Mevalonate-derived quinonemethide triterpenoid from in vitro roots of Peritassa laevigata and their localization in root tissue by MALDI imaging

Biosynthetic investigation of quinonemethide triterpenoid 22β-hydroxy-maytenin (2) from in vitro root cultures of Peritassa laevigata (Celastraceae) was conducted using 13C-precursor. The mevalonate pathway in P. laevigata is responsible for the synthesis of the quinonemethide triterpenoid scaffold. Moreover, anatomical analysis of P. laevigata roots cultured in vitro and in situ showed the presence of 22β-hydroxy-maytenin (2) and maytenin (1) in the tissues from transverse or longitudinal sections with an intense orange color. MALDI-MS imaging confirmed the distribution of (2) and (1) in the more distal portions of the root cap, the outer cell layers, and near the vascular cylinder of P. laevigata in vitro roots suggesting a role in plant defense against infection by microorganisms as well as in the root exudation processes.

The metabolic engineering of plants has been a relevant biotechnological tool for the production of secondary metabolites 1 . Root cultures can provide an alternative approach for producing important phytochemicals, as well as for understanding their biosynthetic pathways 2 . Camptothecin, vinblastine and ginsenosides are examples of important secondary metabolites stored in roots 3,4 . Hence, roots have been studied to induce and culture in vitro systems, such as adventitious root cultures, that are not infected with Agrobacterium rhizogenes 2 . Adventitious roots are natural and genetically well-established organs, in contrast to cultured plant cell suspensions, and they are useful for biosynthetic investigation as well as for biotechnological applications 5,6 . Cytotoxic triterpenoids that accumulate in adventitious roots are of great interest due to their extensive range of biological activity, especially in their potential effects against human tumor cell lines 7,8 . Recently, a subclass of terpenoids, namely quinonemethide triterpenoids, has been found to display notable antitumor activity against myeloma cell proliferation 9 , hepatocellular carcinoma cells 10 , prostate cancer cells 11 , glioblastoma cells 12 and pancreatic cancer cells 13 . A terpenoid is synthesized in nature through one of two pathways involved in the biosynthesis of isoprene units (IPP), the mevalonate (MVA) and methylerythritol phosphate (MEP) pathways 14 . Both the MVA and MEP pathways are localized in individual cellular compartments: the MVA pathway is located in the cytosol and the MEP pathway in the plastids 15 . The MVA pathway has long been known for triterpene biosynthesis 16 and despite several studies on the MEP pathway, its complex regulation mechanism has not yet been fully clarified 14 . Although quinonemethide triterpenoid biosynthesis has not yet been Biosynthetic origin of quinonemethide triterpenoids from P. laevigata roots cultured in vitro. To observe the biosynthetic pathway of the target quinonemethide triterpenoids, in vitro roots were cultured in Murashige & Skoog medium 18 supplemented with 1-D-13 C-glucose precursor for 30 days. A chloroform extract from fresh root cultures of P. laevigata was then prepared and fractioned by column chromatography to yield 2. The incorporation pattern was determined by quantitative 13 C NMR by comparing the relative intensities of the labeled and non-labeled signals for 2. After 1-13 C-D-glucose metabolism, the 13 C-enrichment pattern of 2 showed that the positions C-1, C-3, C-5, C-7, C-9, C-13, C-15, C-18, C-19, C-22, C-23, C-25, C-26, C-27, C-28 and C-30 (Fig. S2, Table S1 -Supplementary Information) were highly labeled with 13 C (3.1% to 6.3% range). The MVA pathway generates an IPP unit enriched in C-2, C-4 and C-5 while an IPP unit from the MEP pathway is enriched in C-1 and C-5. Obtained data confirm that the IPP building units were biosynthesized exclusively by the MVA pathway since quinonamethide triterpenes are biosynthesized by 6 IPP units, and therefore 18 C-positions would be labeled, however only 16 C-positions labeled were found ( Fig. 2 and Table S1-Supplementary Information). A hypothesis could be that two methyl groups undergo further descarboxylation In situ roots 22β-hydroxy-maytenin (mg.g −1 ) maytenin(mg.g −1 ) 3-year-old 0.47 a 7.76 a 10-year-old 0.54 a 8.54 a

Table 1. Quantification of maytenin (1) and 22β-hydroxy-maytenin (2) in roots from 3-year-old and
10-year-old P. laevigata cultured in situ. Scott-Knott test (p < 0.05). reaction ( Fig. S3- Supplementary Information). The initial precursor chair-chair-chair-boat conformation of 2,3-oxidosqualene undergoes a series of Wagner-Meerwein rearrangements, first hydride migration generating a new cation followed by 1,2-methyl rearrangement 19 . The dammarenyl cation (tertiary cation) then undergoes ring expansion, giving the baccharenyl cation. The baccharenyl cation is converted to a 5-membered ring followed by the formation of the tertiary lupanyl carbocation. Wagner-Meerwein 1,2-methyl rearrangement of the lupanyl cation occurs, leading to an oleanyl cation 19 . The oleanyl cation is converted to friedelin, a key precursor of quinonemethide triterpenes 20 . Regarding friedelin, the hypothetical pathway also involves sequential oxidations, most likely by cytochrome P450 enzymes, which may catalyze more than one oxidation reaction leading to intermediates such as celastrol and maytenin (1). Then, maytenin (1) is converted to 22β -hydroxy-maytenin (2) through one stereospecific hydroxylation at position C-22, and the presence of both is confirmed in root cultures ( Fig. 2  Anatomical studies of P. laevigata roots. It has been reported that quinonemethide triterpenoids are produced and accumulate in roots of several Celastraceae species 17 , and understanding their unexplored biosynthesis is a fundamental step to manipulate their production. In addition, their tissue distribution, which has not yet been investigated, could contribute to identify their functions in the tissue and in root the exudation process 21 , and also to improve biotechnological manipulation. The in vitro root culture showed typical root anatomy (  Moreover, higher ion intensity was observed in the endoderm and outer cell layers (Fig. 3A,B), but they were extensively detectable in tissues such as the cortical parenchyma.

Discussion
Unlike most plant species, the maximum accumulation of quinonemethide triterpenes in P. laevigata in vitro roots preceded biomass enhancement, corroborating the results obtained with Celastraceae species cultured in vitro [30][31][32] . P. laevigata roots cultured in vitro produced 1.3 times more compound 2 compared to roots of 10-year-old plants cultured in situ, although yields of compound 1 was 3.3 fold higher in P. laevigata roots grown in situ. Our data suggest that sub-cultivation of roots could be highly advantageous, considering that roots cultured in vitro for four months produced the same amount of 1 and five times more the amount of 2 compared to the production in roots of 10-year-old plants cultured in situ. There is no statistical significance between the quinonemethide triterpenes amount of 3-year-old and 10-year-old in situ roots. Moreover, the production of 2 in P. laevigata roots cultured in vitro was superior compared to other Celastraceae species, including Peritassa campestres and Maytenus ilicifolia [30][31][32] . The cultivation of differentiated organs, such as root culture obtained from P. laevigata, can significantly improve the accumulation of secondary metabolites often considered cytotoxic, such as quinonemethide triterpenes, which are present in cells specialized for storage of compounds at specific stages of development. Compounds 1 and 2 could be related to protection against microorganisms, justifying their tissue distributions (higher ion intensities) in the outer cell layers and near the vascular cylinder, as they showed significant antimicrobial activity 8,33 , as well as facilitating the exudation processes toward growth medium, as observed in Catharanthus roseus 34 and other species 21 . Roots are able to secrete defense compounds into the rhizosphere and this process is regulated by endogenous and exogenous stimuli. In fact, cap and border cells are involved in the development of roots and these cells act as a defensive barrier of roots protecting the plant against pathogen invasion 21,35 . More recently, it was reported that exudation of the isoflavonoid pisatin and the construction of the root border cell is stimulated in pea when root tips are challenged with a plant pathogen. Besides, exogenous pisatin leads to the upregulation of border cell production in vitro 35 . The production of antimicrobial naftoquinones and epidermal and outer layer cells increased, after the fungal elicitation in Lithospermum erythrorhizon roots 36 . Similar tissue distribution was observed in our study, and these findings together with the effective antimicrobial activity observed for maytenin and 22-β -hydroxy-maytenin 33,37 , suggest there are functions related to exudation and the antimicrobial protection. Altogether, our results regarding the biosynthesis of quinonemethide triterpenoids in P. laevigata have shown that they are constructed by the MVA pathway, and the route is compartmentalized in the cytosol. This study provides the first experimental evidence of quinonemethide triterpenoid biosynthesis using a 13 C-precursor and shows the production and tissue distribution of quinonemethide triterpenoids in P. laevigata roots cultured in vitro using the MALDI imaging mass spectrometry. Obtained data present the possibility of developing large-scale production of quinonemethide triterpenoids, achieving greater levels than that produced by plants in situ, to supply the pharmaceutical industry with anticancer compounds and also to provide support for the manipulation of the biosynthesis of those compounds by biotechnological processes.

Methods
Chemical. The  Cotyledon segments presenting adventitious root formation were transferred to Erlenmeyer flasks containing the same culture medium described above (without Phytagel ® ) and placed on an orbital shaker (90 rpm) in the dark.
In vitro root cultures were subcultured in a 60-day interval. Quinonemethide triterpenes 1 and 2 were quantified by biomass growth curves in the same culture medium used for root induction. For each sample, an initial inoculum of P. laevigata roots (2.00 ± 0.20 g) was placed into an Erlenmeyer flask (250 mL) containing 100 mL of culture medium and kept in growth chamber at 25 ± 2 °C, in the dark, on an orbital shaker under 90 rpm agitation. Sample root collection was performed for 84 days starting at day 0 (cultivation start). Roots cultured in vitro were removed from the culture medium (n = 9), weighed and dried in a circulating air oven for 48 h at 43 °C at seven-day intervals. Roots from 3-year-old (n = 3) and 10-year-old (n = 3) plants grown in situ collected from the field (UNAERP, Ribeirao Preto, SP -Brazil), were dried in circulating air oven for 48 h at 43 °C for seven days.

General procedure for biosynthesis experiments from P. laevigata roots cultured in vitro.
In vitro root cultures (twenty Erlenmeyers containing 3.0 g each) were maintained for a 60-day interval in basal WPM medium supplemented with IBA (19.68 μM), PVP (899.74 μM) and 1-D-13 C-glucose (20 g.L −1 ) and pH adjusted to 6.0. In vitro roots (control) were cultivated with the same culture medium described above using D-glucose. The medium was able to induce younger roots after 2 weeks of culture under a cycle of 16 h light/8 h dark with continuous growth until 35 days.
General procedure for the isolation of quinonemethide triterpenoids (1 and 2). After  HPLC analysis. Dried plant material from roots cultured in vitro and in situ (1 g), was extracted with chloroform for 12 h to obtain the crude extract, and the sample preparation for quantitative HPLC analysis of 1 and 2 was performed as previously described by our group and colaborators 17,31 . HPLC analyses were carried out using the Shimadzu instrument (LC-10-AVP) system. The quantification of compounds 1 and 2 was performed on a Phenomenex -Luna (C-18) column 250 × 4.6 mm, 5 μ particle size, using a isocratic mobile phase, methanol/ water/formic acid (80 : 20 : 0.1, v/v/v) for 20 min., 1.0 mL.min −1 flow rate; detection at 420 nm). NMR analysis. NMR spectra were recorded on a Bruker 400 MHz spectrometer using CDCl 3 as the solvent and internal standard. The relative 13 C enrichments were obtained by comparing the relative intensity of the labeled signal and the natural abundance of quinonemethide triterpenoids. The combination of 1D and 2D NMR experiments allowed complete elucidation consistent with the literature values 38 . All NMR spectra are described in the supplementary material.
Sample preparation. The roots were transversely sectioned in a Leica RM2245 microtome at a thickness of 30 μm, and photomicrographs were obtained with a Leica DM 500 photomicroscope.

MALDI imaging analyses.
The MALDI imaging analyses were performed using a MALDI-TOF/TOF UltrafleXtreme (Bruker Daltonics, Bremen, Germany), equipped with an 1KHz smartbeam II laser and operating in reflectron positive ion mode. The transverse sections of roots were adhered with double-sided tape (3 M Co., USA) to indium tin oxide-coated conductive slides (Bruker Daltonics) for MALDI analysis. The matrix (DHB:CHCA 7:3 (w/w) was prepared at a concentration of 10 mg/mL with the addition of 0.15 mg/mL NaCl using acetonitrile and deionized water (9:1, v/v). The matrix was applied to the tissue by an ImagePrep station, and N 2 flux was used in the entire spraying process. The instrumental conditions employed were as follows: ion source 1 of 25.00 kV, ion source 2 of 22.55 kV, pulsed ion extraction 110 ns, laser frequency 1000 Hz, minimum laser setting and 800 shots. The external calibration was performed using a flavonoid mixture (galangin, rutin, quercetin and isoquercetin). The images were collected at 25 μm spatial resolution in both the x and y directions. The spectra were calibrated internally using matrix ions and a corrected baseline. The tissue analysis of different parts of the roots (differentiating region and root primary structure) were analyzed together using different region of interest to compare the ion intensities and to apply the same processing parameters. The images were normalized, and a logarithm ion intensity scale was applied.
Histochemical analyses. The root anatomy was analyzed in transverse and longitudinal sections stained with safranine/astra blue (safrablau) and lugol to detect starch and total lipids and terpenes into the root tissues, respectively. Lugol solution yielded a blue-black color in the presence of starch, while safrablau solution yielded a blue color in the presence of primary cell walls and a pink color for secondary cell walls (xylematic elements). Starch was also checked using polarized light. Unstained sections were also analyzed and photographed. All the images are presented in the supplementary material.