Comparative morpho-anatomical standardization and chemical profiling of root drugs for distinction of fourteen species of family Apocynaceae

Background The root drugs of the family Apocynaceae are medicinally important and used in Indian Systems of Medicine (ISM). There is often a problem of misidentification and adulteration of genuine samples with other samples in the market trade. Keeping in view the adulteration problem of raw drug material, comparative macroscopic and microscopic (qualitative and quantitative) characterisation and chemical analysis (TLC and LC–MS profiling) of a total of 14 economically important root drugs of family Apocynaceae were done for practical and rapid identification. A total of 33 qualitative botanical characteristics of root samples were subjected to Principal Component Analysis (PCA) and Cluster analysis to identify taxonomically significant characteristics in the distinction of root drug samples at the species level. Results Comparative qualitative and quantitative data on morphological, macroscopic, and microscopic characters were generated for the studied 14 species. Despite the similarity in some root characters, a combined study involving the surface, anatomical, and powder features helped distinguish root samples at the species level. The relative relationship between selected species was represented as clustering or grouping in the dendrogram. PCA analysis determined significant characters leading to species grouping and identification. Results showed that clustering of xylem vessels in cross-section, pore size, and distribution in the cut root, the shape of starch grains, the thickness of cork zone were among the most notable characters in species distinction. Chemical profiling revealed unique fingerprints and content of chemical compounds, which were significant in identification of root drug samples. Conclusions The comparative botanical standards and chemical profiles developed in the present study can be used as future reference standards for the quick, easy, and correct identification of root drug samples to be used in the herbal drug industry. Further, the identified significant microscopic characters have the potential for taxonomic studies in species delimitation. Supplementary Information The online version contains supplementary material available at 10.1186/s40529-022-00342-z.

2050 (Nirmal et al. 2013). India is known as the second-largest exporter of medicinal plants after China (Dhanabalan 2011). Around 960 medicinal plants are traded in India, of which 178 are known with high trade value with annual consumption of more than 100 metric tonnes (NMPB 2010). The Apocynaceae, belonging to the order Gentianales, also known as 'Dogbane family' or 'Toxic plant's family, ' is considered one of the largest and economically most important angiosperms family. It comprises about 5100 species belonging to 366 genera in five subfamilies, plants are generally trees, shrubs, and vines distributed mainly in tropical and subtropical regions, with several genera widely occurring in various regions of India (Endress and Bruyns 2000;Lens et al. 2009;Nazar et al. 2013;Endress et al. 2014;eFI 2020). Plants of the family Apocynaceae are characterized by latex and are rich in several metabolites, such as alkaloids, triterpenoids, flavonoids, steroids, phenols, lactones, and glycosides (Hofling et al. 2010;Bhadane et al. 2018). Plants of this family possess many pharmacological properties (Endress 1997;Yarnell and Abascal 2002;Bhadane et al. 2018). Roots of several Apocynaceae species are widely used in Indian Systems of Medicine (ISM) such as Ayurveda, Siddha, and Yunani systems (Khare 2007;Devi et al. 2017;Jeewandara et al. 2017). Only a few selected species are cultivated commercially, and most of the traded raw plant material is collected from wild sources. Due to widespread use in Indian traditional medicines and the similar appearance of plants, several species of Apocynaceae are often prone to adulteration (Devi et al. 2017).
There are identification problems with raw herbal root drugs of Apocynaceae due to similar or confusing names, similar physical appearance, lack of an organized plant collection and procurement chain. The use of the wrong species for medicinal purposes can be harmful to end-users. Correct identification and authentication of herbal drug samples are essential to ensure traditional medicines' efficacy, purity, and quality (Sahoo et al. 2010). Botanical and chromatographic fingerprint, reference standards are helpful to identify and to determine the purity and quality of herbal drugs (Zafar et al. 2010;Folashade et al. 2012;Upton et al. 2020). For identification and ensuring consistent quality of plant raw materials and botanicals of herbal products, several chemical identification methods (qualitative and quantitative) are accepted.
Reference standards are helpful in the correct identification and distinction of different root drug samples. Microscopic methods are known to be taxonomically significant for the identification of fragmented herbal samples and are used for sample identification in the various traditional pharmacopeia, for taxonomic characterization and systematic studies in many plants (Kraemer 1920;Metcalfe and Chalk 1957;Carlsward et al. 1997;Scatena et al. 2005;Aldasoro et al. 2005;Matias et al. 2007;Figueroa et al. 2008;Zarrei et al. 2010;Ginko et al. 2016). Botanical identification by macroscopic and microscopic studies of herbal plants is known to be simple and easy (Apraj et al. 2011). In chemical based identification, Thin Layer Chromatography (TLC) remains the simplest, efficient, with low cost and rapid tool to check and identify known markers compounds in plant extract (Pascual et al. 2002). TLC is a common chromatographic technique for separating non-volatile substances and quite valuable for assessing quality of herbal remedies (Yuen and Lau-Cam 1985). Michael Tsweet was the first to introduce the separation and identification of plant constituents using chromatography (Ettre and Sakodynskii 1993). Now a day, chromatographic and spectroscopic techniques are used for the quantitative estimation and quality control of herbal drugs (Balekundri and Mannur 2020).
Considering the medicinal and trade value and the authentication problem of raw drug material, root drug samples of 14 species of family Apocynaceae were selected for the present study. The selected plants are used in different Ayurvedic formulations, reported with high estimated annual trade value, and often have adulteration problems. The present study aims to develop a detailed comparative morphological, macroscopic, and microscopic standard (qualitative and quantitative) along with chemical profiling for practical and rapid identification of the highly traded fourteen root drugs of the family Apocynaceae.

Plant material
For the present study, authentic dry raw root drug samples (RDS) of the fourteen species of the family Apocynaceae available at Crude Drug Repository (CDR) were selected. CDR is a national referral facility (a sub-section of Janaki Ammal Herbarium (RRLH) at CSIR-IIIM, Jammu, which is an internationally recognised Herbarium), having a collection of > 4200 authentic raw plant drug specimens collected from different parts of India. Accepted botanical names and synonyms of the selected species were verified from theplantlist.org (TPL 2013).

Botanical studies
Surface characters of root drug samples (such as color, texture, appearance, nature, etc.) and transverse cut root surface characters (such as surface appearance, color, thickness, and nature of various zones) were analyzed by hand lens and by stereomicroscope (Leica S9i).
For the anatomical study, dry raw root drug samples (RDS) were kept in FAA fixative; Formalin Few (1)/ Abundant (2) 33 Cork cells Few (1)/ Abundant (2)  MT Metric Tonnes, NA not available stain (5-10 min), decolorized in 70% alcohol (5 min), stained in fast green (2-3 min), decolorized in 70% alcohol (5 min), dehydrated in 90% alcohol and absolute alcohol (each for 2 min) and finally cleared in xylene for 1-2 min. Xylene cleared sections were carefully mounted in Canada balsam and then observed under the compound light microscope. In powder study, crushed dried root drug samples were characterized for organoleptic characters (color, odor, taste, and texture) and microscopic characters (cell types and cell contents). An iodine test was performed in root powder and T.S. of the root to study the shape and size of starch grains. Microscopic characters were observed using a compound microscope (LEICA DM 750) with an associated camera (LEICA ICC50E). Histological measurements were also done for various tissue zones, cells, and cell contents using Leica software (LEICA LAS V 4.9.0 software).

Statistical analysis
The botanical data were subjected to variance analysis, Principal Component Analysis (PCA), and Cluster analysis. Variance analysis of selected quantitative characters was done using descriptive statistics such as mean and standard deviation by Tukey's post hoc test using Minitab 17 (Minitab, LLC, State College, PA, USA). Among various studied root botanical characters, a total of 33 qualitative macroscopic and microscopic characters (Table 1) appearing in more than one state were selected, coded in binary (20 characters), and multistate (13 characters) numerical values for creating a data matrix (Additional file 1: Table S1). Selected root characters, their types, and codes for identifying the studied RDS are given in Table 1. Botanical traits were subjected to PCA and Cluster analysis with Paleontological Statistics Software (PAST) (Version 3.26) to study species grouping and determine the taxonomically significant characters for species grouping (Hammer et al. 2001). Cluster analysis was done by Ward's hierarchical clustering method based on Euclidean metric distances. Results of PCA are presented as two-dimensional scatter plots representing species and character states.

Chemical identification Extraction procedure for chromatographic fingerprinting
The root samples were air dried at temperature 25 °C ± 2 °C and relative humidity of 65% ± 5%. The dried material was powdered using pestle and mortar. The 10.0 gm of dried powdered material was socked in methanol, kept under sonication for 2 h, and kept overnight. A similar extraction procedure was repeated 24 h with the same solvent until a clear and colorless solvent was obtained. The combined extract was then filtered through Whatman filter paper (No.2) and dried under a vacuum evaporator at 40 °C. Dried extract was stored at 0 °C in an airtight container until used.

Thin Layer Chromatography (TLC) fingerprint
In present study, TLC profiles were developed for the root samples of the selected species. For the TLC fingerprint, methanolic extract of samples was used. The sample (2 gm) was dissolved in 10 ml methanol with continuous stirring at room temperature for 24 h. The extract was filtered through Whatman filter paper No. 2. Subsequently, the extract was diluted (1 ml extract) in 25 ml of methanol and was later used for TLC fingerprinting. The root extract was spotted with a capillary tube onto a silica-gel TLC plate with F 254 fluorescent indicator and developed in suitable solvent polarity for resolution (Factor 1991). The developed plates were then stained in anisaldehyde reagent and heated at 105 °C for 5 min. The movement of the active compound was expressed and recorded by the retention factor value (R f ).
For development of TLC, methanolic crude extract (10 µl) of root samples (100 mg in 10 ml) was applied on to a silica-gel TLC plate with F254 fluorescent indicator. Prior to chromatography, the chamber was saturated with mobile phase for 15 min. The loaded plate was placed in a developing chamber with a mobile phase until the mobile phase rose to 7 cm in height. The TLC was developed in varied solvent combinations for each plant ( Table 6). The developed plate was air-dried to remove the solvent from the container, stained with anisaldehyde reagent, heated at 105 °C for 5 min and then examined at white light for the varied band patterns.

Adsorbent
Chromatographic Silica gel F 254 mixture with an average particle size of 5 µm.

Application volume
10 µl each of the sample solution as 7-mm bands.
Relative humidity: Condition the plate to a relative humidity of about 33% using a suitable device.
Developing distance: 7 cm. Derivatization reagent: Anisaldehyde reagent: add 20 ml of acetic acid and 10 ml of sulphuric acid to 170 ml of cold methanol and mix well. After cooling to room temperature, add 1 ml of anisaldehyde to the mixture.

LC-MS analysis
The chemicals used for the LC-MS analysis were MSgrade acetonitrile, water, acetic acid, and formic acid; all were purchased from Merck, Germany. Other solvents and chemicals used for the extraction were of analytical grade and procured from Merk, Germany.
The sample for LC-MS analysis was prepared in a volumetric flask in methanol-water (1:1, v/v). The crude extract was filtered through a 0.25 μm disposal membrane filter (Millipore) and made appropriate dilutions using methanol. The stock and working solution were stored at + 4 °C. An Agilent 1260 liquid chromatography system (Agilent, USA) equipped with a quaternary solvent delivery system, an autosampler, and a column heater was used. The chromatographic separation was performed on Merck Chromolith fast gradient RP18 e column (100 mm × 4.6 mm) protected by a Chromolith guard column. The mobile phase consisted of A (0.1% aq. formic acid: 1.0% ACN, v/v/v) and B (Acetonitrile). A gradient elution was performed with mobile phase started with B-0%; 4.0 min B-20%; 15 min B 50%; 20 min B-50%; 25 min B 70%; 35 min B-70% 38 min B-85%; 42 min B-85%; 45 min B-0% and at 47 min B-0%. The flow rate was monitored at 0.5 mL/min. The injection volume was 1 μL, and the column temperature was maintained at 30 0 C. A 6410B triple quad LC/MS system from Agilent was used to detect a hybrid triple quadrupole mass spectrometer equipped with Turbo V sources. The analyses were performed using electrospray ionization (ESI) sources in positive and negative modes. The operation conditions were as follows: scan range of 110-1300 amu, V charging 4000 V, ion source temperature 300 0 C, nebulizer 50psi, gas flow 13L/min, capillary voltage 4000, and a step size of 0.1 amu. Nitrogen was used in all cases. Agilent Mass Hunter software (version B.04.00) was used for data acquisition and processing.

Results
The scientific literature on taxonomic, medicinal, and commercial aspects was searched from various sources such as scientific journals, edited books, floras, scientific databases, eFloras, online databases, etc. Raw root drug samples of selected species in the present study are essential ingredients in different Ayurvedic formulations, reported with much high annual trade value, and are among widely traded RPD's from India ( Table 2). The literature review revealed that several closely related species have similar names. Similarity and confusion in local or trade names of many species are often reported with adulteration problems. For example, C. procera and C. gigantea have the same ayurvedic name, i.e., Alarka. Similarly, the roots of three selected plants, viz., C. dubia, H. indicus, and I. frutescens are known as "Sariva" in Sanskrit. Due to the similar common name, the official part of true 'Sariva' (H. indicus) is known to be adulterated by the other two plants of the same common name.

Botanical characterisation of root samples
In the comparative morphological study, sample appearance, surface, and cut root appearance were studied. Comparative morphological characteristics of the studied RDS are shown in Figs. 1, 2. Root drug samples of studied species appeared similar in physical appearance and morphological features, while some surface and cut root features were characteristic. RDS of most species appeared elongated or cylindrical, less branched, twisted, or bent, but A. curassavica was observed with secondary and tertiary fibrous branches. The root surface of most drug samples was rough with wrinkles (in C. procera, C. dubia, M. tenacissima, T. divaricata), cracks (in C. gigantea, C. carandas, C. dubia, H. pubescens, M. tenacissima, N. oleander), some with a powdery mass on scraping (in C. gigantea, C. procera, C. spinarum, R. serpentina, T. divaricata), and sloughed off bark (in C. carandas, C. dubia, H. indicus), and nodule like protuberances (in C. spinarum). Some species were with smooth root surfaces (A. curassavica, C. roseus, M. tenacissima). Root surface was of variable color such as greenish (in A. curassavica), light cream (in C. gigantea, R. serpentina), buff-colored (in C. procera, M. tenacissima, T. divaricata), dark brown (in C. dubia, H. indicus, H. pubescens, I. frutescens), dark brown with light patches (in C. carandas), brown (in N. indicum), light brown (in C. spinarum) to light green (in C. roseus). Cut root surface was circular in most species while irregular outline in C. procera, H. indicus, I. frutescens, R. serpentina and circular to oval in C. dubia. The bark and woody region showed variability in thickness ( Fig. 1, 2, Fig. 3). Woody region showed variation in pore size and characteristic pattern of pores. Pores in most species were of varying size, having well-distinguished large pores, while some had very small pores (in A. curassavica, C. roseus, H. pubescens, R. serpentina) ( Table 3).
Xylem vessel arrangement varied from solitary (in C. carandas, C. spinarum), grouped (in C. gigantea, C.  Powder microscopic study of most species showed cork cell fragments, parenchyma cell fragments, sclereid fragments, coloured fragments, prismatic crystals, rosette crystals, starch grains, xylem vessel fragments. However, variability was observed in cells and cell contents such as starch grains and crystals. Starch grains of most species were solitary to compound (3-4 units), some up to 2 units (T. divaricata), and some up to 9 units (M. tenacissima). The shape of starch grains varied from spherical (in C. dubia), oval to spherical (C. gigantea, C. procera, C. carandas, C. spinarum, C. roseus, I. frutescens, M. tenacissima, T. divaricata), oval to elongated (H. indicus, H. pubescens, N. oleander), to more than one shape (A. curassavica and R. serpentina). Among studied species, prismatic crystals were present in all root samples except M. tenacissima. Apart from these, rosette crystals were also observed in some species (A. curassavica, C. procera, H. pubescens, M. tenacissima, N. oleander, and T. divaricata). The size of starch grains and prismatic crystals are provided in Table 4.
In the current study, some characters were shared in studied species, while some features were also observed as characteristics useful in species distinction. Statistical analysis of studied botanical characters by the mean-variance analysis (Tables 4, 5), PCA, and Cluster analysis was observed to resolve the complexity in species distinction and identification of significant characters. The cluster analysis results are represented in a dendrogram, which shows closely related species' grouping ( Fig. 4). A Scatter plot diagram of PC1 versus PC2 showed significant characters with taxonomic value in the grouping and distinction of various species (Fig. 5). PCA analysis showed that the first three components accounted for nearly 64% of the total variance (30.49%, 16.96%, and 16.58%, respectively). According to the first three PCA's, the following characters including clustering of xylem vessel, the shape of starch grains, pores in woody part (size, arrangement, and distribution in the cut root),            One-way ANOVA's were carried out separately for each quantitative character to figure out the differences among different species. The same letters after values in a column denote a lack of statistically significant differences, according to Tukey's post hoc test (p < 0.05) the width of medullary rays, the thickness of dominating tissues in T.S. (cork, cortex and xylem zones), cork lignification and cork colour in the cut root, type of crystals, starch distribution test in T.S. of the root, surface characters of raw drug sample such as texture, wrinkles, fissures, etc., were observed as major contributors in the whole variation. Also, mean-variance analysis of quantitative features revealed the mean thickness of tissue zones in cross-section, vessel

TLC chromatogram profile
TLC was employed for the preliminary phytochemical investigation of the crude extracts of root samples under study. For development of TLC profile of root Table 6 Table showing Table 6. The TLC of root samples of studied 14 species were observed with prominent bands with different retention factor (Rf) values ( Table 6). The migration profiles of constituents of the root samples are shown in Fig. 6.

LC-MS profile
The dataset generated by Liquid Chromatography mass spectrometry measurement of raw plant materials can be used for authentication of plant species. In present study, LC-MS analysis compared the phytochemical contents of the methanol extracts of root samples. Some plant metabolites were identified for each of the 14 species in a single analytical run (Table 7), which helped in species identification with high accuracy. Although it was difficult to identify each peak in the LC-MS chromatogram, some major constituents were identified for studied species. The characteristic compounds from crude extracts of the given plants have been identified by the LC-MS technique. In LC-MS studies of root extract, the major compounds were identified based on its mass data and UV pattern. The chemical constituents for 14 different species at specific retention time are given in Figs. 7,8,9,10,11,12,13,14,15,16,17,18,19,20.

Discussion
Morphological features (shape, size, color, surface feature, texture, fracture, and appearance) and anatomical features are considered of diagnostic value in the identification and distinction of herbal drug samples in several plant groups (Fritz and Saukel 2011;Manohan et al. 2013;Ginko et al. 2016;Park et al. 2019). Surface characters may not be used for species authentication; however, the combination of some surface and cut root surface characters can be used in the preliminary distinction of samples. According to Park et al. (2019), the only morphological character-based distinction of root drug samples is challenging and also requires anatomical characterisation. In the anatomical study, principal dominating tissue in cross-section and other tissue zone was considered a suitable character for distinguishing herbal drug samples (Fritz and Saukel 2011;Hassan et al. 2015;Ginko et al. 2016). In the present anatomical study, the relative thickness, number & arrangement of cell layers of outer bark (cork region), inner bark (cortex and phloem), and woody zone (xylem) to the total radius of the studied T.S. were observed varying for studied species (Fig. 3). Some bark anatomical features such as the structure of cork, number, and thickness of cork layers, the occurrence of sclereids, type, structure, and arrangement of secretory ducts, presence of crystals were also known of taxonomic value in species characterization (Fritz and Saukel 2011;Ginko et al. 2016;Park et al. 2019). In the present study, the cork zone was variable in colour and lignifications of cell walls, cortex zone varied in the cell composition, cell contents, mean lumen size of secretory canals, and occurrence of sclereids. The cortex of some species was observed with characteristic anatomical  Kumar et al. Botanical Studies (2022) (Figs. 1, 2). A comparative morphometric study in the present study revealed variation in mean lumen diameter of secretory canals in both Carissa spp. (comparatively broader in C. spinarum). In the present study, anatomical characters of root bark of M. tenacissima corresponded with anatomical structure observed by Tripathi et al. (2014). Anatomical characters of the Carissa genus corresponded with anatomical characters in some previous studies (Salunke and Ghate 2013;Khalil et al. 2015;Allam et al. 2016). Some vascular anatomical characters such types, arrangement, and grouping of xylem vessels, vessel outline, the dimension of largest vessels, frequency of vessels per square area; the appearance of medullary rays in secondary xylem, the thickness of medullary rays, and laticifers in rays, etc. were reported as significant characters in discrimination of different root samples (Lens et al. 2008;Fritz and Saukel 2011;Ginko et al. 2016;Park et al. 2019). In the present study, xylem anatomical characters such as mean lumen diameter of xylem vessels, number of vessels per square area, xylem vessel arrangement, and distribution were observed as the variable for various species under study. The mean lumen diameter of xylem vessels in the studied species ranged from 30.95 µm (C. roseus) to 134.54 µm (C. gigantea). The mean number of vessels per square area ranged from 22.8 (C. gigantea) to 481.5 (C. roseus). Quantitative xylem anatomical characters for other species are shown in Table 5.
Powder microscopy helps identify broken or powdered plant samples (Sereena and Sreeja 2014). Several microscopic features, including starch grains and crystals types, are considered helpful in identifying some herbal material (Cortella and Pochettino 1994;Lens et al. 2008 Ginko et al. 2016;Ya'ni et al. 2018). The organoleptic examination of root powder samples showed variation in color, odor, texture, and taste in the present study.
Organoleptic and microscopic characters of powder are provided in Table 4 and Additional file 1: Fig. S1-S2. Species belonging to the same genus were observed with comparatively more similarities in powder characters. Organoleptic and microscopic powder characters of two Carissa spp. and Calotropis spp. were nearly identical. In powder study, variation was observed in quantitative and quantitative microscopic features such as the shape and size of starch grains, the grouping of starch grains; the type of crystals; the size of prismatic crystals, the abundance of coloured fragments. An iodine test in T.S. of the root also showed variation in abundance and distribution of starch grains in a different zone of T.S. of the root (Additional file 1: Fig. S1-S2).
The powder sample of two Calotropis spp. showed variation only in taste, mean size of starch grains and colorful crystals, and abundance of starch grains in medullary rays. Some other genera (such as C. dubia, H. indicus, and I. frutescens, all three with the common name 'Sariva') were similar in some organoleptic and microscopic powder characters. However, the color of powder samples and the shape of starch grains also varied (circular in C. dubia, oval to slightly elongated in H. indicus, and oval to circular in I. frutescens).
Among the studied species, some previous studies had been done on some species. Out of the 14 species studied, macro and microscopic identification studies for 11 species were conducted earlier by various researchers. Root anatomical studies were performed earlier in A. curassavica (Hassan et al. 1952;Kalidass et al. 2009a;Ramesh et al. 2014); C. gigantea (Shirsat et al. 2011 two genera with the similar common name (Jeewandara et al. 2017). Other species belonging to the same genus (Calotropis spp. and Carissa spp.) were observed in a close clade. Such closely grouped species can be distinguished based on some unique combination of botanical characters identified in the present study.
In the present study, some surface and anatomical characters such as the appearance of bark and the presence or absence of pith, etc. observed as less stable and should be carefully considered for the identification of herbal samples. For example, root bark was observed sloughed off in dried root samples of C. dubia. Similarly, Jeewandara et al. (2017) observed pith in older roots of C. dubia; however, in the present study, pith was not observed. The physical integrity of raw herbal samples is considered essential as identifying herbal drugs only from powdered samples can be challenging (Ginko et al. 2016). In addition, a single botanical character may not be considered unique in describing a species. For plant species with similar botanical features, a combination of diagnostic microscopic characters is essential for species identification and distinction of herbal samples (Lens et al. 2008(Lens et al. , 2009Ginko et al. 2016). Detailed taxonomic information provided in the present study can be helpful in taxonomic identification and distinction of genuine raw herbal drugs from contaminants to be used for herbal drug preparations. Chemical profiling of herbal samples in addition to botanical characterization is helpful and is more authentic in the identification of raw herbal drugs.
In the present study, the TLC fingerprinting profile was done for methanolic extracts of root samples of selected 14 species of family Apocynaceae. The R f values acquired from TLC chromatograms provided essential information regarding their polarity of phytochemicals as well as important clues in the separation process. The usage of multiple solvent systems for TLC investigations could be essential for selecting the suitable solvent system since different R f values of the molecule reflect a notion about their polarity. This knowledge will aid in the selection of a suitable solvent system for subsequent compound separation from these plant extracts. However, the TLC results were not sufficient to determine their profile and the chemical complexity of the crude extracts. Thus to identify phytoconstituents in root extracts, Liquid Chromatography-Mass Spectrometry (LC-MS) studies were also carried out in present work. LC-MS analysis is now a routine technique employed to identify phytoconstituents present in a wide range of botanical samples (Zhao et al. 2005;Lai et al. 2015;Park et al. 2019). Park et al. (2019) performed LC-MS profiling alongwith anatomical studies to develop identification standards of roots of Adenophora sp. In the present study, the chemical compounds identified were major metabolites present in 14 species and were comparable with literature reports (Table 7). In addition, the NMR data is also obtained for the identified compounds which are comparable to published reports (data provided in supplementary file as ' Additional file 1'). These compounds provide supportive data can be used as the chemical markers for the identification of raw herbal drugs in addition to botanical data.
While modern testing techniques for evaluating plant drugs are available in today's scientific age, microscopic analysis remains one of the most basic and cost-effective methods for correctly identifying source materials (Kumar et al. 2011). Anatomical studies are helpful in the distinction of herbal samples with similar morphological characters (Traiperm et al. 2017). The combined approach involving botanical and chemical identifcaiton adopted in the present study ensures more authenticity in sample identification irrespective the physical form of herbal sample. The identification standards thus help overcome the adulteration and misidentification problems.

Conclusions
Detailed comparative botanical characterisation (qualitative and quantitative features) of root drug samples was found helpful in identifying and distinguishing similar-looking adulterant samples. Statistical analysis of botanical characters helped in identification of some of taxonomically significant characters in distinction of root samples. Among various characters, the clustering of xylem vessels was observed as the most significant character in Apocynaceae species' distinction from PCA values. The unique chromatographic fingerprint profiles and major chemical constituents identified for studied species further aid in distinction of root samples of closely related species. The combined study including botanical and chemical characterization in the present study provide a reference database for future identification of raw root samples. The studies performed in present study will help the herbal industry in quality control of raw herbal drugs and the botanical characters further help as a reference guide for future taxonomic studies of herbal drugs.