°C), sum of precipitation (mm) and sunshine duration (s) of the meteorological conditions during the whole study period (from the beginning of April to the end of October 2023) that could have influenced the accumulation of metabolites in E. hirsutum were obtained from Sutton, Weather Station from Open-Meteo.com Weather API (Zippenfenig, 2023). Data are given in Supplementary Table S1.

The chosen approach of collecting raw materials from natural habitats allowed for indirect control of plants’ growth conditions through the careful selection of sites and systematic recording of environmental and botanical data.

It also ensured the traceability of plant materials (which is not always possible from commercial suppliers) and prevented potential contamination or adulteration. It ensured that the collected plant material accurately reflected natural vegetative and phytochemical changes.

At the Grange Gardens site, the E. hirsutum stand occupied mesic fringe grassland surrounded by sparse Quercus robur woodland. As a result, the grassland was partially shaded by trees in the morning and evening. The soil at this site was moderately moist, with no standing water on the surface at any time of the year. At the Green Hackbridge site, the study species grew in a wet grassland habitat close to the river. They are surrounded by a stand of Alnus glutinosa, which provides light shade to E. hirsutum in the morning and evening. The soil at this site was waterlogged in spring, damp after heavy rainfall and wet during the dry months. At the Manor Garden site, the stand of E. hirsutum occupied the completely open shore of a eutrophic artificial lake. The roots of the sampled plants were in the waterlogged or damp soil of the lake shore throughout the growing season. Herb cover was similar at all study sites, ranging from 60% to 70% of the surface (Table 1), and the study species represented approximately 30% of all herbs at all sites.

To assess the dynamics and variability of the chemical profile of E. hirsutum, plants were sampled at regularly 2-week intervals from the beginning of the growing season at the end of April (28 April 2023) until the onset of wilt at the beginning of October (3 October 2023). Five well-defined phenological phases of E. hirsutum were defined: intensive growth, flower formation, flowering, seed dispersal and wilting. The flowering phase was further subdivided into three subphases: beginning of flowering, intensive flowering and end of flowering. Each phenological phase was defined by observable morphological traits (e.g., shoot height, leaf size, presence of flower, seed set) and occurred over a consistent period of time across sites, with only ±3–5 days variation depending on microhabitat conditions (Figure 2). All samples were collected on the same day. For each phenophase, 3–4 individual plants were sampled at each site, ensuring a minimum of three biological replicates per group. All sampled individuals at a given site were synchronized in phenophase to reduce developmental variability, and only individuals at the same stage were included in each sampling. A total of 78 samples of leaf and stem were collected at 13 sampling events, representing all three habitats and all defined phenophases with sufficient biological replication (minimum of three replicates per phenophase per site in almost all cases).

The plants grew rapidly during the intensive growth phase, which lasted from the end of April to the end of May. At the end of April, the shoots were 10–15 cm high, and a month later, they had grown to 80–100 cm but had not yet started to develop flowers (Figure 1). The upper part of the shoot began to branch. The first flowers began to form in early June, and this phase continued until mid-June (Figure 2). In the second half of June, the plants started to flower, and 2 weeks later, they reached a phase of intensive flowering that lasted until the second half of July. By early August, the flowering of E. hirsutum plants had almost ceased, with only single flowers remaining on the main and secondary branches. By early September, all flowers on the main branches had produced fruits, and some of the fruits were already ripe and starting to release seeds (Figures 1, 2). By early October, most of the seeds had been released, and the leaves on the upper part of the plant had begun to turn yellow. A total of 78 Epilobium hirsutism samples were collected during the 2023 growing season and used for analysis. Plants were identified by Dr O. Mykhailenko, and voucher specimens (OM1–OM78) were deposited in the herbarium of the School of Pharmacy, University College London.

Plants in wild populations were randomly sampled on dry, windless days. The upper part (approximately two-thirds of the plant) was cut off and labelled, indicating on the label the stage of development, the location and the sampling date. In the laboratory, leaves with flowers were separated from the stems and the collected raw material was dried at an ambient temperature of 20°C–24°C in a well-ventilated room. Thus, each analytical group, representing habitat type and developmental stage, consisted of two samples: leaves (including flowers) and stems. Plant samples were crushed immediately before chemical analysis.

Gallic acid, ellagic acid, chlorogenic acid, caffeic acid, oenothein B, hyperoside, isoquercitrin (syn. isoquercetin), avicularin, guaijaverin, myricetin were obtained from Inc. ChromaDex (Santa Ana, United States) and Sigma-Aldrich (Saint Louis, United States). All solvents (acetonitrile, methanol, glacial acetic acid) were of HPLC grade and were purchased from Merck KGaA, Fisher Scientific Ltd. And VWR International LLC, except of deuterated methanol, which was purchased from Cambridge Isotope Laboratories Inc. Analytical- and chromatographic-grade chemicals and solvents were used for this study: acetonitrile, methanol, glacial acetic acid from Sigma-Aldrich (GmbH, Karlsruhe, Germany). The water was purified using the ULTRAPURE water system (Millipore, Germany). All other chemicals were of analytical grade.

For the HPTLC analysis, air-dried powdered plant material (0.50 g) of each sampling date sample was extracted with 50% (v/v) methanol (1:10) in an ultrasound-assisted extraction for 20 min at 40 °C. The sample preparation for HPLC analysis was the following: a precise weight (0.20 g) of each sample powder was extracted with 50% (v/v) methanol (1:50) in an ultrasonic bath (WiseClean) at 45°C ± 2 °C for 20 min. Extraction yields ranged from 7.4% to 6.1% for leaves and from 5.3% to 3.4% for stems extracts. All extracts for both analyses were filtered using a Millex R Syringe filter unit 0.45 mm. The references (avicularin, guajaverin, hyperoside, oenothein B, isoquercitrin, gallic acid, ellagic acid, chlorogenic acid, myricetin) were dissolved in methanol, with a concentration of 1 mg/μL, then sonicated for 10 min. Both the reference standards and the test samples were stored at 4°C.

The samples and reference standards were injected individually using a CAMAG Linomat 5 semi-automated sampler (100 μl syringe). The injection volumes of 3 μL reference standards and 7 μL sample solutions were applied. HPTLC glass plates Silica gel 60 F₂₅₄ used for the stationary phase were purchased from Merck KGaA (Darmstadt, Germany). Samples and reference standards were applied to the plates as bands 8 mm wide. The space between bands was 2.0 mm, and the application rate was 90 nL·s⁻¹.

Plates were developed in the ADC 2 with chamber saturation (with filter paper) for 20 min and after activation at 33% relative humidity for 10 min using a saturated solution of magnesium chloride, development with ethyl acetate–formic acid–water 68:8:8 (v/v) to the migration distance of 70 mm (from the lower edge), followed by drying for 5 min. The temperature and relative humidity were controlled to 21°C–24°C and 33%, respectively. The development distance was 70.0 mm from the lower edge.

A total of 78 methanolic 50% (v/v) extracts (39 leaf and 39 stem samples) were applied across 8 HPTLC plates, with each plate accommodating up to 15 lanes. Each lane represented a different extract corresponding to a specific plant part (leaf or stem), habitat (mesic grassland, wet grassland, lake shore), and phenological stage (from intensive growth to wilting). Reference standards were applied in duplicate to each plate for comparison.

The developed plates were derivatised by dipping in Natural Product A reagent (2 g of 2-aminoetheyl diphenylborinate in 200 mL of methanol) followed by PEG400 (macrogol 400; 10 g polyethylene glycol 400 in 200 mL dichloromethane) using a CAMAG chromatogram immersion device and heated at 100°C on a plate heater for 5 min. The plates were visualised using CAMAG Visualizer under white light, UV 254 nm and UV 366 nm, and the data was analysed using CAMAG visionCATS 3.1 software.

Chromatographic separation of polyphenols was performed using the Waters e2695 Alliance HPLC system coupled with a 2998 PDA detector (Waters, Milford, MA, United States). Phenolic compounds were separated on an ACE Super C₁₈ (250 mm × 4.6 mm, 3 µm) column (ACT, Aberdeen, United Kingdom) with a full run time of 81 min. Column temperature was 25 °C. The gradient elution mode consisting of 0.1% (v/v) trifluoroacetic acid in pure water (A) and acetonitrile (B) was as follows: 0 min, 5% B; 8–30 min, 20% B; 30–48 min, 40% B; 48–58 min, 50% B; 58–65 min, 50% B; 65–66 min, 95% B; 66–70 min, 95% B; 70–81 min, 5% B. The flow rate was 1 mL/min, and the injection volume was 10 µL. The samples were subjected to two different analyses. The quantity of oenothein A was obtained by recalculation through oenothein B. The quantity of myricetin 3-β-D-glucopyranoside (syn. isomyricitrin) was obtained by recalculation through myricetin. Further in the text, the trivial name of the compound–isomyricitrin–will be used in order to maintain uniformity in naming all identified compounds. The analytical HPLC-PDA method for polyphenols was validated according to the ICH Q2 (R1) guidelines and described at (Ivanauskas et al., 2024). The calibration curve, limits of detection (LOD), limits of quantification (LOQ), and the linear range for each analyte are provided in Table 2; precision, repeatability, and recovery for each compound presented in Table 3; and specificity of the compounds is presented in Table 4 and was verified by comparing the retention times and UV-Vis spectra of the compounds in 78 extracts in comparison with the reference standard.

All data was processed using the LabSolutions Analysis Data System (Shimadzu Corporation). The results are mean ± standard deviation (SD) of three replicates (Microsoft Office Excel 2010, JAV). The value of p < 0.05 was taken as the significance level. The distribution of the data sets was assessed using the Shapiro-Wilk test. As all data were non-normally distributed, non-parametric statistical analysis methods were used. The significance level was set at p < 0.05. The Kruskal–Wallis test was used to detect differences between sets of samples, and the Dunn’s post hoc test was used for pairwise comparisons between samples. The effect of habitat, plant part and time of harvest of the raw material on the content of bioactive metabolites in the tissues was evaluated by two-way permutation analysis of variance (Euclidean similarity index and 9,999 permutations). Statistical analysis of the data was performed using PAST 5.0.2 software (Hammer, Harper, and Ryan, 2001).

The optimal conditions for the extraction of components and the chromatographic system for analysis were optimised. As there are no official monographs for Epilobium species in the national pharmacopoeias, we adapted the HPTLC method to assess the qualitative profile of the metabolites.

Initially, the method for extraction of the phenolic compounds in the methanolic extracts of E. hirsutum raw materials was identified using a modified pharmacopoeial Thin-Layer Chromatography (TLC) method, as specified in the European Pharmacopoeia (Ph. Eur.) 9.0 Gingo leaf monographs (Wichtl, 2002). Given the selected standards of phenolic compounds, three solvent systems were used (Table 5). The first method (ethyl acetate-formic acid-purified water (68:8:8 v/v)) showed the best separation of plant components from hydroxycinnamic acids and flavonoids’ groups, and the standards were also clearly visible on the plate. The detection of the zones was achieved using diphenylboric acid aminoethyl ether P and Macrogol as derivatising reagents. Chromatograms were observed under UV light at 254 nm and 366 nm wavelengths and white light before and after derivatisation.

To select the optimal solvent for sample extraction, we compared the results of chromatographic analysis using water, 70% ethanol, 50% methanol, and 100% methanol extracts. The colour intensity of the identified compound zones was similar across extracts, but in the water, 70% ethanol and 100% methanol extracts, zone separation was relatively poor, likely due to intense chlorophyll extraction, which interfered with separation. Additionally, extraction at 40°C yielded better metabolite release. Therefore, 50% methanol was selected as the optimal solvent for extracting phenolics from Epilobium.

At UV 366 nm, the reference standards appear at the increasing retention factors (R _f ) values, e.g., hyperoside at R _f 0,41), isoquercitrin at (R _f 0,45), guaijaverin (R _f 0,52), avicularin (R _f 0,65) (Figure 3) with distinctive characteristic colours (yellow) respectively and after derivatisation at UV 366 nm with PEG and macrogol 400, those reference compounds appear yellow bands. The remaining reference standards, such as gallic acid (R _f 0,81), caffeic acid (R _f 0,87), (R _f 0,81) and chlorogenic acid (R _f 0,38), show good visibility under UV 254 nm and UV 366 nm.

HPTLC analysis showed a difference in chemical content between the leaves and stems of the plant (Figure 4). The methanol solution of the stems was light yellow, whereas the leaf extract was deep green. The chromatographic fingerprint showed very low levels of detectable constituents in the stems and the absence of caffeic acid, avicularin, guajaverin, isoquercitrin, hyperoside. The leaves contained all the components (caffeic acid, gallic acid, chlorogenic acid, avicularin, guajaverin, isoquercitrin, hyperoside) with visible differences in band intensity and colouration depending on location and the phenological phase. For example, hyperoside and isoquercitrin appeared as bright-yellow fluorescent zones under UV 366 nm, with increased intensity during the flowering stage in mesic grassland samples. Gallic acid was more prominent in early vegetative stages, typically appearing as a bluish-green zone with moderate intensity, especially in wet grassland specimens.

It should be noted that based on the visual (i.e., qualitative) assessment of the intensity of zones of marker compounds at different stages of plant vegetation, the dynamics of accumulation, decline, and maximum component content were noted (Figure 5), which was subject to further HPLC analysis. This is especially visible on the chromatogram with white light after derivatisation.

Some challenges were encountered in accurately identifying oenothein B, which has an R _f of approximately 0.1 – a low value for exact identification. This low R _f is probably due to the large molecular size of this ellagitannin, which perhaps affects its mobility. It should also be noted that the available literature describing the analysis of Epilobium species by HPTLC does not provide visual TLC data for oenothein B. In this study, chromatograms showing the observed zone separations are published for the first time.

The dynamics of the accumulation of phenolic compounds in E. hirsutum samples was analysed at different phenological stages to determine the maximum period of their accumulation. Samples were collected from three different habitats and the environment definitely plays a key role in the accumulation of metabolites.

The applied HPLC method showed a higher separation of compounds with good peak symmetry. In our study, we used 50% methanol as the extraction solvent which provides a higher extraction yield of highly polar compounds such as phenols and glycosides and an ultrasonic bath to enhance the extraction efficacy (Uminska et al., 2025). Chromatographic separation of the extracts was carried out using a Waters e2695 Alliance HPLC system equipped with an ACE C₁₈ column. Gradient elution was applied with 0.1% trifluoroacetic acid in pure water and acetonitrile with increasing polarity from 5% to 95%. The HPLC fingerprint of the compounds is shown in Figure 6.

The developed method was fully validated according to the ICH Q2 (R1) guidelines. All standards and templates were introduced in three integrations. High reproducibility and low standard error were observed in multiple injections. All compounds showed good linearity (r ² ≥ 0.997) across the tested ranges (Table 2). The method demonstrated satisfactory repeatability and precision, with RSD values for retention time and peak area ranging from 0.1% to 1.2%. The limits of detection (LOD) and limit of quantification (LOQ) were determined based on signal-to-noise ratios of 3:1 and 10:1, respectively.

In addition, the precision and accuracy at the LOQ level were assessed to confirm the method’s suitability for quantifying low-concentration analytes. The accuracy of the method was confirmed by a regeneration experiment. A sample of Epilobium extract impregnated with a known amount of a standard compound was analysed. The intra- and inter-day variation of the analysis was found to be less than 0.24%, and the maximum RSD did not exceed 1.31%, and recovery values ranged from 97.8% to 101.91%, with RSDs below 1.05%, demonstrating high trueness, repeatability, and robustness of the analytical procedure. The analysis was carried out in triplicate on three different days and each concentration point was given in triplicate (Table 3).

The identification of analytes in Epilobium samples was done by comparing the peaks retention times and the UV-spectrum obtained from the chromatogram of the standard solution (Table 4). Results revealed repeatability, accuracy, high sensitivity and good linearity of the method.

During the early stages of plant development, chlorophyll and secondary metabolites are produced due to photosynthetic activity. Therefore, this stage is expected to have high levels of phenolic compounds, mainly phenolic acid derivatives (Kumar and Goel, 2019), typically synthesised in young plant tissues. These metabolites are primarily derived from the shikimic acid and phenylpropanoid pathways. However, at this stage, the biomass is still too low for large-scale collection for the pharmaceutical industry.

During the intensive growth phase, E. hirsutum focuses on vegetative growth, producing leaves and stems. During this phase, the accumulation of marker compounds (particularly chlorogenic acid and gallic acid) is generally associated with the plant’s defence mechanisms against oxidative stress and growth regulation. The primary phenolic acids in the tissues of E. hirsutum plants, gallic, ellagic and chlorogenic acids, influence the formation of tannins in the plant’s raw material (Jürgenson, Matto, and Raal, 2012). In general, the accumulation pattern during the vegetation phases in the three different habitats was the same; namely, the highest concentration was observed during the growth phase with a gradual decrease in content towards the seed formation phase. The highest accumulation was found during the growth phase (April-May), and a decrease was observed during the flowering phase (June-July-August), probably due to their conversion into flavonoids and lignin precursors. During the seed formation phase (September-October), the content of phenolic acids did not exceed 0.2 mg/g DW for stems and 0.8 mg/g DW for leaves. Compared to well-lit dry habitats, higher concentrations were observed in plants collected from shaded wet grasslands and lakeshores.

Hydroxycinnamic acid content varied between habitats and growth stages, highlighting the adaptability of E. hirsutum to environmental conditions. Hydroxycinnamic acid content varied across habitats and growth stages, highlighting the adaptability of E. hirsutum to environmental conditions. For example, elevated levels of chlorogenic acid were noted in samples from shaded habitats. Changes in the accumulation of phenolic compounds in E. hirsutum were strongly influenced by meteorological conditions. In our analysis, increased precipitation in early spring (up to 3.47 mm in April, Supplementary Material, Figure 13) promoted the synthesis of phenolic acids in E. hirsutum, consistent with results obtained in E. angustifolium harvested from the Carpathian Mountains, Ukraine, 2019, which reported similar trends in hydroxycinnamic acid dynamics under varying phenological phases using the same HPLC method (Ivanauskas et al., 2024). In another study, samples of E. angustifolium herb samples collected during flowering (June 2015) and after flowering (September 2015) were analysed (Gryszczyńska et al., 2018). In the samples collected in September, the gallic acid content was 7.04 mg/100 g, while the June samples had a gallic acid content of 3.91 mg/100 g by UPLC-MS/MS. The plants were grown in vitro cultures and micro-propagated from seeds from Germany and Poland; the resulting roots were transferred to a substrate and later planted in the soil. Therefore, in this case, the data comparison is not entirely correct due to the difference in the species and the nature of the origin of the samples. This is only a comparison of the dynamics of the formation of individual substances in representatives of the genus Epilobium.

However, analytical data of E. hirsutum aerial parts collected during the flowering period at the end of June from natural habitats of Iran (Mohammadi Bazargani, Falahati-Anbaran, and Rohloff, 2021) showed that gallic acid had the highest concentration among phenolic compounds (quinic acid, gallic acid, (E)-caffeic acid) and also among all other secondary metabolites with a mean value of 15.12 and 23.36 mg/g DW by GC/MS method after derivatisation with MSTFA reagent.

The metabolic variability of E. hirsutum highlights the importance of selecting the optimal growth stage, habitat, and plant part to maximise the desired metabolite yield for pharmaceutical applications. The results suggest that using E. hirsutum leaves rather than stems may be more feasible. Selection of appropriate habitats and simulation of favourable conditions (e.g., shaded environments) may enhance phenolic acid production, providing a sustainable and scalable source for industrial use.

In Epilobium species, gallic and ellagic acids serve as precursors of hydrolysable tannins, and particularly a cyclic dimeric ellagitannin oenothein B, the dominant pharmacologically active metabolite of this genus (Ramstead et al., 2012; Okuyama et al., 2021). These compounds accumulate during reproductive and senescent stages (Gasperotti et al., 2013) and exhibit significant antioxidant, anti-inflammatory, and anticancer properties (Marino et al., 2022). Their pharmacological potential extends to the prevention and treatment of cardiovascular diseases, neurodegenerative disorders (e.g., Alzheimer’s, Parkinson’s), and metabolic diseases such as diabetes.

As E. hirsutum progresses through the growth and reproductive stages, phenolic acids often serve as precursors for more complex secondary metabolites, including flavonoids (Cheynier et al., 2013). Upon entering the flowering phase, the plant produces vibrant pink-purple flowers, signalling increased flavonoid biosynthesis. This phase also triggers the production of lignins and tannins, which are essential in UV protection, pigmentation, and reproduction. Flavonoids, in particular, attract pollinators and protect them from pathogens, underlining their ecological and pharmacological importance.

The accumulation of specific flavonoids such as avicularin, guajaverin, isoquercitrin, hyperoside, myricetin, and isomyricitrin in E. hirsutum is influenced by the growth stages of the plant, a phenomenon observed in several Epilobium species (Ak et al., 2021). The concentration of flavonoids increases as the plant matures, with the flowering stage often showing higher levels of these metabolites than earlier vegetative stages. This shift is likely to be related to the developmental priorities of the plant, with energy allocation shifting from growth to reproduction, promoting higher synthesis of secondary metabolites such as flavonoids during flowering. Therefore, the HPLC method is recommended for accurate quantitative analysis due to its high precision in measuring individual metabolites in complex plant mixtures.

Glycosylated flavonoids such as hyperoside and isoquercitrin were absent from E. hirsutum stems, and avicularin appeared only during the intense growth phase (April-May) and seed dispersal phase (September) but its content did not exceed 0.02 mg/g DW in samples. Guajaverin content in stems is very low, averaging no more than 0.05–0.1 mg/g DW across samples, gradually increasing with the onset of flowering (June) and peaking at 0.2–0.3 mg/g DW during seed formation (September).

The distribution of metabolites in descending order in Epilobium leaf samples is isoquercitrin > isomyricitrin > hyperoside > guajaverin > avicularin > myricetin. The maximum accumulation of these flavonoid glycosides occurs during the plant’s flowering period (June-August). The maximum accumulation of isoquercitrin found during this period is 1.40–3.08 mg/g DW.

The distribution of the compounds in the stems was the same regardless of the habitat type. However, in the plant samples from the wet grassland and the lake shore, the content of guajaverin was higher in two times (0.28–0.30 mg/g DW for 64S and 66S samples) compared to the mesic grassland (0.12 mg/g DW, 68S) in August-September. Of course, the qualitative and quantitative profile of metabolites in the leaf samples was significantly higher than in the stems, which is also in agreement with the data of other authors (Remmel et al., 2012) on the assessment of the total polyphenols, tannins and flavonoids content by UV spectroscopy in the organs (roots, stems, leaves, and flowers) of E. hirsutum and similar species. This highlights the importance of selecting appropriate plant parts for raw material collection to optimise secondary metabolite yield.

High levels of glycosylated derivatives of quercetin were observed in the E. hirsutum leaf samples collected from the wet grassland and lakeshore habitats, probably due to the moderate environmental light conditions and moisture availability, which favoured the accumulation of metabolites. In our case, samples collected from late May to June-July-August were exposed to the sunshine duration for an average of 10.13–13.07 h/day (Supplementary Material). It should be noted that in the samples (39L-75L) from the lake shore, the second dominant metabolite after isoquercitrin (1.05–4.54 mg/g DW) was hyperoside (0.6–2.5 mg/g DW), possibly due to increased exposure to sunlight and reduced competition.

Myricetin (0.03–0.265 mg/g) and its glucoside (0.04–1.985 mg/g) were also found in the samples and were reported in E. hirsutum from Estonia by (Ain et al., 2024) and from Turkey by (Ak et al., 2021) and Iran by (Mohammadi Bazargani, Falahati-Anbaran, and Rohloff, 2021).

The variability in flavonoid content between E. hirsutum samples highlight the influence of environmental factors, plant part, and phenological stage on secondary metabolite biosynthesis. Higher flavonoid levels during the active flowering phase in samples collected in shaded or moderately illuminated humid habitats suggest a balance between light-induced synthesis and regulation of oxidative stress. These results highlight the role of plant adaptation to the environment in secondary metabolite production, with implications for plant-based pharmaceutical applications.

Understanding the optimal timing of isoquercitrin accumulation is crucial as this flavonoid glycoside exhibits antioxidant, anti-inflammatory, cardioprotective (Valentová et al., 2014) and antiviral (Agrawal, Blunden, and Jacob, 2024) effects, making it beneficial for various health conditions. Most of the identified flavonoids are active constituents of many medicinal plants used in traditional medicine for their neuroprotective, anti-inflammatory, antioxidant, antiproliferative and other pharmacological properties (Schepetkin et al., 2016).

After the flowering phase, E. hirsutum enters the seed development phase. During this phase, the plant focuses on seed maturation (Sripathy and Groot, 2023). Marker compounds associated with this phase include fatty acids, particularly γ-linolenic acid (Mendes et al., 2013), and tannins (Santos et al., 2021) such as oenothein B. During the later stages of plant development, metabolites are degraded, and their levels decrease significantly (up to 3–9 mg/g DW).

Oenothein A and B, the major macrocyclic ellagitannins, are the main bioactive components of E. hirsutum (Yoshida, Yoshimura, and Amakura, 2018), and are known for their antioxidant, immunomodulatory, antiviral, tumour cell cytotoxicity properties (Ramstead et al., 2012). Considering the renewable resources of the plant and the high content of oenothein B depending on the stage of vegetation, the development of herbal preparations based on Epilobium species is promising.

The content of oenothein A in Epilobium stems did not exceed 0.03–0.2 mg/g DW and was the highest in the Epilobium samples from the wet grassland habitat. It should be noted that in the sample 78S, collected in October in the wet grassland during the wilting phase, the content of oenothein A reaches the maximum value of all samples, namely, 8.43 mg/g DW, which can be attributed to the increased occurrence of rainfall during this period (3,69 mm of precipitation in October). In the leaves, the content of oenothein A was slightly higher than in the stems and was at the level of other phenolic compounds, ranging from 0.03 to 1.26 mg/g DW, with the maximum accumulation in June: 0.38 mg/g DW for the 25 L sample from the mesic grassland, 1.26 mg/g DW for the 27L sample from the lakeshore and 2.26 mg/g DW for the 41L sample from the wet grassland. As for the 78S sample of E. hirsutum stems from the wet grassland, the oenothein content in the 77L sample leaves during the rainy season in October was comparatively high at four times (7.52 mg/g DW).

The concentration of oenothein B in E. hirsutum samples was significantly higher than that of other phenolic compounds. It ranges from 5.11 to 53.75 mg/g DW for stems and from 4.27 to 73.97 mg/g DW for leaves, depending on the habitats and the development phase. The highest accumulation of oenothein B was observed in E. hirsutum samples collected during the flowering period of the plant in June in the United Kingdom, which correlates with the data of the authors (Jürgenson, Matto, and Raal, 2012) who studied the total tannin content in E. angustifolium organs (roots, stems, flowers) during the whole growth period from May to October in Estonia and indicated that the highest tannin content was found in small growing plants in May. The highest average temperature (up to 18 °C, Supplementary Material) is also observed in June and September in Wallington area, United Kingdom in 2023. Due to geographical differences between countries, the flowering period of the plant falls in different months (Sell and Murrell, 2009). Based on the comparative analysis, the above-ground organs without stems, collected in the United Kingdom in July-August, are the best choice for obtaining E. hirsutum plant material with a consistently high phenolic compound content.

Leaf samples from the mesic grassland had lower levels of oenothein B, possibly due to reduced water availability. Plants from the wet grassland and lakeshore habitats had the highest levels of oenothein B, reflecting optimal growth conditions, with samples from the lake shore habitat having slightly higher levels, possibly due to the higher light intensity, which favours ellagitannin biosynthesis (Niemetz and Gross, 2005). Generally, a pattern of biosynthesis of phenolic compounds was observed in the samples, with a decrease in the levels of phenolic acids and an increase in the levels of flavonoids and tannins as products of their transformation.

The biosynthesis of phenolic acids, flavonoids and tannins in E. hirsutum is regulated by environmental factors such as light intensity, temperature and water availability. For example, UV exposure is a known signal for flavonoid accumulation, particularly quercetin derivatives, which act as UV protectors and antioxidants. Similarly, phenolic acid synthesis is often increased in response to abiotic stress such as drought or oxidative stress, acting as part of the plant’s defence mechanism. In our study, the increased flavonoid and tannin levels observed in shaded or moist habitats suggest that moderate stress and light exposure may positively influence secondary metabolite accumulation, possibly through regulation of the phenylpropanoid pathway. In Understanding these variations is crucial for optimising raw material collection to ensure sustainable harvesting and high-quality pharmaceutical products.

Due to climate change, the predictability of crop seasons has become more uncertain in recent decades. Changes in meteorological conditions are the main factor affecting harvesting time, productivity, and the quality of plant raw materials (Yeleliere, Antwi-Agyei, and Guodaar, 2023). From 1880 to 2012, the planet’s temperature increased by 1.54°C (Copernicus, 2025). Heat waves, which are increasing in frequency and duration, can cause premature ripening of crops and medicinal plants, leading to changes in the yield and quality of plant raw materials.

The meteorological data for the selected sampling sites is very similar, as they were collected in the same county as England (Figure 13). It can only be noted that Lake Shore (site 2 in the Supplementary Material) and wet grassland (site 1) had higher average temperatures (during June–September 17°C–18°C) and slightly longer sunshine duration (up to 13 h in June) than mesic grassland (site 0), which was in a shaded area. However, all three sites showed similar trends in sunshine duration (from 6.5 to 13 h, depending on the month), indicating that they are exposed to comparable regional weather conditions. Lake shore showed slightly greater temperature variations, probably due to the influence of the lake, which moderates temperatures but can also create sharper contrasts. Precipitation was similar at all three sites, but soil moisture retention was longer at wet grassland and lake shore. General meteorological observations were as follows: (1) lake shore was the wettest and most variable due to the influence of water; (2) wet grassland has the most consistent wet conditions with slightly higher temperatures; (3) mesic grassland was the driest and coolest of the three, with more consistent conditions but less precipitation. These differences highlight how local ecosystems respond to different climates, affecting plant growth, soil moisture, and biodiversity at each site. While E. hirsutism can be sustainably sourced due to its weediness, our finding supports and highlights the importance of an integrative approach to ecological assessment and sustainable harvesting practices, including the development of targeted strategies to reduce waste and optimise resource use. Identifying the optimal phenological stages for harvesting reduces unnecessary collection of plant material, thereby conserving biodiversity and reducing environmental impact. The observed variability in bioactive compound accumulation across phenological stages and habitats highlights the need for evidence-based harvesting strategies. Our results support the need for individual harvesting protocols for E. hirsutum, with a focus on harvesting leaves during the flowering stage under moist, shaded habitats to maximise the yield of bioactive components and, as a consequence, the pharmacological potential of plant application.