Unraveling the Bioactive Potential of Camellia japonica Edible Flowers: Profiling Antioxidant Substances and In Vitro Bioactivity Assessment

In recent years, the search for novel natural-based ingredients by food and related industries has sparked extensive research aimed at discovering new sources of functional molecules. Camellia japonica, traditionally known as an ornamental plant, has gained attention due to its diverse array of bioactive compounds with potential industrial applications. Although C. japonica flowers are edible, their phytochemical profile has not been thoroughly investigated. In this study, a phenolic profile screening through an HPLC–ESI-QQQ-MS/MS approach was applied to C. japonica flower extracts, revealing a total of 36 compounds, including anthocyanins, curcuminoids, dihydrochalcones, dihydroflavonols, flavonols, flavones, hydroxybenzoic acids, hydroxycinnamic acids, isoflavonoids, stilbenes, and tyrosols. Following extract profiling, their bioactivity was assessed by means of in vitro antioxidant, antimicrobial, cytotoxic, and neuroprotective activities. The results showed a multifaceted high correlation of phenolic compounds with all the tested bioactivities according to Pearson’s correlation analysis, unraveling the potential of C. japonica flowers as promising sources of nutraceuticals. Overall, these findings provide insight into the valorization of C. japonica flowers from different unexplored cultivars thus diversifying their industrial outcome.


Introduction
The growing trend in consumer preference for natural-based products, avoiding synthetic additives, has led to a paradigm shift in various industries, including food, pharmaceuticals, and cosmetics.This transformative trend underscores the imperative to explore innovative, natural, and sustainable sources of health-promoting ingredients [1].
Specifically, compounds with antioxidant properties are currently in high demand, sought not only for their efficacy in extending product shelf life but also for their pivotal role in neutralizing free radicals.This functionality is crucial in preventing a spectrum of disorders and diseases, including cancer, cardiovascular diseases, neurodegeneration, hypertension, and diabetes mellitus [2].Phenolic compounds, renowned for their antioxidant Pharmaceuticals 2024, 17, 946 2 of 23 capabilities, frequently manifest additional bioactivities, encompassing anti-inflammatory, antimicrobial, and anti-proliferative effects [2].This diverse spectrum of bioactivities positions phenolic compounds as pivotal molecules with a myriad of applications across diverse industries.Consequently, there has been a surge in investigations over the last decade focusing on evaluating new natural sources of phenolic compounds.
In this context, Camellia genus (family Theaceae) has emerged as a potential alternative.Comprising approximately 280 species native to Asia, this globally distributed genus is considered economically significant.Among their most popular species, Camellia sinensis (tea beverages) and Camellia japonica (ornamental), also called winter rose, stand out.Particularly, over 10,000 cultivars and 3000 hybrids of C. japonica are currently known, with a cultivar of beautiful flowers of different colors and shapes [3,4].The ornamental camellias were brought to Europe and Americas in late 1870s and are now popular flowering and landscaping shrubs in many regions across the world with a mild climate, such as Northwestern Spain [4,5].Therefore, nowadays, they are globally distributed and considered an economically important group of perennial evergreen flowering plants [6].
Regarding the chemical composition of Camellia sp., the most abundant biomolecules are phenolic compounds, terpenoids, fatty acids, and a few key minor compounds that include pigments and biosugars [7].This rich chemical profile, mainly associated with Camellia leaves, has prompted its traditional use as medicine in Asia (Japan, China, Korea) for the treatment of stomachic illness, bleeding, and inflammation [8].Nowadays, this traditional application as therapy has been supported by several studies that have demonstrated the beneficial properties of Camellia sp., such as antioxidant, anti-inflammatory, or antimicrobial properties, among others [7,9].
Therefore, this species may represent not only a source of antioxidant phenolic compounds, but also a source of compounds with antimicrobial activity.This is of great interest since the antibacterial potential of plant-derived molecules has gained particular relevance in addressing the escalating challenge of antibiotic resistance-a significant concern in public health [10].Antibiotic resistance not only poses a substantial threat to public health but has also been suggested to impede advancements in cancer treatment, contributing to the escalating incidence of cancer cases worldwide [11,12].Within this context, the antiproliferative capacity of bioactive compounds has emerged as a promising attribute, offering alternative avenues for treatment [13].
Despite these advancements, the flowers of C. japonica, the most representative species in the study area, lack studies analyzing the use of this raw material.While information on the bright and large flowers of this genus is scarce compared to leaves and seeds, they present an attractive option for research, given their ability to synthesize phenolic compounds and other bioactive secondary metabolites [14].Camellia flowers are considered to possess astringent, antihemorrhagic, hemostatic, and salve properties [15].Despite these health-enhancing bioactivities, camellias are still considered an underexploited raw material, especially regarding their flowers.Therefore, greater efforts are needed to fully achieve their chemical and bioactive characterization.
In this work, extracts of eight cultivar of C. japonica flowers obtained by an easy and profitable extraction technique (heat assisted extraction) have been analyzed to determine their chemical profile and assess their bioactivities with the aim of providing the basis for their further exploitation as sustainable sources natural ingredients.
The high presence of hydroxybenzoic acids contrasts with the previous results from experiments performed with C. japonica flower extracts [16,17], which supports the idea that C. japonica can be a promising source of this type of compound as reported by Pereira et al. (2023) [18].Regarding the class of tyrosols, the most abundant molecule found in this study was oleuropein, with concentrations ranging from 10 to 79 µg per 100 g of extract.In prior investigations, the concentration of tyrosols accounted for 1.1% of the peak area in C. japonica petal wine, designating them as samples rich in potential antioxidative volatiles [19].Additionally, it was noted that such tyrosol-rich extracts could potentially exhibit utility in mitigating diabetes [19].Furthermore, it has been demonstrated that tyrosol-rich extracts obtained from camellias exhibited a superior antioxidant effect in camellia oils compared to exogenous reference antioxidants such as caffeic acid [20].Therefore, the presence of these compounds in camellia extracts is of great interest, as they not only hold the potential for bioactivities but also contribute to enhanced preservation of the extracts [20].Nevertheless, a scarcity of studies persists in documenting the presence of tyrosol in C. japonica raw materials.
Anthocyanins, which have been previously reported in C. japonica flowers, are responsible for the different colorations of camellia cultivars, as supported by different studies [18,21].Nevertheless, the molecular mechanisms governing flower pigmentation in camellias remain unexplored, and the current breeding program for camellias relies on serendipitous seedling selection.Molecular insights into the regulation of camellia flower traits are essential for the development of an effective and targeted breeding system for the plant [22].This lack of understanding, coupled with the extensive diversity of C. japonica cultivars and their diverse colorations, leads to widely varying results in the available studies on anthocyanin content, both in terms of reported molecules and their concentrations.In this regard, previous studies have identified cyanidin 3-O-glycosides [23], cyanidin 3-galactoside [22,24], cyanidin 3-p-coumaroylglucoside [25] and leucoanthocyanins [26,26] as the primary anthocyanins in C. japonica.In other species of the same genus, such as C. sinensis, cyanidin, malvidin, petunidin, and pelargonidin have also been reported as the main anthocyanins among cultivars with pink flowers [27].These findings align with the results of our study (Table 1).It was observed that Eugenia de Montijo (EM) exhibited the most vibrant pink coloration according to the color chart of the Royal Horticultural Society of London (RHS).This vibrant pink coloration could be attributed to its significantly higher content of anthocyanins, including peonidin, peonidin derivative, malvidin 3-O-(6 ′′ -caffeoyl-glucoside), and cyanidin 3-O-glucosyl-rutinoside, in comparison to the other cultivars studied.Nevertheless, our study also noted a limited presence of pelargonidin glucosides in C. japonica, consistent with the findings of previous authors [23].In all instances, the camellia samples consistently exhibited a noteworthy presence of anthocyanins, comprising a substantial proportion of the total identified compounds (ranging from 4.16% to 17.13% of the mass of the total compounds).This class of bioactive compounds has been recognized for its significant biological properties, including potent antioxidant, anti-tumor, and neuroprotective effects [28].Therefore, this chemical composition could partly justify the biological effects reported in this raw material.In addition to anthocyanins, the flavonoid profile of C. japonica flowers includes the presence of three flavones, including two apigenin glycosides and nobiletin (Table 1).Each sample manifested a distinctive flavone profile, underscoring the chemical diversity among various cultivars.Notably, CT and HA demonstrated elevated concentrations of nobiletin, positioning them as significant contributors of this flavone.EM stood out for its high content in apigenin 7-O-glucuronide.While prior studies conducted on flowers and fruit shells of other C. japonica cultivars indicated only minimal concentrations of flavones [23,29], more recent research involving the cultivars examined in this study reveals substantial levels [18].Among flavonols, the two compounds identified were quercetin-3-O-arabinose and kaempferol 3-O-acetyl-glucoside, both compounds having been previously reported in C. japonica var Eugenia de Montijo flowers [30].The presence of these compounds was previously reported in the leaves of C. japonica [8], with concentrations ranging from 10.4 to 75.1 mg catechin/g dry leaf, depending on the cultivar studied.In these leaves, the predominant compounds are three aglycone flavonoids (quercetin, kaempferol, and apigenin) and two glycosylated flavonoids (rutin and quercitrin) [31].However, it is anticipated that the concentration of flavonoids in C. japonica leaves is higher than in the flowers due to the extensive body of research emphasizing tea leaves (produced from various Camellia sp.) as a rich source of flavonoids [32,33].
Regarding dihydrochalcones, this is a compound family previously documented in camellias [18,34].In the case of these camellia cultivars, dihydrochalcones do not represent a predominant group of compounds, comprising only 0.15-0.80% of the total identified compounds.This observation is consistent with findings from previous studies involving the Camellia species, where dihydrochalcones were also not identified as major compounds [18,35].However, it is noteworthy that one of the identified compounds, phloridzin, which had previously been reported in C. japonica [36,37], holds significance for the commercialization of camellia products.This is because phloridzin is recognized as one of the primary sweet taste compounds, and its concentration plays a vital role in ensuring tea quality.The recognition threshold for phloridzin is 0.01 mg/mL, and it demonstrates synergistic effects when combined with dulcitol, l-alanine, and sucrose [38,39].
Within the dihydroflavonol family, four isomers of dihydroquercetin were successfully isolated.These compounds had previously been identified in extracts from C. japonica var.Eugenia de Montijo (EM) using various extraction techniques, including ultrasounds and microwave [30].Previous research in different Camellia species, such as Camellia nitidissima and C. japonica, regarded this compound as an intermediate in the production of colorinducing compounds like quercetin, cyanidin, and their respective derivatives [40,41].This compound was also associated with high antioxidant, capillary-protective, and antiinflammatory activity [42].Therefore, it could be interesting to conduct an optimization of the extraction to increase phenolic compound yield.
Regarding the disparities observed among the different cultivars investigated in this study, the EM cultivar emerged as particularly noteworthy due to its elevated concentrations of anthocyanins, dihydrochalcones, hydroxybenzoic acids, curcuminoids, hydroxycinnamic acids, and stilbenes.Notably, the content of stilbenes was at least six times higher, while dihydrochalcones and hydroxybenzoic acids were a minimum of three times more abundant in this cultivar compared to others.As a result, this cultivar is a promising candidate for future optimization studies.The CT and HA cultivars are notable for their substantial dihydroflavonol content, which was ten times higher than that of the other cultivars considered in this study.The EM and CT cultivars are the richest in flavonols and flavones, with a content at least three times higher for flavanols.In terms of isoflavonoids, the cultivars with the highest concentration are EM, CT, and HA, boasting a content at least two times greater than that of the other cultivars.Additionally, the CT cultivar stands out for its high tyrosol content, with concentrations reaching up to eight times those found in the other cultivars.These findings underscore the potential of EM, CT, and HA as valuable candidates for future optimization studies.
In addition, for future investigations, it would be valuable to assess the presence of bound phenolic compounds in the samples, also referred to as non-extractable, unextractable, or insoluble phenolics [43].Generally, these compounds are ester-linked to structural cell wall polymers.Liberation of these compounds typically occurs through alkali, acid, or enzymatic treatment of samples prior to extraction [44].However, operations conducted with the extracts, such as extrusion for incorporating the extracts into a food matrix, could also result in the release of these bound phenolics [45].This is of considerable interest, as various studies substantiate that the primary contributor to antioxidant activity in both raw and processed vegetables is in the bound form.Thus, the potential of these plants as antioxidant agents may be even greater, considering that these compounds were not considered in this study (due to the removal of such large compounds by the used filters).Nonetheless, it is important to note that the bioaccessibility of bound phenolics is lower than that of free compounds [45,46].Therefore, understanding the precise composition of bound phenolics is crucial for evaluating their physicochemical properties, nutritional values, potential applications, and for conducting epidemiological and clinical studies addressing their potential health effects.

Antioxidant Capacity
The antioxidant capacity of the different cultivars of C. japonica analyzed shows a significant variability between them (Table 2).The DPPH values of the cultivars under study oscillate between 29.71 and 39.18 µg/mL for Eugenia de Montijo (EM) and Conde de la Torre (CT), respectively.These results are significantly lower than those reported in previous studies involving these specific cultivars, in which values as high as 136.5 µg/mL were reported for the Elegans variegated cultivar [18].Therefore, all samples can be considered potential antioxidant agents.These values are similar to those reported in other studies, with DPPH values of 45 µg/mL in C. japonica flowers [47] and 38.53 µg/mL in C. japonica leaves [8].However, our results are lower than the outcomes obtained from the pink petals of C. japonica extracted with 0.1% of trifluoroacetic acid in ethanol (3.8 µg/mL), while the red and white petals showed higher values (7.31 and 43 µg/mL, respectively) [48].Similar outcomes were obtained for the purified red pigments extracted from C. japonica flowers that showed DPPH scavenging activity comparable to that of the standard of butylated hydroxyanisole (BHA) (4.55 and 4.17 µg/mL, respectively) [49].These values, significantly better than those obtained in this study, may be due to the utilization of the purified pigments in Zhang's study.However, other purified fractions of C. japonica such as polysaccharides still showed similar or even higher values (36-119 µg/mL) than those from the present work [50].Regarding other species of the same genus, such as black tea, C. sinensis, the antioxidant results are two orders of magnitude higher even when different solvents such as the following are used: 9 mg/mL for 50% MeOH or 42 mg/mL for water [51].While an aqueous extract obtained from C. nitidissima leaves and further fractionated using several solvents showed values similar to ours (from 37 to >60 µg/mL) [52].In general terms, the antioxidant potential of C. japonica is higher than other plants recognized as medicinal like Terminalia glaucescens (145.54 µg/mL) [53], Zygophyllum album (96 µg/mL), [54], or Sida rhombifolia (546.1-1222.5 µg/mL) [55].
The ABTS values of the cultivars analyzed ranged from 77.23 to 120.82 µg/mL (CT, Conde de la Torre and TU, Carolyn Tuttle, respectively).According to the previous reports, the C. japonica flower ethanolic extracts had a better ABTS radical scavenging capacity than vitamin C at the same concentration (50 µg/mL) [47].These values were significantly better than those obtained with other parts of C. japonica, such as the roots, which showed ABTS values in the range of 200-2000 µg/mL [56] and the leaves, with an ABTS scavenging activity of 96.10 mg/mL [57].Other species such as C. nitidissima showed values ranging from 15 to 24 µg/mL [52].However, the high variability observed from the experimental conditions of ABTS assay, together with the variability regarding the presentation of the results, hinders the proper comparison of the results.For instance, several works have provided data in different units, such as in the case of Zheng's study where flowers of C. japonica were proven to have an ABTS activity of 165.89 µmol Trolox equivalents/g sample [58], or in the case of Luo's study where different species of the genus Camellia were assessed but the results were provided as percentages (7.79-89.35%)[59].

Antioxidant Activity
Regarding the comparison of CBA, SRSA, OHSA, NOSA, and H 2 O 2 , results are scarcer since these assays seem not to have been widely applied for the study of Camellia sp.Concerning the crocin bleaching inhibition assay (Figure 1), CBA values of the analyzed cultivars ranged from 0.01 to 0.43 min/g dw for Grandiflora Superba (GS) and Hagoromo (HA), respectively.The lower activity values were obtained from the Grandiflora Superba (GS) and Carolyn Tuttle (TU) (Table 2).As this is not a widely used method, its comparison with the results obtained from other camellias or plant matrices has not been possible.The SRSA values of the different cultivars analyzed in this study ranged between 0.02 and 0.11 mg/mL (Figure 2), with significant differences among the samples.EV, DD, AA, and TU showed no significant differences, with their EC 50 values near 0.02, accounting for the best results for SRSA.These outcomes are much lower than the ones reported in previous studies, such as in the case of the fermented C. japonica leaves (IC 50 0.23 mg/mL) [60].
However, purified polysaccharides from the same species provided values within our range (0.043 to 0.067 mg/mL) [50], while the SRSA values for the petals extracted with an acidified ethanol solution were much lower at around 0.003 mg/mL [48].The OHSA values obtained in our work (EC 50 from 0.6 to 1.3 mg/mL) are slightly lower than those reported in previous studies for polysaccharides from C. oleifera, which were in the range of 1.3-3.7 mg/mL [61], or those from C. japonica that ranged from 0.32 to 1.04 mg/mL [50].Only one study showed much lower OHSA results from analyzing colorful petals extracted under acidified conditions, with values ranging between 6.2 and 12.6 µg/mL [48].DT and CT showed the lowest values and had no significant difference among them.The NOSA data obtained in the present work (EC 50 from 0.6 to 2.2 mg/mL) are higher than those reported in previous studies with C. japonica stems, in which a NOSA value of 0.35 mg/mL was detected [60].H 2 O 2 scavenging activity ranged from 0.031 to 0.161 mg/mL for Hagoromo (HA) and Carolyn Tuttle (TU), respectively.These values are in the same range than those obtained with C. sinensis flowers, which had IC 50 values of 6.09-169 µg/mL depending on the solvent used during the extraction [62].In general terms, our results, which are mainly in concordance with the other previously published results, point to the different cultivars of C. japonica flowers as potential sources of antioxidant agents.

Antimicrobial Activity
Among the eight cultivars of C. japonica, Dr Tinsley (DT), Eugenia de Montijo (EM), and Conde de la Torre (CT) were demonstrated to be the most effective ones for the inhibition of bacterial growth (Table 2).These three cultivars could inhibit all bacterial species except for E. coli.However, other studies carried out with C. japonica flowers have proven that some cultivars of C. japonica have antimicrobial activity against E. coli with inhibition halos of 17-18 mm [63].Similar values were reported for C. sinensis (16.66-23.33mm) [64].Less common camellia species, such as C. assamica, have shown lower levels of activity against E. coli (3-4.3 mm) [65].
The strongest inhibiting zones were observed against S. epidermidis, with diameters ranging between 10.9 and 14.0 mm.To the best of our knowledge, these are the first results reported of C. japonica activity against S. epidermidis.The only other record found in the scientific literature was with regard to magnesium oxide nanoparticles (MgO-NPs) synthesized with the leaf extract of C. sinensis (5-25 µg/mL), which exhibited antibacterial activity (10-19 mm inhibition zone) against S. epidermidis [66].
On the contrary, the smallest halos were achieved for B. cereus (5.1-7.4 mm) although this strain was attacked by two additional cultivars, Donation dentada (DD) and Hagoromo (HA), which showed a broader halo of ~9 mm (Table 2).These inhibition halos were lower than those obtained with other camellia species such as C. sinensis, with an inhibition zone of 25.38 mm even when using a lower concentration extract (5 mg/mL) [67].However, this study used an organic solvent, methanol, for extraction which may have recovered biomolecules with different biological properties.In addition, the concentration of biocompounds may vary notably among tea species; indeed, C. sinensis has been suggested to be more effective against Gram-positive bacteria than C. japonica [68].The correlation analysis allowed us to determine a correlation of 0.88 between the activity against this microorganism and the CBA assay (Figure 3).
The only bacterial strain inhibited by all the C. japonica cultivars was P. aeruginosa, with inhibiting halos of around 9-10 mm, except for the Elegans variegated cultivar (EV), whose halo was around 7 mm.Indeed, this cultivar was just capable of inhibiting the growth of P. aeruginosa.In fact, this cultivar, Elegans variegated, displayed no inhibition against the remaining bacterial strains.Within the previously published studies, some cultivars of C. japonica had already shown their effectiveness against P. aeruginosa.A greater activity was described with halos from 13 to 25 mm, which were dependent on the solvent used, with acetone displaying the strongest inhibition [64].The only bacterial strain inhibited by all the C. japonica cultivars was P. aeruginosa, with inhibiting halos of around 9-10 mm, except for the Elegans variegated cultivar (EV), whose halo was around 7 mm.Indeed, this cultivar was just capable of inhibiting the growth of P. aeruginosa.In fact, this cultivar, Elegans variegated, displayed no inhibition against the remaining bacterial strains.Within the previously published studies, some cultivars of C. japonica had already shown their effectiveness against P. aeruginosa.A greater activity was described with halos from 13 to 25 mm, which were dependent on the solvent used, with acetone displaying the strongest inhibition [64].
Among all the analyzed cultivars, only two have been proven to be ineffective against S. aureus, namely Elegans variegated (EV) and Donation dentada (DD).The rest of the cultivars had inhibition zones in the range of 9.37-10.98mm, values slightly lower than those reported by Kim (14-15 mm) [63].This may have been due to the higher percentage of methanol used in Kim's study, since other studies have shown that ethanol extracts tend to exhibit higher antimicrobial activity [60].In another study, the minimum inhibitory concentration (MIC) against S. aureus of young leaf C. japonica extract were revealed to be around 15 µg/mL [69], while for C. sinensis, the MIC was much lower at 0.23 mg/mL [64], which again supports the stronger capacity of C. sinensis against Grampositive bacteria.Among all the analyzed cultivars, only two have been proven to be ineffective against S. aureus, namely Elegans variegated (EV) and Donation dentada (DD).The rest of the cultivars had inhibition zones in the range of 9.37-10.98mm, values slightly lower than those reported by Kim (14-15 mm) [63].This may have been due to the higher percentage of methanol used in Kim's study, since other studies have shown that ethanol extracts tend to exhibit higher antimicrobial activity [60].In another study, the minimum inhibitory concentration (MIC) against S. aureus of young leaf C. japonica extract were revealed to be around 15 µg/mL [69], while for C. sinensis, the MIC was much lower at 0.23 mg/mL [64], which again supports the stronger capacity of C. sinensis against Gram-positive bacteria.

Cytotoxic Activity
The different extracts of C. japonica flowers showed significant inhibitory effects on the proliferation of different cancerous cell lines (AGS, HepG2, A549) (Table 2).The highest cytotoxicity was observed against A549, whose IC 50 never exceeded 22.2 µg/mL, whereas the IC 50 values were 8.69-51.67µg/mL for the AGS cells and 13.37-45.05µg/mL for the HepG2 cells.Data obtained in the previous studies reported similar levels of activity of the C. japonica flower extracts against the AGS cells, with IC 50 values of 14.1-79.4µg/mL, depending on the cultivar analyzed [18].In contrast, all the cultivars had cytotoxic effects on non-tumor Vero cells, with IC 50 values ranging between 6.23 and 18.40 µg/mL.CT (Conde de la Torre) and GS (Grandiflora Superba) showed the lowest toxicity against Vero cells (IC 50 = 51.67 and 68.99 µg/mL, respectively).From the results, six cultivars, Elegans variegated (EV), Dr Tinsley (DT), Eugenia de Montijo (EM), Donation dentada (DD), Hagoromo (HA), and Carolyn Tuttle (TU), showed significantly more cytotoxic activity against the cancer cell lines at different degrees.Cultivar DT (Dr.Tinsley) showed the lowest IC 50 = 1.65 µg/mL and thus had the best cytotoxic activity against A549 cells, improving the capacity of the ellipticine used as positive control.This high activity may have been due to the presence of catechins, a compound present in the Camellia genus that has been shown to have cytotoxicity against A549 cell line, even though units of data hinder their comparison (inhibitory rate of 19.76% at a concentration of 600 µmol/L, 24 h incubation) [70].This activity may also have been due to the high concentration of saponins of the Camellia genus [71].Data obtained from other species of camellias revealed IC 50 values of 64.4-113.3mg GAE/L for C. sinensis [72], and from <10 µM [73] to 18.5 µM [74] for C. oleifera, although the variability of units once again do not allow the comparison of the cytotoxic effect.
To our knowledge, there are no studies that analyze the cytotoxic activity of C. japonica against the other cell lines used in this study.However, C. oleifera [75] and C. ptilophylla (IC 50 292 µg/mL at 72 h) [76] have been proven to be effective against the HepG2 cell line.C. sinensis has no significant activity against HepG2 in comparison with the standard (doxorubicin, IC 50 of 18.8 µM) [77].Therefore, C. japonica seems to have more potential as a source of cytotoxic compounds than the other Camellia species, especially the cultivars DT (Dr Tinsley) and DD (Donation dentada) against the cell line A549.Nevertheless, C. japonica has less activity than the other reference compounds used against the same cell line.
However, it is imperative to persist in this line of research, as recent studies have demonstrated how phenolic compounds in food undergo a series of alterations by the intestinal microflora and enterocytes.This process leads to the presence of molecules in the plasma that possess distinct molecular structures and bioactivity compared to the original molecules present in the consumed foods.This concept holds significant relevance, especially in the context of studies involving cultured cells, particularly liver and lung lines, in which the phenolic compounds were supplemented that are not actually present at the level of these tissues in vivo [78].

Neuroprotective Effects
The data obtained in this study show low neuroprotective activities for the analyzed samples, making it impossible to determine EC 50 (Table 2).However, previous studies have reported that some noroleanane triterpenoids isolated from the fruit peels of C. japonica had potential neuroprotective and anti-inflammatory effects against Aβ-induced neuronal damage [79].Other cultivars of camellias (C.nitidissima) have also been proven to have a neuroprotective effect via synergistically boosting the endogenous antioxidant defenses and neurotrophic signaling pathway.The main active phytochemicals seemed to be catechins, quercetin, kaempferol, and their derivatives [80].Green tea from C. sinensis presented a strong neuroprotective capacity in the Alzheimer-like rat model, avoiding memory deficits and increases in ROS and TBARS levels in the hippocampus.This activity was attributed to the high content of epigallocatechin gallate [81].Therefore, in our case, it could be interesting to purify/isolate the compounds present in the sample to better assess this bioactivity.

Pearson's Correlation Analysis
Additionally, Pearson's correlation analysis was performed to elucidate the concealed patterns and relations between the different bioactivities assessed (Figure 3).According to these results, antimicrobial activity may be due to the presence of high concentrations of antioxidant compounds.This can also be observed in Table 2, in which the camellias with the highest antioxidant capacity are EM (Eugenia de Montijo) and HA (Hagoromo), which are also the ones with the highest inhibition halos.This correlation between antioxidant and antimicrobial activity had already been observed in previous studies with other vegetal matrices, such as Myrtus spp., Eucalyptus oleosa, and rosemary [82,83].

Camellia Sampling and Preparation
Studies were conducted using eight different cultivars of 20-year-old C. japonica flowers, four of them had the white phenotype (Conde de la Torre, Dr Tinsley, Grandiflora Superba, Hagoromo) and four had the pink phenotype (Elegans variegated, Donation dentada, Eugenia de Montijo and Carolyn Tuttle).All of them were manually harvested in NW Spain (42.431 • N, 8.6444 • W) in January 2020 at their optimal state of maturity (stage 5, full bloom), harvesting a total quantity of 400 g of fresh flowers (around 15 flowers for each cultivar).The different flowers were botanically identified by official germplasm banks and identification guide resources, including the American Camellia Society, the International Camellia Register, and previous studies of identification [84].Fresh flowers were immediately frozen at −80 • C to prevent their degradation.Once frozen, they were lyophilized (LyoAlfa 10/15 from Telstar, Thermo Fisher Scientific, Shanghai, China) and pulverized into a fine powder (~63 µm).The resulting powder was mixed to guarantee the homogeneity of the samples and stored at −20 • C until their analysis.To facilitate experimental work and data analysis, each cultivar was named as follows: EV, Elegans variegated; DT, Dr Tinsley; EM, Eugenia de Montijo; CT, Conde de la Torre; DD, Donation dentada; HA, Hagoromo; GS, Grandiflora Superba; and TU, Carolyn Tuttle.

Extraction Procedure
Heat-assisted extraction was carried out following a previously established protocol [10] with some modifications.Briefly, 2 g of lyophilized samples were weighed and placed into dark glass bottles.A total of 50 mL of solvent, a mixture of methanol (MeOH) and distilled water in a proportion of 60:40 (v/v), was then added.The extraction was conducted in a water bath at 50 • C for 1 h using constant magnetic stirring.Subsequently, the samples were collected in Falcon tubes and centrifuged at 4000× g for 15 min at room temperature.Finally, the supernatants were filtered using a 0.2 µm syringe filter and subjected to subsequent analyses.The extracts used for antimicrobial analysis were lyophilized and stored at −80 • C until analysis.

Phytochemical Characterization
Separation and detection of the phenolic compounds present in the EV-TU extracts obtained from the eight cultivars of the C. japonica flowers were performed by highperformance liquid chromatography coupled to a triple quadrupole tandem mass spectrometer via electrospray ionization (HPLC-ESI-QQQ-MS/MS, Thermo Scientific TSQ Quantis) in positive ion mode.HPLC separation was conducted using a Waters Spherisorb S3 ODS-2C 18 column (150 × 3.9 mm, 4 µm particle) coupled to an in-line filter kit.Mobile phases A and B were water and acetonitrile, respectively, acidified with 0.1% formic acid (v/v).The HPLC gradient started with 100% A for 1 min, 95% A for 4 min, 85% A for 5 min, 70% A for 5 min, 40% for 5 min, 20% A for 5 min, and 5% A for 3 min; it was then switched back to 100% A for 2 min to ensure good HPLC column equilibration.The flow rate was 0.350 mL/min, and column temperature was maintained at 35 • C. The injection volume was 500 µL.Tandem mass analysis was performed after optimizing the tube lens and collision energy for each compound separately.On the other hand, the following ESI conditions were automatically adjusted to the set flow: nebulizing (N 2 ) gas flow, 1.5 L/min; curved desolvation line and heat block temperatures, 200 • C; drying gas pressure, 100 kPa; and detector voltage, 1.65 kV.The profiling of phytochemicals was carried out using the comprehensive databases Phenol-Explorer3.6 (version) and MassBank and subsequently integrated with compounds obtained from the literature.The annotation process was conducted with a level two confidence, which entailed putative identification using the isotopic profile of each compound, ensuring a mass accuracy of less than 5 ppm.Data acquisition and HPLC-MS/MS analysis interpretation were conducted by means of Xcalibur software (version 4.1, Thermo Fisher Scientific, Waltham, MA, USA).
Following annotation, the chemical compounds were categorized into various phenolic classes and subclasses and then assessed, employing a semi-quantitative methodology.The quantification of each class was performed using a calibration curve established with pure analytical standards, all of which were of HPLC-grade quality and procured from Sigma-Aldrich (USA).The results were expressed as equivalents in micrograms (µg) of the representative compound per 100 g of dry weight (µg of equivalents/100 g dw).Thus, anthocyanin content was represented as cyanidin equivalents; flavanol content as (+)-catechin equivalents; flavone, isoflavonoids, and dihydrochalcone content as luteolin equivalents; flavonols and dihydroflavonols content as quercetin equivalents; hydroxybenzoic acids as gallic acid equivalents; hydroxycinnamic acids as ferulic acid equivalents; tyrosols and curcuminoids as tyrosol equivalents; and stilbene content as trans-resveratrol equivalents.The determinations were performed in quadruplicate (n = 4).

DPPH Radical Scavenging Activity
This assay was founded on the oxidative stability and free radical release of α,αdiphenyl-β-picrylhydrazyl (DPPH) in the presence of an antioxidant-rich extract [85].The DPPH RSA activity of the extracts of the C. japonica flowers was spectrophotometrically determined at 515 nm, against the radical DPPH as previously described [86].Briefly, a volume of 50 µL of each C. japonica flower extract (serial dilutions using MeOH) was added to a microplate and mixed with 200 µL of DPPH methanol solution (190 mM).Once this mixture was homogenized, the absorbance was quantified (t = 0 min) immediately and after one hour (t = 60 min) using a wavelength of 515 nm in a Synergy HTX multi-mode reader (Bio-Tek, Winooski, VT, USA).The control of this assay consisted of 50 mL of MeOH and 200 µL of DPPH (190 mM).A calibration curve using a methanolic solution of Trolox (5-100 µg/mL) was generated to compare the scavenging activity ability of C. japonica flower extracts using the following Equation (1): where Ac is the absorbance of control at t = 0 and As is the absorbance of sample at t = 60.Absorbance data were expressed as the inhibitory percentage against the extract concentration (mg/mL).Data adjusted to a linear regression allowed for the determination of the half inhibitory concentration (IC 50 ), which indicated the concentration of C. japonica flowers to scavenge half of DPPH radicals.Results were expressed as µg of Trolox equivalent (TE) per mL of extract.The determination was performed in quintuplicate and expressed as the mean ± standard deviation (SD) (n = 5).

ABTS Radical Scavenging Activity
This assay was based on the quantification of the discoloration of the 2,2 ′ -azino-bis(3ethylbenzothiazoline-6-sulfonic acid (ABTS •+ ) cationic radical due to interaction with hydrogen or electron donor species [87].The applied protocol was previously published [86].Briefly, the activated ABTS •+ radical solution was prepared with 5 mL of 3.86 mg/mL of ABTS (Alfa Aesar, Thermo Fisher, Kandel, Germany) and 88 µL of a potassium persulfate solution at 37.8 mg/mL (Carlo Erba Reagents, Milan, Italy).This mixture was incubated overnight at room temperature and then diluted to 1:40 (v/v) using ethanol as solvent.Afterward, 200 µL of this activated ABTS •+ radical solution was mixed with 50 µL of each C. japonica flower extract (serial dilutions using ethanol) into a microplate.The absorbances of the mixtures were then measured spectrophotometrically at 734 nm (Synergy HTX multi-mode reader) at time zero (t = 0 min) and after 1 h of incubation at room temperature (t = 60 min).Ethanol was used as control.Trolox was used for the calibration curve, and the data analysis applied was the same as with DPPH.Thus, the inhibitory half concentration value (IC 50 , µg/mL), which refers to the C. japonica flowers extract concentration required to scavenge half of ABTS •+ radicals, was calculated.The determination was performed in quintuplicate and expressed as the mean ± standard deviation (SD) (n = 5).

Crocin Bleaching Assay (CBA)
This method was based on the use of crocin as an oxidizable substrate and 2,2 ′ -azobis 2-amidinopropane dihydrochloride (AAPH) as a source of free radicals.Through this assay, the competition established between crocin and the antioxidants present in the sample would be assessed.Therefore, the bleaching inhibition of crocin in the presence of C. japonica flowers was detected [88].A solution of crocin (100 µmol/L) (Tokyo Chemical Industry C.O., Tokyo, Japan) and AAPH (7.68 mmol/L) (Acros organics, Morris Planes, NJ, USA) were added into 30 mL of preheated Milli-Q water (40 • C) with 100 mM of phosphate buffer (pH = 5.5), according to the methodology proposed by Prieto [88].Thereafter, 50 µL of each C. japonica flower extract (serially diluted with Mili-Q water) was placed in a microplate with 200 µL of the previously prepared mixture of reagents.The absorbance was measured at 450 nm at time zero (t = 0 min) using a Synergy HTX multi-mode reader.Then, the samples were incubated on an orbital shaker (85/100 rpm) at 37 • C and the absorbance of each sample was regularly measured every 20 min until a total incubation time of 200 min.The control sample was prepared using a mixture of the above-mentioned reagents and Milli-Q water.The obtained data were adjusted to a kinetic profile based on the Weibull mass function, according to Equation (2) as suggested by Prieto [88].
where K is the asymptote, τ is the time when half of the oxidation is achieved, α is the shape parameter associated with the maximum slope of the response (v max ).The assay was performed in quadruplicate and expressed as the mean ± standard deviation (SD) (n = 4).
3.6.Antioxidant Activity Determination 3.6.1.Superoxide Radical Scavenging Activity (SRSA) This assay was based on the generation superoxide radicals (O 2 •− ) by applying nonenzymatic phenazine methosulfate-nicotinamide adenine dinucleotide (PMS/NADH).These superoxide radicals are meant to reduce nitro blue tetrazolium (NBT) into a purple formazan [89].The determination of the superoxide radical scavenging activity was assayed according to a previously published method [90,91], with some modifications.Briefly, 50 µL of each C. japonica flower extract dissolved in phosphate buffer, ranking from 2.0 to 0.03125 mg/mL in concentration, were applied to a microplate.Then, 50 µL of NADH (166 µM), 50 µL of NBT (43 µM), and 50 µL of PMS (2.7 µM) were added to each well.Absorbance was recorded at 560 nm for 2 min using a microplate reader, Synergy HT (BioTek Instruments).Ascorbic acid was used as positive control.The obtained data were adjusted to a kinetic profile based on the Weibull mass function, according to Equation (2), as suggested by Prieto et al. [88].The results were expressed in mg/mL as the mean ± standard deviation (SD) (n = 4).

Hydroxyl Radical Scavenging Activity (OHSA)
Hydroxyl radical scavenging activity was measured by the salicylic acid method [92], with some modifications.Briefly, the lyophilized extract was resuspended in deionized water at a concentration of 8 mg/mL, after which 6 serial dilutions were prepared up to the final concentration of 312.5 µg/mL.Then, 70 µL of this extract was transferred to a microplate and mixed with 70 µL of salicylic acid (9 mM), 70 µL of iron sulphate (9 mM), and 70 µL of hydrogen peroxide (H 2 O 2 , 9 mM).The reaction was carried out at 37 • C for 60 min [93].The absorbance was then determined using a microplate reader, Synergy HT (BioTek Instruments), at 510 nm.For the control group, the extract was replaced with distilled water.The blank sample did not contain H 2 O 2 , whereas ascorbic acid was used as the positive control.The obtained data were adjusted to a kinetic profile based on the Weibull mass function, according to Equation (2), as suggested by Prieto et al. [88].Results were expressed as mg/mL (n = 3).

Nitric Oxide Scavenging Assay (NOSA)
The nitric oxide scavenging assay was evaluated based on a diazotization reaction as described in previous works [90].Briefly, C. japonica flower extracts were dissolved in a potassium phosphate buffer (0.1 M, pH = 7.4) at 20 mg/mL, and six serial dilutions were prepared.A volume of 100 µL of each C. japonica flower extract and 100 µL of sodium nitroprusside (20 mM) was added to each well, and the microplate was incubated at room temperature for 1 h.Then, 100 µL of Griess reagent (1% sulphanilamide and 0.1% naphthylethylenediamine in 2% phosphoric acid) was added to the mixture and left to stand for 10 min.After incubation, the absorbance of the mixtures was measured at 560 nm.The control group was made with potassium phosphate buffer instead of C. japonica flower extracts, while 2% phosphoric acid instead of Griess reagent was added to the blanks.Ascorbic acid was used as a positive control.The obtained data were adjusted to a kinetic profile based on the Weibull mass function, according to Equation (2), as suggested by Prieto et al. [88].Results were expressed as mg/mL (n = 3).

H 2 O 2 Scavenging Assay
The scavenging activity of the C. japonica flower extracts towards H 2 O 2 was assessed by measuring the signal attenuation at 230 nm [94].For that, 75 µL of a 40 mM H 2 O 2 solution was added to 425 µL of sample extract, previously dissolved in a 0.1 M potassium phosphate buffer (pH 7.4), at concentrations ranging from 27.2 to 870 µg/mL.The absorbance was measured after a 10 min incubation [95] using a Shimadzu UV-260 spectrophotometer.Ascorbic acid was employed as the positive control, while the buffer solution was used as the negative control.The scavenging activity of hydrogen peroxide was then calculated as previously explained, by adjusting data to a kinetic profile based on the Weibull mass function, according to Equation (2), as suggested by Prieto et al. [88].Results were expressed as mg/mL (n = 3).

Antibacterial Tests
The determination of the antimicrobial activity was assayed according to the agar disk diffusion method, which consists of the cultivation of bacteria in a Petri dish and the subsequent measurement of inhibition zones produced by the addition of the extracts [96,97].

Microorganisms and Culture Conditions
The antimicrobial activity of the different C. japonica flower extracts was assessed against the following Gram-positive bacterial strains: S. aureus (ATCC 25923), S. epidermidis (NCTC 11047), and B. cereus (ATCC 14579); and the following Gram-negative strains: P. aeruginosa (ATCC 10145), S. enteritidis (ATCC 13076), and E. coli (NCTC 9001).With the exception of S. epidermidis, an opportunistic bacteria [98], the microbes used are some of the most common food poisoning microorganisms [99].The initial number of colony forming units was normalized (0.5 McFarland scale) by measuring the turbidity at 600 nm [100].The lyophilized C. japonica flower extracts were dissolved in DMSO to the final concentration of 20 mg/mL and sterilized by filtration using a 0.22 µm syringe filter.

Agar Diffusion Assay (ADA)
To determine the antimicrobial activity of the C. japonica flower extracts, the agar diffusion method [97] was adopted, with some modifications.Briefly, 50 µL of the previously cultivated inoculum were deposited in a Petri dish containing Mueller-Hinton Agar, where it was seeded and spread by streaking in 4 quadrants with sterile swabs.Next, 15 µL of dimethyl sulfoxide (DMSO) was added as negative control in one quadrant, 15 µL of lactic acid (40%) was added as positive control in another quadrant; and 15 µL of the extract (20 mg/mL in DMSO) was added in yet another quadrant [100].The Petri dishes were incubated at 37 • C for 24 h and the inhibition zone diameters were determined with a digital caliper rule.The assay was conducted in triplicate, and the experimental data were expressed as the mean ± standard deviation (SD) (n = 3).

Cytotoxic Activity
The antiproliferative activity of the C. japonica flowers extracts was evaluated using the Sulforhodamine B (SRB) colorimetric assay, as previously described by Vichai [101].The cytological models used in this study were the Vero, human gastric cancer (AGS), human hepatocellular carcinoma (HepG2), and human lung cancer (A549) cell lines.All these cell lines were provided by Frilabo (Porto, Portugal).
Briefly, 190 µL of cells at a concentration of 50,000 cells/mL were placed in a microplate and treated with 10 µL of each C. japonica flower extract.The starting concentration was 8 mg/mL and reached a minimum concentration of 31.25 µg/mL after several serial dilutions.The negative control consisted of 10 µL of non-supplemented medium.Plates were first incubated at room temperature for 1 h and then for 48 h at 37 • C with 5% CO 2 .Then, 100 µL of cold 10% trichloroacetic acid (Fisher Chemical Reagents, Pittsburgh, PA, USA) was added, and the plate was incubated at 4 • C for 1 h.After a washing step, another 100 µL of sulforhodamine B (0.057% m/v in 1% acetic acid) were added and incubated at room temperature for 30 min.After a final washing and drying process, 200 µL of 10 mM Tris were added and the plate was incubated at room temperature on an orbital shaker for 30 min.Absorbances were measured in a microplate reader at 540 nm.The determination of IC 50 was performed in triplicate and expressed as the mean ± standard deviation (SD) (n = 3).

Neuroprotective Activity
The neuroprotective activity of the samples was evaluated using a previously developed colorimetric method, as described by Ellman [102], with minor modifications.This assay was based on the detection of the inhibition of the activity of the acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE) enzymes, by evaluating the increase in yellow coloration as a result of the production of thiocholine [103].Briefly, extracts were dissolved in Tris-HCl buffer (50 mM, pH = 8), and 25 µL of each mixture was added to the wells, along with 125 µL of 5,50-dithiobis(2-nitrobenzoic acid) reagent (DTNB), 50 µL of buffer B (Tris-HCl buffer + 0.1% albumin), 25 µL of acetylthiocholine iodide/S-butyrylthiocholine iodide (ATCI/BTCI), and 25 µL of 0.44 U/mL of AChE or 0.40 U/mL of BuChE solutions.The slopes were calculated from the kinetic curve obtained from six absorbance measurements at 405 nm, within a total reaction time of 1 min 44 s, after subtracting the blanks (wells containing all reagents except the enzyme).Galantamine was used as a positive control (effective half concentration value, EC 50 = 0.92 g/mL against AChE and EC 50 = 4.92 g/mL against BuChE) and Tris-HCl buffer as a negative control.The inhibitory activity values were calculated in accordance with Equation (3): where AbsC is the mean absorbance per minute of the control, AbsCB is the absorbance per minute of the control blank; and AbsM is the absorbance per minute of the sample.
The determination was performed in triplicate and data were expressed as the mean ± standard deviation (SD) (n = 3).

Statistical Analysis
Data analysis was performed using R software version 4.2.0 (R Development Core Team, 2011).Before conducting parametric tests, the normality of the data distributions was assessed using the Shapiro-Wilk test.Analysis of variance (one-way ANOVA, p < 0.05) was employed to study the matrix effect on antioxidant capacity and activity of samples, provided that the data met the assumptions of normality and homogeneity of variances.Tukey-Kramer comparison tests were then conducted to study the matrix effect on antioxidant capacity and activity of the samples.To analyze the correlations between the different biological capacities and activities assessed, Pearson's correlation analysis was performed using R software.The results were visualized as a correlation matrix in the form of a heatmap.A significance level of α = 0.01 was applied to determine statistically significant correlations.All statistical procedures were carried out ensuring that the assumptions for parametric testing were met, with normality verified prior to ANOVA and correlation analyses.

Conclusions
C. japonica flowers are a source of interesting natural bioactive compounds that could be employed in the development of new industrial applications.Our research was focused on the assessment of the antioxidant, antimicrobial, cytotoxic, anti-inflammatory, and neuroprotective activities of eight different cultivars.As expected, due to their similar composition, not all parameters were affected by the cultivar analyzed, showing a possible correlation between antioxidant capacity and the rest of bioactivities.In terms of antimicrobial activity, no cultivar showed effects against E. coli, with only three cultivars (DT, EM, and CT) showing activity against the rest of the microorganisms assessed.Among the evaluated cultivars, GS stands out for possessing a higher cytotoxic activity, with the HepG2 and A549 activity being approximately two times higher than the AGS activity.None of the cultivars under study had promising neuroprotective effects.Therefore, the flowers of C. japonica can be considered a potential source of various pharmacological compounds against microorganisms or tumors.However, though the outcome of the study and the bibliographic resources seem to be very promising, there are still certain obstacles being faced in developing these extracts into a drug.An interdisciplinary approach of various fields is needed such that a chemical modification strategy or nano-drug delivery mechanisms can be introduced to enhance its activity.

Figure 1 .
Figure 1.Results of crocin assay.(A) Raw experimental signal for the Elegans Variegated cultivarcultivar, with the experimental data fitted to Equation (2) to determine IC 50 values.(B) Compilation of IC 50 values for all cultivars, providing a comparative view of crocin assay outcomes across the different C. japonica flower cultivars.

Figure 1 .
Figure 1.Results of crocin assay.(A) Raw experimental signal for the Elegans Variegate cultivarcultivar, with the experimental data fitted to Equation (2) to determine IC50 values.( Compilation of IC50 values for all cultivars, providing a comparative view of crocin assay outcom across the different C. japonica flower cultivars.

Figure 2 .
Figure 2. Graphs of in vitro antioxidant studies.Column 1 illustrates hydroxyl radical scavenging activity, column 2 showcases H 2 O 2 scavenging activity, column 3 demonstrates superoxide radical scavenging activity, and column 4 displays nitric oxide scavenging activity.Symbols represent experimental data, while lines depict model values.The bar graphs in the final row present the IC 50 values (mg/mL) for each assay across all the cultivars examined in this study.

Table 1 .
Compounds annotated in C. japonica flowers using HPLC-MS-MS in positive mode.Conde de la Torre; DD, Donation dentada; HA, Hagoromo; GS, Grandiflora Superba; TU, Carolyn Tuttle.Note: all values presented in the table exhibit a standard deviation consistently below 5%.

Table 2 .
Bioactive properties of the methanolic extract of the 8 flowers cultivars of C. japonica.