Metabolite Profiles in Various Plant Organs of Justicia gendarussa Burm.f. and Its in Vitro Cultures

Metabolite profiles of plant organs and their in vitro cultures of Justicia gendarussa have been studied by using Ultra Performance Liquid Chromatography-Quadrupole Time-of-Flight-Mass Spectrometry (UPLC-Qtof-MS). Samples of leaves, stems, roots, and shoot cultures showed similar patterns of metabolites, while samples of root cultures gave very different profiles. Concentrations of secondary metabolites in shoot cultures were relatively low compared to those in the leaves and roots of the plants. The results suggested that secondary metabolites in J. gendarussa were biosynthetized in the leaves, then transported to the roots.

The variations in metabolites of different parts of the plant or its in vitro cultures have been reported in the literature for several plant species. Diosgenin was detected in plantlet and shoot cultures of Costus speciosus, but it could not be found in its callus cultures and root cultures [26]. Concentrations of phenolics, flavonoids, alkaloids, and phytosterols in callus cultures of J. gendarussa were equal or slightly increased compared to the original plants [27]. Ma and Gang reported that different tissues of turmeric possessed different metabolites profiles [28]. Accumulations of alkaloids in different organs of Lycoris chinensis were different [29]. Variations of secondary metabolites were observed in Juniperus communis [30].
For commercial production of herbal drugs, it is essential to determine where secondary metabolites are accumulated within a plant. The objective of the present study was to investigate the profiles of secondary metabolites in different plant organs of J. gendarussa and its in vitro cultures.

Materials and Chemicals
The J. gendarussa used in this work was of Papua origin and was planted at the campus of Airlangga University, Surabaya, Indonesia. This plant was identified by the Department of Pharmacognosy and Phytochemistry, Faculty of Pharmacy, Airlangga University (voucher no. 22/H3.1.5/DT/2013). Three plants (6 months old) derived from a single plant were cultivated in three different pots (plant 1, 2, and 3), and used as samples. Mature, dark green leaves and stems were collected 4-5 internodes from the terminal bud. Roots of 0-15 cm length were collected from the main trunk. Murashige Skoog (MS) media supplemented with 0.1 g·L −1 myo-inositol, 30 g·L −1 sucrose, and different hormone combinations were used for the in vitro cultures; media A: 6 mg·L −1 benzylaminopurine (BAP), media B: 6 mg·L −1 naphthaleneacetic acid (NAA), and media C: 6 mg·L −1 indolebutyric acid (IBA). Cultures were incubated under continuous light in a growth room illuminated with cool white fluorescent tubes (Philip Lifemax Cool Daylight TLD 36W/54-765) (Philip Lighting, Jakarta, Indonesia) at 25 ± 2 • C. The subculturing period was 21 days. Plant parts and in vitro cultures were air-dried (Loss on Drying were 1.1% ± 0.3%, n = 63) and powdered. Table 1 summarizes codes of the samples. Methanol, ethanol, and formic acid (analytical reagent grade) were from Merck (Darmstadt, Germany). Purified water was from Sigma-Aldrich (St. Louis, MO, USA), acetic acid from J.T. Baker (Phillipsburg, NJ, USA), and NaOH from Agilent (Agilent solution for HPCE) (Mulgrave, Victoria, Australia). All samples were filtered through the 0.2 µm Agilent econo filter polyvinylidene difluoride (PVDF) 13 mm.

Preparation of Extracts and Quality Control (QC) Samples
All the samples of leaves, stems, roots and in vitro cultures of J. gendarussa (Table 1) were extracted in triplicate as described before [25]. The QC samples were prepared according to the published method [31].

Instrumentation
Samples were analyzed using the UPLC Dionex Ultimate 3000 RS LC (Dionex Suftron, GmbH, Thermo Fischer Scientific, Germening, Germany) coupled to the QTOF Bruker Maxis Impact HD (Bruker Daltonik, Bremen, Germany), equipped with an Enclosure services interface operating in negative ion mode. It had a mass range of m/z 50-1000, the capillary voltage was 2500 V, dry N 2 gas flow of 8.0 L/min (200 • C), nebulizer pressure 2.0 bars, end plate offset 500 V, collision energy 25 eV, and an acquisition time factor of 1 s.

Data Processing
Automatic time alignment was performed on retention time (RT)-m/z pairs of 0.4 to 20 min. Data were grouped automatically into buckets with RT-m/z pairs of 0.5035 min and m/z 30.3587; the mass range was 200-700 Da with a mass tolerance 0.05 Da, normalized with the sum of bucket values, pareto-scaled, and a bucket filter of 2% as described before [25].
The proposed molecular formula was performed using SmartFormula based on the exact mass and isotopic pattern; the proposed fragmentation of the compound was generated using SmartFormula 3D. Then, the fragmentation pattern of the compounds were generated using MetFrag [32] and Fragmentation Explorer.

Analytical Method Validation
Stability testing and method validation (intra-day variability) were performed by injecting sample SC2A at different times: 0 h, 12 h, 18 h, and 24 h in triplicate. Principal component analysis (PCA) confirmed that the extracts were stable for at least 24 h, and showed acceptable intra-and inter-day variability.
For checking the reliability of the method for each series of experiments, the QC sample was injected three times at the beginning of the analysis, then regularly every 6-7 samples. Coefficient variations (CV) of the data set were evaluated according to the published method [31]. Our data showed >85.75% of the bucket data that showed the CV <30%. PCA models were cross-validated with full cross-validation and showed no outliers. The tight clustering of the QC samples in the PCA analysis showed the reliability of the method.

Results and Discussion
PCA analysis of pairs RT and m/z ( Figure 1) showed definite discrimination of samples leaves (L), roots (R), stems (S), shoot cultures (SC), and samples of root cultures (RC). Samples L, R, and SC were not well-separated. The total explained variant for the three principle components (PC) were 39.1%. Score plots constructed by using more PCs (up to PC 8) showed similar discrimination patterns (the total explained variants PC 1 to PC 8 was 62.4%). The PCA score plot revealed that different combinations of plant growth hormones (media B and C) have relatively no influence on metabolite profiles of the root cultures. PCA score plots showed closeness among the cluster of metabolite profiles of leaves, roots, stems, and shoot cultures, while root cultures produced a very different metabolite profile. These were confirmed by their total ion chromatogram (TIC) patterns ( Figure 2). TICs of leaves, roots, stems, and shoot cultures showed almost similar patterns, but root cultures yielded a distinctive profile. Relative intensities of the metabolites as shown by their TICs were confirmed with their bucket statistic plots. TICs also showed that the concentration of metabolites in the shoot cultures was relatively low compared to the leaves and roots of the plants.
The results suggest that secondary metabolites in J. gendarussa are biosynthesized in the leaves and then transported to the stems and roots. The PCA score plot revealed that different combinations of plant growth hormones (media B and C) have relatively no influence on metabolite profiles of the root cultures. PCA score plots showed closeness among the cluster of metabolite profiles of leaves, roots, stems, and shoot cultures, while root cultures produced a very different metabolite profile. These were confirmed by their total ion chromatogram (TIC) patterns ( Figure 2). TICs of leaves, roots, stems, and shoot cultures showed almost similar patterns, but root cultures yielded a distinctive profile. Relative intensities of the metabolites as shown by their TICs were confirmed with their bucket statistic plots. TICs also showed that the concentration of metabolites in the shoot cultures was relatively low compared to the leaves and roots of the plants.
The results suggest that secondary metabolites in J. gendarussa are biosynthesized in the leaves and then transported to the stems and roots.

Identification of Metabolites
Loading plots (Figure 1) showed that 12 significant metabolites (1-12) affected the clustering of the samples. Proposed metabolites and their fragmentation patterns are shown in Tables 2 and 3 and Figure 3.
Metabolites 9 and 10 gave the highest intensity in leaves, roots, and stem samples. The presence of metabolites a and b in leaves have previously been reported [16,24,25]. Metabolites 3 and 4 were proposed as fatty acids; stearic acid and 9,12-octadecadienoic acid (Z, Z) have been reported to come from the methanolic extract of J. wynaadensis analyzed by gas chromatography-mass spectrometry
Metabolites 9 and 10 gave the highest intensity in leaves, roots, and stem samples. The presence of metabolites a and b in leaves have previously been reported [16,24,25]. Metabolites 3 and 4 were proposed as fatty acids; stearic acid and 9,12-octadecadienoic acid (Z, Z) have been reported to come from the methanolic extract of J. wynaadensis analyzed by gas chromatography-mass spectrometry (GC-MS) [35]. Metabolite 6 was proposed as protoberberine alkaloid; different protoberberine alkaloids were previously isolated from aerial parts of Gendarussa vulgaris Nees (synonym of J. gendarussa) [36].
In conclusion, this present work has shown that metabolite profiles in the roots and leaves of J. gendarussa are almost identical, but the concentrations of metabolites in shoot cultures seemed very low compared to the leaves. Therefore, it is suggested to use leaves of J. gendarussa as the source for herbal drug raw materials; it seems that the application of tissue cultures as an alternative source for herbal drug production of J. gendarussa is not recommended.