Castanea sativa Mill. Leaf: UHPLC-HR MS/MS Analysis and Effects on In Vitro Rumen Fermentation and Methanogenesis

Castanea sativa Mill. (Fagaceae) is a deciduous tree grown for its wood and edible fruits. Chestnut processing produces residues (burs, shells, and leaves) exploitable for their diversity in bioactive compounds in animal nutrition. In fact, plant-specialized metabolites likely act as rumen modifiers. Thus, the recovery of residual plant parts as feed ingredients is an evaluable strategy. In this context, European chestnut leaves from northern Germany have been investigated, proving to be a good source of flavonoids as well as gallo- and ellagitannins. To this purpose, an alcoholic extract was obtained and an untargeted profiling carried out, mainly by means of ultra-high-performance liquid chromatography/high-resolution tandem mass spectrometry (UHPLC-HR MS/MS) techniques. To better unravel the polyphenol constituents, fractionation strategies were employed to obtain a lipophilic fraction and a polar one. This latter was highly responsive to total phenolic and flavonoid content analyses, as well as to antiradical (DPPH● and ABTS+●) and reducing activity (PFRAP) assays. The effect of the alcoholic extract and its fractions on rumen liquor was also evaluated in vitro in terms of fermentative parameter changes and impact on methanogenesis. The data acquired confirm that chestnut leaf extract and the fractions therefrom promote an increase in total volatile fatty acids, while decreasing acetate/propionate ratio and CH4 production.


Introduction
The genus Castanea (Fagaceae family) consists of twelve species of deciduous trees native to temperate regions of the northern hemisphere. Among them, Castanea sativa Mill. (European or sweet chestnut) is the most representative species in Southern Europe, with Italy representing the largest chestnut-producing area, although its worldwide distribution area is very extensive ( Figure S1), ranging from Southern Europe and North Africa to North-Western Europe and eastward to Western Asia [1].
C. sativa Mill. is of interest for its precious wood and its edible seed (nut), which is composed of the fruit, the pericarp (outer shell), the integument (inner shell), and the surrounding bur. In recent years, the consumption of fresh or processed chestnuts has significantly increased due to their positive effects on health. In fact, chestnut fruits can be used as an energy source due to their content of carbohydrates (especially starch) and fatty acids (mainly linoleic, oleic, and palmitic acids). In addition, chestnuts, as purée and chestnut flour, are finding a new nutritional application as ingredients in gluten-free diets [2]. Indeed, because only the nuts are used for food processing, the shells, leaves, and burs represent common waste products of the chestnut food industry, even if these The chemical composition assessment, preliminarily carried out using colorimetric assays, markedly differentiated Cs/1/1 and its fractions Cs/2/1 and Cs/3/2. The first one showed TPC and TFC values equal to 288.1 ± 1.1 gallic acid equivalents (GAEs) and 52.9 ± 4.4 mg quercetin equivalents (QUEs) per g of extract, respectively (Figure 2A(a,b)). These values are in line with those by Cerulli et al. [22], who, evaluating the TPC and TFC of a chestnut leaf methanol extract, found that total phenols were 299.0 ± 14.8 mg GAEs, while flavonoids were 45.54 ± 0.99 mg RUEs per g of extract. Indeed, due to fractionation, Cs/3/2 appeared to be enriched in phenolic and flavonoid compounds, exhibiting a TPC value threefold higher than Cs/1/1, while flavonoids accounted for 83.2 ± 10.7 mg QUEs per g of extract. To clarify the main actors of Cs/3/2 TP content, which could be attributable to The chemical composition assessment, preliminarily carried out using colorimetric assays, markedly differentiated Cs/1/1 and its fractions Cs/2/1 and Cs/3/2. The first one showed TPC and TFC values equal to 288.1 ± 1.1 gallic acid equivalents (GAEs) and 52.9 ± 4.4 mg quercetin equivalents (QUEs) per g of extract, respectively (Figure 2A(a,b)). These values are in line with those by Cerulli et al. [22], who, evaluating the TPC and TFC of a chestnut leaf methanol extract, found that total phenols were 299.0 ± 14.8 mg GAEs, while flavonoids were 45.54 ± 0.99 mg RUEs per g of extract. Indeed, due to fractionation, Cs/3/2 appeared to be enriched in phenolic and flavonoid compounds, exhibiting a TPC value threefold higher than Cs/1/1, while flavonoids accounted for 83.2 ± 10.7 mg QUEs per g of extract. To clarify the main actors of Cs/3/2 TP content, which could be attributable to gallo-and ellagitannins [23,24], UHPLC-ESI-QqTOF tandem mass spectrometry analysis was carried out, as well as antiradical and reducing power assessment. In this latter regard, Cs/3/2 showed marked scavenging activity with both ABTS •+ and DPPH • , with relative ID 50 values equal to 11.2 ± 0.3 µg/mL and 30.3 ± 0.8 µg/mL. Moreover, Cs/3/2 effectively reduced ferric ions at the lowest tested doses, so much so that ID 50 and TEAC (Trolox Equivalents Antioxidant Capacity) values were 0.37 ± 0.02 µg/mL and 4.7, respectively ( Figure 2B(a-c)). gallo-and ellagitannins [23,24], UHPLC-ESI-QqTOF tandem mass spectrometry analysis was carried out, as well as antiradical and reducing power assessment. In this latter regard, Cs/3/2 showed marked scavenging activity with both ABTS •+ and DPPH • , with relative ID50 values equal to 11.2 ± 0.3 µg/mL and 30.3 ± 0.8 µg/mL. Moreover, Cs/3/2 effectively reduced ferric ions at the lowest tested doses, so much so that ID50 and TEAC (Trolox Equivalents Antioxidant Capacity) values were 0.37 ± 0.02 µg/mL and 4.7, respectively ( Figure 2B(a,b,c)).

Chemical Investigation of Alcoholic Leaf Extract and Cs/2/1 and Cs/3/2 Fractions
According to literature data, C. sativa leaf extracts, obtained using different alcoholic or hydroalcoholic extractants, are a promising source of natural antioxidants [25,26]. The bioactivity of chestnut leaf extracts is involved in suppressing the Nrf2-mediated antioxidant system in cancer stem cells through ROS accumulation and apoptosisinducing effects [25], as well as by attenuating TLR2-and TLR4-induced inflammatory responses [26]. These and other properties made chestnut leaves usable in topical applications for preventing and/or treating oxidative stress-mediated diseases and photoageing [6,[27][28][29]. In addition, Castanea sativa leaf showed antimicrobial activity against such multiresistant bacteria as Staphylococcus epidermidis, S. aureus, and Pseudomonas aeruginosa [30]. Herein, in order to hypothesize the use of this forest residue in ruminant feeding, the chemical composition in terms of specialized metabolites was tentatively determined by means of UHPLC-HR MS/MS analyses. These latter, firstly carried out on the alcoholic chestnut leaf extract, highlighted the occurrence of ellagitannins, and flavonoid mono-and di-glycosides, beyond fatty acids and terpene compounds. When fractionation was carried out, tannins and flavonoids appeared to constitute Cs/3/2. This

Chemical Investigation of Alcoholic Leaf Extract and Cs/2/1 and Cs/3/2 Fractions
According to literature data, C. sativa leaf extracts, obtained using different alcoholic or hydroalcoholic extractants, are a promising source of natural antioxidants [25,26]. The bioactivity of chestnut leaf extracts is involved in suppressing the Nrf2-mediated antioxidant system in cancer stem cells through ROS accumulation and apoptosis-inducing effects [25], as well as by attenuating TLR2-and TLR4-induced inflammatory responses [26]. These and other properties made chestnut leaves usable in topical applications for preventing and/or treating oxidative stress-mediated diseases and photo-ageing [6,[27][28][29]. In addition, Castanea sativa leaf showed antimicrobial activity against such multiresistant bacteria as Staphylococcus epidermidis, S. aureus, and Pseudomonas aeruginosa [30]. Herein, in order to hypothesize the use of this forest residue in ruminant feeding, the chemical composition in terms of specialized metabolites was tentatively determined by means of UHPLC-HR MS/MS analyses. These latter, firstly carried out on the alcoholic chestnut leaf extract, highlighted the occurrence of ellagitannins, and flavonoid mono-and di-glycosides, beyond fatty acids and terpene compounds. When fractionation was carried out, tannins and flavonoids appeared to constitute Cs/3/2. This was confirmed by the UV-Vis spectrum of the fraction ( Figure S2, Supplementary Material), whose bands at 365, 267, and 215 nm were in line with flavonoid and gallo-and ellagitannin electronic transitions [31]. Different spectroscopic and spectrometric behavior was noted for Cs/2/1 fraction. In fact, the UV-Vis absorption bands occurred at 667 and 415, attributable to chlorophylls and carotenoids, and at 236 and 205 nm. These latter are commonly ascribed to fatty acids and their oxidative derivatives, as well as to triterpenoids. Indeed, the mass spectra of Cs/2/1 compounds ascertained a complexity greater than that of Cs/1/1, due to the triterpenes, acylated flavonoids, and ellagic acid methyl derivatives. In Tables 1-4, as well as Table S1, data from UHPLC-QqTOF-MS/MS analyses on all three extracts are reported. The relative quantification assessment was also performed and plotted using principal component analysis (PCA; Figure 3C). was confirmed by the UV-Vis spectrum of the fraction ( Figure S2, Supplementary Material), whose bands at 365, 267, and 215 nm were in line with flavonoid and gallo-and ellagitannin electronic transitions [31]. Different spectroscopic and spectrometric behavior was noted for Cs/2/1 fraction. In fact, the UV-Vis absorption bands occurred at 667 and 415, attributable to chlorophylls and carotenoids, and at 236 and 205 nm. These latter are commonly ascribed to fatty acids and their oxidative derivatives, as well as to triterpenoids. Indeed, the mass spectra of Cs/2/1 compounds ascertained a complexity greater than that of Cs/1/1, due to the triterpenes, acylated flavonoids, and ellagic acid methyl derivatives. In Tables 1-4, as well as Table S1, data from UHPLC-QqTOF -MS/MS analyses on all three extracts are reported. The relative quantification assessment was also performed and plotted using principal component analysis (PCA; Figure 3C).      Among the polar minor constituents of the chestnut alcoholic extract (Tables 1 and 2), the greater part was found only in the Cs/3/2 fraction together with gallo-and ellagitannins.

Chestnut Leaf Flavonoids
Flavonoids in the investigated chestnut leaf extract and derived fractions are listed in Table 3

Effects of Chestnut Leaf Alcoholic Extract and Its Fractions on In Vitro Rumen Fermentative Activity
The use of bioactive compounds as natural feed supplements has been extensively studied as a strategy to manipulate rumen fermentative activities, and the effects of flavonoid and tannin extract as supplements in animal nutrition have been increasingly investigated, especially for ruminants. The availability of literature for condensed tannins (CTs) is extensive, whereas hydrolysable tannins (HTs) have been less extensively explored in animal nutrition, even if little attention is given to the chemical composition of the treated extracts. Data about the rumen fermentation effects of Castanea sativa extracts leaf and/or only leaves are scarce [42]; however, different studies reported the effects of chestnut involucre or tannin wood extracts on ruminal fermentation [13,[43][44][45].
Nonetheless, herein, the presence of other bioactive compounds, such as flavonoids, triterpenes, or fatty acids, in the Cs/1/1 and Cs/3/2 fractions might interfere and modulate the tannin effects. Tannins generally exert a dual mechanism: high doses reduce voluntary feed intake and nutrient digestibility, whereas moderate concentrations can improve nutrient utilization. Moreover, tannins modify the digestive processes of ruminants not only by binding dietary protein, but also through modulation of rumen microbiota; not surprisingly, hydrolysable and condensed tannins as well as flavonoids are considered as antimicrobial feed additives, due to their antibacterial and antiparasitic activity [46]. Furthermore, these secondary metabolites can inhibit the enzymatic activity of microbial protease and urease into the rumen [47]. The supplementation of polyphenols in ruminant diet can modulate the diversity and activity of rumen microorganisms, the nutrient degradability and rumen methanogenesis. The differences between substrates in the fermentation process are clearly illustrated in Table 5 and Figure 6, where the gas production rate and in vitro fermentation rate over time are shown. The organic matter degradability (OMD) after 120 h of incubation showed, for the all-chestnut-based diet, values lower than the control diet. It was observed that chestnut tannins linearly decreased OMD [48] as well as Castanea sativa leaves [42]. Analyzing the fermentation effects of tannic acid (HT) and quebracho tannins (CTs) on wheat and corn grain, Martìnez et al. [49] highlighted a reduction in gas production after 24 h of incubation by means of a physical mechanism. Scanning electron microscopy (SEM) revealed that both sources of tannins slowed starch hydrolysis through degradation of the endosperm protein matrix. Due to their chemical structure, tannins can bind protein through hydrogen bonds forming tannin-protein complexes that are stable at the rumen pH (5.0-7.0) [50]. Similarly, except for Cs/1/1 at the 200-mg dose level, the tested samples produced a lower amount of gas (OMCV) than the control diet. In particular, the lipophilic fraction Cs/2/1 exhibited the lowest values (p < 0.001) for OMD and OMCV at both tested dose levels. Furthermore, the Cs/2/1-200 mg sample showed the lowest value for T max and a significative decrease (p < 0.05) for R max . Beyond the Cs/2/1-200 mg dose level, for the Cs/3/2-50 mg sample, a reduction in both parameters T max and R max was also noted.   The results after 24 h of incubation showed that supplementation with the extract does not depress degradability; even the 200 mg dose levels of three extracts seemed to significantly improve organic matter degradability (OMD). The rumen pH did not appear to change after 24 h. By contrast, Hassanat and Benchaar [43] demonstrated a significant increase in rumen pH after only 24 h of incubation with different tannin extracts sources (i.e., chestnuts, acacia, quebracho, and valonia). In our case, after 120h, the rumen pH was significantly modified by the chestnut alcoholic extract and its fractions. In particular, the Cs/2/1-50 mg sample showed the highest value of rumen pH followed by Cs/1/1-50 mg and Cs/1/1-200 mg samples. Only the Cs/3/2-200 mg sample showed a slight increase (p < 0.05) in rumen pH. Dìaz-Carrasco et al. [51], evaluating the effects of chestnuts and quebracho tannins on rumen microbiota, highlighted that tannin addiction increased ruminal pH.   The results after 24 h of incubation showed that supplementation with the extract does not depress degradability; even the 200 mg dose levels of three extracts seemed to significantly improve organic matter degradability (OMD). The rumen pH did not appear to change after 24 h. By contrast, Hassanat and Benchaar [43] demonstrated a significant increase in rumen pH after only 24 h of incubation with different tannin extracts sources (i.e., chestnuts, acacia, quebracho, and valonia). In our case, after 120h, the rumen pH was significantly modified by the chestnut alcoholic extract and its fractions. In particular, the Cs/2/1-50 mg sample showed the highest value of rumen pH followed by Cs/1/1-50 mg and Cs/1/1-200 mg samples. Only the Cs/3/2-200 mg sample showed a slight increase (p < 0.05) in rumen pH. Dìaz-Carrasco et al. [51], evaluating the effects of chestnuts and quebracho tannins on rumen microbiota, highlighted that tannin addiction increased ruminal pH.  The results after 24 h of incubation showed that supplementation with the extract does not depress degradability; even the 200 mg dose levels of three extracts seemed to significantly improve organic matter degradability (OMD). The rumen pH did not appear to change after 24 h. By contrast, Hassanat and Benchaar [43] demonstrated a significant increase in rumen pH after only 24 h of incubation with different tannin extracts sources (i.e., chestnuts, acacia, quebracho, and valonia). In our case, after 120h, the rumen pH was significantly modified by the chestnut alcoholic extract and its fractions. In particular, the Cs/2/1-50 mg sample showed the highest value of rumen pH followed by Cs/1/1-50 mg and Cs/1/1-200 mg samples. Only the Cs/3/2-200 mg sample showed a slight increase (p < 0.05) in rumen pH. Dìaz-Carrasco et al. [51], evaluating the effects of chestnuts and quebracho tannins on rumen microbiota, highlighted that tannin addiction increased ruminal pH.  Furthermore, all the extracts depressed methane production related to incubated organic matter and organic matter degradability. This was particularly true of Cs/1/1 and Cs/2/1 supplementation, and appeared to correlate with acetate and propionate production data (Figure 7). Hydrolysable tannins, as well as condensed ones, decreased the total methanogen population, with a positive effect on methane production [16]. Generally, the ability of polyphenols to reduce methane emissions could be due to the decrease in NDF degradation, which leads to a reduction in acetate synthesis, and eventually, to a decrease in the availability of electron donors for the methanogens [52]. Indeed, the reduction in acetate with the increase in propionate plays a key role in establishing an internal hydrogen balance by regulating methanogenesis. The formation of acetate from pyruvate produces metabolic hydrogen, the main substrate of methanogenesis, which is consumed by the formation of propionate from pyruvate [53]. After 24 h of incubation, acetic acid was reduced in all samples tested, with the sole exception of Cs/3/2-50 mg, which favors a 3.7% increase over the control diet. Instead, propionate increased significantly following treatment with Cs/1/1, and decreased only by Cs/3/2-50 mg. The higher molar proportion of propionate in the in vitro fermentation system may be associated with a lower count of protozoa produced by tannins [47,50]. Unsurprisingly, the Cs/3/2-50 mg fraction showed the lowest inhibition of methanogenesis and the highest increase in the A/P ratio, which was reduced by all treatments. Total VFAs increased to the highest percentage with Cs/2/1-200 mg treatment, followed by alcoholic extract at both dosages, which increased by between 32.2% and 30.5 %. In contrast, Castanea sativa leaf reduced total VFAs with a higher percentage of propionate after 24 h incubation, producing low total gas and CH4/digestible organic matter. This, as previously hypothesized, could be achieved without a significant reduction in bacterial and protozoal counts [42]. Unlike the 120 h of incubation, during The fermentation rates after 24 h of incubation following the different C. sativa extract supplementations are reported in Table 6, while data acquired in terms of pH and the relative concentration of the fermentation end products are listed in Table S2 and represented by the scatter plot in Figure 7. In vitro fermentation end products after 120 h of incubation were reported in Table 7, and according to the single volatile fatty acids, are reported as proportion (%) of single volatile fatty acids within the total volatile fatty acid (VFAs) content, expressed as mmol/L (Table 7). Table 6. In vitro gas production and fermentation rate of different C. sativa extracts after 24 h of incubation. OMD: organic matter degradability; CH 4 (mL/g iOM): methane production related to incubate organic matter; CH 4 (mL/d OM): mL of methane production related to degraded OM. Along the row ** p < 0.01 and *** p < 0.001; NS : not significant; MSE: mean square error.  After 120 h of incubation, while some parameters maintained their trend (e.g., total VFAs, acetate, propionate, iso-valerate), others showed an inversion (e.g., pH, butyrate, valerate, iso-butyrate). Furthermore, the total VFAs increased, and the observed effect for all the tested samples appeared to be dose-dependent (Table 7), except for Cs/3/2. Cs/2/1-200 mg increased the VFA content by a percentage of 54.2% compared to the control diet,  Table 7. Effects of C. sativa Cs/1/1 extract and its fractions Cs/2/1 and Cs/3/2 at 50-mg and 200-mg dose levels on fermentation end products after 120 h of incubation. Total VFA: total volatile fatty acids (acetate + propionate + butyrate + iso-butyrate + valerate + iso-valerate); AcA = acetic acid; PrA = propionic acid; ButA = Butyric acid; ValA = valeric acid; iso-ButA = iso-butyric acid; iso-ValA = iso-valeric acid; BCFA= branched chain fatty acids (iso-butyrate + iso-valerate/total VFAs); A/P = acetate/propionate. Along the row * p < 0.05, ** p < 0.01 and *** p < 0.001; NS: not significant; MSE: mean square error. In the lower panel, the percentage increase or decrease in each volatile fatty acid is plotted for different tested dose levels (• 50 mg and • 200 mg) vs. FA% in the control diet.

Plant Collection, Fractionation, and Evaluation of Leaf Chemical Composition
The leaves of C. sativa Mill. were collected in August 2021. Voucher specimens were deposited in the private herbarium of CZ (CZ-20190628A-1) and in the Herbarium of Kiel University (KIEL0005015); leaves for phytochemical analyses and twigs for the vouchers were collected in the Botanical Garden of the Pharmaceutical Institute of Kiel University in Kiel/Schleswig-Holstein/Germany, N 54°20′00.8″, E 10°06′58.3″, 33 m a.m.s.l., 28.06.2019, leg.: C. Zidorn, det.: C. Zidorn. The identity of the single cultivated tree was confirmed using the German standard reference flora by Jäger [59].

Plant Collection, Fractionation, and Evaluation of Leaf Chemical Composition
The leaves of C. sativa Mill. were collected in August 2021. Voucher specimens were deposited in the private herbarium of CZ (CZ-20190628A-1) and in the Herbarium of Kiel University (KIEL0005015); leaves for phytochemical analyses and twigs for the vouchers were collected in the Botanical Garden of the Pharmaceutical Institute of Kiel University in Kiel/Schleswig-Holstein/Germany, N 54°20′00.8″, E 10°06′58.3″, 33 m a.m.s.l., 28.06.2019, leg.: C. Zidorn, det.: C. Zidorn. The identity of the single cultivated tree was confirmed using the German standard reference flora by Jäger [59].

Plant Collection, Fractionation, and Evaluation of Leaf Chemical Composition
The leaves of C. sativa Mill. were collected in August 2021. Voucher specimens were deposited in the private herbarium of CZ (CZ-20190628A-1) and in the Herbarium of Kiel University (KIEL0005015); leaves for phytochemical analyses and twigs for the vouchers were collected in the Botanical Garden of the Pharmaceutical Institute of Kiel University in Kiel/Schleswig-Holstein/Germany, N 54°20′00.8″, E 10°06′58.3″, 33 m a.m.s.l., 28.06.2019, leg.: C. Zidorn, det.: C. Zidorn. The identity of the single cultivated tree was confirmed using the German standard reference flora by Jäger [59].
The leaves were first lyophilized and pulverized using a rotating knife homogenizer. Dried leaves underwent ultrasound-assisted maceration (UAM; Branson UltrasonicsTM BransonicTM M3800-E; Danbury, CT, USA) using methanol as extractive solvent. The raw material/solvent ratio was 1:5 (g raw material: mL solvent). Three sonication cycles were carried out, each one of 30 min, to obtain the Cs/1/1 extract. The extract yield (%) was equal to 24.3 % (103.7 g). The alcoholic extract was then dissolved in a biphasic solution CHCl3:MeOH:H2O (13:7:6, v:v:v), and discontinuous liquid-liquid extraction (LLE) was performed. Thus, an organic fraction (Cs/2/1; 16.2% of Cs/1/1) and a hydroalcoholic one (Cs/2/2) were obtained. The fraction Cs/2/2 was further chromatographed on XAD-4 resin, using water first and then methanol. The alcoholic fraction Cs/3/2 was obtained with a 26.6% yield. The results after 24 h of incubation showed that supplementation with the extract does not depress degradability; even the 200 mg dose levels of three extracts seemed to significantly improve organic matter degradability (OMD). The rumen pH did not appear to change after 24 h. By contrast, Hassanat and Benchaar [43] demonstrated a significant increase in rumen pH after only 24 h of incubation with different tannin extracts sources (i.e., chestnuts, acacia, quebracho, and valonia). In our case, after 120 h, the rumen pH was significantly modified by the chestnut alcoholic extract and its fractions. In particular, the Cs/2/1-50 mg sample showed the highest value of rumen pH followed by Cs/1/1-50 mg and Cs/1/1-200 mg samples. Only the Cs/3/2-200 mg sample showed a slight increase (p < 0.05) in rumen pH. Dìaz-Carrasco et al. [51], evaluating the effects of chestnuts and quebracho tannins on rumen microbiota, highlighted that tannin addiction increased ruminal pH.
Furthermore, all the extracts depressed methane production related to incubated organic matter and organic matter degradability. This was particularly true of Cs/1/1 and Cs/2/1 supplementation, and appeared to correlate with acetate and propionate production data (Figure 7). Hydrolysable tannins, as well as condensed ones, decreased the total methanogen population, with a positive effect on methane production [16]. Generally, the ability of polyphenols to reduce methane emissions could be due to the decrease in NDF degradation, which leads to a reduction in acetate synthesis, and eventually, to a decrease in the availability of electron donors for the methanogens [52]. Indeed, the reduction in acetate with the increase in propionate plays a key role in establishing an internal hydrogen balance by regulating methanogenesis. The formation of acetate from pyruvate produces metabolic hydrogen, the main substrate of methanogenesis, which is consumed by the formation of propionate from pyruvate [53]. After 24 h of incubation, acetic acid was reduced in all samples tested, with the sole exception of Cs/3/2-50 mg, which favors a 3.7% increase over the control diet. Instead, propionate increased significantly following treatment with Cs/1/1, and decreased only by Cs/3/2-50 mg. The higher molar proportion of propionate in the in vitro fermentation system may be associated with a lower count of protozoa produced by tannins [47,50]. Unsurprisingly, the Cs/3/2-50 mg fraction showed the lowest inhibition of methanogenesis and the highest increase in the A/P ratio, which was reduced by all treatments. Total VFAs increased to the highest percentage with Cs/2/1-200 mg treatment, followed by alcoholic extract at both dosages, which increased by between 32.2% and 30.5%. In contrast, Castanea sativa leaf reduced total VFAs with a higher percentage of propionate after 24 h incubation, producing low total gas and CH 4 /digestible organic matter. This, as previously hypothesized, could be achieved without a significant reduction in bacterial and protozoal counts [42]. Unlike the 120 h of incubation, during the first 24 h, butyric acid underwent a notable decrease, mainly at the dose of 200 mg of both the Cs/1/1 extract and its Cs/3/2 fraction. These results are only partially in line with Castro-Montoya et al. [48], who, testing different doses of chestnut tannins (0.5, 0.75, 1 mg/mL) for 24 h of incubation, found a significant decrease in acetate and A/P with a linear increase in propionate. An opposite trend was obtained for total VFAs, butyrate, valerate, and branched-chain fatty acids (BCFA). In addition, valerate increased at all doses tested by up to 50.3% after treatment with Cs/2/1-200 mg.
After 120 h of incubation, while some parameters maintained their trend (e.g., total VFAs, acetate, propionate, iso-valerate), others showed an inversion (e.g., pH, butyrate, valerate, iso-butyrate). Furthermore, the total VFAs increased, and the observed effect for all the tested samples appeared to be dose-dependent (Table 7), except for Cs/3/2. Cs/2/1-200 mg increased the VFA content by a percentage of 54.2% compared to the control diet, followed by Cs/3/2-50 mg, which was able to increase its VFAs by 1.4 times. However, previous reports on chestnut wood tannins (CTWs) highlighted that CTWs did not significantly increase total VFAs [45], or even observed a reduction in VFAs with CTWs [43,54]. The significant effect on VFAs in this study was probably attributable to the high flavonoid content of chestnut leaves [15]. Among the main volatile fatty acids (acetate, propionate, and butyrate), only for propionic and butyric acid was an increase recorded for all the dose fractions tested, while acetate decreased. Propionate increased in various degrees; the Cs/1/1 extract at 50 and 200 mg dose levels was able to increase it by 42.4 and 64.9%, respectively. Both dose levels of Cs/2/1 fraction enhanced propionic acid production, whereas Cs/3/2 resulted in a dose-dependent reduction. Data on acetate levels could be justified by the ability of compounds therein to inhibit acetate-producing bacteria (e.g., Ruminococcus albus, Butyrivibrio fibrisolvens), either by direct action or by inhibiting the production of their substrates [48]. To support this hypothesis, in a previous study testing increasing levels (0, 5, 10, 15, and 20 g/kg of diet DM) of A. mearnsii CTs in the diets of Jersey steers, it was found that the ruminal protozoa population was not affected by CTs, while protein digestibility, ruminal pH, and acetate proportion decreased [55]. The A:P ratio appeared to significantly (p < 0.001) decrease at all diet dose levels. Furthermore, Cs/1/1-50 mg and Cs/2/1 at both doses showed a similar increase in butyrate. Moreover, a non-significant variation intra-dose of butyrate was recorded for the Cs/3/2 fraction. The Cs/3/2-200 mg dose level showed the lowest decrease in acetate, with a value equal to 60.3% VFA. It was seen that a chestnut/quebracho tannin supplementation (ratio 1:2) was effective in reducing ruminal NH 3 -N concentration and molar proportion of branched-chain fatty acids, providing a decrease in amino acid deamination [56]. Herein, the same trend was observed for BCFAs, which decreased at all the tested doses (p < 0.001), mainly at the Cs/3/2-50 mg dose level. Iso-valerate is a branched-chain VFA from the deamination of leucine, and in the present study, its proportion decreased after only 24 h of incubation at greater doses of polyphenols (fraction Cs/3/2). The decrease in iso-valerate with a higher dose of tannins may indicate a reduction in ruminal protein degradation [54]. Analogously, iso-butyrate, which showed an increase after 24 h, decreased (p < 0.05) after 120 h, but not as significantly as iso-valerate. Fraction Cs/2/1 further modified some fermentation parameters (e.g., pH, total VFAs, propionic acid, butyric acid, and valeric acid) for its terpenoid constituents, which are able to influence fermentation process with effects also on the methanogenesis [57] and milk production [58]. Valerate, as opposed to 24 h of incubation, was significantly reduced in all tested fractions, mainly by means of the Cs/3/2-200 mg sample.
The dendrograms related to the results of total VFAs and related to a diet at the dose level of 50 and 200 mg are in Figure S11A. Both dendrograms showed three clusters: the first group included total VFA and acetic acid, the second consisted of propionic and butyric acid, and the third included two subgroups with valeric acid, A/P, BCFA, iso-valeric acid (IIIa) and iso-butyric acid (IIIb). Principal component analysis (PCA) ( Figure S11B) at 50 mg showed the positive correlation of total VFA with the enriched polyphenol fraction (Cs/3/2), which showed the highest amount of TFC (r = 0.998) and TPC (r = 0.913) and acetate at the apolar fraction (named Cs/2/1) (r = 0.207 in correlation with TFC; r = −0.153 in correlation with TPC). An opposite trend was displayed at the 200 mg dosage. Other volatile fatty acids showed similar distribution at both strengths, with the exception of propionate and butyrate.

Plant Collection, Fractionation, and Evaluation of Leaf Chemical Composition
The leaves of C. sativa Mill. were collected in August 2021. Voucher specimens were deposited in the private herbarium of CZ (CZ-20190628A-1) and in the Herbarium of Kiel University (KIEL0005015); leaves for phytochemical analyses and twigs for the vouchers were collected in the Botanical Garden of the Pharmaceutical Institute of Kiel University in Kiel/Schleswig-Holstein/Germany, N 54 • 20 00.8", E 10 • 06 58.3", 33 m a.m.s.l., 28.06.2019, leg.: C. Zidorn, det.: C. Zidorn. The identity of the single cultivated tree was confirmed using the German standard reference flora by Jäger [59].
Chestnut leaves were also analyzed according to the procedures of the Association of Official Agricultural Chemists [60] to determine dry matter (DM), ether extract (EE), crude protein (CP) and ash. The fiber fractions (neutral detergent fiber on organic matter basis, NDFom; acid detergent fiber on organic matter basis, ADFom; acid detergent lignin, ADL) were also determined according to Van Soest et al. [61].

UHPLC-HRMS and MS/MS Parameters and UV-Vis Analyses
The alcoholic extract, Cs/1/1, and the fractions therefrom were first analyzed by UV-Vis spectrophotometry in the range 200-800 nm by a Cary 100 spectrophotometer. The three samples (10 mg/mL) were profiled by a NEXERA UHPLC system (Shimadzu; Tokyo, Japan) equipped with Luna ® Omega C-18 column (1.6-µm particle size, 50
3.3.1. Determination of DPPH (2,2 -Diphenyl-1-Picrylhydrazyl) Radical Scavenging Capacity Samples were mixed into a DPPH • methanol solution (9.4 × 10 −5 M). Mixtures were stirred for 15 min; after that, the absorption was read at 517 nm by a Wallac Victor3 spectrophotometer with reference to a blank. The results were expressed in terms of the percentage reduction in the initial radical adsorption by the tested samples [62]. Trolox (2,4,8,16, and 32 µM) was used as positive standard.

Determination of Potassium Ferricyanide Reducing Power (PFRAP)
The ability to reduce the Fe(III) of chestnut leaf Cs/1/1 extract and its fractions were estimated using the ferricyanide FRAP assay, according to PFRAP procedure. The absorbance was measured at 700 nm and the increase in absorbance with reference to the blank was considered to value the reducing power [63]. Trolox (2,4,8,16,32 µM) was used as positive standard.

Determination of Total Phenolic Content
The total phenolic content (TPC) was measured according to the Folin-Ciocalteau procedure [62]. Samples (0.25 and 0.125 mg) were mixed with Na 2 CO 3 (2.25 mL; 7.5% w/v) and Folin-Ciocalteu reagent (0.25 mL) and allowed to stand for 3 h at room temperature. The absorbance was read at 765 nm using a Synergy spectrophotometer (Biotek, Winooski, VT, USA). Data were expressed as milligrams of gallic acid equivalents (GAEs) per g of extract. To this purpose, a gallic acid calibration curve (y = 0.0247x − 0.0063; R 2 = 0.9998) was built up in the range 0.78-25 µg/mL (final concentration levels).

Determination of Total Flavonoid Content
The total flavonoid content (TFC) was determined, with NaNO 2 (5%, w/v; 0.3 mL) previously solubilized into 5 mL of distillate water being added to the samples. After 10 min, AlCl 3 solution (10%, w/v; 0.6 mL) was added. The reaction was carried out for 6 min. Then, NaOH aqueous solution (1.0 M, 2.0 mL) was added, and the mixture was further diluted to 10 mL with distillate water. The absorbance was read at 510 nm against the blank (water) using a Synergy spectrophotometer (Biotek, Winooski, VT, USA). The flavonoid content was expressed as milligrams of quercetin equivalents (QUEs) per g of extract. To this purpose, a quercetin calibration curve (y = 0.0243x − 0.0038; R 2 = 0.9978) was built up in the range 0.78-50 µg/mL (final concentration levels).

In Vitro Fermentation
The alcoholic extract of C. sativa Mill. (Cs/1/1), fractions Cs/2/1 and Cs/3/2, and a standard diet (as control), were tested to evaluate fermentation characteristics and kinetics by means of the in vitro gas production technique [64]. The extracts were incubated for 24 and 120 h at two dosage levels (50 and 200 mg) with the control diet (1.0055 ± 0.0024 g). This latter consisted in corn silage, oat hay, and concentrate (NDF: 44.2% and CP: 13.7%). All the samples were incubated in hermetically closed serum flasks (120 mL each, three replications for each extract and dosage (3 × 3 × 2)) with rumen fluid (10 mL) at 39 • C under anaerobic conditions, and buffered medium (75 mL) and reducing agent (4 mL) were added [65]. The rumen fluid was collected in a pre-warmed thermos at a slaughterhouse, as authorized according to EU legislation [66], from six healthy young bulls (Bos taurus) to increase the variability of the microbial population. All procedures concerning animals were accepted by the Ethical Animal Care and Use Committee of the University of Napoli Federico II (Prot. 2019/0013729 of 08/02/2019). Then, the collected rumen liquor was transferred to the laboratory of the Department of Veterinary Medicine and Animal Production. There, it was pooled, flushing with CO 2 , filtered through a cheesecloth, and added to the flasks. After 24 h of incubation, the gas-phase from each flask was sampled (3 mL) in duplicate with a gastight syringe and injected into a gas chromatograph (ThermoQuest 8000top Italia SpA, Rodano, Milan, Italy), equipped with a loop TC detector and a packed column (HaySepQ SUPELCO, 3/16-inch, 80/100 mesh) to detect the methane (CH 4 ) production related to incubate organic matter (CH 4 , mL/g iOM) and degraded organic matter (CH 4 (mL/g dOM). The gas produced during the 120 h of incubation was reordered using a manual pressure transducer (Cole and Palmer Instrument Co, Vernon Hills, IL, USA) and related to incubated OM (OMCV, mL/g). After 120 h of incubation, the fermentation was stopped and the pH of the fermentation liquor was measured using a pH meter (ThermoOrion 720 A+, Fort Collins, CO, USA). Then, the organic matter degradability (OMD, %) at both the incubation times was assessed by the weight differences of the incubated OM and the undegraded filtered (sintered glass crucibles; Schott Duran, Mainz, Germany, porosity # 2) residue burned at 550 • C overnight [67].

Data Processing and Statistical Analysis
Colorimetric tests were carried out in three replicate measurements for three samples (n = 3) of the extracts (in total, 3 × 3 measurements). All data were expressed as mean ± SD values.
To estimate the fermentation kinetic parameters, the gas production profiles were fitted to the sigmoidal model [69]: G is the total gas produced (mL/g of incubated OM) at time t (h), A is the asymptotic gas production (mL/g), B is the time at which one-half of A is reached (h), and C is the curve switch. The maximum fermentation rate (R max , mL/h) and the time at which it occurred (T max , h) were determined using model parameters [70]: Statistical analyses were performed by ANOVA (JMP ® , Version 14 SW, SAS Institute Inc., Cary, NC, USA, 1989-2019). A post-hoc Dunnett's test was performed to observe the differences between the control and experimental diets. The significance level was verified at p < 0.05, p < 0.01, and p < 0.001. A statistical comparison Shapiro-Wilk test for normally distributed data was performed. The correlations between the colorimetric assay values and fermentation parameters were also evaluated using the Pearson correlation coefficient (Tables S3 and S4). Dendrograms of fermentation end-product data were made based on mean values of parameters obtained by three fractions (Cs/1/1, Cs/2/1 and Cs/3/2) at both doses (50 and 200 mg). In addition, principal component analyses (PCA), as well as scatter plots, were carried out for two doses by Origin2015 and GraphPad Prism 8.4.2, respectively.

Conclusions
Chestnut leaf extracts can provide modulation of rumen parameters and can be used as supplements in the livestock diet, thus creating added economic value for the chestnut industry. Chestnut leaves have been shown to be a "reservoir" of structurally different metabolites capable of exerting effects both at 24 and 120 h of incubation, with an increase in total VFAs and propionate and a reduction in BCFA, A/P, acetate, and CH 4 . As for the positive effects on CH 4 production, one of the most important challenges in livestock is to mitigate the production and emission of this greenhouse gas (GHG). Herein, it has been shown that the alcoholic extract and its polar fraction (Cs/3/2) are rich in polyphenols, including flavonoids and tannins (gallo-and ellagitannins). These are known to be considered "anti-nutritional" because they influence the feed palatability in ruminants by exerting negative effects on rumen fermentation. However, the effects of tannins can be variable depending on the sub-class (condensed or hydrolysable), the plant source and the molecular structure, which plays a key role in determining "structure-effect" in any complex biological system. Furthermore, their effects were also modulated by the synergistic action of flavonoids, which are present in a moderate percentage in the alcoholic extract and Cs/3/2 fraction. However, the non-polar fraction (Cs/2/1) also gave positive results, which were attributable to the presence of fatty acids and triterpenes. Therefore, in the light of the data obtained, it is essential to accurately attribute specific effects to an extract consisting of a well-defined class of metabolites, even better if they are pure molecules. The dose-response efficacy of the chestnut extract/fractions and the evaluation of pure isolated compounds are in line with our future perspectives.