Identification of annotated metabolites in the extract of Centella asiatica

Plants contribute to 75% of molecular medicines (MM) either directly or indirectly. Centella asiatica (CA) is being extensively used in experimental and clinical studies. However, its scientific approval is not forthcoming. It is well known that every plant contains useful as well as some harmful compounds. Subjecting whole plant extracts containing harmful compounds to modern pharmacological methods will only demonstrate that they are not safe for use as medicines. To ascertain both the useful and harmful compounds in CA extract, all the compounds of the extract must be identified. In the present study, a methanol extract was prepared from the whole plant. The compounds were identified using liquid chromatography-mass spectrometry with database confirmation. 3,201 compounds were identified using the METLIN database. Database searches yielded 1,187 biological compounds of which 154 were for human/human cell lines. These 154 compounds were classified based on their already reported effects. Two contemporary medicines found in this extract were quantified. Here, we report both beneficial and harmful compounds in the methanol extract of CA. We propose that the harmful compounds can be removed to yield safe medicines from CA.


INTRODUCTION
Traditional medicine (TM) is the result of accumulated knowledge and practices based on past experiences.As different TMs evolved 5000 years ago, they mainly depended on natural products.However, the TM system appears to have become stagnant over the past millennium with regard to new theories and practices.TM lacks a suitable means to review its principles and practices.Cochrane reviews of available studies reveal that clinical trials in TM have poor controls and lack statistical power or comparisons (Nahin and Straus, 2001;Narahari et al., 2010;Telles et al., 2014).
Conversely, traditional herbal medicines are claimed to be effective and safe.This assertion has existed throughout history and is a common claim by millions who have different medical options.About one-fifth of the US population uses some form of TM (Barnes et al., 2004), more than 40% of adults with neuropsychiatric symptoms in the US used complementary or alternative medicine in the year 2007 (Purohit et al., 2012), and 35.3% of persons diagnosed with cancer in the US took aid of complementary health approaches (Clarke, 2018).
Importantly, plants appear to have molecules that may be helpful in certain clinical applications for which contemporary medicine has no solution, such as neurodegenerative disorders.Though it evolved before the concepts and tools of modern pharmacology, TM remains empirical, and the molecules devised by it can be used therapeutically without significant manipulation.Rather than blaming TM, these molecules should be engineered for efficacy and to avoid unwanted interactions.There were early attempts to identify molecules from the extracts of many medicinal plants in different contexts.Some studies identified plant extract molecules in the serum because only the absorbed molecules matter as medicines (Sun et al., 2012).
Few studies have reported the compounds that are present in specific fractions (Ding et al., 2014;Sun et al., 2016).Numerous interaction studies on herbal formulations with conventional drugs are available (Chen et al., 2012).Knowledge about herbal drugs has grown in recent years, but the recommendation of their use is not forthcoming.They are acceptable only if there is no contemporary treatment available and no herb-drug interaction is known (Mörike and Gleiter, 2014).The number, concentrations, and identities of all compounds consumed as herbal medicines are not known.
These preparations along with many beneficial compounds may contain compounds such as the phytotoxins, excess heavy metals, apoptotic inducers, and psychoactive and drug interacting chemicals.Rani et al. 113 Available evidence has shown that several herbal products that have been put to extensive use in traditional medications for generations may possess carcinogenic, hepatotoxic, cardiotoxic and other harmful activities (Gromek et al., 2015).Thus, such harmful compounds should be removed to make traditional herbal medicines safer and more acceptable.It is imperative that we identify all of the compounds present in an extract.Only then can we know which useful, harmful or conditionally useful compounds are present in traditional medicines.The fingerprinting of extracts is helpful for molecular medicine to better understand the active components the plants contain and for TM to avoid unwanted reactions (Miyata, 2007;Yuan et al., 2016).
Centella asiatica, a member of Apiaceae family, is native to the tropical countries of India, China, Malaysia, Sri Lanka, Indonesia, South Africa, and Madagascar (Orhan, 2012;Sabaragamuwa et al., 2018).This perennial, herbaceous creeper has small fan-shaped green leaves, white flowers and bears small oval fruits.The medicinal qualities of this plant have been utilized in Ayurvedic and Chinese traditional medicines for centuries (Meulenbeld and Wujastyk, 2001;Yuan et al., 2016).The whole plant of CA can be used for medicinal purposes.
CA is a popular medicinal herb that is widely utilized for its therapeutic properties in a number of traditional medicinal systems.However, individual compounds responsible for different therapeutic physiological effects or compounds that might have adverse effects are not known.Since it is a widely consumed plant, there is an urgent need for in-depth analysis of compounds present in this plant rather than only focusing on the gross effects of the herb.Often prescribed in many traditional medicine systems, the whole plant extract of CA was used to identify all the compounds present in this plant during the present study.
Methanolic extraction was employed to get high percentage yield of extracts from the plant (Dhawan and Gupta, 2017) and extract maximum bioactive compounds such as alkaloids, steroids, flavinoids, saponins, and tannins from CA.
Here, we report the identification of a large number of compounds in the methanol extract of CA (Gotu Kola) using mass spectrometry and database confirmation, an effective way of performing large scale untargeted plant screening.We understand that this knowledge is critically needed to realize the medicinal value of a plant by identifying the presence of (1) the active therapeutic molecules (efficiency) and ( 2) the harmful molecules that should be removed from the preparation.

The extract
CA was obtained from the Mother India Nursery, Najafgarh, New Delhi and identified by Dr. Vandana Mishra, Department of Environmental Studies, University of Delhi.The plant was submitted to the herbarium in the same department with the herbarium voucher DUH 14337.The plants were collected from the same place within a period of 15 days to avoid the metabolite variability due to seasons, geography and environmental causes.Surface disinfection was performed by brushing the fresh plants with a soft brush under running tap water before processing.The whole plant was cut into small pieces and dried in the laboratory.The methanol extract of the dried plant material was prepared by incubating it with 100% methanol for 24 h at room temperature with mechanical stirring.The extract was centrifuged at 10,000 g for 10 min at room temperature.The supernatant was collected.The methanol from the supernatant was removed in a centrifugal evaporator.The dried sample was weighed and stored at -20°C in an airtight vial

Identification of compounds
The dried methanolic extract was re-suspended in 0.1% formic acid, centrifuged and injected into a Shimadzu Prominence SIL HTC system equipped with a C18 column, 1.8 µm, 2.1×100 mm (ACQUITY UPLC HSS T3 Column).A gradient elution using water and acetonitrile containing 0.1% formic acid was used at a flow rate of 0.3 ml/min for 20 min.The injection volume was 5 µl.An AB SCIEX TripleTOF ® 5600 LC-MS/MS system with a DuoSpray TM Source and electrospray ionization (ESI) probe was used for data acquisition.TOF MS scan mode acquisition with simultaneous information dependent acquisition MS/MS was performed.Each sample was analyzed in positive polarity.Data were acquired over a mass range of 100 to 1100 m/z with IDA MS/MS performed with a collision energy of 35 eV and a spread of ±20 eV.From the MS/MS data collected, the compounds corresponding to 100 to 1100 m/z were searched.Only the monoisotopic molecule ions were considered.The generated peak list with a peak above the quality of 30 was searched in METLIN database using its m/z ratio.All the possible compounds for a given monoisotopic mass were screened for known biological activity using Pubmed and Google.The data of shortlisted compounds was processed using PeakView® Software, AB Sciex for features identification and ID confirmation.The structure elucidation of the compound was confirmed by matching experimental fragments of the m/z to theoretical fragments of the structure.The molfile file of the possible structure was uploaded in PeakView software which then generated the fragmentation pattern for that structure.The fragments generated from molfile were matched with the experimental fragments.The percentage of matching was generated from the software.Only the compounds having 100% match were considered in the present analysis.

Quantification of compounds
In order to substantiate that the compounds identified by software are actually present in the extract, two compounds, methotrexate and cytarabine were selected at random for quantification.The stock solutions of methotrexate, cytarabine, and homatropine (internal standard) were separately prepared at a concentration of 1 mg/ml in acetonitrile.The stock solutions were appropriately diluted with 75% acetonitrile containing 0.1% formic acid to reach the required lower working concentrations.
The calibration curves of methotrexate and cytarabine were plotted with concentrations ranging from 3.9 to 62.5 ng/ml.A working internal standard solution was used at a concentration of 250 ng/ml in 75% acetonitrile.A ZIC HILIC 4.6×50 mm column was used for analytical separation with an isocratic mobile phase consisting of 10 mM ammonium formate and acetonitrile containing 0.1% formic acid at a ratio of 25:75 with a flow rate of 0.5 ml/min.The autosampler tray and the column were kept at 22°C.20 µl of the sample was injected into the UPLC with a run time of 5 min.
Chromquest software version 4.1 was used to control all UPLC parameters.The tandem mass spectrometric detection of analyte and internal standard (IS) was performed on a 4000QTrap (AB SCIEX) equipped with a Turbo Ion Spray (ESI) source operating in positive ion mode.Data acquisition and integration were performed by Analyst 1.5.2 software (AB SCIEX).The source dependent and compound dependent parameters were optimized in positive ion mode using the inbuilt algorithm of Analyst 1.5.2 software to yield the maximum intensity for precise detection.

Identification of compounds
The non-redundant number of compounds containing both MS and MS2 spectra was found to be 3,201 in positive polarity.About half of the compounds (46.8%) had less than 1 ppm mass error when compared to the METLIN database.More than three-fourths of the compounds (77.4%) were within 5 ppm mass error.The rest of the compounds had a mass error between 6 and 10 ppm.The methanolic extract of CA was separated on UPLC coupled to mass spectrometer.The duplicate runs are shown in Figure 1.
Molecular identification of m/z 244.09 is as shown in Figure 2. The molfile of each compound was then used to generate the theoretical fragments and was compared with the experimental MS/MS spectrum for identification

Categorization of compounds
Among these 1,187 compounds identified in the extract, 154 compounds have been studied in humans or human cell lines and have 100% matches with the theoretical and observed MS/MS spectra (fragments).These compounds had varying effects on living systems.They had properties that were inflammatory, anti-oxidative, anti-bacterial, anti-nociceptive, anti-viral, anti-fungal, antiapoptotic, anti-parasitic, anti-insecticidal, neuroprotective, anti-tumor, anti-diabetic, etc.They also had volatile compound, including insect attractants and repellants, some of which were quorum sensing compounds for plants (Supplementary Information, Tables S1, S2, S3, S4, S5, and S6).These 154 compounds were categorized based on their effect on normal physiology.The present report is based on the analysis of these compounds.Table 1 lists 37 compounds that are listed as drugs in the Drugbank database (DrugBank Version 4.3) (Wishart et al., 2006).17 toxins were detected in the extract (Table 2).In addition to these compounds, 11 environmental pollutants were identified (Table 3).

Quantification of compounds
The quantification of methotrexate and cytarabine was performed in multiple reaction monitoring (MRM) mode based on molecular adduct ion and its fragment ion with the transitions of 244.ng/mg of methanol extract, respectively.

The toxins
Seventeen toxins (Table 2) have been identified in the  CA extract.Five of these toxins were of plant origin, nine toxins were from fungi and three toxins were from marine origin.The plant origin toxins could be inherent but fungi and marine origin toxins seem to be acquired.Plant origin toxins ipomeamaronol, huratoxin, daphnin, gitoxin, and miserotoxin have been reported in diverse plants in the literature.It is known that a large number of endophytic fungi and bacteria reside in CA (Rakotoniriana et al., 2008).45 different taxa of fungi and 31 endophytic bacteria were isolated from healthy leaves of CA.Endophytes alter the composition of secondary metabolites in their plant hosts (Gao et al., 2015;Zhang et al., 2013).Climatic stress, wounding, metals, bacterial infections, fungal infections, and environmental chemicals also control and contribute to plant metabolomics (Ramakrishna and Ravishankar, 2011;Ncube et al., 2013;Nasim and Dhir, 2010).Plant secondary metabolites have microbicidal, insecticidal or herbicidal functions.It is surprising that the toxins oscillatoxin A (blue-green algae), 1-desulfoyessotoxin (dinoflagellate) and onchidal (onchidioideans) are present in CA considering that these three species are of marine origin.These three species are also known to thrive in freshwater environments.The early land-dwelling plants were effective in the uptake of DNA (broken by the harsh conditions), which might have helped in the acquisition of many self-defense functions (Cove, 2015).Notwith-standing the apical meristem, plants do acquire foreign entities through various mechanisms.These toxin molecules

Environmental pollutants
Eleven environmental pollutants were identified in the present study of CA extract (Table 3).The soil of medicinal plants is also a major cause for concern since the toxicity of traditional medicinal plants is associated with environmental pollutants.The unacceptable level of heavy metals in traditional medicine is well known.Plants take up nutrients, fertilizers and other chemicals from their environment.Environmental pollutants such as synthetic pesticides, rodenticide, etc., can be taken up.Some of the pollutants detected are organo-phosphorous and c-organochlorine compounds.Long-term exposure of experimental animals to organo-phosphorous compounds resulted in glucose intolerance with hyperinsulinemia, a hallmark of insulin resistance (Nagaraju et al., 2015).Some of the pollutants are acetylcholine esterase inhibitors and carcinogens.A previous study reported 198 types of pesticides in 120 types of traditional Chinese medicine (Wang et al., 2014).Furthermore, the types of pesticides were not the same in different parts of the plant.Typically, any acute effect (toxicity of pesticides, the benefit of medicines, etc.) has been largely measured at the acute median lethal dose or concentration.In addition to direct mortality induced by pesticides, their sub-lethal effects on physiology and behavior must be considered for a complete analysis of their impact.To appreciate the potential effects of a pest or pollutant, ecologically relevant experimental approaches that take into account the sub-lethal exposure over the long-term should be developed.The results of one such study indicated that insecticide mixtures continue to affect natural systems over many weeks, despite no traces of the mixture being detectable in the environment.Direct and indirect consequences across many levels could be observed (Hasenbein et al., 2016).The data of the present study revealed that care should be exercised while disposing biological and industrial waste.

Quantification of compounds
Cytarabine and methotrexate are prescribed for cancer treatment.As the beginning of therapy, a low dose of cytarabine produced maximal anti-leukemic effects a evidenced by all response endpoints.This suggests saturation in the dose-response relationship at this dose level.High-dose cytarabine results in excessive toxic effects without the therapeutic benefit (NTR 230).When used for induction, methotrexate in doses of 3.3 mg/m 2 in combination with 60 mg/m 2 of prednisone, given daily, produced remission produced remission in 50% of treated patients, usually within a period of 4 to 6 weeks.These concentrations are at least ten times greater than the concentrations of cytarabine and methotrexate estimated in the CA extract.The two drugs were detected at concentrations of 0.4 µg and 0.06 µg/mg of methanol extract.A normal prescribed dose of up to 1 to 10 g/kg/day of CA extract is recommended for enhanced cognition over a 52 week period (Manyam et al., 2004;Kumari et al., 2016).
Over this span, 2.0 and 0.3 mg of these drugs will be consumed, respectively.Methotrexate is at the top of the list of high-risk drugs in a hospital setting causing prolonged hospitalization (Saedder et al., 2014).Methotrexate and cytarabine can exert their effects at low doses (Hocaoglu et al., 2008).Another set of compounds that could be actively harmful are hormones.There are plant and insect related hormones in CA.Four compounds, megestrol acetate, norethindrone acetate, conivaptan and ∆4-tibolone, are hormonal drugs used in humans.There are also compounds that stimulate hormones.In TM, CA extract is taken over a long period.Effectively, one is consuming over twenty cancer drugs at sub-lethal doses over an extended period.If the intention is neuroprotection or anti-oxidation, the consumption of cytotoxic compounds is unnecessary.
CA has been actively used as a medicinal herb in Ayurveda for various ailments from ancient times.For many years, CA extracts have been considered acutely and chronically non-toxic, even at doses as high as 5000 mg/kg (Chivapat et al., 2011;Chauhan and Singh, 2012;Deshpande et al., 2015).A clinical case study has reported the development of hepatotoxicity in three women after consuming CA tablets for 30, 20 and 60 days in order to lose weight (Jorge and Jorge, 2005).In the same study, it was found that not only did the discontinuation of the tablets help cure the symptoms such as jaundice, hepatitis, hepatomegaly, choluria, etc., but the commencement of the tablets again reinstated the same symptoms.Another study reported a 15-year girl to develop hepatotoxicity after ingesting lymecycline and a herbal medicine with active ingredient CA (20 mg/day) for 6 weeks to treating acne (Dantuluri et al., 2011).
The study concluded that the deranged liver function and coagulation profile were in fact caused by herbal medicine containing CA. Methotrexate and Cytarabine are well-known hepatotoxicity inducers (Sotoudehmanesh et al., 2010;Thatishetty et al., 2013).Consequently, chemotherapic drugs methotrexate and cytarabine might play a role in hepatotoxicity caused by CA ingestion in different cases.Ingestion of crude plant extract can cause many such interactions.

Library of secondary metabolites
Since the pharmacological activities of 154 compounds have been reported in human or human cell lines, we have the name, chemical formulae, therapeutic activity in humans as well as the MS/MS data for these 154 plant products (many given under different categories).This list should be combined with compounds from aqueous and other extracts of CA in the future.This compilation should be continued to include all medicinal plants.A detailed database of secondary metabolites from all medicinal plants will be of great pharmaceutical use.To complete the knowledge, (1) all the compounds are to be studied in vitro/ex vivo/in vivo and (2) the peaks with MS and MS/MS data (including compounds with less than 100% matching) should be identified de novo as well as their effects should be studied.Natural products are the most proven source of novel drug candidates.An integrated approach involving virtual screening, automated high content assays and high impact technologies for fractionation as well as assay development is urgently needed in drug discovery research (Cremin and Zeng, 2002;Koehn, 2008).The availability of purified natural products and experimental high-throughput screening can synergize to yield new therapies.

Conclusions
TM should make sure of the identity of the compounds present in the extract.This will identify the harmful compounds.The presence of the active molecules indicates the efficacy of the extract.TM should undergo pharmacological tests to gain acceptance.Equally important issues are quality assurance and Good Manufacturing Practices (GMP).TM can assure the safety of plant compounds after enzyme inhibitors, toxins, environmental pollutants, and cytotoxic substances are removed.The approach used in this study is affordable, efficient, and safe.

Figure 2 ,
Panel A shows chromatogram of CA methanolic extract while Panel B shows the compound eluted at 1.43 min and had m/z value of 244.09.The identification of the compound was done based on the fragment matching of the theoretical and experimental MS/MS pattern.The m/z of the peak was uploaded into METLIN software and a list of all possible compounds was generated from METLIN.

Figure 1 .
Figure 1.Chromatogram of the Centella asaitica methanolic extract.Two different runs are indicated in red and blue.
3 112.1 for cytarabine (Figure 3, Panel A) and m/z 455.2 308.3 for methotrexate (Figure 3, Panel B).The transition of m/z 276.1142.2 was used for the internal standard (homatropine).All of the calibration curves showed good linearity (r > 0.999) within the test ranges.The precision was evaluated by intra and inter day tests, which revealed relative standard deviation (RSD) values of less than 3.88%.The recoveries for the quantified compounds were between 96.3 and 103.7%, with RSD values below 2.89%.Methotrexate and cytarabine were detected at concentrations of 60 and 400

Figure 2 .
Figure 2. Structural Elucidation of m/z 244.09 in PeakView™ Software.Panel (A) shows the elution pattern, panel (B) shows the m/z and retention time, panel (C) indicates the fragmentation pattern, panel (D) shows the identity, and panel (E) shows peak matching of both experimental and theoretical fragments for the compound shown, cytarabine.

Figure 3 .
Figure 3. Quantification of cytarabine and methotrexate using LC-ESI-MS/MS.(A) representative chromatogram of cytarabine m/z 244.3 → 112.1 at 2.67 min and internal standard homatropine m/z 276.1 → 142.2 at 1.83 min, respectively, in the plant extract.(B) Representative chromatogram of methotrexate m/z 455.2→ 308.3 at 2.06 min and the internal standard homatropine at m/z 276.1 → 142 at 1.83 min in the plant extract.

Table 1 .
Contemporary medicines identified in the Centella asiatica extract.

Table 3 .
Environmental pollutants seen in Centella asiatica.

Table S1 .
Anti-cancer compounds from CA extract:

Table S2 .
Antioxidant compounds from CA extract.

Table S3 .
Cytotoxic compounds from CA extract.

Table S4 .
Apoptotic compounds from CA extracts.

Table S5 .
Anti-Inflammatory compounds from CA.

Table S6 .
Agonist /Antagonist of receptors from CA extract.