Evolutionary Aspects of Hypericin Productivity and Endogenous Phytohormone Pools Evidenced in Hypericum Species In Vitro Culture Model

Shoot cultures of hypericin non-producing H. calycinum L. (primitive Ascyreia section), hypericin-producing H. perforatum L., H. tetrapterum Fries (section Hypericum) and H. richeri Vill. (the evolutionarily most advanced section Drosocarpium in our study) were developed and investigated for their growth, development, hypericin content and endogenous phytohormone levels. Hypericins in wild-growing H. richeri significantly exceeded those in H. perforatum and H. tetrapterum. H. richeri also had the highest hypericin productivity in vitro in medium supplemented with 0.2 mg/L N6-benzyladenine and 0.1 mg/L indole-3-butyric acid and H. tetrapterum—the lowest one in all media modifications. In shoot culture conditions, the evolutionarily oldest H. calycinum had the highest content of salicylic acid and total jasmonates in some of its treatments, as well as dominance of the storage form of abscisic acid (ABA-glucose ester) and lowest cytokinin ribosides and cytokinin O-glucosides as compared with the other three species. In addition, the evolutionarily youngest H. richeri was characterized by the highest total amount of cytokinin ribosides. Thus, both evolutionary development and the hypericin production capacity seemed to interact closely with the physiological parameters of the plant organism, such as endogenous phytohormones, leading to the possible hypothesis that hypericin productivity may have arisen in the evolution of Hypericum as a means to adapt to environmental changes.


Introduction
Hypericum perforatum L. is one of the most widely used and studied medicinal plant species, with origins dating back to the first century AD [1]. The pharmacologically used aerial parts of the plant (Herba hyperici) are characterized by containing a remarkably broad spectrum of phytochemical constituents such as polyphenolic compounds, flavonoids, naphthodianthrones, phloroglucinols and terpenes [2][3][4]. This is the prerequisite for the wide range of pharmacological activities characteristic of preparations of this species, such as antidepressive, antitumor, antiviral and antibiotic [5][6][7]. In addition to the well-known H. perforatum L., the genus Hypericum includes over 484 species distributed worldwide and classified into 36 sections. Either naturally occurring or introduced Hypericum species are found on all continents of the world except Antarctica, with habitus ranging from trees to herbaceous plants [8].
Hypericin and pseudohypericin are condensed naphthodianthrones found only in the genus Hypericum. They were first isolated from H. perforatum about 80 years ago, and

Effect of Plant Growth Regulators on In Vitro Culture Development
The translucent glands characteristic of essential oil and phloroglucinol accumulation in hypericin-non-producing Hypericum species were clearly visible in H. calycinum (Figures 1 and 2B), whereas the dark hypericin glands associated with hypericin accumulation were evident in the other three Hypericum species (see Figures 1 and 2H,J,K,M,N). Cultivation of H. richeri for a period longer than 4 weeks in medium without plant growth regulators (PGRs) was not possible, because no multiplication occurred and explants gradually decayed during the cultivation period of the entire experiment ( Figure 2K). The addition of PGRs strongly stimulated biomass formation compared to the PGR-free control in all four species of Hypericum shoot cultures.   Estimation of the FW/DW ratio of the samples as a result of the different treatments was used as a simple physical characterization of the extent of water retention in plant tissues. H. calycinum showed the lowest tissue hydricity in M_0 (PGR-free) and M_3 (0.1 mg/L BA + 0.2 mg/L IBA) treatments compared with the other three Hypericum species (Figure 3). The highest plant tissue hydration for all PGR-treated samples was generally observed in H. tetrapterum compared to the other three species (Figure 3). H. richeri and H. perforatum showed a comparable ability to accumulate water in plant tissues, with the lowest values for this parameter observed in the M_3 treatment of H. richeri. Water retention in plant tissues is an important physiological parameter in plant cell tissue and organ culture development. Foliar water content (estimated as the difference between fresh and dry leaf weights) does not proportionally relate to the leaf dry weight during growth, and has been shown to play an important role in photosynthesis. Studies have shown that dependencies obtained upon scaling between fresh and dry weight are species specific [25]. This feature was also confirmed by our experimental results outlining H. calycinum as the species with the lowest and H. tetrapterum with the highest water retention capacity of the plant tissues.  The morphological observations of the translucent and dark glands in the four Hypericum in vitro models in the present work confirm the literature data on their morphological structures and relationship with the secondary metabolites produced. Previous studies have shown that the phloroglucinol hyperforin is biosynthesized in the translucent glands and their surrounding cells, with its isoprenoid moiety synthesized via the biogenetic pathway of monoterpenoids, sharing the same constituents with the essential oils also present in the translucent glands [26,27]. The site of hyperforin biogenesis is thought to be the chloroplasts of the cells surrounding the translucent glands [28].
Although it is widely accepted that hypericin accumulation is undoubtedly related to dark glands in Hypericum tissues [9,20], its biosynthetic pathway is still unknown. A recent work by Pradeep and Franklin [29] shed more light on the matter by conducting an experiment to reduce the number and size of dark glands in H. perforatum shoot cultures by treatment with a sub-lethal concentration of glyphosate. As a result, the accumulation of hypericins and their precursors (emodin and bisanthrones) was significantly inhibited, and this process was reversible after glyphosate was removed from the medium. Polyketide synthases and phenolic oxidative coupling proteins were downregulated after treatment and upregulated again after glyphosate removal. Interestingly, phenolic compounds were largely unaffected by treatment.

Total Hypericins Content of Field and In Vitro Collected Samples
Wild collections of H. richeri from both Vitosha and Rila mountain wild accessions proved superior to the other two Hypericum species in terms of total hypericin production. The lowest hypericin productivity was found for H. tetrapterum ( Figure 4).

Total Hypericins Content of Field and In Vitro Collected Samples
Wild collections of H. richeri from both Vitosha and Rila mountain wild accessions proved superior to the other two Hypericum species in terms of total hypericin production. The lowest hypericin productivity was found for H. tetrapterum ( Figure 4). When considering the total hypericins content of H. perforatum and H. tetrapterum in vitro, it became clear that in both species, PGR supplements decreased the total hypericins content so that M_0 treatment proved to be optimal for in vitro hypericin production. In general, H. tetrapterum was found to have the lowest in vitro hypericin production potential of all three hypericin-producing species. For H. richeri, the intragenic comparison could only be based on the M_1, M_2 and M_3 treatments because no M_0 sample was icins content so that M_0 treatment proved to be optimal for in vitro hypericin production. In general, H. tetrapterum was found to have the lowest in vitro hypericin production potential of all three hypericin-producing species. For H. richeri, the intragenic comparison could only be based on the M_1, M_2 and M_3 treatments because no M_0 sample was available. When comparing the same PGR treatments of the three species, the highest hypericin production was found in H. richeri in the sample M_2 (0.2 mg/L BA + 0.1 mg/L IBA) ( Figure 5). Plant cell tissue and organ cultures of Hypericum species have been intensively studied in the literature from 1992 with the pioneering work of Zdunek and Alfermann [30] and Čellarova et al. [31] to date. Various in vitro model systems have been investigated, with the vast majority of literature reports on the production of secondary metabolites in tissue cultures focusing primarily on H. perforatum ([1] and references cited therein). Important tools for improving biotechnological hypericin production have been the multiple approaches to biotic and abiotic elicitation in various models of Hypericum plant cell tissue and organ culture ( [32] and references cited therein). Research was also conducted on Hy- HP-H. perforatum L., HR-H. richeri Vill. and HT-H. tetrapterum Fries. All measurements were performed in triplicate and the values are given as mean ± SD. Same letters denote non-significant differences and different letters denote statistically significant differences of the means.
Plant cell tissue and organ cultures of Hypericum species have been intensively studied in the literature from 1992 with the pioneering work of Zdunek and Alfermann [30] anď Cellarova et al. [31] to date. Various in vitro model systems have been investigated, with the vast majority of literature reports on the production of secondary metabolites in tissue cul-tures focusing primarily on H. perforatum ( [1] and references cited therein). Important tools for improving biotechnological hypericin production have been the multiple approaches to biotic and abiotic elicitation in various models of Hypericum plant cell tissue and organ culture ( [32] and references cited therein). Research was also conducted on Hypericum species from other sections of the genus, although on a much smaller scale. The effects of various elicitors, as well as PGRs, on the production of secondary metabolites of H. hirsutum (section XVIII. Taeniocarpium) and H. maculatum (section IX. Hypericum) were studied, and showed that salicylic acid (SA) increased the production of hypericin (7.98-fold) and pseudohypericin (13.58-fold) in H. hirsutum at 50 µM and at 200 µM increased the production of hypericin (2.2-fold) and pseudohypericin (3.94-fold) in H. maculatum [33]. By applying the processes of de-differentiation in vitro, valuable secondary metabolites of lower molecular weight have been obtained. Thus, the callus cultures of H. perforatum have proved to be a promising new source for the production of secondary metabolites with xanthone structure [34]. The potential of representatives of the more highly developed Hypericum sections to produce high amounts of naphthodianthrone derivatives has been discussed in the literature. For example, in a comparative work with 36 Hypericum species, Kitanov [18] found the highest hypericin content in taxa of the sections Drosocarpium, Hypericum, and Thasia, which were the most evolutionarily advanced in the study. It was concluded that hypericin and pseudohypericin play an important role in the intragenic classification of Hypericum, being characteristic of the phylogenetically advanced sections (IX-XV, XXVII) and absent in representatives of the primitive sections (III, V-VIII) of the genus. However, the role of plant origin and environmental factors also appears to correlate with hypericin production. Thus, in a comparative analysis of five samples of H. perforatum (section IX. Hypericum) collected from different geographical locations with four Hypericum species native to the Balkan Peninsula (H. maculatum subsp. immaculatum-section IX. Hypericum, H. olympicum-section X. Olympia, H. richeri subsp. grisebachii and H. barbatum-both section XIII. Drosocarpium), Božin et al. [35] found that samples of H. perforatum showed remarkable variability depending on their origin, with the highest hypericin concentration among the probes being comparable to that of H. barbatum. The samples of H. olympicum had lower hypericin concentrations, and interestingly, no hypericin was detected in the accessions of H. maculatum and H. richeri, studied in the latter work.
The hypericin values determined in the present work for the samples of the wildcollected Hypericum species confirm the above literature data and show that H. richeri, which belongs to the most advanced Drosocarpium section, has the highest capacity. The lowest hypericin levels were found for H. tetrapterum in both in situ and in vitro samples. When comparing the PGR-treated samples of the Hypericum species producing hypericin, the in vitro capacity of H. richeri to produce the highest levels of hypericin was found for the M_2 treatment.

Levels of Abscisic Acid and Its Conjugates In Vitro
As a general observation (with the exception of M_0 and M_3 H. calycinum treatments) the total sum of abscisic acid (ABA) and its glucose ester (ABA-GE) did not show much variation within the same species as a result of the applied PGR treatments ( Figure 6). Apart from the exceptions mentioned above, the overall lowest values were recorded for H. tetrapterum (1906-3579 pmol/g DW) and H. richeri (2557-6034 pmol/g DW) and the highest for H. perforatum (10,158-12,708 pmol/g DW) and H. calycinum (8444-57,096 pmol/g DW) ( Figure 6). Thus, the total concentration of ABA conjugates appears to be similar for H. tetrapetrum and H. richeri (which have a lower accumulation capacity) on the one hand, and H. perforatum and H. calycinum (which have a higher accumulation capacity) on the other. Interestingly, ABA-GE was not detected at all in four H. tetrapterum treatments (Figure S1). In H. richeri, ABA-GE was observed only in the M_2 treatment, where it accounted for about one-third of the total ABA sum.
The widely accepted concept of the role of ABA in plants is the response to biotic and abiotic stresses such as salt, cold, heat, drought and wounding, as well as exposure to A tendency toward general analogic representation between H. perforatum and H. calycinum on one side, and H. richeri and H. tetrapterum on the other was also noted in the allocation of the two ABA forms ( Figure S1). In the most primitive species, H. calycinum, the proportion of ABA-GE ranged from 62% to 83% in the different treatments, and ABA-GE was the dominant conjugate of ABA in all PGR treatments examined. In H. perforatum, the proportion of ABA-GE ranged from 41% to 54%, indicating a tendency toward equilibrium between the two forms studied in the present experiment.
Interestingly, ABA-GE was not detected at all in four H. tetrapterum treatments ( Figure S1). In H. richeri, ABA-GE was observed only in the M_2 treatment, where it accounted for about one-third of the total ABA sum.
The widely accepted concept of the role of ABA in plants is the response to biotic and abiotic stresses such as salt, cold, heat, drought and wounding, as well as exposure to pathogens and the regulation of various physiological processes in the absence of stress, such as dormancy, heterophylly, pollination and leaf senescence ( [36] and references cited therein). The role of ABA in regulating plant cell turgor is expressed by prevention of opening of closed stomata and by closing open stomata, as illustrated by a model of a diurnal monitoring of ABA levels (peak at the beginning of the dark phase to ensure stomata closure and rehydration during the night, and a sharp decrease after the beginning of the light period) [37].
ABA levels in plants have been shown to fluctuate constantly due to its production, inactivation, and transport pathways in response to different developmental stages or to changes occurring in the environment [38]. Studies have shown that ABA is predominantly synthesized de novo in higher plants via the carotenoid pathway. ABA-GE is considered to be an inactivation end product of ABA metabolism, and the majority of this conjugate remains at constant levels in vacuoles and does not decrease in response to stress [39,40]. On the other hand, an increase in the concentration of ABA-GE was detected in the xylem of stressed plants, suggesting that it may also act as a potential stress signal from roots to shoots [41].
In the present experiment, for the lowest hypericin producing H. tetrapterum there was a relation between its highest tissue water retention of all four Hypericum species (its physical estimation expressed as the highest FW/DW ratio, Figure 3) and the lowest total concentrations of ABA conjugates ( Figure 6). Moreover, no detectable ABA inactivation process was observed in this species (a lack of ABA-GE detection in HT, Figure 7). These results could indicate intensive physiologically-active ABA mobilization in the process of water retention. In contrast, in H. calycinum, which has the lowest FW/DW ratios of all four Hypericum species (Figure 3), was found to have higher total ABA pools ( Figure 6) and a clear ABA-GE dominance ( Figure S1), suggesting a less intense ABA mobilization and a process of ongoing ABA inactivation. In the other two Hypericum species, characterized by a higher hypericin productivity, the experiment showed a clear similarity in tissue water retention capacity (Figure 3), which, however, did not correlate with the dynamics of ABA and its inactivation form (ABA-GE) as was the case for HT and HC. Thus, as far as ABA is concerned, hypericin production capacity rather than evolutionary development may be involved in the processes regulating water retention in plant tissues.

Levels of Benzoic and Salicylic Acid In Vitro
Phenolic compounds, BzA and SA, were found in relatively high amounts in the shoot cultures of four Hypericum species. Their total concentrations ranged from 11,823 to 181,761 pmol/g DW (H. perforatum at M_2 and M_0, respectively) ( Figure 7). Observation of total SA and BzA concentrations within species and treatments showed that H. tetrapterum had the least variability as a result of PGR treatments compared with the other three species (Figure 7). In contrast, the highest variability was observed in H. perforatum, with the PGR-free control significantly exceeding the PGR treatments. In general, this parameter does not appear to be related to the evolutionary stage or hypericin productivity of the samples studied. Further elaboration of the individual contents of BzA and SA within the species/PGR treatments showed a significant dominance of BzA over SA in all samples ( Figure S2). The concentration of BzA significantly exceeded that of SA, and the highest BzA levels (99% of the total amount) were measured in H. perforatum at M_0 (179,303 pmol/g DW). H. perforatum exhibited the highest values of SA stimulation as a result of the PGR treatments (from 1% of the total amount at M_0 up to 35% at M_2) compared to the other three Hypericum species. The evolutionarily oldest H. calycinum had the highest proportion of SA in the PGR-free control M_0 (12% of the total amount) compared with the other Hypericum species. The plant phenolic hormone SA, derived from the building blocks of BzA, is produced in plastids and transported to the cytosol. Thus, it is an important endogenous signal for coping with environmental stress as well as for controlling vital processes such as flowering and senescence [42]. The process of SA biogenesis appeared to be strongly stimulated by PGR treatments in the hypericin-producing Hypericum species in the present experiment and constitutively high in the PGR-free control in the most primitive H. calycinum.

Endogenous Levels of Jasmonates In Vitro
Jasmonates are known to influence many physiological processes of plant growth and development, and especially mediating plant responses to biotic and abiotic stresses [43,44]. They have been extensively used for elicitation studies within in vitro cultures [45], and their effect on the production of secondary metabolites in culture systems has been widely described ( [46] and references cited therein).
In the present work, the total pool of jasmonates ranged from 258 pmol/g DW in H. richeri (M_1) to 9521 pmol/g DW in H. calycinum (M_1) (Figure 8). It consisted of jasmonic acid (JA) and its conjugate JA-isoleucine (JA-Ileu), which, however, was much less abundant or even absent (Figure 8). JA-Ileu is known to be a bioactive JA metabolite that is recognized by the jasmonate receptor [47] and its formation thus represents an important step in JA signaling.
When analyzing the data obtained for total jasmonates, i.e., the sum of JA and JA-Ileu, the greatest variability was observed when comparing the PGR-free and PGR-treated samples in H. calycinum, as the treatments significantly increased the values of the studied parameter ( Figure 8). The jasmonate levels of the PGR-treated H. calycinum samples exceeded those of the other three Hypericum species. The lowest variability between treatments was observed in H. perforatum, which displayed intermediate values of total jasmonates. H. richeri and H. tetrapterum had the lowest values for this parameter (Figure 8).
Analysis of the ratios between JA and JA-Ileu showed that in all samples the levels of JA significantly exceeded those of the bioactive JA conjugate ( Figure S3). Although there were no substantial differences in the proportion of JA-Ileu between PGR treatments, it appears to be greatest in H. tetrapetrum and H. calycinum (up to 10 and 11%, respectively) and smallest in H. perforatum (max. 7%).

Contents of Endogenous Cytokinins In Vitro
In our previous work, we determined the total amount of endogenous CKs and compared the pools and ratios between cZand tZ-types in Hypericum in vitro cultures [24]. We detected the highest pool of total CKs in the evolutionarily most advanced H. richeri, while the lowest amounts of total CKs, as well as cZand tZ-types, were found in the most primitive H. calycinum [24]. Therefore, we aimed here to investigate other metabolic forms of CKs (ribosides, Nand O-glucosides and methylthioderivatives) in the described Hypericum in vitro experimental model. [45], and their effect on the production of secondary metabolites in culture systems has been widely described ( [46] and references cited therein).
In the present work, the total pool of jasmonates ranged from 258 pmol/g DW in H. richeri (M_1) to 9521 pmol/g DW in H. calycinum (M_1) (Figure 8). It consisted of jasmonic acid (JA) and its conjugate JA-isoleucine (JA-Ileu), which, however, was much less abundant or even absent (Figure 8). JA-Ileu is known to be a bioactive JA metabolite that is recognized by the jasmonate receptor [47] and its formation thus represents an important step in JA signaling.  The data of total CK ribosides in vitro shown in Figure 9 represent the sums of trans-zeatin riboside (tZR), dihydrozeatin riboside (DZR), cis-zeatin riboside (cZR), N 6 -(∆ 2 -isopentenyl)adenosine (iPR) and 2-methylthio zeatin riboside (MeSZR) analyzed in this work. The content of total CK ribosides in Hypericum in vitro cultures varied from 242 pmol/g DW in H. calycinum (M_1) to 1313 pmol/g DW in H. richeri (M_3). In general, it was observed in all Hypericum species that PGR treatment reduced the content of total CK ribosides compared to the PGR-free control (data for the PGR-free control of H. richeri are not available because the species cannot be maintained long-term in PGR-free medium). Intragenic comparison of the three PGR treatments showed that H. richeri had the highest values of the studied parameter for all treatments compared to the other three species (Figure 9). In contrast, the lowest determined values of the total CK ribosides in the experiment were found in the PGR treatments of H. calycinum. The pool of endogenous CK N-glucosides in Hypericum in vitro cultures consisted only of N9-glucosides, namely trans-zeatin 9-glucoside (tZ9G) and dihydrozeatin 9-glucoside (DZ9G), and ranged from 18 pmol/g DW in H. calycinum (M_2) to 202 pmol/g DW in H. perforatum (M_0). No CK N7-glucosides were detected. The levels of tZ9G and DZ9G were obviously always higher in the PGR-free control samples (for H. perforatum, H. tetrapterum, and H. calycinum) compared to the PGR-treated variants ( Figure 10). Thus, the PGR treatments significantly reduced the pool of CK N-glucosides, but no significant general trend was observed as a function of medium composition among the four species. When the profiles of the individual CK ribosides were elaborated, it was found that MeSZR was the most abundant metabolite in all four Hypericum species ( Figure S4). In general, PGR treatments increased the content of MeSZR compared to controls (in H. perforatum, H. tetrapterum and H. calycinum except M_1). MeSZR had its highest proportion in the total CK pool in the evolutionarily oldest H. calycinum and its lowest in the evolutionarily youngest H. richeri (in both species except M_1). In the evolutionarily most advanced H. richeri, the second most abundant CK riboside was tZR (19-33%) whereas cZR had the lowest proportion (1-2%). In the other three Hypericum species, the second most abundant riboside was iPR, followed by tZR. Remarkably, in the evolutionarily oldest H. calycinum, tZR was detected as a minor component (1-3%).
The pool of endogenous CK N-glucosides in Hypericum in vitro cultures consisted only of N9-glucosides, namely trans-zeatin 9-glucoside (tZ9G) and dihydrozeatin 9-glucoside (DZ9G), and ranged from 18 pmol/g DW in H. calycinum (M_2) to 202 pmol/g DW in H. perforatum (M_0). No CK N7-glucosides were detected. The levels of tZ9G and DZ9G were obviously always higher in the PGR-free control samples (for H. perforatum, H. tetrapterum, and H. calycinum) compared to the PGR-treated variants ( Figure 10). Thus, the PGR treatments significantly reduced the pool of CK N-glucosides, but no significant general trend was observed as a function of medium composition among the four species.  Five metabolites contributed to the CK O-glucoside pool in Hypericum in vitro cultures, namely trans-zeatin O-glucoside (tZOG) plus its riboside (tZROG), dihydrozeatin riboside 9-glucoside (DZROG) and cis-zeatin O-glucoside (cZOG) plus its riboside (cZROG). Their content varied from 62 pmol/g DW in H. calycinum (M_3) to 1651 pmol/g DW in H. perforatum (M_0). When analyzing the data expressing the total sum of CK O-glucosides, the highest values for each species were again obtained for the PGR-free control, with H. perforatum showing a prevalence ahead of the other species ( Figure 10). The evolutionarily oldest H. calycinum had significantly lower values in both the PGR-free and PGR-treated samples compared to all three species.
In analogy with our previous data demonstrating total CK pools in Hypericum in vitro cultures [24], we clearly indicate here that the evolutionary factors may also affect portions of particular CK metabolic forms such as cZ and tZ ribosides, methythioderivatives and O-glucosides in the Hypericum in vitro experimental model. The highest abundance of MeSZR in the total pool of CKs in the evolutionarily oldest H. calycinum and the lowest in the evolutionarily youngest H. richeri is consistent with the reported biosynthesis of CK methylthioderivatives via the tRNA degradation pathway typical of evolutionarily less advanced organisms ( [48] and references cited therein). Also, the striking differences in the levels of cZ and tZ ribosides in Hypericum species of different evolutionary ages are consistent with our previous findings that tZ-type CKs are mainly found in evolutionarily more advanced plants, whereas cZ-types are typical of evolutionarily older cyanobacteria and algae [49], mosses [50], ferns [51] and some less evolved seed plants [52].

Contents of Phenylacetic Acid In Vitro
A non-indole phenolic compound, phenylacetic acid (PAA), was detected in shoot cultures of four Hypericum species in vitro at relatively high concentrations, ranging from 640 pmol/g DW in H. perforatum (M_2) to 33108 pmol/g DW in H. tetrapterum (M_0). When comparing PAA content within species and PGR treatments, H. tetrapterum proved to be the species with the highest values of this parameter, whereas no particular trends were evident in the other species/PGR treatment variants ( Figure 11). PAA is widely distributed in vascular and non-vascular plants, with higher endogenous levels and lower auxin activity compared with the more widely studied natural indole auxin indole-3-acetic acid (IAA). Therefore, it is thought to play a different role than IAA, as the transport and biosynthesis of PAA occur independently of IAA [53,54]. , 11, x FOR PEER REVIEW 18 of 24

Plant Material
The plant material for tissue culture of the four Hypericum species was collected in the respective floristic regions [55]

Plant Material
The plant material for tissue culture of the four Hypericum species was collected in the respective floristic regions [55] Figure 1. In addition, for the phytochemical analyses, material of H. richeri Vill. was also collected from its natural habitat in the Rila Mountain floristic region.

Tissue Culture Initiation
For shoot culture initiation, the stems of plants growing in situ were cut into 2-3 nodules segments, washed in 70% ethanol for 30 s, sterilized in 0.1% HgCl 2 for 5 min, and washed in sterile distilled water in triplicate. The sterilized explants were placed in culture medium containing Murashige & Skoog [56] macro-and microelements, Gamborg [57] vitamins, supplemented with 20 g/L sucrose and 0.5 mg/L N 6 -benzyladenine (BA), at 25 ± 0.2 • C and 16/8 h photoperiod. After induction of axillary shoot growth, sterile stem explants of 3-4 stem nodules were placed in Murashige and Skoog basic medium, supplemented with 30 g/L sucrose, for long-term maintenance. For H. richeri, the addition of 0.2 mg/L BA and 0.1 mg/L indole-3-butyric acid (IBA) to the culture medium was necessary because the plant could not grow and develop in a culture medium without plant growth regulators (PGRs). The subculture period of the stock shoot cultures described in this way was 20 weeks. In addition, the reversal of the cytokinin-auxin ratio was also tested (M_3, 0.1 mg/L BA + 0.2 mg/L IBA). The selection of the specific plant growth regulators applied (namely benzyladenine and indole-3-butyric acid) was based on our previously reported experiment with genus Hypericum [58,59], as well as unpublished experimental in-house laboratory data.

Estimation of Hydricity of the Plant Tissues In Vitro
The fresh biomass of the samples was collected as an average of at least five culture vessels and after obtaining the air dry weight, samples were kept in a desiccator until constant weight was reached. The extent of tissue hydricity was estimated by comparing the ratio of fresh weight to dry weight (FW/DW), as the increase in this parameter characterizes the extent of water accumulation per unit of plant tissue (g).

Extraction of Plant Material
The air-dried plant material of the in vitro cultivated and collected from the wild H. richeri, H. perforatum and H. tetrapterum was kept in a desiccator with silica gel as a desiccant before analysis. To obtain the extracts, the ground plant material was initially macerated in chloroform in order to remove chlorophylls and then with methanol in order to obtain hypericin-enriched extract. Maceration with both solvents was at room temperature for 24 h, followed by ultrasonic extraction at room temperature for 15 min in replicate. The extracts obtained were filtered, combined and concentrated in a rotary vacuum evaporator at 40 • C.

Determination of Total Hypericins
The total hypericin content of was determined spectrophotometrically at 588 nm according to the following formula: (1) A = measured absorbance, m = dry weight of the plant material [g], V = volume of the methanol extract [mL], 870 = specific absorption coefficient of hypericin [60].
The measurements were performed in triplicate and the values are given as the mean of three replicates ± standard deviation.

Endogenous Phytohormone Levels Analyses
Samples of H. richeri, H. perforatum, H. tetrapterum and H. calycinum growing in vitro were collected and averaged from at least five culture vessels for each treatment and immediately subjected to freeze-drying. Then, 7-17 mg of the freeze-dried material was subjected to extraction and purification with the addition of isotope-labeled internal standards as previously described [61]. Using reversed-phase and ion-exchange chromatography, two fractions were separated: fraction A-containing hormones of acidic and neutral character (auxins, abscisic acid [ABA], salicylic acid [SA], benzoic acid [BzA], jasmonates and gibberellins), and fraction B-the hormones of basic character (such as CKs). For quantification of endogenous phytohormones, HPLC (Ultimate 3000, Dionex, Sunnyvale, CA, USA) coupled with a hybrid triple quadrupole/Linear ion trap mass spectrometer (3200 Q TRAP, Applied Biosystems, Waltham, MA, USA) was used, as previously described [62,63]. Apart from the gradient programme, the chromatographic conditions for both fractions were the same and included: HPLC column Kinetex C18 (100 × 3 mm, 2.6 µm, Phenomenex, Torrance, CA, USA), flow rate of 0.3 mL/min, solvent A (50 mM acetic acid in water), solvent B (water), and solvent C (95% acetonitrile and 5% water), where solvent A was kept constant at 5%. Linear gradient of 10% to 45% C for fraction A, or 5% to 25%C for fraction B was run in 7 min, followed by 1 min wash at 95% C, and 5 min equilibration to initial gradient conditions. The abbreviations for the CKs were adopted and modified according to Kamínek et al. [64].

Statistical Analyses
Number of repetitions, plant culture vessels, plant individuals, extractions and measurements are given for each respective method. The means were compared using the t test of unequal variances at p ≤ 0.05. Results are presented as means ± standard errors of the means, unless stated otherwise. Differences were considered statistically significant at p ≤ 0.05. In the figures, statistically nonsignificant means were marked with the same letters.

Conclusions
The relationships between intragenic evolutionary development and the ability to produce hypericin in the genus are a much debated topic in Hypericum research. The present study was motivated by the phytopharmacological value of the condensed naphthodianthrones, hypericin and pseudohypericin, and the potential for their biotechnological production by using other Hypericum representatives of different evolutionary ages in addition to the widely known H. perforatum. It has been shown that the evolutionarily youngest sections of the genus Hypericum have the highest potential for hypericin production. The choice of Hypericum species of evolutionarily more advanced sections (such as Drosocarpium) in the tissue culture protocol could be a prospective tool for enhancing biotechnological delivery of the valuable naphthodianthrones hypericin and pseudohypericin. The distinctive physiological features of Hypericum species belonging to different evolutionary stages is crucial to the development of these processes. Thus, for the highest in hypericin potential productivity, H. richeri, it was shown that PGRs are a necessary condition to obtain in vitro culture biomass formation. On the contrary, the lowest in hypericin productivity, H. tetrapterum, expressed vigorous growth and biomass formation even in media without any PGR treatments. Subtle physiological processes such as tissue water retention, endogenous phytohormone accumulation and conjugation were also shown to interplay with evolutionary level and hypericin productivity of the studied species. The observed relationships between hypericin productivity potential and particular physiological traits suggest that hypericin production may have arisen in the evolutionary development of the genus as an adaptation feature, required by species to adapt to changing conditions in their environment. Data Availability Statement: Data is available from the corresponding co-authors upon request.

Conflicts of Interest:
The authors declare no conflict of interest.