MALDI-HRMS Imaging Maps the Localization of Skyrin, the Precursor of Hypericin, and Pathway Intermediates in Leaves of Hypericum Species

Hypericum perforatum and related species (Hypericaceae) are a reservoir of pharmacologically important secondary metabolites, including the well-known naphthodianthrone hypericin. However, the exact biosynthetic steps in the hypericin biosynthetic pathway, vis-à-vis the essential precursors and their localization in plants, remain unestablished. Recently, we proposed a novel biosynthetic pathway of hypericin, not through emodin and emodin anthrone, but skyrin. However, the localization of skyrin and its precursors in Hypericum plants, as well as the correlation between their spatial distribution with the hypericin pathway intermediates and the produced naphthodianthrones, are not known. Herein, we report the spatial distribution of skyrin and its precursors in leaves of five in vitro cultivated Hypericum plant species concomitant to hypericin, its analogs, as well as its previously proposed precursors emodin and emodin anthrone, using MALDI-HRMS imaging. Firstly, we employed HPLC-HRMS to confirm the presence of skyrin in all analyzed species, namely H. humifusum, H. bupleuroides, H. annulatum, H. tetrapterum, and H. rumeliacum. Thereafter, MALDI-HRMS imaging of the skyrin-containing leaves revealed a species-specific distribution and localization pattern of skyrin. Skyrin is localized in the dark glands in H. humifusum and H. tetrapterum leaves together with hypericin but remains scattered throughout the leaves in H. annulatum, H. bupleuroides, and H. rumeliacum. The distribution and localization of related compounds were also mapped and are discussed concomitant to the incidence of skyrin. Taken together, our study establishes and correlates for the first time, the high spatial distribution of skyrin and its precursors, as well as of hypericin, its analogs, and previously proposed precursors emodin and emodin anthrone in the leaves of Hypericum plants.


Introduction
Secondary metabolites actively participate in a plethora of physiological activities in plants, which includes imparting stress tolerance and accessory functions, unlike primary metabolites [1]. Additionally, these plant-derived specialized metabolites exhibit a wide array of pharmacological activities, which has opened gates to explore plant communities for novel compounds [2]. Hypericaceae is a central, ethnomedicinal plant family, within which Hypericum perforatum has been extensively studied for its bioactive metabolites [3]. H. perforatum, commonly known as St. John's wort, accumulate Figure 1. The proposed biosynthetic pathways of hypericin. The biosynthesis of hypericin and its protoforms using emodin (1) and emodin anthrone (2) as precursors is represented on the left side with green colored arrows. The localization of compounds is represented in blue color according to our previous work [7]. The biosynthesis of skyrin (7) and proposed hypericin (6) production through skyrin (7) as a precursor is represented on the right with violet arrows [17]. Our present work reports for the first time, the occurrence and spatial distribution of skyrin (7) and its precursors in Hypericum leaves concomitant to hypericin (6), its protoforms (3-5), as well as its previously proposed precursors emodin (1) and emodin anthrone (2).
In order to answer the aforementioned questions, we employed a combination of HPLC-HRMS and matrix-assisted laser desorption/ionization high-resolution mass spectrometry imaging (MALDI-HRMS imaging). In the present study, we first confirmed the occurrence of skyrin (7) and its precursors in leaves of five in vitro cultivated Hypericum plant species concomitant to emodin (1), emodin anthrone (2), protohypericin (5), pseudohypericin (4), protopseudohypericin (3), and hypericin (6) by HPLC-HRMS. After that, using MALDI-HRMS imaging, we visualized the distribution and dynamics of skyrin (7) and its precursors in the leaves in high spatial resolution, compared to hypericin (6) and its analogs, as well as its proposed precursors (emodin (1) and emodin anthrone (2)). In particular, both the dorsal and ventral sides of the leaves were mapped with emphasis on the dark glands, where hypericin (6) is localized [7], as well as the tissues surrounding The proposed biosynthetic pathways of hypericin. The biosynthesis of hypericin and its protoforms using emodin (1) and emodin anthrone (2) as precursors is represented on the left side with green colored arrows. The localization of compounds is represented in blue color according to our previous work [7]. The biosynthesis of skyrin (7) and proposed hypericin (6) production through skyrin (7) as a precursor is represented on the right with violet arrows [17]. Our present work reports for the first time, the occurrence and spatial distribution of skyrin (7) and its precursors in Hypericum leaves concomitant to hypericin (6), its protoforms (3-5), as well as its previously proposed precursors emodin (1) and emodin anthrone (2).
In order to answer the aforementioned questions, we employed a combination of HPLC-HRMS and matrix-assisted laser desorption/ionization high-resolution mass spectrometry imaging (MALDI-HRMS imaging). In the present study, we first confirmed the occurrence of skyrin (7) and its precursors in leaves of five in vitro cultivated Hypericum plant species concomitant to emodin (1), emodin anthrone (2), protohypericin (5), pseudohypericin (4), protopseudohypericin (3), and hypericin (6) by HPLC-HRMS. After that, using MALDI-HRMS imaging, we visualized the distribution and dynamics of skyrin (7) and its precursors in the leaves in high spatial resolution, compared to hypericin (6) and its analogs, as well as its proposed precursors (emodin (1) and emodin anthrone (2)). In particular, both the dorsal and ventral sides of the leaves were mapped with emphasis on the dark glands, where hypericin (6) is localized [7], as well as the tissues surrounding the glands. Our study unravels for the first time, the occurrence, distribution, and dynamics of skyrin (7) and its precursors, versus the accumulation of hypericin (6), its analogs, and its possible precursors in the leaves of Hypericum plants.
Emodin (1) was detected in four samples except in H. bupleuroides (<LOD). Our results were in accordance with previous reports where emodin was not found in the Hypericum species mentioned above [24,25]. Conversely, emodin anthrone (2) was typically either not detected or in low abundance in the samples where emodin (1) was detected in higher abundance. Emodin (1) and emodin anthrone (2) condenses to form an unstable compound called emodin dianthrone, which is converted to hypericin (6) in two subsequent steps ( Figure 1). However, it has been previously reported that emodin (1) levels in leaves do not correlate with hypericin (6) accumulation [12], and the spatial distribution of emodin (1) is not restricted to dark glands, unlike to that of hypericin (6) [7]. Hence, our observed dissimilarities in the levels of these two compounds could be attributed to the fact that flux is directed towards the production of hypericin (6) and its analogs, pseudohypericin (4) and protopseudohypericin (3) ( Table S1).
Hypericin (6) was detected in all the analyzed samples except in H. bupleuroides in which skyrin (7) was produced. Further, emodin (1) was not detected in H. bupleuroides, although emodin anthrone (2) was detected. On the one hand, the presence of skyrin (7) coupled to the absence of hypericin (6) or its protoforms in H. bupleuroides lends possible hints that the biosynthesis of hypericin (6) through emodin (1) and related intermediates might be species-specific (Table S1). On the other hand, it is possible that skyrin production occurs in the vegetative stage, followed by its utilization for the biosynthesis of hypericin (final product of the pathway) during floral development and generation of dark glands.

MALDI-HRMS Imaging Reveals That Skyrin Is Localized in the Dark Glands in H. humifusum and H. tetrapterum
Selective metabolic profiling using HPLC-HRMS revealed the presence of skyrin (7) in the five species of Hypericum under investigation. Therefore, we used MALDI-HRMS imaging to map both the dorsal and ventral leaf surfaces, in high spatial resolution and with minimum sample preparation, the localization of skyrin (7) vis-à-vis its precursors, hypericin (6), emodin (1), emodin anthrone (2), pseudohypericin (4), protopseudohypericin (3), and protohypericin (5). While hypericin (6) and its analogs are synthesized and accumulated in the dark glands [7,8,26], hyperforin is present in translucent glands [27,28]. We analyzed both the dorsal and ventral sides of the leaves, and our critical focus was on the dark glands where hypericin (6) accumulates. Primarily, our emphasis was to determine the spatial distribution of skyrin (7) and its precursors. Localization of these compounds could help to understand whether skyrin (7) and its metabolic precursors are explicitly localized to the dark glands along with hypericin (6) or distributed throughout the leaves. Furthermore, HRMS 2 was also performed during MALDI-HRMS imaging on the selected portion of leaves, both from the dorsal and ventral sides, to reconfirm the identity of compounds in the imaging mode following our previously established method [7,17]. We segregated and interpreted the results centering primarily on the distribution of skyrin (7).

MALDI-HRMS Imaging Reveals A Scattered Distribution of Skyrin in H. annulatum, H. bupleuroides, and H. rumeliacum
Dark glands of H. annulatum are localized on the ventral side of the leaf along with numerous leaf hairs rather than the dorsal side ( Figure 5A). As anticipated, hypericin (6) was found to accumulate in the dark glands with higher abundances in dark glands at the leaf margins ( Figure  5A). Strikingly, the distribution and localization of skyrin (7) were entirely dissimilar to what was observed in H. humifusum and H. tetrapterum plants. A typically scattered pattern of distribution of skyrin (7) was observed near or around the dark glands; however, skyrin (7) did not localize in the dark glands ( Figure 5A). Interestingly, skyrin-6-O-β-glucopyranoside (9) accumulated in the dark glands similar to its distribution observed in H. tetrapterum (Figure 6; ventral leaf surface only, not dorsal). Emodin (1) displayed a similar pattern distribution as that of skyrin (7) ( Figure 5A). It is wellknown that in plants, secondary metabolites synthesized at a particular site are distributed across different parts of the tissues during adverse conditions such as biotic and abiotic stresses [29,30]. Hence, it might be possible that in these Hypericum species, skyrin (7) was produced in the dark glands and later translocated into the surrounding leaf tissues. Emodin anthrone (2) also exhibited a scattered distribution pattern around the dark glands, similar to emodin (1) (Figures 5A and 6). Furthermore, the analogs of hypericin (6), namely pseudohypericin (4), protopseudohypericin (3), and protohypericin (5), were all found to localize in the dark glands, corroborating the previous results [7,8,24] (Figure 6). Dark glands of H. annulatum are localized on the ventral side of the leaf along with numerous leaf hairs rather than the dorsal side ( Figure 5A). As anticipated, hypericin (6) was found to accumulate in the dark glands with higher abundances in dark glands at the leaf margins ( Figure 5A). Strikingly, the distribution and localization of skyrin (7) were entirely dissimilar to what was observed in H. humifusum and H. tetrapterum plants. A typically scattered pattern of distribution of skyrin (7) was observed near or around the dark glands; however, skyrin (7) did not localize in the dark glands ( Figure 5A). Interestingly, skyrin-6-O-β-glucopyranoside (9) accumulated in the dark glands similar to its distribution observed in H. tetrapterum (Figure 6; ventral leaf surface only, not dorsal). Emodin (1) displayed a similar pattern distribution as that of skyrin (7) ( Figure 5A). It is well-known that in plants, secondary metabolites synthesized at a particular site are distributed across different parts of the tissues during adverse conditions such as biotic and abiotic stresses [29,30]. Hence, it might be possible that in these Hypericum species, skyrin (7) was produced in the dark glands and later translocated into the surrounding leaf tissues. Emodin anthrone (2) also exhibited a scattered distribution pattern around the dark glands, similar to emodin (1) (Figures 5A and 6). Furthermore, the analogs of hypericin (6), Surprisingly, HPLC-HRMS analysis of H. bupleuroides revealed the absence of all target compounds except emodin anthrone (2) and skyrin (7).  reported the absence of emodin (1), hypericin (6), and its analogs in H. bupleuroides [26]. Besides, our present observations were in partial agreement with our earlier results [17] in which emodin (1) was detected, whereas emodin anthrone (2) could not be detected. In our MALDI-HRMS imaging analyses, we were not able to detect intensities of compounds on the ventral side of the leaf (<LOD), except in the leaf veins where emodin (1) was observed ( Figure 5B). Whereas, when imaging from the dorsal side, skyrin (7), emodin (1), and hypericin (6) were observed in low intensities ( Figures 5B and 7).

Plant Material and Growth Conditions
For the experiments, 5 different in vitro grown Hypericum species in the vegetative stage of development were used. The stock cultures of H. humifusum L., H. bupleuroides Stef., H. annulatum Moris L., H. tetrapterum Fr., and H. rumeliacum Boiss. were derived from seeds obtained through the Index Seminum exchange program and characterized by DNA barcoding [12,Bruňáková et al. unpublished]. The shoot cultures were cultivated in solid MS media (Duchefa Biochemie, Haarlem, Netherlands) containing a 4.4 g L −1 salt mixture according to Murashige and Skoog [31] with Gamborg's B5 vitamins [32], 30 g L −1 sucrose (CentralChem, Banská Bystrica, Slovakia), 7 g L −1 agar (REMI M. B., Proseč nad Nisou, Czech Republic), and 2 mg L −1 glycine with pH adjusted to 5.65 before autoclaving. The cultures were grown at 23 ± 2 °C temperature under 16/8 h photoperiod at 90 μmol m −2 s −1 artificial irradiance. The subculture interval was 5 to 6 weeks.

Extraction of Metabolites From Leaves
The extraction of aboveground tissues of Hypericum species was performed according to our previously established procedures [7].

Plant Material and Growth Conditions
For the experiments, 5 different in vitro grown Hypericum species in the vegetative stage of development were used. The stock cultures of H. humifusum L., H. bupleuroides Stef., H. annulatum Moris L., H. tetrapterum Fr., and H. rumeliacum Boiss. were derived from seeds obtained through the Index Seminum exchange program and characterized by DNA barcoding [12,Bruňáková et al. unpublished]. The shoot cultures were cultivated in solid MS media (Duchefa Biochemie, Haarlem, Netherlands) containing a 4.4 g L −1 salt mixture according to Murashige and Skoog [31] with Gamborg's B5 vitamins [32], 30 g L −1 sucrose (CentralChem, Banská Bystrica, Slovakia), 7 g L −1 agar (REMI M. B., Proseč nad Nisou, Czech Republic), and 2 mg L −1 glycine with pH adjusted to 5.65 before autoclaving. The cultures were grown at 23 ± 2 • C temperature under 16/8 h photoperiod at 90 µmol m −2 s −1 artificial irradiance. The subculture interval was 5 to 6 weeks.

Extraction of Metabolites From Leaves
The extraction of aboveground tissues of Hypericum species was performed according to our previously established procedures [7].

Sample Preparation for MALDI-HRMS Imaging
Fresh leaves were harvested from healthy plants and subjected to sample preparation. In each case, the second set of 2 leaves from the shoot apex was harvested and used for analysis from both the ventral and dorsal sides. Leaves were fixed on glass slides using adhesive tapes. The samples were sprayed uniformly with matrix HCCA (alpha-cyano-4-hydroxycinnamic acid; 7 mg/mL) prepared in a 1:1 ratio of acetonitrile and distilled water with 0.1% FA. A SMALDI Prep spray device (TransMIT GmbH, Giessen, Germany) was utilized for matrix spraying. Before proceeding with MALDI-HRMS imaging, a photographic image was taken for each sample using a specialized digital microscope (VHX-5000, Keyence Deutschland GMBH, Neu-Isenburg, Germany) to evaluate the measured area and record the optical image.

MALDI-HRMS Imaging
MALDI-HRMS imaging experiments of the leaf samples were carried out with an atmospheric pressure scanning microprobe matrix-assisted laser desorption/ionization source (AP-SMALDI; TransMIT GmbH, Giessen, Germany) coupled with a Q-Exactive high-resolution mass spectrometer (Thermo Scientific Inc., Bremen, Germany). The parameters used were according to our previously established protocol [7], with slight modifications. A 60 Hz pulsed N 2 laser MNL 100 series (LTB Lasertechnik GmbH, Berlin, Germany) was used for the UV beam generation at 337.1 nm. The resolution of measurement was adjusted to 10-15 µm, and measurements were made in full scan negative ion mode at m/z 100-800 mass range with an internal lock mass correction utilizing m/z 333.08808, corresponding to the HCCA matrix ion signal [2M−H−CO 2 ] − . Furthermore, measurements were performed with a mass resolution of 140,000 at m/z 200, and the source spray voltage was set at 3000 V. For HRMS 2 measurements in the imaging mode, the isolation width of m/z 1.5 and collision energy of 50 eV was used. HRMS 2 measurements for the skyrin (7) were recorded within a mass range of m/z 500-800 in the negative-ion mode. Processing of data and mapping of mass pixels of the target compounds was done with the software package ImageQuest (v. 1.1.0; Thermo Fisher Scientific, Bremen, Germany). Ion images were generated with a bin width of ±2.0 ppm for full scans. The mass pixels are shown color-coded (Figures 2-9), starting with blue, indicating lower intensities and red, indicating the highest intensities.
Besides, the present results lend a scientific handle to support further that skyrin (7) is an immediate precursor of hypericin (6) (Figure 1) due to their typical colocalization in the dark glands. Emodin (1) accumulation in all Hypericum species irrespective of hypericin (6) production supports the possible role of skyrin (7) in hypericin (6) biosynthesis. Skyrin (7) and its precursors are not abundantly available in plants, but these compounds are produced by different classes of endophytic filamentous fungi [34]. It could be possible that Hypericum plant-associated endophytes produce these metabolites in planta and contribute to hypericin production, given that endophytes are known to produce secondary metabolites found in their host plants [23,35]. Besides, native endophytes might have acquired the skyrin (7) producing gene machinery through horizontal gene transfer in the course of co-evolution with Hypericum host plants. It would be interesting to identify candidate genes responsible for converting skyrin (7) to hypericin (6), and their biological validation would bring more light into understanding the final steps in the biosynthetic pathway of hypericin (6).
Supplementary Materials: The followings are available online. Table S1. The phytochemical composition of leaves of the five Hypericum species under study by HPLC-HRMS.